Summary: | [Qt]REGRESSION(r120107): It made http/tests/xmlhttprequest/origin-exact-matching.html fails | ||||||
---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | Csaba Osztrogonác <ossy> | ||||
Component: | Tools / Tests | Assignee: | Pablo Flouret <pf> | ||||
Status: | RESOLVED FIXED | ||||||
Severity: | Normal | CC: | darin, ossy, pf | ||||
Priority: | P2 | Keywords: | Qt, QtTriaged | ||||
Version: | 528+ (Nightly build) | ||||||
Hardware: | All | ||||||
OS: | All | ||||||
Bug Depends on: | |||||||
Bug Blocks: | 79666, 88896 | ||||||
Attachments: |
|
Description
Csaba Osztrogonác
2012-06-12 13:57:07 PDT
I skipped it on Qt to paint the bot green. (r120126) Please unskip it with the proper fix. The only suspect i can find is this line: $origin = isset($_GET['origin']) ? $_GET['origin'] : $_SERVER['HTTP_ORIGIN']; The tests that fail are the ones where the origin starts with a null character. Maybe that's fooling isset() in that platform? *preventively curses php* I can make a patch to use the origin directly without checking and see if that's the problem. I don't have the ability to build a linux64 qt, is there a way to test if it would work or is landing a patch and looking at what happens the only way to move forward? The bug is still valid now. Created attachment 167849 [details]
Patch
Would you mind trying this patch somewhere and checking if the test still fails?
Comment on attachment 167849 [details]
Patch
It works for me now, r=me. I'll land it manually with unskipping this test.
|