Slot element's wrappers getting collected |
|||
Issue descriptionReport from rniwa@apple.com ---- Hi, It looks like both WebKit and Blink have a problem where we end up collecting JS wrappers of slot elements enqueued for a slotchange event You can see that this test case fails on both Safari & Chrome where it passes always on Firefox: https://bug-167652-attachments.webkit.org/attachment.cgi?id=350479 (You may need to reload multiple times / tweak the numbers in the test to reproduce the issue on Chrome) I was under the impression that oilpan solved all these wrapper lifecycle related issues but perhaps you still manually keep JS wrappers alive? If so, you probably want to make sure you don’t have the same bug for custom element reaction queues as we did in WebKit: https://trac.webkit.org/browser/webkit/trunk/LayoutTests/fast/custom-elements/custom-elements-reaction-queue-retains-js-wrapper.html?rev=236376 Again, you might need to tweak the test to induce the issue in Chrome if one exists.
,
Sep 25
The issue can be reproduced easily: https://bug-167652-attachments.webkit.org/attachment.cgi?id=350479
,
Sep 25
I chatted with yukishiino@. Per the discussion, this issue will be resolved once unified GC is enabled in Blink. Given that unified GC will be shipped sooner or later, we won't take any action on this issue.
,
Sep 25
We might close this as WONTFIX after unified GC is shipped. |
|||
►
Sign in to add a comment |
|||
Comment 1 by hayato@chromium.org
, Sep 25