Bug 128409 – Unify JSLock implementation for iOS and non-iOS ports.

• Status: RESOLVED
• Resolution: FIXED
• Component: JavaScriptCore

[Mark Lam]

Patch coming.  Details will in the ChangeLog in the patch.

[Mark Lam]

Created attachment 223522 [details]
the patch.

[Mark Hahnenberg]

Comment on attachment 223522 [details]
the patch.

We don't need a special case for dropAllLocks() now that we're on the C stack. You might be able to just make dropAllLocksUnconditionally be the only dropAllLocks.

[Mark Lam]

Comment on attachment 223522 [details]
the patch.

View in context: https://bugs.webkit.org/attachment.cgi?id=223522&action=review

> Source/JavaScriptCore/ChangeLog:12
> +        In the process of switching to iOS implementation, I compared the

typo: to "the" iOS implementation.  Will fix.

> Source/JavaScriptCore/ChangeLog:14
> +        implementation has a bug in dropAllLocks() and dropAllLocksUnconditionally().().

Redundant "().".  Will fix.

[Mark Lam]

Comment on attachment 223522 [details]
the patch.

View in context: https://bugs.webkit.org/attachment.cgi?id=223522&action=review

> Source/JavaScriptCore/ChangeLog:16
> +        The bug is that they do not clear our m_vm->stackPointerAtVMEntry and

typo: our ==> out.

> Source/JavaScriptCore/ChangeLog:18
> +        JSLock::unlock() to dropAllLocks() and dropAllLocksUnconditionally().

Missing ")" after "JSLock::unlock()".  Will fix.

[Geoffrey Garen]

Comment on attachment 223522 [details]
the patch.

Split out the bug fix, please.

[Mark Lam]

Created attachment 223534 [details]
patch 2: just there part that makes everyone do drop/grab locks the iOS way.

[Michael Saboff]

Comment on attachment 223534 [details]
patch 2: just there part that makes everyone do drop/grab locks the iOS way.

r=me

[Mark Lam]

Thanks.  Landed in r163685: <http://trac.webkit.org/r163685>.