Adding a text node to a shadow root crashes; it should instead present the text in the shadow DOM.
Created attachment 90885 [details] Patch
Attachment 90885 [details] did not pass chromium-ews: Output: http://queues.webkit.org/results/8507105 Unexpected failures: fast/dom/shadow/gc-shadow.html fast/dom/shadow/append-child-text.html
I’m setting up a Linux build environment to investigate the cr-linux EWS failures. I don’t know why fast/dom/shadow/append-child-text.html is an unexpected failure… the linked log output indicates that it passed :/
s/append-child-text/gc-shadow/
(In reply to comment #4) > s/append-child-text/gc-shadow/ Hum... Interesting. Running tests on the EWS is still very experimental. I wouldn't be surprised of there were some bugs.
(In reply to comment #3) > I’m setting up a Linux build environment to investigate the cr-linux EWS failures. I don’t know why fast/dom/shadow/append-child-text.html is an unexpected failure… the linked log output indicates that it passed :/ I don't think you need a Linux env to test this. There's nothing Linuxey about this code.
Comment on attachment 90885 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=90885&action=review > LayoutTests/fast/dom/shadow/append-child-text-expected.txt:13 > + text run at (0,0) width 39: "PASS" Can this test be dumpAsText?
Comment on attachment 90885 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=90885&action=review >> LayoutTests/fast/dom/shadow/append-child-text-expected.txt:13 >> + text run at (0,0) width 39: "PASS" > > Can this test be dumpAsText? No, because shadow DOM is not dumped as text. That it doesn’t crash is observable, but that the text appears is not observable with a text result. WDYT?
cr-linux’s complaint is probably the lack of pixel test results.
(In reply to comment #9) > cr-linux’s complaint is probably the lack of pixel test results. Yes, and it will need its own expectations due to text metric differences. Let's make this a dumpAsText. Perhaps event file a bug to teach the dumper about shadow DOM.
Created attachment 90886 [details] Patch
Comment on attachment 90886 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=90886&action=review the only nit remaining: > LayoutTests/fast/dom/shadow/append-child-text.html:4 > +<script src="../../js/resources/js-test-pre.js"></script> don't really need the harness for this test.
Created attachment 90887 [details] Patch
For the fossil record: filed bug 59305 for adding text in shadow DOM to layout test text output.
Attachment 90887 [details] did not pass chromium-ews: Output: http://queues.webkit.org/results/8504488 Unexpected failures: fast/dom/shadow/gc-shadow.html
Comment on attachment 90887 [details] Patch Clearing flags on attachment: 90887 Committed r84759: <http://trac.webkit.org/changeset/84759>
All reviewed patches have been landed. Closing bug.
(In reply to comment #2) > Attachment 90885 [details] did not pass chromium-ews: > Output: http://queues.webkit.org/results/8507105 > Unexpected failures: > fast/dom/shadow/gc-shadow.html > fast/dom/shadow/append-child-text.html EWS, you were right. gc-shadow is now crashing on canaries. Dominic, Adam, can you rollout on #webkit?