Support SVGUseElement child within SVGClipPathElement?

[dirkschulze]

SVG 1.1 allows a <use> element within a <clipPath> element as long as the reference chain ends with an SVGShapeElement.

As far as I recall, SVG Native wasn't supposed to support an SVGUseElement as child. However, the spec does not disallow it. @litherum, is this an issue in the spec or is a SVGUseElement child supported in SVG Native?

[anthrotype]

I would also like to know the answer to this question.

In our nanoemoji color font compiler we would like to use <use> element to reuse similar shapes within/across glyphs to reduce the file size. However, neither Adobe apps nor Safari seem to like that: https://github.com/googlefonts/nanoemoji/issues/264#issuecomment-817983791 https://github.com/googlefonts/nanoemoji/issues/266#issuecomment-817975544

@dirkschulze question was specifically about allowing <use> element within a <clipPath>. My question is more general, is <use> actually supported at all, not just as a child of <clipPath>?

[dirkschulze]

@anthrotype Yes, <use> is supported in general. I can not speak for the macOS/Safari implementation but to my knowledge this is the case for Safari and its OT-SVG fonts implementation as well. I am unsure why Adobe Illustrator is not able to render your file as expected. This might be an issue in the used SVG parser or the with the SVG content itself.

[anthrotype]

as @litherum pointed out, we are using the unqualified href attribute whereas Safari expects the xlink:href (with appropriate xmlns declared) which the spec says it's deprecated, but apparently is still valid for SVG 1.1 content.

https://github.com/googlefonts/nanoemoji/issues/266#issuecomment-818289140

[moazin]

@dirkschulze I would like to point out that NotoColorEmoji glyphs have a use element inside a clipPath element and it's very common. Given that it's so common, we'd like to have support for this.

[anthrotype]

90% of the glyphs

that's a lot, where/how did you get those metrics?

[moazin]

I should make a correction, it's not 90% and I don't have an exact metric. I iterated through ~3000 of those glyphs manually and almost every other glyph would not render properly due to this limitation of SNV. So it's a lot of glyphs but definitely not 90%. I confused this with another problem of SNV which I've pointed out in #154. That affected almost every glyph in NotoColor. Thanks for pointing out. :laughing: