745941 - CSS media query (orientation) evaluates true

Status: RESOLVED INVALID

(Core :: CSS Parsing and Computation, defect)

Description

[Eric A. Meyer]

User Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.6; rv:11.0) Gecko/20100101 Firefox/11.0
Build ID: 20120312181643

Steps to reproduce:

Created a CSS rule like so:

@media (orientation) { body {background: green;} }


Actual results:

The body background was green.


Expected results:

The rule should have been ignored.  Per http://www.w3.org/TR/css3-mediaqueries/#media1:

“For a media feature feature, (feature) will evaluate to true if (feature:x) will evaluate to true for a value x other than zero or zero followed by a unit identifier (i.e., other than 0, 0px, 0em, etc.).”

Per http://www.w3.org/TR/css3-mediaqueries/#orientation, the only valid values for 'orientation' are 'portrait' and 'landscape'.  Can it evaluate to true under the defined constraints?

(For comparison, WebKit evaluates to false; Opera to true.  Was not able to test in Explorer.)

Comment 1

[Boris Zbarsky [:bzbarsky]]

> Can it evaluate to true under the defined constraints?

Seems like the spec is pretty clear.  If there is any value x other than 0 or a zero length
that makes (orientation: x) evaluate to true, then (orientation) evaluates to true.  There is certainly such a value at all times: either x == landscape or x == portrait.

I'm not sure what part of the spec made you think anything here should be ignored.  Nothing in the spec says to ignore this query, as far as I can see.

Please do file a bug on WebKit for their buggy implementation.  Or let me know and I'll do it; either way.

Comment 2

[Eric A. Meyer]

Ah, I see.  The wording there is (to me, and probably most non-CS veterans) very misleading.  What it strongly implies (again, to me) is that it only applies when values are or can be numeric with or without a unit, which would preclude keyword-based values from ever matching.  I believe I now grasp the intended logic.  Apologies!

Comment 3

[Eric A. Meyer]

Oh, and I’ll go see about filing a WebKit bug.  I think I might have a login over there…

Comment 4

[Eric A. Meyer]

Filed: https://bugs.webkit.org/show_bug.cgi?id=84151

Comment 5

[Boris Zbarsky [:bzbarsky]]

No need to apologize.  It might be worth raising a spec issue here too; I agree the spec could be clearer here.