1376019 - calc() expressions are rejected, in radial-gradient() radius components

Status: RESOLVED FIXED

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

Description

[Daniel Holbert [:dholbert]]

Attachment: testcase 1

STR:
 1. Load attached testcase.

EXPECTED RESULTS: Every box should be blue with a green circle (not all necessarily the same size).

ACTUAL RESULTS: The second group of boxes (under the "calc() in radii" label) are all empty.

It looks like we're rejecting calc() expressions when parsing the radial-gradient radius (or radii, in the case of ellipse)

Comment 1

[Daniel Holbert [:dholbert]]

Chrome 61, Edge 15, and Safari 10.1 all give EXPECTED RESULTS.

Firefox Nightly 56.0a1 (2017-06-23) gives ACTUAL RESULTS.  So we're on our own on this.

Comment 2

[Emilio Cobos Álvarez (:emilio)]

Stylo also gives expected results here. Do we want to do anything before Stylo? If not, we should probably mark it as a dependency of bug 1243581.

Comment 3

[Daniel Holbert [:dholbert]]

(In reply to Emilio Cobos Álvarez [:emilio] from comment #2)
> Stylo also gives expected results here. Do we want to do anything before
> Stylo?

As in bug 1363349 comment 8: I don't think we *need* to do anything before stylo. But it might also be pretty straightforward to fix this before stylo, if we want to.

Comment 4

[Chris Mills [:cmills]]

Marking as part of documentation plan for intentional stylo differences

Comment 5

[Chris Mills [:cmills]]

Documented:

https://developer.mozilla.org/en-US/docs/Web/CSS/calc#Quantum_CSS_notes
https://developer.mozilla.org/en-US/docs/Web/CSS/radial-gradient#Quantum_CSS_notes
https://developer.mozilla.org/en-US/Firefox/Releases/57#Quantum_CSS_notes

Comment 6

[Daniel Holbert [:dholbert]]

As expected, this is fixed by stylo (i.e. you get expected results if layout.css.servo.enabled is set to true) -- and stylo seems to be riding the trains with 57.

So, I don't think there's anything to gain by leaving this open, so resolving as FIXED by stylo.

BTW, I tried to mark this as depending on meta bug 1243581 (stylo), but I can't, because that'd create a cycle in bugzilla. This bug here is already marked as blocking bug 1365771 (stylo-behavior-changes), which blocks bug 1374034 (stylo-release), which blocks bug 1243581 (stylo).  So if this bug were to depend on bug 1243581 (stylo), that'd be a cycle.

To the extent we care about tracking the fact that this was fixed-by-stylo: I'll let someone who knows more about our stylo tracking bugs sort out how best to update the dependencies here to reflect this without creating a cycle.