Programming Language Extended

Rexx

Dallas Draft
November 1998

Caution: The developers of this standard have requested that holders of patents that may be required for
the implementation of the standard disclose such patents to the publisher. However, neither the
developers nor the publisher have undertaken a patent search to identify which, if any, patents may apply
to this standard.
As of the date of publication of this standard and following calls for identification of patents that may be
required for the implementation of the standard, no such claims have been made. No further patent
search is conducted by the developer or publisher in respect of any standard it processes. No
representation is made or implied that licenses are not required to avoid infringement in the use of this
standard.
Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Purpose . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
0 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1 Scope, purpose, and application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.1 Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.2 Purpose . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.3 Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.4 Recommendation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2 Normative references . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3 Definitions and document notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.1 Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.2 Document notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3.2.1 Rexx Code . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3.2.2 Italics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
4 Conformance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
4.1 Conformance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
4.2 Limits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
5 Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
5.1 Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
5.1.1 Notation for completion response and
conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
5.2 Processing initiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
5.2.1 API_Start . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
5.3 Source programs and character sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
5.3.1 Syntactic_characters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
5.3.2 Extra_letters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
5.3.3 Other_blank_characters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
5.3.4 Other_negators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
5.3.5 Other_characters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
5.4 Configuration characters and encoding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
5.4.1 Config_SourceChar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
5.4.2 Config_OtherBlankCharacters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
5.4.3 Config_Upper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
5.4.4 Config_Lower . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
5.4.5 Config_Compare . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
5.4.6 Config_B2C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
5.4.7 Config_C2B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
5.4.8 Config_Substr . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
5.4.9 Config_Length . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
5.4.10 Config_Xrange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
5.5 Objects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
5.5.1 Config_ObjectNew . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
5.5.2 Config_Array_Size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
5.5.3 Config_Array_Put . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
5.5.4 Config_Array_At . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
5.5.5 Config_Array_Hasindex . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
5.5.6 Config_Array_Remove . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
5.6 Commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
5.6.1 Config_Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
5.7 External routines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
5.7.1 Config_ExternalRoutine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
5.7.2 Config_ExternalMethod . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
5.8 External data queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
5.8.1 Config_Push . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
5.8.2 Config_Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
5.8.3 Config_Pull . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
5.8.4 Config_Queued . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
5.9 Streams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
5.9.1 Config_Stream_Charin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
5.9.2 Config_Stream_Position . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
5.9.3 Config_Stream_Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
5.9.4 Config_Stream_State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
5.9.5 Config_Stream_Charout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
5.9.6 Config_Stream_Qualified . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
5.9.7 Config_Stream_Unique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
5.9.8 Config_Stream_Query . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
5.9.9 Config_Stream_Close . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
5.9.10 Config_Stream_Count . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
5.10 External variable pools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
5.10.1 Config_Get . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
5.10.2 Config_Set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
5.11 Configuration characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
5.11.1 Config_Constants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
5.12 Configuration routines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
5.12.1 Config_Trace_Query . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
5.12.2 Config_Trace_Input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
5.12.3 Config_Trace_Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
5.12.4 Config_Default_Input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
5.12.5 Config_Default_Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
5.12.6 Config_Initialization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
5.12.7 Config_Termination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
5.12.8 Config_Halt_Query . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
5.12.9 Config_Halt_Reset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
5.12.10 Config_NoSource . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
5.12.11 Config_Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
5.12.12 Config_Random_Seed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
5.12.13 Config_Random_Next . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
5.12.14 Config_Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
5.13 Traps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
5.14 Variable pool . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
5.14.1 API_Set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
5.14.2 API_Value . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
5.14.3 API_Drop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
5.14.4 API_SetDirect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
5.14.5 API_ValueDirect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
5.14.6 API_DropDirect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
5.14.7 API_ValueOther . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
5.14.8 API_Next . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
5.14.9 API_NextVariable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
6 Syntax constructs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
6.1 Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
6.1.1 Backus-Naur Form (BNF) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
6.1.2 Operands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
6.1.3 Operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
6.1.4 Grouping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
6.1.5 BNF syntax definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
6.1.6 Syntactic errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
6.2 Lexical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
6.2.1 Lexical elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
6.2.1.1 Events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
6.2.1.2 Actions and tokens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
6.2.1.3 Source characters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
6.2.1.4 Rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
6.2.2 Lexical level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
6.2.3 Interaction between levels of syntax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
6.2.3.1 Reserved symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
6.2.3.2 Function name syntax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
6.3 Syntax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
6.3.1 Syntax elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
6.3.2 Syntax level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
6.1 Syntactic information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
6.1.1 VAR_SYMBOL matching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
6.1.2 Trace-only labels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
6.1.3 Clauses and line numbers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
6.1.4 Nested IF instructions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
6.1.5 Choice of messages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
6.1.6 Creation of messages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
6.1.6.1 Error message prefix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
6.2 Replacement of insertions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
6.3 Syntactic equivalence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
7 Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
7.1 Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
7.1.1 Var_Empty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
7.1.2 Var_Set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
7.1.3 Var_Value . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
7.1.4 Var_Drop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
7.1.5 Var_Expose . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
7.1.6 Var_Reset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
7.2 Symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
7.3 Value of a variable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
7.3.1 Derived names . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
7.3.2 Value of a reserved symbol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
7.4 Expressions and operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
7.4.1 The value of a term . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
7.4.2 The value of a prefix_expression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
7.4.3 The value of a power_expression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
7.4.4 The value of a multiplication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
7.4.5 The value of an addition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
7.4.6 The value of a concatenation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
7.4.7 The value of a comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
7.4.8 The value of an and_expression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
7.4.9 The value of an expression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
7.4.10 Arithmetic operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
7.5 Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
7.5.1 Invocation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
7.5.2 Evaluation of arguments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
7.5.3 The value of a label . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
7.5.4 The value of a function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
7.5.5 The value of a method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
7.5.6 The value of a message term . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
7.1.1 Use of Config_ExternalRoutine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
8 Directives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
8.1 Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
8.2 Initializing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
8.2.1 Program initialization and message texts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
8.3 REQUIRES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
8.4 CLASS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
8.5 METHOD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
8.6 ROUTINE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
9 Instructions .......................................................... 69
9.1 Method initialization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
9.2 Routine initialization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
9.3 Clause initialization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
9.4 Clause termination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
9.5 Instruction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
9.5.1 ADDRESS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
9.5.2 ARG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
9.5.3 Assignment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
9.5.4 CALL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
9.5.5 Command to the configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
9.5.6 DO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
9.5.7 DO loop tracing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
9.5.8 DROP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
9.5.9 EXIT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
9.5.10 EXPOSE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
9.5.11 FORWARD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
9.5.12 GUARD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
9.5.13 IF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
9.5.14 INTERPRET . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
9.5.15 ITERATE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
9.5.16 Execution of labels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
9.5.17 LEAVE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
9.5.18 Message term . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
9.5.19 LOOP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
9.5.20 NOP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
9.5.21 NUMERIC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
9.5.21.1 NUMERIC DIGITS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
9.5.21.2 NUMERIC FORM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
9.5.21.3 NUMERIC FUZZ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
9.5.22 OPTIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
9.5.23 PARSE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
9.5.24 PROCEDURE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
9.5.25 PULL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
9.5.26 PUSH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
9.5.27 QUEUE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
9.5.28 RAISE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
9.5.29 REPLY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
9.5.30 RETURN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
9.5.31 SAY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
9.5.32 SELECT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
9.5.33 SIGNAL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
9.5.34 TRACE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
9.5.35 Trace output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
9.5.36 USE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
9.6 Conditions and Messages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
9.6.1 Raising of conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
9.6.2 Messages during execution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
10 Built-in functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
10.1 Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
10.2 Routines used by built-in functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
10.2.1 Argument checking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
10.2.2 Date calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
10.2.1 Radix conversion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
10.2.2 Raising the SYNTAX condition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
10.1 Character built-in functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
10.1.1 ABBREV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
10.1.2 CENTER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
10.1.3 CENTRE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
10.1.4 CHANGESTR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
10.1.5 COMPARE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
10.1.6 COPIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
10.1.7 COUNTSTR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
10.1.8 DATATYPE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
10.1.9 DELSTR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
10.1.10 DELWORD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
10.1.11 INSERT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
10.1.12 LASTPOS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
10.1.13 LEFT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
10.1.14 LENGTH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
10.1.15 OVERLAY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
10.1.16 POS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
10.1.17 REVERSE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
10.1.18 RIGHT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
10.1.19 SPACE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
10.1.20 STRIP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
10.1.21 SUBSTR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
10.1.22 SUBWORD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
10.1.23 TRANSLATE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
10.1.24 VERIFY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
10.1.25 WORD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
10.1.26 WORDINDEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
10.1.27 WORDLENGTH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
10.1.28 WORDPOS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
10.1.29 WORDS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
10.1.30 XRANGE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
10.2 Arithmetic built-in functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
10.2.1 ABS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
10.2.2 FORMAT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
10.2.3 MAX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
10.2.4 MIN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
10.2.5 SIGN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
10.2.6 TRUNC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
10.3 State built-in functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
10.3.1 ADDRESS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
10.3.2 ARG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
10.3.3 CONDITION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
10.3.4 DIGITS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
10.3.5 ERRORTEXT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
10.3.6 FORM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
10.3.7 FUZZ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
10.3.8 SOURCELINE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
10.3.9 TRACE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
10.4 Conversion built-in functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
10.4.1 B2X . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
10.4.2 BITAND . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
10.4.3 BITOR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
10.4.4 BITXOR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
10.4.5 C2D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
10.4.6 C2X . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
10.4.7 D2C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
10.4.8 D2X . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
10.4.9 X2B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
10.4.10 X2C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
10.4.11 X2D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
10.5 Input/Output built-in functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
10.5.1 CHARIN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
10.5.2 CHAROUT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
10.5.3 CHARS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
10.5.4 LINEIN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
10.5.5 LINEOUT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
10.5.6 LINES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
10.5.7 QUALIFY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
10.5.8 STREAM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
10.6 Other built-in functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
10.6.1 DATE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
10.6.1 QUEUED . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
10.6.2 RANDOM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
10.6.3 SYMBOL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
10.6.4 TIME . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
10.6.1 VALUE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
10.6.1 QUEUED . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
10.6.2 RANDOM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
10.6.3 SYMBOL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
10.6.4 TIME . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
10.6.5 VALUE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
11 Built-in classes ................................................... 132
11.1 Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
11.2 Object, class and method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
11.2.1 The object class . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
11.2.2 The class class . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
11.2.3 The method class . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
11.3 The string class . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
11.3.1 The array class . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
11.4 The supplier class . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
11.5 The message class . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
12 Provided classes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
12.1 Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
12.2 The Collection Classes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
12.2.1 Collection Class Routines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
12.2.2 The collection class . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
12.2.2.1 INIT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
12.2.2.2 EXPOSED . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
12.2.2.3 FINDINDEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
12.2.2.4 AT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
12.2.2.5 [] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
12.2.2.6 PUT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
12.2.2.7 []= . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
12.2.2.8 HASINDEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
12.2.2.9 ITEMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
12.2.2.10 REMOVE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
12.2.2.11 REMOVEIT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
12.2.2.12 MAKEARRYA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
12.2.2.13 MAKEARRAYX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
12.2.2.14 SUPPLIER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
12.2.3 Class list . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
12.2.3.1 PUT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
12.2.3.2 OF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
12.2.3.3 INSERT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
12.2.3.4 FIRST . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
12.2.3.5 LAST . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
12.2.3.6 FIRSTITEM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
12.2.3.7 LASTITEM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
12.2.3.8 NEXT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
12.2.3.9 PREVIOUS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
12.2.3.10 SECTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
12.2.4 Class queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
12.2.4.1 PUSH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
12.2.4.2 PULL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
12.2.4.3 QUEUE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
12.2.4.4 PEEK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
12.2.4.5 REMOVE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
12.2.5 Class table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
12.2.5.1 MAKEARRAY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
12.2.5.2 UNION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
12.2.5.3 INTERSECTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
12.2.5.4 XOR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
12.2.5.5 DIFFERENCE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
12.2.5.6 SUBSET . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
12.2.6 Class set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
12.2.6.1 PUT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
12.2.6.2 OF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
12.2.6.3 UNION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
12.2.6.4 INTERSECTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
12.2.6.5 XOR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
12.2.6.6 DIFFERENCE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
12.2.7 Class relation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
12.2.7.1 PUT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
 1521.577...O3 0 10
12.2.9.14 UNKNOWN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
12.3 The stem class . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
12.1 The stream class . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
12.2 The alarm class . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
12.3 The monitor class . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
12.3.1 INIT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
12.3.2 CURRENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
12.3.3 DESTINATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
12.3.4 UNKNOWN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
Annex A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
Rationale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
Incompatibilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
Call . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
Concurrency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
Guard . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
To be processed: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
Annex B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
Method of definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
Conformance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
Notation for completion response and
conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .181
Source programs and character sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .181
Lexical level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
Syntax level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
Data Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .181
Evaluation (Definitions written as code) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
Annex C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
Foreword
Purpose
This standard provides an unambiguous definition of the programming language Rexx. Its purpose is to
facilitate portability of Rexx programs for use on a wide variety of computer systems.
History
The computer programming language Rexx was designed by Mike Cowlishaw to satisfy the following
principal aims:
- to provide a highly readable command programming language for the benefit of programmers and
program readers, users and maintainers;
- to incorporate within this language program design features such as natural data typing and control
structures which would contribute to rapid, efficient and accurate program development;
- to define a language whose implementations could be both reliable and efficient on a wide variety of
computing platforms.

In November, 1990, X3 announced the formation of a new technical committee, X3J18, to develop an
American National Standard for Rexx. This standard was published as ANSI X3.274-1996.

The popularity of "Object Oriented" programming, and the need for Rexx to work with objects created in
various ways, led to Rexx extensions and to a second X3J18 project which produced this standard. (Ed -
hopefully)

Committee lists
(Here)

This standard was prepared by the Technical Development Committee for Rexx, X3J18.
There are annexes in this standard; they are informative and are not considered part of this standard.

Suggestions for improvement of this standard will be welcome. They should be sent to the
Information Technology Industry Council, 1250 Eye Street, NW, Washington DC 20005-3922.
This standard was processed and approved for submittal to ANSI by the Accredited Standards
Committee on Information Processing Systems, NCITS. Committee approval of this standard does not
necessarily imply that all committee members voted for its approval. At the time it approved this standard,
the NCITS Committee had the following members:

To be inserted
The people who contributed to Technical Committee J18 on Rexx, which developed this standard,
include:
0 Introduction
This standard provides an unambiguous definition 2 Normative references
of the programming language Rexx. There are no standards which constitute provisions
1 Scope, purpose, and application of this American National Standard.
1.1 Scope 3 Definitions and document notation
This standard specifies the semantics and syntax Lots more for NetRexx
of the programming language Rexx by specifying 3.1 Definitions
requirements for a conforming language processor. 3.1.1 application programming interface:
The scope of this standard includes A set of functions which allow access to some
- the syntax and constraints of the Rexx Rexx facilities from non-Rexx programs.
language; 3.1.2 arguments:
- the semantic rules for interpreting Rexx The expressions (separated by commas) between
programs; the parentheses of a function call or following the
- the restrictions and limitations that a name on a CALL instruction. Also the
conforming language processor may impose; corresponding values which may be accessed by a
- the semantics of configuration interfaces. function or routine, however invoked.
This standard does not specify 3.1.3 built-in function:
- the mechanism by which Rexx programs are A function (which may be called as a subroutine)
transformed for use by a data-processing that is defined in section nnn of this standard and
system; can be used directly from a program.
- the mechanism by which Rexx programs are 3.1.4 character string:
invoked for use by a data-processing system; A sequence of zero or more characters.
- the mechanism by which input data are 3.1.5 clause:
transformed for use by a Rexx program; A section of the program, ended by a semicolon.
- the mechanism by which output data are The semicolon may be implied by the end of a line
transformed after being produced by a Rexx or by some other constructs.
program; 3.1.6 coded:
- the encoding of Rexx programs; A coded string is a string which is not necessarily
- the encoding of data to be processed by Rexx comprised of characters. Coded strings can occur
programs; as arguments to a program, results of external
- the encoding of output produced by Rexx routines and commands, and the results of some
programs; built-in functions, such as D2C.
- the size or complexity of a program and its 3.1.7 command:
data that will exceed the capacity of any specific A clause consisting of just an expression is an
data-processing system or the capacity of a instruction known as a command. The expression
particular language processor; is evaluated and the result is passed as a
- all minimal requirements of a data-processing command string to some external environment.
system that is capable of supporting a 3.1.8 condition:
conforming language processor; A specific event, or state, which can be trapped by
- the syntax of the configuration interfaces. CALL ON or SIGNAL ON.
1.2 Purpose 3.1.9 configuration:
The purpose of this standard is to facilitate Any data-processing system, operating system and
portability of Rexx programs for use on a wide software used to operate a language processor.
variety of configurations. 3.1.10 conforming language processor:
1.3 Application A language processor which obeys all the
This standard is applicable to Rexx language provisions of this standard.
processors. 3.1.11 construct:
1.4 Recommendation A named syntax grouping, for example
It is recommended that before detailed reading of "expression", "do_specification".
this standard, a reader should first be familiar with 3.1.12 default error stream:
the Rexx language, for example through reading An output stream, determined by the configuration,
one of the books about Rexx. It is also on which error messages are written.
recommended that the annexes should be read in 3.1.13 default input stream:
conjunction with this standard. An input stream having a name which is the null
string. The use of this stream may be implied.

14
3.1.14 default output stream: to the queue whenever REXX relinquishes control
An output stream having a name which is the null to some other program.
string. The use of this stream may be implied. 3.1.26 external routine:
3.1.15 direct symbol: A function or subroutine that is neither built-in nor
A symbol which, without any modification, names a in the same program as the CALL instruction or
variable in a variable pool. function call that invokes it.
3.1.16 directive: 3.1.27 external variable pool:
Clauses which begin with two colons are A named variable pool supplied by the
directives. Directives are not executable, they configuration which can be accessed by the
indicate the structure of the program. Directives VALUE built-in function.
may also be written with the two colons implied. 3.1.28 function:
3.1.17 dropped: Some processing which can be invoked by name
A symbol which is in an unitialized state, as and will produce a result. This term is used for
opposed to having had a value assigned to it, is both Rexx functions (see nnn) and functions
described as dropped. The names in a variable provided by the configuration (see n).
pool have an attribute of 'dropped' or 'not-dropped'. 3.1.29 identifier:
The name of a construct.
3.1.18 encoding: 3.1.30 implicit variable:
The relation between a character string and a A tailed variable which is in a variable pool solely
corresponding number. The encoding of character as a result of an operation on its stem. The names
strings is determined by the configuration. in a variable pool have an attribute of 'implicit' or
3.1.19 end-of-line: 'not-implicit'.
An event that occurs during the scanning of a 3.1.31 instruction:
source program. Normally the end-of-lines will One or more clauses that describe some course of
relate to the lines shown if the configuration lists action to be taken by the language processor.
the program. They may, or may not, correspond to 3.1.32 internal routine:
characters in the source program. A function or subroutine that is in the same
3.1.20 environment: program as the CALL instruction or function call
The context in which a command may be that invokes it.
executed. This is comprised of the environment 3.1.33 keyword:
name, details of the resource that will provide input This standard specifies special meaning for some
to the command, and details of the resources that tokens which consist of letters and have particular
will receive output of the command. spellings, when used in particular contexts. Such
3.1.21 environment name: tokens, in these contexts, are keywords.
The name of an external procedure or process 3.1.34 label:
that can execute commands. Commands are sent A clause that consists of a single symbol or a literal
to the current named environment, initially selected followed by a colon.
externally but then alterable by using the 3.1.35 language processor:
ADDRESS instruction. Compiler, translator or interpreter working in
3.1.22 error number: combination with a configuration.
A number which identifies a particular situation 3.1.36 notation function:
which has occurred during processing. The A function with the sole purpose of providing a
message prose associated with such a number is notation for describing semantics, within this
defined by this standard. standard. No Rexx program can invoke a notation
3.1.23 exposed: function.
Normally, a symbol refers to a variable in the most 3.1.37 null clause:
recently established variable pool. When this is not A clause which has no tokens.
the case the variable is referred to as an exposed 3.1.38 null string:
variable. A character string with no characters, that is, a
3.1.24 expression: string of length zero.
The most general of the constructs which can be 3.1.39 production:
evaluated to produce a single string value. The definition of a construct, in Backus-Naur form.
3.1.25 external data queue: 3.1.40 return code:
A queue of strings that is external to REXX A string that conveys some information about the
programs in that other programs may have access command that has been executed. Return codes
usually indicate the success or failure of the

15
command but can also be used to represent other A collection of the names of variables and their
information. associated values.
3.1.41 routine:
Some processing which can be invoked by name. 3.2 Document notation
3.1.42 state variable: 3.2.1 Rexx Code
A component of the state of progress in processing Some Rexx code is used in this standard. This
a program, described in this standard by a named code shall be assumed to have its private set of
variable. No Rexx program can directly access a variables. Variables used in this code are not
state variable. directly accessible by the program to be
3.1.43 stem: processed. Comments in the code are not part of
If a symbol naming a variable contains a period the provisions of this standard.
which is not the first character, the part of the 3.2.2 Italics
symbol up to and including the first period is the Throughout this standard, except in Rexx code,
stem. references to the constructs defined in section nnn
3.1.44 stream: are italicized.
Named streams are used as the sources of input 4 Conformance
and the targets of output. The total semantics of 4.1 Conformance
such a stream are not defined in this standard and A conforming language processor shall not
will depend on the configuration. A stream may be implement any variation of this standard except
a permanent file in the configuration or may be where this standard permits. Such permitted
something else, for example the input from a variations shall be implemented in the manner
keyboard. prescribed by this standard and noted in the
3.1.45 string: documentation accompanying the processor.
For many operations the unit of data is a string. It A conforming processor shall include in its
may, or may not, be comprised of a sequence of accompanying documentation
characters which can be accessed individually. - a list of all definitions or values for the
3.1.46 subcode: features in this standard which are specified to
The decimal part of an error number. be dependent on the configuration.
3.1.47 subroutine: - a statement of conformity, giving the complete
An internal, built-in, or external routine that may or reference of this standard (ANSI X3.274-1996)
may not return a result string and is invoked by the with which conformity is claimed.
CALL instruction. If it returns a result string the 4.2 Limits
subroutine can also be invoked by a function call, Aside from the items listed here (and the assumed
in which case it is being called as a function. limitation in resources of the configuration), a
3.1.48 symbol: conforming language processor shall not put
A sequence of characters used as a name, see numerical limits on the content of a program.
nnn. Symbols are used to name variables, Where a limit expresses the limit on a number of
functions, etc. digits, it shall be a multiple of three. Other limits
3.1.49 tailed name: shall be one of the numbers one, five or twenty
The names in a variable pool have an attribute of five, or any of these multiplied by some power of
'tailed' or 'non-tailed'. Otherwise identical names ten.
are distinct if their attributes differ. Tailed names Limitations that conforming language processors
are normally the result of replacements in the tail of may impose are:
a symbol, the part that follows a stem. - NUMERIC DIGITS values shall be supported
3.1.50 token: up to a value of at least nine hundred and
The unit of low-level syntax from which high-level ninety nine.
constructs are built. Tokens are literal strings, - Exponents shall be supported. The limit of
symbols, operators, or special characters. the absolute value of an exponent shall be at
3.1.51 trace: least as large as the largest number that can be
A description of some or all of the clauses of a expressed without an exponent in nine digits.
program, produced as each is executed. - String lengths shall be supported. The limit
3.1.52 trap: on the length shall be at least as large as the
A function provided by the user which replaces or largest number that can be expressed without
augments some normal function of the language an exponent in nine digits.- String literal length
processor. shall be supported up to at least two hundred
3.1.53 variable pool: and fifty.

16
 - Symbol length shall be supported up to at
least two hundred and fifty.

17
5 Configuration
Any implementation of this standard will be functioning within a configuration. In practice, the boundary
between what is implemented especially to support Rexx and what is provided by the system will vary
from system to system. This clause describes what they shall together do to provide the configuration for
the Rexx language processing which is described in this standard.
We don't want to add undue "magic" to this section. It seems we will need the concept of a "reference" (equivalent to
a machine address) so that this section can at least have composite objects as arguments. (As it already does but
these are not Rexx objects)
Possibly we could unify "reference" with "variable pool number" since object one-to-one with its variable pool is a fair
model. That way we don't need a new primitive for comparison of two references.

JAVA is only a "reference" for NetRexx so some generalized JAVA-like support is needed for that. It would provide
the answers to what classes were in the context, what their method signatures were etc.

5.1 Notation
The interface to the configuration is described in terms of functions. The notation for describing the
interface functionally uses the name given to the function, followed by any arguments. This does not
constrain how a specific implementation provides the function, nor does it imply that the order of
arguments is significant for a specific implementation.
The names of the functions are used throughout this standard; the names used for the arguments are
used only in this clause and nnn.
The name of a function refers to its usage. A function whose name starts with
- Config_ is used only from the language processor when processing programs;
- API_ is part of the application programming interface and is accessible from programs which are not
written in the Rexx language;
- Trap_ is not provided by the language processor but may be invoked by the language processor.
As its result, each function shall return a completion Response. This is a string indicating how the function
behaved. The completion response may be the character 'N' indicating the normal behavior occurred;
otherwise the first character is an indicator of a different behavior and the remainder shall be suitable as a
human-readable description of the function's behavior.
This standard defines any additional results from Config_ functions as made available to the language
processor in variables. This does not constrain how a particular implementation should return these
results.
5.1.1 Notation for completion response and conditions
As alternatives to the normal indicator 'N', each function may return a completion response with indicator
'X' or 'S'; other possible indicators are described for each function explicitly. The indicator 'X' means that
the function failed because resources were exhausted. The indicator 'S' shows that the configuration was
unable to perform the function.
Certain indicators cause conditions to be raised. The possible raising of these conditions is implicit in the
use of the function; it is not shown explicitly when the functions are used in this standard.
The implicit action is
call #Raise 'SYNTAX', Message, Description
where:
#Raise raises the condition, see nnn.
Message is determined by the indicator in the completion response. If the indicator is 'X' then
Message is 5.1. If the indicator is 'S' then Message is 48.1.
Description is the description in the completion response.
The 'SYNTAX' condition 5.1 can also be raised by any other activity of the language processor.
5.2 Processing initiation
The processing initiation interface consists of a function which the configuration shall provide to invoke
the language processor.
We could do REQUIRES in a macro-expansion way by adding an argument to Config_SourceChar to specify the
source file. However, I'm assuming we will prefer to recursively "run" each required file. One of the results of that will
be the classes and methods made public by that REQUIRES subject.
5.2.1 API_Start
Syntax:

18
API_Start(How, Source, Environment, Arguments, Streams, Traps, Provides)
where:
How is one of 'COMMAND', 'FUNCTION', or 'SUBROUTINE' and indicates how the program is
invoked.
What does OOI say for How when running REQUIREd files?
Source is an identification of the source of the program to be processed.
Environment is the initial value of the environment to be used in processing commands. This has
components for the name of the environment and how the input and output of commands is to be
directed.
Arguments is the initial argument list to be used in processing. This has components to specify the
number of arguments, which arguments are omitted, and the values of arguments that are not
omitted.
Streams has components for the default input stream to be used and the default output streams to
be used.
Traps is the list of traps to be used in processing (see nnn). This has components to specify
whether each trap is omitted or not.
Semantics:
This function starts the execution of a Rexx program.
If the program was terminated due to a RETURN or EXIT instruction without an expression the
completion response is 'N'.
If the program was terminated due to a RETURN or EXIT instruction with an expression the indicator
in the completion response is 'R' and the description of the completion response is the value of the
expression.
If the program was terminated due to an error the indicator in the completion response is 'E' and the
description in the completion response comprises information about the error that terminated
processing.
If How was 'REQUIRED' and the completion response was not 'E', the Provides argument is set to
reference classes made available. See nnn for the semantics of these classes.
5.3 Source programs and character sets
The configuration shall provide the ability to access source programs (see nnn).
Source programs consist of characters belonging to the following categories:
- syntactic_characters;
- extra_letters;
- other_blank_characters;
- other_negators;
- other_characters.
A character shall belong to only one category.
5.3.1 Syntactic_characters
The following characters represent the category of characters called syntactic_characters, identified by
their names. The glyphs used to represent them in this document are also shown. Syntactic_characters
shall be available in every configuration:
- & ampersand;
- ' apostrophe, single quotation mark, single quote;
- * asterisk, star;
- blank, space;
- A-Z capital letters A through Z;
- : colon;
- , comma;
- 0-9 digits zero through nine;
- = equal sign;
- ! exclamation point, exclamation mark;
- > greater-than sign;
- - hyphen, minus sign;
- < less-than sign;
- [ left bracket, left square bracket;
- ( left parenthesis;

19
 - % percent sign;
- . period, decimal point, full stop, dot;
- + plus sign;
- ? question mark;
- " quotation mark, double quote;
- \ reverse slant, reverse solidus, backslash;
- ] right bracket, right square bracket;
- ) right parenthesis;
- ; semicolon;
- / slant, solidus, slash;
- a-z small letters a through z;
- ~ tilde, twiddle;
- _ underline, low line, underscore;
- | vertical line, bar, vertical bar.
5.3.2 Extra_letters
A configuration may have a category of characters in source programs called extra_letters. Extra_letters
are determined by the configuration.
5.3.3 Other_blank_characters
A configuration may have a category of characters in source programs called other_blank_characters.
Other_blank_characters are determined by the configuration. Only the following characters represent
possible characters of this category:
- carriage return;
- form feed;
- horizontal tabulation;
- new line;
- vertical tabulation.
5.3.4 Other_negators
A configuration may have a category of characters in source programs called other_negators.
Other_negators are determined by the configuration. Only the following characters represent possible
characters of this category. The glyphs used to represent them in this document are also shown:
- ^ circumflex accent, caret;
- ¬ not sign.
5.3.5 Other_characters
A configuration may have a category of characters in source programs called other_characters.
Other_characters are determined by the configuration.
5.4 Configuration characters and encoding
The configuration characters and encoding interface consists of functions which the configuration shall
provide which are concerned with the encoding of characters.
The following functions shall be provided:
- Config_SourceChar;
- Config_OtherBlankCharacters;
- Config_Upper;
- Config_Compare;
- Config_B2C;
- Config_C2B;
- Config_Substr;
- Config_Length;
- Config_Xrange.
5.4.1 Config_SourceChar
Syntax:
Config_SourceChar()
Semantics:
Supply the characters of the source program in sequence, together with the EOL and EOS events.
The EOL event represents the end of a line. The EOS event represents the end of the source
program. The EOS event must only occur immediately after an EOL event. Either a character or an
event is supplied on each invocation, by setting #Outcome.

20
 If this function is unable to supply a character because the source program encoding is incorrect the
indicator of the completion response is 'E' and the description of the completion information is the
encoding which is incorrect, in hexadecimal notation.
5.4.2 Config_OtherBlankCharacters
Syntax:
Config_OtherBlankCharacters()
Semantics:
Get other_blank_characters (see nnn).
Set #Outcome to a string of zero or more distinct characters in arbitrary order. Each character is one
that the configuration considers equivalent to the character Blank for the purposes of parsing.
5.4.3 Config_Upper
Syntax:
Config_Upper(Character)
where:
Character is the character to be translated to uppercase.
Semantics:
Translate Character to uppercase. Set #Outcome to the translated character. Characters which have
been subject to this translation are referred to as being in uppercase. Config_Upper applied to a
character in uppercase must not change the character.
5.4.4 Config_Lower
Syntax:
Config_Lower(Character)
where:
Character is the character to be translated to lowercase.
Semantics:
Translate Character to lowercase. Set #Outcome to the translated character. Characters which have
been subject to this translation are referred to as being in lowercase. Config_Lower applied to a
character in lowercase must not change the character. Config_Upper of the outcome of
Config_Lower(Character) shall be the original character.
5.4.5 Config_Compare
Syntax:
Config_Compare(Character1, Character2)
where:
Character1 is the character to be compared with Character2.
Character2 is the character to be compared with Character1.
Semantics:
Compare two characters. Set #Outcome to
- 'equal' if Character1 is equal to Character2;
- 'greater' if Character1 is greater than Character2;
- 'lesser' if Character1 is less than Character2.
The function shall exhibit the following characteristics. If Config_Compare(a,b) produces
- 'equal' then Config_Compare(b,a) produces 'equal';
- 'greater' then Config_Compare(b,a) produces 'lesser';
- 'lesser' then Config_Compare(b,a) produces 'greater';
- 'equal' and Config_Compare(b,c) produces 'equal' then Config_Compare(a,c) produces 'equal';
- 'greater' and Config_Compare(b,c) produces 'greater' then Config_Compare(a,c) produces
'greater';
- 'lesser' and Config_Compare(b,c) produces 'lesser' then Config_Compare(a,c) produces 'lesser';
- 'equal' then Config_Compare(a,c) and Config_Compare(b,c) produce the same value.
Syntactic characters which are different characters shall not compare equal by Config_Compare, see
nnn.
5.4.6 Config_B2C
Syntax:
Config_B2C(Binary)
where:
Binary is a sequence of digits, each '0' or '1'. The number of digits shall be a multiple of eight.
Semantics:

21
 Translate Binary to a coded string. Set #Outcome to the resulting string. The string may, or may not,
correspond to a sequence of characters.
5.4.7 Config_C2B
Syntax:
Config_C2B(String)
where:
String is a string.
Semantics:
Translate String to a sequence of digits, each '0' or '1'. Set #Outcome to the result. This function is the
inverse of Config_B2C.
5.4.8 Config_Substr
Syntax:
Config_Substr(String, n)
where:
String is a string.
n is an integer identifying a position within String.
Semantics:
Copy the n-th character from String. The leftmost character is the first character. Set #Outcome to the
resulting character.
If this function is unable to supply a character because there is no n-th character in String the indicator
of the completion response is 'M'.
If this function is unable to supply a character because the encoding of String is incorrect the indicator
of the completion response is 'E' and the description of the completion information is the encoding
which is incorrect, in hexadecimal notation.
5.4.9 Config_Length
Syntax:
Config_Length(String)
where:
String is a string.
Semantics:
Set #Outcome to the length of the string, that is, the number of characters in the string.
If this function is unable to determine a length because the encoding of String is incorrect, the
indicator of the completion response is 'E' and the description of the completion information is the
encoding which is incorrect, in hexadecimal notation.
5.4.10 Config_Xrange
Syntax:
Config_Xrange(Character1, Character2)
where:
Character1 is the null string, or a single character.
Character2 is the null string, or a single character.
Semantics:
If Character1 is the null string then let LowBound be a lowest ranked character in the character set
according to the ranking order provided by Config_Compare; otherwise let LowBound be Character1.
If Character2 is the null string then let HighBound be a highest ranked character in the character set
according to the ranking order provided by Config_Compare; otherwise let HighBound be Character2If
#Outcome after Config_Compare(LowBound,HighBound) has a value of
- 'equal' then #Outcome is set to LowBound;
- 'lesser' then #Outcome is set to the sequence of characters between LowBound and HighBound
inclusively, in ranking order;
- 'greater' then #Outcome is set to the sequence of characters HighBound and larger, in ranking
order, followed by the sequence of characters LowBound and smaller, in ranking order.
5.5 Objects
The objects interface consists of functions which the configuration shall provide for creating objects.
5.5.1 Config_ObjectNew
Syntax:
Config_ObjectNew
Semantics:

22
 Set #Outcome to be a reference to an object. The object shall be suitable for use as a variable pool,
see nnn. This function shall never return a value in #Outcome which compares equal with the value
returned on another invokation of the function.
5.5.2 Config_Array_Size
Syntax:
Config_Array_Size(Object, size)
where:
Object is an object.
Size is an integer greater or equal to 0.
Semantics:
The configuration should prepare to deal efficiently with the object as an array with indexes having
values up to the value of size.
5.5.3 Config_Array_Put
Syntax:
Config_Array_Put(Array, Item, Index)
where:
Array is an array.
Item is an object
Index is an integer greater or equal to 1.
Semantics:
The configuration shall record that the array has Item associated with Index.
5.5.4 Config_Array_At
Syntax:
Config_Array_At(Array, Index)
where:
Array is an array.
Index is an integer greater or equal to 1.
Semantics:
The configuration shall return the item that the array has associated with Index.
5.5.5 Config_Array_Hasindex
Syntax:
Config_Array_At(Array, Index)
where:
Array is an array.
Index is an integer greater or equal to 1.
Semantics:
Return '1' if there is an item in Array associated with Index, '0' otherwise.
5.5.6 Config_Array_Remove
Syntax:
Config_Array_At(Array, Index)
where:
Array is an array.
Index is an integer greater or equal to 1.
Semantics:
After this operation, no item is associated with the Index in the Array.

5.6 Commands
The commands interface consists of a function which the configuration shall provide for strings to be
passed as commands to an environment.
See nnn and nnn for a description of language features that use commands.
5.6.1 Config_Command
Syntax:
Config_Command(Environment, Command)
where:
Environment is the environment to be addressed. It has components for:
- the name of the environment;

23
 - the name of a stream from which the command will read its input. The null string indicates use of
the default input stream;
- the name of a stream onto which the command will write its output. The null string indicates use
of the default output stream. There is an indication of whether writing is to APPEND or REPLACE;
- the name of a stream onto which the command will write its error output. The null string indicates
use of the default error output stream. There is an indication of whether writing is to APPEND or
REPLACE.
Command is the command to be executed.
Semantics:
Perform a command.
- set the indicator to 'E' or 'F' if the command ended with an ERROR condition, or a FAILURE
condition, respectively;
- set #RC to the return code string of the command.
5.7 External routines
The external routines interface consists of a function which the configuration shall provide to invoke
external routines.
See nnn and nnn for a description of the language features that use external routines.
5.7.1 Config_ExternalRoutine
Syntax:
Config_ExternalRoutine(How, NameType, Name, Environment, Arguments, Streams, Traps)
where:
How is one of 'FUNCTION' or 'SUBROUTINE' and indicates how the external routine is to be
invoked.
NameType is a specification of whether the name was provided as a symbol or as a string literal.
Name is the name of the routine to be invoked.
Environment is an environment value with the same components as on API_Start.
Arguments is a specification of the arguments to the routine, with the same components as on
API_Start.
Streams is a specification of the default streams, with the same components as on API_Start.
Traps is the list of traps to be used in processing, with the same components as on API_Start.
Semantics:
Invoke an external routine. Set #Outcome to the result of the external routine, or set the indicator of
the completion response to 'D' if the external routine did not provide a result.
If this function is unable to locate the routine the indicator of the completion response is 'U'. As a result
SYNTAX condition 43.1 is raised implicitly.
If How indicated that a result from the routine was required but the routine did not provide one the
indicator of the completion response is 'H'. As a result SYNTAX condition 44.1 is raised implicitly.
If How indicated that a result from the routine was required but the routine provided one that was too
long (see #Limit_String in nnn) the indicator of the completion response is 'L'. As a result SYNTAX
condition 52 is raised implicitly.
If the routine failed in a way not indicated by some other indicator the indicator of the completion
response is 'F'. As a result SYNTAX condition 40.1 is raised implicitly.
5.7.2 Config_ExternalMethod
OOI has external classes explicitly via the ::CLASS abc EXTERNAL mechanism. Analogy with classic would also
allow the subject of ::REQUIRES to be coded in non-Rexx. However ::REQUIRES subject is coded, we need to
gather in knowledge of its method names because of the search algorithm that determines which method is called.
Hence reasonable that the ultimate external call is to a method. Perhaps combine Config_ExternalRoutine with
Config_ExternalMethod.
There is a terminology clash on "environment". Perhaps easiest to change the classic to "address_environment".
(And make it part of new "environment"?)
There are terminology decisions to make about "files", "programs", and "packages". Possibly "program" is the thing
you run (and we don't say what it means physically), "file" is a unit of scope (ROUTINEs in current file before those in
REQUIREd), and "package" we don't use (since a software package from a shop would probably have several files
but not everything to run a program.) Using "file" this way may not be too bad since we used "stream" rather than
"file" in the classic definition.
The How parameter will need 'METHOD' as a value. Should API_Start also allow 'METHOD'. If we pass the new
Environment we don't have to pass Streams separately.

24
Text of Config_ExternalMethod waiting on such decisions.
Syntax:
Config_ExternalMethod(How, NameType, Name, Environment, Arguments, Streams, Traps)
where:
How is one of 'FUNCTION' or 'SUBROUTINE' and indicates how the external routine is to be
invoked.
NameType is a specification of whether the name was provided as a symbol or as a string literal.
Name is the name of the routine to be invoked.
Environment is an environment value with the same components as on API_Start.
Arguments is a specification of the arguments to the routine, with the same components as on
API_Start.
Streams is a specification of the default streams, with the same components as on API_Start.
Traps is the list of traps to be used in processing, with the same components as on API_Start.
Semantics:
Invoke an external routine. Set #Outcome to the result of the external routine, or set the indicator of
the completion response to 'D' if the external routine did not provide a result.
If this function is unable to locate the routine the indicator of the completion response is 'U'. As a result
SYNTAX condition 43.1 is raised implicitly.
If How indicated that a result from the routine was required but the routine did not provide one the
indicator of the completion response is 'H'. As a result SYNTAX condition 44.1 is raised implicitly.
If How indicated that a result from the routine was required but the routine provided one that was too
long (see #Limit_String in nnn) the indicator of the completion response is 'L'. As a result SYNTAX
condition 52 is raised implicitly.
If the routine failed in a way not indicated by some other indicator the indicator of the completion
response is 'F'. As a result SYNTAX condition 40.1 is raised implicitly.
5.8 External data queue
The external data queue interface consists of functions which the configuration shall provide to
manipulate an external data queue mechanism.
See nnn, nnn, nnn, nnn, and nnn for a description of language features that use the external data queue.
The configuration shall provide an external data queue mechanism. The following functions shall be
provided:
- Config_Push;
- Config_Queue;
- Config_Pull;
- Config_Queued.
The configuration may permit the external data queue to be altered in other ways. In the absence of such
alterations the external data queue shall be an ordered list. Config_Push adds the specified string to one
end of the list, Config_Queue to the other. Config_Pull removes a string from the end that Config_Push
adds to unless the list is empty.
5.8.1 Config_Push
Syntax:
Config_Push(String)
where:
String is the value to be retained in the external data queue.
Semantics:
Add String as an item to the end of the external data queue from which Config_Pull will remove an
item.
5.8.2 Config_Queue
Syntax:
Config_Queue(String)
where:
String is the value to be retained in the external data queue.
Semantics:
Add String as an item to the opposite end of the external data queue from which Config_Pull will
remove an item.
5.8.3 Config_Pull

25
Syntax:
Config_Pull()
Semantics:
Retrieve an item from the end of the external data queue to which Config_Push adds an element to
the list. Set #Outcome to the value of the retrieved item.
If no item could be retrieved the indicator of the completion response is 'F'.
5.8.4 Config_Queued
Syntax:
Config_Queued()
Semantics:
Get the count of items in the external data queue. Set #Outcome to that number.
5.9 Streams
The streams interface consists of functions which the configuration shall provide to manipulate streams.
See nnn, nnn, and nnn for a description of language features which use streams.
Streams are identified by names and provide for the reading and writing of data. They shall support the
concepts of characters, lines, positioning, default input stream and default output stream.
The concept of a persistent stream shall be supported and the concept of a transient stream may be
supported. A persistent stream is one where the content is not expected to change except when the
stream is explicitly acted on. A transient stream is one where the data available is expected to vary with
time.
The concepts of binary and character streams shall be supported. The content of a character stream is
expected to be characters.
The null string is used as a name for both the default input stream and the default output stream. The null
string names the default output stream only when it is an argument to the Config_Stream_Charout
operation.
The following functions shall be provided:
- Config_Stream_Charin;
- Config_Stream_Position;
- Config_Stream_Command;
- Config_Stream_State;
- Config_Stream_Charout;
- Config_Stream_Qualified;
- Config_Stream_Unique;
- Config_Stream_Query;
- Config_Stream_Close;
- Config_Stream_Count.
The results of these functions are described in terms of the following stems with tails which are stream
names:
- #Charin_Position.Stream;
- #Charout_Position.Stream;
- #Linein_Position.Stream;
- #Lineout_Position.Stream.
5.9.1 Config_Stream_Charin
Syntax:
Config_Stream_Charin(Stream, OperationType)
where:
Stream is the name of the stream to be processed.
OperationType is one of 'CHARIN', 'LINEIN', or 'NULL'.
Semantics:
Read from a stream. Increase #Linein_Position.Stream by one when the end-of-line indication is
encountered. Increase #Charin_Position.Stream when the indicator will be 'N'.
If OperationType is 'CHARIN' the state variables describing the stream will be affected as follows:
- when the configuration is able to provide data from a transient stream or the character at position
#Charin_Position.Stream of a persistent stream then #Outcome shall be set to contain the data.
The indicator of the response shall be 'N';

26
 - when the configuration is unable to return data because the read position is at the end of a
persistent stream then the indicator of the response shall be 'O';
- when the configuration is unable to return data from a transient stream because no data is
available and no data is expected to become available then the indicator of the response shall be
'O';
- otherwise the configuration is unable to return data and does not expect to be able to return data
by waiting; the indicator of the response shall be 'E'.
The data set in #Outcome will either be a single character or will be a sequence of eight characters,
each '0' or '1'. The choice is decided by the configuration. The eight character sequence indicates a
binary stream, see nnn.
If OperationType is 'LINEIN' then the action is the same as if Operation had been 'CHARIN' with the
following additional possibility. If end-of-line is detected any character (or character sequence) which
is an embedded indication of the end-of-line is skipped. The characters skipped contribute to the
change of #Charin_Position.Stream. #Outcome is the null string.
If OperationType is 'NULL' then the stream is accessed but no data is read.
5.9.2 Config_Stream_Position
Syntax:
Config_Stream_Position(Stream, OperationType, Position)
where:
Stream is the name of the stream to be processed.
Operation is 'CHARIN', 'LINEIN', 'CHAROUT', or 'LINEOUT'.
Position indicates where to position the stream.
Semantics:
If the operation is 'CHARIN' or 'CHAROUT' then Position is a character position, otherwise Position is
a line position.
If Operation is 'CHARIN' or 'LINEIN' and the Position is beyond the limit of the existing data then the
indicator of the completion response shall be 'R'. Otherwise if Operation is 'CHARIN' or 'LINEIN' set
#Charin_Position.Stream to the position from which the next Config_Stream_Charin on the stream
shall read, as indicated by Position. Set #Linein_Position.Stream to correspond with this position.
If Operation is 'CHAROUT' or 'LINEOUT' and the Position is more than one beyond the limit of existing
data then the indicator of the response shall be 'R'. Otherwise if Operation is 'CHAROUT' or
'LINEOUT' then #Charout_Position.Stream is set to the position at which the next
Config_Stream_Charout on the stream shall write, as indicated by Position. Set
#Lineout_Position.Stream to correspond with this position.
If this function is unable to position the stream because the stream is transient then the indicator of the
completion response shall be 'T'.
5.9.3 Config_Stream_Command
Syntax:
Config_Stream_Command(Stream, Command)
where:
Stream is the name of the stream to be processed.
Command is a configuration-specific command to be performed against the stream.
Semantics:
Issue a configuration-specific command against a stream. This may affect all state variables
describing Stream which hold position information. It may alter the effect of any subsequent operation
on the specified stream. If the indicator is set to 'N', #Outcome shall be set to information from the
command.
5.9.4 Config_Stream_State
Syntax:
Config_Stream_State(Stream)
where:
Stream is the name of the stream to be queried.
Semantics:
Set the indicator to reflect the state of the stream. Return an indicator equal to the indicator that an
immediately subsequent Config_Stream_Charin(Stream, 'CHARIN') would return. Alternatively, return
an indicator of 'U'.

27
 The remainder of the response shall be a configuration-dependent description of the state of the
stream.
5.9.5 Config_Stream_Charout
Syntax:
Config_Stream_Charout(Stream, Data)
where:
Stream is the name of the stream to be processed.
Data is the data to be written, or 'EOL' to indicate that an end-of-line indication is to be written, or a
null string. In the first case, if the stream is a binary stream then Data will be eight characters, each
'0' or '1', otherwise Data will be a single character.
Semantics:
When Data is the null string, no data is written.
Otherwise write to the stream. The state variables describing the stream will be affected as follows:
- when the configuration is able to write Data to a transient stream or at position
#Charout_Position.Stream of a persistent stream then the indicator in the response shall be 'N'.
When Data is not 'EOL' then #Charout_Position.Stream is increased by one. When Data is 'EOL',
then #Lineout_Position.Stream is increased by one and #Charout_Position.Stream is increased as
necessary to account for any end-of-line indication embedded in the stream;
- when the configuration is unable to write Data the indicator is set to 'E'.
5.9.6 Config_Stream_Qualified
Syntax:
Config_Stream_Qualified(Stream)
where:
Stream is the name of the stream to be processed.
Semantics:
Set #Outcome to some name which identifies Stream.
Return a completion response with indicator 'B' if the argument is not acceptable to the configuration
as identifying a stream.
5.9.7 Config_Stream_Unique
Syntax:
Config_Stream_Unique()
Semantics:
Set #Outcome to a name that the configuration recognizes as a stream name. The name shall not be
a name that the configuration associates with any existing data.
5.9.8 Config_Stream_Query
Syntax:
Config_Stream_Query(Stream)
where:
Stream is the name of the stream to be queried.
Semantics:
Set #Outcome to 'B' if the stream is a binary stream, or to 'C' if it is a character stream.
5.9.9 Config_Stream_Close
Syntax:
Config_Stream_Close(Stream)
where:
Stream is the name of the stream to be closed.
Semantics:
#Charout_Position.Stream and #Lineout_Position.Stream are set to 1 unless the stream has existing
data, in which case they are set ready to write immediately after the existing data.
If this function is unable to position the stream because the stream is transient then the indicator of the
completion response shall be 'T'.
5.9.10 Config_Stream_Count
Syntax:
Config_Stream_Count(Stream, Operation, Option)
where:
Stream is the name of the stream to be counted.
Operation is 'CHARS', or 'LINES'.

