REOPENED167594
[ios-simulator] API test WebKit2.WKWebProcessPlugInRangeHandle timing out 
https://bugs.webkit.org/show_bug.cgi?id=167594
Summary [ios-simulator] API test WebKit2.WKWebProcessPlugInRangeHandle timing out
Ryan Haddad
Reported 2017-01-30 09:55:29 PST
[ios-simulator] API test WebKit2.WKWebProcessPlugInRangeHandle timing out https://build.webkit.org/builders/Apple%20iOS%2010%20Simulator%20Release%20WK2%20%28Tests%29/builds/3262
Attachments
Patch (1.57 KB, patch)
2017-04-04 23:49 PDT, Andy Estes 		no flags
DetailsFormatted DiffDiff
Show Obsolete (1) View All Add attachment proposed patch, testcase, etc.
Ryan Haddad
Comment 1 2017-01-30 10:02:16 PST
Test was added with https://trac.webkit.org/changeset/211354
mitz
Comment 2 2017-01-30 10:09:18 PST
The test passes for me in the iOS Simulator. Not sure what’s going on.
Ryan Haddad
Comment 3 2017-01-30 12:15:12 PST
(In reply to comment #2) > The test passes for me in the iOS Simulator. Not sure what’s going on. This is making the bots red, so I'd like to roll out if necessary.
mitz
Comment 4 2017-01-30 12:27:15 PST
If possible, I’d just disable the test.
mitz
Comment 5 2017-01-30 12:27:43 PST
Ryan, are you able to reproduce this locally?
Radar WebKit Bug Importer
Comment 6 2017-01-30 15:25:24 PST
<rdar://problem/30274363>
Ryan Haddad
Comment 7 2017-01-31 19:57:57 PST
Disabled test in http://trac.webkit.org/projects/webkit/changeset/211467.
mitz
Comment 8 2017-01-31 20:28:47 PST
On a computer that’s reproducing this failure, I see that the Web Content process is crashing with this backtrace: * frame #0: 0x000000010d113569 CoreFoundation`CFStringGetLength + 9 frame #1: 0x0000000123a611bd DataDetectorsCore`_typesAreEqual + 43 frame #2: 0x0000000123a613e4 DataDetectorsCore`_DDResultTypeIsAnyPhoneKind + 22 frame #3: 0x0000000123a61341 DataDetectorsCore`_DDGetRelevantResultsForUrlificationRange + 28 frame #4: 0x0000000123a61432 DataDetectorsCore`DDResultGetQueryRangeForURLification + 25 frame #5: 0x000000011538ac23 WebCore`WebCore::softLink_DataDetectorsCore_DDResultGetQueryRangeForURLification(result=0x00007fb1ab10acc0) + 35 at DataDetectorsCoreSoftLink.h:49 frame #6: 0x0000000115388411 WebCore`WebCore::DataDetection::detectContentInRange(contextRange=0x00007fff56249be8, types=DataDetectorTypeAll, context=0x00007fb1ab0009c0) + 5569 at DataDetection.mm:536
mitz
Comment 9 2017-01-31 20:49:09 PST
The crash seems to happen because we somehow mangle the value passed to DDResultGetQueryRangeForURLification. It is not the value of result seen in frame #5 above. Perhaps something about the “soft linking” code or the forward declarations used when not using the Apple Internal SDK is causing it.
mitz
Comment 10 2017-04-04 23:24:45 PDT
Looks like Andy found the cause in bug 170496.
Andy Estes
Comment 11 2017-04-04 23:41:45 PDT
*** Bug 170496 has been marked as a duplicate of this bug. ***
Andy Estes
Comment 12 2017-04-04 23:49:08 PDT
Created attachment 306265 [details] Patch
Andy Estes
Comment 13 2017-04-05 00:08:17 PDT
Comment on attachment 306265 [details] Patch Nevermind, this is just gardening. No need for a review.
Andy Estes
Comment 14 2017-04-05 00:40:45 PDT
Committed r214933: <https://trac.webkit.org/r214933>
Ryan Haddad
Comment 15 2017-04-07 11:11:21 PDT
This test is still timing out on the bots: https://build.webkit.org/builders/Apple%20iOS%2010%20Simulator%20Release%20WK2%20(Tests)/builds/501
Ryan Haddad
Comment 16 2017-04-07 13:13:03 PDT
I disabled test again in https://trac.webkit.org/changeset/215112/webkit
Andy Estes
Comment 17 2017-04-14 13:35:14 PDT
Committed r215372: <https://trac.webkit.org/changeset/215372>
Alexey Proskuryakov
Comment 18 2017-04-16 00:19:09 PDT
What was the new fix after 2017-04-07?
Andy Estes
Comment 19 2017-04-17 12:07:38 PDT
(In reply to Alexey Proskuryakov from comment #18) > What was the new fix after 2017-04-07? See bug #161967
Ryan Haddad
Comment 20 2017-04-21 15:48:23 PDT
This API test is still a flaky timeout.
Ryan Haddad
Comment 21 2017-04-21 15:56:17 PDT
Disabled test in http://trac.webkit.org/projects/webkit/changeset/215651
Note You need to log in before you can comment on or make changes to this bug.