Bug 132022 – REGRESSION: JSRegress's js/slow-stress/new-spread.html fails sometimes

• Status: RESOLVED
• Resolution: FIXED
• Component: JavaScriptCore

[Tim Horton]

Starting some time yesterday, js/slow-stress/new-spread started failing like so:

-PASS no exception thrown
-PASS successfullyParsed is true
+FAIL caught exception: Result was -1743984020 expected -2371153088
+FAIL successfullyParsed should be true (of type boolean). Was undefined (of type undefined).

(see http://build.webkit.org/results/Apple%20Mavericks%20Release%20WK1%20%28Tests%29/r167675%20%285412%29/results.html for example)

[Tim Horton]

This is a sometimes fail, as you can see here, but just started: http://webkit-test-results.appspot.com/dashboards/flakiness_dashboard.html#showAllRuns=true&tests=js%2Fslow-stress%2Fnew-spread.html

[Radar WebKit Bug Importer]

<rdar://problem/16690671>

[Tim Horton]

Marked as flaky in http://trac.webkit.org/changeset/167680

[Michael Saboff]

Running repeatedly, I can get this to fail ~25% of the time:

$ for i in 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20; do echo "Try #$i"; DYLD_FRAMEWORK_PATH=WebKitBuild/Release WebKitBuild/Release/jsc LayoutTests/js/slow-stress/script-tests/new-spread.js ; done
Try #1
Try #2
Exception: Result was -872245192 expected -2371153088
Try #3
Try #4
Try #5
Exception: Result was 1614215892 expected -2371153088
Try #6
Try #7
Try #8
Try #9
Try #10
Try #11
Try #12
Exception: Result was 1614215892 expected -2371153088
Try #13
Try #14
Try #15
Try #16
Try #17
Exception: Result was -872245192 expected -2371153088
Try #18
Try #19
Try #20
Exception: Result was 1614215892 expected -2371153088

It happens less frequently with a debug build, roughly 10% of the time.

[Tim Horton]

Is this fixed by bug 132079? Can we un-flaky the test?

[Alexey Proskuryakov]

Ping?

[Oliver Hunt]

(In reply to comment #5)
> Is this fixed by bug 132079? Can we un-flaky the test?

yes1

[Tim Horton]

(In reply to comment #7)
> (In reply to comment #5)
> > Is this fixed by bug 132079? Can we un-flaky the test?
> 
> yes1

un-flaky'd in http://trac.webkit.org/changeset/167812