RESOLVED FIXED 7718
run-webkit-tests fast/dom/HTMLObjectElement/ hangs
https://bugs.webkit.org/show_bug.cgi?id=7718
Summary run-webkit-tests fast/dom/HTMLObjectElement/ hangs
Alexey Proskuryakov
Reported 2006-03-11 04:19:15 PST
run-webkit-tests fast/dom/HTMLObjectElement/ This runs two tests, each of those runs fine on its own. Here, object-as-frame.html hangs while waiting for notifyDone(), because an exception happens earlier: ReferenceError: Can't find variable: shouldBe This seems to indicate that onLoad is called before external scripts are loaded.
Attachments
proposed fix (690 bytes, patch)
2006-03-13 09:17 PST, Alexey Proskuryakov
no flags
proposed fix (695 bytes, patch)
2006-03-13 10:50 PST, Alexey Proskuryakov
darin: review+
Darin Adler
Comment 1 2006-03-12 22:02:11 PST
Maybe this is the reason some layout tests intermittently fail on the server. Those also seem to indicate that onload is firing before scripts run, although they are internal scripts, not external.
mitz
Comment 2 2006-03-13 08:18:51 PST
It's all about the trailing slash. This hangs: run-webkit-tests fast/dom/HTMLObjectElement//object-as-frame.html but this doesn't: run-webkit-tests fast/dom/HTMLObjectElement/object-as-frame.html
Alexey Proskuryakov
Comment 3 2006-03-13 09:17:03 PST
Created attachment 7048 [details] proposed fix Mitz has explained it to me that sequences of slashes don't collapse in URLs (unlike they do in POSIX paths), so just patching run-webkit-tests.
Alexey Proskuryakov
Comment 4 2006-03-13 09:43:53 PST
DumpRenderTree uses CFURLCreateWithFileSystemPath() for file paths, and it sounds like this function should canonicalize the paths; <rdar://4475631>.
David Kilzer (:ddkilzer)
Comment 5 2006-03-13 10:33:05 PST
Comment on attachment 7048 [details] proposed fix >- print OUT "$testPath\n"; >+ print OUT File::Spec->canonpath("$testPath\n"); I don't think the "\n" should be included in the method call since it's not part of the path! Something more like this: >+ print OUT File::Spec->canonpath($testPath), "\n";
Alexey Proskuryakov
Comment 6 2006-03-13 10:50:30 PST
Created attachment 7049 [details] proposed fix I don't think it changes the behavior in any way, but if it reads better, why not :)
Darin Adler
Comment 7 2006-03-13 11:29:20 PST
Comment on attachment 7049 [details] proposed fix r=me
Note You need to log in before you can comment on or make changes to this bug.