Bug 225787 – m_calleeSaveRegisters should not be a pointer to a pointer

• Status: RESOLVED
• Resolution: FIXED
• Component: New Bugs

[Geoffrey Garen]

m_calleeSaveRegisters should not be a pointer to a pointer

[Geoffrey Garen]

Created attachment 428573 [details]
Patch

[Keith Miller]

Comment on attachment 428573 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=428573&action=review

r=me with nits.

> Source/JavaScriptCore/bytecode/CodeBlock.h:287
> +        bool m_hasCalleeSaveRegisters { false };

Can't we just add an operator bool to RegisterAtOffsetList which is true when the FixedVector has a pointer?

[Geoffrey Garen]

> > Source/JavaScriptCore/bytecode/CodeBlock.h:287
> > +        bool m_hasCalleeSaveRegisters { false };
> 
> Can't we just add an operator bool to RegisterAtOffsetList which is true
> when the FixedVector has a pointer?

If we did that, we wouldn't be able to distinguish "not initialized" from "initialized with 0 size". I'm not sure if that matters or not.

[Geoffrey Garen]

....so I took the conservative approach and maintained support for "initialized with 0 size".

[Geoffrey Garen]

Comment on attachment 428573 [details]
Patch

CQ+

[Keith Miller]

(In reply to Geoffrey Garen from comment #4)
> ....so I took the conservative approach and maintained support for
> "initialized with 0 size".

Ah yes, all those platforms with 0 calleeSavedRegisters we support 🙃

[EWS]

Committed r277475 (237710@main): <https://commits.webkit.org/237710@main>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 428573 [details].

[Radar WebKit Bug Importer]

<rdar://problem/77998921>