28
 Option is 'N' or 'C'.
Semantics:
If the option is 'N', #Outcome is set to zero if:
- the file is transient and no more characters (or no more lines if the Operation is 'LINES') are
expected to be available, even after waiting;
- the file is persistent and no more characters (or no more lines if the Operation is 'LINES') can be
obtained from this stream by Config_Stream_Charin before use of some function which resets
#Charin_Position.Stream and #Linein_Position.Stream.
If the option is 'N' and #Outcome is set nonzero, #Outcome shall be 1, or be the number of characters
(or the number of lines if Operation is 'LINES') which could be read from the stream before resetting.
If the option is 'C', #Outcome is set to zero if:
- the file is transient and no characters (or no lines if the Operation is 'LINES') are available without
waiting;
- the file is persistent and no more characters (or no more lines if the Operation is 'LINES') can be
obtained from this stream by Config_Stream_Charin before use of some function which resets
#Charin_Position.Stream and #Linein_Position.Stream.
If the option is 'C' and #Outcome is set nonzero, #Outcome shall be the number of characters (or the
number of lines if the Operation is 'LINES') which can be read from the stream without delay and
before resetting.
5.10 External variable pools
The external variable pools interface consists of functions which the configuration shall provide to
manipulate variables in external variable pools.
See nnn for the VALUE built-in function which uses external variable pools.
The configuration shall provide an external variable pools mechanism. The following functions shall be
provided:
- Config_Get;
- Config_Set.
The configuration may permit the external variable pools to be altered in other ways.
5.10.1 Config_Get
Syntax:
Config_Get(Poolid, Name)
where:
Poolid is an identification of the external variable pool.
Name is the name of a variable.
Semantics:
Get the value of a variable with name Name in the external variable pool Poolid. Set #Outcome to this
value.
If Poolid does not identify an external pool provided by this configuration, the indicator of the
completion response is 'P'.
If Name is not a valid name of a variable in the external pool, the indicator of the completion response
is 'F'.
5.10.2 Config_Set
Syntax:
Config_Set(Poolid, Name, Value)
where:
Poolid is an identification of the external variable pool.
Name is the name of a variable.
Value is the value to be assigned to the variable.
Semantics:
Set a variable with name Name in the external variable pool Poolid to Value.
If Poolid does not identify an external pool provided by this configuration, the indicator of the
completion response is 'P'.
If Name is not a valid name of a variable in the external pool, the indicator of the completion response
is 'F'.
5.11 Configuration characteristics

29
The configuration characteristics interface consists of a function which the configuration shall provide
which indicates choices decided by the configuration.
5.11.1 Config_Constants
Syntax:
Config_Constants()
Semantics:
Set the values of the following state variables:
- if there are any built-in functions which do not operate at NUMERIC DIGITS 9, then set variables
#Bif_Digits. (with various tails which are the names of those built-in functions) to the values to be
used;
- set variables #Limit_Digits, #Limit_EnvironmentName, #Limit_ExponentDigits, #Limit_Literal,
#Limit_MessageInsert, #Limit_Name, #Limit_String, #Limit_TraceData to the relevant limits. A
configuration shall allow a #Limit_MessageInsert value of 50 to be specified. A configuration shall
allow a #Limit_TraceData value of 250 to be specified;
- set #Configuration to a string identifying the configuration;
- set #Version to a string identifying the language processor. It shall have five words. Successive
words shall be separated by a blank character. The first four letters of the first word shall be
'REXX'. The second word shall be the four characters '5.00'. The last three words comprise a date.
This shall be in the format which is the default for the DATE() built-in function.
- set .nil to a value which compares unequal with any other value that can occur in execution.
- set .local .kernel .system?
5.12 Configuration routines
The configuration routines interface consists of functions which the configuration shall provide which
provide functions for a language processor.
The following functions shall be provided:
- Config_Trace_Query;
- Config_Trace_Input;
- Config_Trace_Output;
- Config_Default_Input;
- Config_Default_Output;
- Config_Initialization;
- Config_Termination;
- Config_Halt_Query;
- Config_Halt_Reset;
- Config_NoSource;
- Config_Time;
- Config_Random_Seed;
- Config_Random_Next.
5.12.1 Config_Trace_Query
Syntax:
Config_Trace_Query()
Semantics:
Indicate whether external activity is requesting interactive tracing. Set #Outcome to 'Yes' if interactive
tracing is currently requested. Otherwise set #Outcome to 'No'.
5.12.2 Config_Trace_Input
Syntax:
Config_Trace_Input()
Semantics:
Set #Outcome to a value from the source of trace input. The source of trace input is determined by the
configuration.
5.12.3 Config_Trace_Output
Syntax:
Config_Trace_Output(Line)
where:
Line is a string.
Semantics:

30
 Write String as a line to the destination of trace output. The destination of trace output is defined by
the configuration.
5.12.4 Config_Default_Input
Syntax:
Config_Default_Input()
Semantics:
Set #Outcome to the value that LINEIN( ) would return.
5.12.5 Config_Default_Output
Syntax:
Config_Default_Output(Line)
where:
Line is a string.
Semantics:
Write the string as a line in the manner of LINEOUT( ,Line).
5.12.6 Config_Initialization
Syntax:
Config_Initialization()
Semantics:
This function is provided only as a counterpart to Trap_Initialization; in itself it does nothing except
return the response. An indicator of 'F' gives rise to Msg3.1.
5.12.7 Config_Termination
Syntax:
Config_Termination()
Semantics:
This function is provided only as a counterpart to Trap_Termination; in itself it does nothing except
return the response. An indicator of 'F' gives rise to Msg2.1.
5.12.8 Config_Halt_Query
Syntax:
Config_Halt_Query()
Semantics:
Indicate whether external activity has requested a HALT condition to be raised. Set #Outcome to 'Yes'
if HALT is requested. Otherwise set #Outcome to 'No'.
5.12.9 Config_Halt_Reset
Syntax:
Config_Halt_Reset()
Semantics:
Reset the configuration so that further attempts to cause a HALT condition will be recognized.
5.12.10 Config_NoSource
Syntax:
Config_NoSource()
Semantics:
Indicate whether the source of the program may or may not be output by the language processor.
Set #NoSource to '1' to indicate that the source of the program may not be output by the language
processor, at various points in processing where it would otherwise be output. Otherwise, set
#NoSource to '0'.
A configuration shall allow any program to be processed in such a way that Config_NoSource() sets
#NoSource to '0'. A configuration may allow any program to be processed in such a way that
Config_NoSource() sets #NoSource to '1'.
5.12.11 Config_Time
Syntax:
Config_Time()
Semantics:
Get a time stamp. Set #Time to a string whose value is the integer number of microseconds that have
elapsed between 00:00:00 on January first 0001 and the time that Config_Time is called, at longitude
zero. Values sufficient to allow for any date in the year 9999 shall be supported. The value returned
may be an approximation but shall not be smaller than the value returned by a previous use of the
function.

31
 Set #Adjust<Index "#Adjust" # "" > to an integer number of microseconds. #Adjust<Index "#Adjust" #
"" > reflects the difference between the local date/time and the date/time corresponding to #Time.
#Time + #Adjust<Index "#Adjust" # "" > is the local date/time.
5.12.12 Config_Random_Seed
Syntax:
Config_Random_Seed(Seed)
where:
Seed is a sequence of up to #Bif_Digits.RANDOM digits.
Semantics:
Set a seed, so that subsequent uses of Config_Random_Next will reproducibly return quasi-random
numbers.
5.12.13 Config_Random_Next
Syntax:
Config_Random_Next(Min, Max)
where:
Min is the lower bound, inclusive, on the number returned in #Outcome.
Max is the upper bound, inclusive, on the number returned in #Outcome.
Semantics:
Set #Outcome to a quasi-random nonnegative integer in the range Min to Max.
5.12.14 Config_Options
Syntax:
Config_Options(String)
where:
String is a string.
Semantics:
No effect beyond the effects common to all Config_ invocations. The value of the string will have
come from an OPTIONS instruction, see nnn.
5.13 Traps
The trapping interface consists of functions which may be provided by the caller of API_Start (see nnn) as
a list of traps. Each trap may be specified or omitted. The language processor shall invoke a specified
trap before, or instead of, using the corresponding feature of the language processor itself. This
correspondence is implied by the choice of names; that is, a name beginning Trap_ will correspond to a
name beginning Config_ when the remainder of the name is the same. Corresponding functions are
called with the same interface, with one exception. The exception is that a trap may return a null string.
When a trap returns a null string, the corresponding Config_ function is invoked; otherwise the invocation
of the trap replaces the potential invocation of the Config_ function.
In the rest of this standard, the trapping mechanism is not shown explicitly. It is implied by the use of a
Config_ function.
The names of the traps are
- Trap_Command;
- Trap_ExternalRoutine;
- Trap_Push;
- Trap_Queue;
- Trap_Pull;
- Trap_Queued;
- Trap_Trace_Query;
- Trap_Trace_Input;
- Trap_Trace_Output;
- Trap_Default_Input;
- Trap_Default_Output;
- Trap_Initialization;
- Trap_Termination;
- Trap_Halt_Query;
- Trap_Halt_Reset.
5.14 Variable pool

32
How does this fit with variables as properties?

The variable pool interface consists of functions which the configuration shall provide to manipulate the
variables and to obtain some characteristics of a Rexx program.
These functions can be called from programs not written in Rexx _ commands and external routines
invoked from a Rexx program, or traps invoked from the language processor.
All the functions comprising the variable pool interface shall return with an indication of whether an error
occurred. They shall return indicating an error and have no other effect, if #API_Enabled has a value of '0'
or if the arguments to them fail to meet the defined syntactic constraints.
These functions interact with the processing of clauses. To define this interaction, the functions are
described here in terms of the processing of variables, see nnn.
Some of these functions have an argument which is a symbol. A symbol is a string. The content of the
string shall meet the syntactic constraints of the left hand side of an assignment. Conversion to
uppercase and substitution in compound symbols occurs as it does for the left hand side of an
assignment. The symbol identifies the variable to be operated upon.
Some of the functions have an argument which is a direct symbol. A direct symbol is a string. The content
of this string shall meet the syntactic constraints of a VAR_SYMBOL in uppercase with no periods or it
shall be the concatenation of a part meeting the syntactic constraints of a stem in uppercase, and a part
that is any string. In the former case the symbol identifies the variable to be operated upon. In the latter
case the variable to be operated on is one with the specified stem and a tail which is the remainder of the
direct symbol.
Functions that have an argument which is symbol or direct symbol shall return an indication of whether
the identified variable existed before the function was executed.
Clause nnn defines functions which manipulate Rexx variable pools. Where possible the functions
comprising the variable pool interface are described in terms of the appropriate invocations of the
functions defined in nnn. The first parameter on these calls is the state variable #Pool. If these Var_
functions do not return an indicator 'N', 'R', or 'D' then the API_ function shall return an error indication.
5.14.1 API_Set
Syntax:
API_Set(Symbol, Value)
where:
Symbol is a symbol.
Value is the string whose value is to be assigned to the variable.
Semantics:
Assign the value of Value to the variable identified by Symbol. If Symbol contains no periods or
contains one period as its last character:
Var_Set(#Pool, Symbol, '0', Value)
Otherwise:
Var_Set(#Pool, #Symbol, '1', Value)
where:
#Symbol is Symbol after any replacements in the tail as described by nnn.
5.14.2 API_Value
Syntax:
API_Value(Symbol)
where:
Symbol is a symbol.
Semantics:
Return the value of the variable identified by Symbol. If Symbol contains no periods or contains one
period as its last character this is the value of #Outcome after:
Var_Value(#Pool, Symbol, '0')
Otherwise the value of #Outcome after:
Var_Value(#Pool, #Symbol, '1')
where:
#Symbol is Symbol after any replacements in the tail as described by nnn.
5.14.3 API_Drop
Syntax:
API_Drop(Symbol)

33
 where:
Symbol is a symbol.
Semantics:
Drop the variable identified by Symbol. If Symbol contains no periods or contains one period as its last
character:
Var_Drop(#Pool, Symbol, '0')
Otherwise:
Var_Drop(#Pool, #Symbol, '1')
where:
#Symbol is Symbol after any replacements in the tail as described by nnn.
5.14.4 API_SetDirect
Syntax:
API_SetDirect(Symbol, Value)
where:
Symbol is a direct symbol.
Value is the string whose value is to be assigned to the variable.
Semantics:
Assign the value of Value to the variable identified by Symbol. If the Symbol contains no period:
Var_Set(#Pool, Symbol, '0', Value)
Otherwise:
Var_Set(#Pool, Symbol, '1', Value)
5.14.5 API_ValueDirect
Syntax:
API_ValueDirect(Symbol)
where:
Symbol is a direct symbol.
Semantics:
Return the value of the variable identified by Symbol. If the Symbol contains no period:
Var_Value(#Pool, Symbol, '0')
Otherwise:
Var_Value(#Pool, Symbol, '1')
5.14.6 API_DropDirect
Syntax:
API_DropDirect(Symbol)
where:
Symbol is a direct symbol.
Semantics:
Drop the variable identified by Symbol. If the Symbol contains no period:
Var_Drop(#Pool, Symbol, '0')
Otherwise:
Var_Drop(#Pool, Symbol, '1')
5.14.7 API_ValueOther
Syntax:
API_ValueOther(Qualifier)
where:
Qualifier is an indication distinguishing the result to be returned including any necessary further
qualification.
Semantics:
Return characteristics of the program, depending on the value of Qualifier. The possibilities for the
value to be returned are:
- the value of #Source;
- the value of #Version;
- the largest value of n such that #ArgExists.1.n is '1', see nnn;
- the value of #Arg.1.n where n is an integer value provided as input.
5.14.8 API_Next
Syntax:
API_Next()
Semantics:

34
 Returns both the name and the value of some variable in the variable pool that does not have the
attribute 'dropped' or the attribute 'implicit' and is not a stem; alternatively return an indication that
there is no suitable name to return. When API_Next is called it will return a name that has not
previously been returned; the order is undefined. This process of returning different names will restart
whenever the Rexx processor executes Var_Reset.
5.14.9 API_NextVariable
Syntax:
API_NextVariable()
Semantics:
Returns both the name and the value of some variable in the variable pool that does not have the
attribute 'dropped' or the attribute 'implicit'; alternatively, return an indication that there is no suitable name
to return. When API_NextVariable is called it will return data about a variable that has not previously been
returned; the order is undefined. This process of returning different names will restart whenever the Rexx
processor executes Var_Reset. In addition to the name and value, an indication of whether the variable
was 'tailed' will be returned.

35
6 Syntax constructs
6.1 Notation
6.1.1 Backus-Naur Form (BNF)
The syntax constructs in this standard are defined in Backus-Naur Form (BNF). The syntax used in these
BNF productions has
- a left-hand side (called identifier);
- the characters ':=';
- a right-hand side (called bnf_expression).
The left-hand side identifies syntactic constructs. The right-hand side describes valid ways of writing a
specific syntactic construct.
The right-hand side consists of operands and operators, and may be grouped.
6.1.2 Operands
Operands may be terminals or non-terminals. If an operand appears as identifier in some other
production it is called a non-terminal, otherwise it is called a terminal. Terminals are either literal or
symbolic.
Literal terminals are enclosed in quotes and represent literally (apart from case) what must be present in
the source being described.
Symbolic terminals formed with lower case characters represent something which the configuration may,
or may not, allow in the source program, see nnn, nnn, nnn, nnn.
Symbolic terminals formed with uppercase characters represent events and tokens, see nnn and nnn.
6.1.3 Operators
The following lists the valid operators, their meaning, and their precedence; the operator listed first has
the highest precedence; apart from precedence recognition is from left to right:
- the postfix plus operator specifies one or more repetitions of the preceding construct;
- abuttal specifies that the preceding and the following construct must appear in the given order;
- the operator '|' specifies alternatives between the preceding and the following constructs.
6.1.4 Grouping
Parentheses and square brackets are used to group constructs. Parentheses are used for the purpose of
grouping only. Square brackets specify that the enclosed construct is optional.
6.1.5 BNF syntax definition
The BNF syntax, described in BNF, is:
production := identifier ':=' bnf_expression
bnf_expression := abuttal | bnf_expression '|' abuttal
abuttal := [abuttal] bnf_primary
bnf_primary := '[' bnf_expression ']' | '(' bnf_expression ')' | literal |
identifier | message_identifier | bnf_primary '+'
6.1.6 Syntactic errors
The syntax descriptions (see nnn and nnn) make use of message_identifiers which are shown as
Msgnn.nn or Msgnn, where nn is a number. These actions produce the correspondingly numbered error
messages (see nnn and nnn).
6.2 Lexical
The lexical level processes the source and provides tokens for further recognition by the top syntax level.
6.2.1 Lexical elements
6.2.1.1 Events
The fully-capitalized identifiers in the BNF syntax (see nnn) represent events. An event is either supplied
by the configuration or occurs as result of a look-ahead in left-to-right parsing. The following events are
defined:
- EOL occurs at the end of a line of the source. It is provided by Config_SourceChar, see nnn;
- EOS occurs at the end of the source program. It is provided by Config_SourceChar;
- RADIX occurs when the character about to be scanned is 'X' or 'x' or 'B' or 'b' not followed by a
general_letter, or a digit, or '.';
- CONTINUE occurs when the character about to be scanned is ',', and the characters after the ',' up
to EOL represent a repetition of comment or blank, and the EOL is not immediately followed by an
EOS;

36
 - EXPONENT_SIGN occurs when the character about to be scanned is '+' or '-', and the characters to
the left of the sign, currently parsed as part of Const_symbol, represent a plain_number followed by 'E'
or 'e', and the characters to the right of the sign represent a repetition of digit not followed by a
general_letter or '.'.
- I would put ASSIGN here for the leftmost '=' in a clause that is not within parentheses or brackets. But Simon not
happy with message term being an assignment?
6.2.1.2 Actions and tokens
Mixed case identifiers with an initial capital letter cause an action when they appear as operands in a
production. These actions perform further tests and create tokens for use by the top syntax level. The
following actions are defined:
- Special supplies the source recognized as special to the top syntax level;
- Eol supplies a semicolon to the top syntax level;
- Eos supplies an end of source indication to the top syntax level;
- Var_symbol supplies the source recognized as Var_symbol to the top syntax level, as keywords or
VAR_SYMBOL tokens, see nnn. The characters in a Var_symbol are converted by Config_Upper to
uppercase. Msg30.1 shall be produced if Var_symbol contains more than #Limit_Name characters,
see nnn;
- Const_symbol supplies the source recognized as Const_symbol to the top syntax level. If it is a
number it is passed as a NUMBER token, otherwise it is passed as a CONST_SYMBOL token. The
characters in a Const_symbol are converted by Config_Upper to become the characters that comprise
that NUMBER or CONST_SYMBOL. Msg30.1 shall be produced if Const_symbol contains more than
#Limit_Name characters;
- Embedded_quotation_mark records an occurrence of two consecutive quotation marks within a
string delimited by quotation marks for further processing by the String action;
- Embedded_apostrophe records an occurrence of two consecutive apostrophes within a string
delimited by apostrophes for further processing by the String action;
- String supplies the source recognized as String to the top syntax level as a STRING token. Any
occurrence of Embedded_quotation_mark or Embedded_apostrophe is replaced by a single quotation
mark or apostrophe, respectively. Msg30.2 shall be produced if the resulting string contains more than
#Limit_Literal characters;
- Binary_string supplies the converted binary string to the top syntax level as a STRING token, after
checking conformance to the binary_string syntax. If the binary_string does not contain any
occurrence of a binary_digit, a string of length 0 is passed to the top syntax level. The occurrences of
binary_digit are concatenated to form a number in radix 2. Zero or 4 digits are added at the left if
necessary to make the number of digits a multiple of 8. If the resulting number of digits exceeds 8
times #Limit_Literal then Msg30.2 shall be produced. The binary digits are converted to an encoding,
see nnn. The encoding is supplied to the top syntax level as a STRING token;
- Hex_string supplies the converted hexadecimal string to the top syntax level as a STRING token,
after checking conformance to the hex_string syntax. If the hex_string does not contain any
occurrence of a hex_digit, a string of length 0 is passed to the top syntax level. The occurrences of
hex_digit are each converted to a number with four binary digits and concatenated. 0 to 7 digits are
added at the left if necessary to make the number of digits a multiple of 8. If the resulting number of
digits exceeds 8 times #Limit_Literal then Msg30.2 shall be produced. The binary digits are converted
to an encoding. The encoding is supplied to the top syntax level as a STRING token;
- Operator supplies the source recognized as Operator (excluding characters that are not
operator_char ) to the top syntax level. Any occurrence of an other_negator within Operator is
supplied as '\';
- Blank records the presence of a blank. This may subsequently be tested (see nnn).
Constructions of type Number, Const_symbol, Var_symbol or String are called operands.
6.2.1.3 Source characters
The source is obtained from the configuration by the use of Config_SourceChar (see nnn). If no character
is available because the source is not a correct encoding of characters, message Msg22.1 shall be
produced.
The terms extra_letter, other_blank_character, other_negator, and other_character used in the
productions of the lexical level refer to characters of the groups extra_letters (see nnn),

37
other_blank_characters (see nnn), other_negators (see nnn) and other_characters (see nnn),
respectively.
6.2.1.4 Rules
In scanning, recognition that causes an action (see nnn) only occurs if no other recognition is possible,
except that Embedded_apostrophe and Embedded_quotation_mark actions occur wherever possible.
6.2.2 Lexical level
6.2.3 Interaction between levels of syntax
When the lexical process recognizes tokens to be supplied to the top level, there can be changes made
or tokens added. Recognition is performed by the lexical process and the top level process in a
synchronized way. The tokens produced by the lexical level can be affected by what the top level syntax
has recognized. Those tokens will affect subsequent recognition by the top level. Both processes operate
on the characters and the tokens in the order they are produced. The term "context" refers to the
progress of the recognition at some point, without consideration of unprocessed characters and tokens.
If a token which is '+', '-', '\' or '(' appears in a lexical level context (other than after the keyword 'PARSE')
where the keyword 'VALUE' could appear in the corresponding top level context, then 'VALUE' is passed
to the top level before the token is passed.
If an '=' operator_char appears in a lexical level context where it could be the '=' of an assignment or
message_instruction in the corresponding top level context then it is recognized as the '=' of that
instruction. (It will be outside of brackets and parentheses, and any Var_symbol immediately preceding it
is passed as a VAR_SYMBOL). If an operand is followed by a colon token in the lexical level context then
the operand only is passed to the top level syntax as a LABEL, provided the context permits a LABEL.
Except where the rules above determine the token passed, a Var_symbol is passed as a terminal (a
keyword) rather than as a VAR_SYMBOL under the following circumstances:
- if the symbol is spelled 'WHILE' or 'UNTIL' it is a keyword wherever a VAR_SYMBOL would be part
of an expression within a do_specification;
- if the symbol is spelled 'TO' , 'BY', or 'FOR' it is a keyword wherever a VAR_SYMBOL would be part
of an expression within a do_rep;
- if the symbol is spelled 'WITH' it is a keyword wherever a VAR_SYMBOL would be part of a
parsevalue, or part of an expression or taken_constant within address;
- if the symbol is spelled 'THEN' it is keyword wherever a VAR_SYMBOL would be part of an
expression immediately following the keyword 'IF' or 'WHEN'.
Except where the rules above determine the token passed, a Var_symbol is passed as a keyword if the
spelling of it matches a keyword which the top level syntax recognizes in its current context, otherwise the
Var_symbol is passed as a VAR_SYMBOL token.
In a context where the top level syntax could accept a '||' token as the next token, a '||' operator or a ' '
operator may be inferred and passed to the top level provided that the next token from the lexical level is
a left parenthesis or an operand that is not a keyword. If the blank action has recorded the presence of
one or more blanks to the left of the next token then the ' ' operator is inferred. Otherwise, a '||' operator is
inferred, except if the next token is a left parenthesis following an operand (see nnn); in this case no
operator is inferred.
When any of the keywords 'OTHERWISE', 'THEN', or 'ELSE' is recognized, a semicolon token is supplied
as the following token. A semicolon token is supplied as the previous token when the 'THEN' keyword is
recognized. A semicolon token is supplied as the token following a LABEL.
6.2.3.1 Reserved symbols
A Const_symbol which starts with a period and is not a Number shall be spelled .MN, .RESULT, .RC,
.RS, or .SIGL otherwise Msg50.1 is issued.
6.2.3.2 Function name syntax
A symbol which is the leftmost component of a function shall not end with a period, otherwise Msg51.1 is
issued.
6.3 Syntax
6.3.1 Syntax elements
The tokens generated by the actions described in nnn form the basis for recognizing larger constructs.
6.3.2 Syntax level

starter:=x3j18

38
x3j18:=program Eos | Msg35.1
program := [label_list][ncl][requires+][prolog_instruction+]
(class_definition [requires+])+
requires := 'REQUIRES' ( taken_constant | Msg19.8 ) ';'+
prolog_instruction:= (package | import | options) ncl
package := 'PACKAGE'( NAME | Msgnn )
import := 'IMPORT' ( NAME | Msgnn ) ['.']
options := 'OPTIONS' ( symbol+ | Msgnn )
ncl := null_clause+ | Msg21.1
null_clause := ';' [label_list]
label_list := (LABEL ';')+
class_definition := class [property_info][method_definition+]
class := 'CLASS' ( taken_constant | Msg19.12 ) [class_option+]
['INHERIT' ( taken_constant | Msg19.13 )+] ncl
class_option := visibility | modifier | 'BINARY' | 'DEPRECATED'
| 'EXTENDS' ( NAME | Msgnn )
| 'USES' ( NAMElist | Msgnn )
| 'IMPLEMENTS' ( NAMElist | Msgnn )
| external | metaclass | submix /* | 'PUBLIC' */
external := 'EXTERNAL' (STRING | Msg19.14)
metaclass := 'METACLASS' ( taken_constant | Msg19.15 )
submix := 'MIXINCLASS' ( taken_constant | Msg19.16 )
| 'SUBCLASS' ( taken_constant | Msg19.17 )
visibility := 'PUBLIC' | 'PRIVATE'
modifier := 'ABSTRACT' | 'FINAL' | 'INTERFACE' | 'ADAPTER'
NAMElist := NAME [(',' ( NAME | Msgnn ) )+]
property_info := numeric | property_assignment | properties | trace
numeric := 'NUMERIC' (numeric_digits | numeric_form | Msg25.15)
numeric_digits:= 'DIGITS' [expression]
numeric_form := 'FORM' ['ENGINEERING' | 'SCIENTIFIC']
property_assignment := NAME | assignment
properties := 'PROPERTIES' ( properties_option+ | Msgnn)
properties_option := properties_visibility | properties_modifier
properties_visibility := 'INHERITABLE' | 'PRIVATE' | 'PUBLIC' |
'INDIRECT'
properties_modifier := 'CONSTANT' | 'STATIC' | 'VOLATILE' | 'TRANSIENT'
trace := 'TRACE' ['ALL' | 'METHODS' | 'OFF' | 'RESULTS']
method_definition := (method [expose ncl]| routine)
balanced
expose := 'EXPOSE' variable_list
method := 'METHOD' (taken_constant | Msg19.9)
[ '(' assigncommalist | Msgnn ( ')' | Msgnn )]
[method_option+] ncl
assigncommalist := assignment [(',' ( assignment | Msgnn ) )+]
method_option := method_visibility | method_modifier | 'PROTECT'
| 'RETURNS' ( term | Msgnn )
| 'SIGNAL' ( termcommalist | Msgnn )
| 'DEPRECATED'
| 'CLASS' | 'ATTRIBUTE' | /*'PRIVATE' | */ guarded
guarded := 'GUARDED' | 'UNGUARDED'
method_visibility := 'INHERITABLE' | 'PRIVATE' | 'PUBLIC' | 'SHARED'
method_modifier := 'ABSTRACT' | 'CONSTANT' | 'FINAL' | 'NATIVE' |
'STATIC'
termcommalist := term [(',' ( term | Msgnn ) )+]
routine := 'ROUTINE' ( taken_constant | Msg19.11 ) ['PUBLIC'] ncl
balanced:= instruction_list ['END' Msg10.1]
instruction_list:= instruction+
/* The second part is about groups */
instruction := group | single_instruction ncl
group := do ncl | if | loop ncl | select ncl
do := do_specification ncl [instruction+] [group_handler]
('END' [NAME] | Eos Msg14.1 | Msg35.1)
group_option := 'LABEL' ( NAME | Msgnn ) | 'PROTECT' ( term | Msgnn )

39
 group handler := catch | finally | catch finally
catch := 'CATCH' [ NAME '=' ] ( NAME | Msgnn) ncl [instruction+]
/* FINALLY implies a semicolon. */
finally := 'FINALLY' ncl ( instruction+ | Msgnn )
if := 'IF' expression [ncl] (then | Msg18.1)
[else]
then := 'THEN' ncl
(instruction | EOS Msg14.3 | 'END' Msg10.5)
else := 'ELSE' ncl
(instruction | EOS Msg14.4 | 'END' Msg10.6)
loop := 'LOOP' [group_option+] [repetitor] [conditional] ncl
instruction+ [group_handler]
loop_ending
loop_ending := 'END' [VAR_SYMBOL] | EOS Msg14.n | Msg35.1
conditional := 'WHILE' whileexpr | 'UNTIL' untilexpr
untilexpr := expression
whileexpr := expression
repetitor := assignment [count_option+] | expression | over |
'FOREVER'
count_option := loopt | loopb | loopf
loopt := 'TO' expression
loopb := 'BY' expression
loopf := 'FOR' expression
over := VAR_SYMBOL 'OVER' expression
| NUMBER 'OVER' Msg31.1
| CONST_SYMBOL 'OVER' (Msg31.2 | Msg31.3)
select := 'SELECT' [group_option+] ncl select_body [group_handler]
('END' [NAME Msg10.4] | EOS Msg14.2 | Msg7.2)
select_body := (when | Msg7.1) [when+] [otherwise]
when := 'WHEN' expression [ncl] (then | Msg18.2)
otherwise := 'OTHERWISE' ncl [instruction+]

/* Third part is for single instructions. */

single_instruction:= assignment | message_instruction | keyword_instruction
|command
assignment := VAR_SYMBOL '#' expression
| NUMBER '#' Msg31.1
| CONST_SYMBOL '#' (Msg31.2 | Msg31.3)
message_instruction := message_term | message_term '#' expression
keyword_instruction:= address | arg | call | drop | exit
| interpret | iterate | leave
| nop | numeric | options
| parse | procedure | pull | push | queue
| raise | reply | return | say | signal | trace | use
| 'THEN' Msg8.1 | 'ELSE' Msg8.2
| 'WHEN' Msg9.1 | 'OTHERWISE' Msg9.2
command := expression
address := 'ADDRESS' [(taken_constant [expression]
| Msg19.1 | valueexp) [ 'WITH' connection]]
taken_constant := symbol | STRING
valueexp := 'VALUE' expression
connection := ad_option+
ad_option := error | input | output | Msg25.5
error := 'ERROR' (resourceo | Msg25.14)
input := 'INPUT' (resourcei | Msg25.6)
resourcei := resources | 'NORMAL'
output := 'OUTPUT' (resourceo | Msg25.7)
resourceo := 'APPEND' (resources | Msg25.8)
| 'REPLACE' (resources | Msg25.9)
| resources | 'NORMAL'
resources := 'STREAM' (VAR_SYMBOL | Msg53.1)
| 'STEM' (VAR_SYMBOL | Msg53.2)
vref := '(' var_symbol (')' | Msg46.1)
var_symbol := VAR_SYMBOL | Msg20.1

40
arg := 'ARG' [template list]
call := 'CALL' (callon_spec|
(taken_constant | vref | Msg19.2)[expression_list])
callon_spec := 'ON' (callable_condition | Msg25.1)
['NAME' (symbol_constant_term | Msg19.3)]
| 'OFF' (callable_condition | Msg25.2)
symbol_constant_term := term
callable_condition:= 'ANY' | 'ERROR' | 'FAILURE' | 'HALT' | 'NOTREADY'
| 'USER' ( symbol_constant_term | Msg19.18 )
condition := callable_condition | 'LOSTDIGITS'
| 'NOMETHOD' | 'NOSTRING' | 'NOVALUE' | 'SYNTAX'
expression_list := expr | [expr] ',' [expression_list]
do_specification := do_simple | do_repetitive
do_simple := 'DO' [group_option+]
do_repetitive := do_simple (dorep | conditional | dorep conditional)
dorep := 'FOREVER' | repetitor
drop := 'DROP' variable_list
variable_list := (vref | var_symbol)+
exit := 'EXIT' [expression]
forward := 'FORWARD' [forward_option+ | Msg25.18]
forward_option := 'CONTINUE' | ArrayArgOption |
MessageOption | ClassOption | ToOption
ArrayArgOption:='ARRAY' arguments | 'ARGUMENTS' term
MessageOption :='MESSAGE' term
ClassOption :='CLASS' term
ToOption :='TO' term
guard := 'GUARD' ('ON' | Msg25.22) [('WHEN' | Msg25.21)
expression]
| ( 'OFF' | Msg25.19) [('WHEN' | Msg25.21)
expression]
interpret := 'INTERPRET' expression
iterate := 'ITERATE' [VAR_SYMBOL | Msg20.2]
leave := 'LEAVE' [VAR_SYMBOL | Msg20.2]
nop := 'NOP'
numeric := 'NUMERIC' (numeric_digits | numeric_form
| numeric_fuzz | Msg25.15)
numeric_digits := 'DIGITS' [expression]
numeric_form := 'FORM' [numeric_form_suffix]
numeric_form_suffix:=('ENGINEERING'|'SCIENTIFIC'|valueexp | Msg25.11)
numeric_fuzz := 'FUZZ' [expression]
options := 'OPTIONS' expression
parse := 'PARSE' [translations] (parse_type
|Msg25.12)[template_list]
translations := 'CASELESS' ['UPPER' | 'LOWER']
| ('UPPER' | 'LOWER') ['CASELESS']
parse_type := parse_key | parse_value | parse_var | term
parse_key := 'ARG' | 'PULL' | 'SOURCE' | 'LINEIN'
| 'VERSION'
parse_value := 'VALUE' [expression] ('WITH' | Msg38.3)
parse_var := 'VAR' var_symbol
template := NAME [( [pattern] NAME)+]
pattern:= STRING | [indicator] NUMBER | [indicator] '(' symbol ')'
indicator := '+' | '-' | '='
procedure := 'PROCEDURE' [expose | Msg25.17]
pull := 'PULL' [template_list]
push := 'PUSH' [expression]
queue := 'QUEUE' [expression]
raise := 'RAISE' conditions (raise_option | Msg25.24)
conditions := 'ANY' | 'ERROR' term | 'FAILURE' term
| 'HALT'| 'LOSTDIGITS' | 'NOMETHOD' | 'NOSTRING' |
'NOTREADY'
| 'NOVALUE' | 'PROPAGATE' | 'SYNTAX' term
| 'USER' ( symbol_constant_term | Msg19.18) | Msg25.23
raise_option := ExitRetOption | Description | ArrayOption

41
 ExitRetOption := 'EXIT' [term] | 'RETURN' [term]
Description :='DESCRIPTION' term
ArrayOption := 'ADDITIONAL' term | 'ARRAY' arguments
reply := 'REPLY' [ expression]
return := 'RETURN' [expression]
say := 'SAY' [expression]
signal := 'SIGNAL' (signal_spec | valueexp
| symbol_constant_term | Msg19.4)
signal_spec := 'ON' (condition | Msg25.3)
['NAME' (symbol_constant_term | Msg19.3)]
| 'OFF' (condition | Msg25.4)
trace := 'TRACE' [(taken_constant | Msg19.6) | valueexp]
use := 'USE' ('ARG' | Msg25.26) [use_list]
use_list := VAR_SYMBOL | [VAR_SYMBOL] ',' [use_list]
/* Note: The next part describes templates. */
template_list := template | [template] ',' [template_list]
template := (trigger | target | Msg38.1)+
target := VAR_SYMBOL | '.'
trigger := pattern | positional
pattern := STRING | vrefp
vrefp := '(' (VAR_SYMBOL | Msg19.7) (')' | Msg46.1)
positional := absolute_positional | relative_positional
absolute_positional:= NUMBER | '=' position
position := NUMBER | vrefp | Msg38.2
relative_positional:= ('+' | '-') position
/* Note: The final part specifies the various forms of symbol, and
expression. */
symbol := VAR_SYMBOL | CONST_SYMBOL | NUMBER
expression := expr [(',' Msg37.1) | (')' Msg37.2 )]
expr := expr_alias
expr_alias := and_expression
| expr_alias or_operator and_expression
or_operator := '|' | '&&'
and_expression := comparison | and_expression '&' comparison
comparison := concatenation
| comparison comparison_operator concatenation
comparison_operator:= normal_compare | strict_compare
normal_compare:= '=' | '\=' | '<>' | '><' | '>' | '<' | '>='
| '<=' | '\>' | '\<'
strict_compare:= '==' | '\==' | '>>' | '<<' | '>>=' | '<<='
| '\>>' | '\<<'
concatenation := addition
| concatenation (' ' | '||') addition
addition := multiplication
| addition additive_operator multiplication
additive_operator:= '+' | '-'
multiplication := power_expression
| multiplication multiplicative_operator
power_expression
multiplicative_operator:= '*' | '/' | '//' | '%'
power_expression := prefix_expression
| power_expression '**' prefix_expression
prefix_expression := ('+' | '-' | '\') prefix_expression
| term | Msg35.1
/* "Stub" has to be identified semantically? */
term := simple_term [ '.' ( term | Msgnn )]
simple_term := symbol | STRING | invoke | indexed
| '(' expression ( ')' | Msg36 )
| initializer
| message_term '##'
message_term:= term ('~' | '~~') method_name [arguments]
| term '['[ expression_list ] (']' | Msg36.2)

42
 method name:=(taken constant | Msg19.19)
[':' ( VAR_SYMBOL | Msg19.21 )]
/* Method-call without arguments is syntactically like symbol. */
/* Editor - not sure of my notes about here. */
invoke := (symbol | STRING) arguments
arguments := '#(' [expression_list] (')' | Msg36)
expression_list := expression | [expression] ',' [expression_list]
indexed := (symbol | STRING) indices
indices := '#[' [expression_list] (']' | Msg36.n)
initializer := '['expression_list (']' | Msg36.n)

6.1 Syntactic information

6.1.1 VAR_SYMBOL matching
Any VAR_SYMBOL in a do_ending must be matched by the same VAR_SYMBOL occurring at the start
of an assignment contained in the do_specification of the do that contains both the do_specification and
the do_ending, as described in nnn.
If there is a VAR_SYMBOL in a do_ending for which there is no assignment in the corresponding
do_specification then message Msg10.3 is produced and no further activity is defined.
If there is a VAR_SYMBOL in a do_ending which does not match the one occurring in the assignment
then message Msg10.2 is produced and no further activity is defined.
An iterate or leave must be contained in the instruction_list of some do with a do_specification which is
do_repetitive, otherwise a message (Msg28.2 or Msg28.1 respectively) is produced and no further activity
is defined.
If an iterate or leave contains a VAR_SYMBOL there must be a matching VAR_SYMBOL in a
do_specification, otherwise a message (Msg28.1, Msg28.2, Msg28.3 or Msg28.4 appropriately) is
produced and no further activity is defined. The matching VAR_SYMBOL will occur at the start of an
assignment in the do_specification. The do_specification will be associated with a do by nnn. The iterate
or leave will be a single instruction in an instruction_list associated with a do by nnn. These two dos shall
be the same, or the latter nested one or more levels within the former. The number of levels is called the
nesting_correction and influences the semantics of the iterate or leave. It is zero if the two dos are the
same. The nesting_correction for iterates or leaves that do not contain VAR_SYMBOL is zero.
6.1.2 Trace-only labels
Instances of LABEL which occur within a grouping_instruction and are not in a ncl at the end of that
grouping_instruction are instances of trace-only labels.
6.1.3 Clauses and line numbers
The activity of tracing execution is defined in terms of clauses. A program consists of clauses, each
clause ended by a semicolon special token. The semicolon may be explicit in the program or inferred.
The line number of a clause is one more than the number of EOL events recognized before the first token
of the clause was recognized.
6.1.4 Nested IF instructions
The syntax specification nnn allows 'IF' instructions to be nested and does not fully specify the
association of an 'ELSE' keyword with an 'IF' keyword. An 'ELSE' associates with the closest prior 'IF' that
it can associate with in conformance with the syntax.
6.1.5 Choice of messages
The specifications nnn and nnn permit two alternative messages in some circumstances. The following
rules apply:
- Msg15.1 shall be preferred to Msg15.3 if the choice of Msg15.3 would result in the replacement for
the insertion being a blank character;
- Msg15.2 shall be preferred to Msg15.4 if the choice of Msg15.4 would result in the replacement for
the insertion being a blank character;
- Msg31.3 shall be preferred to Msg31.2 if the replacement for the insertion in the message starts with
a period;
- Preference is given to the message that appears later in the list: Msg21.1, Msg27.1, Msg25.16,
Msg36, Msg38.3, Msg35.1, other messages.
6.1.6 Creation of messages

43
If a message_term contains a '[' it is regarded as an equivalent message_term without a '[', for execution.
The equivalent is term~'[]'(expression_list). See nnn. If a message_instruction has the construction
message_term '=' expression it is regarded as equivalent to a message_term with the same components
as the message_term left of the '=', except that the taken_constant has an '=' character appended and
arguments has the expression from the right of the '=' as an extra first argument. See nnn.

45
7 Evaluation
The syntax section describes how expressions and the components of expressions are written in a
program. It also describes how operators can be associated with the strings, symbols and function results
which are their operands.
This evaluation section describes what values these components have in execution, or how they have no
value because a condition is raised.
This section refers to the DATATYPE built-in function when checking operands, see nnn. Except for
considerations of limits on the values of exponents, the test:
datatype(Subject) == 'NUM'
is equivalent to testing whether the subject matches the syntax:
num := [blank+] ['+' | '-'] [blank+] number [blank+]
For the syntax of number see nnn.
When the matching subject does not include a '-' the value is the value of the number in the match,
otherwise the value is the value of the expression (0 - number).
The test:
datatype(Subject , 'W')
is a test that the Subject matches that syntax and also has a value that is "whole", that is has no non-zero
fractional part.
When these two tests are made and the Subject matches the constraints but has an exponent that is not
in the correct range of values then a condition is raised:
call #Raise 'SYNTAX', 41.7, Subject
This possibility is implied by the uses of DATATYPE and not shown explicitly in the rest of this section
nnn.
7.1 Variables
The values of variables are held in variable pools. The capabilities of variable pools are listed here,
together with the way each function will be referenced in this definition.
The notation used here is the same as that defined in sections nnn and nnn, including the fact that the
Var_ routines may return an indicator of 'N', 'S' or 'X'.
Each possible name in a variable pool is qualified as tailed or non-tailed name; names with different
qualification and the same spelling are different items in the pool. For those Var_ functions with a third
argument this argument indicates the qualification; it is '1' when addressing tailed names or '0' when
addressing non-tailed names.
Each item in a variable pool is associated with three attributes and a value. The attributes are 'dropped' or
'not-dropped', 'exposed' or 'not-exposed' and 'implicit' or 'not-implicit'.
A variable pool is associated with a reference denoted by the first argument, with name Pool. The value
of Pool may alter during execution. The same name, in conjunction with different values of Pool, can
correspond to different values.
7.1.1 Var_Empty
Var_Empty(Pool)
The function sets the variable pool associated with the specified reference to the state where every name
is associated with attributes 'dropped', 'implicit' and 'not-exposed'.
7.1.2 Var_Set
Var_Set(Pool, Name, '0', Value)
The function operates on the variable pool with the specified reference. The name is a non-tailed name. If
the specified name has the 'exposed' attribute then Var_Set operates on the variable pool referenced by
#Upper in this pool and this rule is applied to that pool. When the pool with attribute 'not-exposed' for this
name is determined the specified value is associated with the specified name. It also associates the
attributes 'not-dropped' and 'not-implicit'. If that attribute was previously 'not-dropped' then the indicator
returned is 'R'. The name is a stem if it contains just one period, as its rightmost character. When the
name is a stem Var_Set(Pool,TailedName, '1',Value) is executed for all possible valid tailed names which
have Name as their stem, and then those tailed-names are given the attribute 'implicit'.
Var_Set(Pool, Name, '1', Value)
The function operates on the variable pool with the specified reference. The name is a tailed name. The
left part of the name, up to and including the first period, is the stem. The stem is a non-tailed name. If the
specified stem has the 'exposed' attribute then Var_Set operates on the variable pool referenced by
#Upper in this pool and this rule is applied to that pool. When the pool with attribute 'not-exposed' for the
stem is determined the name is considered in that pool. If the name has the 'exposed' attribute then the

46
variable pool referenced by #Upper in the pool is considered and this rule applied to that pool. When the
pool with attribute 'not-exposed' is determined the specified value is associated with the specified name.
It also associates the attributes 'not-dropped' and 'not-implicit' . If that attribute was previously
'not-dropped' then the indicator returned is 'R'.
7.1.3 Var_Value
Var_Value(Pool, Name, '0')
The function operates on the variable pool with the specified reference. The name is a non-tailed name. If
the specified name has the 'exposed' attribute then Var_Value operates on the variable pool referenced
by #Upper in this pool and this rule is applied to that pool. When the pool with attribute 'not-exposed' for
this name is determined the indicator returned is 'D' if the name has 'dropped' associated, 'N' otherwise.
In the former case #Outcome is set equal to Name, in the latter case #Outcome is set to the value most
recently associated with the name by Var_Set.
Var_Value(Pool, Name, '1')
The function operates on the variable pool with the specified reference. The name is a tailed name. The
left part of the name, up to and including the first period, is the stem. The stem is a non-tailed name. If the
specified stem has the 'exposed' attribute then Var_Value operates on the variable pool referenced by
#Upper in this pool and this rule is applied to that pool. When the pool with attribute 'not-exposed' for the
stem is determined the name is considered in that pool. If the name has the 'exposed' attribute then the
variable pool referenced by #Upper in the pool is considered and this rule applied to that pool. When the
pool with attribute 'not-exposed' is determined the indicator returned is 'D' if the name has 'dropped'
associated, 'N' otherwise. In the former case #Outcome is set equal to Name, in the latter case #Outcome
is set to the value most recently associated with the name by Var_Set.
7.1.4 Var_Drop
Var_Drop(Pool, Name, '0')
The function operates on the variable pool with the specified reference. The name is a non-tailed name. If
the specified name has the 'exposed' attribute then Var_Drop operates on the variable pool referenced by
#Upper in this pool and this rule is applied to that pool. When the pool with attribute 'not-exposed' for this
name is determined the attribute 'dropped' is associated with the specified name. Also, when the name is
a stem, Var_Drop(Pool,TailedName,'1') is executed for all possible valid tailed names which have Name
as a stem.
Var_Drop(Pool, Name, '1')
The function operates on the variable pool with the specified reference. The name is a tailed name. The
left part of the name, up to and including the first period, is the stem. The stem is a non-tailed name. If the
specified stem has the 'exposed' attribute then Var_Drop operates on the variable pool referenced by
#Upper in this pool and this rule is applied to that pool. When the pool with attribute 'not-exposed' for the
stem is determined the name is considered in that pool. If the name has the 'exposed' attribute then the
variable pool referenced by #Upper in the pool is considered and this rule applied to that pool. When the
pool with attribute 'not-exposed' is determined the attribute 'dropped' is associated with the specified
name.
7.1.5 Var_Expose
Var_Expose(Pool, Name, '0')
The function operates on the variable pool with the specified reference. The name is a non-tailed name.
The attribute 'exposed' is associated with the specified name. Also, when the name is a stem,
Var_Expose(Pool,TailedName,'1') is executed for all possible valid tailed names which have Name as a
stem.
Var_Expose(Pool, Name, '1')
The function operates on the variable pool with the specified reference. The name is a tailed name. The
attribute 'exposed' is associated with the specified name.
7.1.6 Var_Reset
Var_Reset(Pool)
The function operates on the variable pool with the specified reference. It establishes the effect of
subsequent API_Next and API_NextVariable functions (see sections nnn and nnn). A Var_Reset is
implied by any API_ operation other than API_Next and API_NextVariable.
7.2 Symbols
For the syntax of a symbol see nnn.
The value of a symbol which is a NUMBER or a CONST_SYMBOL which is not a reserved symbol is the
content of the appropriate token.

