Bug 49740 – [NRWT] '--new-baseline' doesn't work

• Status: RESOLVED
• Resolution: FIXED
• Component: Tools / Tests

[Hayato Ito]

Including the '--new-baseline' flag in new-run-webkit-test causes the script to fail.

[Hayato Ito]

Created attachment 74249 [details]
fix-new-baseline

[Shinichiro Hamaji]

Comment on attachment 74249 [details]
fix-new-baseline

Thanks for the fix.

[Hayato Ito]

Comment on attachment 74249 [details]
fix-new-baseline

Clearing flags on attachment: 74249

Committed r72307: <http://trac.webkit.org/changeset/72307>

[Hayato Ito]

All reviewed patches have been landed.  Closing bug.

[Dirk Pranke]

I was wondering why the unit tests didn't catch this ... of course, we disabled them because they wrote to the real file system :( I really need to land the patches in 48072.

[Hayato Ito]

As far as I remember, I ran ./WebKitTools/Scripts/test-webkit-scripts and it passed all tests.
Could you tell me how do you guys run tests about new-run-webkit-tests?

(In reply to comment #5)
> I was wondering why the unit tests didn't catch this ... of course, we disabled them because they wrote to the real file system :( I really need to land the patches in 48072.

[Dirk Pranke]

(In reply to comment #6)
> As far as I remember, I ran ./WebKitTools/Scripts/test-webkit-scripts and it passed all tests.
> Could you tell me how do you guys run tests about new-run-webkit-tests?
> 

That would be sufficient. Like I said, the tests that would've caught at least part of this were disabled (and had bit-rotted accordingly). The new patch will bring them back online.

Typically I run a script I have that just runs the subset of test-webkitpy that includes the tests under layout_tests/*, and if that passes, then I'll usually run test-webkitpy to be sure nothing weird happened.

I have some other patches to test-webkitpy that should make things a little better so that I don't my separate script, and of course the patch above will allow test-webkitpy to complete in ~ 4 seconds, so that'll help, too.