Bug 226432 – Drop UncheckedCondition / UncheckedLock

• Status: RESOLVED
• Resolution: FIXED
• Component: Web Template Framework

[Chris Dumez]

Drop UncheckedCondition / UncheckedLock now that they are unused.

[Chris Dumez]

Created attachment 430137 [details]
Patch

[Darin Adler]

Comment on attachment 430137 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=430137&action=review

> Source/WTF/wtf/Lock.h:146
> +template<typename LockType> inline void assertIsHeld(const LockType& lock) { ASSERT_UNUSED(lock, lock.isLocked()); }

Why is this template needed? What classes is this used on besides Lock? Is a template the best way to do this, or would simple overloading be better?

[Chris Dumez]

Comment on attachment 430137 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=430137&action=review

> Source/WTF/wtf/Condition.h:201
> +                    assertIsHeld(lock);

I had to add an assertIsHeld() here to make clang happy (even though I already marked the whole function as WTF_IGNORES_THREAD_SAFETY_ANALYSIS). I am not sure why but anyway. WTF::Condition is templated and supposed to work with all kind of locks (RecursiveLock comes to mind but this is not limited AFAIK).
As a result, I had to add a templated assertIsHeld() so that this would build no matter what lock type the condition is used with.

[Chris Dumez]

(In reply to Chris Dumez from comment #3)
> Comment on attachment 430137 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=430137&action=review
> 
> > Source/WTF/wtf/Condition.h:201
> > +                    assertIsHeld(lock);
> 
> I had to add an assertIsHeld() here to make clang happy (even though I
> already marked the whole function as WTF_IGNORES_THREAD_SAFETY_ANALYSIS). I
> am not sure why but anyway. WTF::Condition is templated and supposed to work
> with all kind of locks (RecursiveLock comes to mind but this is not limited
> AFAIK).
> As a result, I had to add a templated assertIsHeld() so that this would
> build no matter what lock type the condition is used with.

Oh, based on something I just saw online, I may be able to add WTF_IGNORES_THREAD_SAFETY_ANALYSIS on the lambda. Let me try that. If that works, then I wouldn't need this new templated assertIsHeld at all.

[Chris Dumez]

Created attachment 430140 [details]
Patch

[EWS]

Committed r278257 (238294@main): <https://commits.webkit.org/238294@main>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 430140 [details].

[Radar WebKit Bug Importer]

<rdar://problem/78667488>