47
The value of a VAR_SYMBOL which is "taken as a constant" is the VAR_SYMBOL itself, otherwise the
VAR_SYMBOL identifies a variable and its value may vary during execution.
Accessing the value of a symbol which is not "taken as a constant" shall result in trace output, see nnn:
if #Tracing.#Level == 'I' then call #Trace Tag
where Tag is '>L>' unless the symbol is a VAR_SYMBOL which, when used as an argument to
Var_Value, does not yield an indicator 'D'. In that case, the Tag is '>V>'.
7.3 Value of a variable
If VAR_SYMBOL does not contain a period, or contains only one period as its last character, the value of
the variable is the value associated with VAR_SYMBOL in the variable pool, that is #Outcome after
Var_Value(Pool,VAR_SYMBOL,'0')
If the indicator is 'D', indicating the variable has the 'dropped' attribute, the NOVALUE condition is raised;
see nnn and nnn for exceptions to this.
#Response = Var_Value(Pool, VAR_SYMBOL, '0')
if left(#Response,1) == 'D' then
call #Raise 'NOVALUE', VAR_SYMBOL, ''
If VAR_SYMBOL contains a period which is not its last character, the value of the variable is the value
associated with the derived name.
7.3.1 Derived names
A derived name is derived from a VAR_SYMBOL as follows:
VAR_SYMBOL := Stem Tail
Stem := PlainSymbol '.'
Tail := (PlainSymbol | '.' [PlainSymbol]) ['.' [PlainSymbol]]+
PlainSymbol := (general_letter | digit)+
The derived name is the concatenation of:
- the Stem, without further evaluation;
- the Tail, with the PlainSymbols replaced by the values of the symbols. The value of a PlainSymbol
which does not start with a digit is #Outcome after
Var_Value(Pool,PlainSymbol,'0')
These values are obtained without raising the NOVALUE condition.
If the indicator from the Var_Value was not 'D' then:
if #Tracing.#Level == 'I' then call #Trace '>C>'
The value associated with a derived name is obtained from the variable pool, that is #Outcome after:
Var_Value(Pool,Derived_Name,'1')
If the indicator is 'D', indicating the variable has the 'dropped' attribute, the NOVALUE condition is raised;
see nnn for an exception.
7.3.2 Value of a reserved symbol
The value of a reserved symbol is the value of a variable with the corresponding name in the reserved
pool, see nnn.
7.4 Expressions and operators
Add a load of string coercions. Equality can operate on non-strings. What if one operand non-string?
7.4.1 The value of a term
See nnn for the syntax of a term.
The value of a STRING is the content of the token; see nnn.
The value of a function is the value it returns, see nnn.
If a term is a symbol or STRING then the value of the term is the value of that symbol or STRING.
If a term contains an expr_alias the value of the term is the value of the expr_alias, see nnn.
7.4.2 The value of a prefix_expression
If the prefix_expression is a term then the value of the prefix_expression is the value of the term,
otherwise let rhs be the value of the prefix_expression within it _ see nnn
If the prefix_expression has the form '+' prefix_expression then a check is made:
if datatype(rhs)\=='NUM' then
call #Raise 'SYNTAX',41.3, rhs, '+'
and the value is the value of (0 + rhs).
If the prefix_expression has the form '-' prefix_expression then a check is made:
if datatype(rhs)\=='NUM' then
call #Raise 'SYNTAX',41.3,rhs, '-'
and the value is the value of (0 - rhs).
If a prefix_expression has the form not prefix_expression then
if rhs \== '0' then if rhs \=='1' then call #Raise 'SYNTAX', 34.6, not, rhs

48
See nnn for the value of the third argument to that #Raise.
If the value of rhs is '0' then the value of the prefix_expression value is '1', otherwise it is '0'.
If the prefix_expression is not a term then:
if #Tracing.#Level == 'I' then call #Trace '>P>'
7.4.3 The value of a power_expression
See nnn for the syntax of a power_expression.
If the power_expression is a prefix_expression then the value of the power_expression is the value of
the prefix_expression.
Otherwise, let lhs be the value of power_expression within it, and rhs be the value of prefix_expression
within it.
if datatype(lhs)\=='NUM' then call #Raise 'SYNTAX',41.1,lhs,'**'
if \datatype(rhs,'W') then call #Raise 'SYNTAX',26.1,rhs,'**'
The value of the power_expression is
ArithOp(lhs,'**',rhs)
If the power_expression is not a prefix_expression then:
if #Tracing.#Level == 'I' then call #Trace '>O>'
7.4.4 The value of a multiplication
See nnn for the syntax of a multiplication.
If the multiplication is a power_expression then the value of the multiplication is the value of the
power_expression.
Otherwise, let lhs be the value of multiplication within it, and rhs be the value of power_expression within
it.
if datatype(lhs)\=='NUM' then
call #Raise 'SYNTAX',41.1,lhs,multiplicative_operation
if datatype(rhs)\=='NUM' then
call #Raise 'SYNTAX',41.2,rhs,multiplicative_operation
The value of the multiplication is
ArithOp(lhs,multiplicative_operation,rhs)
If the multiplication is not a power_expression then:
if #Tracing.#Level == 'I' then call #Trace '>O>'
7.4.5 The value of an addition
See nnn for the syntax of addition.
If the addition is a multiplication then the value of the addition is the value of the multiplication.
Otherwise, let lhs be the value of addition within it, and rhs be the value of the multiplication within it. Let
operation be the additive_operator.
if datatype(lhs)\=='NUM' then
call #Raise 'SYNTAX', 41.1, lhs, operation
if datatype(rhs)\=='NUM' then
call #Raise 'SYNTAX', 41.2, rhs, operation
If either of rhs or lhs is not an integer then the value of the addition is
ArithOp(lhs,operation,rhs)
Otherwise if the operation is '+' and the length of the integer lhs+rhs is not greater than #Digits.#Level
then the value of addition is
lhs+rhs
Otherwise if the operation is '-' and the length of the integer lhs-rhs is not greater than #Digits.#Level then
the value of addition is
lhs-rhs
Otherwise the value of the addition is
ArithOp(lhs,operation,rhs)
If the addition is not a multiplication then:
if #Tracing.#Level == 'I' then call #Trace '>O>'
7.4.6 The value of a concatenation
See nnn for the syntax of a concatenation.
If the concatenation is an addition then the value of the concatenation is the value of the addition.
Otherwise, let lhs be the value of concatenation within it, and rhs be the value of the additive_expression
within it.
If the concatenation contains '||' then the value of the concatenation will have the following characteristics:
- Config_Length(Value) will be equal to Config_Length(lhs)+Config_Length(rhs).
- #Outcome will be 'equal' after each of:

49
 - Config_Compare(Config_Substr(lhs,n),Config_Substr(Value,n)) for values of n not less than 1
and not more than Config_Length(lhs);
- Config_Compare(Config_Substr(rhs,n),Config_Substr(Value,Config_Length(lhs)+n)) for values of
n not less than 1 and not more than Config_Length(rhs).
Otherwise the value of the concatenation will have the following characteristics:
- Config_Length(Value) will be equal to Config_Length(lhs)+1+Config_Length(rhs).
- #Outcome will be 'equal' after each of:
- Config_Compare(Config_Substr(lhs,n),Config_Substr(Value,n)) for values of n not less than 1
and not more than Config_Length(lhs);
- Config_Compare(' ',Config_Substr(Value,Config_Length(lhs)+1));
- Config_Compare(Config_Substr(rhs,n),Config_Substr(Value,Config_Length(lhs)+1+n)) for values
of n not less than 1 and not more than Config_Length(rhs).
If the concatenation is not an addition then:
if #Tracing.#Level == 'I' then call #Trace '>O>'
7.4.7 The value of a comparison
See nnn for the syntax of a comparison.
If the comparison is a concatenation then the value of the comparison is the value of the concatenation.
Otherwise, let lhs be the value of the comparison within it, and rhs be the value of the concatenation
within it.
If the comparison has a comparison_operator that is a strict_compare then the variable #Test is set as
follows:
#Test is set to 'E'. Let Length be the smaller of Config_Length(lhs) and Config_Length(rhs). For values of
n greater than 0 and not greater than Length, if any, in ascending order, #Test is set to the uppercased
first character of #Outcome after:
Config_Compare(Config_Substr(lhs),Config_Substr(rhs)).
If at any stage this sets #Test to a value other than 'E' then the setting of #Test is complete. Otherwise, if
Config_Length(lhs) is greater than Config_Length(rhs) then #Test is set to 'G' or if Config_Length(lhs) is
less than Config_Length(rhs) then #Test is set to 'L'.
If the comparison has a comparison_operator that is a normal_compare then the variable #Test is set as
follows:
if datatype(lhs)\== 'NUM' | datatype(rhs)\== 'NUM' then do
/* Non-numeric non-strict comparison */
lhs=strip(lhs, 'B', ' ') /* ExtraBlanks not stripped */
rhs=strip(rhs, 'B', ' ')
if length(lhs)>length(rhs) then rhs=left(rhs,length(lhs))
else lhs=left(lhs,length(rhs))
if lhs>>rhs then #Test='G'
else if lhs<<rhs then #Test='L'
else #Test='E'
end
else do /* Numeric comparison */
if left(-lhs,1) == '-' & left(+rhs,1) \== '-' then #Test='G'
else if left(-rhs,1) == '-' & left(+lhs,1) \== '-' then #Test='L'
else do
Difference=lhs - rhs /* Will never raise an arithmetic condition. */
if Difference > 0 then #Test='G'
else if Difference < 0 then #Test='L'
else #Test='E'
end
end
The value of #Test, in conjunction with the operator in the comparison, determines the value of the
comparison.
The value of the comparison is '1' if
- #Test is 'E' and the operator is one of '=', '==', '>=', '<=', '\>', '\<', '>>=', '<<=', '\>>', or '\<<';
- #Test is 'G' and the operator is one of '>', '>=', '\<', '\=', '<>', '><', '\==', '>>', '>>=', or '\<<';
- #Test is 'L' and the operator is one of '<', '<=', '\>', '\=', '<>', '><', '\==', '<<', '<<=', or '\>>'.
In all other cases the value of the comparison is '0'.
If the comparison is not a concatenation then:
if #Tracing.#Level == 'I' then call #Trace '>O>'
7.4.8 The value of an and_expression
See nnn for the syntax of an and_expression.

50
If the and_expression is a comparison then the value of the and_expression is the value of the
comparison.
Otherwise, let lhs be the value of the and_expression within it, and rhs be the value of the comparison
within it.
if lhs \== '0' then if lhs \== '1' then call #Raise 'SYNTAX',34.5,lhs,'&'
if rhs \== '0' then if rhs \== '1' then call #Raise 'SYNTAX',34.6,rhs,'&'
Value='0'
if lhs == '1' then if rhs == '1' then Value='1'
If the and_expression is not a comparison then:
if #Tracing.#Level == 'I' then call #Trace '>O>'
7.4.9 The value of an expression
See nnn for the syntax of an expression.
The value of an expression, or an expr, is the value of the expr_alias within it.
If the expr_alias is an and_expression then the value of the expr_alias is the value of the and_expression.
Otherwise, let lhs be the value of the expr_alias within it, and rhs be the value of the and_expression
within it.
if lhs \== '0' then if lhs \== '1' then
call #Raise 'SYNTAX',34.5,lhs,or_operator
if rhs \== '0' then if rhs \== '1' then
call #Raise 'SYNTAX',34.6,rhs,or_operator
Value='1'
if lhs == '0' then if rhs == '0' then Value='0'
If the or_operator is '&&' then
if lhs == '1' then if rhs == '1' then Value='0'
If the expr_alias is not an and_expression then:
if #Tracing.#Level == 'I' then call #Trace '>O>'
The value of an expression or expr shall be traced when #Tracing.#Level is 'R'. The tag is '>=>' when
the value is used by an assignment and '>>>' when it is not.
if #Tracing.#Level == 'R' then call #Trace Tag
7.4.10 Arithmetic operations
The user of this standard is assumed to know the results of the binary operators '+' and '-' applied to
signed or unsigned integers.
The code of ArithOp itself is assumed to operate under a sufficiently high setting of numeric digits to
avoid exponential notation.
ArithOp:

arg Number1, Operator, Number2

/* The Operator will be applied to Number1 and Number2 under the numeric
settings #Digits.#Level, #Form.#Level, #Fuzz.#Level */

/* The result is the result of the operation, or the raising of a 'SYNTAX' or

'LOSTDIGITS' condition. */

/* Variables with digit 1 in their names refer to the first argument of the
operation. Variables with digit 2 refer to the second argument. Variables
with digit 3 refer to the result. */

/* The quotations and page numbers are from the first reference in
Annex C of this standard. */

/* The operands are prepared first. (Page 130) Function Prepare does this,
separating sign, mantissa and exponent. */

v = Prepare(Number1,#Digits.#Level)
parse var v Sign1 Mantissa1 Exponent1

v = Prepare(Number2,#Digits.#Level)
parse var v Sign2 Mantissa2 Exponent2

/* The calculation depends on the operator. The routines set Sign3

Mantissa3 and Exponent3. */

Comparator = ''
select
when Operator == '*' then call Multiply

51
 when Operator == '/' then call DivType
when Operator == '**' then call Power
when Operator == '%' then call DivType
when Operator == '//' then call DivType
otherwise call AddSubComp
end

call PostOp /* Assembles Number3 */

if Comparator \== '' then do

/* Comparison requires the result of subtraction made into a logical */

/* value. */

t = '0'
select
when left(Number3,1) == '-' then
if wordpos(Comparator,'< <= <> >< \= \>') > 0 then t = '1'
when Number3 \== '0' then
if wordpos(Comparator,'> >= <> >< \= \<') > 0 then t = '1'
otherwise
if wordpos(Comparator,'>= = =< \< \>') > 0 then t = '1'
end
Number3 = t
end

return Number3 /* From ArithOp */

/* Activity before every operation: */

Prepare: /* Returns Sign Mantissa and Exponent */

/* Preparation of operands, Page 130 */
/* "...terms being operated upon have leading zeros removed (noting the
position of any decimal point, and leaving just one zero if all the digits in
the number are zeros) and are then truncated to DIGITS+1 significant digits
(if necessary)..." */

arg Number, Digits

/* Blanks are not significant. */

/* The exponent is separated */
parse upper value space(Number,0) with Mantissa 'E' Exponent
if Exponent == '' then Exponent = '0'

/* The sign is separated and made explicit. */

Sign = '+' /* By default */
if left(Mantissa,1) == '-' then Sign = '-'
if verify(left(Mantissa,1),'+-') = 0 then Mantissa = substr(Mantissa,2)

/* Make the decimal point implicit; remove any actual Point from the
mantissa. */
p = pos('.',Mantissa)
if p > 0 then Mantissa = delstr(Mantissa,p,1)
else p = 1+length(Mantissa)

/* Drop the leading zeros */

do q = 1 to length(Mantissa) - 1
if substr(Mantissa,q,1) \== '0' then leave
p = p - 1
end q
Mantissa = substr(Mantissa,q)

/* Detect if Mantissa suggests more significant digits than DIGITS

caters for. */
do j = Digits+1 to length(Mantissa)
if substr(Mantissa,j,1) \== '0' then call #Raise 'LOSTDIGITS', Number
end j

/* Combine exponent with decimal point position, Page 127 */

/* "Exponential notation means that the number includes a power of ten
following an 'E' that indicates how the decimal point will be shifted. Thus
4E9 is just a shorthand way of writing 4000000000 " */

52
 /* Adjust the exponent so that decimal point would be at right of
the Mantissa. */
Exponent = Exponent - (length(Mantissa) - p + 1)

/* Truncate if necessary */
t = length(Mantissa) - (Digits+1)
if t > 0 then do
Exponent = Exponent + t
Mantissa = left(Mantissa,Digits+1)
end

if Mantissa == '0' then Exponent = 0

return Sign Mantissa Exponent

/* Activity after every operation. */

/* The parts of the value are composed into a single string, Number3. */

PostOp:
/* Page 130 */
/* 'traditional' rounding */
t = length(Mantissa3) - #Digits.#Level
if t > 0 then do
/* 'traditional' rounding */
Mantissa3 = left(Mantissa3,#Digits.#Level+1) + 5
if length(Mantissa3) > #Digits.#Level+1 then
/* There was 'carry' */
Exponent3 = Exponent3 + 1
Mantissa3 = left(Mantissa3,#Digits.#Level)
Exponent3 = Exponent3 + t
end
/* "A result of zero is always expressed as a single character '0' "*/
if verify(Mantissa3,'0') = 0 then Number3 = '0'
else do
if Operator == '/' | Operator == '**' then do
/* Page 130 "For division, insignificant trailing zeros are removed
after rounding." */
/* Page 133 "... insignificant trailing zeros are removed." */
do q = length(Mantissa3) by -1 to 2
if substr(Mantissa3,q,1) \== '0' then leave
Exponent3 = Exponent3 + 1
end q
Mantissa3 = substr(Mantissa3,1,q)
end

if Floating() == 'E' then do /* Exponential format */

Exponent3 = Exponent3 + (length(Mantissa3)-1)

/* Page 136 "Engineering notation causes powers of ten to be expressed as a

multiple of 3 - the integer part may therefore range from 1 through
999." */
g = 1
if #Form.#Level == 'E' then do
/* Adjustment to make exponent a multiple of 3 */
g = Exponent3//3 /* Recursively using ArithOp as
an external routine. */
if g < 0 then g = g + 3
Exponent3 = Exponent3 - g
g = g + 1
if length(Mantissa3) < g then
Mantissa3 = left(Mantissa3,g,'0')
end /* Engineering */

/* Exact check on the exponent. */

if Exponent3 > #Limit_ExponentDigits then
call #Raise 'SYNTAX', 42.1, Number1, Operator, Number2
if -#Limit_ExponentDigits > Exponent3 then
call #Raise 'SYNTAX', 42.2, Number1, Operator, Number2

/* Insert any decimal [point. */

53
 if length(Mantissa3) \= g then Mantissa3 = insert('.',Mantissa3,g)
/* Insert the E */
if Exponent3 >= 0 then Number3 = Mantissa3'E+'Exponent3
else Number3 = Mantissa3'E'Exponent3
end /* Exponent format */
else do /* 'pure number' notation */
p = length(Mantissa3) + Exponent3 /* Position of the point within
Mantissa */
/* Add extra zeros needed on the left of the point. */
if p < 1 then do
Mantissa3 = copies('0',1 - p)||Mantissa3
p = 1
end
/* Add needed zeros on the right. */
if p > length(Mantissa3) then
Mantissa3 = Mantissa3||copies('0',p-length(Mantissa3))
/* Format with decimal point. */
Number3 = Mantissa3
if p < length(Number3) then Number3 = insert('.',Mantissa3,p)
else Number3 = Mantissa3
end /* pure */
if Sign3 == '-' then Number3 = '-'Number3
end /* Non-Zero */
return

/* This tests whether exponential notation is needed. */

Floating:
/* The rule in the reference has been improved upon. */
t = ''
if Exponent3+length(Mantissa3) > #Digits.#Level then t = 'E'

if length(Mantissa3) + Exponent3 < -5 then t = 'E'

return t

/* Add, Subtract and Compare. */

AddSubComp: /* Page 130 */

/* This routine is used for comparisons since comparison is
defined in terms of subtraction. Page 134 */
/* "Numeric comparison is affected by subtracting the two numbers(calculating
the difference) and then comparing the result with '0'." */
NowDigits = #Digits.#Level
if Operator \=='+' & Operator \== '-' then do
Comparator = Operator
/* Page 135 "The effect of NUMERIC FUZZ is to temporarily reduce the value
of NUMERIC DIGITS by the NUMERIC FUZZ value for each numeric comparison" */
NowDigits = NowDigits - #Fuzz.#Level
end

/* Page 130 "If either number is zero then the other number ... is used as
the result (with sign adjustment as appropriate). */
if Mantissa2 == '0' then do /* Result is the 1st operand */
Sign3=Sign1; Mantissa3 = Mantissa1; Exponent3 = Exponent1
return ''
end

if Mantissa1 == '0' then do /* Result is the 2nd operand */

Sign3 = Sign2; Mantissa3 = Mantissa2; Exponent3 = Exponent2
if Operator \== '+' then if Sign3 = '+' then Sign3 = '-'
else Sign3 = '+'
return ''
end

/* The numbers may need to be shifted into alignment. */

/* Change to make the exponent to reflect a decimal point on the left,
so that right truncation/extension of mantissa doesn't alter exponent. */
Exponent1 = Exponent1 + length(Mantissa1)
Exponent2 = Exponent2 + length(Mantissa2)
/* Deduce the implied zeros on the left to provide alignment. */

54
 Align1 = 0
Align2 = Exponent1 - Exponent2
if Align2 > 0 then do /* Arg 1 provides a more significant digit */
Align2 = min(Align2,NowDigits+1) /* No point in shifting further. */
/* Shift to give Arg2 the same exponent as Arg1 */
Mantissa2 = copies('0',Align2) || Mantissa2
Exponent2 = Exponent1
end
if Align2 < 0 then do /* Arg 2 provides a more significant digit */
/* Shift to give Arg1 the same exponent as Arg2 */
Align1 = -Align2
Align1 = min(Align1,NowDigits+1) /* No point in shifting further. */
Align2 = 0
Mantissa1 = copies('0',Align1) || Mantissa1
Exponent1 = Exponent2
end

/* Maximum working digits is NowDigits+1. Footnote 41. */

SigDigits = max(length(Mantissa1),length(Mantissa2))
SigDigits = min(SigDigits,NowDigits+1)

/* Extend a mantissa with right zeros, if necessary. */

Mantissa1 = left(Mantissa1,SigDigits,'0')
Mantissa2 = left(Mantissa2,SigDigits,'0')

/* The exponents are adjusted so that

the working numbers are integers, ie decimal point on the right. */
Exponent3 = Exponent1-SigDigits
Exponent1 = Exponent3
Exponent2 = Exponent3

if Operator = '+' then

Mantissa3 = (Sign1 || Mantissa1) + (Sign2 || Mantissa2)
else Mantissa3 = (Sign1 || Mantissa1) - (Sign2 || Mantissa2)

/* Separate the sign */

if Mantissa3 < 0 then do
Sign3 = '-'
Mantissa3 = substr(Mantissa3,2)
end
else Sign3 = '+'

/* "The result is then rounded to NUMERIC DIGITS digits if necessary,

taking into account any extra (carry) digit on the left after addition,
but otherwise counting from the position corresponding to the most
significant digit of the terms being added or subtracted." */

if length(Mantissa3) > SigDigits then SigDigits = SigDigits+1

d = SigDigits - NowDigits /* Digits to drop. */
if d <= 0 then return
t = length(Mantissa3) - d /* Digits to keep. */
/* Page 130. "values of 5 through 9 are rounded up, values of 0 through 4 are
rounded down." */
if t > 0 then do
/* 'traditional' rounding */
Mantissa3 = left(Mantissa3, t + 1) + 5
if length(Mantissa3) > t+1 then
/* There was 'carry' */
/* Keep the extra digit unless it takes us over the limit. */
if t < NowDigits then t = t+1
else Exponent3 = Exponent3+1
Mantissa3 = left(Mantissa3,t)
Exponent3 = Exponent3 + d
end /* Rounding */
else Mantissa3 = '0'
return /* From AddSubComp */

/* Multiply operation: */

Multiply: /* p 131 */

55
/* Note the sign of the result */
if Sign1 == Sign2 then Sign3 = '+'
else Sign3 = '-'
/* Note the exponent */
Exponent3 = Exponent1 + Exponent2
if Mantissa1 == '0' then do
Mantissa3 = '0'
return
end
/* Multiply the Mantissas */
Mantissa3 = ''
do q=1 to length(Mantissa2)
Mantissa3 = Mantissa3'0'
do substr(Mantissa2,q,1)
Mantissa3 = Mantissa3 + Mantissa1
end
end q
return /* From Multiply */

/* Types of Division: */

DivType: /* p 131 */
/* Check for divide-by-zero */
if Mantissa2 == '0' then call #Raise 'SYNTAX', 42.3
/* Note the exponent of the result */
Exponent3 = Exponent1 - Exponent2
/* Compute (one less than) how many digits will be in the integer
part of the result. */
IntDigits = length(Mantissa1) - Length(Mantissa2) + Exponent3
/* In some cases, the result is known to be zero. */
if Mantissa1 = 0 | (IntDigits < 0 & Operator = '%') then do
Mantissa3 = 0
Sign3 = '+'
Exponent3 = 0
return
end
/* In some cases, the result is known to be to be the first argument. */
if IntDigits < 0 & Operator == '//' then do
Mantissa3 = Mantissa1
Sign3 = Sign1
Exponent3 = Exponent1
return
end
/* Note the sign of the result. */
if Sign1 == Sign2 then Sign3 = '+'
else Sign3 = '-'
/* Make Mantissa1 at least as large as Mantissa2 so Mantissa2 can be
subtracted without causing leading zero to result. Page 131 */
a = 0
do while Mantissa2 > Mantissa1
Mantissa1 = Mantissa1'0'
Exponent3 = Exponent3 - 1
a = a + 1
end
/* Traditional divide */
Mantissa3 = ''
/* Subtract from part of Mantissa1 that has length of Mantissa2 */
x = left(Mantissa1,length(Mantissa2))
y = substr(Mantissa1,length(Mantissa2)+1)
do forever
/* Develop a single digit in z by repeated subtraction. */
z = 0
do forever
x = x - Mantissa2
if left(x,1) == '-' then leave
z = z + 1
end
x = x + Mantissa2 /* Recover from over-subtraction */
/* The digit becomes part of the result */
Mantissa3 = Mantissa3 || z
if Mantissa3 == '0' then Mantissa3 = '' /* A single leading

56
 zero can happen. */
/* x||y is the current residue */
if y == '' then if x = 0 then leave /* Remainder is zero */
if length(Mantissa3) > #Digits.#Level then leave /* Enough digits
in the result */
/* Check type of division */
if Operator \== '/' then do
if IntDigits = 0 then leave
IntDigits = IntDigits - 1
end
/* Prepare for next digit */
/* Digits come from y, until that is exhausted. */
/* When y is exhausted an extra zero is added to Mantissa1 */
if y == '' then do
y = '0'
Exponent3 = Exponent3 - 1
a = a + 1
end
x = x || left(y,1)
y = substr(y,2)
end /* Iterate for next digit. */
Remainder = x || y
Exponent3 = Exponent3 + length(y) /* The loop may have been left early. */
/* Leading zeros are taken off the Remainder. */
do while length(Remainder) > 1 & Left(Remainder,1) == '0'
Remainder = substr(Remainder,2)
end
if Operator \== '/' then do
/* Check whether % would fail, even if operation is // */
/* Page 133. % could fail by needing exponential notation */
if Floating() == 'E' then do
if Operator == '%' then MsgNum = 26.11
else MsgNum = 26.12
call #Raise 'SYNTAX', MsgNum, Number1 , Number2, #Digits.#Level
end
end
if Operator == '//' then do
/* We need the remainder */
Sign3 = Sign1
Mantissa3 = Remainder
Exponent3 = Exponent1 - a
end
return /* From DivType */

/* The Power operation: */

Power: /* page 132 */

/* The second argument should be an integer */
if \WholeNumber2() then call #Raise 'SYNTAX', 26.8, Number2
/* Lhs to power zero is always 1 */
if Mantissa2 == '0' then do
Sign3 = '+'
Mantissa3 = '1'
Exponent3 = '0'
return
end

/* Pages 132-133 The Power algorithm */

Rhs = left(Mantissa2,length(Mantissa2)+Exponent2,'0')/* Explicit
integer form */
L = length(Rhs)
b = X2B(D2X(Rhs)) /* Makes Rhs in binary notation */
/* Ignore initial zeros */
do q = 1 by 1
if substr(b,q,1) \== '0' then leave
end q
a = 1
do forever
/* Page 133 "Using a precision of DIGITS+L+1" */
if substr(b,q,1) == '1' then do
a = Recursion('*',Sign1 || Mantissa1'E'Exponent1)

57
 if left(a,2) == 'MN' then signal PowerFailed
end
/* Check for finished */
if q = length(b) then leave
/* Square a */
a = Recursion('*',a)
if left(a,2) == 'MN' then signal PowerFailed
q = q + 1
end
/* Divide into one for negative power */
if Sign2 == '-' then do
Sign2 = '+'
a = Recursion('/')
if left(a,2) == 'MN' then signal PowerFailed
end
/* Split the value up so that PostOp can put it together with rounding */
Parse value Prepare(a,#Digits.#Level+L+1) with Sign3 Mantissa3 Exponent3
return
PowerFailed:
/* Distinquish overflow and underflow */
RcWas = substr(a,4)
if Sign2 = '-' then if RcWas == '42.1' then RcWas = '42.2'
else RcWas = '42.1'
call #Raise 'SYNTAX', RcWas, Number1, '**', Number2
/* No return */

WholeNumber2:
numeric digits Digits
if #Form.#Level == 'S' then numeric form scientific
else numeric form engineering
return datatype(Number2,'W')

Recursion: /* Called only from '**' */

numeric digits #Digits.#Level + L + 1
signal on syntax name Overflowed
/* Uses ArithOp again under new numeric settings. */
if arg(1) == '/' then return 1 / a
else return a * arg(2)
Overflowed:
return 'MN '.MN

7.5 Functions
7.5.1 Invocation
Invocation occurs when a function or a message_term or a call is evaluated. Invocation of a function
may result in a value, in which case:
if #Tracing.#Level == 'I' then call #Trace '>F>'
Invocation of a message_term may result in a value, in which case:
if #Tracing.#Level == 'I' then call #Trace '>M>'
7.5.2 Evaluation of arguments
The argument positions are the positions in the expression_list where syntactically an expression occurs
or could have occurred. Let ArgNumber be the number of an argument position, counting from 1 at the
left; the range of ArgNumber is all whole numbers greater than zero.
For each value of ArgNumber, #ArgExists.#NewLevel.ArgNumber is set '1' if there is an expression
present, '0' if not.
From the left, if #ArgExists.#NewLevel.ArgNumber is '1' then #Arg.#NewLevel.ArgNumber is set to the
value of the corresponding expression. If #ArgExists.#NewLevel.ArgNumber is '0' then
#Arg.#NewLevel.ArgNumber is set to the null string.
#ArgExists.#NewLevel.0 is set to the largest ArgNumber for which #ArgExists.#NewLevel.ArgNumber is
'1', or to zero if there is no such value of ArgNumber.
7.5.3 The value of a label
The value of a LABEL, or of the taken_constant in the function or call_instruction, is taken as a constant,
see nnn. If the taken_constant is not a string_literal it is a reference to the first LABEL in the program
which has the same value. The comparison is made with the '==' operator.
If there is such a matching label and the label is trace-only (see nnn) then a condition is raised:
call #Raise 'SYNTAX', 16.3, taken_constant

58
If there is such a matching label, and the label is not trace-only, execution continues at the label with
routine initialization (see nnn). This is execution of an internal routine.
If there is no such matching label, or if the taken_constant is a string_literal, further comparisons are
made.
If the value of the taken_constant matches the name of some built-in function then that built-in function is
invoked. The names of the built-in functions are defined in section nnn and are in uppercase.
If the value does not match any built-in function name, Config_ExternalRoutine is used to invoke an
external routine.
Whenever a matching label is found, the variables SIGL and .SIGL are assigned the value of the line
number of the clause which caused the search for the label. In the case of an invocation resulting from a
condition occurring that shall be the clause in which the condition occurred.
Var_Set(#Pool, 'SIGL', '0', #LineNumber)
Var_Set(0 , '.SIGL', '0', #LineNumber)
The name used in the invocation is held in #Name.#Level for possible use in an error message from the
RETURN clause, see nnn
7.5.4 The value of a function
A built-in function completes when it returns from the activity defined in section nnn. The value of a
built-in function is defined in section nnn.
An internal routine completes when #Level returns to the value it had when the routine was invoked. The
value of the internal function is the value of the expression on the return which completed the routine.
The value of an external function is determined by Config_ExternalRoutine.
7.5.5 The value of a method
A built-in method completes when it returns from the activity defined in section n. The value of a built-in
method is defined in section n.
An internal method completes when #Level returns to the value it had when the routine was invoked. The
value of the internal method is the value of the expression on the return which completed the method.
The value of an external method is determined by Config_ExternalMethod.
7.5.6 The value of a message term
See nnn for the syntax of a message_term. The value of the term within a message_term is called the
receiver.
The receiver and any arguments of the term are evaluated, in left to right order.
r = #evaluate(message_term, term)
If the message term contains '~~' the value of the message term is the receiver.
Any effect on .Result?
Otherwise the value of a message_term is the value of the method it invokes. The method invoked is
determined by the receiver and the taken_constant and symbol.
t = #Instance(message_term, taken_constant)
If there is a symbol, it is subject to a constraints.
if #contains(message_term, symbol) then do
if r <> #Self then
call #Raise 'SYNTAX', nn.n
/* OOI: "Message search overrides can only be used from methods of the target
object." */

The search will progress from the object to its class and superclasses.
/* This is going to be circular because it describes the message lookup
algorithm and also uses messages. However for the messages in this code
the message names are chosen to be unique to a method so there is no need
to use this algorithm in deciding which method is intended. */

/* message_term ::= receiver '~' taken_constant ':' VAR_SYMBOL arguments */

/* This code reflects OOI - the arguments on the message don't affect
the method choice. */

/* This code selects a method based on its arguments, receiver,

taken_constant, and symbol. */

/* This code is used in a context where #Self is the receiver of the

method invocation which the subject message_term is running under. */

59
SelectMethod:

/* If symbol given, receiver must be self. */

if arg(3,'E') then if arg(1)\==#Self then signal error /* syntax number? */

t = arg(2) /* Will have been uppercased, unless a literal. */

x = arg(1) /* Cursor through places to look for the method. */
Mixing = 1 /* Off for potential mixins ignored because symbol given. */
Mixins = .array~new /* to note any Mixins involved. */

/* Look in the method table of the object, if no 'symbol' given. */

if arg(3,'E') then do
Mixing = 0
end
else do
m = x~#MethodTable[t]
if m \== .nil then return m
end

do until x==.object
/* Follow the class hierarchy. */
x = x~class
/* Note any mixins for later reference. */
Mix = x~Inherited /* An array, ordered as the directive left-to-right. */
if Mix \== .nil then /* Append to the record. */
do j=1 to Mix~dimension(1)
Mixins[Mixins~dimension(1)+1] = Mix[j]
end

if Mixing do
/* Consider mixins only for superclasses of 'symbol'. */
do j=1 to Mixins~dimension(1)
/* Look at the baseclass of each. */
/* That is closest superclass not a mixin. */
s = Mixins[j]~class
do while s~Mixin /* Assert stop at .object if not before. */
s=s~class
end
if s==x then do
m=Mixins[j]~#MethodTable[t]
if m \== .nil then return m
end
end j
end /* Mixing */
if arg(3,'E') then if arg(3)==x then do
Mixing=1
end
if Mixing do
/* Consider non-Mixins */
m = x~#InstanceMethodTable[t]
if m \== .nil then return m
end
x=x~superclass
end

/* Try for UNKNOWN instead */

if t == 'UNKNOWN' then return .nil
if \arg(3,'E') then return = SelectMethod arg(1),'UNKNOWN'
else return = SelectMethod arg(1),'UNKNOWN',arg(3)

7.1.1 Use of Config_ExternalRoutine

The values of the arguments to the use of Config_ExternalRoutine, in order, are:
The argument How is 'SUBROUTINE' if the invocation is from a call, 'FUNCTION' if the invocation is from
a function.
The argument NameType is '1' if the taken_constant is a string_literal, '0' otherwise.
The argument Name is the value of the taken_constant.
The argument Environment is the value of this argument on the API_Start which started this execution.
The argument Arguments is the #Arg. and #ArgExists. data.

60
The argument Streams is the value of this argument on the API_Start which started this execution.
The argument Traps is the value of this argument on the API_Start which started this execution.
Var_Reset is invoked and #API_Enabled set to '1' before use of Config_ExternalRoutine. #API_Enabled
is set to '0' after.
The response from Config_ExternalRoutine is processed. If no conditions are (implicitly) raised,
#Outcome is the value of the function.

61
8 Directives
The syntax constructs which are introduced by the optional '::' token are known as directives.
8.1 Notation
Notation functions are functions which are not directly accessible as functions in a program but are used
in this standard as a notation for defining semantics.
Some notation functions allow reference to syntax constructs defined in nnn. Which instance of the
syntax construct in the program is being referred to is implied; it is the one for which the semantics are
being specified.
The BNF_primary referenced may be directly in the production or in some component referenced in the
production, recursively. The components are considered in left to right order.
#Contains(Identifier, BNF_primary)
where:
Identifier is an identifier in a production (see nnn) defined in nnn.
BNF_primary is a bnf_primary (see nnn) in a production defined in nnn.
Return '1' if the production identified by Identifier contained a bnf_primary identified by BNF_primary,
otherwise return '0'.

#Instance(Identifier, BNF_primary)
where:
Identifier is an identifier in a production defined in nnn.
BNF_primary is a bnf_primary in a production defined in nnn.
Returns the content of the particular instance of the BNF_primary. If the BNF_primary is a
VAR_SYMBOL this is referred to as the symbol "taken as a constant."

#Evaluate(Identifier, BNF_primary)
where:
Identifier is an identifier in a production defined in nnn.
BNF_primary is a bnf_primary in a production defined in nnn.
Return the value of the BNF_primary in the production identified by Identifier.

#Execute(Identifier, BNF_primary)
where:
Identifier is an identifier in a production defined in nnn.
BNF_primary is a bnf_primary in a production defined in nnn.
Perform the instructions identified by the BNF_primary in the production identified by Identifier.

#Parses(Value, BNF_primary)
where:
Value is a string
BNF_primary is a bnf_primary in a production defined in nnn.
Return '1' if Value matches the definition of the BNF_primary, by the rules of clause 6, '0' otherwise.

#Clause(Label)
where:
Label is a label in code used by this standard to describe processing.
Return an identification of that label. The value of this identification is used only by the #Goto notation
function.

#Goto(Value)
where:
Value identifies a label in code used by this standard to describe processing.
The description of processing continues at the identified label.

#Retry()

62
This notation is used in the description of interactive tracing to specify re-execution of the clause just
previously executed. It has the effect of transferring execution to the beginning of that clause, with state
variable #Loop set to the value it had when that clause was previously executed.
8.2 Initializing
Some of the initializing, now grouped in classic section 8.2.1 will have to come here so that we have picked up
anything from the START_API that needs to be passed on to the execution of REQUIRES subject.
We will be using some operations that are forward reference to what was section nnn.
8.2.1 Program initialization and message texts
Processing of a program begins when API_Start is executed. A pool becomes current for the reserved
variables.
call Config_ObjectNew
#ReservedPool = #Outcome
#Pool = #ReservedPool
Is it correct to make the reserved variables and the builtin objects in the same pool?

Some of the values which affect processing of the program are parameters of API_Start:
#HowInvoked is set to 'COMMAND', 'FUNCTION' or 'SUBROUTINE' according to the first parameter of
API_Start.
#Source is set to the value of the second parameter of API_Start.
The third parameter of API_Start is used to determine the initial active environment.
The fourth parameter of API_Start is used to determine the arguments. For each argument position
#ArgExists.1.ArgNumber is set '1' if there is an argument present, '0' if not. ArgNumber is the number of
the argument position, counting from 1. If #ArgExists.1.ArgNumber is '1' then #Arg.1.ArgNumber is set to
the value of the corresponding argument. If #ArgExists.1.ArgNumber is '0' then #Arg.1.Arg is set to the
null string. #ArgExists.1.0 is set to the largest n for which #ArgExists.1.n is '1', or to zero if there is no
such value of n.
Some of the values which affect processing of the program are provided by the configuration:
call Config_OtherBlankCharacters
#AllBlanks<Index "#AllBlanks" # "" > = ' '#Outcome /* "Real" blank concatenated with
others */
#Bif_Digits. = 9
call Config_Constants
.true = '1'
.false = '0'
Objects in our model are only distinquished by the values within their pool so we can construct the builtin classes
incomplete and then complete them with directives.
Can we initialize the methods of .nil by directives?
call Config_ObjectNew
.List = #Outcome
call var_set .List, #IsClass, '0', '1'
call var_set .List, #ID, '0', 'List'
Some of the state variables set by this call are limits, and appear in the text of error messages. The
relation between message numbers and message text is defined by the following list, where the message
number appears immediately before an '=' and the message text follows in quotes.
#ErrorText. = ''

#ErrorText.0.1 = 'Error <value> running <source>, line <linenumber>: '

#ErrorText.0.2 = 'Error <value> in interactive trace: '
#ErrorText.0.3 = 'Interactive trace. "Trace Off" to end debug. ',
'ENTER to continue.'
#ErrorText.2 = 'Failure during finalization'
#ErrorText.2.1 = 'Failure during finalization: <description>'

#ErrorText.3 = 'Failure during initialization'

#ErrorText.3.1 = 'Failure during initialization: <description>'

#ErrorText.4 = 'Program interrupted'

#ErrorText.4.1 = 'Program interrupted with HALT condition: <description>'

#ErrorText.5 = 'System resources exhausted'

#ErrorText.5.1 = 'System resources exhausted: <description>'

#ErrorText.6 = 'Unmatched "/*" or quote'

#ErrorText.6.1 = 'Unmatched comment delimiter ("/*")'

63
 #ErrorText.6.2 = "Unmatched single quote (')"
#ErrorText.6.3 = 'Unmatched double quote (")'

#ErrorText.7 = 'WHEN or OTHERWISE expected'

#ErrorText.7.1 = 'SELECT on line <linenumber> requires WHEN;',
'found "<token>"'
#ErrorText.7.2 = 'SELECT on line <linenumber> requires WHEN, OTHERWISE,',
'or END; found "<token>"'
#ErrorText.7.3 = 'All WHEN expressions of SELECT on line <linenumber> are',
'false; OTHERWISE expected'

#ErrorText.8 = 'Unexpected THEN or ELSE'

#ErrorText.8.1 = 'THEN has no corresponding IF or WHEN clause'
#ErrorText.8.2 = 'ELSE has no corresponding THEN clause'

#ErrorText.9 = 'Unexpected WHEN or OTHERWISE'

#ErrorText.9.1 = 'WHEN has no corresponding SELECT'
#ErrorText.9.2 = 'OTHERWISE has no corresponding SELECT'

#ErrorText.10 = 'Unexpected or unmatched END'

#ErrorText.10.1= 'END has no corresponding DO or SELECT'
#ErrorText.10.2= 'END corresponding to DO on line <linenumber>',
'must have a symbol following that matches',
'the control variable (or no symbol);',
'found "<token>"'
#ErrorText.10.3= 'END corresponding to DO on line <linenumber>',
'must not have a symbol following it because',
'there is no control variable;',
'found "<token>"'
#ErrorText.10.4= 'END corresponding to SELECT on line <linenumber>',
'must not have a symbol following;',
'found "<token>"'
#ErrorText.10.5= 'END must not immediately follow THEN'
#ErrorText.10.6= 'END must not immediately follow ELSE'

#ErrorText.13 = 'Invalid character in program'

#ErrorText.13.1= 'Invalid character "('<hex-encoding>'X)" in program'

#ErrorText.14 = 'Incomplete DO/SELECT/IF'

#ErrorText.14.1= 'DO instruction requires a matching END'
#ErrorText.14.2= 'SELECT instruction requires a matching END'
#ErrorText.14.3= 'THEN requires a following instruction'
#ErrorText.14.4= 'ELSE requires a following instruction'

#ErrorText.15 = 'Invalid hexadecimal or binary string'

#ErrorText.15.1= 'Invalid location of blank in position',
'<position> in hexadecimal string'
#ErrorText.15.2= 'Invalid location of blank in position',
'<position> in binary string'
#ErrorText.15.3= 'Only 0-9, a-f, A-F, and blank are valid in a',
'hexadecimal string; found "<char>"'
#ErrorText.15.4= 'Only 0, 1, and blank are valid in a',
'binary string; found "<char>"'

#ErrorText.16 = 'Label not found'

#ErrorText.16.1= 'Label "<name>" not found'
#ErrorText.16.2= 'Cannot SIGNAL to label "<name>" because it is',
'inside an IF, SELECT or DO group'
#ErrorText.16.3= 'Cannot invoke label "<name>" because it is',
'inside an IF, SELECT or DO group'

#ErrorText.17 = 'Unexpected PROCEDURE'

#ErrorText.17.1= 'PROCEDURE is valid only when it is the first',
'instruction executed after an internal CALL',
'or function invocation'
#ErrorText.17.2= 'The EXPOSE instruction is valid only when it is the first',
'instruction executed after a method invocation'

#ErrorText.18 = 'THEN expected'

#ErrorText.18.1= 'IF keyword on line <linenumber> requires',
'matching THEN clause; found "<token>"'
#ErrorText.18.2= 'WHEN keyword on line <linenumber> requires',

64
 'matching THEN clause; found "<token>"'

#ErrorText.19 = 'String or symbol expected'

#ErrorText.19.1= 'String or symbol expected after ADDRESS keyword;',
'found "<token>"'
#ErrorText.19.2= 'String or symbol expected after CALL keyword;',
'found "<token>"'
#ErrorText.19.3= 'String or symbol expected after NAME keyword;',
'found "<token>"'
#ErrorText.19.4= 'String or symbol expected after SIGNAL keyword;',
'found "<token>"'
#ErrorText.19.6= 'String or symbol expected after TRACE keyword;',
'found "<token>"'
#ErrorText.19.7= 'Symbol expected in parsing pattern;',
'found "<token>"'
#ErrorText.19.8= 'String or symbol expected after REQUIRES;',
'found "<token>"'
#ErrorText.19.9= 'String or symbol expected after METHOD;',
'found "<token>"'
#ErrorText.19.11='String or symbol expected after ROUTINE;',
'found "<token>"'
#ErrorText.19.12='String or symbol expected after CLASS;',
'found "<token>"'
#ErrorText.19.13='String or symbol expected after INHERIT;',
'found "<token>"'
#ErrorText.19.15='String or symbol expected after METACLASS;',
'found "<token>"'
#ErrorText.19.16='String or symbol expected after MIXINCLASS;',
'found "<token>"'
#ErrorText.19.17='String or symbol expected after SUBCLASS;',
'found "<token>"'
Unsound now we are using 'term'?

#ErrorText.20 = 'Name expected'

#ErrorText.20.1= 'Name required; found "<token>"'
#ErrorText.20.2= 'Found "<token>" where only a name is valid'
#ErrorText.20.3= 'Found "<token>" where only a name or '(' is valid'

#ErrorText.21 = 'Invalid data on end of clause'

#ErrorText.21.1= 'The clause ended at an unexpected token;',
'found "<token>"'

#ErrorText.22 = 'Invalid character string'

#ErrorText.22.1= "Invalid character string '<hex-encoding>'X"

#ErrorText.23 = 'Invalid data string'

#ErrorText.23.1= "Invalid data string '<hex-encoding>'X"

#ErrorText.24 = 'Invalid TRACE request'

#ErrorText.24.1= 'TRACE request letter must be one of',
'"ACEFILNOR"; found "<value>"'

#ErrorText.25 = 'Invalid sub-keyword found'

#ErrorText.25.1= 'CALL ON must be followed by one of the',
'keywords <keywords>; found "<token>"'
#ErrorText.25.2= 'CALL OFF must be followed by one of the',
'keywords <keywords>; found "<token>"'
#ErrorText.25.3= 'SIGNAL ON must be followed by one of the',
'keywords <keywords>; found "<token>"'
#ErrorText.25.4= 'SIGNAL OFF must be followed by one of the',
'keywords <keywords>; found "<token>"'
#ErrorText.25.5= 'ADDRESS WITH must be followed by one of the',
'keywords <keywords>; found "<token>"'
#ErrorText.25.6= 'INPUT must be followed by one of the',
'keywords <keywords>; found "<token>"'
#ErrorText.25.7= 'OUTPUT must be followed by one of the',
'keywords <keywords>; found "<token>"'
#ErrorText.25.8= 'APPEND must be followed by one of the',
'keywords <keywords>; found "<token>"'

65
 #ErrorText.25.9= 'REPLACE must be followed by one of the',
'keywords <keywords>; found "<token>"'
#ErrorText.25.11='NUMERIC FORM must be followed by one of the',
'keywords <keywords>; found "<token>"'
#ErrorText.25.12='PARSE must be followed by one of the',
'keywords <keywords>; found "<token>"'
#ErrorText.25.13='UPPER must be followed by one of the',
'keywords <keywords>; found "<token>"'
#ErrorText.25.14='ERROR must be followed by one of the',
'keywords <keywords>; found "<token>"'
#ErrorText.25.15='NUMERIC must be followed by one of the',
'keywords <keywords>; found "<token>"'
#ErrorText.25.16='FOREVER must be followed by one of the',
'keywords <keywords> or nothing; found "<token>"'
#ErrorText.25.17='PROCEDURE must be followed by the keyword',
'EXPOSE or nothing; found "<token>"'
#ErrorText.25.18='FORWARD must be followed by one of the the keywords',
'<keywords>; found "<token>"'
#ErrorText.26 = 'Invalid whole number'
#ErrorText.26.1= 'Whole numbers must fit within current DIGITS',
'setting(<value>); found "<value>"'
#ErrorText.26.2= 'Value of repetition count expression in DO instruction',
'must be zero or a positive whole number;',
'found "<value>"'
#ErrorText.26.3= 'Value of FOR expression in DO instruction',
'must be zero or a positive whole number;',
'found "<value>"'
#ErrorText.26.4= 'Positional pattern of parsing template',
'must be a whole number; found "<value>"'
#ErrorText.26.5= 'NUMERIC DIGITS value',
'must be a positive whole number; found "<value>"'
#ErrorText.26.6= 'NUMERIC FUZZ value',
'must be zero or a positive whole number;',
'found "<value>"'
#ErrorText.26.7= 'Number used in TRACE setting',
'must be a whole number; found "<value>"'
#ErrorText.26.8= 'Operand to right of the power operator ("**")',
'must be a whole number; found "<value>"'
#ErrorText.26.11='Result of <value> % <value> operation would need',
'exponential notation at current NUMERIC DIGITS <value>'
#ErrorText.26.12='Result of % operation used for <value> // <value>',
'operation would need',
'exponential notation at current NUMERIC DIGITS <value>'

#ErrorText.27 = 'Invalid DO syntax'

#ErrorText.27.1= 'Invalid use of keyword "<token>" in DO clause'

#ErrorText.28 = 'Invalid LEAVE or ITERATE'

#ErrorText.28.1= 'LEAVE is valid only within a repetitive DO loop'
#ErrorText.28.2= 'ITERATE is valid only within a repetitive DO loop'
#ErrorText.28.3= 'Symbol following LEAVE ("<token>") must',
'either match control variable of a current',
'DO loop or be omitted'
#ErrorText.28.4= 'Symbol following ITERATE ("<token>") must',
'either match control variable of a current',
'DO loop or be omitted'

#ErrorText.29 = 'Environment name too long'

#ErrorText.29.1= 'Environment name exceeds',
#Limit_EnvironmentName 'characters; found "<name>"'

#ErrorText.30 = 'Name or string too long'

#ErrorText.30.1= 'Name exceeds' #Limit_Name 'characters'
#ErrorText.30.2= 'Literal string exceeds' #Limit_Literal 'characters'

#ErrorText.31 = 'Name starts with number or "."'

