Add "in select" insertion mode to parser.
Created attachment 70815 [details] Patch
I couldn't figure out how to contrive a test case to exercise this. If you know how, please let me know.
Comment on attachment 70815 [details] Patch From the ASSERT, it looks like you'll need to parse a fragment. Maybe try setting the innerHTML of an <option> element to "</select>" ? <select> <option>xxx</option> </select> You might need to mix in some optgroup elements or something to see the difference between table scope and select scope. r- for no test. Tests for these sorts of changes are essential.
I don't think it's possible to test. The innerHTML of <option> doesn't work because that's processed in "in body" mode. Only <select> starts "in select" mode. From "in select" mode, there doesn't seem to be any way to insert an element that would change the scope, but not the mode.
Comment on attachment 70815 [details] Patch We looked at this change in detail. We don't believe it's observable because it's not possible to have a select element in the stack of open elements when in fragment mode.
Comment on attachment 70815 [details] Patch Rejecting patch 70815 from commit-queue. Failed to run "['./WebKitTools/Scripts/webkit-patch', '--status-host=queues.webkit.org', '--bot-id=abarth-cq-sl', 'land-attachment', '--force-clean', '--ignore-builders', '--quiet', '--non-interactive', '--parent-command=commit-queue', 70815]" exit_code: 2 Last 500 characters of output: atform/win/Skipped M LayoutTests/ChangeLog r70080 = 6b74d470f67a7767229ffea11b1571769f0f5ce5 (refs/remotes/trunk) First, rewinding head to replay your work on top of it... error: Untracked working tree file 'LayoutTests/fast/dom/nodesFromRect-links-and-text-expected.txt' would be overwritten by merge. could not detach HEAD rebase refs/remotes/trunk: command returned error: 1 Died at WebKitTools/Scripts/update-webkit line 129. Failed to run "['WebKitTools/Scripts/update-webkit']" exit_code: 2 Full output: http://queues.webkit.org/results/4558003
Eek. I'lll see if I can figure out which bot has the screwy checkout.
(In reply to comment #7) > Eek. I'lll see if I can figure out which bot has the screwy checkout. Is there a command line argument we can pass to run-webkit-tests to tell it not to generated results for new tests?
--[no-]new-test-results Generate results for new tests
Comment on attachment 70815 [details] Patch Adam's machine's checkout was wedged (eseidel)
Comment on attachment 70815 [details] Patch Clearing flags on attachment: 70815 Committed r70094: <http://trac.webkit.org/changeset/70094>
All reviewed patches have been landed. Closing bug.
(In reply to comment #9) > --[no-]new-test-results Generate results for new tests We should definitely add that.