[css-transforms-2] Should individual transform properties accept unitless 0 angle as a value?

[emilio]

Right now they don't per spec.

This came up over https://bugzilla.mozilla.org/show_bug.cgi?id=1207734.

@birtles seemed to think it was good for consistency. I tend to agree, but I also dislike having to add yet another exception to that rule, so I don't have any strong opinion.

Chromium's impl doesn't allow it.

[emilio]

cc @ericwilligers

[ewilligers]

I think the guideline is to never support unitless 0 angle unless we need to for reasons of legacy compatibility.

Allowing unitless 0 also makes 0 1 0 0 harder to read - it isn't instantly obvious if the author intended rotation around the x or y axis.

The axis() with reordering proposal in #1269 would help here - the author's intent is clear with axis(0 1 0) 0deg or 0deg axis(1 0 0) - but it also weakens the compatibility argument for unitless 0.

[emilio]

I think the guideline is to never support unitless 0 angle unless we need to for reasons of legacy compatibility.

Yeah, that was what I expected too here. I'm glad I caught that on review then :P

Allowing unitless 0 also makes 0 1 0 0 harder to read - it isn't instantly obvious if the author intended rotation around the x or y axis.

The axis() with reordering proposal in #1269 would help here - the author's intent is clear with axis(0 1 0) 0deg or 0deg axis(1 0 0) - but it also weakens the compatibility argument for unitless 0.

Indeed, those are both fair points.

@birtles would you be fine with closing this without a spec change?

[birtles]

Yeah, if the guideline is not to add 0 unless absolutely necessary, that sounds fair enough.