Bug 117141
Summary: | [Qt] REGRESSION(r151021): fast/js/custom-constructors.html made next running test crashing | ||
---|---|---|---|
Product: | WebKit | Reporter: | Zoltan Arvai <zarvai> |
Component: | Tools / Tests | Assignee: | Nobody <webkit-unassigned> |
Status: | RESOLVED INVALID | ||
Severity: | Normal | CC: | allan.jensen, andrunko, jturcotte, kadam, pnormand |
Priority: | P2 | Keywords: | Qt |
Version: | 528+ (Nightly build) | ||
Hardware: | Unspecified | ||
OS: | Unspecified | ||
Bug Depends on: | |||
Bug Blocks: | 79668, 115352 |
Zoltan Arvai
It seems r151020 made fast/js/cyclic-proto.html to crash on Qt bots. It not crashing on all Qt bots, but there are crashes on 32 / 64 bit and WK1 / WK2 builds, too.
Crash log in Qt Release WK1 32bit
http://build.webkit.org/results/Qt%20Linux%20Release/r151100%20%2860605%29/fast/js/cyclic-proto-crash-log.txt
Backtrace from Qt WK1 64bit:
https://gist.github.com/azbesthu/5697017
Crash log on Qt Release WK2 32bit:
http://build.webkit.sed.hu/results/x86-32%20Linux%20Qt%20Release%20WebKit2/r151100%20%2835948%29/fast/js/cyclic-proto-crash-log.txt
Backtrace from Qt WK2 64bit:
https://gist.github.com/azbesthu/5697094
Can you check it please?
Attachments | ||
---|---|---|
Add attachment proposed patch, testcase, etc. |
Zoltan Arvai
Skipped on Qt in http://trac.webkit.org/changeset/151105.
Please unskip it with proper fix.
Zoltan Arvai
(In reply to comment #1)
> Skipped on Qt in http://trac.webkit.org/changeset/151105.
> Please unskip it with proper fix.
After the skip now fast/js/cyclic-prototypes.html is crashing.
The last test run befor these is fast/js/custom-constructors.html, maybe that test is responsible for the crashes.
Zoltan Arvai
fast/js/cyclic-proto.html is unskipped and
fast/js/custom-constructors.html is skipped in http://trac.webkit.org/changeset/151171.
It's hard to reproduce the crash on local environment, let see what happens on the bots.
Allan Sandfeld Jensen
(In reply to comment #0)
> Backtrace from Qt WK1 64bit:
> https://gist.github.com/azbesthu/5697017
>
This backtrace would fit r151021 better. Were the crashes reliable on the bots, or did the first one just start at r151020?
Zoltan Arvai
(In reply to comment #4)
> (In reply to comment #0)
> > Backtrace from Qt WK1 64bit:
> > https://gist.github.com/azbesthu/5697017
> >
> This backtrace would fit r151021 better. Were the crashes reliable on the bots, or did the first one just start at r151020?
If I remember right r151020 was the first revision that crashed locally. I had to re-run some times the tests to get a crash.
Somehow the problem has a connection with fast/js/custom-constructors.html, because after skipping that test no more crashes occurred, as I saw.
Zoltan Arvai
I re-run the tests with 151020:
../run-in-xvfb.sh Tools/Scripts/run-webkit-tests --no-new-test-results --debug-rwt-logging --no-show-results --no-sample-on-timeout --results-directory layout-test-results --debug --webkit-test-runner --qt fast/js/custom-constructors.html fast/js/cyclic-proto.html --iteration=100
07:17:44.710 3295 worker/0 fast/js/custom-constructors.html output stderr lines:
07:17:44.710 3295 ERROR: Unabled to create LocalStorage database path
07:17:44.710 3295 /home/azbest/webkit/WebKit/Source/WebKit2/UIProcess/Storage/LocalStorageDatabaseTracker.cpp(146) : WTF::String WebKit::LocalStorageDatabaseTracker::databasePath(const WTF::String&) const
07:17:44.710 3295 ERROR: Icon database already has a path and is already open. We don't currently support changing its path and reopening.
