Bug 281038
| Summary: | [GTK] wxPython fails test suite on i586 with webkitgtk 2.46.0 | ||
|---|---|---|---|
| Product: | WebKit | Reporter: | Dominique Leuenberger <dimstar> |
| Component: | WebKitGTK | Assignee: | Nobody <webkit-unassigned> |
| Status: | RESOLVED DUPLICATE | ||
| Severity: | Normal | CC: | bugs-noreply, mcatanzaro |
| Priority: | P3 | Keywords: | Gtk |
| Version: | WebKit Nightly Build | ||
| Hardware: | PC | ||
| OS: | Linux | ||
Dominique Leuenberger
While preparing the integration of GNOME 47 into openSUSE Tumbleweed, we have noticed that with the update of webkitgtk to 2.46.0 (or 2.46.1), the test suite of wxPython starts failing on i586 (passes on x86_64)
[ 1312s] =========================== short test summary info ============================
[ 1312s] FAILED unittests/test_webview.py::webview_Tests::test_webview1
[ 1312s] FAILED unittests/test_webview.py::webview_Tests::test_webview2
[ 1312s] FAILED unittests/test_webview.py::webview_Tests::test_webview3
[ 1312s] = 3 failed, 1689 passed, 31 skipped, 4 xfailed, 4 warnings in 71.17s (0:01:11) =
Python error: Segmentation fault
[ 1309s]
[ 1309s] Current thread 0xf77c3700 (most recent call first):
[ 1309s] File "/home/abuild/rpmbuild/BUILDROOT/python311-wxPython-4.2.1-9.3.i386/usr/lib/python3.11/site-packages/wx/core.py", line 2252 in MainLoop
[ 1309s] File "/home/abuild/rpmbuild/BUILD/wxPython-4.2.1/unittests/wtc.py", line 34 in tearDown
[ 1309s] File "/usr/lib/python3.11/unittest/case.py", line 584 in _callTearDown
[ 1309s] File "/usr/lib/python3.11/unittest/case.py", line 626 in run
[ 1309s] File "/usr/lib/python3.11/unittest/case.py", line 678 in __call__
Reverting to 2.45.92 gave us a successful result for that test suite
This is just the early findings to have it documented - I'll be trying to
* get some better debug output (stack trace of the segfault if possible)
* bisect 2.45.92..2.46.0 to hopefully identify the failing commit
Unless of course somebody already knows what's wrong based on that little information
| Attachments | ||
|---|---|---|
| Add attachment proposed patch, testcase, etc. |
Dominique Leuenberger
It's not limited to i586, but was also observed on ppc64le and aarch64 (which might make it slightly easier to produce traces)
Michael Catanzaro
If you're lucky, it might be bug #281495, which I'll provide a fix for soon.
Dominique Leuenberger
(In reply to Michael Catanzaro from comment #2)
> If you're lucky, it might be bug #281495, which I'll provide a fix for soon.
the 'works on x86_64 by luck but fails on aarch64 and ppcle64' sounds promising
Michael Catanzaro
(In reply to Dominique Leuenberger from comment #3)
> the 'works on x86_64 by luck but fails on aarch64 and ppcle64' sounds
> promising
Mike just reported bug #281565. Not sure whether that's a coincidence, or whether he's investigating the same problem as you.
Michael Catanzaro
Wait, that bug mentions wxPython as well. Must be the same.
*** This bug has been marked as a duplicate of bug 281495 ***