30127 - -moz-fixed not friendly with the cascade

Status: VERIFIED FIXED

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

Description

[Hixie (not reading bugmail)]

You don't seem to be able to override, using the 'inherit' keyword, the font of 
elements which have been styled with the -moz-fixed font in html.css.

The test document contains a piece of emphasised text and a piece of code. They 
should both have the same font as the parent paragraph, exactly. (They are also 
both blue.) However, while the emphasised text does indeed lose its default 
italic, the piece of code does not lose its font.

   http://www.bath.ac.uk/%7Epy8ieh/m/mozfixed-universal.html

FYI, the page contains this stylesheet:
   em, code, pre { font: inherit; color: blue; background: white; }
    
David, Erik: I've cc'ed you in case you are aware of this bug already and know
of a DUP. I could not find any. And a brief persual through lxr.mozilla.org did 
little to enlighten me as to the cause of this problem.

Comment 1

[Hixie (not reading bugmail)]

BTW, this bug is blocking one of my projects. If anyone has a workaround I'm 
all ears...

Marking {font} for pierre. Downgrading to P4.

Comment 2

[Hixie (not reading bugmail)]

More data:
   code { font: 1em monospace; }
...works as expected (so I'm using it as a workaround).

Strange. Very strange. I'm sure there is a perfectly rational answer to this...

Comment 3

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Are you sure this is a problem with -moz-fixed and not a problem with:
 * the cascade (could be tested by changing -moz-fixed to monospace in ua.css)
 * the inherit value on 'font' or 'font-family' (could be tested lots of ways)?

I may have the chance to test myself...

Comment 4

[Pierre Saslawsky]

Someone (Paul Hangas, I think) reported about 2 weeks ago in a bug that I can't 

find back that when an element uses "font-family:-moz-fixed" (like input[type=

text] in html.css), the font cannot be overriden. His workaround was to use 

Courier or monospace instead of -moz-fixed. I think I noted in the aforementioned 

bug that I had already fixed something like that for bug 17303 (see my comment 

from 2000-01-11 02:07).



Help would be appreciated to find the bug, mark it as dup, and write a testcase 

for this one.

Comment 5

[Pierre Saslawsky]

Forget what I wrote: it wasn't a bug but a discussion we had by email between 
hangas, rods, akkana, buster, german and myself. The problem described by hangas 
was:

The text field is totally skinable now.  You can set the font-family and font-
size in html.css and everything works fine.  The problem is that if you set the
font-family in html.css to -moz-fixed then we cannot override this style in 
global.css for html:input fields.  If we can change the html.css style to a
normal font family like Courier or something then everything will work for beta1.  
I don't know why we use -moz-fixed but choosing another font there
would allow us to fix the chrome.  Later we someone can fix the bug that prevents 
this font-family from being overridden in another css file when it has
been set to -moz-fixed.

So the present bug is valid. A testcase would be nice...

Comment 6

[Hixie (not reading bugmail)]

Um, see the original entry for this bug. :-) Or, just look at:
   http://www.bath.ac.uk/%7Epy8ieh/m/mozfixed-universal.html

Comment 7

[Marc Attinasi]

Assigning to ekrock: 1/3 of Pierre's NEW bugs to help reduce his doomage factor

Comment 8

[Pierre Saslawsky]

Reassigned back to me these bugs that shouldn't have left my list.

Comment 9

[Michael Kay]

Is this bug I'm seeing related? It's form HotBot.com. It seems that nesting some

css code within font tags screws up the css. A span sets the font-color to red

but the red "bleeds" beyond that span. I believe this is a regression bug since

it didn't appear until M13 and M14.

See this test case:
http://staff.hotwired.com/mikeque/mozilla/nesting_bug.html

Comment 10

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Mikeque - That's a completely separate bug, probably in residual style handling,
and should be filed against the Parser.

Comment 11

[Michael Kay]

I opened a new bug for that bug (31421).
Thanks,
Mike

Comment 12

[Pierre Saslawsky]

*** Bug 31744 has been marked as a duplicate of this bug. ***

Comment 13

[Ben Bucksch (:BenB)]

A simple test case is attached to bug #31744:
<http://bugzilla.mozilla.org/createattachment.cgi?id=31744>

Need this for bug #31047 (I need "<pre>" for the "whitespace: pre", but want to
conditionally overwrite the "font-family" in Mozilla).

Comment 14

[Ben Bucksch (:BenB)]

> I need "<pre>" for the "whitespace: pre"

In *other* UAs

(Sorry for the spam.)

Comment 15

[Pierre Saslawsky]

Fix checked in nsCSSStyleRule.cpp.

Comment 16

[Ben Bucksch (:BenB)]

wow, that was fast! tnx.

Comment 17

[Hixie (not reading bugmail)]

Original test case renders fine now. Thanks pierre.