liburlpattern: Add missing check for U8_NEXT return to Tokenizer
The ICU macro U8_NEXT's documentation [0] indicates that it may return a
negative codepoint on failure.
This CL annotates the Tokenizer::Next() and NextAt() methods with
WARN_UNUSED_RESULT, and adds appropriate error-handling to affected call
sites.
Prior to this change, the original input (below) was parsed and
canonicalized to a string that failed the canonicalization idempotency
property check. With this change, the original input is rejected at
position 1 because 0xcd is an illegal UTF8 codepoint.
original :
* * { } ? * ?
2a cd 2a 7b 7d 3f 2a 3f
canonical:
* * * ?
2a cd 2a 2a 3f
[0]: https://source.chromium.org/chromium/chromium/src/+/main:third_party/icu/source/common/unicode/utf8.h;l=352;drc=f90543d272e2e2edc5b3cdf8ead0b5b3eebceef5
Bug: 1263594
Change-Id: Ib1c66cb322f0c48db2ae7e264cdfd5818975ac9a
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/3248041
Commit-Queue: Dan McArdle <[email protected]>
Reviewed-by: Ben Kelly <[email protected]>
Cr-Commit-Position: refs/heads/main@{#935863}
diff --git a/third_party/liburlpattern/parse.h b/third_party/liburlpattern/parse.h
index add6eec..9af64518 100644
--- a/third_party/liburlpattern/parse.h
+++ b/third_party/liburlpattern/parse.h
@@ -29,10 +29,10 @@
typedef std::function<absl::StatusOr<std::string>(absl::string_view)>
EncodeCallback;
-// Parse a pattern string and return the result. The input |pattern| must
-// consist of UTF-8 characters. Currently only group names may actually
-// contain non-ASCII characters, however. Unicode characters in other parts
-// of the pattern will cause an error to be returned. A |callback| must be
+// Parse a pattern string and return the result. The parse will fail if the
+// input |pattern| is not valid UTF-8. Currently only group names may actually
+// contain non-ASCII characters, however. Unicode characters in other parts of
+// the pattern will cause an error to be returned. A |callback| must be
// provided to validate and encode plain text parts of the pattern. An
// |options| value may be provided to override default behavior.
COMPONENT_EXPORT(LIBURLPATTERN)
