[JSC] Do not use FTLOutput::weakPointer directly
Created attachment 378051 [details] Patch
Comment on attachment 378051 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=378051&action=review > Source/JavaScriptCore/ChangeLog:12 > + For FrozenValue, we should use frozenPointer helper function. Note that DFG implementation of CreatePromise/NewPromise are using DFGSpeculativeJIT::TrustedImmPtr::weakPointer correctly. And I checked NewPromise and it was OK.
<rdar://problem/55055463>
Ensures this patch fixes the debug assertion in jsc-stress-tests. https://build.webkit.org/builders/Apple%20High%20Sierra%20Debug%20JSC%20%28Tests%29/builds/3589
Comment on attachment 378051 [details] Patch Can you do a follow-up that renames FTL::Output::weakPointer so that folks don't make the same mistake? Maybe Output::alreadyRegisteredWeakPointer or something that makes it clear what the assumptions are.
Comment on attachment 378051 [details] Patch Clearing flags on attachment: 378051 Committed r249530: <https://trac.webkit.org/changeset/249530>
All reviewed patches have been landed. Closing bug.
(In reply to Filip Pizlo from comment #5) > Comment on attachment 378051 [details] > Patch > > Can you do a follow-up that renames FTL::Output::weakPointer so that folks > don't make the same mistake? Maybe Output::alreadyRegisteredWeakPointer or > something that makes it clear what the assumptions are. Sounds nice! I'll change this.
Committed r249552: <https://trac.webkit.org/changeset/249552>