1041994 - calc( 0 ) is not accepted as a <length>

Status: RESOLVED INVALID

(Core :: Layout, defect)

Description

[Marek ‘saji’ Augustynowicz]

User Agent: Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/36.0.1985.125 Safari/537.36

Steps to reproduce:

Use e.g. `margin-right: calc( 0 )`.

Example use case: http://codepen.io/marek-saji/full/KGmze/


Actual results:

Propery is ignored.


Expected results:

`calc( 0 )` works exactly like `0`.

Comment 1

[Paul Silaghi, QA [:pauly]]

Opera and Safari seem to show the 'good' behavior.

Comment 2

[Boris Zbarsky [:bzbarsky]]

Effectively, we're not applying the "Treat 0 as a length" special-case to 0 inside calc(), right?  Note that this "treat 0 as a length" is a syntactic shortcut, not a semantic one, according to http://dev.w3.org/csswg/css-values/#length-value

It's not clear to me what the right behavior here is per spec, but at least the following cases should be considered:

  calc(1 - 1)
  calc(0 * 1)
  calc((1 - 1) * 2)

Comment 3

[David Baron :dbaron: (⌚️UTC-5, no longer working on Mozilla)]

The above also combines with http://dev.w3.org/csswg/css-values/#calc-type-checking :

The resolved type of the expression is determined by the types of the values it contains. <number-token>s are of type <number> or <integer>. A <dimension-token>’s type is given by its unit (cm is <length>, deg is <angle>, etc.).


So per the current spec, I think our behavior is correct.  It also matches the general pattern that the 0-as-length behavior doesn't happen in contexts where <number> values are accepted.

Comment 4

[David Baron :dbaron: (⌚️UTC-5, no longer working on Mozilla)]

Chrome seems to have the same behavior we do.

I added a test for this case (and a few others) to the CSSWG test repository:
https://hg.csswg.org/test/rev/60c7ad462121
https://hg.csswg.org/test/rev/78fd41d6c90a