Scribd
Upload
197 views

Introduction To PDF Programming: Leonard Rosenthol Lazerware

This document provides an overview and introduction to programming with PDF files. It discusses common tasks that can be done with PDF like creation, manipulation, and extraction. It reviews several popular libraries for working with PDF including PDFlib, the Acrobat SDK, and SPDF. It covers the basic structure and objects that make up a PDF document like the catalog, pages, and objects. It also discusses important concepts like error handling and memory management when programming with PDF.

Uploaded by

Bieliūnaitė Karolina
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
197 views

Introduction To PDF Programming: Leonard Rosenthol Lazerware

This document provides an overview and introduction to programming with PDF files. It discusses common tasks that can be done with PDF like creation, manipulation, and extraction. It reviews several popular libraries for working with PDF including PDFlib, the Acrobat SDK, and SPDF. It covers the basic structure and objects that make up a PDF document like the catalog, pages, and objects. It also discusses important concepts like error handling and memory management when programming with PDF.

Uploaded by

Bieliūnaitė Karolina
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 33

Introduction to PDF

Programming
Leonard Rosenthol
Lazerware

Overview
What might you want to do with PDF?
Review of available libraries
Review of the PDF file format
Developing with the Acrobat API
Developing with PDFlib

You are here because


Youre a programmer looking to expand in
doing stuff with PDF.
Youre already programming PDF using some
library and wanted to hear about other
libraries.
There wasnt anything else interesting to do.
Youre a friend of mine and wanted to heckle

How I do things
You should all have copies of the presentation
that you received when you walked in.
There is also an electronic copy of this
presentation (PDF format, of course!) on my
website at http://www.lazerware.com/
Ive left time at the end for Q&A, but please
feel free to ask questions at any time!

What to do with PDF?

Creation

Report generation
Content repurposing
Document Conversion

Manipulation

Adding text or images


Form filling
Append or removing pages
Imposition
Adding structural elements
Bookmarks, hyperlinks, etc.

Securing and signing

What else can you do?

Imaging
Printing
Rasterization (conversion to bitmap)

Content extraction/conversion
Text, HTML, XML
Postscript

Review of Libraries

Creation Only

PDFlib
ClibPDF (FastIO)
Panda (StillHQ)
PDF File Creator (FyTek)
PDF in a Box (Synactis)
PDFever (Perl Script Studio)
SanFace PDFLibrary (SanFace)
ReportLab

Libraries (cont)

Creation Only

retepPDF (Peter Mount)


Root River Delta (Root River Systems)
The Big Faceless PDF Library (Big Faceless)
iText (Lowagie)

Creation & Manipulation

PDFLibrary (Glance)
Life*JOVE (Corena)
PJ (Etymon)
activePDF Toolkit (ActivePDF)

Libraries (cont)

Imaging
5D PDFLibrary (Global Graphics)
Ghostscript (Artifex)

Everything
Acrobat SDK
Adobe PDFLibrary
DocuCom PDF Core Library (Zeon)
SPDF (Appligent)

Whats in a PDF?

Peeling the layers of PDF

PDF file

PDF document (aka page description)

physical container in a file system containing


the PDF document and other data
Contains one or more pages, where each page
consists of text, graphics and/or images as
well as hyperlinks, sounds, etc.

other data

PDF version, object catalog, etc.

PDF Document Layout

Header

Body

Sequence of objects

XREF

Specifies PDF version

Where to find each object

Trailer

Tells where to find XREF

Structure of a PDF
document
Imagable
Content

Thumbnail

Outline entry
1

...

Outline entry
n

Thread 1

Bead 1

...

Page 1

Pages tree

Annotations

...

Page n

Outline tree

Catalog

Article
threads

...

Thread n

Named
destinations

AcroForm

Bead n

Smallest PDF
%PDF-1.1

xref

1 0 obj

0 5

<<

0000000000 65535 f

/Pages 3 0 R

0000000015 00000 n

/Type /Catalog

0000000085 00000 n

>>

0000000136 00000 n

endobj

0000000227 00000 n

2 0 obj

trailer

<<

<<

/Type /Page

/Size 5

/Parent 3 0 R

/Root 1 0 R

>>

/ID[<5181383ede94727bcb32ac27ded71c68><5181383ede94727bcb32ac27ded71c68>]

endobj
3 0 obj

>>

<<

startxref

/Kids [ 2 0 R ]

277

/Count 1

%%EOF

/Type /Pages
/MediaBox [ 0 0 612 792 ]
>>
endobj

A look at the SDK

Where to find the SDK?

Acrobat Plugins

Adobe PDFLibrary

Mac OS, Windows, Linux x86, Solaris

SPDF (Appligent)

Mac OS & Windows

Mac OS, Windows, Linux (x86 & PPC), Solaris,


AIX, HP/UX, Digital Unix, IBM System 390

DocuCom PDF Core (Zeon)??

Windows

Whats in there?

