[Webkit-unassigned] [Bug 88936] StringImpl::characters can return NULL for an empty string

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=88936





--- Comment #19 from Myles C. Maxfield <mmaxfield at google.com>  2012-06-20 12:40:06 PST ---
I found this by running V8's mjsunit tests against JavaScriptCore. This was uncovered by http://v8.googlecode.com/svn/trunk/test/mjsunit/deep-recursion.js lines 52 and 55.

I suppose there are a couple options here:
1) Re-write the test as a C++ unit test, creating a Collator, calling collate with (null, 0, "a", 1) and making sure it doesn't return "Equal." I could also reverse the arguments and make sure the output is reversed. I'm not sure where this unit test should end up living.

2) Add all of mjsunit to the JavaScriptCore tests inside Source/JavaScriptCore/tests

3) Add just deep-recursion.js to the JavaScriptCore tests inside Source/JavaScriptCore/tests

4) Add a comment saying this is tested in v8's mjsunit ;-)

This is complicated by the fact that it looks like the only tests in Source/JavaScriptCore/tests are regular expression tests (not applicable here), performance tests (also not applicable here) and the mozilla test suite. This test doesn't belong in any of those places. Also, adding existing source might get complicated because of possibly-conflicting licenses.

I'm willing to do whatever the community thinks is best.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.