Bug 264752 – Negative SVGTransform scale values should be correctly stringified

• Status: RESOLVED
• Resolution: FIXED
• Component: SVG

[Ahmad Saleem]

Hi Team,

While going through Blink's commit, I came across another failing test case:

Test Case: https://jsfiddle.net/7nqg2L01/show

^ Firefox Nightly 121:

FAIL g.getAttribute('transform') should be scale(-2 -4). Was scale(-2, -4).

^ Chrome Canary 121:

PASS g.getAttribute('transform') is "scale(-2 -4)"

^ Safari Technology Preview 182 (Don't show negative):

FAIL g.getAttribute('transform') should be scale(-2 -4). Was scale(2 4).

___

Blink Commit: https://chromium.googlesource.com/chromium/blink/+/8a7252a2602bd8067c83d48c8fd525ef322708c2

Just wanted to raise, so we can fix it.

Thanks!

[Radar WebKit Bug Importer]

<rdar://problem/118656892>

[Karl Dubost]

There are potentially two different issues 
* negative to positive number
* serialization with/without comma
as the 3 browsers return something different. 



         IN             OUT
Safari:  scale(-2, -4)  scale(2 4).   No comma, positive numbers
Firefox: scale(-2, -4)  scale(-2, -4) Comma, negative numbers
Chrome:  scale(-2, -4)  scale(-2 -4)  No comma, negative numbers

I have the feeling that Firefox is right here for the comma.
https://drafts.csswg.org/css-transforms/#funcdef-transform-scale
I didn't find a reference why it should be serialize differently, aka without the comma. 


But WebKit definitely needs to return negative numbers.

[Karl Dubost]

The test:

var svg = document.createElementNS('http://www.w3.org/2000/svg', 'svg');
var g = document.createElementNS('http://www.w3.org/2000/svg', 'g');
var transform = svg.createSVGTransform();
transform.setScale(-2, -4);
g.transform.baseVal.appendItem(transform)
g.getAttribute('transform')


​

[Ahmad Saleem]

We need to update this:

https://searchfox.org/wubkat/rev/c22ca98693565713efde37e8311190dd00e165ef/Source/WebCore/svg/SVGTransformValue.h#267

void appendScale(StringBuilder& builder) const
    {
        appendFixedPrecisionNumbers(builder, m_matrix->value().xScale(), m_matrix->value().yScale());
    }

TO:

void appendScale(StringBuilder& builder) const
    {
        appendFixedPrecisionNumbers(builder, m_matrix->a(), m_matrix->d());
    }

__

Compiles and make us pass but match 'Chrome'.

As for ',' issue highlighted in Comment 02. I think it might be 'Parser' issue? I am happy to fix this and then have separate bug to discuss ',' issue.

[Ahmad Saleem]

Draft PR - https://github.com/WebKit/WebKit/pull/22475

To see if it fails anything else on EWS. :-)

[EWS]

Committed 272885@main (2efdea0ba354): <https://commits.webkit.org/272885@main>

Reviewed commits have been landed. Closing PR #22475 and removing active labels.