Bug 226145 – Drop WTF::tryHoldLock() as it is incompatible with Clang Thread Safety Analysis

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

[Chris Dumez]

Drop WTF::tryHoldLock() as it is incompatible with Clang Thread Safety Analysis. Instead, use the following pattern:
```
Locker locker { DeferLock, m_lock };
if (locker.tryLock())
```

[Chris Dumez]

Created attachment 429429 [details]
Patch

[EWS Watchlist]

Thanks for the patch. If this patch contains new public API please make sure it follows the guidelines for new WebKit2 GTK+ API. See https://trac.webkit.org/wiki/WebKitGTK/AddingNewWebKit2API

[Chris Dumez]

Created attachment 429430 [details]
Patch

[Chris Dumez]

Created attachment 429432 [details]
Patch

[Chris Dumez]

Created attachment 429433 [details]
Patch

[Chris Dumez]

Created attachment 429436 [details]
Patch

[Chris Dumez]

Created attachment 429437 [details]
Patch

[Chris Dumez]

Created attachment 429439 [details]
Patch

[Chris Dumez]

Created attachment 429440 [details]
Patch

[Chris Dumez]

Not sure what's going on, it builds fine locally :/

[Chris Dumez]

Created attachment 429443 [details]
Patch

[Chris Dumez]

Created attachment 429446 [details]
Patch

[Chris Dumez]

Created attachment 429447 [details]
Patch

[Chris Dumez]

Created attachment 429448 [details]
Patch

[Chris Dumez]

Created attachment 429449 [details]
Patch

[Darin Adler]

Comment on attachment 429449 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=429449&action=review

> Source/WTF/wtf/Locker.h:59
> +constexpr AdoptLockTag AdoptLock = AdoptLockTag();

I am surprised that we need the "= AdoptLockTag()" here. Can we omit it?

[Chris Dumez]

(In reply to Darin Adler from comment #16)
> Comment on attachment 429449 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=429449&action=review
> 
> > Source/WTF/wtf/Locker.h:59
> > +constexpr AdoptLockTag AdoptLock = AdoptLockTag();
> 
> I am surprised that we need the "= AdoptLockTag()" here. Can we omit it?

Will check and land without it if it builds. I had copied the definition from <mutex>:
/* _LIBCPP_INLINE_VAR */ constexpr adopt_lock_t  adopt_lock  = adopt_lock_t();

[Chris Dumez]

Committed r277932 (238062@main): <https://commits.webkit.org/238062@main>

[Radar WebKit Bug Importer]

<rdar://problem/78371802>

[Chris Dumez]

(In reply to Chris Dumez from comment #17)
> (In reply to Darin Adler from comment #16)
> > Comment on attachment 429449 [details]
> > Patch
> > 
> > View in context:
> > https://bugs.webkit.org/attachment.cgi?id=429449&action=review
> > 
> > > Source/WTF/wtf/Locker.h:59
> > > +constexpr AdoptLockTag AdoptLock = AdoptLockTag();
> > 
> > I am surprised that we need the "= AdoptLockTag()" here. Can we omit it?
> 
> Will check and land without it if it builds. I had copied the definition
> from <mutex>:
> /* _LIBCPP_INLINE_VAR */ constexpr adopt_lock_t  adopt_lock  =
> adopt_lock_t();

Yes, it built fine without it.