#ErrorText.31.1= 'A value cannot be assigned to a number;',
'found "<token>"'
#ErrorText.31.2= 'Variable symbol must not start with a number;',
'found "<token>"'
#ErrorText.31.3= 'Variable symbol must not start with a ".";',
'found "<token>"'

66
 #ErrorText.33 = 'Invalid expression result'
#ErrorText.33.1= 'Value of NUMERIC DIGITS ("<value>")',
'must exceed value of NUMERIC FUZZ "(<value>)"'
#ErrorText.33.2= 'Value of NUMERIC DIGITS ("<value>")',
'must not exceed' #Limit_Digits
#ErrorText.33.3= 'Result of expression following NUMERIC FORM',
'must start with "E" or "S"; found "<value>"'

#ErrorText.34 = 'Logical value not "0" or "1"'

#ErrorText.34.1= 'Value of expression following IF keyword',
'must be exactly "0" or "1"; found "<value>"'
#ErrorText.34.2= 'Value of expression following WHEN keyword',
'must be exactly "0" or "1"; found "<value>"'
#ErrorText.34.3= 'Value of expression following WHILE keyword',
'must be exactly "0" or "1"; found "<value>"'
#ErrorText.34.4= 'Value of expression following UNTIL keyword',
'must be exactly "0" or "1"; found "<value>"'
#ErrorText.34.5= 'Value of expression to left',
'of logical operator "<operator>"',
'must be exactly "0" or "1"; found "<value>"'
#ErrorText.34.6= 'Value of expression to right',
'of logical operator "<operator>"',
'must be exactly "0" or "1"; found "<value>"'

#ErrorText.35 = 'Invalid expression'

#ErrorText.35.1= 'Invalid expression detected at "<token>"'

#ErrorText.36 = 'Unmatched "(" in expression'

#ErrorText.37 = 'Unexpected "," or ")"'

#ErrorText.37.1= 'Unexpected ","'
#ErrorText.37.2= 'Unmatched ")" in expression'

#ErrorText.38 = 'Invalid template or pattern'

#ErrorText.38.1= 'Invalid parsing template detected at "<token>"'
#ErrorText.38.2= 'Invalid parsing position detected at "<token>"'
#ErrorText.38.3= 'PARSE VALUE instruction requires WITH keyword'

#ErrorText.40 = 'Incorrect call to routine'

#ErrorText.40.1= 'External routine "<name>" failed'
#ErrorText.40.3= 'Not enough arguments in invocation of <bif>;',
'minimum expected is <argnumber>'
#ErrorText.40.4= 'Too many arguments in invocation of <bif>;',
'maximum expected is <argnumber>'
#ErrorText.40.5= 'Missing argument in invocation of <bif>;',
'argument <argnumber> is required'
#ErrorText.40.9= '<bif> argument <argnumber>',
'exponent exceeds' #Limit_ExponentDigits 'digits;',
'found "<value>"'
#ErrorText.40.11='<bif> argument <argnumber>',
'must be a number; found "<value>"'
#ErrorText.40.12='<bif> argument <argnumber>',
'must be a whole number; found "<value>"'
#ErrorText.40.13='<bif> argument <argnumber>',
'must be zero or positive; found "<value>"'
#ErrorText.40.14='<bif> argument <argnumber>',
'must be positive; found "<value>"'
#ErrorText.40.17='<bif> argument 1',
'must have an integer part in the range 0:90 and a',
'decimal part no larger than .9; found "<value>"'
#ErrorText.40.18='<bif> conversion must',
'have a year in the range 0001 to 9999'
#ErrorText.40.19='<bif> argument 2, "<value>", is not in the format',
'described by argument 3, "<value>"'
#ErrorText.40.21='<bif> argument <argnumber> must not be null'
#ErrorText.40.23='<bif> argument <argnumber>',
'must be a single character; found "<value>"'
#ErrorText.40.24='<bif> argument 1',
'must be a binary string; found "<value>"'
#ErrorText.40.25='<bif> argument 1',
'must be a hexadecimal string; found "<value>"'

67
 #ErrorText.40.26='<bif> argument 1',
'must be a valid symbol; found "<value>"'
#ErrorText.40.27='<bif> argument 1',
'must be a valid stream name; found "<value>"'
#ErrorText.40.28='<bif> argument <argnumber>,',
'option must start with one of "<optionslist>";',
'found "<value>"'
#ErrorText.40.29='<bif> conversion to format "<value>" is not allowed'
#ErrorText.40.31='<bif> argument 1 ("<value>") must not exceed 100000'
#ErrorText.40.32='<bif> the difference between argument 1 ("<value>") and',
'argument 2 ("<value>") must not exceed 100000'
#ErrorText.40.33='<bif> argument 1 ("<value>") must be less than',
'or equal to argument 2 ("<value>")'
#ErrorText.40.34='<bif> argument 1 ("<value>") must be less than',
'or equal to the number of lines',
'in the program (<sourceline()>)'
#ErrorText.40.35='<bif> argument 1 cannot be expressed as a whole number;',
'found "<value>"'
#ErrorText.40.36='<bif> argument 1',
'must be the name of a variable in the pool;',
'found "<value>"'
#ErrorText.40.37='<bif> argument 3',
'must be the name of a pool; found "<value>"'
#ErrorText.40.38='<bif> argument <argnumber>',
'is not large enough to format "<value>"'
#ErrorText.40.39='<bif> argument 3 is not zero or one; found "<value>"'
#ErrorText.40.41='<bif> argument <argnumber>',
'must be within the bounds of the stream;',
'found "<value>"'
#ErrorText.40.42='<bif> argument 1; cannot position on this stream;',
'found "<value>"'
#ErrorText.40.45='<bif> argument <argnumber> must be a single',
'non-alphanumeric character or the null string; ',
'"found <value>"'
#ErrorText.40.46='<bif> argument 3, "<value>", is a format incompatible',
'with separator specified in argument <argnumber>'

#ErrorText.41 = 'Bad arithmetic conversion'

#ErrorText.41.1= 'Non-numeric value ("<value>")',
'to left of arithmetic operation "<operator>"'
#ErrorText.41.2= 'Non-numeric value ("<value>")',
'to right of arithmetic operation "<operator>"'
#ErrorText.41.3= 'Non-numeric value ("<value>")',
'used with prefix operator "<operator>"'
#ErrorText.41.4= 'Value of TO expression in DO instruction',
'must be numeric; found "<value>"'
#ErrorText.41.5= 'Value of BY expression in DO instruction',
'must be numeric; found "<value>"'
#ErrorText.41.6= 'Value of control variable expression of DO instruction',
'must be numeric; found "<value>"'
#ErrorText.41.7= 'Exponent exceeds' #Limit_ExponentDigits 'digits;',
'found "<value>"'

#ErrorText.42 = 'Arithmetic overflow/underflow'

#ErrorText.42.1= 'Arithmetic overflow detected at',
'"<value> <operation> <value>";',
'exponent of result requires more than',
#Limit_ExponentDigits 'digits'
#ErrorText.42.2= 'Arithmetic underflow detected at',
'"<value> <operation> <value>";',
'exponent of result requires more than',
#Limit_ExponentDigits 'digits'
#ErrorText.42.3= 'Arithmetic overflow; divisor must not be zero'

#ErrorText.43 = 'Routine not found'

#ErrorText.43.1= 'Could not find routine "<name>"'

#ErrorText.44 = 'Function did not return data'

#ErrorText.44.1= 'No data returned from function "<name>"'

#ErrorText.45 = 'No data specified on function RETURN'

#ErrorText.45.1= 'Data expected on RETURN instruction because',

68
 'routine "<name>" was called as a function'

#ErrorText.46 = 'Invalid variable reference'

#ErrorText.46.1= 'Extra token ("<token>") found in variable',
'reference; ")" expected'

#ErrorText.47 = 'Unexpected label'

#ErrorText.47.1= 'INTERPRET data must not contain labels;',
'found "<name>"'

#ErrorText.48 = 'Failure in system service'

#ErrorText.48.1= 'Failure in system service: <description>'

#ErrorText.49 = 'Interpretation Error'

#ErrorText.49.1= 'Interpretation Error: <description>'

#ErrorText.50 = 'Unrecognized reserved symbol'

#ErrorText.50.1= 'Unrecognized reserved symbol "<token>"'

#ErrorText.51 = 'Invalid function name'

#ErrorText.51.1= 'Unquoted function names must not end with a period;',
'found "<token>"'

#ErrorText.52 = 'Result returned by "<name>" is longer than',

#Limit_String 'characters'

#ErrorText.53 = 'Invalid option'

#ErrorText.53.1= 'Variable reference expected',
'after STREAM keyword; found "<token>"'
#ErrorText.53.2= 'Variable reference expected',
'after STEM keyword; found "<token>"'
#ErrorText.53.3= 'Argument to STEM must have one period,',
'as its last character; found "<name>"'
#ErrorText.54 = 'Invalid STEM value'
#ErrorText.54.1= 'For this use of STEM, the value of "<name>" must be a',
'count of lines; found: "<value>"'
If the activity defined by clause 6 does not produce any error message, execution of the program
continues.
call Config_NoSource
If Config_NoSource has set #NoSource to '0' the lines of source processed by clause 6 are copied to
#SourceLine. , with #SourceLine.0 being a count of the lines and #SourceLine.n for n=1 to #SourceLine.0
being the source lines in order.
If Config_NoSource has set #NoSource to '1' then #SourceLine.0 is set to 0.
The following state variables affect tracing:
#InhibitPauses = 0
#InhibitTrace = 0
#AtPause = 0 /* Off until interactive input being received. */
#Trace_QueryPrior = 'No'
An initial variable pool is established:
call Config_ObjectNew
#Pool = #Outcome
#Pool1 = #Pool
call Var_Empty #Pool
call Var_Reset #Pool

#Level = 1 /* Level of invocation */

#NewLevel = 2
#IsFunction.#Level = (#HowInvoked == 'FUNCTION')
For this first level, there is no previous level from which values are inherited. The relevant fields are
initialized.
#Digits.#Level = 9 /* Numeric Digits */
#Form.#Level = 'SCIENTIFIC' /* Numeric Form */
#Fuzz.#Level = 0 /* Numeric Fuzz */

#StartTime.#Level = '' /* Elapsed time boundary */

#LineNumber = ''
#Tracing.#Level = 'N'
#Interactive.#Level = '0'

69
An environment is provided by the API_Start to become the initial active environment to which commands
will be addressed. The alternate environment is made the same:
/* Call the environments ACTIVE, ALTERNATE, TRANSIENT where these are
never-initialized state variables.
Similarly call the redirections I O and E */
call EnvAssign ALTERNATE, #Level, ACTIVE, #Level
Conditions are initially disabled:
#Enabling.SYNTAX.#Level = 'OFF'
#Enabling.HALT.#Level = 'OFF'
#Enabling.ERROR.#Level = 'OFF'
#Enabling.FAILURE.#Level = 'OFF'
#Enabling.NOTREADY.#Level = 'OFF'
#Enabling.NOVALUE.#Level = 'OFF'
#Enabling.LOSTDIGITS.#Level = 'OFF'
#PendingNow.HALT.#Level = 0
#PendingNow.ERROR.#Level = 0
#PendingNow.FAILURE.#Level = 0
#PendingNow.NOTREADY.#Level = 0
/* The following field corresponds to the results from the CONDITION built-in
function. */
#Condition.#Level = ''
The opportunity is provided for a trap to initialize the pool.
#API_Enabled = '1'
call Var_Reset #Pool
call Config_Initialization
#API_Enabled = '0'
8.3 REQUIRES
For each requires in order of appearence:
A use of Start_API with #instance(requires, taken_constant). Msg40.1 or a new if completion 'E'. Add Provides to an
ordered collection. Not cyclic because .LIST can be defined without defining REQUIRES but a fairly profound forward
reference.
8.4 CLASS
For each class in order of appearence:
#ClassName = #Instance(class, taken_constant)
call var_value #ReservedPool, '#CLASSES.'ClassName, '1'
if #Indicator == 'D' then do
call Config_ObjectNew
#Class = #Outcome
call var_set #ReservedPool, '#CLASSES.'ClassName, '1', #Class
end
else call #Raise 'SYNTAX', nn.nn, #ClassName
New instance of CLASS class added to list. Msg "Duplicate ::CLASS directive instruction"(?)
8.5 METHOD
For each method in order of appearence:
call Config_ObjectNew
#Pool = #Outcome
call Config_ObjectSource(#Pool)
#MethodName = #Instance(method, taken_constant)
call var_value #Class, '#METHODS.'#MethodName, '1'
if #Indicator == 'D' then
call var_set #Class, '#METHODS.'#MethodName, '1',#Pool
else call #Raise 'SYNTAX', nn.nn, #MethodName, #ClassName

GUARDED & public is default. if #contains(method, 'PRIVATE') then m~setprivate; if #contains(method,

'UNGUARDED') then m~setunguarded
Why is there a keyword for GUARDED but not for PUBLIC here?
Does CLASS option mean ENHANCE with Class class methods?
#CurrentClass~class(#instance(method, taken_constant), m)
For ATTRIBUTE, should we actually construct source for two methods? ATTRIBUTE case needs test of null body.
OOI doesn't have source (because it actually traps UNKNOWN?).
For EXTERNAL test for null body. Simon Nash doc says "Accessibility to external methods ... is
implementation-defined". Left like that it doesn't even tell us about search order. We will need a
Config_ExternalClass to import the public names of the class.
8.6 ROUTINE

70
For each routine in order of appearence:
Add name (with duplicate check) to list for this file.
Extra step needed in the invokation search order. Although this is nominally EXTERNAL we presumably wont use
the external call mechanism. (Except perhaps when the routine was made available by a REQUIRES; in that case
the PARSE SOURCE answer has to change.)
I have the builtins-defined-by-directives elsewhere; it would make sense if they wound up about here.

71
9 Instructions
This clause describes the execution of instructions, and how the sequence of execution can vary from the
normal execution in order of appearance in the program.

Execution of the program begins with its first clause.

If we left Routine initialization to here we can leave method initialization.
9.1 Method initialization
There is a pool for local variables.
call Config_ObjectNew
#Pool = #Outcome
Set self and super

9.2 Routine initialization

If the routine is invoked as a function, #IsFunction.#NewLevel shall be set to '1', otherwise to '0'; this
affects the processing of a subsequent RETURN instruction.
#AllowProcedure.#NewLevel = '1'
Many of the initial values for a new invocation are inherited from the caller's values.
#Digits.#NewLevel = #Digits.#Level
#Form.#NewLevel = #Form.#Level
#Fuzz.#NewLevel = #Fuzz.#Level

#StartTime.#NewLevel = #StartTime.#Level

#Tracing.#NewLevel = #Tracing.#Level
#Interactive.#NewLevel = #Interactive.#Level

call EnvAssign ACTIVE, #NewLevel, ACTIVE, #Level

call EnvAssign ALTERNATE, #NewLevel, ALTERNATE, #Level

do t=1 to 7
Condition = word('SYNTAX HALT ERROR FAILURE NOTREADY NOVALUE LOSTDIGITS',t)
#Enabling.Condition.#NewLevel = #Enabling.Condition.#Level
#Instruction.Condition.#NewLevel = #Instruction.Condition.#Level
#TrapName.Condition.#NewLevel = #TrapName.Condition.#Level
#EventLevel.Condition.#NewLevel = #EventLevel.Condition.#Level
end t
If this invocation is not caused by a condition occurring, see nnn, the state variables for the CONDITION
built-in function are copied.
#Condition.#NewLevel = #Condition.#Level
#ConditionDescription.#NewLevel = #ConditionDescription.#Level
#ConditionExtra.#NewLevel = #ConditionExtra.#Level
#ConditionInstruction.#NewLevel = #ConditionInstruction.#Level
Execution of the initialized routine continues at the new level of invocation.
#Level = #NewLevel
#NewLevel = #Level + 1

9.3 Clause initialization

The clause is traced before execution:
if pos(#Tracing.#Level, 'AIR') > 0 then call #TraceSource
The time of the first use of DATE or TIME will be retained throughout the clause.
#ClauseTime.#Level = ''
The state variable #LineNumber is set to the line number of the clause, see nnn.
A clause other than a null clause or label or procedure instruction sets:
#AllowProcedure.#Level = '0' /* See message 17.1 */

9.4 Clause termination

if #InhibitTrace > 0 then #InhibitTrace = #InhibitTrace - 1

Polling for a HALT condition occurs:

#Response = Config_Halt_Query()
if #Outcome == 'Yes' then do
call Config_Halt_Reset

72
 call #Raise 'HALT', substr(#Response,2) /* May return */
end
At the end of each clause there is a check for conditions which occurred and were delayed. It is acted on
if this is the clause in which the condition arose.
do t=1 to 4
#Condition=WORD('HALT FAILURE ERROR NOTREADY',t)
/* HALT can be established during HALT handling. */
do while #PendingNow.#Condition.#Level
#PendingNow.#Condition.#Level = '0'
call #Raise
end
end
Interactive tracing may be turned on via the configuration. Only a change in the setting is significant.
call Config_Trace_Query
if #AtPause = 0 & #Outcome == 'Yes' & #Trace_QueryPrior == 'No' then do
/* External request for Trace '?R' */
#Interactive.#Level = '1'
#Tracing.#Level = 'R'
end
#TraceQueryPrior = #Outcome

Tracing just not the same with NetRexx.

When tracing interactively, pauses occur after the execution of each clause except for CALL, DO the
second or subsequent time through the loop, END, ELSE, EXIT, ITERATE, LEAVE, OTHERWISE,
RETURN, SIGNAL, THEN and null clauses.
If the character '=' is entered in response to a pause, the prior clause is re-executed.
Anything else entered will be treated as a string of one or more clauses and executed by the language
processor. The same rules apply to the contents of the string executed by interactive trace as do for
strings executed by the INTERPRET instruction. If the execution of the string generates a syntax error,
the standard message is displayed but no condition is raised. All condition traps are disabled during
execution of the string. During execution of the string, no tracing takes place other than error or failure
return codes from commands. The special variable RC is not set by commands executed within the
string, nor is .RC.
If a TRACE instruction is executed within the string, the language processor immediately alters the trace
setting according to the TRACE instruction encountered and leaves this pause point. If no TRACE
instruction is executed within the string, the language processor simply pauses again at the same point in
the program.
At a pause point:
if #AtPause = 0 & #Interactive.#Level & #InhibitTrace = 0 then do
if #InhibitPauses > 0 then #InhibitPauses = #InhibitPauses-1
else do
#TraceInstruction = '0'
do forever
call Config_Trace_Query
if #Outcome == 'No' & #Trace_QueryPrior == 'Yes' then do
/* External request to stop tracing. */
#Trace_QueryPrior=#Outcome
#Interactive.#Level = '0'
#Tracing.#Level = 'N'
leave
end
if #Outcome == 'Yes' & #Trace_QueryPrior == 'No' then do
/* External request for Trace '?R' */
#Trace_QueryPrior = #Outcome
#Interactive.#Level = '1'
#Tracing.#Level = 'R'
leave
end
if \#Interactive.#Level | #TraceInstruction then leave

/* Accept input for immediate execution. */

call Config_Trace_Input
if length(#Outcome) = 0 | #Outcome == '=' then leave
#AtPause = #Level
interpret #Outcome

73
 #AtPause = 0
end /* forever loop */
if #Outcome == '=' then call #Retry /* With no return */
end
end
9.5 Instruction
9.5.1 ADDRESS
For a definition of the syntax of this instruction, see nnn.
An external environment to which commands can be submitted is identified by an environment name.
Environment names are specified in the ADDRESS instruction to identify the environment to which a
command should be sent.
I/O can be redirected when submitting commands to an external environment. The submitted command's
input stream can be taken from an existing stream or from a set of compound variables with a common
stem. In the latter case (that is, when a stem is specified as the source for the commands input stream)
whole number tails are used to order input for presentation to the submitted command. Stem.0 must
contain a whole number indicating the number of compound variables to be presented, and stem.1
through stem.n (where n=stem.0) are the compound variables to be presented to the submitted
command.
Similarly, the submitted command's output stream can be directed to a stream, or to a set of compound
variables with a given stem. In the latter case (i.e., when a stem is specified as the destination)
compound variables will be created to hold the standard output, using whole number tails as described
above. Output redirection can specify a REPLACE or APPEND option, which controls positioning prior to
the command's execution. REPLACE is the default.
I/O redirection can be persistently associated with an environment name. The term "environment" is used
to refer to an environment name together with the I/O redirections.
At any given time, there will be two environments, the active environment and the alternate environment.
When an ADDRESS instruction specifies a command to the environment, any specified I/O redirection
applies to that command's execution only, providing a third environment for the duration of the instruction.
When an ADDRESS command does not contain a command, that ADDRESS command creates a new
active environment, which includes the specified I/O redirection.
The redirections specified on the ADDRESS instruction may not be possible. If the configuration is aware
that the command processor named does not perform I/O in a manner compatible with the request, the
value of #Env_Type. may be set to 'UNUSED' as an alternative to 'STEM' and 'STREAM' where those
values are assigned in the following code.
In the following code the particular use of #Contains(address, expression) refers to an expression
immediately contained in the address.

AddrInstr:
/* If ADDRESS keyword alone, environments are swapped. */
if \#Contains(address,taken_constant),
& \#Contains(address,valueexp),
& \#Contains(address,'WITH') then do
call EnvAssign TRANSIENT, #Level, ACTIVE, #Level
call EnvAssign ACTIVE, #Level, ALTERNATE, #Level
call EnvAssign ALTERNATE, #Level, TRANSIENT, #Level
return
end
/* The environment name will be explicitly specified. */
if #Contains(address,taken_constant) then
Name = #Instance(address, taken_constant)
else
Name = #Evaluate(valueexp, expression)
if length(Name) > #LimitEnvironmentName then
call #Raise 'SYNTAX', 29.1, Name

if #Contains(address,expression) then do
/* The command is evaluated (but not issued) at this point. */
Command = #Evaluate(address,expression)
if #Tracing.#Level == 'C' | #Tracing.#Level == 'A' then do
call #Trace '>>>'
end
end

74
 call AddressSetup /* Note what is specified on the ADDRESS instruction. */
/* If there is no command, the persistent environment is being set. */
if \#Contains(address,expression) then do
call EnvAssign ACTIVE, #Level, TRANSIENT, #Level
return
end

call CommandIssue Command /* See nnn */

return /* From AddrInstr */

AddressSetup:
/* Note what is specified on the ADDRESS instruction,
into the TRANSIENT environment. */
EnvTail = 'TRANSIENT.'#Level
/* Initialize with defaults. */
#Env_Name.EnvTail = ''
#Env_Type.I.EnvTail = 'NORMAL'
#Env_Type.O.EnvTail = 'NORMAL'
#Env_Type.E.EnvTail = 'NORMAL'
#Env_Resource.I.EnvTail = ''
#Env_Resource.O.EnvTail = ''
#Env_Resource.E.EnvTail = ''
/* APPEND / REPLACE does not apply to input. */
#Env_Position.I.EnvTail = 'INPUT'
#Env_Position.O.EnvTail = 'REPLACE'
#Env_Position.E.EnvTail = 'REPLACE'

/* If anything follows ADDRESS, it will include the command processor name.*/

#Env_Name.EnvTail = Name

/* Connections may be explicitly specified. */

if #Contains(address,connection) then do
if #Contains(connection,input) then do /* input redirection */
if #Contains(resourcei, 'STREAM') then do
#Env_Type.I.EnvTail = 'STREAM'
#Env_Resource.I.EnvTail=#Evaluate(resourcei, VAR_SYMBOL)
end
if #Contains(resourcei,'STEM') then do
#Env_Type.I.EnvTail = 'STEM'
Temp=#Instance(resourcei,VAR_SYMBOL)
if \#Parses(Temp, stem /* See nnn */) then
call #Raise 'SYNTAX', 53.3, Temp
#Env_Resource.I.EnvTail=Temp
end
end /* Input */

if #Contains(connection,output) then /* output redirection */

call NoteTarget O

if #Contains(connection,error) then /* error redirection */

/* The prose on the description of #Contains specifies that the
relevant resourceo is used in NoteTarget. */
call NoteTarget E
end /* Connection */

return /* from AddressSetup */

NoteTarget:
/* Note the characteristics of an output resource. */
arg Which /* O or E */
if #Contains(resourceo,'STREAM') then do
#Env_Type.Which.EnvTail='STREAM'
#Env_Resource.Which.EnvTail=#Evaluate(resourceo, VAR_SYMBOL)
end
if #Contains(resourceo,'STEM') then do
#Env_Type.Which.EnvTail='STEM'
Temp=#Instance(resourceo,VAR_SYMBOL)
if \#Parses(Temp, stem /* See nnn */) then
call #Raise 'SYNTAX', 53.3, Temp
#Env_Resource.Which.EnvTail=Temp
end

75
 if #Contains(resourceo,append) then
#Env_Position.Which.EnvTail='APPEND'
return /* From NoteTarget */

EnvAssign:
/* Copy the values that name an environment and describe its
redirections. */
arg Lhs, LhsLevel, Rhs, RhsLevel
#Env_Name.Lhs.LhsLevel = #Env_Name.Rhs.RhsLevel
#Env_Type.I.Lhs.LhsLevel = #Env_Type.I.Rhs.RhsLevel
#Env_Resource.I.Lhs.LhsLevel = #Env_Resource.I.Rhs.RhsLevel
#Env_Position.I.Lhs.LhsLevel = #Env_Position.I.Rhs.RhsLevel
#Env_Type.O.Lhs.LhsLevel = #Env_Type.O.Rhs.RhsLevel
#Env_Resource.O.Lhs.LhsLevel = #Env_Resource.O.Rhs.RhsLevel
#Env_Position.O.Lhs.LhsLevel = #Env_Position.O.Rhs.RhsLevel
#Env_Type.E.Lhs.LhsLevel = #Env_Type.E.Rhs.RhsLevel
#Env_Resource.E.Lhs.LhsLevel = #Env_Resource.E.Rhs.RhsLevel
#Env_Position.E.Lhs.LhsLevel = #Env_Position.E.Rhs.RhsLevel
return
9.5.2 ARG
For a definition of the syntax of this instruction, see nnn.
The ARG instruction is a shorter form of the equivalent instruction:
PARSE UPPER ARG template_list
9.5.3 Assignment
Assignment can occur as the result of executing a clause containing an assignment (see nnn and nnn),
or as a result of executing the VALUE built-in function, or as part of the execution of a PARSE instruction.
Assignment involves an expression and a VAR_SYMBOL. The value of the expression is determined;
see nnn.
If the VAR_SYMBOL does not contain a period, or contains only one period as its last character, the
value is associated with the VAR_SYMBOL:
call Var_Set #Pool,VAR_SYMBOL,'0',Value
Otherwise, a name is derived, see nnn. The value is associated with the derived name:
call Var_Set #Pool,Derived_Name,'1',Value
9.5.4 CALL
For a definition of the syntax of this instruction, see nnn.
The CALL instruction is used to invoke a routine, or is used to control trapping of conditions.
If a vref is specified that value is the name of the routine to invoke:
if #Contains(call, vref) then
Name = #Evaluate(vref, var_symbol)
If a taken_constant is specified, that name is used.
if #Contains(call, taken_constant) then
Name = #Instance(call, taken_constant)
The name is used to invoke a routine, see nnn. If that routine does not return a result the RESULT and
.RESULT variables become uninitialized:
call Var_Drop #Pool, 'RESULT', '0'
call Var_Drop #ReservedPool, '.RESULT', '0'
If the routine does return a result that value is assigned to RESULT and .RESULT. See nnn for an
exception to assigning results.
If the routine returns a result and the trace setting is 'R' or 'I' then a trace with that result and a tag '>>>'
shall be produced, associated with the call instruction.
If a callon_spec is specified:
If #Contains(call,callon_spec) then do
Condition = #Instance(callon_spec,callable_condition)
#Instruction.Condition.#Level = 'CALL'
If #Contains(callon_spec,'OFF') then
#Enabling.Condition.#Level = 'OFF'
else
#Enabling.Condition.#Level = 'ON'
/* Note whether NAME supplied. */
If Contains(callon_spec,taken_constant) then
Name = #Instance(callable_condition,taken_constant)
else
Name = Condition
#TrapName.Condition.#Level = Name
end

76
9.5.5 Command to the configuration
For a definition of the syntax of a command, see nnn.
A command that is not part of an ADDRESS instruction is processed in the ACTIVE environment.
Command = #Evaluate(command, expression)
if #Tracing.#Level == 'C' | #Tracing.#Level == 'A' then
call #Trace '>>>'
call EnvAssign TRANSIENT, #Level, ACTIVE, #Level
call CommandIssue Command

CommandIssue is also used to describe the ADDRESS instruction:

CommandIssue:
parse arg Cmd
/* Issues the command, requested environment is TRANSIENT */
/* This description does not require the command processor to understand
stems, so it uses an altered environment. */
call EnvAssign PASSED, #Level, TRANSIENT, #Level
EnvTail = 'TRANSIENT.'#Level

/* Note the command input. */

if #Env_Type.I.EnvTail = 'STEM' then do
/* Check reasonableness of the stem. */
Stem = #Env_Resource.I.EnvTail
Lines = value(Stem'0')
if \datatype(Lines,'W') then
call #Raise 'SYNTAX',54.1,Stem'0', Lines
if Lines<0 then
call #Raise 'SYNTAX',54.1,Stem'0', Lines
/* Use a stream for the stem */
#Env_Type.I.PASSED.#Level = 'STREAM'
call Config_Stream_Unique
InputStream = #Outcome
#Env_Resource.I.PASSED.#Level = InputStream
call charout InputStream , ,1
do j = 1 to Lines
call lineout InputStream, value(Stem || j)
end j
call lineout InputStream
end

/* Note the command output. */

if #Env_Type.O.EnvTail = 'STEM' then do
Stem = #Env_Resource.O.EnvTail
if #Env_Position.O.EnvTail == 'APPEND' then do
/* Check that Stem.0 will accept incrementing. */
Lines=value(Stem'0');
if \datatype(Lines,'W') then
call #Raise 'SYNTAX',54.1,Stem'0', Lines
if Lines<0 then
call #Raise 'SYNTAX',54.1,Stem'0', Lines
end
else call value Stem'0',0
/* Use a stream for the stem */
#Env_Type.O.PASSED.#Level = 'STREAM'
call Config_Stream_Unique
#Env_Resource.O.PASSED.#Level = #Outcome
end

/* Note the command error stream. */

if #Env_Type.E.EnvTail = 'STEM' then do
Stem = #Env_Resource.E.EnvTail
if #Env_Position.E.EnvTail == 'APPEND' then do
/* Check that Stem.0 will accept incrementing. */
Lines=value(Stem'0');
if \datatype(Lines,'W') then
call #Raise 'SYNTAX',54.1,Stem'0', Lines
if Lines<0 then
call #Raise 'SYNTAX',54.1,Stem'0', Lines

77
 end
else call value Stem'0',0
/* Use a stream for the stem */
#Env_Type.E.PASSED.#Level = 'STREAM'
call Config_Stream_Unique
#Env_Resource.E.PASSED.#Level = #Outcome
end

#API_Enabled = '1'
call Var_Reset #Pool
/* Specifying PASSED here implies all the
components of that environment. */
#Response = Config_Command(PASSED, Cmd)
#Indicator = left(#Response,1)
Description = substr(#Response,2)
#API_Enabled = '0'
/* Recognize success and failure. */
if #AtPause = 0 then do
call value 'RC', #RC
call Var_Set 0, '.RC', 0, #RC
end
select
when #Indicator=='N' then Temp=0
when #Indicator=='F' then Temp=-1 /* Failure */
when #Indicator=='E' then Temp=1 /* Error */
end
call Var_Set 0, '.RS', 0, Temp
/* Process the output */
if #Env_Type.O.EnvTail='STEM' then do /* get output into stem. */
Stem = #Env_Resource.O.EnvTail
OutputStream = #Env_Resource.O.PASSED.#Level
do while lines(OutputStream) > 0
call value Stem'0',value(Stem'0')+1
call value Stem||value(Stem'0'),linein(OutputStream)
end
end /* Stemmed Output */
if #Env_Type.E.EnvTail='STEM' then do /* get error output into stem. */
Stem = #Env_Resource.E.EnvTail
OutputStream = #Env_Resource.E.PASSED.#Level
do while lines(OutputStream)> 0
call value Stem'0',value(Stem'0')+1
call value Stem||value(Stem'0'),linein(OutputStream)
end
end /* Stemmed Error output */
if #Indicator \== 'N' & pos(#Tracing.#Level, 'CAIR') > 0 then
call #Trace '+++'
if (#Indicator \== 'N' & #Tracing.#Level=='E'),
| (#Indicator=='F' & (#Tracing.#Level=='F' | #Tracing.#Level=='N')) then do
call #Trace '>>>'
call #Trace '+++'
end
#Condition='FAILURE'
if #Indicator='F' & #Enabling.#Condition.#Level \== 'OFF' then call #Raise 'FAILURE'
, Cmd
else if #Indicator='E' | #Indicator='F' then call #Raise 'ERROR', Cmd
return /* From CommandIssue */

The configuration may choose to perform the test for message 54.1 before or after issuing the command.

9.5.6 DO
For a definition of the syntax of this instruction, see nnn.
The DO instructions is used to group instructions together and optionally to execute them repeatedly.
Executing a do_simple has the same effect as executing a nop, except in its trace output. Executing the
do_ending associated with a do_simple has the same effect as executing a nop, except in its trace
output.
A do_instruction that does not contain a do_simple is equivalent, except for trace output, to a sequence of
instructions in the following order.

78
#Loop = #Loop+1
#Iterate.#Loop = #Clause(IterateLabel)
#Once.#Loop = #Clause(OnceLabel)
#Leave.#Loop = #Clause(LeaveLabel)
if #Contains(do_specification,assignment) then
#Identity.#Loop = #Instance(assignment, VAR_SYMBOL)
if #Contains(do_specification, repexpr) then
if \datatype(repexpr,'W') then
call #Raise 'SYNTAX', 26.2,repexpr
else do
#Repeat.#Loop = repexpr+0
if #Repeat.#Loop<0 then
call #Raise 'SYNTAX',26.2,#Repeat.#Loop
end
if #Contains(do_specification,assignment) then do
#StartValue.#Loop = #Evaluate(assignment,expression)
if datatype(#StartValue.#Loop) \== 'NUM' then
call #Raise 'SYNTAX', 41.6, #StartValue.#Loop
#StartValue.#Loop = #StartValue.#Loop + 0
if \#Contains(do_specification,byexpr) then
#By.#Loop = 1
end

The following three assignments are made in the order in which 'TO', 'BY' and 'FOR' appear in docount;
see nnn.

if #Contains(do_specification, toexpr) then do

if datatype(toexpr) \== 'NUM' then
call #Raise 'SYNTAX', 41.4, toexpr
#To.#Loop = toexpr+0
if #Contains(do_specification, byexpr) then do
if datatype(byexpr)\=='NUM' then
call #Raise 'SYNTAX', 41.5, byexpr
#By.#Loop = byexpr+0
if #Contains(do_specification, forexpr) then do
if \datatype(forexpr, 'W') then
call #Raise 'SYNTAX', 26.3, forexpr
#For.#Loop = forexpr+0
if #For.#Loop <0 then
call #Raise 'SYNTAX', 26.3, #For.#Loop
end
if #Contains(do_specification,assignment) then do
call value #Identity.#Loop, #StartValue.#Loop
end
if #Contains(do_specification,'OVER') then do
Value = #Evaluate(dorep, expression)
#OverArray.#Loop = Value ~ makearray
#Repeat.#Loop = #OverArray~items /* Count this downwards as if repexpr. */
#Identity.#Loop = #Instance(dorep, VAR_SYMBOL)
end
call #Goto #Once.#Loop /* to OnceLabel */
IterateLabel:
if #Contains(do_specification, untilexpr) then do
Value = #Evaluate(untilexp, expression)
if Value == '1' then leave
if Value \== '0' then call #Raise 'SYNTAX', 34.4, Value
end
if #Contains(do_specification, assignment) then do
t = value(#Identity.#Loop)
if #Indicator == 'D' then call #Raise 'NOVALUE', #Identity.#Loop
call value #Identity.#Loop, t + #By.#Loop
end

OnceLabel:
if #Contains(do_specification, toexpr) then do
if #By.#Loop>=0 then do
if value(#Identity.#Loop) > #To.#Loop then leave
end
else do if value(#Identity.#Loop) < #To.#Loop then leave
end
end

79
if #Contains(dorep, repexpr) | #Contains(dorep, 'OVER') then do
if #Repeat.#Loop = 0 then leave
#Repeat.#Loop = #Repeat.#Loop-1
if #Contains(dorep, 'OVER') then
call value #Identity.#Loop, #OverArray[#OverArray~items - #Repeat.#Loop]
end
if #Contains(do_specification, forexpr) then do
if #For.#Loop = 0 then leave
#For.#Loop = #For.#Loop - 1
end
if #Contains(do_specification, whileexpr) then do
Value = #Evaluate(whileexp, expression)
if Value == '0' then leave
if Value \== '1' then call #Raise 'SYNTAX', 34.3, Value
end
#Execute(do_instruction, instruction_list)
TraceOfEnd:
call #Goto #Iterate.#Loop /* to IterateLabel */
LeaveLabel:
#Loop = #Loop - 1
9.5.7 DO loop tracing
When clauses are being traced by #TraceSource, due to pos(#Tracing.#Level, 'AIR') > 0, the DO
instruction shall be traced when it is encountered and again each time the IterateLabel (see nnn) is
encountered. The END instruction shall be traced when the TraceOfEnd label is encountered.
When expressions or intermediates are being traced they shall be traced in the order specified by nnn.
Hence, in the absence of conditions arising, those executed prior to the first execution of OnceLabel shall
be shown once per execution of the DO instruction; others shall be shown depending on the outcome of
the tests.
The code in the DO description:
t = value(#Identity.#Loop)
if #Indicator == 'D' then call #Raise 'NOVALUE', #Identity.#Loop
call value #Identity.#Loop, t + #By.#Loop
represents updating the control variable of the loop. That assignment is subject to tracing, and other
expressions involving state variables are not. When tracing intermediates, the BY value will have a tag of
'>+>'.
9.5.8 DROP
For a definition of the syntax of this instruction, see nnn.
The DROP instruction restores variables to an uninitialized state.
The words of the variable_list are processed from left to right.
A word which is a VAR_SYMBOL, not contained in parentheses, specifies a variable to be dropped. If
VAR_SYMBOL does not contain a period, or has only a single period as its last character, the variable
associated with VAR_SYMBOL by the variable pool is dropped:
#Response = Var_Drop(#Pool,VAR_SYMBOL,'0')
If VAR_SYMBOL has a period other than as the last character, the variable associated with
VAR_SYMBOL by the variable pool is dropped by:
#Response = Var_Drop (#Pool,VAR_SYMBOL,'1')
If the word of the variable_list is a VAR_SYMBOL enclosed in parentheses then the value of the
VAR_SYMBOL is processed. The value is considered in uppercase:
#Value = Config_Upper(#Value)
Each word in that value found by the WORD built-in function, from left to right, is subjected to this
process:
If the word does not have the syntax of VAR_SYMBOL a condition is raised:
call #Raise 'SYNTAX', 20.1, word
Otherwise the VAR_SYMBOL indicated by the word is dropped, as if that VAR_SYMBOL were a word of
the variable_list.
9.5.9 EXIT
For a definition of the syntax of this instruction, see nnn.
The EXIT instruction is used to unconditionally complete execution of a program.
Any expression is evaluated:
if #Contains(exit, expression) then Value = #Evaluate(exit, expression)
#Level = 1
#Pool = #Pool1

80
The opportunity is provided for a final trap.
#API_Enabled = '1'
call Var_Reset #Pool
call Config_Termination
#API_Enabled = '0'
The processing of the program is complete. See nnn for what API_Start returns as the result.
If the normal sequence of execution "falls through" the end of the program; that is, would execute a
further statement if one were appended to the program, then the program is terminated in the same
manner as an EXIT instruction with no argument.
9.5.10 EXPOSE
The expose instruction identifies variables that are not local to the method.
We need a check that this starts method; similarities with PROCEDURE.
For a definition of the syntax of this instruction, see nnn.
It is used at the start of a method, after method initialization, to make variables in the receiver's pool
accessible:
if \#AllowExpose then call #Raise 'SYNTAX', 17.2
The words of the variable_list are processed from left to right.
A word which is a VAR_SYMBOL, not contained in parentheses, specifies a variable to be made
accessible. If VAR_SYMBOL does not contain a period, or has only a single period as its last character,
the variable associated with VAR_SYMBOL by the variable pool (as a non-tailed name) is given the
attribute 'exposed'.
call Var_Expose #Pool, VAR_SYMBOL, '0'
If VAR_SYMBOL has a period other than as last character, the variable associated with VAR_SYMBOL
in the variable pool ( by the name derived from VAR_SYMBOL, see nnn) is given the attribute 'exposed'.
call Var_Expose #Pool, Derived_Name, '1'
If the word from the variable_list is a VAR_SYMBOL enclosed in parentheses then the VAR_SYMBOL is
exposed, as if that VAR_SYMBOL was a word in the variable_list. The value of the VAR_SYMBOL is
processed. The value is considered in uppercase:
#Value = Config_Upper(#Value)
Each word in that value found by the WORD built-in function, from left to right, is subjected to this
process:
If the word does not have the syntax of VAR_SYMBOL a condition is raised:
call #Raise 'SYNTAX', 20.1, word
Otherwise the VAR_SYMBOL indicated by the word is exposed, as if that VAR_SYMBOL were a word of
the variable_list.
9.5.11 FORWARD
For a definition of the syntax of this instruction, see nnn.
The FORWARD instruction is used to send a message based on the current message.
if #Contains(forward, 'ARRAY') & #Contains(forward, 'ARGUMENTS') then
call #Raise 'SYNTAX', nn.n
9.5.12 GUARD
For a definition of the syntax of this instruction, see nnn.
The GUARD instruction is used to conditionally delay the execution of a method.
do forever
Value = #Evaluate( guard, expression)
if Value == '1' then leave
if Value \== '0' then call #Raise 'SYNTAX', 34.7, Value
Drop exclusive access and wait for change
end
9.5.13 IF
For a definition of the syntax of this instruction, see nnn.
The IF instruction is used to conditionally execute an instruction, or to select between two alternatives.
The expression is evaluated. If the value is neither '0' nor '1' error 34.1 occurs. If the value is '1', the
instruction in the then is executed. If the value is '0' and else is specified, the instruction in the else is
executed.
In the former case, if tracing clauses, the clause consisting of the THEN keyword shall be traced in
addition to the instructions.
In the latter case, if tracing clauses, the clause consisting of the ELSE keyword shall be traced in addition
to the instructions.

81
9.5.14 INTERPRET
For a definition of the syntax of this instruction, see nnn.
The INTERPRET instruction is used to execute instructions that have been built dynamically by
evaluating an expression.
The expression is evaluated.
The HALT condition is tested for, and may be raised, in the same way it is tested at clause termination,
see nnn.
The process of syntactic recognition described in clause 6 is applied, with Config_SourceChar obtaining
its results from the characters of the value, in left-to-right order, without producing any EOL or EOS
events. When the characters are exhausted, the event EOL occurs, followed by the event EOS.
If that recognition would produce any message then the interpret raises the corresponding 'SYNTAX'
condition.
If the program recognized contains any LABELs then the interpret raises a condition:
call #Raise 'SYNTAX',47.1,Label
where Label is the first LABEL in the program.
Otherwise the instruction_list in the program is executed.
9.5.15 ITERATE
For a definition of the syntax of this instruction, see nnn.
The ITERATE instruction is used to alter the flow of control within a repetitive DO.
For a definition of the nesting correction, see nnn.
#Loop = #Loop - NestingCorrection
call #Goto #Iterate.#Loop

9.5.16 Execution of labels

The execution of a label has no effect, other than clause termination activity and any tracing.
if #Tracing.#Level=='L' then call #TraceSource

9.5.17 LEAVE
For a definition of the syntax of this instruction, see nnn.
The LEAVE instruction is used to immediately exit one or more repetitive DOs.
For a definition of the nesting correction, see nnn.
#Loop = #Loop - NestingCorrection
call #Goto #Leave.#Loop

9.5.18 Message term

We can do this by reference to method invokation, just as we do CALL by reference to invoking a function.

9.5.19 LOOP

Shares most of it's definition with repetitive DO.

9.5.20 NOP
For a definition of the syntax of this instruction, see nnn.
The NOP instruction has no effect other than the effects associated with all instructions.

9.5.21 NUMERIC
For a definition of the syntax of this instruction, see nnn.
The NUMERIC instruction is used to change the way in which arithmetic operations are carried out.

9.5.21.1 NUMERIC DIGITS

For a definition of the syntax of this instruction, see nnn.
NUMERIC DIGITS controls the precision under which arithmetic operations and arithmetic built-in
functions will be evaluated.
if #Contains(numericdigits, expression) then
Value = #Evaluate(numericdigits, expression)
else Value = 9

82
if \datatype(Value,'W') then
call #Raise 'SYNTAX',26.5,Value
Value = Value + 0
if Value<=#Fuzz.#Level then
call #Raise 'SYNTAX',33.1,Value
if Value>#Limit_Digits then
call #Raise 'SYNTAX',33.2,Value
#Digits.#Level = Value

9.5.21.2 NUMERIC FORM

For a definition of the syntax of this instruction, see nnn.
NUMERIC FORM controls which form of exponential notation is to be used for the results of operations
and arithmetic built-in functions.
The value of form is either taken directly from the SCIENTIFIC or ENGINEERING keywords, or by
evaluating valueexp .
if \#Contains(numeric,numericsuffix) then
Value = 'SCIENTIFIC'
else if #Contains(numericformsuffix,'SCIENTIFIC') then
Value = 'SCIENTIFIC'
else
if #Contains(numericformsuffix,'ENGINEERING') then
Value = 'ENGINEERING'
else do
Value = #Evaluate(numericformsuffix,valueexp)
Value = translate(left(Value,1))
select
when Value == 'S' then Value = 'SCIENTIFIC'
when Value == 'E' then Value = 'ENGINEERING'
otherwise call #Raise 'SYNTAX',33.3,Value
end
end
#Form.#Level = Value

9.5.21.3 NUMERIC FUZZ

For a definition of the syntax of this instruction, see nnn.
NUMERIC FUZZ controls how many digits, at full precision, will be ignored during a numeric comparison.
If #Contains(numericfuzz,expression) then
Value = #Evaluate(numericfuzz,expression)
else
Value = 0
If \datatype(Value,'W') then
call #Raise 'SYNTAX',26.6,Value
Value = Value+0
If Value < 0 then
call #Raise 'SYNTAX',26.6,Value
If Value >= #Digits.#Level then
call #Raise 'SYNTAX',33.1,#Digits.#Level,Value
#Fuzz.#Level = Value

9.5.22 OPTIONS
For a definition of the syntax of this instruction, see nnn.
The OPTIONS instruction is used to pass special requests to the language processor.
The expression is evaluated and the value is passed to the language processor. The language processor
treats the value as a series of blank delimited words. Any words in the value that are not recognized by
the language processor are ignored without producing an error.
call Config_Options(Expression)

9.5.23 PARSE
For a definition of the syntax of this instruction, see nnn.
The PARSE instruction is used to assign data from various sources to variables.
The purpose of the PARSE instruction is to select substrings of the parse_type under control of the
template_list. If the template_list is omitted, or a template in the list is omitted, then a template which is
the null string is implied.
Processing for the PARSE instruction begins by constructing a value, the source to be parsed.

83
 ArgNum = 0
select
when #Contains(parse_type, 'ARG') then do
ArgNum = 1
ToParse = #Arg.#Level.ArgNum
end
when #Contains(parse_type, 'LINEIN') then ToParse = linein('')
when #Contains(parse_type, 'PULL') then do
/* Acquire from external queue or default input. */
#Response = Config_Pull()
if left(#Response, 1) == 'F' then
call Config_Default_Input
ToParse = #Outcome
end
when #Contains(parse_type, 'SOURCE') then
ToParse = #Configuration #HowInvoked #Source
when #Contains(parse_type, 'VALUE') then
if \#Contains(parse_value, expression) then ToParse = ''
else ToParse = #Evaluate(parse_value, expression)
when #Contains(parse_type, 'VAR') then
ToParse = #Evaluate(parse_var,VAR_SYMBOL)
when #Contains(parse_type, 'VERSION') then ToParse = #Version
end
Uppering = #Contains(parse, 'UPPER')
The first template is associated with this source. If there are further templates, they are matched against
null strings unless 'ARG' is specified, when they are matched against further arguments.
The parsing process is defined by the following routine, ParseData. The template_list is accessed by
ParseData as a stemmed variable. This variable Template. has elements which are null strings except
for any elements with tails 1,2,3,... corresponding to the tokens of the template_list from left to right.

