1234357 - Process 0 as 0deg in CSS angular properties (e.g., gradients, transforms syntax)

Status: NEW

(Core :: CSS Parsing and Computation, defect, P3)

Description

[Jet Villegas (inactive)]

Chrome, Edge, and Safari process 0 as 0deg in CSS gradients. Although this is invalid per this CSS spec, the other major browser engines interoperate:

http://www.w3.org/TR/css3-values/#lengths

I filed a Chromium bug to track the issue: 
https://code.google.com/p/chromium/issues/detail?id=569703

Comment 1

[Jet Villegas (inactive)]

Updated bug title to reflect CSSWG resolution from 2/1/2016 F2F meeting.

Comment 3

[Xidorn Quan [:xidorn] UTC+10]

We have tests for the old behavior, while Servo's style system has the new behavior, so make it block stylo-style-mochitest.

Comment 4

[Emilio Cobos Álvarez (:emilio)]

Attachment: Bug 1234357: Process 0 as 0deg in CSS angular properties.

Review commit: https://reviewboard.mozilla.org/r/125934/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/125934/

Comment 5

[Cameron McCormack (:heycam)]

Comment on attachment 8853912 [details]
Bug 1234357: Process 0 as 0deg in CSS angular properties.

https://reviewboard.mozilla.org/r/125934/#review128468

::: layout/style/test/property_database.js:188
(Diff revision 1)
> +  "-moz-linear-gradient(top left 0, red, blue)",
> +  "-moz-linear-gradient(5px 5px 0, red, blue)",

Is it that a unitless zero angle is allows in a -moz-prefixed gradient only if we've already parsed the position?  Is that the current behaviour, or do we disallow unitless angles anywhere in legacy gradient syntax?

Comment 6

[Emilio Cobos Álvarez (:emilio)]

Comment on attachment 8853912 [details]
Bug 1234357: Process 0 as 0deg in CSS angular properties.

https://reviewboard.mozilla.org/r/125934/#review128468

> Is it that a unitless zero angle is allows in a -moz-prefixed gradient only if we've already parsed the position?  Is that the current behaviour, or do we disallow unitless angles anywhere in legacy gradient syntax?

Right, we disallow it at the first position (where it'd be ambiguous). Otherwise I've just kept VARIANT_ANGLE, which now allows unitless angles. Happy to tweak it as you see fit.

Comment 7

[Emilio Cobos Álvarez (:emilio)]

Perhaps worth waiting until https://github.com/w3c/csswg-drafts/issues/1162 gets resolved.

Comment 8

[Cameron McCormack (:heycam)]

OK, holding off for now.  Please re-request review once the issue is discussed in the WG.

Comment 9

[yisibl]

RESOLVED: Only spec the unitless 0 quirk for transforms & gradients

Spec changes: https://github.com/w3c/csswg-drafts/commit/602789171429b2231223ab1e5acf8f7f11652eb3