142592 - color attributes on MathML elements can't be updated sometime

Status: RESOLVED FIXED

(Core :: MathML, defect)

Description

[steve.swanson]

1. Open the testcase 
2. Open the DOM Inspector and navigate to the first MathML <mi> entity in the 
document
3. Change the "mathcolor" attribute to value "red"
4. Notice how the color changes to black!

If you change the color back to what it started at (green), the color does 
change to green.

Similar behavior happens with the mathbackground attribute, but not the 
mathsize attribute. (mathvariant, the other MathML base attribute, is not 
supported in Mozilla).

Strangely, the behavior does not occur in a standalone document. If I build a 
document containing JavaScript which modifies the mathcolor, the color changes!

It seems the problem has something to do with XUL, since I have XUL code which 
tries to modify these attributes but fails.  (The DOM is modified, but the 
appearance doesn't match.)

Comment 1

[steve.swanson]

Attachment: Very short test case for use with DOM Inspector

Comment 2

[rbs]

Do you have your JS example handily available?

Comment 3

[rbs]

Attachment: fix for the bug

The mapping of attributes to CSS just need to be kept in sync (bug 69409#c4).

Comment 4

[rbs]

Cc:ing the usual suspects in the tiny/closed circle.

steve, is this bug important for your JS MathML editor? If so, feel free to 
nominate for m1.0 to bring it to the consideration of drivers.

Comment 5

[steve.swanson]

This is somewhat important to me, but it's not something I'd lose sleep over. 
My MathML editor doesn't have a direct interface to color, so you have to go 
through the general attribute dialog.  (And if you save and reload, the correct 
color does show up.)

The principal argument I have for fixing this is that these attributes are 
probably the first ones people will try to use when running the MathML editor. 
That is, they're prominent in the MathML spec and (should) have an obvious 
effect.

Bug 135895 was much more important to me, and I'm glad to see it's fixed in RC2.

Comment 6

[rbs]

--> ok, moving to m1.0; there are no blatant topcrashers in MathML, and the
special artistic issues of MathML are pretty much what need to be addressed
properly while time permits (and there is now a planned RC3).

roc/attinasi: requesting your kind r=/sr=, thanks.

Comment 7

[steve.swanson]

I applied the patch and it fixed the bug.  It's so nice when something you 
expect to work, works.

Comment 8

[rbs]

Attachment: updated patch (oops... really skip over duplicate rules)

Comment 9

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Comment on attachment 83452 [details] [diff] [review]
updated patch (oops... really skip over duplicate rules)

r=roc+moz

Comment 10

[Marc Attinasi]

Comment on attachment 83452 [details] [diff] [review]
updated patch (oops... really skip over duplicate rules)

sr=attinasi

Comment 11

[rbs]

Fixed on the trunk. Have requested drivers' approval for the m1.0 branch.

Comment 12

[Brendan Eich [:brendan]]

Let's get this into the 1.0 branch.  I'll update the patch with approved status
in a second.

/be

Comment 13

[Dawn Endico]

Comment on attachment 83452 [details] [diff] [review]
updated patch (oops... really skip over duplicate rules)

a=brendan,shaver,chofmann

please check this in on the branch asap for rc3

Comment 14

[Mike Shaver (:shaver -- probably not reading bugmail closely)]

I checked this in on the branch, for extra weekend baking.  Thanks!