1708113 - SpiderMonkey job exits with 0 status even if there's failure in some case

Status: RESOLVED FIXED

(Core :: JavaScript Engine, defect, P1)

Description

[Tooru Fujisawa [:arai]]

https://treeherder.mozilla.org/jobs?repo=try&group_state=expanded&revision=41d4245d7d9a05f6f96f334d5aefdfebcf5aba19&selectedTaskRun=Ce90P6ajR72-QQlwnDmSyA.0

This try job fails with many testcase hitting TEST-UNEXPECTED-FAIL, but its shown as green.

[task 2021-04-27T22:05:59.259Z] exit status 0 for '(make-nonempty)'
[task 2021-04-27T22:05:59.259Z] exit status 0 for 'make check'
[task 2021-04-27T22:05:59.259Z] exit status 2 for 'make check-jit-test'
[task 2021-04-27T22:05:59.259Z] exit status 2147483651 for 'jsapi-tests'
[task 2021-04-27T22:05:59.259Z] exit status 2 for 'jstests'
...
[task 2021-04-27T22:05:59.767Z] + exit 0
...
[taskcluster 2021-04-27T22:05:59.790Z]    Exit Code: 0
...
[taskcluster 2021-04-27T22:05:59.790Z]       Result: SUCCEEDED

it sounds like the exit code is somehow overridden, or overlooked

Comment 1

[Steve Fink [:sfink] [:s:]]

Whoa, it even prints out that it has a nonzero exit status. I must have broken this very recently.

Comment 2

[Steve Fink [:sfink] [:s:]]

My guess is that my fix for pretty much the same problem must be broken:

sys.exit(max((st for _, st in results), key=abs))

The 2147483651 exit status must be throwing it off.

It seems like that ought to work when I tried it in a local python, but maybe there's something weird going on on Windows. I'll look tomorrow. I am not arai, so I sleep sometimes.

Comment 3

[Steve Fink [:sfink] [:s:]]

Attachment: Bug 1708113 - sys.exit(2147483651L) on Windows exits with status 0

Comment 4

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/7b854d162a85
sys.exit(2147483651L) on Windows exits with status 0 r=arai

Comment 5

[Cristian Brindusan [:cbrindusan]]

https://hg.mozilla.org/mozilla-central/rev/7b854d162a85