Not every implementation of the SDK


has 100% of the same features (even
between Acrobat and PDFLibrary).

Access to everything in a PDF file

Content extraction
PDF rendering

Read, Add, Modify

to bitmap or platform window

Printing

Everything is an object

CosObj

PDDoc

PDPage, PDBookmark, PDAnnot

AVDoc

CosString, CosInteger, CosArray, CosDict

AVWindow, AVPageView, AVTool

PDEObject

PDEText, PDEImage, PDEPath

PDF Objects

Acrobat treats the objects as opaque,


while SPDF lets you view their contents in
the debugger (incl. objectID!)
All objects are NOT
created equal!

PDDoc != AVDoc != CosObj

Although Acrobat allows you to use them


interchangeably, SPDF does not and in
fact will generate compile time errors

PDDoc == CPDDoc, CosObj == CCosObj


But there are API calls to go between them
PDDocGetCosObj()

ASAtoms

Rather than working with literal strings all


the time, many SDK calls take ASAtoms.
Think of them as a list of name/values
pairs which are keyed by strings.

improved memory management & ease of use


As such, many developers use a single set of
global ASAtom variables.
SPDF even includes macros for doing this

ASAtomFromString()
ASAtomGetString()
ASAtomExistsForString()

Fun with File Systems

ASFileSys

A base class the represents a way for the


SDK to read & write the data of a PDF file. (a
fancy Stream)
Acrobat provides only file-based ones
SPDF also provides memory, FTP & HTTP

ASPathName

ASFileSysCreatePathName (const ASFileSys


fileSys, ASAtom pathSpecType, const void*
pathSpec, const void* mustBeZero);
ASPathFromPlatformPath(void* platformPath)

Error Handling

DURING/HANDLER/ENDHANDLER

In Acrobat itself, these map to


something akin to setjmp/longjmp
Trying to mix them with C++ exceptions
can be a problem.
You cant nest them!

SPDF actually defines them as try/catch


blocks
ERRORCODE

More on Error Handling

Unfortunately, Acrobat does NOT


always throw. Sometimes you
have to use other methods

foo == NULL, PDxxxIsValid(), etc.

CosNull != NULL

If want a null CosObject, you can call


CosNewNull() to get one. BUT that
should be treated as a valid object and
NOT as NULL.

Error Handling Sample


DURING
theASPathName = ASPathFromPlatformPath( inPDF ) ; // Create the ASPathName
thePDDoc = PDDocOpen( theASPathName, NULL, NULL, true ) ; // Open the PDDoc
if ( thePDDoc == (PDDoc)NULL ) {
fprintf( gOutputFile, "# Unable to open PDF file - %s\n", inPDF ) ;
ASRaise ( ASFileError( fileErrOpenFailed ) ) ;
}
HANDLER
theError = ERRORCODE ;
if ( theASPathName != NULL ) {
ASFileSysReleasePath( NULL, theASPathName ) ;
theASPathName = ( ASPathName )NULL ;
}
ASGetErrorString( theError, theAcrobatMessage, sizeof( theAcrobatMessage ) ) ;
fprintf( stderr, "# Error: %s\n", theAcrobatMessage ) ;
return ;
END_HANDLER

Thread Safety?

Acrobat, nor the Adobe PDFLibrary, are


thread safe! As such, you should not try
to use them in a threaded environment
OR make your own threads outside the
SDK.

There are some exceptions to this rule if you


are VERY careful, but youre playing with fire.

SPDF comes in both thread safe and nonthread safe versions.

If you know you dont need threads, then why


take the performance overhead!

SPDF Memory Tracker


SPDF object usage table:
created

freed

leaked

high water mark

Array

17

17

16

HashTable

HashtableEntriesTable

ASAtom

145

145

124

ASFile

CosArray

CosBoolean

CosDict

CosDoc

CosDocRevision

CosName

23

23

23

CosNull

CosNumber

LZWFilter

FlateFilter

PDBookmark

PDBead

PDDoc

PDPage

PDPath

PDFileSpec

PDFont

Splitter Example (SDK)

PDFlib

Whats in there?

PDF Creation/Generation

Text, images, vectors, bookmarks, links, etc.

Allows importing of pages from other PDFs


as XObjects with accompanying PDI library

Accessible from C/C++, Java, Perl, PHP, etc.


Available as an ActiveX/COM component
Available as platform-neutral C source

Everything is a PDF?

You initialize PDFlib and get back a reference


to an opaque PDF structure.

PDF *p = PDF_new();

Each PDF can have only a single PDF open


at any one time for generation, BUT you can
have as many PDFs around as you want
(eg. One per thread).

Error Handling

Each language binding uses its native error handling


mechanism

For C, you can specify a function to be called

Eg. C++ & Java == exceptions


Provides you with the type/class of error and a string
describing it.
You decide whether a given error is fatal or can be ignore
(more of a warning)

You can also specify globally how you want to deal


with warnings (treat as errors or not).

Hello (PDFlib)

Q&A

You might also like