[Webkit-unassigned] [Bug 67176] JavaScriptCore does not have tiered compilation

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=67176





--- Comment #34 from Filip Pizlo <fpizlo at apple.com>  2011-09-05 17:51:26 PST ---
(In reply to comment #32)
> > Just remembered why I did it this way: CodeBlock is not ref counted.  But do we really want to make it ref counted just for this case?
> 
> Gah, sorry, meant OwnPtr rather than RefPtr.  I think the suggestion still stands? - code block could be a local OwnPtr, released & assigned to the executable only if compilation is successful.

But then we still need releaseAlternative().

We need CodeBlock::m_alternative to be set prior to invoking optimized compilation, because optimized compilation reads this field.

So if optimized compilation fails, we need releaseAlternative().  I know, it's nasty.  But that leaves us with these choices:

1) Leave it as-is.

2) Make CodeBlock ref-counted to satisfy this case.

3) Don't set m_fooBarCodeBlock, but still release() it, prior to optimized compilation.  That might make the code make a bit more sense, but it wouldn't be too different from what I have right now.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.