Fix WebCore library path used on DRT.
Created attachment 141688 [details] Patch
Comment on attachment 141688 [details] Patch looks fine, though I'm not an EFL expert so I'll trust that this is right.
Comment on attachment 141688 [details] Patch Clearing flags on attachment: 141688 Committed r116981: <http://trac.webkit.org/changeset/116981>
All reviewed patches have been landed. Closing bug.
103 new failures on the build bot now: http://build.webkit.org/builders/EFL%20Linux%20Debug/builds/1639 I'm not 100% sure this is caused by this patch because there were 2 others in that batch. However, this seems the most likely. Any idea?
It looks like those are all compositor failures, and according to http://build.webkit.org/builders/EFL%20Linux%20Debug/builds/1639, http://trac.webkit.org/changeset/116979 also was a part of that build. It seems more likely to me that the compositing change broke the compositing tests?
(In reply to comment #6) > It looks like those are all compositor failures, and according to http://build.webkit.org/builders/EFL%20Linux%20Debug/builds/1639, > > http://trac.webkit.org/changeset/116979 also was a part of that build. It seems more likely to me that the compositing change broke the compositing tests? My change is a Chromium-only change that adds asserts, so it seems unlikely that it would cause anything other than crashes on debug Chromium bots.
(In reply to comment #7) > (In reply to comment #6) > > It looks like those are all compositor failures, and according to http://build.webkit.org/builders/EFL%20Linux%20Debug/builds/1639, > > > > http://trac.webkit.org/changeset/116979 also was a part of that build. It seems more likely to me that the compositing change broke the compositing tests? > > My change is a Chromium-only change that adds asserts, so it seems unlikely that it would cause anything other than crashes on debug Chromium bots. Okay, that's what I get for not looking. Well, it's easy enough to revert this change and see if things green up.
Re-opened since this is blocked by 86402
Reopening. I think adding myself as CC to the bug collided with the buildbot changing the status of this bug.
For posterity: the problem is that since <http://trac.webkit.org/changeset/107020> the shared libraries are being built in a different place in the build directory. However, that change forgot to do what the current patch does. But this also meant webcore_efl.so was never found, so webkitpy always considered compositing, mathml and friends were not enabled and the respective tests were not run, and they were thus not rebased. tl;dr: Kangil's patch is correct, but he should have rebaselined the failing tests. For now, we have rolled his patch out to keep the bot green.
Created attachment 142473 [details] Patch
(In reply to comment #11) > For posterity: the problem is that since <http://trac.webkit.org/changeset/107020> the shared libraries are being built in a different place in the build directory. However, that change forgot to do what the current patch does. But this also meant webcore_efl.so was never found, so webkitpy always considered compositing, mathml and friends were not enabled and the respective tests were not run, and they were thus not rebased. > > tl;dr: Kangil's patch is correct, but he should have rebaselined the failing tests. For now, we have rolled his patch out to keep the bot green. Done!
Comment on attachment 142473 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=142473&action=review Thanks for working on this. We also have PNGs in compositing/ and mathml/ that should be updated. Please run r-w-t with --pixel-tests for these directories. Doing so even helps checking these new results are correct with `webkit-patch rebaseline-server'. > LayoutTests/platform/efl/test_expectations.txt:-463 > -// This test fails if mixer is not at 100% > -BUGWK83704 : media/event-attributes.html = TEXT This is wrong and unrelated to this patch.
Please commit BUG 86956, 86957 first.
Created attachment 143261 [details] Patch
Attachment 143261 [details] did not pass style-queue: Failed to run "['Tools/Scripts/check-webkit-style', '--diff-files', u'LayoutTests/ChangeLog', u'LayoutTests/plat..." exit_code: 1 LayoutTests/platform/efl/test_expectations.txt:119: fast/events/show-modal-dialog-onblur-onfocus.html is also in a Skipped file. [test/expectations] [5] Total errors found: 1 in 4 files If any of these errors are false positives, please file a bug against check-webkit-style.
The error was not come from my patch. I think it should be fixed in different patch.
LGTM. Just rebase to make the style checker happy. It doesn't fail for me here locally.
Created attachment 143552 [details] Patch
Comment on attachment 143552 [details] Patch Informal r+, thank you all for fixing this!
Comment on attachment 143552 [details] Patch Clearing flags on attachment: 143552 Committed r118311: <http://trac.webkit.org/changeset/118311>