ParseData:
/* Targets will be flagged as the template is examined. */
Target.='0'
/* Token is a cursor on the components of the template,
moved by FindNextBreak. */
Token = 1
/* Tok is a cursor on the components of the template
moved through the target variables by routine WordParse. */
Tok = 1
do forever /* Until commas dealt with. */
/* BreakStart and BreakEnd indicate the position in the source
string where there is a break that divides the source. When the break
is a pattern they are the start of the pattern and the position just
beyond it. */
BreakStart = 1
BreakEnd = 1
SourceEnd = length(ToParse) + 1
If Uppering then ToParse = translate(ToParse)

do while Template.Tok \== '' & Template.Tok \== ','

/* Isolate the data to be processed on this iteration. */

call FindNextBreak /* Also marks targets. */

/* Results have been set in DataStart which indicates the start

of the isolated data and BreakStart and BreakEnd which are ready
for the next iteration. Tok has not changed. */

/* If a positional takes the break leftwards from the end of the

previous selection, the source selected is the rest of the string, */

if BreakEnd <= DataStart then

DataEnd = SourceEnd
else
DataEnd = BreakStart

/* Isolated data, to be assigned from: */

Data=substr(ToParse,DataStart,DataEnd-DataStart)
call WordParse /* Does the assignments. */

84
 end /* while */
if Template.Tok \== ',' then leave
/* Continue with next source. */
Token=Token+1
Tok=Token
if ArgNum <> 0 then do
ArgNum = ArgNum+1
ToParse = #Arg.ArgNum
end
else ToParse=''
end

return /* from ParseData */

FindNextBreak:
do while Template.Token \== '' & Template.Token \== ','
Type=left(Template.Token,1)
/* The source data to be processed next will normally start at the end of
the break that ended the previous piece. (However, the relative
positionals alter this.) */
DataStart = BreakEnd
select

when Type='"' | Type="'" | Type='(' then do

if Type='(' then do
/* A parenthesis introduces a pattern which is not a constant. */
Token = Token+1
Pattern = value(Template.Token)
if #Indicator == 'D' then call #Raise 'NOVALUE', Template.Token
Token = Token+1
end
else
/* The following removes the outer quotes from the
literal pattern */
interpret "Pattern="Template.Token
Token = Token+1
/* Is that pattern in the remaining source? */
PatternPos=pos(Pattern,ToParse,DataStart)
if PatternPos>0 then do
/* Selected source runs up to the pattern. */
BreakStart=PatternPos
BreakEnd=PatternPos+length(Pattern)
return
end
leave /* The rest of the source is selected. */
end

when datatype(Template.Token,'W') | pos(Type,'+-=') > 0 then do

/* A positional specifies where the relevant piece of the subject
ends. */
if pos(Type,'+-=') = 0 then do
/* Whole number positional */
BreakStart = Template.Token
Token = Token+1
end
else do
/* Other forms of positional. */
Direction=Template.Token
Token = Token + 1
/* For a relative positional, the position is relative to the start
of the previous trigger, and the source segment starts there. */
if Direction \== '=' then
DataStart = BreakStart
/* The adjustment can be given as a number or a variable in
parentheses. */
if Template.Token ='(' then do
Token=Token + 1
BreakStart = value(Template.Token)
if #Indicator == 'D' then call #Raise 'NOVALUE', Template.Token
Token=Token + 1
end
else BreakStart = Template.Token

85
 if \datatype(BreakStart,'W')
then call #Raise 'SYNTAX', 26.4,BreakStart
Token = Token+1
if Direction='+'
then BreakStart=DataStart+BreakStart
else if Direction='-'
then BreakStart=DataStart-BreakStart
end
/* Adjustment should remain within the ToParse */
BreakStart = max(1, BreakStart)
BreakStart = min(SourceEnd, BreakStart)
BreakEnd = BreakStart /* No actual literal marks the boundary. */
return
end

when Template.Token \== '.' & pos(Type,'0123456789.')>0 then

/* A number that isn't a whole number. */
call #Raise 'SYNTAX', 26.4, Template.Token
/* Raise will not return */

otherwise do /* It is a target, not a pattern */

Target.Token='1'
Token = Token+1
end

end /* select */
end /* while */
/* When no more explicit breaks, break is at the end of the source. */
DataStart=BreakEnd
BreakStart=SourceEnd
BreakEnd=SourceEnd
return /* From FindNextBreak */

WordParse:
/* The names in the template are assigned blank-delimited values from the
source string. */

do while Target.Tok /* Until no more targets for this data. */

/* Last target gets all the residue of the Data. */
NextTok = Tok + 1
if \Target.NextTok then do
call Assign(Data)
leave
end
/* Not last target; assign a word. */
Data = strip(Data,'L')
if Data == '' then call Assign('')
else do
Word = word(Data,1)
call Assign Word
Data = substr(Data,length(Word) + 1)
/* The word terminator is not part of the residual data: */
if Data \== '' then Data = substr(Data,2)
end
Tok = Tok + 1
end
Tok=Token /* Next time start on new part of template. */
return

Assign:
if Template.Tok=='.' then Tag='>.>'
else do
Tag='>=>'
call value Template.Tok,arg(1)
end
/* Arg(1) is an implied argument of the tracing. */
if #Tracing.#Level == 'R' | #Tracing.#Level == 'I' then call #Trace Tag
return

9.5.24 PROCEDURE
For a definition of the syntax of this instruction, see nnn.

86
The PROCEDURE instruction is used within an internal routine to protect all the existing variables by
making them unknown to following instructions. Selected variables may be exposed.
It is used at the start of a routine, after routine initialization:
if \#AllowProcedure.#Level then call #Raise 'SYNTAX', 17.1
#AllowProcedure.#Level = 0
/* It introduces a new variable pool: */
call #Config_ObjectNew
call var_set(#Outcome,'#UPPER', '0', #Pool) /* Previous #Pool is upper from the new
#Pool. */
#Pool=#Outcome
IsProcedure.#Level='1'
call Var_Empty #Pool

If there is a variable_list, it provides access to a previous variable pool.

The words of the variable_list are processed from left to right.
A word which is a VAR_SYMBOL, not contained in parentheses, specifies a variable to be made
accessible. If VAR_SYMBOL does not contain a period, or has only a single period as its last character,
the variable associated with VAR_SYMBOL by the variable pool (as a non-tailed name) is given the
attribute 'exposed'.
call Var_Expose #Pool, VAR_SYMBOL, '0'
If VAR_SYMBOL has a period other than as last character, the variable associated with VAR_SYMBOL
in the variable pool ( by the name derived from VAR_SYMBOL, see nnn) is given the attribute 'exposed'.
call Var_Expose #Pool, Derived_Name, '1'
If the word from the variable_list is a VAR_SYMBOL enclosed in parentheses then the VAR_SYMBOL is
exposed, as if that VAR_SYMBOL was a word in the variable_list. The value of the VAR_SYMBOL is
processed. The value is considered in uppercase:
#Value = Config_Upper(#Value)
Each word in that value found by the WORD built-in function, from left to right, is subjected to this
process:
If the word does not have the syntax of VAR_SYMBOL a condition is raised:
call #Raise 'SYNTAX', 20.1, word
Otherwise the VAR_SYMBOL indicated by the word is exposed, as if that VAR_SYMBOL were a word of
the variable_list.
9.5.25 PULL
For a definition of the syntax of this instruction, see nnn.
A PULL instruction is a shorter form of the equivalent instruction:
PARSE UPPER PULL template_list
9.5.26 PUSH
For a definition of the syntax of this instruction, see nnn.
The PUSH instruction is used to place a value on top of the stack.
If #Contains(push,expression) then
Value = #Evaluate(push,expression)
else
Value = ''
call Config_Push Value

9.5.27 QUEUE
For a definition of the syntax of this instruction, see nnn.
The QUEUE instruction is used to place a value on the bottom of the stack.
If #Contains(queue,expression) then
Value = #Evaluate(queue,expression)
else
Value = ''
call Config_Queue Value

9.5.28 RAISE
The RAISE instruction returns from the current method or routine and raises a condition.
9.5.29 REPLY
The REPLY instruction is used to allow both the invoker of a method, and the replying method, to
continue executing.

87
Must set up for error of expression on subsequent RETURN.
9.5.30 RETURN
For a definition of the syntax of this instruction, see nnn.
The RETURN instruction is used to return control and possibly a result from a program or internal routine
to the point of its invocation.
The RETURN keyword may be followed by an optional expression, which will be evaluated and returned
as a result to the caller of the routine.
Any expression is evaluated:
if #Contains(return,expression) then
#Outcome = #Evaluate(return, expression)
else if #IsFunction.#Level then
call #Raise 'SYNTAX', 45.1, #Name.#Level

At this point the clause termination occurs and then the following:

If the routine started with a PROCEDURE instruction then the associated pool is taken out of use:
if #IsProcedure.#Level then #Pool = #Upper
A RETURN instruction which is interactively entered at a pause point leaves the pause point.
if #Level = #AtPause then #AtPause = 0
The activity at this level is complete:
#Level = #Level-1
#NewLevel = #Level+1
If #Level is not zero, the processing of the RETURN instruction and the invocation is complete.
Otherwise processing of the program is completed:
The opportunity is provided for a final trap.
#API_Enabled = '1'
call Var_Reset #Pool
call Config_Termination
#API_Enabled = '0'
The processing of the program is complete. See nnn for what API_Start returns as the result.
9.5.31 SAY
For a definition of the syntax of this instruction, see nnn.
The SAY instruction is used to write a line to the default output stream.
If #Contains(say,expression) then
Value = Evaluate(say,expression)
else
Value = ''
call Config_Default_Output Value

9.5.32 SELECT
For a definition of the syntax of this instruction, see nnn.
The SELECT instruction is used to conditionally execute one of several alternative instructions.
When tracing, the clause containing the keyword SELECT is traced at this point.
The #Contains(select_body, when) test in the following description refers to the items of the optional
when repetition in order:
LineNum = #LineNumber
Ending = #Clause(EndLabel)
Value=#Evaluate(select_body, expression) /* In the required WHEN */
if Value \== '1' & Value \== '0' then
call #Raise 'SYNTAX',34.2,Value
If Value=='1' then
call #Execute when, instruction
else do
do while #Contains(select_body, when)
Value = #Evaluate(when, expression)
If Value=='1' then do
call #Execute when, instruction
call #Goto Ending
end
if Value \== '0' then
call #Raise 'SYNTAX', 34.2, Value
end /* Of each when */

88
 If \#Contains(select_body, 'OTHERWISE') then
call #Raise 'SYNTAX', 7.3, LineNum
If #Contains(select_body, instruction_list) then
call #Execute select_body, instruction_list
end
EndLabel:
When tracing, the clause containing the END keyword is traced at this point.

9.5.33 SIGNAL
For a definition of the syntax of this instruction, see nnn.
The SIGNAL instruction is used to cause a change in the flow of control or is used with the ON and OFF
keywords to control the trapping of conditions.
If #Contains(signal,signal_spec) then do
Condition = #Instance(signal_spec,condition)
#Instruction.Condition.#Level = 'SIGNAL'
If #Contains(signal_spec,'OFF') then
#Enabling.Condition.#Level = 'OFF'
else
#Enabling.Condition.#Level = 'ON'
If Contains(signal_spec,taken_constant) then
Name = #Instance(condition,taken_constant)
else
Name = Condition
#TrapName.Condition.#Level = Name
end
If there was a signal_spec this complete the processing of the signal instruction. Otherwise:
if #Contains(signal,valueexp)
then Name = #Evaluate(valueexp, expression)
else Name = #Instance(signal,taken_constant)

The Name matches the first LABEL in the program which has that value. The comparison is made with
the '==' operator.
If no label matches then a condition is raised:
call #Raise 'SYNTAX',16.1, Name
If the name is a trace-only label then a condition is raised:
call #Raise 'SYNTAX', 16.2, Name
If the name matches a label, execution continues at that label after these settings:
#Loop.#Level = 0
/* A SIGNAL interactively entered leaves the pause point. */
if #Level = #AtPause then #AtPause = 0
9.5.34 TRACE
For a definition of the syntax of this instruction, see nnn.
The TRACE instruction is used to control the trace setting which in turn controls the tracing of execution
of the program.
The TRACE instruction is ignored if it occurs within the program (as opposed to source obtained by
Config_Trace_Input) and interactive trace is requested (#Interactive.#Level = '1'). Otherwise:
#TraceInstruction = '1'
value = ''
if #Contains(trace, valueexp) then Value = #Evaluate(valueexp, expression)
if #Contains(trace, taken_constant) then Value = #Instance(trace,taken_constant)
if datatype(Value) == 'NUM' & \datatype(Value,'W') then
call #Raise 'SYNTAX', 26.7, Value
if datatype(Value,'W') then do
/* Numbers are used for skipping. */
if Value>=0 then #InhibitPauses = Value
else #InhibitTrace = -Value
end
else do
if length(Value) = 0 then do
#Interactive.#Level = '0'
Value = 'N'
end
/* Each question mark toggles the interacting. */
do while left(Value,1)=='?'
#Interactive.#Level = \#Interactive.#Level
Value = substr(Value,2)

89
 end
if length(Value) \= 0 then do
Value = translate( left(Value,1) )
if verify(Value, 'ACEFILNOR') > 0 then
call #Raise 'SYNTAX', 24.1, Value
if Value=='O' then #Interactive.#Level='0'
end
#Tracing.#Level = Value
end

9.5.35 Trace output

If #NoSource is '1' there is no trace output.
The routines #TraceSource and #Trace specify the output that results from the trace settings. That
output is presented to the configuration by Config_Trace_Output as lines. Each line has a clause
identifier at the left, followed by a blank, followed by a three character tag, followed by a blank, followed
by the trace data.
The width of the clause identifier shall be large enough to hold the line number of the last line in the
program, and no larger. The clause identifier is the source program line number, or all blank if the line
number is the same as the previous line number indicated and no execution with trace Off has occurred
since. The line number is right-aligned with leading zeros replaced by blank characters.
When input at a pause is being executed (#AtPause \= 0 ), #Trace does nothing when the tag is not '+++'.

When input at a pause is being executed, #TraceSource does nothing.

If #InhibitTrace is greater than zero, #TraceSource does nothing except decrement #InhibitTrace.
Otherwise, unless the current clause is a null clause, #TraceSource outputs all lines of the source
program which contain any part of the current clause, with any characters in those lines which are not
part of the current clause and not other_blank_characters replaced by blank characters. The possible
replacement of other_blank_characters is defined by the configuration. The tag is '*-*', or if the line is not
the first line of the clause. '*,*'.
#Trace output also has a clause identifier and has a tag which is the argument to the #Trace invocation.
The data is truncated, if necessary, to #Limit_TraceData characters. The data is enclosed by quotation
marks and the quoted data preceded by two blanks. If the data is truncated, the trailing quote has the
three characters '...' appended.
_ when #Tracing.#Level is 'C' or 'E' or 'F' or 'N' or 'A' and the tag is '>>>' then the data is the value of
the command passed to the environment;
_ when the tag is '+++' then the data is the four characters 'RC "' concatenated with #RC
concatenated with the character '"';
_ when #Tracing.#Level is 'I' or 'R' the data is the most recently evaluated value.
Trace output can also appear as the result of a 'SYNTAX' condition occurring, irrespective of the trace
setting. If a 'SYNTAX' condition occurs and it is not trapped by SIGNAL ON SYNTAX, then the clause in
error shall be traced, along with a traceback. A traceback is a display of each active CALL and
INTERPRET instruction, and function invocation, displayed in reverse order of execution, each with a tag
of '+++'.
9.5.36 USE
For a definition of the syntax of this instruction, see nnn.
The USE instruction assigns the values of arguments to variables.
Better not say copies since COPY method has different semantics.
The optional VAR_SYMBOL positions, positions 1, 2, ..., of the instruction are considered from left to
right. If the position has a VAR_SYMBOL then its value is assigned to:
if #ArgExists.Position then
call Value VAR_SYMBOL, #Arg.Position
else
Messy because VALUE bif won't DROP and var_drop needs to know if compound.
9.6 Conditions and Messages
When an error occurs during execution of a program, an error number and message are associated with
it. The error number has two parts, the error code and the error subcode. These are the integer and
decimal parts of the error number. Subcodes beginning or ending in zero are not used.
Error codes in the range 1 to 90 and error subcodes up to .9 are reserved for errors described here and
for future extensions of this standard.

90
Error number 3 is available to report error conditions occuring during the initialization phase; error
number 2 is available to report error conditions during the termination phase. These are error conditions
recognized by the language processor, but the circumstances of their detection is outside of the scope of
this standard.
The ERRORTEXT built-in function returns the text as initialized in nnn when called with the 'Standard'
option. When the 'Standard' option is omitted, implementation-dependent text may be returned.
When messages are issued any message inserts are replaced by actual values.
The notation for detection of a condition is:
call #Raise Condition, Arg2, Arg3, Arg4, Arg5, Arg6
Some of the arguments may be omitted. In the case of condition 'SYNTAX' the arguments are the
message number and the inserts for the message. In other cases the argument is a further description of
the condition.
The action of the program as a result of a condition is dependent on any signal_spec and callon_spec in
the program.
9.6.1 Raising of conditions
The routine #Raise corresponds to raising a condition. In the following definition, the instructions
containing SIGNAL VALUE and INTERPRET denote transfers of control in the program being
processed. The instruction EXIT denotes termination. If not at an interactive pause, this will be
termination of the program, see nnn, and there will be output by Config_Trace_Output of the message
(with prefix _ see nnn) and tracing (see nnn). If at an interactive pause (#AtPause \= 0), this will be
termination of the interpretation of the interactive input; there will be output by Config_Trace_Output of
the message (without traceback) before continuing. The description of the continuation is in nnn after
the "interpret #Outcome" instruction.
The instruction "interpret 'CALL' #TrapName.#Condition.#Level" below does not set the variables
RESULT and .RESULT; any result returned is discarded.

#Raise:
/* If there is no argument, this is an action which has been delayed from the time the
condition occurred until an appropriate clause boundary. */
if \arg(1,'E') then do
Description = #PendingDescription.#Condition.#Level
Extra = #PendingExtra.#Condition.#Level
end
else do
#Condition = arg(1)
if #Condition \== 'SYNTAX' then do
Description = arg(2)
Extra = arg(3)
end
else do
Description = #Message(arg(2),arg(3),arg(4),arg(5))
call Var_Set #ReservedPool, '.MN', 0, arg(2)
Extra = ''
end
end

/* The events for disabled conditions are ignored or cause termination. */

if #Enabling.#Condition.#Level == 'OFF' | #AtPause \= 0 then do
if #Condition \== 'SYNTAX' & #Condition \== 'HALT' then
return /* To after use of #Raise. */
if #Condition == 'HALT' then Description = #Message(4.1, Description)
exit /* Terminate with Description as the message. */
end

/* SIGNAL actions occur as soon as the condition is raised. */

if #Instruction.#Condition.#Level == 'SIGNAL' then do
#ConditionDescription.#Level = Description
#ConditionExtra.#Level = Extra
#ConditionInstruction.#Level = 'SIGNAL'
#Enabling.#Condition.#Level = 'OFF'
signal value #TrapName.#Condition.#Level
end

/* All CALL actions are initially delayed until a clause boundary. */

91
if arg(1,'E') then do
/* Events within the handler are not stacked up, except for one
extra HALT while a first is being handled. */
EventLevel = #Level
if #Enabling.#Condition.#Level == 'DELAYED' then do
if #Condition \== 'HALT' then return
EventLevel = #EventLevel.#Condition.#Level
if #PendingNow.#Condition.EventLevel then return
/* Setup a HALT to come after the one being handled. */
end
/* Record a delayed event. */
#PendingNow.#Condition.EventLevel = '1'
#PendingDescription.#Condition.EventLevel = Description
#PendingExtra.#Condition.EventLevel = Extra
#Enabling.#Condition.EventLevel = 'DELAYED'
return
end
/* Here for CALL action after delay. */
/* Values for the CONDITION built-in function. */
#Condition.#NewLevel = #Condition
#ConditionDescription.#NewLevel = #PendingDescription.#Condition.#Level
#ConditionExtra.#NewLevel = #PendingExtra.#Condition.#Level
#ConditionInstruction.#NewLevel = 'CALL'
interpret 'CALL' #TrapName.#Condition.#Level
#Enabling.#Condition.#Level = 'ON'
return /* To clause termination */

9.6.2 Messages during execution

The state function #Message corresponds to constructing a message.
This definition is for the message text in nnn. Translations in which the message inserts are in a different
order are permitted.
In addition to the result defined below, the values of MsgNumber and #LineNumber shall be shown when
a message is output. Also there shall be an indication of whether the error occurred in code executed at
an interactive pause, see nnn.
Messages are shown by writing them to the default error stream.
#Message:
MsgNumber = arg(1)
if #NoSource then MsgNumber = MsgNumber % 1 /* And hence no inserts */
Text = #ErrorText.MsgNumber
Expanded = ''

do Index = 2
parse var Text Begin '<' Insert '>' +1 Text
if Insert = '' then leave
Insert = arg(Index)
if length(Insert) > #Limit_MessageInsert then
Insert = left(Insert,#Limit_MessageInsert)'...'
Expanded = Expanded || Begin || Insert
end

Expanded = Expanded || Begin

say Expanded
return

92
10 Built-in functions
10.1 Notation
The built-in functions are defined mainly through code. The code refers to state variables. This is solely a
notation used in this standard.
The code refers to functions with names that start with 'Config_'; these are the functions described in
section nnn.
The code is specified as an external routine that produces a result from the values #Bif (which is the
name of the built-in function), #Bif_Arg.0 (the number of arguments), #Bif_Arg.i and #Bif_ArgExists.i
(which are the argument data.)
The value of #Level is the value for the clause which invoked the built-in function.
The code either returns the result of the built-in or exits with an indication of a condition that the
invocation of the built-in raises.
The code below uses built-in functions. Such a use invokes another use of this code with a new value of
#Level. On these invocations, the CheckArgs function is not relevant.
Numeric settings as follows are used in the code. When an argument is being checked as a number by
'NUM' or 'WHOLENUM' the settings are those current in the caller. When an argument is being checked
as an integer by an item containing 'WHOLE' the settings are those for the particular built-in function.
Elsewhere the settings have sufficient numeric digits to avoid values which would require exponential
notation.
10.2 Routines used by built-in functions
The routine CheckArgs is concerned with checking the arguments to the built-in. The routines Time2Date
and Leap are for date calculations. ReRadix is used for radix conversion. The routine Raise raises a
condition and does not return.
10.2.1 Argument checking
/* Check arguments. Some further checks will be made in particular built-ins.*/
/* The argument to CheckArgs is a checklist for the allowable arguments. */
/* NUM, WHOLENUM and WHOLE have a side-effect, 'normalizing' the number. */
/* Calls to raise syntax conditions will not return. */

CheckArgs:
CheckList = arg(1) /* This refers to the argument of CheckArgs. */

/* Move the checklist information from a string to individual variables */

ArgType. = ''
ArgPos = 0 /* To count arguments */
MinArgs = 0
do j = 1 to length(CheckList)
ArgPos = ArgPos+1
/* Count the required arguments. */
if substr(CheckList,j,1) == 'r' then MinArgs = MinArgs + 1
/* Collect type information. */
do while j < length(CheckList)
j = j + 1
t = substr(CheckList,j,1)
if t==' ' then leave
ArgType.ArgPos = ArgType.ArgPos || t
end
/* A single space delimits parts. */
end j
MaxArgs = ArgPos

/* Check the number of arguments to the built-in, in this instance. */

NumArgs = #Bif_Arg.0
if NumArgs < MinArgs then call Raise 40.3, MinArgs
if NumArgs > MaxArgs then call Raise 40.4, MaxArgs

/* Check the type(s) of the arguments to the built-in. */

do ArgPos = 1 to NumArgs
if #Bif_ArgExists.ArgPos then
call CheckType
else
if ArgPos <= MinArgs then call Raise 40.5, ArgPos
end ArgPos

93
 /* No errors found by CheckArgs. */
return

CheckType:

Value = #Bif_Arg.ArgPos
Type = ArgType.ArgPos

select
when Type == 'ANY' then nop /* Any string */

when Type == 'NUM' then do /* Any number */

/* This check is made with the caller's digits setting. */
if \Cdatatype(Value, 'N') then
if #DatatypeResult=='E' then call Raise 40.9, ArgPos, Value
else call Raise 40.11, ArgPos, Value
#Bif_Arg.ArgPos=#DatatypeResult /* Update argument copy. */
end

when Type == 'WHOLE' then do /* Whole number */

/* This check is made with digits setting for the built-in. */
if \Edatatype(Value,'W') then
call Raise 40.12, ArgPos, Value
#Bif_Arg.ArgPos=#DatatypeResult
end

when Type == 'WHOLE>=0' then do /* Non-negative whole number */

if \Edatatype(Value,'W') then
call Raise 40.12, ArgPos, Value
if #DatatypeResult < 0 then
call Raise 40.13, ArgPos, Value
#Bif_Arg.ArgPos=#DatatypeResult
end

when Type == 'WHOLE>0' then do /* Positive whole number */

if \Edatatype(Value,'W') then
call Raise 40.12, ArgPos, Value
if #DatatypeResult <= 0 then
call Raise 40.14, ArgPos, Value
#Bif_Arg.ArgPos=#DatatypeResult
end

when Type == 'WHOLENUM' then do /* D2X type whole number */

/* This check is made with digits setting of the caller. */
if \Cdatatype(Value,'W') then
call Raise 40.12, ArgPos, Value
#Bif_Arg.ArgPos=#DatatypeResult
end

when Type == 'WHOLENUM>=0' then do /* D2X Non-negative whole number */

if \Cdatatype(Value,'W') then
call Raise 40.12, ArgPos, Value
if #DatatypeResult < 0 then
call Raise 40.13, ArgPos, Value
#Bif_Arg.ArgPos=#DatatypeResult
end

when Type == '0_90' then do /* Errortext */

if \Edatatype(Value,'N') then
call Raise 40.11, ArgPos, Value
Value=#DatatypeResult
#Bif_Arg.ArgPos=Value
Major=Value % 1
Minor=Value - Major
if Major < 0 | Major > 90 | Minor > .9 | pos('E',Value)>0 then
call Raise 40.17, Value /* ArgPos will be 1 */
end

when Type == 'PAD' then do /* Single character, usually a pad. */

if length(Value) \= 1 then
call Raise 40.23, ArgPos, Value

94
 end

when Type == 'HEX' then /* Hexadecimal string */

if \datatype(Value, 'X') then
call Raise 40.25, Value /* ArgPos will be 1 */

when Type == 'BIN' then /* Binary string */

if \datatype(Value,'B') then
call Raise 40.24, Value /* ArgPos will be 1 */

when Type == 'SYM' then /* Symbol */

if \datatype(Value, 'S') then
call Raise 40.26, Value /* ArgPos will be 1 */

when Type == 'STREAM' then do

call Config_Stream_Qualify Value
if left(#Response, 1) == 'B' then
call Raise 40.27, Value /* ArgPos will be 1 */
end

when Type = 'ACEFILNOR' then do /* Trace */

Val = Value
/* Allow '?' alone */
if Val \== '?' then do
/* Allow leading '?' */
if left(Val,1) == '?' then Val = substr(Val,2)
if pos(translate(left(Val, 1)), 'ACEFILNOR') = 0 then
call Raise 40.28, ArgPos, Type, Val
end
end

otherwise do /* Options */
/* The checklist item is a list of allowed characters */
if Value == '' then
call Raise 40.21, ArgPos
#Bif_Arg.ArgPos = translate(left(Value, 1))
if pos(#Bif_Arg.ArgPos, Type) = 0 then
call Raise 40.28, ArgPos, Type, Value
end

end /* Select */

return

Cdatatype:
/* This check is made with the digits setting of the caller. */
/* #DatatypeResult will be set by use of datatype() */
numeric digits #Digits.#Level
numeric form value #Form.#Level
return datatype(arg(1), arg(2))

Edatatype:
/* This check is made with digits setting for the particular built-in. */
/* #DatatypeResult will be set by use of datatype() */
numeric digits #Bif_Digits.#Bif
numeric form scientific
return datatype(arg(1),arg(2))

10.2.2 Date calculations

Time2Date:
if arg(1) < 0 then
call Raise 40.18
if arg(1) >= 315537897600000000 then
call Raise 40.18
return Time2Date2(arg(1))

Time: procedure
/* This routine is essentially the code from the standard, put in

95
stand-alone form. The only 'tricky bit' is that there is no Rexx way
for it to fail with the same error codes as a "real" implementation
would. It can however give a SYNTAX error, albeit not the desirable
one. This causing of an error is done by returning with no value.
Since the routine will have been called as a function, this produces
an error. */

/* Backslash is avoided as some systems don't handle that negation sign. */

if arg()>3 then
return
numeric digits 18
if arg(1,'E') then
if pos(translate(left(arg(1),1)),"CEHLMNRS")=0 then
return
/* (The standard would also allow 'O' but what this code is running
on would not.) */
if arg(3,'E') then
if pos(translate(left(arg(3),1)),"CHLMNS")=0 then
return
/* If the third argument is given then the second is mandatory. */
if arg(3,'E') & arg(2,'E')=0 then
return
/* Default the first argument. */
if arg(1,'E') then
Option = translate(left(arg(1),1))
else
Option = 'N'
/* If there is no second argument, the current time is returned. */
if arg(2,'E') = 0 then
if arg(1,'E') then
return 'TIME'(arg(1))
else
return 'TIME'()
/* One cannot convert to elapsed times. */
if pos(Option, 'ERO') > 0 then
return
InValue = arg(2)
if arg(3,'E') then
InOption = arg(3)
else
InOption = 'N'
HH = 0
MM = 0
SS = 0
HourAdjust = 0
select

96
 when InOption == 'C' then do
parse var InValue HH ':' . +1 MM +2 XX
if HH = 12 then
HH = 0
if XX == 'pm' then
HourAdjust = 12
end
when InOption == 'H' then
HH = InValue
when InOption == 'L' | InOption == 'N' then
parse var InValue HH ':' MM ':' SS
when InOption == 'M' then
MM = InValue
otherwise
SS = InValue
end
if datatype(HH,'W')=0 | datatype(MM,'W')=0 | datatype(SS,'N')=0 then
return
HH = HH + HourAdjust
/* Convert to microseconds */
Micro = trunc((((HH * 60) + MM) * 60 + SS) * 1000000)
/* There is no special message for time-out-of-range; the bad-format
message is used. */
if Micro<0 | Micro > 24*3600*1000000 then
return
/* Reconvert to further check the original. */
if TimeFormat(Micro,InOption) == InValue then
return TimeFormat(Micro, Option)
return

TimeFormat: procedure
/* Convert from microseconds to given format. */
/* The day will be irrelevant; actually it will be the first day possible. */
x = Time2Date2(arg(1))
parse value x with Year Month Day Hour Minute Second Microsecond Base Days
select
when arg(2) == 'C' then
select
when Hour>12 then
return Hour-12':'right(Minute,2,'0')'pm'
when Hour=12 then
return '12:'right(Minute,2,'0')'pm'
when Hour>0 then
return Hour':'right(Minute,2,'0')'am'
when Hour=0 then
return '12:'right(Minute,2,'0')'am'

97
 end
when arg(2) == 'H' then return Hour
when arg(2) == 'L' then
return right(Hour,2,'0')':'right(Minute,2,'0')':'right(Second,2,'0'),
|| '.'right(Microsecond,6,'0')
when arg(2) == 'M' then
return 60*Hour+Minute
when arg(2) == 'N' then
return right(Hour,2,'0')':'right(Minute,2,'0')':'right(Second,2,'0')
otherwise /* arg(2) == 'S' */
return 3600*Hour+60*Minute+Second
end

Time2Date2: Procedure
/* Convert a timestamp to a date.
Argument is a timestamp (the number of microseconds relative to
0001 01 01 00:00:00.000000)
Returns a date in the form:
year month day hour minute second microsecond base days */

/* Argument is relative to the virtual date 0001 01 01 00:00:00.000000 */

Time = arg(1)

Second = Time % 1000000 ; Microsecond = Time // 1000000

Minute = Second % 60 ; Second = Second // 60
Hour = Minute % 60 ; Minute = Minute // 60
Day = Hour % 24 ; Hour = Hour // 24

/* At this point, the days are the days since the 0001 base date. */
BaseDays = Day
Day = Day + 1

/* Compute either the fitting year, or some year not too far earlier.
Compute the number of days left on the first of January of this year. */
Year = Day % 366
Day = Day - (Year*365 + Year%4 - Year%100 + Year%400)
Year = Year + 1

/* Now if the number of days left is larger than the number of days
in the year we computed, increment the year, and decrement the
number of days accordingly. */
do while Day > (365 + Leap(Year))
Day = Day - (365 + Leap(Year))
Year = Year + 1
end

/* At this point, the days left pertain to this year. */

YearDays = Day

/* Now step through the months, increment the number of the month,
and decrement the number of days accordingly (taking into
consideration that in a leap year February has 29 days), until
further reducing the number of days and incrementing the month
would lead to a negative number of days */
Days = '31 28 31 30 31 30 31 31 30 31 30 31'
do Month = 1 to words(Days)
ThisMonth = Word(Days, Month) + (Month = 2) * Leap(Year)
if Day <= ThisMonth then leave
Day = Day - ThisMonth
end

return Year Month Day Hour Minute Second Microsecond BaseDays YearDays

98
Leap: procedure
/* Return 1 if the year given as argument is a leap year, or 0
otherwise. */
return (arg(1)//4 = 0) & ((arg(1)//100 <> 0) | (arg(1)//400 = 0))

10.2.1 Radix conversion

ReRadix: /* Converts Arg(1) from radix Arg(2) to radix Arg(3) */
procedure
Subject=arg(1)
FromRadix=arg(2)
ToRadix=arg(3)
/* Radix range is 2-16. Conversion is via decimal */
Integer=0
do j=1 to length(Subject)
/* Individual digits have already been checked for range. */
Integer=Integer*FromRadix+pos(substr(Subject,j,1),'0123456789ABCDEF')-1
end
r = ''
do while Integer>0
r = substr('0123456789ABCDEF',1 + Integer // ToRadix, 1) || r
Integer = Integer % ToRadix
end
/* When between 2 and 16, there is no zero suppression. */
if FromRadix = 2 & ToRadix = 16 then
r=right(r, (length(Subject)+3) % 4, '0')
else if FromRadix = 16 & ToRadix = 2 then
r=right(r, length(Subject) * 4, '0')
return r

10.2.2 Raising the SYNTAX condition

Raise:
/* These 40.nn messages always include the built-in name as an insert.*/
call #Raise 'SYNTAX', arg(1), #Bif, arg(2), arg(3), arg(4)
/* #Raise does not return. */

10.1 Character built-in functions

These functions process characters or words in strings. Character positions are numbered from one at
the left. Words are delimited by blanks and their equivalents, word positions are counted from one at the
left.
10.1.1 ABBREV
ABBREV returns '1' if the second argument is equal to the leading characters of the first and the length of
the second argument is not less than the third argument.
call CheckArgs 'rANY rANY oWHOLE>=0'

Subject = #Bif_Arg.1
Subj = #Bif_Arg.2
if #Bif_ArgExists.3 then Length = #Bif_Arg.3
else Length = length(Subj)
Cond1 = length(Subject) >= length(Subj)
Cond2 = length(Subj) >= Length
Cond3 = substr(Subject, 1, length(Subj)) == Subj
return Cond1 & Cond2 & Cond3

10.1.2 CENTER
CENTER returns a string with the first argument centered in it. The length of the result is the second
argument and the third argument specifies the character to be used for padding.
call CheckArgs 'rANY rWHOLE>=0 oPAD'

String = #Bif_Arg.1
Length = #Bif_Arg.2
if #Bif_ArgExists.3 then Pad = #Bif_Arg.3
else Pad = ' '

99
 Trim = length(String) - Length

if Trim > 0 then

return substr(String, Trim % 2 + 1, Length)

return overlay(String, copies(Pad, Length), -Trim % 2 + 1)

10.1.3 CENTRE
This is an alternative spelling for the CENTER built-in function.
10.1.4 CHANGESTR
CHANGESTR replaces all occurrences of the first argument within the second argument, replacing them
with the third argument.
call CheckArgs 'rANY rANY rANY'

Output = ''
Position = 1
do forever
FoundPos = pos(#Bif_Arg.1, #Bif_Arg.2, Position)
if FoundPos = 0 then leave
Output = Output || substr(#Bif_Arg.2, Position, FoundPos - Position),
|| #Bif_Arg.3
Position = FoundPos + length(#Bif_Arg.1)
end
return Output || substr(#Bif_Arg.2, Position)

10.1.5 COMPARE
COMPARE returns '0' if the first and second arguments have the same value. Otherwise, the result is the
position of the first character that is not the same in both strings.
call CheckArgs 'rANY rANY oPAD'

Str1 = #Bif_Arg.1
Str2 = #Bif_Arg.2
if #Bif_ArgExists.3 then Pad = #Bif_Arg.3
else Pad = ' '

/* Compare the strings from left to right one character at a time */

if length(Str1) > length(Str2) then do
Length = length(Str1)
Str2=left(Str2,Length,Pad)
end
else do
Length = length(Str2)
Str1=left(Str1,Length,Pad)
end

do i = 1 to Length
if substr(Str1, i, 1) \== substr(Str2, i, 1) then return i
end
return 0

10.1.6 COPIES
COPIES returns concatenated copies of the first argument. The second argument is the number of
copies.
call CheckArgs 'rANY rWHOLE>=0'

Output = ''
do #Bif_Arg.2
Output = Output || #Bif_Arg.1
end
return Output

10.1.7 COUNTSTR
COUNTSTR counts the appearances of the first argument in the second argument.

100
 call CheckArgs 'rANY rANY'

Output = 0
Position = pos(#Bif_Arg.1,#Bif_Arg.2)
do while Position > 0
Output = Output + 1
Position = pos(#Bif_Arg.1, #Bif_Arg.2, Position + length(#Bif_Arg.1))
end
return Output

10.1.8 DATATYPE
DATATYPE tests for characteristics of the first argument. The second argument specifies the particular
test.
call CheckArgs 'rANY oABLMNSUWX'

/* As well as returning the type, the value for a 'NUM' is set in

#DatatypeResult. This is a convenience when DATATYPE is used
by CHECKARGS. */

String = #Bif_Arg.1

/* If no second argument, DATATYPE checks whether the first is a number. */

if \#Bif_ArgExists.2 then return DtypeOne()

Type = #Bif_Arg.2
/* Null strings are a special case. */
if String == '' then do
if Type == "X" then return 1
if Type == "B" then return 1
return 0
end

/* Several of the options are shorthands for VERIFY */

azl="abcdefghijklmnopqrstuvwxyz"
AZU="ABCDEFGHIJKLMNOPQRSTUVWXYZ"
D09="0123456789"
if Type == "A" then return verify(String,azl||AZU||D09)=0
if Type == "B" then do
/* Check blanks in allowed places. */
if pos(left(String,1),#AllBlanks)>0 then return 0
if pos(right(String,1),#AllBlanks)>0 then return 0
BinaryDigits=0
do j = length(String) by -1 to 1
c = substr(String,j,1)
if pos(c,#AllBlanks)>0 then do
/* Blanks need four BinaryDigits to the right of them. */
if BinaryDigits//4 \= 0 then return 0
end
else do
if verify(c,"01") \= 0 then return 0
BinaryDigits = BinaryDigits + 1
end
end j
return 1
end /* B */
if Type == "L" then return(verify(String,azl)=0)
if Type == "M" then return(verify(String,azl||AZU)=0)
if Type == "N" then return(datatype(String)=="NUM")
if Type == "S" then return(symbol(String)\=='BAD')
if Type == "U" then return(verify(String,AZU)=0)
if Type == "W" then do
/* It may not be a number. */
if DtypeOne(String) == 'CHAR' then return '0'
/* It can be "Whole" even if originally in exponential notation,
provided it can be written as non-exponential. */
if pos('E',#DatatypeResult)>0 then return '0'
/* It won't be "Whole" if there is a non-zero after the decimal point. */
InFraction='0'

101
 do j = 1 to length(String)
c = substr(String,j,1)
if pos(c,'Ee')>0 then leave j
if InFraction & pos(c,'+-')>0 then leave j
if c == '.' then InFraction='1'
else if InFraction & c\=='0' then return 0
end j
/* All tests for Whole passed. */
#DatatypeResult = #DatatypeResult % 1
return 1
end /* W */
/* Type will be "X" */
if pos(left(String,1),#AllBlanks)>0 then return 0
if pos(right(String,1),#AllBlanks)>0 then return 0
HexDigits=0
do j=length(String) by -1 to 1
c=substr(String,j,1)
if pos(c,#AllBlanks)>0 then do
/* Blanks need a pair of HexDigits to the right of them. */
if HexDigits//2 \= 0 then return 0
end
else do
if verify(c,"abcdefABCDEF"D09) \= 0 then return 0
HexDigits=HexDigits+1
end
end
return 1
/* end X */

DtypeOne:
/* See section nnn for the syntax of a number. */
#DatatypeResult = 'S' /* If not syntactically a number */
Residue = strip(String) /* Blanks are allowed at both ends. */
if Residue == '' then return "CHAR"
Sign = ''
if left(Residue,1) == '+' | left(Residue,1) == '-' then do
Sign = left(Residue, 1)
Residue = strip(substr(Residue,2),'L') /* Blanks after sign */
end
if Residue == '' then return "CHAR"
/* Now testing Number, section nnn */
if left(Residue,1) == '.' then do
Residue = substr(Residue, 2)
Before = ''
After = DigitRun()
if After == '' then return "CHAR"
end
else do
Before = DigitRun()
if Before == '' then return "CHAR"
if left(Residue,1) == '.' then do
Residue = substr(Residue, 2)
After = DigitRun()
end
end
Exponent = 0
if Residue \== '' then do
if left(Residue, 1) \== 'e' & left(Residue, 1) \== 'E' then
return "CHAR"
Residue = substr(Residue, 2)
if Residue == '' then return "CHAR"
Esign = ''
if left(Residue, 1) == '+' | left(Residue, 1) == '-' then do
Esign = left(Residue, 1)
Residue = substr(Residue, 2)
if Residue == '' then return "CHAR"
end
Exponent = DigitRun()
if Exponent == '' then return "CHAR"
Exponent = Esign || Exponent
end
if Residue \== '' then return "CHAR"

102
 /*DATATYPE tests for exponent out of range. */
#DatatypeResult = 'E' /* If exponent out of range */
Before = strip(Before,'L','0')
if Before == '' then Before = '0'
Exponent = Exponent + length(Before) -1 /* For SCIENTIFIC */

/* "Engineering notation causes powers of ten to expressed as a

multiple of 3 - the integer part may therefore range from 1 through 9910." */
g = 1
if #Form.#Level == 'E' then do
/* Adjustment to make exponent a multiple of 3 */
g = Exponent//3
if g < 0 then g = g + 3
Exponent = Exponent - g
end

/* Check on the exponent. */

if Exponent > #Limit_ExponentDigits then return "CHAR"
if -#Limit_ExponentDigits > Exponent then return "CHAR"

/* Format to the numeric setting of the caller of DATATYPE */

numeric digits #Digits.#Level
numeric form value #Form.#Level
#DatatypeResult = 0 + #Bif_Arg.1
return "NUM"

DigitRun:
Outcome = ''
do while Residue \== ''
if pos(left(Residue, 1), '0123456789') = 0 then leave
Outcome = Outcome || left(Residue, 1)
Residue = substr(Residue, 2)
end
return Outcome

10.1.9 DELSTR
DELSTR deletes the sub-string of the first argument which begins at the position given by the second
argument. The third argument is the length of the deletion.
call CheckArgs 'rANY rWHOLE>0 oWHOLE>=0'

String = #Bif_Arg.1
Num = #Bif_Arg.2
if #Bif_ArgExists.3 then Len = #Bif_Arg.3

if Num > length(String) then return String

Output = substr(String, 1, Num - 1)

if #Bif_ArgExists.3 then
if Num + Len <= length(String) then
Output = Output || substr(String, Num + Len)
return Output

10.1.10 DELWORD
DELWORD deletes words from the first argument. The second argument specifies position of the first
word to be deleted and the third argument specifies the number of words.
call CheckArgs 'rANY rWHOLE>0 oWHOLE>=0'

String = #Bif_Arg.1
Num = #Bif_Arg.2
if #Bif_ArgExists.3 then Len = #Bif_Arg.3

if Num > words(String) then return String

EndLeft = wordindex(String, Num) - 1

Output = left(String, EndLeft)
if #Bif_ArgExists.3 then do
BeginRight = wordindex(String, Num + Len)

103
 if BeginRight>0 then
Output = Output || substr(String, BeginRight)
end
return Output

10.1.11 INSERT
INSERT insets the first argument into the second. The third argument gives the position of the character
before the insert and the fourth gives the length of the insert. The fifth is the padding character.
call CheckArgs 'rANY rANY oWHOLE>=0 oWHOLE>=0 oPAD'

New = #Bif_Arg.1
Target = #Bif_Arg.2
if #Bif_ArgExists.3 then Num = #Bif_Arg.3
else Num = 0
if #Bif_ArgExists.4 then Length = #Bif_Arg.4
else Length = length(New)
if #Bif_ArgExists.5 then Pad = #Bif_Arg.5
else Pad = ' '

return left(Target, Num, Pad), /* To left of insert */

|| left(New, Length, Pad), /* New string inserted */
|| substr(Target, Num + 1) /* To right of insert */

10.1.12 LASTPOS
LASTPOS returns the position of the last occurrence of the first argument within the second. The third
argument is a starting position for the search.
call CheckArgs 'rANY rANY oWHOLE>0'

Needle = #Bif_Arg.1
Haystack = #Bif_Arg.2
if #Bif_ArgExists.3 then Start = #Bif_Arg.3
else Start = length(Haystack)

NeedleLength = length(Needle)
if NeedleLength = 0 then return 0
Start = Start - NeedleLength + 1
do i = Start by -1 while i > 0
if substr(Haystack, i, NeedleLength) == Needle then return i
end i
return 0

10.1.13 LEFT
LEFT returns characters that are on the left of the first argument. The second argument specifies the
length of the result and the third is the padding character.
call CheckArgs 'rANY rWHOLE>=0 oPAD'

if #Bif_ArgExists.3 then Pad = #Bif_Arg.3

else Pad = ' '

return substr(#Bif_Arg.1, 1, #Bif_Arg.2, Pad)

10.1.14 LENGTH
Length returns a count of the number of characters in the argument.
call CheckArgs 'rANY'

String = #Bif_Arg.1

#Response = Config_Length(String)
Length = #Outcome
call Config_Substr #Response, 1
if #Outcome \== 'E' then return Length
/* Here if argument was not a character string. */
call Config_C2B String
call #Raise 'SYNTAX', 23.1, b2x(#Outcome)

104
 /* No return to here */

10.1.15 OVERLAY
OVERLAY overlays the first argument onto the second. The third argument is the starting position of the
overlay. The fourth argument is the length of the overlay and the fifth is the padding character.
call CheckArgs 'rANY rANY oWHOLE>0 oWHOLE>=0 oPAD'

New = #Bif_Arg.1
Target = #Bif_Arg.2
if #Bif_ArgExists.3 then Num = #Bif_Arg.3
else Num = 1
if #Bif_ArgExists.4 then Length = #Bif_Arg.4
else Length = length(New)
if #Bif_ArgExists.5 then Pad = #Bif_Arg.5
else Pad = ' '

return left(Target, Num - 1, Pad), /* To left of overlay */

|| left(New, Length, Pad), /* New string overlaid */
|| substr(Target, Num + Length) /* To right of overlay */

10.1.16 POS
POS returns the position of the first argument within the second.
call CheckArgs 'rANY rANY oWHOLE>0'

Needle = #Bif_Arg.1
Haystack = #Bif_Arg.2
if #Bif_ArgExists.3 then Start = #Bif_Arg.3
else Start = 1

if length(Needle) = 0 then return 0

do i = Start to length(Haystack)+1-length(Needle)
if substr(Haystack, i, length(Needle)) == Needle then return i
end i
return 0

10.1.17 REVERSE
REVERSE returns its argument, swapped end for end.
call CheckArgs 'rANY'

String = #Bif_Arg.1

Output = ''
do i = 1 to length(String)
Output = substr(String,i,1) || Output
end
return Output

10.1.18 RIGHT
RIGHT returns characters that are on the right of the first argument. The second argument specifies the
length of the result and the third is the padding character.
call CheckArgs 'rANY rWHOLE>=0 oPAD'

String = #Bif_Arg.1
Length = #Bif_Arg.2
if #Bif_ArgExists.3 then Pad = #Bif_Arg.3
else Pad = ' '

Trim = length(String) - Length

if Trim >= 0 then return substr(String,Trim + 1)
return copies(Pad, -Trim) || String /* Pad string on the left */

10.1.19 SPACE

105
SPACE formats the blank-delimited words in the first argument with pad characters between each word.
The second argument is the number of pad characters between each word and the third is the pad
character.
call CheckArgs 'rANY oWHOLE>=0 oPAD'

String = #Bif_Arg.1
if #Bif_ArgExists.2 then Num = #Bif_Arg.2
else Num = 1
if #Bif_ArgExists.3 then Pad = #Bif_Arg.3
else Pad = ' '

Padding = copies(Pad, Num)

Output = subword(String, 1, 1)
do i = 2 to words(String)
Output = Output || Padding || subword(String, i, 1)
end
return Output

10.1.20 STRIP
STRIP removes characters from its first argument. The second argument specifies whether the deletions
are leading characters, trailing characters or both. Each character deleted is equal to the third argument,
or equivalent to a blank if the third argument is omitted.
call CheckArgs 'rANY oLTB oPAD'

