Bug 128001 – Feature flag for shape-inside

• Component: CSS
• Status: RESOLVED FIXED

[Dean Jackson]

I'd like to separate the shape-inside implementation feature defines from the shape-outside ones.

[Radar WebKit Bug Importer]

<rdar://problem/15958487>

[Dean Jackson]

Created attachment 223004 [details]
Patch

[Dean Jackson]

Comment on attachment 223004 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=223004&action=review

> Source/WebCore/css/CSSParser.cpp:5885
> -        || (valueId == CSSValueOutsideShape && propId == CSSPropertyWebkitShapeInside)) {
> +        || (valueId == CSSValueOutsideShape
> +#if ENABLE(CSS_SHAPE_INSIDE)
> +            && propId == CSSPropertyWebkitShapeInside
> +#endif

I'm not sure if this is correct.

[Dean Jackson]

Created attachment 223005 [details]
Patch

[Bear Travis]

(In reply to comment #3)
> (From update of attachment 223004 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=223004&action=review
> 
> > Source/WebCore/css/CSSParser.cpp:5885
> > -        || (valueId == CSSValueOutsideShape && propId == CSSPropertyWebkitShapeInside)) {
> > +        || (valueId == CSSValueOutsideShape
> > +#if ENABLE(CSS_SHAPE_INSIDE)
> > +            && propId == CSSPropertyWebkitShapeInside
> > +#endif
> 
> I'm not sure if this is correct.

The outside-shape value is only valid for shape-inside, so should be wrapped in the compile guard as well. Looks like you already got it in the second patch though.

Thanks for doing this.

[Bem Jones-Bey]

Comment on attachment 223005 [details]
Patch

Two comments:
1) RenderBlock::imageChanged is used by both shape-inside and shape-outside. (It won't let me comment on that line because it's not part of the diff)

2) shape-padding is only used by shape-inside.

Otherwise, looks good to me. Thanks!

[Dean Jackson]

(In reply to comment #6)
> (From update of attachment 223005 [details])
> Two comments:
> 1) RenderBlock::imageChanged is used by both shape-inside and shape-outside. (It won't let me comment on that line because it's not part of the diff)

Great. Fixed this.

> 2) shape-padding is only used by shape-inside.

OK. I'll protect against it in the CSS parser, but I'll leave it in RenderStyle (otherwise it is a bit too intrusive, and I already hate how many #ifs this added).

[Dean Jackson]

Created attachment 223022 [details]
Patch

[Dean Jackson]

I think this one is all good!

[Bem Jones-Bey]

(In reply to comment #9)
> I think this one is all good!

Yeah, it looks good. I should have asked this before, but have you tested that it compiles with the CSS_SHAPE_INSIDE flag turned off? (and maybe even that it passes the shape-outside tests in such a condition?)

[Dean Jackson]

Committed r163333: <http://trac.webkit.org/changeset/163333>

[Darin Adler]

Comment on attachment 223022 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=223022&action=review

> Source/WebCore/css/CSSParser.cpp:5888
> +        || (valueId == CSSValueOutsideShape
> +            && propId == CSSPropertyWebkitShapeInside)

Why broken into two lines? This seems less readable than before.