Use CheckedCondition in even more places to benefit from Clang Thread Safety Analysis.
Created attachment 429454 [details] Patch
Created attachment 429455 [details] Patch
Comment on attachment 429455 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=429455&action=review > Source/WebCore/platform/mediarecorder/cocoa/MediaRecorderPrivateWriterCocoa.mm:373 > +// FIXME: This modifies m_data without grabbing m_dataLock. Hard to fix? > Source/WebCore/worklets/PaintWorkletGlobalScope.h:68 > + HashMap<String, std::unique_ptr<PaintDefinition>>& paintDefinitionMap() WTF_REQUIRES_LOCK(m_paintDefinitionLock) > + { > + ASSERT(m_paintDefinitionLock.isLocked()); > + return m_paintDefinitionMap; > + } Can we put this function body after the class definition to keep the class definition easy to read?
(In reply to Darin Adler from comment #3) > Comment on attachment 429455 [details] > Patch > > View in context: > https://bugs.webkit.org/attachment.cgi?id=429455&action=review > > > Source/WebCore/platform/mediarecorder/cocoa/MediaRecorderPrivateWriterCocoa.mm:373 > > +// FIXME: This modifies m_data without grabbing m_dataLock. > > Hard to fix? I will check in a follow-up. I decided to stay away from fixes in my large refactoring fixes based on earlier feedback. > > > Source/WebCore/worklets/PaintWorkletGlobalScope.h:68 > > + HashMap<String, std::unique_ptr<PaintDefinition>>& paintDefinitionMap() WTF_REQUIRES_LOCK(m_paintDefinitionLock) > > + { > > + ASSERT(m_paintDefinitionLock.isLocked()); > > + return m_paintDefinitionMap; > > + } > > Can we put this function body after the class definition to keep the class > definition easy to read? Will fix.
Created attachment 429471 [details] Patch
Created attachment 429474 [details] Patch
Created attachment 429475 [details] Patch
Comment on attachment 429475 [details] Patch Clearing flags on attachment: 429475 Committed r277934 (238064@main): <https://commits.webkit.org/238064@main>
All reviewed patches have been landed. Closing bug.
<rdar://problem/78372445>