String = #Bif_Arg.1
if #Bif_ArgExists.2 then Option = #Bif_Arg.2
else Option = 'B'
if #Bif_ArgExists.3 then Unwanted = #Bif_Arg.3
else Unwanted = #AllBlanks<Index "#AllBlanks" # "" >

if Option == 'L' | Option == 'B' then do

/* Strip leading characters */
do while String \== '' & pos(left(String, 1), Unwanted) > 0
String = substr(String, 2)
end
end

if Option == 'T' | Option == 'B' then do

/* Strip trailing characters */
do while String \== '' & pos(right(String, 1), Unwanted) > 0
String = left(String, length(String)-1)
end /* of while */
end
return String

10.1.21 SUBSTR
SUBSTR returns a sub-string of the first argument. The second argument specifies the position of the
first character and the third specifies the length of the sub-string. The fourth argument is the padding
character.
call CheckArgs 'rANY rWHOLE>0 oWHOLE>=0 oPAD'

String = #Bif_Arg.1
Num = #Bif_Arg.2
if #Bif_ArgExists.3 then Length = #Bif_Arg.3
else Length = max(length(String)+1-Num,0)
if #Bif_ArgExists.4 then Pad = #Bif_Arg.4
else Pad = ' '

Output = ''
do Length
#Response = Config_Substr(String,Num) /* Attempt to fetch character.*/
Character = #Outcome
Num = Num + 1
call Config_Substr #Response,1 /* Was there such a character? */
if #Outcome == 'E' then do

106
 /* Here if argument was not a character string. */

call Config_C2B String

call #Raise 'SYNTAX', 23.1, b2x(#Outcome)
/* No return to here */
end
if #Outcome == 'M' then Character = Pad
Output=Output||Character
end
return Output

10.1.22 SUBWORD
SUBWORD returns a sub-string of the first argument, comprised of words. The second argument is the
position in the first argument of the first word of the sub-string. The third argument is the number of
words in the sub-string.
call CheckArgs 'rANY rWHOLE>0 oWHOLE>=0'

String = #Bif_Arg.1
Num = #Bif_Arg.2
if #Bif_ArgExists.3 then Length = #Bif_Arg.3
else Length = length(String) /* Avoids call */
/* to WORDS() */
if Length = 0 then return ''
/* Find position of first included word */
Start = wordindex(String,Num)
if Start = 0 then return '' /* Start is beyond end */

/* Find position of first excluded word */

End = wordindex(String,Num+Length)
if End = 0 then End = length(String)+1

Output=substr(String,Start,End-Start)

/* Drop trailing blanks */

do while Output \== ''
if pos(right(Output,1),#AllBlanks) = 0 then leave
Output = left(Output,length(Output)-1)
end
return Output

10.1.23 TRANSLATE
TRANSLATE returns the characters of its first argument with each character either unchanged or
translated to another character.
call CheckArgs 'rANY oANY oANY oPAD'
String = #Bif_Arg.1
/* If neither input nor output tables, uppercase. */
if \#Bif_ArgExists.2 & \#Bif_ArgExists.3 then do
Output = ''
do j=1 to length(String)
#Response = Config_Upper(substr(String,j,1))
Output = Output || #Outcome
end j
return Output
end
/* The input table defaults to all characters. */
if \#Bif_ArgExists.3 then do
#Response = Config_Xrange()
Tablei = #Outcome
end
else Tablei = #Bif_Arg.3
/* The output table defaults to null */
if #Bif_ArgExists.2 then Tableo = #Bif_Arg.2
else Tableo = ''
/* The tables are made the same length */
if #Bif_ArgExists.4 then Pad = #Bif_Arg.4
else Pad = ' '
Tableo=left(Tableo,length(Tablei),Pad)

107
 Output=''
do j=1 to length(String)
c=substr(String,j,1)
k=pos(c,Tablei)
if k=0 then Output=Output||c
else Output=Output||substr(Tableo,k,1)
end j
return Output

10.1.24 VERIFY
VERIFY checks that its first argument contains only characters that are in the second argument, or that it
contains no characters from the second argument; the third argument specifies which check is made.
The result is '0', or the position of the character that failed verification. The fourth argument is a starting
position for the check.
call CheckArgs 'rANY rANY oMN oWHOLE>0'

String = #Bif_Arg.1
Reference = #Bif_Arg.2
if #Bif_ArgExists.3 then Option = #Bif_Arg.3
else Option = 'N'
if #Bif_ArgExists.4 then Start = #Bif_Arg.4
else Start = 1

Last = length(String)
if Start > Last then return 0
if Reference == '' then
if Option == 'N' then return Start
else return 0

do i = Start to Last
t = pos(substr(String, i, 1), Reference)
if Option == 'N' then do
if t = 0 then return i /* Return position of NoMatch character. */
end
else
if t > 0 then return i /* Return position of Matched character. */
end i
return 0

10.1.25 WORD
WORD returns the word from the first argument at the position given by the second argument.
call CheckArgs 'rANY rWHOLE>0'

return subword(#Bif_Arg.1, #Bif_Arg.2, 1)

10.1.26 WORDINDEX
WORDINDEX returns the character position in the first argument of a word in the first argument. The
second argument is the word position of that word.
call CheckArgs 'rANY rWHOLE>0'

String = #Bif_Arg.1
Num = #Bif_Arg.2

/* Find starting position */

Start = 1
Count = 0
do forever
Start = verify(String, #AllBlanks<Index "#AllBlanks" # "" >, 'N', Start) /*
Find non-blank */
if Start = 0 then return 0 /* Start is beyond end */
Count = Count + 1 /* Words found */
if Count = Num then leave
Start = verify(String, #AllBlanks<Index "#AllBlanks" # "" >, 'M', Start + 1) /*
Find blank */

108
 if Start = 0 then return 0 /* Start is beyond end */
end
return Start

10.1.27 WORDLENGTH
WORDLENGTH returns the number of characters in a word from the first argument. The second
argument is the word position of that word.
call CheckArgs 'rANY rWHOLE>0'

return length(subword(#Bif_Arg.1, #Bif_Arg.2, 1))

10.1.28 WORDPOS
WORDPOS finds the leftmost occurrence in the second argument of the sequence of words in the first
argument. The result is '0' or the word position in the second argument of the first word of the matched
sequence. Third argument is a word position for the start of the search.
call CheckArgs 'rANY rANY oWHOLE>0'

Phrase = #Bif_Arg.1
String = #Bif_Arg.2
if #Bif_ArgExists.3 then Start = #Bif_Arg.3
else Start = 1

Phrase = space(Phrase)
PhraseWords = words(Phrase)
if PhraseWords = 0 then return 0
String = space(String)
StringWords = words(String)
do WordNumber = Start to StringWords - PhraseWords + 1
if Phrase == subword(String, WordNumber, PhraseWords) then
return WordNumber
end WordNumber
return 0

10.1.29 WORDS
WORDS counts the number of words in its argument.
call CheckArgs 'rANY'

do Count = 0 by 1
if subword(#Bif_Arg.1, Count + 1) == '' then return Count
end Count

10.1.30 XRANGE
XRANGE returns an ordered string of all valid character encodings in the specified range.
call CheckArgs 'oPAD oPAD'

if \#Bif_ArgExists.1 then #Bif_Arg.1 = ''

if \#Bif_ArgExists.2 then #Bif_Arg.2 = ''
#Response = Config_Xrange(#Bif_Arg.1, #Bif_Arg.2)
return #Outcome

10.2 Arithmetic built-in functions

These functions perform arithmetic at the numeric settings current at the invocation of the built-in
function. Note that CheckArgs formats any 'NUM' (numeric) argument.
10.2.1 ABS
ABS returns the absolute value of its argument.
call CheckArgs 'rNUM'

Number=#Bif_Arg.1
if left(Number,1) = '-' then Number = substr(Number,2)
return Number

10.2.2 FORMAT

109
FORMAT formats its first argument. The second argument specifies the number of characters to be
used for the integer part and the third specifies the number of characters for the decimal part. The fourth
argument specifies the number of characters for the exponent and the fifth determines when exponential
notation is used.
call CheckArgs,
'rNUM oWHOLE>=0 oWHOLE>=0 oWHOLE>=0 oWHOLE>=0'

if #Bif_ArgExists.2 then Before = #Bif_Arg.2

if #Bif_ArgExists.3 then After = #Bif_Arg.3
if #Bif_ArgExists.4 then Expp = #Bif_Arg.4
if #Bif_ArgExists.5 then Expt = #Bif_Arg.5
/* In the simplest case the first is the only argument. */
Number=#Bif_Arg.1
if #Bif_Arg.0 < 2 then return Number

/* Dissect the Number. It is in the normal Rexx format. */

parse var Number Mantissa 'E' Exponent
if Exponent == '' then Exponent = 0
Sign = 0
if left(Mantissa,1) == '-' then do
Sign = 1
Mantissa = substr(Mantissa,2)
end
parse var Mantissa Befo '.' Afte
/* Count from the left for the decimal point. */
Point = length(Befo)
/* Sign Mantissa and Exponent now reflect the Number. Befo Afte and
Point reflect Mantissa. */

/* The fourth and fifth arguments allow for exponential notation. */

/* Decide whether exponential form to be used, setting ShowExp. */
ShowExp = 0
if #Bif_ArgExists.4 | #Bif_ArgExists.5 then do
if \#Bif_ArgExists.5 then Expt = #Digits.#Level
/* Decide whether exponential form to be used. */
if (Point + Exponent) > Expt then ShowExp = 1 /* Digits before rule. */
LeftOfPoint = 0
if length(Befo) > 0 then LeftOfPoint = Befo /* Value left of
the point */

/* Digits after point rule for exponentiation: */

/* Count zeros to right of point. */

z = 0
do while substr(Afte,z+1,1) == '0'
z = z + 1
end
if LeftOfPoint = 0 & (z - Exponent) > 5 then ShowExp = 1

/* An extra rule for exponential form: */

if #Bif_ArgExists.4 then if Expp = 0 then ShowExp = 0

/* Construct the exponential part of the result. */

if ShowExp then do
Exponent = Exponent + ( Point - 1 )
Point = 1 /* As required for 'SCIENTIFIC' */
if #Form.#Level == 'ENGINEERING' then
do while Exponent//3 \= 0
Point = Point+1
Exponent = Exponent-1
end
end
if \ShowExp then Point = Point + Exponent
end /* Expp or Expt given */
else do
/* Even if Expp and Expt are not given, exponential notation will
be used if the original Number+0 done by CheckArgs led to it. */
if Exponent \= 0 then do
ShowExp = 1

110
 end
end

/* ShowExp now indicates whether to show an exponent,

Exponent is its value. */
/* Make this a Number without a point. */
Integer = Befo||Afte
/* Make sure Point position isn't disjoint from Integer. */
if Point<1 then do /* Extra zeros on the left. */
Integer = copies('0',1 - Point) || Integer
Point = 1
end
if Point > length(Integer) then
Integer = left(Integer,Point,'0') /* And maybe on the right. */

/* Deal with right of decimal point first since that can affect the
left. Ensure the requested number of digits there. */
Afters = length(Integer)-Point
if #Bif_ArgExists.3 = 0 then After = Afters /* Note default. */
/* Make Afters match the requested After */
do while Afters < After
Afters = Afters+1
Integer = Integer'0'
end
if Afters > After then do
/* Round by adding 5 at the right place. */
r=substr(Integer, Point + After + 1, 1)
Integer = left(Integer, Point + After)
if r >= '5' then Integer = Integer + 1
/* This can leave the result zero. */
If Integer = 0 then Sign = 0
/* The case when rounding makes the integer longer is an awkward
one. The exponent will have to be adjusted. */
if length(Integer) > Point + After then do
Point = Point+1
end
if ShowExp = 1 then do
Exponent=Exponent + (Point - 1)
Point = 1 /* As required for 'SCIENTIFIC' */
if form() = 'ENGINEERING' then
do while Exponent//3 \= 0
Point = Point+1
Exponent = Exponent-1
end
end
t = Point-length(Integer)
if t > 0 then Integer = Integer||copies('0',t)
end /* Rounded */
/* Right part is final now. */
if After > 0 then Afte = '.'||substr(Integer,Point+1,After)
else Afte = ''

/* Now deal with the integer part of the result. */

Integer = left(Integer,Point)
if #Bif_ArgExists.2 = 0 then Before = Point + Sign /* Note default. */
/* Make Point match Before */
if Point > Before - Sign then call Raise 40.38, 2, #Bif_Arg.1
do while Point<Before
Point = Point+1
Integer = '0'Integer
end

/* Find the Sign position and blank leading zeroes. */

r = ''
Triggered = 0
do j = 1 to length(Integer)
Digit = substr(Integer,j,1)
/* Triggered is set when sign inserted or blanking finished. */
if Triggered = 1 then do
r = r||Digit
iterate
end

111
 /* If before sign insertion point then blank out zero. */
if Digit = '0' then
if substr(Integer,j+1,1) = '0' & j+1<length(Integer) then do
r = r||' '
iterate
end
/* j is the sign insertion point. */
if Digit = '0' & j \= length(Integer) then Digit = ' '
if Sign = 1 then Digit = '-'
r = r||Digit
Triggered = 1
end j
Number = r||Afte

if ShowExp = 1 then do
/* Format the exponent. */
Expart = ''
SignExp = 0
if Exponent<0 then do
SignExp = 1
Exponent = -Exponent
end
/* Make the exponent to the requested width. */
if #Bif_ArgExists.4 = 0 then Expp = length(Exponent)
if length(Exponent) > Expp then
call Raise 40.38, 4, #Bif_Arg.1
Exponent=right(Exponent,Expp,'0')
if Exponent = 0 then do
if #Bif_ArgExists.4 then Expart = copies(' ',expp+2)
end
else if SignExp = 0 then Expart = 'E+'Exponent
else Expart = 'E-'Exponent
Number = Number||Expart
end
return Number

10.2.3 MAX
MAX returns the largest of its arguments.
if #Bif_Arg.0 <1 then
call Raise 40.3, 1
call CheckArgs 'rNUM'||copies(' rNUM', #Bif_Arg.0 - 1)

Max = #Bif_Arg.1
do i = 2 to #Bif_Arg.0 by 1
Next = #Bif_Arg.i
if Max < Next then Max = Next
end i
return Max

10.2.4 MIN
MIN returns the smallest of its arguments.
if #Bif_Arg.0 <1 then
call Raise 40.3, 1
call CheckArgs 'rNUM'||copies(' rNUM', #Bif_Arg.0 - 1)

Min = #Bif_Arg.1
do i = 2 to #Bif_Arg.0 by 1
Next = #Bif_Arg.i
if Min > Next then Min = Next
end i
return Min

10.2.5 SIGN
SIGN returns '1', '0' or '-1' according to whether its argument is greater than, equal to, or less than zero.
call CheckArgs 'rNUM'

Number = #Bif_Arg.1

112
 select
when Number < 0 then Output = -1
when Number = 0 then Output = 0
when Number > 0 then Output = 1
end
return Output

10.2.6 TRUNC
TRUNC returns the integer part of its argument, or the integer part plus a number of digits after the
decimal point, specified by the second argument.
call CheckArgs 'rNUM oWHOLE>=0'

Number = #Bif_Arg.1
if #Bif_ArgExists.2 then Num = #Bif_Arg.2
else Num = 0

Integer =(10**Num * Number)%1

if Num=0 then return Integer

t=length(Integer)-Num
if t<=0 then return '0.'right(Integer,Num,'0')
else return insert('.',Integer,t)

10.3 State built-in functions

These functions return values from the state of the execution.
10.3.1 ADDRESS
ADDRESS returns the name of the environment to which commands are currently being submitted.
Optionally, under control by the argument, it also returns information on the targets of command output
and the source of command input.
call CheckArgs 'oEINO'

if #Bif_ArgExists.1 then Option1 = #Bif_Arg.1

else Option1='N'

if Option1 == 'N' then return #Env_Name.ACTIVE.#Level

Tail = Option1'.ACTIVE.'#Level
return #Env_Position.Tail #Env_Type.Tail #Env_Resource.Tail

10.3.2 ARG
ARG returns information about the argument strings to a program or routine, or the value of one of those
strings.
ArgData = 'oWHOLE>0 oENO'
if #Bif_ArgExists.2 then ArgData = 'rWHOLE>0 rENO'
call CheckArgs ArgData

if \#Bif_ArgExists.1 then return #Arg.#Level.0

ArgNum=#Bif_Arg.1
if \#Bif_ArgExists.2 then return #Arg.#Level.ArgNum
if #Bif_Arg.2 =='O' then return \#ArgExists.#Level.ArgNum
else return #ArgExists.#Level.ArgNum

10.3.3 CONDITION
CONDITION returns information associated with the current condition.
call CheckArgs 'oCDEIS'

/* Values are null if this is not following a condition. */

if #Condition.#Level == '' then do
#ConditionDescription.#Level = ''
#ConditionExtra.#Level = ''
#ConditionInstruction.#Level = ''

113
 end

Option=#Bif_Arg.1
if Option=='C' then return #Condition.#Level
if Option=='D' then return #ConditionDescription.#Level
if Option=='E' then return #ConditionExtra.#Level
if Option=='I' then return #ConditionInstruction.#Level
/* State is the current state. */
if #Condition.#Level = '' then return ""
return #Enabling.#Condition.#Level

10.3.4 DIGITS
DIGITS returns the current setting of NUMERIC DIGITS.
call CheckArgs ''

return #Digits.#Level

10.3.5 ERRORTEXT
ERRORTEXT returns the unexpanded text of the message which is identified by the first argument. A
second argument of 'S' selects the standard English text, otherwise the text may be translated to another
national language. This translation is not shown in the code below.
call CheckArgs 'r0_90 oSN'

msgcode = #Bif_Arg.1
if #Bif_ArgExists.2 then Option = #Bif_Arg.2
else Option = 'N'
return #ErrorText.msgcode

10.3.6 FORM
FORM returns the current setting of NUMERIC FORM.
call CheckArgs ''

return #Form.#Level

10.3.7 FUZZ
FUZZ returns the current setting of NUMERIC FUZZ.
call CheckArgs ''

return #Fuzz.#Level

10.3.8 SOURCELINE
If there is no argument, SOURCELINE returns the number of lines in the program, or '0' if the source
program is not being shown on this execution. If there is an argument it specifies the number of the line
of the source program to be returned.
call CheckArgs 'oWHOLE>0'

if \#Bif_ArgExists.1 then return #SourceLine.0

Num = #Bif_Arg.1
if Num > #SourceLine.0 then
call Raise 40.34, Num, #SourceLine.0
return #SourceLine.Num

10.3.9 TRACE
TRACE returns the trace setting currently in effect, and optionally alters the setting.
call CheckArgs 'oACEFILNOR' /* Also checks for '?' */

/* With no argument, this a simple query. */

Output=#Tracing.#Level
if #Interactive.#Level then Output = '?'||Output
if \#Bif_ArgExists.1 then return Output

114
 Value=#Bif_Arg.1
#Interactive.#Level=0
/* A question mark sets the interactive flag. */
if left(Value,1)=='?' then do
#Interactive.#Level = 1
Value=substr(Value,2)
end
/* Absence of a letter leaves the setting unchanged. */
if Value\=='' then do
Value=translate(left(Value,1))
if Value=='O' then #Interactive.#Level='0'
#Tracing.#Level = Value
end
return Output

10.4 Conversion built-in functions

Conversions between Binary form, Decimal form, and heXadecimal form do not depend on the encoding
(see nnn) of the character data.
Conversion to Coded form gives a result which depends on the encoding. Depending on the encoding,
the result may be a string that does not represent any sequence of characters.
10.4.1 B2X
B2X performs binary to hexadecimal conversion.
call CheckArgs 'rBIN'

String = space(#Bif_Arg.1,0)
return ReRadix(String,2,16)

10.4.2 BITAND
The functions BITAND, BITOR and BITXOR operate on encoded character data. Each binary digit from
the encoding of the first argument is processed in conjunction with the corresponding bit from the second
argument.
call CheckArgs 'rANY oANY oPAD'

String1 = #Bif_Arg.1
if #Bif_ArgExists.2 then String2 = #Bif_Arg.2
else String2 = ''

/* Presence of a pad implies character strings. */

if #Bif_ArgExists.3 then
if length(String1) > length(String2) then
String2=left(String2,length(String1),#Bif_Arg.3)
else
String1=left(String1,length(String2),#Bif_Arg.3)

/* Change to manifest bit representation. */

#Response=Config_C2B(String1)
String1=#Outcome
#Response=Config_C2B(String2)
String2=#Outcome
/* Exchange if necessary to make shorter second. */
if length(String1)<length(String2) then do
t=String1
String1=String2
String2=t
end

/* Operate on common length of those bit strings. */

r=''
do j=1 to length(String2)
b1=substr(String1,j,1)
b2=substr(String2,j,1)
select
when #Bif='BITAND' then
b1=b1&b2
when #Bif='BITOR' then
b1=b1|b2

115
 when #Bif='BITXOR' then
b1=b1&&b2
end
r=r||b1
end j
r=r || right(String1,length(String1)-length(String2))

/* Convert back to encoded characters. */

return x2c(b2x(r))

10.4.3 BITOR
See nnn
10.4.4 BITXOR
See nnn
10.4.5 C2D
C2D performs coded to decimal conversion.
call CheckArgs 'rANY oWHOLE>=0'

if length(#Bif_Arg.1)=0 then return 0

if #Bif_ArgExists.2 then do
/* Size specified */
Size = #Bif_Arg.2
if Size = 0 then return 0
/* Pad will normally be zeros */
t=right(#Bif_Arg.1,Size,left(xrange(),1))
/* Convert to manifest bit */
call Config_C2B t
/* And then to signed decimal. */
Sign = Left(#Outcome,1)
#Outcome = substr(#Outcome,2)
t=ReRadix(#Outcome,2,10)
/* Sign indicates 2s-complement. */
if Sign then t=t-2**length(#Outcome)
if abs(t) > 10 ** #Digits.#Level - 1 then call Raise 40.35, t
return t
end
/* Size not specified. */
call Config_C2B #Bif_Arg.1
t = ReRadix(#Outcome,2,10)
if t > 10 ** #Digits.#Level - 1 then call Raise 40.35, t
return t

10.4.6 C2X
C2X performs coded to hexadecimal conversion.
call CheckArgs 'rANY'

if length(#Bif_Arg.1) = 0 then return ''

call Config_C2B #Bif_Arg.1
return ReRadix(#Outcome,2,16)

10.4.7 D2C
D2C performs decimal to coded conversion.
if \#Bif_ArgExists.2 then ArgData = 'rWHOLENUM>=0'
else ArgData = 'rWHOLENUM rWHOLE>=0'
call CheckArgs ArgData

/* Convert to manifest binary */

Subject = abs(#Bif_Arg.1)
r = ReRadix(Subject,10,2)
/* Make length a multiple of 8, as required for Config_B2C */
Length = length(r)
do while Length//8 \= 0
Length = Length+1
end
r = right(r,Length,'0')

116
 /* 2s-complement for negatives. */
if #Bif_Arg.1<0 then do
Subject = 2**length(r)-Subject
r = ReRadix(Subject,10,2)
end
/* Convert to characters */
#Response = Config_B2C(r)
Output = #Outcome
if \#Bif_ArgExists.2 then return Output

/* Adjust the length with appropriate characters. */

if #Bif_Arg.1>=0 then return right(Output,#Bif_Arg.2,left(xrange(),1))
else return right(Output,#Bif_Arg.2,right(xrange(),1))

10.4.8 D2X
D2X performs decimal to hexadecimal conversion.
if \#Bif_ArgExists.2 then ArgData = 'rWHOLENUM>=0'
else ArgData = 'rWHOLENUM rWHOLE>=0'
call CheckArgs ArgData

/* Convert to manifest hexadecimal */

Subject = abs(#Bif_Arg.1 )
r = ReRadix(Subject,10,16)
/* Twos-complement for negatives */
if #Bif_Arg.1<0 then do
Subject = 16**length(r)-Subject
r = ReRadix(Subject,10,16)
end
if \#Bif_ArgExists.2 then return r
/* Adjust the length with appropriate characters. */
if #Bif_Arg.1>=0 then return right(r,#Bif_Arg.2,'0')
else return right(r,#Bif_Arg.2,'F')

10.4.9 X2B
X2B performs hexadecimal to binary conversion.
call CheckArgs 'rHEX'

Subject = #Bif_Arg.1
if Subject == '' then return ''
/* Blanks were checked by CheckArgs, here they are ignored. */
Subject = space(Subject,0)
return ReRadix(translate(Subject),16,2)

10.4.10 X2C
X2C performs hexadecimal to coded character conversion.
call CheckArgs 'rHEX'

Subject = #Bif_Arg.1
if Subject == '' then return ''
Subject = space(Subject,0)
/* Convert to manifest binary */
r = ReRadix(translate(Subject),16,2)
/* Convert to character */
Length = 8*((length(Subject)+1)%2)
#Response = Config_B2C(right(r,Length,'0'))
return #Outcome

10.4.11 X2D
X2D performs hexadecimal to decimal conversion.
call CheckArgs 'rHEX oWHOLE>=0'

Subject = #Bif_Arg.1
if Subject == '' then return '0'

Subject = translate(space(Subject,0))

117
 if #Bif_ArgExists.2 then
Subject = right(Subject,#Bif_Arg.2,'0')
if Subject =='' then return '0'
/* Note the sign */
if #Bif_ArgExists.2 then SignBit = left(x2b(Subject),1)
else SignBit = '0'
/* Convert to decimal */
r = ReRadix(Subject,16,10)
/* Twos-complement */
if SignBit then r = 2**(4*#Bif_Arg.2) - r
if abs(r)>10 ** #Digits.#Level - 1 then call Raise 40.35, t
return r

10.5 Input/Output built-in functions

The configuration shall provide the ability to access streams. Streams are identified by character string
identifiers and provide for the reading and writing of data. They shall support the concepts of characters,
lines, and positioning. The input/output built-in functions interact with one another, and they make use of
Config_ functions, see nnn. When the operations are successful the following characteristics shall be
exhibited:
- The CHARIN/CHAROUT functions are insensitive to the lengths of the arguments. The data written
to a stream by CHAROUT can be read by a different number of CHARINs.
- The CHARIN/CHAROUT functions are reflective, that is, the concatenation of the data read from a
persistent stream by CHARIN (after positioning to 1, while CHARS(Stream)>0), will be the same as
the concatenation of the data put by CHAROUT.
- All characters can be used as CHARIN/CHAROUT data.
- The CHARS(Stream, 'N') function will return zero only when a subsequent read (without positioning)
is guaranteed to raise the NOTREADY condition.
- The LINEIN/LINEOUT functions are sensitive to the length of the arguments, that is, the length of a
line written by LINEOUT is the same as the length of the string returned by successful LINEIN of the
line.
- Some characters, call them line-banned characters, cannot reliably be used as data for
LINEIN/LINEOUT. If these are not used, LINEIN/LINEOUT is reflective. If they are used, the result is
not defined. The set of characters which are line-barred is a property of the configuration.
- The LINES(Stream, 'N') function will return zero only when a subsequent LINEIN (without
positioning) is guaranteed to raise the NOTREADY condition.
- When a persistent stream is repositioned and written to with CHAROUT, the previously written data
is not lost, except for the data overwritten by this latest CHAROUT.
- When a persistent stream is repositioned and written to with LINEOUT, the previously written data is
not lost, except for the data overwritten by this latest LINEOUT, which may leave lines partially
overwritten.
10.5.1 CHARIN
CHARIN returns a string read from the stream named by the first argument.
call CheckArgs 'oSTREAM oWHOLE>0 oWHOLE>=0'

if #Bif_ArgExists.1 then Stream = #Bif_Arg.1

else Stream = ''
#StreamState.Stream = ''
/* Argument 2 is positioning. */
if #Bif_ArgExists.2 then do
#Response = Config_Stream_Position(Stream,'CHARIN',#Bif_Arg.2)
if left(#Response, 1) == 'R' then call Raise 40.41, 2, #Bif_Arg.2
if left(#Response, 1) == 'T' then call Raise 40.42,Stream
end
/* Argument 3 is how many. */
if #Bif_ArgExists.3 then Count = #Bif_Arg.3
else Count = 1
if Count = 0 then do
call Config_Stream_Charin Stream, 'NULL' /* "Touch" the stream */
return ''
end
/* The unit may be eight bits (as characters) or one character. */
call Config_Stream_Query Stream

118
 Mode = #Outcome
r = ''
do until Count = 0
#Response = Config_Stream_Charin(Stream, 'CHARIN')
if left(#Response, 1) \== 'N' then do
if left(#Response, 1) == 'E' then #StreamState.Stream = 'ERROR'
/* This call will return. */
call #Raise 'NOTREADY', Stream, substr(#Response, 2)
leave
end
r = r||#Outcome
Count = Count-1
end
if Mode == 'B' then do
call Config_B2C r
r = #Outcome
end
return r

10.5.2 CHAROUT
CHAROUT returns the count of characters remaining after attempting to write the second argument to the
stream named by the first argument.
call CheckArgs 'oSTREAM oANY oWHOLE>0'

if #Bif_ArgExists.1 then Stream = #Bif_Arg.1

else Stream = ''

#StreamState.Stream = ''
if \#Bif_ArgExists.2 & \#Bif_ArgExists.3 then do
/* Position to end of stream. */
#Response = Config_Stream_Close(Stream)
if left(#Response,1) == 'T' then call Raise 40.42,Stream
return 0
end

if #Bif_ArgExists.3 then do
/* Explicit positioning. */
#Response = Config_Stream_Position(Stream,'CHAROUT', #Bif_Arg.3)
if left(#Response,1) == 'T' then call Raise 40.42,Stream
if left(#Response, 1) == 'R' then call Raise 40.41, 3, #Bif_Arg.3
end

if \#Bif_ArgExists.2 | #Bif_Arg.2 == '' then do

call Config_Stream_Charout Stream, 'NULL' /* "Touch" the stream */
return 0
end

String = #Bif_Arg.2
call Config_Stream_Query Stream
Mode = #Outcome
if Mode == 'B' then do
call Config_C2B String
String = #Outcome
Stride = 8
Residue = length(String)/8
end
else do
Stride = 1
Residue = length(String)
end

Cursor = 1
do while Residue>0
Piece = substr(String,Cursor,Stride)
Cursor = Cursor+Stride
call Config_Stream_Charout Stream,Piece
if left(#Response, 1) \== 'N' then do
if left(#Response, 1) == 'E' then #StreamState.Stream = 'ERROR'
call #Raise 'NOTREADY', Stream, substr(#Response, 2)

119
 return Residue
end
Residue = Residue - 1
end
return 0

10.5.3 CHARS
CHARS indicates whether there are characters remaining in the named stream. Optionally, it returns a
count of the characters remaining and immediately available.
call CheckArgs 'oSTREAM oCN'

if #Bif_ArgExists.1 then Stream = #Bif_Arg.1

else Stream = ''
if #Bif_ArgExists.2 then Option = #Bif_Arg.2
else Option = 'N'

call Config_Stream_Count Stream, 'CHARS', Option

return #Outcome

10.5.4 LINEIN
LINEIN reads a line from the stream named by the first argument, unless the third argument is zero.
call CheckArgs 'oSTREAM oWHOLE>0 oWHOLE>=0'

if #Bif_ArgExists.1 then Stream = #Bif_Arg.1

else Stream = ''
#StreamState.Stream = ''
if #Bif_ArgExists.2 then do
#Response = Config_Stream_Position(Stream, 'LINEIN', #Bif_Arg2)
if left(#Response, 1) == 'T' then call Raise 40.42,Stream
if left(#Response, 1) == 'R' then call Raise 40.41, 2, #Bif_Arg.2
end
if #Bif_ArgExists.3 then Count = #Bif_Arg.3
else Count = 1
if Count>1 then call Raise 40.39, Count
if Count = 0 then do
call Config_Stream_Charin Stream, 'NULL' /* "Touch" the stream */
return ''
end
/* A configuration may recognise lines even in 'binary' mode. */
call Config_Stream_Query Stream
Mode = #Outcome
r = ''
t = #Linein_Position.Stream
/* Config_Stream_Charin will alter #Linein_Position. */
do until t \= #Linein_Position.Stream
#Response = Config_Stream_Charin(Stream,'LINEIN')
if left(#Response, 1) \== 'N' then do
if left(#Response, 1) == 'E' then #StreamState.Stream = 'ERROR'
call #Raise 'NOTREADY', Stream, substr(#Response, 2)
leave
end
r = r||#Outcome
end
if Mode == 'B' then do
call Config_B2C r
r = #Outcome
end
return r

10.5.5 LINEOUT
LINEOUT returns '1' or '0', indicating whether the second argument has been successfully written to the
stream named by the first argument. A result of '1' means an unsuccessful write.
call CheckArgs 'oSTREAM oANY oWHOLE>0'

if #Bif_ArgExists.1 then Stream = #Bif_Arg.1

else Stream = ''

120
 #StreamState.Stream = ''
if \#Bif_ArgExists.2 & \#Bif_ArgExists.3 then do
/* Position to end of stream. */
#Response = Config_Stream_Close(Stream)
if left(#Response,1) == 'T' then call Raise 40.42,Stream
return 0
end

if #Bif_ArgExists.3 then do
#Response = Config_Stream_Position(Stream,'LINEOUT', #Bif_Arg.3)
if left(#Response, 1) == 'T' then call Raise 40.42,Stream
if left(#Response, 1) == 'R' then call Raise 40.41, 3, #Bif_Arg.3
end

if \#Bif_ArgExists.2 then do
call Config_Stream_Charout Stream, '' /* "Touch" the stream */
return 0
end

String = #Bif_Arg.2
Stride = 1
call Config_Stream_Query Stream
Mode = #Outcome
if Mode == 'B' then do
call Config_C2B String
String = #Outcome
Stride = 8
Residue = length(String)/8
end
else do
Stride = 1
Residue = length(String)
end
Cursor = 1
do while Residue > 0
Piece = substr(String,Cursor,Stride)
Cursor = Cursor+Stride
call Config_Stream_Charout Stream, Piece
if left(#Response, 1) \== 'N' then do
if left(#Response, 1) == 'E' then #StreamState.Stream = 'ERROR'
call #Raise 'NOTREADY', Stream, substr(#Response, 2)
return 1
end
Residue = Residue-1
end
call Config_Stream_Charout Stream, 'EOL'
return 0

10.5.6 LINES
LINES returns the number of lines remaining in the named stream.
call CheckArgs 'oSTREAM oCN'

if #Bif_ArgExists.1 then Stream = #Bif_Arg.1

else Stream = ''
if #Bif_ArgExists.2 then Option = #Bif_Arg.2
else Option = 'N'

Call Config_Stream_Count Stream, 'LINES', Option

return #Outcome

10.5.7 QUALIFY
QUALIFY returns a name for the stream named by the argument. The two names are currently
associated with the same resource and the result of QUALIFY may be more persistently associated with
that resource.
call CheckArgs 'oSTREAM'

if #Bif_ArgExists.1 then Stream = #Bif_Arg.1

else Stream = ''

121
 #Response = Config_Stream_Qualified(Stream)
return #Outcome

10.5.8 STREAM
STREAM returns a description of the state of, or the result of an operation upon, the stream named by
the first argument.
/* Third argument is only correct with 'C' */
if #Bif_ArgExists.2 & translate(left(#Bif_Arg.2, 1)) == 'C' then
ArgData = 'rSTREAM rCDS rANY'
else
ArgData = 'rSTREAM oCDS'
call CheckArgs ArgData

Stream = #Bif_Arg.1

if #Bif_ArgExists.2 then Operation = #Bif_Arg.2

else Operation = 'S'

Select
when Operation == 'C' then do
call Config_Stream_Command Stream,#Bif_Arg.3
return #Outcome
end
when Operation == 'D' then do
#Response = Config_Stream_State(Stream)
return substr(#Response, 2)
end
when Operation == 'S' then do
if StreamState.Stream == 'ERROR' then return 'ERROR'
#Response = Config_Stream_State(Stream)
if left(#Response, 1) == 'N' then return 'READY'
if left(#Response, 1) == 'U' then return 'UNKNOWN'
return 'NOTREADY'
end
end

10.6 Other built-in functions

10.6.1 DATE
DATE with fewer than two arguments returns the local date. Otherwise it converts the second argument
(which has a format given by the third argument) to the format specified by the first argument. If there are
fourth or fifth arguments, they describe the treatment of separators between fields of the date.
call CheckArgs 'oBDEMNOSUW oANY oBDENOSU oSEP oSEP'
/* If the third argument is given then the second is mandatory. */
if #Bif_ArgExists.3 & \#Bif_ArgExists.2 then
call Raise 40.19, '', #Bif_Arg.3

if #Bif_ArgExists.1 then Option = #Bif_Arg.1

else Option = 'N'

/* The date/time is 'frozen' throughout a clause. */

if #ClauseTime.#Level == '' then do
#Response = Config_Time()
#ClauseTime.#Level = #Time
#ClauseLocal.#Level = #Time + #Adjust<Index "#Adjust" # "" >
end
/* English spellings are used, even if messages not in English are used. */
Months = 'January February March April May June July',
'August September October November December'
WeekDays = 'Monday Tuesday Wednesday Thursday Friday Saturday Sunday'

/* If there is no second argument, the current date is returned. */

if \#Bif_ArgExists.2 then
return DateFormat(#ClauseLocal.#Level, Option)

/* If there is a second argument it provides the date to be

converted. */

122
 Value = #Bif_Arg.2
if #Bif_ArgExists.3 then InOption = #Bif_Arg.3
else InOption = 'N'
if Option == 'S' then OutSeparator = ''
else OutSeparator = translate(Option,"xx/x //x","BDEMNOUW")
if #Bif_ArgExists.4 then do
if OutSeparator == 'x' then call Raise 40.46, Option, 4
OutSeparator = #Bif.Arg.4
end
if InOption == 'S' then InSeparator = ''
else InSeparator = translate(InOption,"xx/ //","BDENOU")
if #Bif_ArgExists.5 then do
if InSeparator == 'x' then call Raise 40.46, InOption, 5
InSeparator = #Bif.Arg.5
end
/* First try for Year Month Day */
Logic = 'NS'
select
when InOption == 'N' then do
if InSeparator == '' then do
if length(Value)<9 then return
Year = right(Value,4)
MonthIs = substr(right(Value,7),1,3)
Day = left(Value,length(Value)-7)
end
else
parse var Value Day (InSeparator) MonthIs (InSeparator) Year
do Month = 1 to 12
if left(word(Months, Month), 3) == MonthIs then leave
end Month
end
when InOption == 'S' then
if InSeparator == '' then
parse var Value Year +4 Month +2 Day
else
parse var Value Year (InSeparator) Month (InSeparator) Day
otherwise
Logic = 'EOU' /* or BD */
end

/* Next try for year without century */

if logic = 'EOU' then
Select
when InOption == 'E' then
if InSeparator == '' then
parse var Value Day +2 Month +2 YY
else
parse var Value Day (InSeparator) Month (InSeparator) YY
when InOption == 'O' then
if InSeparator == '' then
parse var Value YY +2 Month +2 Day
else
parse var Value YY (InSeparator) Month (InSeparator) Day
when InOption == 'U' then
if InSeparator == '' then
parse var Value Month +2 Day +2 YY
else
parse var Value Month (InSeparator) Day (InSeparator) YY
otherwise
Logic = 'BD'
end

if Logic = 'EOU' then do

/* The century is assumed, on the basis of the current year. */
if datatype(YY,'W')=0 then
return
YearNow = left('DATE'('S'),4)
Year = YY
do while Year < YearNow-50
Year = Year + 100
end
end /* Century assumption */

123
 if Logic <> 'BD' then do
/* Convert Month & Day to Days of year. */
if datatype(Month,'W')=0 | datatype(Day,'W')=0 | datatype(Year,'W')=0 then
return
Days = word('0 31 59 90 120 151 181 212 243 273 304 334',Month),
+ (Month>2)*Leap(Year) + Day-1
end
else
if datatype(Value,'W')=0 then
return
if InOption == 'D' then do
Year = left('DATE'('S'),4)
Days = Value - 1 /* 'D' includes current day */
end

/* Convert to BaseDays */
if InOption <> 'B' then
BaseDays = (Year-1)*365 + (Year-1)%4 - (Year-1)%100 + (Year-1)%400 + Days
else
Basedays = Value

/* Convert to microseconds from 0001 */

Micro = BaseDays * 86400 * 1000000

/* Reconvert to check the original. (eg for Month = 99) */

if DateFormat(Micro,InOption,InSeparator) \== Value then
call Raise 40.19, Value, InOption
return DateFormat(Micro,Option,OutSeparator)

DateFormat:
/* Convert from microseconds to given format. */
parse value Time2Date(arg(1)) with,
Year Month Day Hour Minute Second Microsecond Base Days
select
when arg(2) == 'B' then
return Base
when arg(2) == 'D' then
return Days
when arg(2) == 'E' then
return right(Day,2,'0')(arg(3))right(Month,2,'0')(arg(3))right(Year,2,'0')
when arg(2) == 'M' then
return word(Months,Month)
when arg(2) == 'N' then
return (Day)(arg(3))left(word(Months,Month),3)(arg(3))right(Year,4,'0')
when arg(2) == 'O' then
return right(Year,2,'0')(arg(3))right(Month,2,'0')(arg(3))right(Day,2,'0')
when arg(2) == 'S' then
return right(Year,4,'0')(arg(3))right(Month,2,'0')(arg(3))right(Day,2,'0')
when arg(2) == 'U' then
return right(Month,2,'0')(arg(3))right(Day,2,'0')(arg(3))right(Year,2,'0')
otherwise /* arg(2) == 'W' */
return word(Weekdays,1+Base//7)
end

10.6.1 QUEUED
QUEUED returns the number of lines remaining in the external data queue.
call CheckArgs ''
#Response = Config_Queued()
return #Outcome

10.6.2 RANDOM
RANDOM returns a quasi-random number.
call CheckArgs 'oWHOLE>=0 oWHOLE>=0 oWHOLE>=0'

if #Bif_Arg.0 = 1 then do
Minimum = 0

124
 Maximum = #Bif_Arg.1
if Maximum>100000 then
call Raise 40.31, Maximum
end
else do
if #Bif_ArgExists.1 then Minimum = #Bif_Arg.1
else Minimum = 0
if #Bif_ArgExists.2 then Maximum = #Bif_Arg.2
else Maximum = 999
end

if Maximum-Minimum>100000 then
call Raise 40.32, Minimum, Maximum

if Maximum-Minimum<0 then
call Raise 40.33, Minimum, Maximum

if #Bif_ArgExists.3 then call Config_Random_Seed #Bif_Arg.3

call Config_Random_Next Minimum, Maximum
return #Outcome

10.6.3 SYMBOL
The function SYMBOL takes one argument, which is evaluated. Let String be the value of that argument.
If Config_Length(String) returns an indicator 'E' then the SYNTAX condition 23.1 shall be raised.
Otherwise, if the syntactic recognition described in section nnn would not recognize String as a symbol
then the result of the function SYMBOL is 'BAD'.
If String would be recognized as a symbol the result of the function SYMBOL depends on the outcome of
accessing the value of that symbol, see nnn. If the final use of Var_Value leaves the indicator with value
'D' then the result of the function SYMBOL is 'LIT', otherwise 'VAR'.
10.6.4 TIME
TIME with less than two arguments returns the local time within the day, or an elapsed time. Otherwise it
converts the second argument (which has a format given by the third argument) to the format specified by
the first argument.
call CheckArgs 'oCEHLMNORS oANY oCHLMNS'
/* If the third argument is given then the second is mandatory. */
if #Bif_ArgExists.3 & \#Bif_ArgExists.2 then
call Raise 40.19, '', #Bif_Arg.3

if #Bif_ArgExists.1 then Option = #Bif_Arg.1

else Option = 'N'

/* The date/time is 'frozen' throughout a clause. */

if #ClauseTime.#Level == '' then do
#Response = Config_Time()
#ClauseTime.#Level = #Time
#ClauseLocal.#Level = #Time + #Adjust<Index "#Adjust" # "" >
end

/* If there is no second argument, the current time is returned. */

if \#Bif_ArgExists.2 then
return TimeFormat(#ClauseLocal.#Level, Option)

/* If there is a second argument it provides the time to be

converted. */
if pos(Option, 'ERO') > 0 then
call Raise 40.29, Option
InValue = #Bif_Arg.2
if #Bif_ArgExists.3 then InOption = #Bif_Arg.3
 HH = 0
if XX == 'pm' then
HourAdjust = 12
end
when InOption == 'H' then
HH = InValue
when InOption == 'L' | InOption == 'N' then
parse var InValue HH ':' MM ':' SS
when InOption == 'M' then
MM = InValue
otherwise
SS = InValue
end
if datatype(HH,'W')=0 | datatype(MM,'W')=0 | datatype(SS,'N')=0 then
call Raise 40.19, InValue, InOption
HH = HH + HourAdjust
/* Convert to microseconds */
Micro = trunc((((HH * 60) + MM) * 60 + SS) * 1000000)
/* There is no special message for time-out-of-range; the bad-format
message is used. */
if Micro<0 | Micro > 24*3600*1000000 then call Raise 40.19, InValue, InOption
/* Reconvert to check the original. (eg for hour = 99) */
if TimeFormat(Micro,InOption) \== InValue then
call Raise 40.19, InValue, InOption
return TimeFormat(Micro, Option)
end /* Conversion */

TimeFormat: procedure
/* Convert from microseconds to given format. */
/* The day will be irrelevant; actually it will be the first day possible. */
x = Time2Date2(arg(1))
parse value x with Year Month Day Hour Minute Second Microsecond Base Days
select
when arg(2) == 'C' then
select
when Hour>12 then
return Hour-12':'right(Minute,2,'0')'pm'
when Hour=12 then
return '12:'right(Minute,2,'0')'pm'
when Hour>0 then
return Hour':'right(Minute,2,'0')'am'
when Hour=0 then
return '12:'right(Minute,2,'0')'am'
end
when arg(2) == 'H' then return Hour
when arg(2) == 'L' then
return right(Hour,2,'0')':'right(Minute,2,'0')':'right(Second,2,'0'),
|| '.'right(Microsecond,6,'0')
when arg(2) == 'M' then
return 60*Hour+Minute
when arg(2) == 'N' then
return right(Hour,2,'0')':'right(Minute,2,'0')':'right(Second,2,'0')
otherwise /* arg(2) == 'S' */
return 3600*Hour+60*Minute+Second
end

Time2Date:
/* These are checks on the range of the date. */
if arg(1) < 0 then
call Raise 40.19, InValue, InOption
if arg(1) >= 315537897600000000 then
call Raise 40.19, InValue, InOption
return Time2Date2(arg(1))
10.6.1 VALUE
VALUE returns the value of the symbol named by the first argument, and optionally assigns it a new
value.
if #Bif_ArgExists.3 then ArgData = 'rANY oANY oANY'
else ArgData = 'rSYM oANY oANY'
call CheckArgs ArgData

126
 Subject = #Bif_Arg.1
if #Bif_ArgExists.3 then do /* An external pool, or the reserved pool. */
/* The reserved pool uses a null string as its pool identifier. */
Pool = #Bif_Arg.3
if Pool == '' then do
Subject = '.' || translate(Subject) /* The dot on the name is implied. */

Value = .environment[Subject] /* Was the translate redundant? */

if #Bif_ArgExists.2 then .environment[Subject] = #Bif_Arg.2
return Value
end
/* Fetch the original value */
#Response = Config_Get(Pool,Subject)
#Indicator = left(#Response,1)
if #Indicator == 'F' then
call Raise 40.36, Subject
if #Indicator == 'P' then
call Raise 40.37, Pool
Value = #Outcome
if #Bif_ArgExists.2 then do
/* Set the new value. */
#Response = Config_Set(Pool,Subject,#Bif_Arg.2)
if #Indicator == 'P' then
call Raise 40.37, Pool
if #Indicator == 'F' then
call Raise 40.36, Subject
end
/* Return the original value. */
return Value
end
/* Not external */
Subject = translate(Subject)
/* See nnn */
p = pos(Subject, '.')
if p = 0 | p = length(Subject) then do
/* Not compound */
#Response = Var_Value(#Pool, Subject, '0')
/* The caller, in the code of the standard, may need
to test whether the Subject was dropped. */
#Indicator = left(#Response, 1)
Value = #Outcome
if #Bif_ArgExists.2 then
#Response = Var_Set(#Pool, Subject, '0', #Bif_Arg.2)
return Value
end
/* Compound */
Expanded = left(Subject,p-1) /* The stem */
do forever
Start = p+1
p = pos(Subject,'.',Start)
if p = 0 then p = length(Subject)
Item = substr(Subject,Start,p-Start) /* Tail component symbol */
if Item\=='' then if pos(left(Item,1),'0123456789') = 0 then do
#Response = Var_Value(#Pool, Item, '0')
Item = #Outcome
end
/* Add tail component. */
Expanded = Expanded'.'Item
end
#Response = Var_Value(#Pool, Expanded, '1')
#Indicator = left(#Response, 1)
Value = #Outcome
if #Bif_ArgExists.2 then
#Response = Var_Set(#Pool, Expanded, '1', #Bif_Arg.2)
return Value

