In order to run the layout tests both with and without accelerated 2D canvas, we should have a flag that controls its behaviour at runtime.
Created attachment 68284 [details] Patch
Comment on attachment 68284 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=68284&action=review Looks good to me other than the nits noted below. I have a patch in progress that adds essentially the same line to port/chromium.py plus handles the Chromium DRT case and the new baseline search path we'll want. I should have it ready inside an hour or so (I would've finished it last night but I got derailed by another bug), so whichever of the two lands first can be coordinated with the other one. > WebKitTools/Scripts/webkitpy/layout_tests/run_webkit_tests.py:577 > + test_args.enable_accelerated_2d_canvas = self._options.enable_accelerated_2d_canvas I don't think you need this. > WebKitTools/Scripts/webkitpy/layout_tests/run_webkit_tests.py:1663 > + dest="enable_accelerated_2d_canvas", I don't think you need this, either. > WebKitTools/Scripts/webkitpy/layout_tests/port/chromium.py:360 > + Here you need to check if enable-accelerated-2d-canvas is not actually set (since there's no default value): if hasattr(self._options, 'enable_accelerated_2d_canvas') and self._options.enable_accelerated_2d_canvas): ...
See also bug 46208 for my patch to new-run-webkit-tests for the full support of GPU testing.
Comment on attachment 68284 [details] Patch It looks like Dirk has some good points here. Other than those the patch looks fine.
(In reply to comment #3) > See also bug 46208 for my patch to new-run-webkit-tests for the full support of GPU testing. Wrong bug ID?
(In reply to comment #5) > (In reply to comment #3) > > See also bug 46208 for my patch to new-run-webkit-tests for the full support of GPU testing. > > Wrong bug ID? Looks like it should have been bug 46225.
(In reply to comment #6) > (In reply to comment #5) > > (In reply to comment #3) > > > See also bug 46208 for my patch to new-run-webkit-tests for the full support of GPU testing. > > > > Wrong bug ID? > > Looks like it should have been bug 46225. Right. Sorry!
(In reply to comment #2) > (From update of attachment 68284 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=68284&action=review > > Looks good to me other than the nits noted below. > > I have a patch in progress that adds essentially the same line to port/chromium.py plus handles the Chromium DRT case and the new baseline search path we'll want. I should have it ready inside an hour or so (I would've finished it last night but I got derailed by another bug), so whichever of the two lands first can be coordinated with the other one. OK, your patch looks good to me, so I'm going to try your patch against the DRT changes in this patch, and if it all works I'll just revert all the python changes in this patch and resubmit. > > WebKitTools/Scripts/webkitpy/layout_tests/run_webkit_tests.py:577 > > + test_args.enable_accelerated_2d_canvas = self._options.enable_accelerated_2d_canvas > > I don't think you need this. > > > WebKitTools/Scripts/webkitpy/layout_tests/run_webkit_tests.py:1663 > > + dest="enable_accelerated_2d_canvas", > > I don't think you need this, either. > > > WebKitTools/Scripts/webkitpy/layout_tests/port/chromium.py:360 > > + > > Here you need to check if enable-accelerated-2d-canvas is not actually set (since there's no default value): > > if hasattr(self._options, 'enable_accelerated_2d_canvas') and self._options.enable_accelerated_2d_canvas): ...
Created attachment 68370 [details] Patch
Comment on attachment 68370 [details] Patch change looks good to me, but I'm not a reviewer.
Committed r68136: <http://trac.webkit.org/changeset/68136>
http://trac.webkit.org/changeset/68136 might have broken Chromium Linux Release The following changes are on the blame list: http://trac.webkit.org/changeset/68136 http://trac.webkit.org/changeset/68137 http://trac.webkit.org/changeset/68138