Bug 189643 – Re-order Node flags based on semantics

• Status: RESOLVED
• Resolution: FIXED
• Component: DOM

[Ryosuke Niwa]

Re-order Node flags so that flags used for similar purposes are grouped together.

[Ryosuke Niwa]

Created attachment 349845 [details]
Cleanup

[Simon Fraser (smfr)]

Comment on attachment 349845 [details]
Cleanup

View in context: https://bugs.webkit.org/attachment.cgi?id=349845&action=review

> Source/WebCore/dom/Node.h:566
> -        IsConnectedFlag = 1 << 6,
> -        IsInShadowTreeFlag = 1 << 7,
> -        IsDocumentNodeFlag = 1 << 8,
> -        IsShadowRootFlag = 1 << 9,
> +        IsDocumentNodeFlag = 1 << 6,
> +        IsShadowRootFlag = 1 << 7,
> +        IsConnectedFlag = 1 << 8,
> +        IsInShadowTreeFlag = 1 << 9,
>          HasRareDataFlag = 1 << 10,

I would prefer all the "= 1 << N" be aligned, so it's easy to see that they are increasing by one, but not everyone agrees.

> Source/WebCore/dom/Node.h:584
> +        StyleValidityShift = 22,

Does this break our ability to use this enum with OptionSet<>?

[Ryosuke Niwa]

(In reply to Simon Fraser (smfr) from comment #2)
> Comment on attachment 349845 [details]
> Cleanup
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=349845&action=review
> 
> > Source/WebCore/dom/Node.h:566
> > -        IsConnectedFlag = 1 << 6,
> > -        IsInShadowTreeFlag = 1 << 7,
> > -        IsDocumentNodeFlag = 1 << 8,
> > -        IsShadowRootFlag = 1 << 9,
> > +        IsDocumentNodeFlag = 1 << 6,
> > +        IsShadowRootFlag = 1 << 7,
> > +        IsConnectedFlag = 1 << 8,
> > +        IsInShadowTreeFlag = 1 << 9,
> >          HasRareDataFlag = 1 << 10,
> 
> I would prefer all the "= 1 << N" be aligned, so it's easy to see that they
> are increasing by one, but not everyone agrees.

Well, if we did that, then we'd have to re-align them whenever a flag with a longer name than the current longest one is added.

> > Source/WebCore/dom/Node.h:584
> > +        StyleValidityShift = 22,
> 
> Does this break our ability to use this enum with OptionSet<>?

Yes. But we can probably just add two vales of StyleValidity directly to NodeFlags and do the manual conversion instead.

[Ryosuke Niwa]

Committed r236083: <https://trac.webkit.org/changeset/236083>

[Radar WebKit Bug Importer]

<rdar://problem/44533292>