Bug 63358 – nrwt: remove --use-apache from the command line

• Status: RESOLVED
• Resolution: FIXED
• Component: New Bugs

[Dirk Pranke]

nrwt: remove --use-apache from the command line

[Dirk Pranke]

Created attachment 98549 [details]
Patch

[Adam Barth]

Comment on attachment 98549 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=98549&action=review

> Tools/Scripts/webkitpy/layout_tests/port/base.py:848
> +        """Override if your port uses something other than Apache for serving
> +        web pages."""

I'd skip this docstring

> Tools/Scripts/webkitpy/layout_tests/port/chromium_win.py:58
> +    USES_APACHE = False

This appears to be unused.

[Dirk Pranke]

Committed r90503: <http://trac.webkit.org/changeset/90503>

[Mihai Parparita]

Reverted r90503 for reason:

Makes Chromium Windows canary bots hang

Committed r90514: <http://trac.webkit.org/changeset/90514>

[Dirk Pranke]

Created attachment 99914 [details]
Patch

[Dirk Pranke]

Comment on attachment 99914 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=99914&action=review

> Tools/Scripts/webkitpy/layout_tests/port/base.py:234
> +            return self._executive.run_command([path, "-v"], env=env, return_exit_code=True) == 0

This change fixes the hang on windows. Note that I had to add the env keyword param to executive.run_command() for this to work.

[Dirk Pranke]

Comment on attachment 99914 [details]
Patch

no unit tests added for executive.run_command, although there probably should be one. I didn't add a test for this chromium_win either, because it would involve a bunch of mocking and I was feeling lazy.

[Adam Barth]

Comment on attachment 99914 [details]
Patch

Ok.

[WebKit Review Bot]

Comment on attachment 99914 [details]
Patch

Clearing flags on attachment: 99914

Committed r90527: <http://trac.webkit.org/changeset/90527>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.

[James Robinson]

A bunch of http tests are still timing out on the Win7 canary:

http://test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=http%2Ftests%2Fappcache%2Fsimple.html%2Chttp%2Ftests%2Fappcache%2Ftop-frame-1.html%2Chttp%2Ftests%2Fappcache%2Ftop-frame-3.html%2Chttp%2Ftests%2Fappcache%2Ftop-frame-4.html%2Chttp%2Ftests%2Fappcache%2Fvideo.html%2Chttp%2Ftests%2Fappcache%2Fwhitelist-wildcard.html%2Chttp%2Ftests%2Fappcache%2Fxhr-foreign-resource.html%2Chttp%2Ftests%2Fcache%2Fhistory-only-cached-subresource-loads-max-age-https.html%2Chttp%2Ftests%2Fcache%2Fhistory-only-cached-subresource-loads.html%2Chttp%2Ftests%2Fcache%2Fpost-redirect-get.php%2Chttp%2Ftests%2Fcache%2Fpost-with-cached-subresources.php%2Chttp%2Ftests%2Fcache%2Fstopped-revalidation.html%2Chttp%2Ftests%2Fcache%2Fsubresource-expiration-1.html%2Chttp%2Ftests%2Fcache%2Fsubresource-expiration-2.html%2Chttp%2Ftests%2Fcache%2Fsubresource-fragment-identifier.html%2Chttp%2Ftests%2Fcanvas%2Fphilip%2Ftests%2Fsecurity.drawImage.canvas.html%2Chttp%2Ftests%2Fcanvas%2Fphilip%2Ftests%2Fsecurity.drawImage.image.html%2Chttp%2Ftests%2Fcanvas%2Fphilip%2Ftests%2Fsecurity.pattern.canvas.fillStyle.html%2Chttp%2Ftests%2Fcanvas%2Fphilip%2Ftests%2Fsecurity.pattern.canvas.strokeStyle.html%2Chttp%2Ftests%2Fcanvas%2Fphilip%2Ftests%2Fsecurity.pattern.canvas.timing.html%2Cplugins%2Fmouse-click-plugin-clears-selection.html&group=%40ToT%20-%20chromium.org

[James Robinson]

Added = TIMEOUT lines to test_expectations in http://trac.webkit.org/changeset/90718

[James Robinson]

Adding those = TIMEOUT lines seemed to just make a different set of tests time out, so I removed those.  At this point I'm suspecting an issue with the bot (http://build.chromium.org/p/chromium.webkit/builders/Webkit%20Win7) but I'm not able to do anything about it right now.

[Mihai Parparita]

(In reply to comment #13)
> Adding those = TIMEOUT lines seemed to just make a different set of tests time out, so I removed those.

Note that's because of the --exit-after-n-crashes-or-timeouts flag, which meant that we stopped running tests after 20 unexpected timeouts.

[Dirk Pranke]

re-closing this.