Bug 61170 – Zombies should "live" forever

• Component: JavaScriptCore
• Status: RESOLVED FIXED

[Zoltan Herczeg]

Some improvements to keep zombies alive.

[Zoltan Herczeg]

Created attachment 94180 [details]
patch

[Oliver Hunt]

Comment on attachment 94180 [details]
patch

View in context: https://bugs.webkit.org/attachment.cgi?id=94180&action=review

> Source/JavaScriptCore/JavaScriptCore.pro:-67
> -    assembler/MacroAssemblerSH4.h \
>      assembler/MacroAssemblerSH4.cpp \
> -    assembler/SH4Assembler.h \

Why remove these?  Or at least why remove them in this patch?

> Source/JavaScriptCore/runtime/JSCell.h:362
> -        return isCell() && asCell() > (JSCell*)0x1ffffffffL && asCell()->isZombie();
> +        return isCell() && asCell()->isZombie();

whoops, i shouldn't have left this in, my bad.

> Source/JavaScriptCore/runtime/WriteBarrier.h:132
> -        ASSERT(!m_cell || !isZombie(m_cell));
> +        ASSERT(!m_cell || value == reinterpret_cast<T*>(1) || !isZombie(m_cell));

What are you testing here?

[Zoltan Herczeg]

> > Source/JavaScriptCore/JavaScriptCore.pro:-67
> > -    assembler/MacroAssemblerSH4.h \
> >      assembler/MacroAssemblerSH4.cpp \
> > -    assembler/SH4Assembler.h \
> 
> Why remove these?  Or at least why remove them in this patch?

True, it is unrelated. That list should contain only .cpp files. It is just a typo.

> > Source/JavaScriptCore/runtime/WriteBarrier.h:132
> > -        ASSERT(!m_cell || !isZombie(m_cell));
> > +        ASSERT(!m_cell || value == reinterpret_cast<T*>(1) || !isZombie(m_cell));
> 
> What are you testing here?

CodeBlock.h : setSeen():
cachedPrototypeStructure.setWithoutWriteBarrier((Structure*)1);

[WebKit Commit Bot]

Comment on attachment 94180 [details]
patch

Clearing flags on attachment: 94180

Committed r86974: <http://trac.webkit.org/changeset/86974>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.