https://treeherder.mozilla.org/logviewer.html#?job_id=155762760&repo=try&lineNumber=4994 In WorkerFetchResolver, we use a SignalProxy in both worker and main-thread. We need to be 100% sure that when used, we don't create race conditions as this ASAN log shows.
Created attachment 8942161 [details] [diff] [review] fetch.patch SignalProxy is already correctly used. It seems to me that the only problem is that it's nullified on the worker side. If we don't do that, we are not introducing any leak. Let's go for this simply approach (we have a similar approach for mPromiseProxy).
Attachment #8942161 - Flags: review?(bkelly)
Attachment #8942161 - Flags: review?(bkelly) → review+
Pushed by firstname.lastname@example.org: https://hg.mozilla.org/integration/mozilla-inbound/rev/37ac87b44ae3 Signal Proxy, in fetch(), doesn't have to be nullified in order to avoid race conditions, r=bkelly
Status: NEW → RESOLVED
Last Resolved: 10 months ago
status-firefox59: --- → fixed
Resolution: --- → FIXED
Target Milestone: --- → mozilla59
1 failures in 788 pushes (0.001 failures/push) were associated with this bug in the last 7 days. Repository breakdown: * try: 1 Platform breakdown: * linux64: 1 For more details, see: https://brasstacks.mozilla.com/orangefactor/?display=Bug&bugid=1430061&startday=2018-01-08&endday=2018-01-14&tree=all
You need to log in before you can comment on or make changes to this bug.