This is until we fix whatever the breakage is.
Created attachment 262647 [details]
Comment on attachment 262647 [details]
Thanks. Landed in r190692: <http://trac.webkit.org/r190692>.
Need to update expected test results to match this change.
Created attachment 262654 [details]
follow up patch.
Thanks. Follow up patch landed in r190695: <http://trac.webkit.org/r190695>.
Created attachment 262662 [details]
follow up 2: need to suppress JSC tail call tests
This should be the last follow up (based on all the tests in http://trac.webkit.org/changeset/190289). This follow up 2 patch has been rubber stamped by Saam. I will land it as soon as my test run confirms that I've addressed the remaining issue.
JSC tests are now passing. Follow up 2 landed in r190699: <http://trac.webkit.org/r190699>.
Looks like follow up 2 changed http/tests/misc/large-js-program.php from crash to timeout on Windows Release. Could you please add an appropriate expectation or skip?
Also, tail call tests apparently need to be disabled: https://build.webkit.org/builders/Apple%20Win%207%20Debug%20%28Tests%29/builds/67893/steps/jscore-test/logs/stdio
(In reply to comment #10)
> Also, tail call tests apparently need to be disabled:
They are already. Build 67893 was built with r190692. The tests were skipped in r190699.
(In reply to comment #9)
> Looks like follow up 2 changed http/tests/misc/large-js-program.php from
> crash to timeout on Windows Release. Could you please add an appropriate
> expectation or skip?
The flakiness dashboard says that the blame list for that result change is https://trac.webkit.org/log/?verbose=on&rev=190700&stop_rev=190698. Of those revisions, r190699 is for follow up 2. But r190699 only changes the JSC tests. It should have no effect on the layout tests.
On the other hand, r190698 is about changing the testing infrastructure for Windows. It is more likely that r190698 triggered the change in test result.
Before they were disabled, I was experiencing crashes due to tail call optimizations on ARM Thumb and Linux. Filed bug #150083 with some information.