10.6.1 QUEUED
QUEUED returns the number of lines remaining in the external data queue.

127
 call CheckArgs ''
#Response = Config_Queued()
return #Outcome

10.6.2 RANDOM
RANDOM returns a quasi-random number.
call CheckArgs 'oWHOLE>=0 oWHOLE>=0 oWHOLE>=0'

if #Bif_Arg.0 = 1 then do
Minimum = 0
Maximum = #Bif_Arg.1
if Maximum>100000 then
call Raise 40.31, Maximum
end
else do
if #Bif_ArgExists.1 then Minimum = #Bif_Arg.1
else Minimum = 0
if #Bif_ArgExists.2 then Maximum = #Bif_Arg.2
else Maximum = 999
end

if Maximum-Minimum>100000 then
call Raise 40.32, Minimum, Maximum

if Maximum-Minimum<0 then
call Raise 40.33, Minimum, Maximum

if #Bif_ArgExists.3 then call Config_Random_Seed #Bif_Arg.3

call Config_Random_Next Minimum, Maximum
return #Outcome

10.6.3 SYMBOL
The function SYMBOL takes one argument, which is evaluated. Let String be the value of that argument.
If Config_Length(String) returns an indicator 'E' then the SYNTAX condition 23.1 shall be raised.
Otherwise, if the syntactic recognition described in section nnn would not recognize String as a symbol
then the result of the function SYMBOL is 'BAD'.
If String would be recognized as a symbol the result of the function SYMBOL depends on the outcome of
accessing the value of that symbol, see nnn. If the final use of Var_Value leaves the indicator with value
'D' then the result of the function SYMBOL is 'LIT', otherwise 'VAR'.
10.6.4 TIME
TIME with less than two arguments returns the local time within the day, or an elapsed time. Otherwise it
converts the second argument (which has a format given by the third argument) to the format specified by
the first argument.
call CheckArgs 'oCEHLMNORS oANY oCHLMNS'
/* If the third argument is given then the second is mandatory. */
if #Bif_ArgExists.3 & \#Bif_ArgExists.2 then
call Raise 40.19, '', #Bif_Arg.3

if #Bif_ArgExists.1 then Option = #Bif_Arg.1

else Option = 'N'

/* The date/time is 'frozen' throughout a clause. */

if #ClauseTime.#Level == '' then do
#Response = Config_Time()
#ClauseTime.#Level = #Time
#ClauseLocal.#Level = #Time + #Adjust<Index "#Adjust" # "" >
end

/* If there is no second argument, the current time is returned. */

if \#Bif_ArgExists.2 then
return TimeFormat(#ClauseLocal.#Level, Option)

/* If there is a second argument it provides the time to be

converted. */
if pos(Option, 'ERO') > 0 then

128
 call Raise 40.29, Option
InValue = #Bif_Arg.2
if #Bif_ArgExists.3 then InOption = #Bif_Arg.3
else InOption = 'N'
HH = 0
MM = 0
SS = 0
HourAdjust = 0
select
when InOption == 'C' then do
parse var InValue HH ':' . +1 MM +2 XX
if XX == 'pm' then HourAdjust = 12
end
when InOption == 'H' then HH = InValue
when InOption == 'L' | InOption == 'N' then
parse var InValue HH ':' MM ':' SS
when InOption == 'M' then MM = InValue
otherwise SS = InValue
end
if \datatype(HH,'W') | \datatype(MM,'W') | \datatype(SS,'N') then
call Raise 40.19, InValue, InOption
HH = HH + HourAdjust
/* Convert to microseconds */
Micro = trunc((((HH * 60) + MM) * 60 + SS) * 1000000)
/* Reconvert to check the original. (eg for hour = 99) */
if TimeFormat(Micro,InOption) \== InValue then
call Raise 40.19, InValue, InOption
return TimeFormat(Micro, Option)
end /* Conversion */

TimeFormat:
/* Convert from microseconds to given format. */
parse value Time2Date(arg(1)) with,
Year Month Day Hour Minute Second Microsecond Base Days
select
when arg(2) == 'C' then
if Hour>12 then
return Hour-12':'right(Minute,2,'0')'pm'
else
return Hour':'right(Minute,2,'0')'am'
when arg(2) == 'E' | arg(2) == 'R' then do
/* Special case first time */
if #StartTime.#Level == '' then do
#StartTime.#Level = #ClauseTime.#Level
return '0'
end
Output = #ClauseTime.#Level-#StartTime.#Level
if arg(2) == 'R' then
#StartTime.#Level = #ClauseTime.#Level
return Output * 1E-6
end /* E or R */
when arg(2) == 'H' then return Hour
when arg(2) == 'L' then
return right(Hour,2,'0')':'right(Minute,2,'0')':'right(Second,2,'0'),
|| '.'right(Microsecond,6,'0')
when arg(2) == 'M' then return 60*Hour+Minute
when arg(2) == 'N' then
return right(Hour,2,'0')':'right(Minute,2,'0')':'right(Second,2,'0')
when arg(2) == 'O' then
return trunc(#ClauseLocal.#Level - #ClauseTime.#Level)
otherwise /* arg(2) == 'S' */
return 3600*Hour+60*Minute+Second
end

10.6.5 VALUE
VALUE returns the value of the symbol named by the first argument, and optionally assigns it a new
value.
if #Bif_ArgExists.3 then ArgData = 'rANY oANY oANY'
else ArgData = 'rSYM oANY oANY'
call CheckArgs ArgData

129
 Subject = #Bif_Arg.1
if #Bif_ArgExists.3 then do /* An external pool, or the reserved pool. */
/* The reserved pool uses a null string as its pool identifier. */
Pool = #Bif_Arg.3
if Pool == '' then do
Subject = '.' || translate(Subject) /* The dot on the name is implied. */

Value = .environment[Subject] /* Was the translate redundant? */

if #Bif_ArgExists.2 then .environment[Subject] = #Bif_Arg.2
return Value
end
/* Fetch the original value */
#Response = Config_Get(Pool,Subject)
#Indicator = left(#Response,1)
if #Indicator == 'F' then
call Raise 40.36, Subject
if #Indicator == 'P' then
call Raise 40.37, Pool
Value = #Outcome
if #Bif_ArgExists.2 then do
/* Set the new value. */
#Response = Config_Set(Pool,Subject,#Bif_Arg.2)
if #Indicator == 'P' then
call Raise 40.37, Pool
if #Indicator == 'F' then
call Raise 40.36, Subject
end
/* Return the original value. */
return Value
end
/* Not external */
Subject = translate(Subject)
/* See nnn */
p = pos(Subject, '.')
if p = 0 | p = length(Subject) then do
/* Not compound */
#Response = Var_Value(#Pool, Subject, '0')
/* The caller, in the code of the standard, may need
to test whether the Subject was dropped. */
#Indicator = left(#Response, 1)
Value = #Outcome
if #Bif_ArgExists.2 then
#Response = Var_Set(#Pool, Subject, '0', #Bif_Arg.2)
return Value
end
/* Compound */
Expanded = left(Subject,p-1) /* The stem */
do forever
Start = p+1
p = pos(Subject,'.',Start)
if p = 0 then p = length(Subject)
Item = substr(Subject,Start,p-Start) /* Tail component symbol */
if Item\=='' then if pos(left(Item,1),'0123456789') = 0 then do
#Response = Var_Value(#Pool, Item, '0')
Item = #Outcome
end
/* Add tail component. */
Expanded = Expanded'.'Item
end
#Response = Var_Value(#Pool, Expanded, '1')
#Indicator = left(#Response, 1)
Value = #Outcome
if #Bif_ArgExists.2 then
#Response = Var_Set(#Pool, Expanded, '1', #Bif_Arg.2)
return Value

130
11 Built-in classes
11.1 Notation
The built-in classes are defined mainly through code. The code refers to state variables. This is solely a
notation used in this standard.
11.2 Object, class and method
These objects provide the basis for class structure.
11.2.1 The object class
::class object

::method new class

Returns a new instance of the receiver class.
call Config_ObjectNew
return #Outcome

::method '=='
'==' with no argument gives a hash value in OOI.
call Config_ObjectCompare #Receiver, #Arg.1
if #Outcome == 'equal' then return '1'
else return '0'
::method '<>'
use arg a
return \self==a

::method '><'
forward message '<>'

::method '='
forward message '=='

::method '\='
forward message '<>'

::method '\=='
forward message '<>'

::method copy
Returns a copy of the receiver object. The copied object has the same methods as the receiver object
and an equivalent set of object variables, with the same values.
call Config_ObjectCopy #Receiver
return #Outcome
Since we have var_empty we could save a primitive by rendering 'new' as 'copy' plus 'empty'.
::method defaultname
Returns a short human-readable string representation for the object.
call var_value #Receiver, '#Human', '0'
return #Outcome
This field would have been filled in at 'NEW' time.

::method 'OBJECTNAME=' /* rSTRING */

Sets the receiver object's name to the specified string.
call var_set #Receiver, #ObjectName, '0', #Arg.1
return
Initialized to #Human? Or ObjectName does forwarding until assigned to?

::method objectname
Returns the receiver object's name (which the OBJECTNAME= method sets).
call var_value #Receiver, #ObjectName, '0'
return #Outcome

::method string
Returns a human-readable string representation for the object.
return #Receiver~ObjectName

::method class
Returns the class object that received the message that created the object.

131
call var_value #Receiver, #IsA, '0'
return #Outcome

::method setmethod /* rSTRING oSTRING/METHOD/ARRAY */

Adds a method to the receiver object's collection of object methods.
Is 'object methods' what is intended; you add to a class without adding to its instance methods? Yes.
if #ArgExists.2 then m = Arg.2
else m = .NIL
call set_var #Receiver, 'METHODS.'#Arg.1, '1', m
return

::method hasmethod /* rSTRING */

Returns 1 (true) if the receiver object has a method with the specified name (translated to uppercase);
otherwise, returns 0 (false).
This presumably means inherited as well as SETMETHOD ones. What about ones set to .NIL?
Need to use the same search as for sending.

::method unsetmethod private

Removes a method from the receiver object's collection of object methods.
Use var_drop
Private means Receiver = Self check.

::method request /* rSTRING */

Returns an object of the specified class, or the NIL object if the request cannot be satisfied.
t = 'MAKE'#Arg.1
if \#Receiver~hasmethod(t) then return .NIL
forward message(t) array()

::method run private /* rMETHOD Ugh keyoptions */

Runs the specified method. The method has access to the object variables of the receiver object, just as
if the receiver object had defined the method by using SETMETHOD.
::method startat Undocumented?

::method start /* rMESSAGE oArglist */

Returns a message object and sends it a START message to start concurrent processing.
::method init
Performs any required object initialization.
11.2.2 The class class
::class class

Lots of these methods are both class and instance. I don't know whether to list them twice.
::method new class /* oARGLIST */
Returns a new instance of the receiver class, whose object methods are the instance methods of the
class. This method initializes a new instance by running its INIT methods.
::method subclass class
Returns a new subclass of the receiver class.
::method subclasses class
Returns the immediate subclasses of the receiver class in the form of a single-index array of the required
size.
::method define class /* rSTRING oMETHOD */
Incorporates the method object in the receiver class's collection of instance methods. The method name
is translated to upper case.
::method delete
Removes the receiver class's definition for the method name specified.
Builtin classes cannot be altered.
::method method class /* rSTRING */
Returns the method object for the receiver class's definition for the method name given.
Do we have to keep saying "method object" as opposed to "method" because "method name" exists?
::method querymixinclass
Returns 1 (true) if the class is a mixin class or 0 (false) otherwise.
::method mixinclass class /* 3 of em */
Returns a new mixin subclass of the receiver class.
::method inherit class /* rCLASS oCLASS */

132
Causes the receiver class to inherit the instance and class methods of the class object specified. The
optional class is a class object that specifies the position of the new superclass in the list of superclasses.
::method uninherit class /* rCLASSOBJ */
Nullifies the effect of any previous INHERIT message sent to the receiver for the class specified.
::method enhanced class /* rCOLLECTION oArgs */
Returns an enhanced new instance of the receiver class, with object methods that are the instance
methods of the class enhanced by the methods in the specified collection of methods.
::method baseclass class
Returns the base class associated with the class. If the class is a mixin class, the base class is the first
superclass that is not also a mixin class. If the class is not a mixin class, then the base class is the class
receiving the BASECLASS message.
::method superclasses class
Returns the immediate superclasses of the receiver class in the form of a single-index array of the
required size.
::method id class
Returns a string that is the class identity (instance SUBCLASS and MIXINCLASS methods.)
::method metaclass class
Returns the receiver class's default metaclass.
::method methods class /* oCLASSOBJECT */
Returns a supplier object for all the instance methods of the receiver class and its superclasses, if no
argument is specified.
11.2.3 The method class
::class method

::method new class /* rSTRING rSOURCE */

Returns a new instance of method class, which is an executable representation of the code contained in
the source.
::method setprivate
Specifies that a method is a private method.
::method setprotected

::method setsecuritymanager

::method setguarded
Reverses any previous SETUNGUARDED messages, restoring the receiver to the default guarded
status.
::method setunguarded
Lets an object run a method even when another method is active on the same object. If a method object
does not receive a SETUNGUARDED message, it requires exclusive use of its object variable pool.
::method source
Returns the method source code as a single index array of source lines.
::method interface

::method setinterface

11.3 The string class

The string class provides conventional strings and numbers.

Some differences from REXX class of NetRexx.

::class string

::method new class

::method '\'
We can do all the operators by appeal to classic section 7.
::method '-'
::method '-'
use arg a
return \a
General problem of making the error message come right.

133
::method '+'
::method '**'
::method '*'
::method '%'
::method '/'
::method '//'
::method ' '
::method '||'
::method '<>'
::method '><'
::method '='
::method '\='
::method '\=='
::method '=='
::method '<'
::method '>'
::method '>='
::method '/<'
::method '<='
::method '/>'
::method '<<'
::method '>>'
::method '>>='
::method '/>>'
::method '<<='
::method '\<<'
::method '&&'
::method '&'
::method abbrev

::method centre
::method center
::method changestr
::method compare
::method copies
::method counstr
::method datatype
::method delstr
::method delword
::method insert
::method lastpos
::method left
::method length
::method overlay
::method pos
::method reverse
::method right
::method space
::method strip
::method substr
::method subword
::method translate
::method verify
::method word
::method wordindex
::method wordlength
::method wordpos
::method words
::method abs
::method format
::method max
::method min
::method sign
::method trunc
::method B2X
::method bitand
::method bitor
::method bitxor
::method C2D
::method C2X
::method D2X

134
::method D2C
::method X2B
::method X2C
::method X2D
::method string

::method makestring

11.3.1 The array class

The main features of a single dimension array are provided by the configuration. This section defines
further methods and multi-dimensional arrays.

To be done. Dimensionality set at first use. Count commas, not classic arg().

::class array

::method new class /* 0 or more WHOLE>=0 */

Returns a new empty array.
::method of class /* 0 or more ANY */
Returns a newly created single-index array containing the specified value objects.
::method put /* rANY one or more WHOLE>0 */
Makes the object value a member item of the array and associates it with the specified index or indexes.
::method '[]=' /* 1 or more WHOLE>0 */
This method is the same as the PUT method.
::method at /* 1 or more WHOLE>0 */
Returns the item associated with the specified index or indexes.
::method '[]' /* 1 or more WHOLE>0 */
Returns the same value as the AT method.
::method remove /* 1 or more WHOLE>0 */
Returns and removes the member item with the specified index or indexes from the array.
::method hasindex /* 1 or more WHOLE>0 */
Returns 1 (true) if the array contains an item associated with the specified index or indexes. Returns 0
(false) otherwise.
::method items /* (None) */
Returns the number of items in the collection.
::method dimension /* oWHOLE>0 */
Returns the current size (upper bound) of dimension specified (a positive whole number). If you omit the
argument this method returns the dimensionality (number of dimensions) of the array.
::method size /* (None) */
Returns the number of items that can be placed in the array before it needs to be extended.
::method first /* (None) */
Returns the index of the first item in the array, or the NIL object if the array is empty.
::method last /* (None) */
Returns the index of the last item in the array, or the NIL object if the array is empty.
::method next /* rWHOLE>0 */
Returns the index of the item that follows the array item having the specified index or returns the NIL
object if the item having that index is last in the array.
::method previous /* rWHOLE>0 */
Returns the index of the item that precedes the array item having index index or the NIL object if the item
having that index is first in the array.
::method makearray /* (None) */
Returns a single-index array with the same number of items as the receiver object. Any index with no
associated item is omitted from the new array.
Returns a new array (of the same class as the receiver) containing selected items from the receiver array.
The first item in the new array is the item corresponding to index start (the first argument) in the receiver
array.
::method supplier /* (None) */
Returns a supplier object for the collection.
::method section /* rWHOLE>0 oWHOLE>=0 */

11.4 The supplier class

135
A supplier object enumerates the items a collection contained at the time of the supplier's creation.
::class supplier

::method new class /* rANYARRAY rINDEXARRAY */

Returns a new supplier object.
::method index
Returns the index of the current item in the collection.
::method next
Moves to the next item in the collection.
::method item
Returns the current item in the collection.
::method available
Returns 1 (true) if an item is available from the supplier (that is, if the ITEM method would return a value).
Returns 0 (false) otherwise.
.
11.5 The message class
::class message

::method init class /* Ugh */

Initializes the message object for sending......
::method completed
Returns 1 if the message object has completed its message; returns 0 otherwise.
::method notify /* rMESSAGE */
Requests notification about the completion of processing for the message SEND or START sends.
::method start /* oANY */
Sends the message for processing concurrently with continued processing of the sender.
::method send /* oANY */
Returns the result (if any) of sending the message.
::method result
Returns the result of the message SEND or START sends.

136
12 Provided classes
(Informative)
12.1 Notation
The provided classes are defined mainly through code.
12.2 The Collection Classes
12.2.1 Collection Class Routines
These routines are used in the definition of the collection classes
::routine CommonXor
/* Returns a new collection that contains all items from self and
the argument except that all indexes that appear in both collections
are removed. */
/* When the target is a bag, there may be an index in the bag that is
duplicated and the same value as an index in the argument. Should one
copy of the index survive in the bag? */
v=1
if (arg(1)~class==.Set & arg(2)~class==.Bag) then v=2
if (arg(1)~class==.Table & arg(2)~class==.Bag) then v=2
if (arg(1)~class==.Table & arg(2)~class==.Relation) then v=2
if (arg(1)~class==.Directory & arg(2)~class==.Bag) then v=2
if (arg(1)~class==.Directory & arg(2)~class==.Relation) then v=2
/* This version it does: */
if v=1 then do
This = arg(1) /* self of caller */
r=This~class~new
ab=MayEnBag(arg(2))
ss=This~supplier
do while ss~available
r[ss~index]=ss~item
ss~next
end
cs=ab~supplier
do while cs~available
if r~hasindex(cs~index) then r~remove(cs~index)
else r[cs~index]=cs~item
cs~next
end
return r
end

/* But following matches practice on Set~XOR(bag) etc. */

This = arg(1) /* self of caller */
r=This~class~new
ab=MayEnBag(arg(2))
ss=This~supplier
do while ss~available
if \ab~hasindex(ss~index) then r[ss~index]=ss~item
ss~next
end
cs=ab~supplier
do while cs~available
if \This~hasindex(cs~index) then r[cs~index]=cs~item
cs~next
end
return r

::routine CommonIntersect
/* Returns a new collection of the same class as SELF that
contains the items from SELF that have indexes also in the
argument. */
/* Actually an index in SELF can only be 'matched' with one in the
argument once. Hence copy and removal. */
This = arg(1) /* self of caller */
w=.Bag~new
sc=This~supplier
do while sc~available
w[sc~index]=sc~index
sc~next
end
r=This~class~new

137
 cs=MayEnBag(arg(2))~supplier
do while cs~available
i=cs~index
if w~hasindex(i) then do
r[i]=This[i]
w~remove(i)
end
cs~next
end
return r

::routine CommonUnion
/* Returns a new collection of the same class as SELF that
contains all the items from SELF and items from the
argument that have an index not in the first. */
/* Best to add them all. By adding non-receiver first we ensure that
receiver takes priority when same indexes. */
This = arg(1) /* self of caller */
r=This~class~new
cs=MayEnBag(arg(2))~supplier
do while cs~available
r[cs~index]=cs~item
cs~next
end
cs=This~supplier
do while cs~available
r[cs~index]=cs~item
cs~next
end
return r

::routine CommonDifference
/* Returns a new collection containing only those index-item pairs from the
SELF whose indexes the other collection does not contain. */
This = arg(1) /* self of caller */
r=This~class~new
cs=This~supplier
do while cs~available
r[cs~index]=cs~item
cs~next
end
cs=MayEnBag(arg(2))~supplier
do while cs~available
r~remove(cs~index)
cs~next
end
return r

::routine MayEnBag
/* For List and Queue the indexes are dropped. */
r=arg(1)
if r~class == .List | r~class == .Queue then r=EnBag(r)
return r

::routine EnBag
r=.Bag~new
s=arg(1)~supplier
do while s~available
if arg(1)~class == .List | arg(1)~class == .Queue then
r[s~item]=s~item
else
/* This case is when the receiver is a Bag. */
r[s~index]=s~index
s~next
end
return r

12.2.2 The collection class

::class 'Collection'

138
12.2.2.1 INIT
::method init
expose a
/* A collection is modelled as using 3 slots in an array for each element.
The first slot holds the item, the second the index, and the third is
used by particular types of collection. This order of slots is arbitary,
chosen to match order of arguments for PUT and SUPPLIER~NEW. */
/* The first set of 3 slots is reserved for other purposes, to avoid
having separate variables which the subclassing would need to access. */
a=.array~new
a[1]/*ItemsCount*/=0
a[2]/*Unique*/=0
return self

12.2.2.2 EXPOSED
::method exposed private
expose a
/* This method allows subclasses to get at the implementation of Collection. */
return a

12.2.2.3 FINDINDEX
::method findindex private
expose a
/* Returns array index if the collection contains any item associated with the
index specified or returns 0 otherwise. */
do j=4 by 3 to 1+3*a[1]/*ItemsCount*/
if a[j+1]==arg(1) then return j
end j
return 0

12.2.2.4 AT
::method at /* rANY */
expose a
/* Returns the item associated with the specified index. */
j=self~findindex(arg(1))
if j=0 then return .nil
return a[j]

12.2.2.5 []
::method '[]'
/* Synonym for the AT method. */
forward message 'AT'

12.2.2.6 PUT
::method put /* rANY rANY */
expose a
use arg item, index
/* Replaces any existing item associated with the specified index with the new
item. Otherwise adds the item-index pair. */
j=self~findindex(index)
if j>0 then do
a[j]=item
return
end
a[1]/*ItemsCount*/=a[1]/*ItemsCount*/+1
j=1+3*a[1]/*ItemsCount*/
a[j]=item
a[j+1]=index
a[j+2]=0
return /* Error 91 in OOI if context requiring result. */

12.2.2.7 []=

139
::method '[]='
/* Synonym for the PUT method. */
forward message 'PUT'

12.2.2.8 HASINDEX
::method hasindex /* rANY */
/* Returns 1 (true) if the collection contains any item associated with the
index specified or returns 0 (false) otherwise. */
return self~findindex(arg(1))>0

12.2.2.9 ITEMS
::method items
expose a
/* Returns the number of items in the collection. */
return a[1]/*ItemsCount*/

12.2.2.10 REMOVE
::method remove /* rANY */
expose a
/* Returns and removes from a collection the member item with the specified
index. */
j=self~findindex(arg(1))
if j=0 then return .nil
r=a[j]
self~removeit(j)
return r

12.2.2.11 REMOVEIT
::method removeit private
expose a
use arg j
/* Remove relevant slots from the array, with compaction. */
do j=j+3 by 3 to 1+3*a[1]/*ItemsCount*/
a[j-3]=a[j];a[j-2]=a[j+1];a[j-1]=a[j+2]
end j
a[1]/*ItemsCount*/=a[1]/*ItemsCount*/-1
return

12.2.2.12 MAKEARRYA
::method makearray
expose a
/* Returns a single-index array containing the receiver list items. */
r = .array~new /* To build result in. */
do j=4 by 3 to 1+3*a[1]/*ItemsCount*/
r[r~dimension(1)+1]=a[j]
end j
return r

12.2.2.13 MAKEARRAYX
::method makearrayx private
expose a
/* Returns a single-index array containing the receiver index items. */
r = .array~new /* To build result in. */
do j=4 by 3 to 1+3*a[1]/*ItemsCount*/
r[r~dimension(1)+1]=a[j+1]
end j
return r

12.2.2.14 SUPPLIER
::method supplier
expose a

140
/* Returns a supplier object for the list. */
return .supplier~new(self~makearray:.collection,self~makearrayx)

12.2.3 Class list

::class 'List' subclass Collection

12.2.3.1 PUT
::method put /* rANY rANY */
use arg item, index
a=self~exposed
/* PUT for a List must not be an insertion. */
j=self~findindex(index)
if j=0 then call Raise 'Syntax',93.918
a[j]=item
return

12.2.3.2 OF
::method of class /* 1 or more oANY Are they omittable? Not in IOO */
/* Returns a newly created list containing the specified value objects in the
order specified. */
r = self ~ new
do j = 1 to arg()
r ~ insert(arg(j))
end j
return r

12.2.3.3 INSERT
::method insert /* rANY oANY */
use arg item, index
a=self~exposed
/* Returns a list-supplied index for a new item, of specified value, which is
added to the list. The new item follows the existing item with the specified
index in the list ordering. */
/* Establish the index of what preceeds the new element. */
/* If there was no index given, the new item becomes the last on list. */
/* .nil argument means first */
if arg(2,'E') then p=arg(2)
else p=self~last
/* Convert from list index to underlying array index. */
if p==.nil then j=1
else j=self~findindex(p)
if j=0 then call Raise 'Syntax',93.918
j=j+3 /* Where new entry will be. */
/* Move space to required place. */
a[1]/*ItemsCount*/=a[1]/*ItemsCount*/+1
do k=1+3*a[1]/*ItemsCount*/ by -3 to j+3
a[k]=a[k-3];a[k+1]=a[k-2];a[k]=a[k-3]
end
/* Insert new element. */
a[j]=item
/* A new, unique, index is needed. */
/* The basic requirement is for something unique, so this would be correct:
i=.object~new /* a unique object, used as a key (the index on the list) */
*/
/* However, a number can be used. (At risk of the user thinking it is
sensible to do arithmetic on it.) */
a[j+1]=a[2]/*Unique*/;a[2]/*Unique*/=a[2]/*Unique*/+1
a[j+2]=0
return a[j+1]

12.2.3.4 FIRST
::method first
a=self~exposed
/* Returns the index of the first item in the list. */
if a[1]/*ItemsCount*/=0 then return .nil

141
 return a[5]

12.2.3.5 LAST
::method last
a=self~exposed
/* Returns the index of the last item in the list. */
if a[1]/*ItemsCount*/=0 then return .nil
return a[3*a[1]/*ItemsCount*/+2]

12.2.3.6 FIRSTITEM
::method firstitem
a=self~exposed
/* Returns the first item in the list. */
if a[1]/*ItemsCount*/=0 then return .nil
return a[4]

12.2.3.7 LASTITEM
::method lastitem
a=self~exposed
/* Returns the last item in the list. */
if a[1]/*ItemsCount*/=0 then return .nil
return a[3*a[1]/*ItemsCount*/+1]

12.2.3.8 NEXT
::method next /* rANY */
a=self~exposed
/* Returns the index of the item that follows the list item having the specified
index. */
j=self~findindex(arg(1))
if j=0 then call Raise 'Syntax',93.918
j=j+3
if j>3*a[1]/*ItemsCount*/ then return .nil /* Next of last was requested. */
return a[j+1]

12.2.3.9 PREVIOUS
::method previous /* rANY */
a=self~exposed
/* Returns the index of the item that precedes the list item having the
specified index. */
j=self~findindex(arg(1))
if j=0 then call Raise 'Syntax',93.918
j=j-3
if j<4 then return .nil /* Previous of first was requested. */
return a[j+1]

12.2.3.10 SECTION
::method section /* rANY oWHOLE>=0 */
a=self~exposed
/* Returns a new list containing selected items from the receiver list. The
first item in the new list is the item corresponding to the index specified,
in the receiver list. */
j=self~findindex(arg(1))
if j=0 then call Raise 'Syntax',93.918
r = .list~new /* To build result in. */
if arg(2,'E') then s = arg(2)
else s = self~items;
do s
r~insert(a[j])
j=j+3
if j>1+3*a[1]/*ItemsCount*/ then leave
end
return r

142
12.2.4 Class queue
::class 'Queue' subclass Collection
/* A queue is a sequenced collection with whole-number indexes. The
indexes specify the position of an item relative to the head (first item) of
the queue. Adding or removing an item changes the association of an index to
its queue item. */

12.2.4.1 PUSH
::method push /* rANY */
/* Adds the object value to the queue at its head. */
a=self~exposed
a[1]/*ItemCount*/=a[1]/*ItemCount*/+1
/* Slide along to make a space. */
do j=1+3*a[1]/*ItemCount*/ by -3 to 7
a[j]=a[j-3]
a[j+1]=a[j-2]+1; /* Index changes */
end j
a[4]=arg(1)
a[5]=1
return

12.2.4.2 PULL
::method pull
/* Returns and removes the item at the head of the queue. */
a=self~exposed
if a[1]/*ItemCount*/=0 then return .nil /* Stays empty */
r=a[4]
a[1]/*ItemCount*/=a[1]/*ItemCount*/-1
do j=4 by 3 to 1+3*a[1]/*ItemCount*/
a[j]=a[j+3]
a[j+1]=a[j+4]-1; /* Index changes */
end j
return r

12.2.4.3 QUEUE
::method queue /* rANY */
/* Adds the object value to the queue at its tail. */
a=self~exposed
a[1]/*ItemCount*/=a[1]/*ItemCount*/+1
a[1+3*a[1]/*ItemCount*/]=arg(1)
a[2+3*a[1]/*ItemCount*/]=a[1]/*ItemCount*/
return

12.2.4.4 PEEK
::method peek
a=self~exposed
/* Returns the item at the head of the queue. The collection remains unchanged.
*/
return a[4]

12.2.4.5 REMOVE
::method remove /* rWHOLE>0 */
/* Returns and removes from a collection the member item with the specified
index. */
a=self~exposed
if a[1]/*ItemCount*/<arg(1) then return .nil
r=self~remove:super(arg(1))
/* Reset the indexes. */
k=0
do j=4 by 3 to 1+3*a[1]/*ItemsCount*/
k=k+1
a[j+1]=k
end j
return r

143
12.2.5 Class table
::class 'Table' subclass Collection

12.2.5.1 MAKEARRAY
::method makearray
/* Returns a single-index array containing the index objects. */
/* This is different from Collection MAKEARRAY where items rather than indexes
are in the returned array. */
forward message 'MAKEARRAYX'

12.2.5.2 UNION
::method union /* rCOLLECTION */
return CommonUnion(self,arg(1))

12.2.5.3 INTERSECTION
::method intersection /* rCOLLECTION */
return CommonIntersect(self,arg(1))

12.2.5.4 XOR
::method xor /* rCOLLECTION */
return CommonXor(self,arg(1))

12.2.5.5 DIFFERENCE
::method difference /* rCOLLECTION */
return CommonDifference(self,arg(1))

12.2.5.6 SUBSET
::method subset /* rCOLLECTION */
return self~difference(arg(1))~items = 0

12.2.6 Class set

::class 'Set' subclass table
/* A set is a collection that restricts the member items to have a value that is
the same as the index. Any object can be placed in a set. There can be only
one occurrence of any object in a set. */

12.2.6.1 PUT
/* Second arg same as first. Committee has dropped second? */
::method put /* rANY oANY */
/* Makes the object value a member item of the collection and associates it with
specified index. */
if arg(2,'E') then
if arg(2)\==arg(1) then signal error /* 949 */
self~put:super(arg(1),arg(1))

12.2.6.2 OF
::method of class /* 1 or more rANY */
/* Returns a newly created set containing the specified value objects. */
r=self~new
do j=1 to arg()
r~put(arg(j))
end j
return r

12.2.6.3 UNION
::method union /* rCOLLECTION */

144
 return CommonUnion(self,EnBag(arg(1)))

12.2.6.4 INTERSECTION
::method intersection /* rCOLLECTION */
return CommonIntersect(self,EnBag(arg(1)))

12.2.6.5 XOR
::method xor /* rCOLLECTION */
return CommonXor(self,EnBag(arg(1)))

12.2.6.6 DIFFERENCE
::method difference /* rCOLLECTION */
return CommonDifference(self,EnBag(arg(1)))

12.2.7 Class relation

::class 'Relation' subclass Collection

12.2.7.1 PUT
::method put /* rANY rANY */
use arg item, index
a=self~exposed
/* Makes the object value a member item of the relation and associates it with
the specified index. If the relation already contains any items with the
specified index, this method adds a new member item value with the same index,
without removing any existing members */
a[1]/*ItemsCount*/=a[1]/*ItemsCount*/+1
j=1+3*a[1]/*ItemsCount*/
a[j]=item
a[j+1]=index
a[j+2]=0
return /* Error 91 in OOI if context requiring result. */

12.2.7.2 ITEMS
::method items /* oANY */
a=self~exposed
/* Returns the number of relation items with the specified index. If you specify
no index, this method returns the total number of items associated with all
indexes in the relation. */
if \arg(1,'E') then return a[1]/*ItemsCount*/
n=0
do j=4 by 3 to 1+3*a[1]/*ItemsCount*/
if arg(1)==a[j+1] then n=n+1
end j
return n

12.2.7.3 MAKEARRAY
::method makearray
forward message 'MAKEARRAYX'

12.2.7.4 SUPPLIER
::method supplier /* oANY */
a=self~exposed
/* Returns a supplier object for the collection. If an index is specified, the
supplier enumerates all of the items in the relation with the specified
index. */
m=.array~new /* For the items */
r=.array~new /* For the indexes */
do j=4 by 3 to 1+3*a[1]/*ItemsCount*/
if arg(1,'E') then
if arg(1)\==a[j+1] then iterate
n=r~dimension(1)+1

145
 m[n]=a[j]
r[n]=a[j+1]
end j
return .supplier~new(m,r)

12.2.7.5 UNION
::method union /* rCOLLECTION */
/* Union for a relation is just all of both. */
r=self~class~new
cs=self~supplier
do while cs~available
r[cs~index]=cs~item
cs~next
end
cs=MayEnBag(arg(1))~supplier
do while cs~available
r[cs~index]=cs~item
cs~next
end
return r

12.2.7.6 INTERSECTION
::method intersection /* rCOLLECTION */
/* Intersection for a relation requires the items as well as the keys to
match. */
r=self~class~new
sc=self~class~new
cs=self~supplier
do while cs~available
sc[cs~index]=cs~item
cs~next
end
cs=MayEnBag(arg(1))~supplier
do while cs~available
if sc~hasitem(cs~item,cs~index) then
r[cs~index]=sc~removeitem(cs~item, cs~index)
cs~next
end
return r

12.2.7.7 XOR
::method xor /* rCOLLECTION */
/* Returns a new relation that contains all items from self and
the argument except that all index-item pairs that appear in both collections
are removed. */
r=self~class~new
cs=self~supplier
do while cs~available
r[cs~index]=cs~item
cs~next
end
cs=MayEnBag(arg(1))~supplier
do while cs~available
if self~hasitem(cs~item,cs~index) then
r~removeitem(cs~item, cs~index)
else
r[cs~index]=cs~item
cs~next
end
return r

12.2.7.8 DIFFERENCE
::method difference /* rCOLLECTION */
/* Returns a new relation containing only those index-item pairs from the
SELF whose indexes the other collection does not contain. */
r=self~class~new

146
 cs=self~supplier
do while cs~available
r[cs~index]=cs~item
cs~next
end
cs=MayEnBag(arg(1))~supplier
do while cs~available
r~removeitem(cs~item,cs~index)
cs~next
end
return r

12.2.7.9 SUBSET
::method subset /* rCOLLECTION */
return self~difference(arg(1))~items = 0

12.2.7.10 REMOVEITEM
::method removeitem /* rANY rANY */
a=self~exposed
/* Returns and removes from a relation the member item value (associated with
the specified index). If value is not a member item associated with index
index, this method returns the NIL object and removes no item. */
do j=4 by 3 to 1+3*a[1]/*ItemsCount*/
if a[j]==arg(1) & a[j+1]==arg(2) then do
self~removeit(j)
return arg(1)
end
end j
return .nil

12.2.7.11 INDEX
::method index /* rANY */
a=self~exposed
/* Returns the index for the specified item. If there is more than one index
associated with the specified item, the one this method returns is not
defined. */
do j=4 by 3 to 1+3*a[1]/*ItemsCount*/
if arg(1)==a[j] then return a[j+1]
end j
return .nil

12.2.7.12 ALLAT
::method allat /* rANY */
a=self~exposed
/* Returns a single-index array containing all the items associated with the
specified index. */
r=.array~new
do j=4 by 3 to 1+3*a[1]/*ItemsCount*/
if arg(1)==a[j+1] then
r[r~dimension(1)+1] = a[j]
end j
return r
12.2.7.13 HASITEM
::method hasitem /* rANY rANY */
a=self~exposed
/* Returns 1 (true) if the relation contains the member item value (associated
with specified index). Returns 0 (false) otherwise. */
do j=4 by 3 to 1+3*a[1]/*ItemsCount*/
if a[j]==arg(1) & a[j+1]==arg(2) then return 1
end j
return 0

12.2.7.14 ALLINDEX
::method allindex /* rANY */

147
 a=self~exposed
/* Returns a single-index array containing all indexes for the specified
item. */
r=.array~new
do j=4 by 3 to 1+3*a[1]/*ItemsCount*/
if a[j]==arg(1) then do
r[r~dimension(1)+1]=a[j+1]
end
end j
return r

12.2.8 The bag class

::class 'Bag' subclass relation
/* A bag is a collection that restricts the member items to having a value that
is the same as the index. Any object can be placed in a bag, and the same
object can be placed in a bag multiple times. */

12.2.8.1 OF
::method of class /* 1 or more rANY */
/* Returns a newly created bag containing the specified value objects. */
r=self~new
do j=1 to arg()
r~put(arg(j))
end j
return r

12.2.8.2 PUT
::method put /* rANY oANY */
/* Committee does away with second argument? */
/* Makes the object value a member item of the collection and associates it with
the specified index. If you specify index, it must be the same as value. */
if arg(2,'E') then
if arg(2)\==arg(1) then signal error
self~put:super(arg(1),arg(1))

12.2.8.3 UNION
::method union /* rCOLLECTION */
return CommonUnion(self,EnBag(arg(1)))

12.2.8.4 INTERSECTION
::method intersection /* rCOLLECTION */
return CommonIntersect(self,EnBag(arg(1)))

12.2.8.5 XOR
::method xor /* rCOLLECTION */
return CommonXor(self,EnBag(arg(1)))

12.2.8.6 DIFFERENCE
::method difference /* rCOLLECTION */
return CommonDifference(self,EnBag(arg(1)))

12.2.9 The directory class

::class 'Directory' subclass Collection

12.2.9.1 AT
::method at /* rANY */
a=self~exposed
/* Returns the item associated with the specified index. */
j=self~findindex(arg(1))

148
 if j=0 then return .nil
/* Run the method if there is one. */
if a[j+2] then return self~run(a[j])
return a[j]

12.2.9.2 PUT
::method put /* rANY rANY */
a=self~exposed
/* Makes the object value a member item of the collection and associates it with
the specified index. */
if \arg(2)~hasmethod('MAKESTRING') then call Raise 'Syntax', 93.938
self~put:super(arg(1),arg(2)~makestring)
return

12.2.9.3 MAKEARRAY
::method makearray
forward message 'MAKEARRAYX'

12.2.9.4 SUPPLIER
::method supplier
a=self~exposed
/* Returns a supplier object for the directory. */
/* Check out what happens to the SETENTRY fields. */
r=.array~new /* For items */
do j=4 by 3 to 1+3*a[1]/*ItemsCount*/
r[r~dimension(1)+1]=a[j]
end j
return .supplier~new(r,self~makearray)

12.2.9.5 UNION
::method union /* rCOLLECTION */
return CommonUnion(self,arg(1))

12.2.9.6 INTERSECTION
::method intersection /* rCOLLECTION */
return CommonIntersect(self,arg(1))

12.2.9.7 XOR
::method xor /* rCOLLECTION */
return CommonXor(self,arg(1))

12.2.9.8 DIFFERENCE
::method difference /* rCOLLECTION */
return CommonDifference(self,arg(1))

12.2.9.9 SUBSET
::method subset /* rCOLLECTION */
return self~difference(arg(1))~items = 0

12.2.9.10 SETENTRY
::method setentry /* rSTRING oANY */
a=self~exposed
/* Sets the directory entry with the specified name (translated to uppercase) to
the second argument, replacing any existing entry or method for the specified
name. */
n=translate(arg(1))
j=self~findindex(n)
if j=0 & \arg(2,'E') then return
if \arg(2,'E') then do /* Removal */
self~removeit(j)

149
 return
end
if j=0 then do /* It's new */
a[1]/*ItemsCount*/=a[1]/*ItemsCount*/ +1
j=1+3*a[1]/*ItemsCount*/
a[j+1]=n
end
a[j]=arg(2)
a[j+2]=0
return

12.2.9.11 ENTRY
::method entry /* rSTRING */
a=self~exposed
/* Returns the directory entry with the specified name (translated to
uppercase). */
n=translate(arg(1))
j=self~findindex(n)
/*if j=0 then signal error according to online */
/* Online has something about running UNKNOWN. */
if j=0 then return .nil
/* If there is an entry decide whether to invoke it. */
if a~hasindex(j) then do
if \a[j+2] then return a[j]
return self~run(a[j])
end

12.2.9.12 HASENTRY
::method hasentry /* rSTRING */
/* Returns 1 (true) if the directory has an entry or a method for the specified
name (translated to uppercase) or 0 (false) otherwise. */
return self~findindex(translate(arg(1)))>0

12.2.9.13 SETMETHOD
::method setmethod /* rSTRING oMETHOD */
a=self~exposed
/* Associates entry with the specified name (translated to uppercase) with
method method. Thus, the language processor returns the result of running
method when you access this entry. */
/* (Part of METHOD checking converts string or array to actual method.) */
n=translate(arg(1))
j=self~findindex(n)
if j=0 & \arg(2,'E') then return
if \arg(2,'E') then do
self~removeit(j)
return
end
if j=0 then do /* It's new */
a[1]/*ItemsCount*/=a[1]/*ItemsCount*/ +1
j=1+3*a[1]/*ItemsCount*/
a[j+1]=n
end
a[j]=arg(2)
a[j+2]=1
return

12.2.9.14 UNKNOWN
::method unknown /* rSTRING rARRAY */
/* Runs either the ENTRY or SETENTRY method, depending on whether the message
name supplied ends with an equal sign. If the message name does not end with an
equal sign, this method runs the ENTRY method, passing the message name as its
argument. */
if right(arg(1),1)\=='=' then
return self~entry(arg(1))
/* ?? Not clear whether second argument is mandatory. */
t=.nil

150
 if arg(2,'E') then t=arg(2)[1]
self~setentry(left(arg(1),length(arg(1))-1),t)
12.3 The stem class
For some reason, the stem class doesn't have PUT and AT methods, which stops us having a general rule about []
synonyms AT, []= synonyms PUT.
Anyway, committee doing without this class as such.

Here is temporary stuff showing how to use algebra in the collection coding.

/* This 1998 version uses Rony's rules for XOR and INTERSECTION based on
UNION and DIFFERENCE */

/* Test Set-Operator-Methods on different collection objects */

/* This top part has some rough parts - not meant for standard. */
/* The dumps put out results sorted, so that comparisons can be made
between implementations that keep collections in different orders. */

/* Invocation example:
settest.cmd 1> tmp.res 2> tmp.err
*/

/* Initial verification that new definitions are in effect */

J18list = .List~new
if \J18list~hasmethod("J18") then signal error

/* Input collections used for the tests */

coll.1 = .array~of(1, 2,, 4)
coll.2 = .list~of(2, 3, 6)
coll.3 = .queue~new~~PUSH(2)~~PUSH(3)~~PUSH(7)
coll.4 = .directory~new~~setentry(1, "eins")~~setentry(3, "drei")
coll.5 = .bag~new~~put(2)~~put(3)~~put(5)~~put(2)
coll.6 = .relation~new~~"[]="("zwei", 2)~~"[]="("drei", 3)~~"[]="("vier", 8)~~"[]="("drei",3)
coll.7 = .set~of(2, 3, 9)
coll.8 = .table~new~~"[]="("zwei", 2)~~"[]="("drei", 3)~~"[]="("vier", 10)
coll.0 = 8

message.1 = "UNION"
message.2 = "INTERSECTION" /* index the same in both */
message.3 = "DIFFERENCE" /* if index only in 1st collection */
message.4 = "XOR" /* unique index among both collections */
message.5 = "SUBSET" /* target is subset of other collection */
message.0 = 5

target. = coll.

hstart = 4
istart = 1
jstart = 1

151
output = 1
setOfTargets = .set~new

SAY "Test Results of Set Operations on Collection Classes -- dated" date('U')

SAY

DO h = hstart TO target.0 /* loop over target */

targetID = target.h~class~id
IF \setOfTargets~hasindex(targetID) THEN
DO
SAY
SAY CENTER(" Target:" targetID "", 70, "=")
setOfTargets~put(targetID)
output = 1
END

DO i = istart TO coll.0 /* loop over other collections */

if output then do
output = 0
argumentID = coll.i~class~id
SAY
SAY CENTER(" argument:" argumentID "", 65, "=")
SAY
SAY "INPUT:"
SAY "contents of" pp(targetID) "target:"
CALL dump_collection target.h
SAY

SAY "contents of" pp(argumentID) "argument:"

CALL dump_collection coll.i
SAY
SAY CENTER(" start set operators ", 65, "-")
end

DO j = jstart TO message.0 /* loop over set operators */

tmpString1 = RIGHT("h" pp(h) "i" pp(i) "j" pp(j), 65)
tmpString2 = pp(targetID "~" message.j || "(" argumentID ")")
SAY OVERLAY( tmpString2, tmpString1 )
/* set resume parameter in case of error*/
jstart = j+1
IF jstart>message.0 THEN DO
istart = i+1

IF istart>coll.0 THEN DO
hstart = h+1
istart = 1
152
 END
jstart = 1
output = 1
END
/* process method invocation */
IF target.h~hasmethod(message.j) THEN DO
tmp = .message~new(target.h, message.j, "I", coll.i)~send
if "The String class"=tmp~class~defaultname then do
if datatype(tmp,"B") then do
if tmp then
SAY " Result is TRUE"
else
SAY " Result is FALSE"
end
end
else CALL dump_collection tmp
END
ELSE
SAY pp(targetID) "does not have method ~" pp(message.j)

SAY LEFT("", 40, "-")

END
jstart = 1
END
jstart = 1
istart = 1
output = 1
END

RETURN

dump_collection:procedure
USE ARG collection
k = .array~new
i = .array~new
tmpSupp = collection~supplier
DO WHILE tmpSupp~AVAILABLE
k[k~dimension(1)+1]=tmpSupp~INDEX
i[i~dimension(1)+1]=tmpSupp~ITEM
tmpSupp~NEXT
END
do until hope
hope=1
do j=1 to k~dimension(1)-1
if k[j]~string>k[j+1]~string |,
(k[j]~string=k[j+1]~string & i[j]~string<i[j+1]~string) then do
153
 t=k[j];k[j]=k[j+1];k[j+1]=t
t=i[j];i[j]=i[j+1];i[j+1]=t
hope=0
end
end
end
if 0=collection~items then
say " The result is empty!"
else
do j=1 to k~dimension(1)
SAY " " "index" pp(k[j]) "item" pp(i[j])
end
RETURN
/* Auxiliary routines */
pp: RETURN "[" || ARG(1)~string || "]"
/*===================================================================
======*/
/* X3J18 Rexx Language Standard proposal for the Set-like operations on the
Collection classes */

/* In the same way that the first standard uses BIFs which are defined using
other BIFs and ultimately the arithmetic and character operators, the second
standard can define classes using other classes and some fundamental basis.

This program gives the definition of some other classes, in a form which
(when thoroughly developed) might be part of the second standard. It also
has a testing mechanism, which will not be part of a standard.

This particular program is implementing collections on top of array.

