The following tests are convertible to dumpAsTest. - fast/multicol/negativeColumnWidth.html - fast/multicol/zeroColumnCount.html Since these tests check if the renderer crashes or not when it opens these evil cases, we can safely convert them so they use dumpAsText().
Created attachment 35035 [details] Convert two layout tests so that they use dumpAsText(). These tests check if the renderer crashes or not. Hence, we do not need to dump the entire render tree. I verified DumpRenderTree at r24513 still crashed when it opened the updated test cases. See bug 14714 for details about these test cases. --- 13 files changed, 54 insertions(+), 54 deletions(-)
Comment on attachment 35035 [details] Convert two layout tests so that they use dumpAsText(). I would rather have "technically invalid" html, than tests full of a bunch of boiler-plate. This test succeeds if it does not crash. <table<td style="-webkit-columns: -9999px;"> <script> if (window.layoutTestController) layoutTestController.dumpAsText() </script> would have been sufficient. WebKit style is 4 space indent. so r-.
Alright. I'm going to give it a second try.
Created attachment 35097 [details] Convert two layout tests so that they use dumpAsText(). These tests check if the renderer crashes or not. Hence, we do not need to dump the entire render tree. I verified DumpRenderTree at r24513 still crashed when it opened the updated test cases. See bug 14714 for details about these test cases. --- 13 files changed, 42 insertions(+), 52 deletions(-)
Comment on attachment 35097 [details] Convert two layout tests so that they use dumpAsText(). Hopefully svn-apply will handle this git-diff correctly. :) LGTM.
Comment on attachment 35097 [details] Convert two layout tests so that they use dumpAsText(). Rejecting patch 35097 from commit-queue. This patch will require manual commit. Patch https://bugs.webkit.org/attachment.cgi?id=35097 from bug 28423 failed to download and apply.
Eric, thank you for your review! Unfortunately auto-commit didn't work (maybe because of file removal?), and I don't have committer's access yet, so I'm waiting for somebody's manual commit.
Applying 35097 from bug 28423. patching file LayoutTests/ChangeLog Hunk #1 succeeded at 1 with fuzz 3. patching file LayoutTests/fast/multicol/negativeColumnWidth-expected.txt patching file LayoutTests/fast/multicol/negativeColumnWidth.html patching file LayoutTests/fast/multicol/zeroColumnCount-expected.txt patching file LayoutTests/fast/multicol/zeroColumnCount.html patching file LayoutTests/platform/mac/fast/multicol/negativeColumnWidth-expected.checksum rm 'LayoutTests/platform/mac/fast/multicol/negativeColumnWidth-expected.checksum' patch: **** Only garbage was found in the patch input. patch -p0 "LayoutTests/platform/mac/fast/multicol/negativeColumnWidth-expected.png" returned 2. Pass --force to ignore patch failures. svn-apply doesn't seem to understand this git diff. Possibly a bug in svn-apply. You can get around this by using svn-create-patch.
OK, I'm going to have a svn checkout and try to create a patch with svn-create-patch. It may take some time.
Created attachment 35114 [details] Convert two layout tests so that they use dumpAsText(). (v3) svn-apply should work fine with this patch.
Comment on attachment 35114 [details] Convert two layout tests so that they use dumpAsText(). (v3) I filed bug 28456 about the svn-apply problem. Looks fine.
Comment on attachment 35114 [details] Convert two layout tests so that they use dumpAsText(). (v3) Clearing flags on attachment: 35114 Committed r47510: <http://trac.webkit.org/changeset/47510>
All reviewed patches have been landed. Closing bug.
Thanks for all your work!