Bug 8951 - REGRESSION: AtomicString hash corrupted by high-bit Latin-1
Summary: REGRESSION: AtomicString hash corrupted by high-bit Latin-1
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: Platform (show other bugs)
Version: 420+
Hardware: Mac OS X 10.4
: P2 Normal
Assignee: Nobody
URL:
Keywords: Regression
Depends on:
Blocks:
 
Reported: 2006-05-16 23:12 PDT by mitz
Modified: 2006-05-16 23:46 PDT (History)
0 users

See Also:

Attachments
Patch (3.32 KB, patch)
2006-05-16 23:15 PDT, mitz 		mjs: review+
Details | Formatted Diff | Diff
View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description mitz 2006-05-16 23:12:52 PDT 
CStringTranslator::equal() is broken for high-bit C strings, which leads to duplicate entries in the AtomicString hash map.
Comment 1 mitz 2006-05-16 23:15:37 PDT 
Created attachment 8361 [details]
Patch

The test is for the fix in StringImpl.cpp. The hash map corruption can't be tested reliably in a normal debug build. I don't know if the corruption happens in shipping WebKit, but the test succeeds, so at least part of this bug is a regression.
Comment 2 Maciej Stachowiak 2006-05-16 23:31:46 PDT 
Comment on attachment 8361 [details]
Patch

r=me
Comment 3 Darin Adler 2006-05-16 23:42:16 PDT 
I broke this when I converted from QChar to UChar!
Comment 4 Darin Adler 2006-05-16 23:46:39 PDT 
Committed revision 14434.