45852 - [CASCADE] UA !important does not override author !important style rules

Status: RESOLVED DUPLICATE of bug 43220

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

Description

[fantasai]

Overview:

        According to both the CSS1 and CSS2 specifications, all rules in the
        the author stylesheet completely override the UA styles, regardless of 
        !important status. This is not happening in Mozilla.

        CSS1:3.2 - http://www.w3.org/TR/REC-CSS1#cascading-order
        CSS2:6.4.1 - http://www.w3.org/TR/REC-CSS2/cascade.html#cascading-order

Steps to Reproduce, Actual Results, and Expected Results:

        Open up testcase (to be attached shortly) and follow directions therein.

Tested on Mozilla nightly build (id: 2000071810) on Windows 2000

Comment 1

[fantasai]

Attachment: Author vs. UA !important - testcase

Comment 2

[Marc Attinasi]

This bug is invalid. !important is not supported by CSS in the UA stylesheet. We 
do use it very sparingly for some special internal purposes, related mostly to 
control aggregation, however this is non-standard. 

Please refer to 6.4.2 of the CSS2 specification: it says that both user and 
author stylesheets may contain !important rules.

Comment 3

[Christine Hoffman]

Adding 'verifyme' keyword

Comment 4

[Hixie (not reading bugmail)]

In fact, per the last face-to-face meeting of the CSS Working Group, Mozilla's
behaviour may even be added to the CSS3 cascade spec.

Comment 5

[fantasai]

> We do use it very sparingly for some special internal purposes, related mostly 
> to control aggregation, however this is non-standard. 

I did not know that. However this does not seem very logical:
  !important in ua.css overrides _normal_ rules in author style
but does not override !important rules in the author styles.
I did not believe this to be by design, therefore I chose to 
file the bug against ua !important overriding normal author rules.
Should you verify that this situation is indeed a problem, I will
either change the summary and attach a modified testcase, or, upon
your request, open a new bug.

Comment 6

[Hixie (not reading bugmail)]

!important rules in ua.css should be un-overridable by author and user rules.
Are you saying this is not the case?

(Removing testcase keyword to remove this bug from the valid testcase radar.)

Comment 7

[fantasai]

Precisely.

Here, I'll attach the testcase I originally used, so you can see for yourself.

Bug 43220, bug 45850, and this bug are very closely related; they stem from the 
same basic problems. You can see why by comparing the two primary sort orders 
posted in bug 43220.


*Changing summary to reflect your comment and removing css1 & css2 keywords.*

Comment 8

[fantasai]

Attachment: original testcase

Comment 9

[Hixie (not reading bugmail)]

Reopening. Fantasai is correct. We are not behaving as advertised.
New, simpler testcase to be attached. Thanks Fantasai!

Comment 10

[Hixie (not reading bugmail)]

Attachment: Simple and self explanatory testcase. Output should be green, not red. Make sure to change your ua.css file and restart when testing this!

Comment 11

[Hixie (not reading bugmail)]

Netscape's standard compliance QA team reorganised itself once again, so taking 
remaining non-tables style bugs. Sorry about the spam. I tried to get this done 
directly at the database level, but apparently that is "not easy because of the 
shadow db", "plus it screws up the audit trail", so no can do...

Comment 12

[karnaze (gone)]

Moving to m0.9.3

Comment 13

[Pierre Saslawsky]

*** This bug has been marked as a duplicate of 43220 ***