Bug 218794 – Support aspect-ratio on non-replaced elements

• Component: CSS
• Status: RESOLVED FIXED

[Rob Buis]

Support aspect-ratio on non-replaced elements

[Rob Buis]

Created attachment 413801 [details]
Patch

[Rob Buis]

Created attachment 413819 [details]
Patch

[Rob Buis]

Created attachment 413853 [details]
Patch

[Rob Buis]

Created attachment 414217 [details]
Patch

[Simon Fraser (smfr)]

Comment on attachment 414217 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=414217&action=review

> Source/WebCore/rendering/RenderBox.cpp:2421
> +static LayoutUnit inlineSizeFromAspectRatio(LayoutUnit borderPaddingInlineSum, LayoutUnit borderPaddingBlockSum, LayoutUnit aspectRatio, BoxSizing boxSizing, LayoutUnit blockSize)

I think aspectRatio should be floating point.

> Source/WebCore/rendering/style/RenderStyle.h:507
> +    LayoutUnit logicalAspectRatio() const

This return a floating point value, not a LayoutUnit.

[Darin Adler]

Comment on attachment 414217 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=414217&action=review

> Source/WebCore/ChangeLog:11
> +        Compute logical width/height for non-replaced elements
> +        taking aspect ratio into account. This is very basic
> +        support and does not consider things like transferred min/max
> +        block size.

Is it a problem to ship very basic support in web browsers? If so, do we need this behind a feature flag so we can turn it on when it’s more complete?

[Radar WebKit Bug Importer]

<rdar://problem/71529770>

[Rob Buis]

Created attachment 414440 [details]
Patch

[Rob Buis]

Comment on attachment 414217 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=414217&action=review

>> Source/WebCore/rendering/RenderBox.cpp:2421
>> +static LayoutUnit inlineSizeFromAspectRatio(LayoutUnit borderPaddingInlineSum, LayoutUnit borderPaddingBlockSum, LayoutUnit aspectRatio, BoxSizing boxSizing, LayoutUnit blockSize)
> 
> I think aspectRatio should be floating point.

Done.

>> Source/WebCore/rendering/style/RenderStyle.h:507
>> +    LayoutUnit logicalAspectRatio() const
> 
> This return a floating point value, not a LayoutUnit.

Done.

[Rob Buis]

Comment on attachment 414217 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=414217&action=review

>> Source/WebCore/ChangeLog:11
>> +        block size.
> 
> Is it a problem to ship very basic support in web browsers? If so, do we need this behind a feature flag so we can turn it on when it’s more complete?

I feel it is safer to let this mature a bit, at least the majority of aspect ratio tests should pass I think. Maybe it can be turned on in a few weeks if people feel it is safe to do so and there is demand for it.

[Rob Buis]

Created attachment 414664 [details]
Patch

[Rob Buis]

Created attachment 414671 [details]
Patch

[Rob Buis]

ping. Anything left for me to do for this bug? I hope I did not misunderstand Darin's comment 6? :)

[Simon Fraser (smfr)]

(In reply to Rob Buis from comment #13)
> ping. Anything left for me to do for this bug? I hope I did not
> misunderstand Darin's comment 6? :)

He was asking whether it's OK to have this shipping in an incomplete form. If not, we should add a feature flag for it.

[Rob Buis]

(In reply to Simon Fraser (smfr) from comment #14)
> (In reply to Rob Buis from comment #13)
> > ping. Anything left for me to do for this bug? I hope I did not
> > misunderstand Darin's comment 6? :)
> 
> He was asking whether it's OK to have this shipping in an incomplete form.
> If not, we should add a feature flag for it.

Thanks! Just to be clear, at the moment aspect-ratio is an experimental preference. If I understand correctly, that does not equal to actually shipping?

The feature is defined here:
Source/WTF/Scripts/Preferences/WebPreferencesExperimental.yaml 

And parsing and handling aspect-ratio depends on that in CSSPropertyParser::parseSingleValue.

[Darin Adler]

(In reply to Rob Buis from comment #15)
> Thanks! Just to be clear, at the moment aspect-ratio is an experimental
> preference. If I understand correctly, that does not equal to actually
> shipping?

Great, that answers my question.

[EWS]

Committed r270551: <https://trac.webkit.org/changeset/270551>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 414440 [details].