07:17:44.710 3295 /home/azbest/webkit/WebKit/Source/WebKit2/UIProcess/WebIconDatabase.cpp(69) : void WebKit::WebIconDatabase::setDatabasePath(const WTF::String&)
once a popup window appeared:
Python (v2.7) requires to install plugins to play media files of the following type: text/html decoder
Sometimes there is a crash:
custom-constructors-crash-log.txt
crash log for WebKitTestRunner (pid 2866):
STDOUT: <empty>
STDERR: ERROR: Unabled to create LocalStorage database path
STDERR: /home/azbest/webkit/WebKit/Source/WebKit2/UIProcess/Storage/LocalStorageDatabaseTracker.cpp(146) : WTF::String WebKit::LocalStorageDatabaseTracker::databasePath(const WTF::String&) const
STDERR: ASSERTION FAILED: !result
STDERR: /home/azbest/webkit/WebKit/Source/WTF/wtf/ThreadingPthreads.cpp(344) : void WTF::Mutex::lock()
STDERR: 1 0x7fa59b781c51 /home/azbest/webkit/WebKit/WebKitBuild/Debug/lib/libQt5WebKit.so.5(WTFCrash+0x1e) [0x7fa59b781c51]
STDERR: 2 0x7fa59b7bfb1d /home/azbest/webkit/WebKit/WebKitBuild/Debug/lib/libQt5WebKit.so.5(_ZN3WTF5Mutex4lockEv+0x45) [0x7fa59b7bfb1d]
STDERR: 3 0x7fa59ab837f5 /home/azbest/webkit/WebKit/WebKitBuild/Debug/lib/libQt5WebKit.so.5(+0x14b77f5) [0x7fa59ab837f5]
STDERR: 4 0x7fa59ab837d4 /home/azbest/webkit/WebKit/WebKitBuild/Debug/lib/libQt5WebKit.so.5(+0x14b77d4) [0x7fa59ab837d4]
STDERR: 5 0x7fa59b7a60d5 /home/azbest/webkit/WebKit/WebKitBuild/Debug/lib/libQt5WebKit.so.5(+0x20da0d5) [0x7fa59b7a60d5]
STDERR: 6 0x7fa59b7bf448 /home/azbest/webkit/WebKit/WebKitBuild/Debug/lib/libQt5WebKit.so.5(+0x20f3448) [0x7fa59b7bf448]
STDERR: 7 0x7fa597580764 /usr/lib/nvidia-current/libGL.so.1(+0xaa764) [0x7fa597580764]
Zoltan Arvai
(In reply to comment #4)
> (In reply to comment #0)
> > Backtrace from Qt WK1 64bit:
> > https://gist.github.com/azbesthu/5697017
> >
> This backtrace would fit r151021 better. Were the crashes reliable on the bots, or did the first one just start at r151020?
It seems you're right. I checked again r151021 and I got the crashes with WK2:
../run-in-xvfb.sh Tools/Scripts/run-webkit-tests --no-new-test-results --debug-rwt-logging --no-show-results --no-sample-on-timeout --results-directory layout-test-results --debug --webkit-test-runner --qt fast/js/custom-constructors.html fast/js/cyclic-proto.html
Regressions: Unexpected crashes (2)
fast/js/custom-constructors.html [ Crash ]
fast/js/cyclic-proto.html [ Crash ]
fast/js/custom-constructors.html is not always crashing, but that test causes to crash the next after it.
Also I got sometimes
*** glibc detected *** /home/azbest/webkit/WebKit/WebKitBuild/Debug/bin/WebKitTestRunner: malloc(): memory corruption (fast): 0x00000000021877e4 ***
Zoltan Arvai
Correcting summary, bock and cc list information after the last updates on this bug.
Jocelyn Turcotte
=== Bulk closing of Qt bugs ===
If you believe that this bug report is still relevant for a non-Qt port of webkit.org, please re-open it and remove [Qt] from the summary.
If you believe that this is still an important QtWebKit bug, please fill a new report at https://bugreports.qt-project.org and add a link to this issue. See http://qt-project.org/wiki/ReportingBugsInQt for additional guidelines.