*/
/* The class Collection is not one builtin, but is used to simplify the
definition. */

::class 'Collection'

::method init
expose a
/* A collection is modelled as using 3 slots in an array for each element.
The first slot holds the item, the second the index, and the third is
used by particular types of collection. This order of slots is arbitary,
chosen to match order of arguments for PUT and SUPPLIER~NEW. */
/* The first set of 3 slots is reserved for other purposes, to avoid
having separate variables which the subclassing would need to access. */
a=.array~new
a[1]/*ItemsCount*/=0
a[2]/*Unique*/=0
154
 return self

::method exposed private

expose a
/* This method allows subclasses to get at the implementation of Collection. */
return a

::method findindex private

expose a
/* Returns array index if the collection contains any item associated with the
index specified or returns 0 otherwise. */
do j=4 by 3 to 1+3*a[1]/*ItemsCount*/
if a[j+1]==arg(1) then return j
end j
return 0

::method at /* rANY */
expose a
/* Returns the item associated with the specified index. */
j=self~findindex(arg(1))
if j=0 then return .nil
return a[j]

::method '[]'
/* Synonym for the AT method. */
forward message 'AT'

::method put /* rANY rANY */

expose a
use arg item, index
/* Replaces any existing item associated with the specified index with the new
item. Otherwise adds the item-index pair. */
j=self~findindex(index)
if j>0 then do
a[j]=item
return
end
a[1]/*ItemsCount*/=a[1]/*ItemsCount*/+1
j=1+3*a[1]/*ItemsCount*/
a[j]=item
a[j+1]=index
a[j+2]=0
return /* Error 91 in OOI if context requiring result. */

::method '[]='
/* Synonym for the PUT method. */
155
 forward message 'PUT'

::method hasindex /* rANY */

/* Returns 1 (true) if the collection contains any item associated with the
index specified or returns 0 (false) otherwise. */
return self~findindex(arg(1))>0

::method items
expose a
/* Returns the number of items in the collection. */
return a[1]/*ItemsCount*/

::method remove /* rANY */

expose a
/* Returns and removes from a collection the member item with the specified
index. */
j=self~findindex(arg(1))
if j=0 then return .nil
r=a[j]
self~removeit(j)
return r

::method removeit private

expose a
use arg j
/* Remove relevant slots from the array, with compaction. */
do j=j+3 by 3 to 1+3*a[1]/*ItemsCount*/
a[j-3]=a[j];a[j-2]=a[j+1];a[j-1]=a[j+2]
end j
a[1]/*ItemsCount*/=a[1]/*ItemsCount*/-1
return

::method makearray
expose a
/* Returns a single-index array containing the receiver list items. */
r = .array~new /* To build result in. */
do j=4 by 3 to 1+3*a[1]/*ItemsCount*/
r[r~dimension(1)+1]=a[j]
end j
return r

::method makearrayx private

expose a
/* Returns a single-index array containing the receiver index items. */
r = .array~new /* To build result in. */
do j=4 by 3 to 1+3*a[1]/*ItemsCount*/
156
 r[r~dimension(1)+1]=a[j+1]
end j
return r

::method supplier
expose a
/* Returns a supplier object for the list. */
return .supplier~new(self~makearray:.collection,self~makearrayx)

::class 'List' subclass Collection

::method J18 /* Here to demonstrate .LIST is replaced */

return

/* List and Queue are special because there is an order to their elements. */

::method put /* rANY rANY */

use arg item, index
a=self~exposed
/* PUT for a List must not be an insertion. */
j=self~findindex(index)
if j=0 then call Raise 'Syntax',93.918
a[j]=item
return

::method of class /* 1 or more oANY Are they omittable? Not in IOO */

/* Returns a newly created list containing the specified value objects in the
order specified. */
r = self ~ new
do j = 1 to arg()
r ~ insert(arg(j))
end j
return r

::method insert /* rANY oANY */

use arg item, index
a=self~exposed
/* Returns a list-supplied index for a new item, of specified value, which is
added to the list. The new item follows the existing item with the specified
index in the list ordering. */
/* Establish the index of what preceeds the new element. */
/* If there was no index given, the new item becomes the last on list. */
/* .nil argument means first */
if arg(2,'E') then p=arg(2)
else p=self~last
/* Convert from list index to underlying array index. */
157
 if p==.nil then j=1
else j=self~findindex(p)
if j=0 then call Raise 'Syntax',93.918
j=j+3 /* Where new entry will be. */
/* Move space to required place. */
a[1]/*ItemsCount*/=a[1]/*ItemsCount*/+1
do k=1+3*a[1]/*ItemsCount*/ by -3 to j+3
a[k]=a[k-3];a[k+1]=a[k-2];a[k]=a[k-3]
end
/* Insert new element. */
a[j]=item
/* A new, unique, index is needed. */
/* The basic requirement is for something unique, so this would be correct:
i=.object~new /* a unique object, used as a key (the index on the list) */
*/
/* However, a number can be used. (At risk of the user thinking it is
sensible to do arithmetic on it.) */
a[j+1]=a[2]/*Unique*/;a[2]/*Unique*/=a[2]/*Unique*/+1
a[j+2]=0
return a[j+1]

::method first
a=self~exposed
/* Returns the index of the first item in the list. */
if a[1]/*ItemsCount*/=0 then return .nil
return a[5]

::method last
a=self~exposed
/* Returns the index of the last item in the list. */
if a[1]/*ItemsCount*/=0 then return .nil
return a[3*a[1]/*ItemsCount*/+2]

::method firstitem
a=self~exposed
/* Returns the first item in the list. */
if a[1]/*ItemsCount*/=0 then return .nil
return a[4]

::method lastitem
a=self~exposed
/* Returns the last item in the list. */
if a[1]/*ItemsCount*/=0 then return .nil
return a[3*a[1]/*ItemsCount*/+1]

::method next /* rANY */

158
 a=self~exposed
/* Returns the index of the item that follows the list item having the specified
index. */
j=self~findindex(arg(1))
if j=0 then call Raise 'Syntax',93.918
j=j+3
if j>3*a[1]/*ItemsCount*/ then return .nil /* Next of last was requested. */
return a[j+1]

::method previous /* rANY */

a=self~exposed
/* Returns the index of the item that precedes the list item having the
specified index. */
j=self~findindex(arg(1))
if j=0 then call Raise 'Syntax',93.918
j=j-3
if j<4 then return .nil /* Previous of first was requested. */
return a[j+1]

::method section /* rANY oWHOLE>=0 */

a=self~exposed
/* Returns a new list containing selected items from the receiver list. The
first item in the new list is the item corresponding to the index specified,
in the receiver list. */
j=self~findindex(arg(1))
if j=0 then call Raise 'Syntax',93.918
r = .list~new /* To build result in. */
if arg(2,'E') then s = arg(2)
else s = self~items;
do s
r~insert(a[j])
j=j+3
if j>1+3*a[1]/*ItemsCount*/ then leave
end
return r

::class 'Queue' subclass Collection

/* A queue is a sequenced collection with whole-number indexes. The
indexes specify the position of an item relative to the head (first item) of
the queue. Adding or removing an item changes the association of an index to
its queue item. */

::method push /* rANY */

/* Adds the object value to the queue at its head. */
a=self~exposed
a[1]/*ItemCount*/=a[1]/*ItemCount*/+1
159
/* Slide along to make a space. */
do j=1+3*a[1]/*ItemCount*/ by -3 to 7
a[j]=a[j-3]
a[j+1]=a[j-2]+1; /* Index changes */
end j
a[4]=arg(1)
a[5]=1
return

::method pull
/* Returns and removes the item at the head of the queue. */
a=self~exposed
if a[1]/*ItemCount*/=0 then return .nil /* Stays empty */
r=a[4]
a[1]/*ItemCount*/=a[1]/*ItemCount*/-1
do j=4 by 3 to 1+3*a[1]/*ItemCount*/
a[j]=a[j+3]
a[j+1]=a[j+4]-1; /* Index changes */
end j
return r

::method queue /* rANY */

/* Adds the object value to the queue at its tail. */
a=self~exposed
a[1]/*ItemCount*/=a[1]/*ItemCount*/+1
a[1+3*a[1]/*ItemCount*/]=arg(1)
a[2+3*a[1]/*ItemCount*/]=a[1]/*ItemCount*/
return

::method peek
a=self~exposed
/* Returns the item at the head of the queue. The collection remains unchanged.
*/
return a[4]

::method remove /* rWHOLE>0 */

/* Returns and removes from a collection the member item with the specified
index. */
a=self~exposed
if a[1]/*ItemCount*/<arg(1) then return .nil
r=self~remove:super(arg(1))
/* Reset the indexes. */
k=0
do j=4 by 3 to 1+3*a[1]/*ItemsCount*/
k=k+1
a[j+1]=k
160
 end j
return r

::class 'Table' subclass Collection

::method makearray
/* Returns a single-index array containing the index objects. */
/* This is different from Collection MAKEARRAY where items rather than indexes
are in the returned array. */
forward message 'MAKEARRAYX'

::method union /* rCOLLECTION */

return CommonUnion(self,arg(1))

::method intersection /* rCOLLECTION */

/* Returns a new collection of the same class as SELF that
contains the items from SELF that have indexes also in the
argument. */
/* Actually an index in SELF can only be 'matched' with one in the
argument once. */
return self~difference(self~difference(arg(1)))

::method xor /* rCOLLECTION */

/* Returns a new relation that contains all items from self and
the argument except that all index-item pairs that appear in both collections
are removed. */
return CommonXor(self,arg(1))

::method difference /* rCOLLECTION */

return CommonDifference(self,arg(1))

::method subset /* rCOLLECTION */

return self~difference(arg(1))~items = 0

::class 'Set' subclass table

/* A set is a collection that restricts the member items to have a value that is
the same as the index. Any object can be placed in a set. There can be only
one occurrence of any object in a set. */

/* Second arg same as first. Committee has dropped second? */

::method put /* rANY oANY */
/* Makes the object value a member item of the collection and associates it with
specified index. */
if arg(2,'E') then
if arg(2)\==arg(1) then signal error /* 949 */
self~put:super(arg(1),arg(1))
161
::method of class /* 1 or more rANY */
/* Returns a newly created set containing the specified value objects. */
r=self~new
do j=1 to arg()
r~put(arg(j))
end j
return r

::method union /* rCOLLECTION */

return CommonUnion(self,EnBag(arg(1)))

::method intersection /* rCOLLECTION */

return self~difference(self~difference(arg(1)))

::method xor /* rCOLLECTION */

return CommonXor(self,EnBag(arg(1)))

::method difference /* rCOLLECTION */

return CommonDifference(self,EnBag(arg(1)))

::class 'Relation' subclass Collection

::method put /* rANY rANY */

use arg item, index
a=self~exposed
/* Makes the object value a member item of the relation and associates it with
the specified index. If the relation already contains any items with the
specified index, this method adds a new member item value with the same index,
without removing any existing members */
a1]/*IitemCount*/= a1]/*IitemCount*/+1j
j=+3* a1]/*IitemCount*/d
aj]= itej
aj+1]= index
aj+2]=0x
return /*Errxor91e i OOI ify conexut rquirtingresults. */

::method items /*orANY */

a=self~exposed
/* Returns thenuember of relation items with the specified index. Ifyoue speciyx
nod index, this methodrReturns thetotalenuember of items associatds withall,
indetes nf the relatios.*/
 if arg(1)==a[j+1] then n=n+1
end j
return n

::method makearray
forward message 'MAKEARRAYX'

::method supplier /* oANY */

a=self~exposed
/* Returns a supplier object for the collection. If an index is specified, the
supplier enumerates all of the items in the relation with the specified
index. */
m=.array~new /* For the items */
r=.array~new /* For the indexes */
do j=4 by 3 to 1+3*a[1]/*ItemsCount*/
if arg(1,'E') then
if arg(1)\==a[j+1] then iterate
n=r~dimension(1)+1
m[n]=a[j]
r[n]=a[j+1]
end j
return .supplier~new(m,r)

::method union /* rCOLLECTION */

/* Union for a relation is just all of both. */
r=self~class~new
cs=self~supplier
do while cs~available
r[cs~index]=cs~item
cs~next
end
cs=MayEnBag(arg(1))~supplier
do while cs~available
r[cs~index]=cs~item
cs~next
end
return r

::method intersection /* rCOLLECTION */

return self~difference(self~difference(arg(1)))

::method xor /* rCOLLECTION */

return CommonXor(self,arg(1))

::method difference /* rCOLLECTION */

/* Returns a new relation containing only those index-item pairs from the
163
SELF whose indexes the other collection does not contain. */
r=self~class~new
cs=self~supplier
do while cs~available
r[cs~index]=cs~item
cs~next
end
cs=MayEnBag(arg(1))~supplier
do while cs~available
r~removeitem(cs~item,cs~index)
cs~next
end
return r

::method subset /* rCOLLECTION */

return self~difference(arg(1))~items = 0

::method removeitem /* rANY rANY */

a=self~exposed
/* Returns and removes from a relation the member item value (associated with
the specified index). If value is not a member item associated with index
index, this method returns the NIL object and removes no item. */
do j=4 by 3 to 1+3*a[1]/*ItemsCount*/
if a[j]==arg(1) & a[j+1]==arg(2) then do
self~removeit(j)
return arg(1)
end
end j
return .nil

::method index /* rANY */

a=self~exposed
/* Returns the index for the specified item. If there is more than one index
associated with the specified item, the one this method returns is not
defined. */
do j=4 by 3 to 1+3*a[1]/*ItemsCount*/
if arg(1)==a[j] then return a[j+1]
end j
return .nil

::method allat /* rANY */

a=self~exposed
/* Returns a single-index array containing all the items associated with the
specified index. */
r=.array~new
do j=4 by 3 to 1+3*a[1]/*ItemsCount*/
164
 if arg(1)==a[j+1] then
r[r~dimension(1)+1] = a[j]
end j
return r

::method hasitem /* rANY rANY */

a=self~exposed
/* Returns 1 (true) if the relation contains the member item value (associated
with specified index). Returns 0 (false) otherwise. */
do j=4 by 3 to 1+3*a[1]/*ItemsCount*/
if a[j]==arg(1) & a[j+1]==arg(2) then return 1
end j
return 0

::method allindex /* rANY */

a=self~exposed
/* Returns a single-index array containing all indexes for the specified
item. */
r=.array~new
do j=4 by 3 to 1+3*a[1]/*ItemsCount*/
if a[j]==arg(1) then do
r[r~dimension(1)+1]=a[j+1]
end
end j
return r

::class 'Bag' subclass relation

/* A bag is a collection that restricts the member items to having a value that
is the same as the index. Any object can be placed in a bag, and the same
object can be placed in a bag multiple times. */

::method of class /* 1 or more rANY */

/* Returns a newly created bag containing the specified value objects. */
r=self~new
do j=1 to arg()
r~put(arg(j))
end j
return r

::method put /* rANY oANY */

/* Committee does away with second argument? */
/* Makes the object value a member item of the collection and associates it with
the specified index. If you specify index, it must be the same as value. */
if arg(2,'E') then
if arg(2)\==arg(1) then signal error
self~put:super(arg(1),arg(1))
165
/* Bag may be a subclass of relation but many methods have different
semantics. */

::method union /* rCOLLECTION */

return CommonUnion(self,EnBag(arg(1)))

::method intersection /* rCOLLECTION */

return self~difference(self~difference(arg(1)))

::method xor /* rCOLLECTION */

return CommonXor(self,EnBag(arg(1)))

::method difference /* rCOLLECTION */

return CommonDifference(self,EnBag(arg(1)))

::class 'Directory' subclass Collection

/* Later we take three array elements for each element in the directory, one
for the item, one to contain the index, one to say if the item is a method
to be run or not. */

::method at /* rANY */
a=self~exposed
/* Returns the item associated with the specified index. */
j=self~findindex(arg(1))
if j=0 then return .nil
/* Run the method if there is one. */
if a[j+2] then return self~run(a[j])
return a[j]

::method put /* rANY rANY */

a=self~exposed
/* Makes the object value a member item of the collection and associates it with
the specified index. */
if \arg(2)~hasmethod('MAKESTRING') then call Raise 'Syntax', 93.938
self~put:super(arg(1),arg(2)~makestring)
return

::method makearray
forward message 'MAKEARRAYX'

::method supplier
a=self~exposed
/* Returns a supplier object for the directory. */
/* Check out what happens to the SETENTRY fields. */
r=.array~new /* For items */
166
 do j=4 by 3 to 1+3*a[1]/*ItemsCount*/
r[r~dimension(1)+1]=a[j]
end j
return .supplier~new(r,self~makearray)

::method union /* rCOLLECTION */

return CommonUnion(self,arg(1))

::method intersection /* rCOLLECTION */

return self~difference(self~difference(arg(1)))

::method xor /* rCOLLECTION */

return CommonXor(self,arg(1))

::method difference /* rCOLLECTION */

return CommonDifference(self,arg(1))

::method subset /* rCOLLECTION */

return self~difference(arg(1))~items = 0

::method setentry /* rSTRING oANY */

a=self~exposed
/* Sets the directory entry with the specified name (translated to uppercase) to
the second argument, replacing any existing entry or method for the specified
name. */
n=translate(arg(1))
j=self~findindex(n)
if j=0 & \arg(2,'E') then return
if \arg(2,'E') then do /* Removal */
self~removeit(j)
return
end
if j=0 then do /* It's new */
a[1]/*ItemsCount*/=a[1]/*ItemsCount*/ +1
j=1+3*a[1]/*ItemsCount*/
a[j+1]=n
end
a[j]=arg(2)
a[j+2]=0
return

::method entry /* rSTRING */

a=self~exposed
/* Returns the directory entry with the specified name (translated to
uppercase). */
n=translate(arg(1))
167
 j=self~findindex(n)
/*if j=0 then signal error according to online */
/* Online has something about running UNKNOWN. */
if j=0 then return .nil
/* If there is an entry decide whether to invoke it. */
if a~hasindex(j) then do
if \a[j+2] then return a[j]
return self~run(a[j])
end

::method hasentry /* rSTRING */

/* Returns 1 (true) if the directory has an entry or a method for the specified
name (translated to uppercase) or 0 (false) otherwise. */
return self~findindex(translate(arg(1)))>0

::method setmethod /* rSTRING oMETHOD */

a=self~exposed
/* Associates entry with the specified name (translated to uppercase) with
method method. Thus, the language processor returns the result of running
method when you access this entry. */
/* (Part of METHOD checking converts string or array to actual method.) */
n=translate(arg(1))
j=self~findindex(n)
if j=0 & \arg(2,'E') then return
if \arg(2,'E') then do
self~removeit(j)
return
end
if j=0 then do /* It's new */
a[1]/*ItemsCount*/=a[1]/*ItemsCount*/ +1
j=1+3*a[1]/*ItemsCount*/
a[j+1]=n
end
a[j]=arg(2)
a[j+2]=1
return

::method unknown /* rSTRING rARRAY */

/* Runs either the ENTRY or SETENTRY method, depending on whether the message
name supplied ends with an equal sign. If the message name does not end with an
equal sign, this method runs the ENTRY method, passing the message name as its
argument. */
if right(arg(1),1)\=='=' then
return self~entry(arg(1))
/* ?? Not clear whether second argument is mandatory. */
t=.nil
168
 if arg(2,'E') then t=arg(2)[1]
self~setentry(left(arg(1),length(arg(1))-1),t)

::routine CommonXor
/* Returns a new collection that contains all items from self and
the argument except that all indexes that appear in both collections
are removed. */
/* When the target is a bag, there may be an index in the bag that is
duplicated and the same value as an index in the argument. Should one
copy of the index survive in the bag? */
lhs=arg(1)~difference(arg(2))
rhs=Cast(arg(1),MayEnBag(arg(2)))~difference(arg(1))
return lhs~union(rhs)

::routine CommonUnion
/* Returns a new collection of the same class as SELF that
contains all the items from SELF and items from the
argument that have an index not in the first. */
/* Best to add them all. By adding non-receiver first we ensure that
receiver takes priority when same indexes. */
This = arg(1) /* self of caller */
r=This~class~new
cs=MayEnBag(arg(2))~supplier
do while cs~available
r[cs~index]=cs~item
cs~next
end
cs=This~supplier
do while cs~available
r[cs~index]=cs~item
cs~next
end
return r

::routine CommonDifference
/* Returns a new collection containing only those index-item pairs from the
SELF whose indexes the other collection does not contain. */
This = arg(1) /* self of caller */
r=This~class~new
cs=This~supplier
do while cs~available
r[cs~index]=cs~item
cs~next
end
cs=MayEnBag(arg(2))~supplier
169
 do while cs~available
r~remove(cs~index)
cs~next
end
return r

::routine MayEnBag
/* For List and Queue the indexes are dropped. */
r=arg(1)
if r~class == .List | r~class == .Queue then r=EnBag(r)
return r

::routine EnBag
r=.Bag~new
s=arg(1)~supplier
do while s~available
if arg(1)~class == .List | arg(1)~class == .Queue then
r[s~item]=s~item
else
/* This case is when the receiver is a Bag. */
r[s~index]=s~index
s~next
end
return r

/* This Cast routine commented away, since replaced by Oct 98 Rony version.
::routine Cast public
use arg Target, Other
TmpColl = Target~class~new /* Create an instance of type Target */
TmpSupp = Other~supplier /* Get supplier from Other */
signal on syntax
do while TmpSupp~available
TmpColl[TmpSupp~index] = TmpSupp~item
TmpSupp~next
end
return TmpColl

/* If syntax error 93.949, then target is an index-only collection like a set.*/

syntax:
if condition( "O" )~code = "93.949" then signal IndexOnly
raise propagate /* Unhandled syntax error, raise in caller */

IndexOnly: /* This for index-only collections. */

do while TmpSupp~available
TmpColl[TmpSupp~index] = TmpSupp~index
TmpSupp~next
170
 end
return TmpColl
End commented away */

/* 98-09-24, ---rgf;
CAST2.CMD
return a collection of type "target" which collected all
item/index pairs of the argument "other" */

:: ROUTINE cast PUBLIC

USE ARG target, other

SIGNAL ON SYNTAX
IF \ other ~ HASMETHOD( "SUPPLIER" ) THEN
RAISE SYNTAX 98.907 ARRAY ( "COLLECTION (i.e. argument2='other'-object must
have a 'SUPPLIER'-method)" )

tmpColl = target ~ CLASS ~ NEW /* create a an instance of type target */

tmpSupp = other ~ SUPPLIER /* get supplier from other */

/* is index of "other" usable ? */

bIndexUsable = other ~ HASMETHOD( "UNION" )

IF .Debug = .true THEN IF \ bIndexUsable THEN

SAY " /// index of 'other' not usable for setlike-operations"

/* possible syntax-error, if index and item must have the same value,
e.g. for sets/bags */
SIGNAL ON SYNTAX NAME INDEX_ONLY
target ~ CLASS ~ NEW ~ PUT( 1, 2 ) /* test, if target-type is index-only
*/

SIGNAL ON SYNTAX
DO WHILE tmpSupp ~ AVAILABLE
IF bIndexUsable THEN tmpColl[ tmpSupp ~ INDEX ] = tmpSupp ~ ITEM
ELSE tmpColl[ tmpSupp ~ ITEM ] = tmpSupp ~ ITEM
tmpSupp ~ NEXT
END
RETURN tmpColl

INDEX_ONLY : /* this is for index-only collections (e.g. sets, bags) */

SIGNAL ON SYNTAX
IF .Debug = .true THEN
SAY " \\\ 'target' is an index-only collection (index==item)"
DO WHILE tmpSupp ~ AVAILABLE
171
 IF bIndexUsable THEN tmpColl[ tmpSupp ~ INDEX ] = tmpSupp ~ INDEX
ELSE tmpColl[ tmpSupp ~ ITEM ] = tmpSupp ~ ITEM
tmpSupp ~ NEXT
END
RETURN tmpColl

SYNTAX: RAISE PROPAGATE /* raise error in caller */

12.1 The stream class

The stream class provides input/output on external streams.
::class stream

::method init /* rString */

Initializes a stream object for a stream named name, but does not open the stream.
::method query /* keywords */
There is also QUERY as command with method COMMAND.
Used with options, the QUERY method returns specific information about a stream.
::method charin

::method charout

::method chars

::method linein

::method lineout

::method lines

::method qualify

::method command /* rString */

Returns a string after performing the specified stream command.
::method open
There is also OPEN as command with method COMMAND.
Opens the stream to which you send the message and returns "READY:".
Committee dropping OPEN POSITION QUERY SEEK as methods in favour of command use.
::method state
Returns a string that indicates the current state of the specified stream.
::method say

::method uninit

::method position /* Ugh */

POSITION is a synonym for SEEK.
::method seek /* Ugh */
Sets the read or write position a specified number (offset) within a persistent stream.
::method flush
Returns "READY:". Forces any data currently buffered for writing to be written to the stream receiving the
message.
There is also FLUSH as command with method COMMAND.
Committee dropping FLUSH.
::method close
Closes the stream that receives the message.
There is also CLOSE as command with method COMMAND.
Semantics are 'seen by other thread'.
::method string

172
::method makearray /* rCHARLINE */
Returns a fixed array that contains the data from the stream in line or character format, starting from the
current read position.
::method supplier
Returns a supplier object for the stream.
::method description

::method arrayin /* rCHARLINE */

Mixed case value works on OOI.
Committee dropping Arrayin & Arrayout . Arrayin == MakeArray
Returns a fixed array that contains the data from the stream in line or character format, starting from the
current read position.
::method arrayout /* rARRAY rCHARLINE */
Returns a stream object that contains the data from array.

12.2 The alarm class

::class alarm

::method init /* Time, Msg */

Sets up an alarm for a future time atime.
::method cancel
Cancels the pending alarm request represented by the receiver. This method takes no action if the
specified time has already been reached.

12.3 The monitor class

The Monitor class forwards messages to a destination object.

.local['OUTPUT'] = .monitor~new(.output)

::class monitor

12.3.1 INIT
Initializes the newly created monitor object.
::method init /* oDESTINATION */
expose Destination
Destination = .queue~new
if arg(1,'E') then Destination~push(arg(1))
return

12.3.2 CURRENT

Returns the current destination object.

::method current
expose Destination
return Destination[1]

12.3.3 DESTINATION

Returns a new destination object.

::method destination /* oDESTINATION */

expose Destination
if arg(1,'E') then Destination~push(arg(1))
else Destination~pull
return Destination[1]

12.3.4 UNKNOWN
Reissues or forwards to the current monitor destination all unknown messages sent to a monitor object

173
::method unknown
expose Destination

Extra parens needed here in original OREXX syntax

forward to destination[1] message arg(1) arguments arg(2)

return

174
Annex A
(informative)
Rationale
This annex explains some of the decisions made by the committee that drafted this standard, and assists
in the understanding of this document. Some of the statements made here are opinions rather than facts.
These should be interpreted as if prefixed by "In the opinion of the X3J18 committee...".
The language described in this standard is, almost entirely, a compatible extension of the language
described by the third reference of Annex C, which we call "Classic Rexx".
The extension allows programs to be written in a less monolithic fashion; "Directives" are introduced to
allow one file to contain several executable units and to allow a program to be written as several files.
The functional extension centers on the addition of objects. Unlike the individual strings which are the
data of Classic Rexx, an object may be composite. The use of identifiers to reference objects is an
indirect reference, that is two identifiers may refer to the same object. Classic Rexx avoided any aliasing,
even to the extent having by-reference parameters, to promote simple error free programming. In the
years since Rexx originated the problems tackled by programmers have become more complex and data
structures larger, so that the benefit of simplicity is outweighed by the power of assignment semantics
that are not simply copying all the data.
Even with the addition of references Rexx remains a typeless language, in the sense that the
programmer need not consider underlying hardware formats such as LONG or FLOAT representations.
Object Rexx does have classes, which are the hardware independent analogy to types. The class of an
object corresponds to the operations that can be performed upon it.

Incompatibilities
The incompatibilities from Classic Rexx are:
Assignment of compound variables, as in ABC. = PQR., is an assignment of references so that ABC.
subsequently refers to the same object as PQR., as opposed to making the default value of ABC. that of
PQR.. This change was necessary to fit compound variables into the object framework, in particular
allowing USE ARG to handle compound variables as by-reference parameters. The first reference of
Annex B discouraged use of this construct in Classic Rexx programs. "Breakage" of programs due to this
incompatibility is rare.
Also something in condition handling that I don't know the reason for.

Call
The call instruction has been extended to allow for a computed name of the callee. Syntax
considerations prevent a similar thing being done for functions.

Concurrency
Meet 17 minutes

Guard
Meet 17 minutes

To be processed:

The following decisons are abstracted from minutes. We need to ensure they are covered in the main standard and
their rationale appropriately reworded for this annex.

Aliasing. Assignment is viewed as making the target reference the same object as the source. Hence
the object (and changes to it) may be accessed through more than one name. For 'immutable' objects a
changed version of an object can only be produced by creating a new object. For compatibility with
classic Rexx, strings are immutable objects. Non-strings may or may not be immutable. Note that there
is an alternative model in which distinction is made between assignments which copy values and
assignments which copy references. This alternative was not chosen; the committee prefered the model
in which all data names are naming references (which may be implicitly followed to values).

175
Arguments 'by-reference'. The introduction of aliasing makes this natural although the detail has
simple-versus-general contentions. (Is it necessary for simple strings to be passed by reference.
Encapsulation. An object may 'own' some variables and access to those may be limited (so that
re-implementation of the object could use different variables without upsetting the usage of the object).
Classess. There will be 'factory' objects capable of creating multiple new objects which have common
characteristics about how they can be used.
Inheritance and hierachy. The semantics of a clas may be specified by adding to the semantics of
another class. This relation is used to form a tree. We prefer a singly rooted tree, rooted in the class
'Object' which is built-in to the language. Other classes will also be built-in. Experience with OOM and
other languages is that unrestricted inheritance by one class from multiple classes does not work in the
way the coder intended (the implementations of the classes do not combine successfully). If multiple
inheritance is added to Rexx at all, it will be in the cautious 'MIXIN' flavor of OOI.
Messaging: Executing some labelled code which is associated with objects of a given class is a form of
invocation that is sufficiently different from classic Rexx to justify a new syntax construct. The new syntax
is Receiver ~ MethodName(Arguments) and implies both a different search for the method to be invoked
and a special role for the receiver as opposed to the other arguments of the invocation.
Packaging: In principle a 'program builder' could be used in developing Rexx programs with many classes
and methods, and that builder could hide from the coder the details of how the configuration held the
methods. However, rather than define a program builder we are choosing to define a simple method of
holding multiple classes & mthods (with specification of their hierarchy) within a single text file. The
non-executable dividers in such a file are known as directives. The files are known as packages and a
package may specify (by directive) that it requires another package in order to function correctly. There
are questions about when initialization of required packages occurs; we intend to find a solution that does
not require the complete graph of requirements to be initialized before other code is executed.
A note on the syntax of directives. When no special token (eg ::) is used to introduce directives the
directives are be recognizable by the spelling of the keyword. (CLASS REQUIRES etc.) The purpose of
the special token is emphasis of directives rather than implementation ease in "pre-processing" the
directives.
Packages in non-Rexx. It is necessary to exploit packages that are not written in Rexx. To invoke their
methods it is necessary that the package makes known to the Rexx method search the names of the
classes and their methods. To do more than invoke the methods (eg to subclass the the external
classes) requires complicated mechanisms and may not be a requirement.
External procedures. To allow Classic internal procedures to be separated into different files with undue
change of semantics, the PROCEDURE statement will be permiteed as the first statement of a routine
which is in a separate file.
Concurrency will be added, that is multiple execution cursors progressing through one program. The
mechanism for creating multiple cursors will be the "early reply" where one cursor becomes two; one of
two progresses by "falling through" the early reply and the other starts its progress after the site of the
current invocation. Multiple cursors carry the risk of execution interleaving in a way which negates the
coder's intentions in writing so that clauses would execute sequentially. The language definition will be
tightened to ensure atomicity of string assignment etc. Additionally, a set of rules about allowing two
cursors on the same method at the same time will provide a reduction of the risk. Since in many cases
the data which have to be maintained consistent will reside in a single object the rules are object-based.
In general a cursor on a method executing against a particular object will delay any other cursor from
executing methods against that object.
This rule provides sensible synchronization without much effort from the programmer but other controls
may be provided:
a) Stronger control, eg only one cursor within the methods of a set of objects.
b) More detailed control, eg division of a method into sections which allow/disallow other cursors into the
section.
Extended Variable Pools. The API for variable pools will need to be extended to reflect the model in
which the named content in a pool is always a reference (and the reference is followed when the value of
a string is required.) We note that OOI adopts a convention that names starting with '!' (shriek) name
objects that are not intended for access by the coder. These objects will not be standardized. Additionally
some objects without shriek names are not candidates for standardising, eg .SYSTEM, .KERNEL.

176
A model is needed for whether changes made to methods are seen by objects created before the
changes. Changes that are seen are preferable where a long-lived object is being brought up-to-date.
Changes that only apply to future objects are preferable if avoiding failure of what "used to work" is the
priority. In view of OOM experience the standard should allow both, on a method by method choice. (eg
perhaps a bug fix applied retrospectively but not an enhancement.)
Multiple inheritance. Study of the 'method search' algorithm, see later, shows that this is an "add-on" that
could readily be retained or omitted. That argues in principle for retention, since the non-user of multiple
inheritance would not suffer from it. On the other hand it adds complexity and can be misused even in
the conservative form that OOI has it.
Signature-based method search. This is not in OOI but is in languages such as Java.
Subclassing of imported classes. It is our intention to say that imported classes can be used in all the
same ways as builtin classes. Because this may be impractical to implement with some external classes,
a conforming language processor will have a list (which may be empty) of external classes it supports.
(And hence nothing of the current SOM interface will be part of the standard.)
Persistent objects. It is our belief that support for very-long-running programs is required. It is a moot
point whether the .ENVIRONMENT directory is enough.
If persistent objects are to converted to a form which is platform independent, ("pickling"), there are
difficulties in deciding what pointers should be followed and further objects included, as opposed to
objects being assumed available on all platforms. This topic is defered.
Locking across a set of objects. In OOI this can only be done by locking the events serially, which has
more risk of deadlock than locking them simultaneously. The decision was made not to add
simultaneous locking.
Critical sections. The GUARD mechanism can be used in a 'critical section' style. Nothing will be added
to the definition.
Old objects seeing new changed methods. When bugs in long running programs are fixed, there can be
a benefit if old objects see the corrected methods. It seems practical to offer a variation of DEFINE for
this - see method lookup discussion.
The committee does not find the current OOI approach to merging 'classic' stems with OO stems
satisfactory. It invalidates some existing programs. (A warning about this was put in A8.3.3 of X3.274.) It
produces surprises for OO programmers, eg a==b after a=.stem~new; b=.stem~new. The proposed
alternative is to make the presence/absence of a dot at the end of the name determine whether coercion
to string is done. The 'classic' meaning of A.=B. would be restored but AA=BB, AA==BB etc. would have
their OO meanings. The meaning of USE ARG with a dotted name would be defined to allow 'by
reference' passing of a stem. Square brackets could be used with both dotted and undotted names. A
further proposal is to note that this leaves few differences between the DIRECTORY class and the
non-dotted STEM class so that it might be a further improvement if the DIRECTORY class was extended
to the extent that the STEM class was unnecessary.
There is a potential problem which the committee has not fully analysed in the OOI treatment of SAY and
streams. OOI has made features (of the STREAM bif) that were configuration determined in X3-274 into
OO language methods, and has made SAY a method (undocumented?). Full analysis may show that
more of I/O could (& should?) be made standard or may show that some OOI I/O language should not be
standardized.
The committee discussed what parts of the OOI implementation were suitable to be defined in a
standard. Potentially, all the builtin classes and objects (which are reachable from .ENVIRONMENT)
might be standardized. However, names which start with an exclamation mark denote unsuitable things.
The committee also thought the following unsuitable:
- Anything specific to SOM. - RX_QUEUE - Stream_Supplier - Parts of .LOCAL other than direct
reference to the default streams. There is a naming problem with this. The names in OOI are STDIN,
STDOUT and STDERR. We would prefer INPUT, OUTPUT, and ERROR to be consistent with the
keywords. OOI has used those names for something else. We will work on the proposal that we use
the prefered names and the MONITOR class is dropped. (Users who want the monitor function can
get it with a few lines of directive.)
The committee feels that OOI over-specifies the index of an item in a LIST. In OOI it is a count giving the
sequence over time of the insertions in the list. The risk in using numbers is that they may be (wrongly)
used as positions, and arithmetic done on them. It is proposed that the index of a list item be of class
OBJECT rather than of class STRING.

177
In OOI, the .ENVIRONMENT is global, not read-only, and contains builtin objects such as .TRUE and
.FALSE. The committee regards this as too risky - suppose that .TRUE was accidentally or maliciously
revalued as 0!
It seems sufficient to add read-only as a characteristic of directories. (This characteristic at the element
level might be expensive to implement.)
Reserved symbols (X3-274 clause 6.2.3.1) also provide a mechanism for preventing the override of
builtin names.
It won't be possible for a standard to exactly define in a system-independent way the scopes/lifetimes of
.ENVIRONMENT and .LOCAL but (as with OOI) the .LOCAL will relate to "One API_START" and
.ENVIRONMENT will have a wider scope. (Power on to power off of some system?).
The proposed "search order" is:
1. Things provided by the system which no user is expected to want to override. Perhaps .TRUE .FALSE
.NIL.
2. The .LOCAL read/write directory, initialized with the default streams, changable by the user for
individual program executions. Perhaps METHODS here.
3. The read-only part of the environment, that is the builtin classes and objects. Also .SYSTEM perhaps.
4. The read/write .ENVIRONMENT directory. Changable by programmers co-operating at the system
level.
Final placement of all builtins needs discussion, but the read-only true&false requirement will be met.
Note that the algorithm of method lookup does not change if "old objects see newest methods" is desired.
What changes is whether the method tables are updated in place or copied-and-updated when they are
changed.
1. There have been sugestions to allow the REQUIRES directive appear in more places. The committee
agrees with this and proposes:
A) All REQUIRES directives must appear together in the file. B) These directives may appear anywhere
the OOI implementation currently allows them to appear.
2. Message numbers and prose are now allocated to messages detected by the syntax, additional to the
messages known to the first standard. Most messages simply involve new minor codes sequential
beyond those defined in the first standard.
3. Proposed language, eg FORWARD, METHOD, and CLASS clauses, allow for many options which can
appear in any order. These can be written in the BNF (in the manner that TO BY FOR were handled in
the first standard) but it is neater to extend the BNF metalanguage.
4. The OOI syntax used in the FORWARD instruction has examples of the 'argument' construct, which is
either a symbol-or-string taken as a constant or is an expression in parentheses. The committee will
define 'term' to be allowed in such places. This is a change to the OOI for valid programs only in the case
where a MESSAGE option used a symbol intending it to be 'taken as a constant'. (As opposed to taken
as a variable with the value defaulting to its name when uninitialized.)
5. In a similar vein to 4 above, some other positions where the "variable reference" notation is used (or
proposed) will be changed. It would be nice to allow "term" in all these places but ambiguity consideration
means some will be "sub-expression", ie parenthesed expression, notation.
6. The colon used for superclass specification will allow symbol-or-string to follow.
DATA:
7. The model of data used in defining the first standard needs changing for OO, to:
- Variable pools are objects, objects are variable pools.
- Variable pool contents are references to objects, not values of strings.
- Pools are not numbered, they are referenced.
- The state variables (those with names beginning '#' used to define processing in the standard) are
present in all pools, as opposed to being in a separate pool.
This data model gives a natural interpretation to the variable pool API applied to local pools. (Local pools
may access non-local pool items by reason of EXPOSE.)
In principle this leads to different threads of execution (resulting from REPLY) being able to execute the
API. (In practice OOI has a restriction to executing the API only on the 'main' thread and the committee
needs to know if this is due to a generally applicable difficulty.)
The committee considered the relevance of IBM's "Object Rexx Programming Guide" G25H-7597-1 to the
Configuration section of the standard. The material there in Appendix A under headings External

178
Function Interface, System Exit Interface, and Variable Pool Interface was deemed material for inclusion,
and the rest not. This is similar to the first standard, although there will be an extra trap, for method calls.
The committee considered the relevance of the STREAM section of IBM's "Object Rexx Reference",
G25H-7598-0. That stream class brings into the language more I/O than the original Rexx, eg an explicit
CLOSE. The new standard will partially follow this trend also.
PEEK on queue unnecessary - same as AT[1]?

Also need to resolve the issues on Monitor class and on run time inspection.

179
Annex B
(informative)
Method of definition
This annex describes the methods chosen to describe Rexx for this standard.
Definitions
Definitions are given for some terms which are both used in this standard and also may be used
elsewhere. This does not include names of syntax constructions; for example, group, which are
distinguished in this standard by the use of italic font.
Conformance
Note that irrespective of how this standard is written, the obligation on a conforming processor is only to
achieve the defined results, not to follow the algorithms in this standard.
Notation
The notation used to describe functions provided by the configuration is like a Rexx function call but it is
not defined as a Rexx function call since a Rexx function call is described in terms of one of these
configuration functions.
Note that the mechanism of a returned string with a distinguishing first character is part of the notation
used in this standard to explain the functions; implementations may use a different mechanism.
Notation for completion response and conditions
The testing of 'X' and 'S' indicators is made implicit, for brevity. Even when written as a subroutine call,
each use of a configuration routine implies the testing. Thus:

call Config_Time
implies
#Response = Config_Time()
if left(#Response,1) == 'X' then call #Raise 'SYNTAX', 5.1, substr(#Response,2)
if left(#Response,1) == 'S' then call #Raise 'SYNTAX', 48.1, substr(#Response,2)
Source programs and character sets
The characters required by Rexx are identified by name, with a glyph associated so that they can be
printed in this standard. Alternative names are shown as a convenience for the reader.
Notation
Note that nnn is not specifying the syntax of a program; it is specifying the notation used in this standard
for describing syntax.
Lexical level
Productions nnn and nnn contain a recursion of comment. Apart from this recursion, the lexical level is a
finite state automaton.
Syntax level
This syntax shows a null_clause list, which is minimally a semicolon, being required in places where
programmers do not normally write semicolons, for example after 'THEN'. This is because the 'THEN'
implies a semicolon. This approach to the syntax was taken to allow the rule 'semicolons separate
clauses' to define 'clauses'.
The precedence rules for the operators are built into this grammar
Data Model
The following explanation of data in terms of Classic Rexx may be helpful. References to clauses of the
existing standard have 274 as a prefix.
We start with the data model from the first Standard - a number of variable pools. Two mechanisms, the
external access of section 274.5.13 (API_Drop etc) and the internal of 274.7.1 (Var_Drop etc). Pools are
numbered, with pool 0 reserved for reserved names (.MN etc) and pool N-1 being related to pool N as the
caller's pool. The symbols which index the pools are distinquished as tailed or non-tailed. The items in the
pool have attributes 'exposed', 'dropped', and 'implicit'. The values in the pools are string values.
An extra scope is used for 'state variables' used in the definition of the standard. These follow the same
lookup rules in a conceptual and separate pool.
The first change necessary is to define the values in the pools as references. For string values this is just
a change in definition style, since a reference always followed to a string value is semantically identical
with the notion of having the value in the pool. However, references open the possibility of referencing

180
non-strings, which can behave in a changed way while still being refered to by the same reference.
(Mutable objects)
It is reasonable that the definition should have the pools reference one another rather than use numbered
pools. It is difficult to have a notion of numbering the pools when any object can have a set of variables
associated with it.
Assignment is defined as assignment of references. The language could have been designed differently,
for example to make assignment behave like the COPY method, but assignment of references is the
natural, powerful, choice.
If pools are not numbered, the notation of the first standard, where some state variables use the #Level
number as part of their names, will not suffice. An appropriate solution is to say that each variable pool
can have state variables and user program variables in it. Placing the state variables that are
per-procedure-level in the variable pool for their level avoids the need to specify #Level in their tails.
There are pre-existing objects such as all possible values that can be written as literals and the objects
accessed by .SYSTEM etc. Further objects are created by the NEW method.
Editorial note: It looks nice to unify: an object *is* a variable pool and a variable pool *is* an object. There is some
awkwardness describing the classic API_ function as applying to an object. There don't seem to be difficulties in
defining any object behaviour we want in terms of state variables that refer from one object to another.

Evaluation (Definitions written as code)

There is no single definitional mechanism for describing semantics that is predominantly used in
standards describing programming languages, except for the use of prose. The committee has chosen to
define some parts of this standard using algorithms written in Rexx. This has the advantages of being
rigorous and familiar to many of the intended readers of this standard. It has the potential disadvantage of
circularity - a definition based on an assumption that the reader already understands what is being
defined.
Circularity has been avoided by:
- specifying the language incrementally, so that the algorithms for more complex details are specified
in code that uses only more simple Rexx. For example, the notion that an expression evaluates to a
result can be understood by the reader even without a complete specification of all operators and
built-in functions that might be used in the expression;
- specifying the valid syntax of Rexx programs without using Rexx coding. The method used, Backus
Normal Form, can adequately be introduced by prose.
Ultimately, some understanding of programming languages is assumed in the reader (just as the ability to
read prose is assumed) but any remaining circularity in this standard is harmless.
The comparison of two single characters is an example of such a circularity; Config_Compare can
compare two characters but the outcome can only be tested by comparing characters. It has to be
assumed that the reader understands such a comparison.
Some of the definition using code is repeating earlier definition in prose. This duplication is to make the
document easier to understand when read from front to back.
Note that the layout of the code, in the choices of instructions-per-line, indentations etc., is not significant.
(The layout style used follows the examples in the base reference and it is deliberate that the DO and
END of a group are not at the same alignment.)
The code is not intended as an example of good programming practice or style.
The variables in this code cannot be directly referenced by any program, even if the spelling of some
VAR_SYMBOL coincides. These variables, referred to as state variables, are referenced throughout
this document; they are not affected by any execution activity involving scopes. Some of more significant
variables and routines are written with # as their first character. The following list of them is intended as
an aid to understanding the code. The index of this standard shows the main usage, but not all usage, of
these names.
The following are constants set by the configuration, by Config_Constants:
#Configuration is used for PARSE SOURCE.
#Version is used for PARSE VERSION.
#Bif_Digits. represents numeric digits settings, tails are built-in function names.
#Limit_Digits is the maximum significant digits.
#Limit_EnvironmentName is a maximum length.

181
#Limit_ExponentDigits is the maximum digits in an exponent.
#Limit_Literal is a maximum length.
#Limit_MessageInsert is a maximum length.
#Limit_Name is a maximum length.
#Limit_String is a maximum length.
#Limit_TraceData is a maximum length.

These are named outputs of configuration routines:

#Response is used to hold the result from a configuration routine.

#Indicator is used to hold the leftmost character of Response.
#Outcome is the main outcome of a configuration routine.
#RC is set by Config_Command.
#NoSource is set by Config_NoSource.
#Time is set by Config_Time
#Adjust<Index "#Adjust" # "" > is set by Config_Time

These variables are set up with output from configuration routines:

#HowInvoked records from API_Start, for use by PARSE SOURCE.

#Source records from API_Start for use by PARSE SOURCE.
#AllBlanks<Index "#AllBlanks" # "" > is a string including Blank and equivalents.
#ErrorText.MsgNumber is the text as altered by limits.
#SourceLine. is a record of the source, retained unless NoSource is set. #SourceLine.0 is a count of
lines.
#Pool is a reference to the current variable pool.

These are variables not initialized from the configuration:

#Level is a count of invocation depth, starting at one.

#NewLevel equals #Level plus one.
#Pool1 is a reference to the variable pool current when the first instruction was executed.
#Upper is a reference to the variable pool which will be current when the current PROCEDURE ends.
#Loop is a count of loop nesting.
#LineNumber is the line number of the current clause.
#Symbol is a symbol after tails replacement.
#API_Enabled determines when the application programming interface for variable pools is available.
#Test is the Greater/Lesser/Equal result.
#InhibitPauses is a numeric trace control.
#InhibitTrace is a numeric trace control.
#AtPause is on when executing interactive input.
#AllowProcedure provides a check for the label needed before a procedure.
#DatatypeResult is a by-product of DATATYPE().
#Condition is a condition, eg 'SYNTAX'.
#Trace_QueryPrior detects an external request for tracing.
#TraceInstruction detects TRACE as interactive input.

These are variables that are per-Level, that is, have #Level as a tail component:

#IsFunction. indicates a function call.

#IsProcedure. indicates indicates the routine is a procedure.
#Condition. indicates whether the routine is handling a condition.
#ArgExists.#Level.ArgNumber indicates whether an argument exists. (Initialized from API_Start for
Level=1)
#Arg.#Level.ArgNumber provides the value of an argument. (Initialized from API_Start for Level=1)
When ArgNumber=0 this gives a count of the arguments.

182
#Tracing. is the trace setting letter.
#Interactive. indicates when tracing is interactive. ('?' trace setting)
#ClauseLocal. ensures that DATE/TIME are consistent across a clause.
#ClauseTime. is the TIME/DATE frozen for the clause.
#StartTime. is for 'Elapsed' time calculations.
#Digits. is the current numeric digits.
#Form. is the current numeric form.
#Fuzz. is the current numeric fuzz.

These are qualified by #Condition as well as #Level:

#Enabling. is 'ON', 'OFF' or 'DELAYED'.

#Instruction. is 'CALL' or 'SIGNAL'
#TrapName. is the label.
#ConditionDescription. is for CONDITION('D')
#ConditionExtra. is for CONDITION('E')
#ConditionInstruction. is for CONDITION('I')
#PendingNow. indicates a DELAYED condition.
#PendingDescription. is the description of a DELAYED condition.
#PendingExtra. is the extra description for a DELAYED condition.
#EventLevel. is the #Level at which an event was DELAYED.

These are qualified by ACTIVE, ALTERNATE, or TRANSIENT as well as #Level:

#Env_Name. is the environment name.

#Env_Type. is the type of a resource, and is additionally qualified by input/output/error distinction.
#Env_Resource. is the name of a resource, and is additionally qualified by input/output/error distinction.
#Env_Position. is INPUT or APPEND or REPLACE, and is additionally qualified by input/output/error
distinction.

These are variables that are per-loop:

#Identity. is the control variable.

#Repeat. is the repetition count.
#By. is the increment.
#To. is the limit.
#For. is that count.
#Iterate. holds a position in code describing DO instruction semantics.
#Once. holds a position in code describing DO instruction semantics.
#Leave. holds a position in code describing DO instruction semantics.

These are variables that are per-stream:

#Charin_Position.
#Charout_Position.
#Linein_Position.
#Lineout_Position.
#StreamState. records ERROR state for return by STREAM built-in function.

These are commonly used prefixes:

Config_ is used for a function provided by the configuration.

API_ is used for an application programming interface.
Trap_ is used for a routine called from the processor, not provided by it.
Var_ is used for the routines operating on the variable pools.

183
These are notation routines, only available to code in this standard:

#Contains checks whether some construct is in the source.

#Instance returns the content of some construct in the source.
#Evaluate returns the value of some construct in the source.
#Execute causes execution of some construct in the source.
#Parses checks whether a string matches some construct.
#Clause notes some position in the code.
#Goto continues execution at some noted position.
#Retry causes execution to continue at a previous clause.

These are frequently used routines:

#Raise is a routine for condition raising.

#Trace is a routine for trace output.
#TraceSource is a routine to trace the source program.
#CheckArgs processes the arguments to a built-in function.

184
Annex C
(informative)
Bibliography
Cowlishaw, Mike F, The Rexx Language, Prentice Hall, Englewood Cliffs, N.J. 07632; second edition
1990. ISBN 0-13-780651-5
Systems Application Architecture Common Programming Interface Rexx Level 2 Reference, IBM
SC24-5549, second edition 1992
American National Standard ANSI X3.274-1996

My NetRexx is pre-ISBN number.

185
0
#Adjust, 29, 111, 113, 145
#AllBlanks, 58, 96, 98, 145