the refactoring out of the Config object had a typo that caused NRWT to stop respecting set-webkit-configuration again :(
Created attachment 73571 [details] Patch
Created attachment 73665 [details] Patch
Comment on attachment 73665 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=73665&action=review > WebKitTools/Scripts/webkitpy/layout_tests/port/config_standalone.py:56 > + else: > + e = executive.Executive() > + fs = filesystem.FileSystem() Is this really needed? It's never used.
(In reply to comment #3) > (From update of attachment 73665 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=73665&action=review > > > WebKitTools/Scripts/webkitpy/layout_tests/port/config_standalone.py:56 > > + else: > > + e = executive.Executive() > > + fs = filesystem.FileSystem() > > Is this really needed? It's never used. Yeah, I've used it by hand to verify that the actual codepath works, and I'll probably use it in the future when we have "integration tests".
Comment on attachment 73665 [details] Patch Sigh. We really need to kill that static. Thanks for fixing.
Comment on attachment 73665 [details] Patch Clearing flags on attachment: 73665 Committed r71858: <http://trac.webkit.org/changeset/71858>
All reviewed patches have been landed. Closing bug.
This broke the NaCl tests, which use new-run-webkit-httpd, which also tries to get the configuration (unclear if it needs it): http://build.chromium.org/p/chromium/builders/NACL%20Tests/builds/1636/steps/nacl_ui_tests/logs/stdio The breakage is preventing WebKit rolls, so I'm going to roll this out for now.
(In reply to comment #8) > The breakage is preventing WebKit rolls, so I'm going to roll this out for now. Rolled out with http://trac.webkit.org/changeset/71916.
Created attachment 73801 [details] Patch
Comment on attachment 73801 [details] Patch Clearing flags on attachment: 73801 Committed r71960: <http://trac.webkit.org/changeset/71960>