@monfera.robert: Could you please provide us the sample file along with the screen-recording for better understanding of this issue, which would help us in triaging it further.

Thank you.

'filterRes' was removed (see  issue 353272 ), and 'kernelUnitLength' doesn't do anything (not implemented/supported.)

fs@, pdr@, do you think we should WontFix this? Regardless, it's low priority.

I was surprised to see we had no existing bug for kernelUnitLength because we've never had it.

I suppose that if we do choose to not implement it we should remove support for it entirely.

These values have been removed from the spec because they didn't really pan out. I think WONTFIX sounds good.

So we should remove the attribute handling and tracking associated with kernelUnitLength.

If we remove an attribute that has no effect, do we need to go through the whole intent to remove process? The attribute doesn't show up anywhere in our usage data.

Removing our vestigial sounds like a good idea. I think we should go through the intent process, but it should be really lightweight and get approved immediately.

FTR, we have https://www.chromestatus.com/metrics/feature/timeline/popularity/1030 which is for all of the "filter DOM" (so an upper bound of sorts.)

I see both these in the spec: https://www.w3.org/TR/SVG11/filters.html
https://www.w3.org/TR/SVG11/filters.html#feConvolveMatrixElementKernelUnitLengthAttribute
https://www.w3.org/TR/SVG11/filters.html#FilterElementFilterResAttribute


Also the utility of these attributes is to be resolution independent. Check out my example that shows that Chrome renders it rather differently for your Macbook Retina screen (pretty) vs. an external monitor with a lower (1) devicePixelRatio: 

http://bl.ocks.org/monfera/c774b8b7196b7e04a5487102b63a330b
http://bl.ocks.org/monfera/21be9bb116a8e5b2423b706155fdb718

They are however removed [1] and deprecated [2] respectively in the Filter Effects spec.

And the reason for that was because they failed to provide the intended resolution independence.

[1] https://drafts.fxtf.org/filters/#element-attrdef-filter-filterres
[2] https://drafts.fxtf.org/filters/#element-attrdef-feconvolvematrix-kernelunitlength

Got it. I haven't come across the removal and deprecation notes, and these attrs had an impact in FF, which led me to think it's a bug. Glad it serves some clarification purpose at least.