Bug 207269 – REGRESSION: ASSERTION FAILED: resetResult == 0 on regression tests in ITP code, resetting SQLite statement

• Status: RESOLVED
• Resolution: FIXED
• Component: WebCore Misc.

[Jason Lawrence]

http/tests/websocket/tests/hybi/workers/close.html

Description:
This test is flaky crashing on Mac debug wk2 and once on Catalina gpuprocess Debug. 

History:
https://results.webkit.org/?limit=50000&suite=layout-tests&test=http%2Ftests%2Fwebsocket%2Ftests%2Fhybi%2Fserver-close.html&style=debug&platform=mac

Crash log:
https://build.webkit.org/results/Apple%20Mojave%20Debug%20WK2%20(Tests)/r254447%20(6866)/http/tests/websocket/tests/hybi/server-close-crash-log.txt

[Jason Lawrence]

Created attachment 389808 [details]
Crash log

[Radar WebKit Bug Importer]

<rdar://problem/59188891>

[Jason Lawrence]

Created attachment 389811 [details]
Update Test Expectations

[Jason Lawrence]

Possible regression at r254421.

[Truitt Savell]

Comment on attachment 389811 [details]
Update Test Expectations

Clearing flags on attachment: 389811

Committed r255816: <https://trac.webkit.org/changeset/255816>

[Truitt Savell]

To confirm, this bug is tracking the crashes for http/tests/websocket/tests/hybi/server-close.html

[Jason Lawrence]

Created attachment 389841 [details]
Update Test Expectations

[Truitt Savell]

Comment on attachment 389841 [details]
Update Test Expectations

Clearing flags on attachment: 389841

Committed r255843: <https://trac.webkit.org/changeset/255843>

[Alexey Proskuryakov]

We should track all such crashes here. It seems very unlikely that this specific test is to blame.

Thread 13 Crashed:: Dispatch queue: WebResourceLoadStatisticsStore Process Data Queue
0   com.apple.JavaScriptCore      	0x000000012a3a8b00 WTFCrash + 16 (Assertions.cpp:305)
1   com.apple.WebKit              	0x00000001091d017b WTFCrashWithInfo(int, char const*, char const*, int) + 27
2   com.apple.WebKit              	0x00000001097b94c8 WebKit::resetStatement(WebCore::SQLiteStatement&) + 104 (ResourceLoadStatisticsDatabaseStore.cpp:326)
3   com.apple.WebKit              	0x00000001097be503 WebKit::ResourceLoadStatisticsDatabaseStore::getNumberOfPrevalentResources() const + 131 (ResourceLoadStatisticsDatabaseStore.cpp:957)
4   com.apple.WebKit              	0x00000001097bec60 WebKit::ResourceLoadStatisticsDatabaseStore::calculateTelemetryData(WebKit::PrevalentResourceDatabaseTelemetry&) const + 32 (ResourceLoadStatisticsDatabaseStore.cpp:1092)
5   com.apple.WebKit              	0x00000001097bf8b3 WebKit::ResourceLoadStatisticsDatabaseStore::calculateAndSubmitTelemetry() const + 163 (ResourceLoadStatisticsDatabaseStore.cpp:1118)
6   com.apple.WebKit              	0x00000001097e7de7 WebKit::ResourceLoadStatisticsDatabaseStore::ResourceLoadStatisticsDatabaseStore(WebKit::WebResourceLoadStatisticsStore&, WTF::WorkQueue&, WebKit::ShouldIncludeLocalhost, WTF::String const&, PAL::SessionID)::$_11::operator()() const + 55 (ResourceLoadStatisticsDatabaseStore.cpp:307)
7   com.apple.WebKit              	0x00000001097e7cc9 WTF::Detail::CallableWrapper<WebKit::ResourceLoadStatisticsDatabaseStore::ResourceLoadStatisticsDatabaseStore(WebKit::WebResourceLoadStatisticsStore&, WTF::WorkQueue&, WebKit::ShouldIncludeLocalhost, WTF::String const&, PAL::SessionID)::$_11, void>::call() + 25 (Function.h:52)

[Kate Cheney]

Created attachment 390124 [details]
Patch

[Kate Cheney]

This will still crash. But this adds more specific logging to know what error occurred, and doesn't try to reset the statement, which would cause another error.

[WebKit Commit Bot]

Comment on attachment 390124 [details]
Patch

Clearing flags on attachment: 390124

Committed r256085: <https://trac.webkit.org/changeset/256085>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.