| Summary: | [Chromium] Worker object may be garbage collected if it doesn't send any messages | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Product: | WebKit | Reporter: | Yury Semikhatsky <yurys> | ||||||
| Component: | WebKit Misc. | Assignee: | Yury Semikhatsky <yurys> | ||||||
| Status: | RESOLVED FIXED | ||||||||
| Severity: | Normal | CC: | ap, caseq, dimich, dslomov, jianli, levin, pfeldman | ||||||
| Priority: | P2 | ||||||||
| Version: | 528+ (Nightly build) | ||||||||
| Hardware: | All | ||||||||
| OS: | All | ||||||||
| Attachments: |
|
||||||||
Created attachment 96729 [details]
Patch
Did you mean r? it or is it still in the works? (In reply to comment #2) > Did you mean r? it or is it still in the works? I wanted to run the test on Chromium to make sure that it passes there. Now the patch is ready for review. Comment on attachment 96729 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=96729&action=review r=me I wonder if it improves flakiness of worker tests run in Chromium. > Source/WebKit/chromium/ChangeLog:11 > + "from it context" -> "from its context" Committed r88914: <http://trac.webkit.org/changeset/88914> (In reply to comment #4) > (From update of attachment 96729 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=96729&action=review > > r=me > I wonder if it improves flakiness of worker tests run in Chromium. > > > Source/WebKit/chromium/ChangeLog:11 > > + > > "from it context" -> "from its context" Done. |
Created attachment 96722 [details] Test case If there are no references to a dedicated worker object and no messages are sent to its worker context(which would increment unconfirmed messages counter which in turn would protect worker from being collected) then it may be garbage collected before a first message from worker context arrives. See attached page where messages from the worker context never reaches worker object. This works fine on other ports that use in-process workers.