57686 - Invisible table borders don't render in composer unless you specify border=0

Status: VERIFIED FIXED

(Core :: DOM: Editor, defect, P3)

Description

[kinmoz]

In composer, it looks like we only display invisible table borders if the table
has a border=0 attribute.

We should also handle the case where border is not specified in the table tag at
all.

Assigning to brade as per beppe, since cmanske is on sabbatical.

Comment 1

[kinmoz]

Attachment: Test case.

Comment 2

[Kathleen :Brade]

accepting bug

Comment 3

[rubydoo123]

setting to mozilla0.9

Comment 4

[Kathleen :Brade]

This is probably a bug in EditorContent.css; reassign to cmanske

Comment 5

[Charles Manske]

This is not trivial to do. We can't set a CSS rule to show a border when an
attribute is missing. We would have to add special attributes just for Composer
during tag parsing. E.g.: if there's no "border" attribute, set border="0" and
some flag (such as a "_moz_no_border" attribute) to tell output system to not
write out the "0" so we can preserve the original HTML.

Comment 6

[Daniel Glazman (:glazou) (not active in Mozilla any more)]

The effect you want to achieve is *exactly* the purpose of the new pseudo-class
:not() in the currently standardized CSS 3 Selectors module. This pseudo
notation takes a simple selector - e.g. an attribute selector - as argument, and
is matched when the argument is not matched by the element.

So what we want to write is table:not([border]) { border : 1px red dotted }

This effect cannot be done only through CSS 2 cascade mechanism.

Comment 7

[Kathleen :Brade]

maybe [e-mail] can help with this bug (thanks to Daniel's tip)?

Comment 8

[Håkan Waara]

No one can fix this before :not is implemented. At least not from CSS.

Comment 9

[Daniel Glazman (:glazou) (not active in Mozilla any more)]

I hope to have a working implementation of :not() by the end of today or at the
beginning of next week (I'll be traveling over the week-end and unable to
work on this bug)

Comment 10

[Daniel Glazman (:glazou) (not active in Mozilla any more)]

Changing dependency from trk bug 65133 to bug 71647.

Comment 11

[Daniel Glazman (:glazou) (not active in Mozilla any more)]

Attachment: patch for 57686

Comment 12

[Daniel Glazman (:glazou) (not active in Mozilla any more)]

Proposing a patch for 57686. This patch can only work with the fix for bug 71647
integrated. Patch for 71647 has been added to Bugzilla a few minutes ago, and
57686 is now fixed in my own build.

r= ?

Comment 13

[kinmoz]

sr=[e-mail]

I'm sure cmanske would love to give his r= on this one!

Comment 14

[Charles Manske]

You're right, I would, though I'm definitely am not qualified to review CSS code!
I would love to just trust Daniel, but would feel better if another CSS guru
gave it a look.

Comment 15

[timeless]

hixie could probably r= this code.

you can have an r=timeless if you value it, however this code can't be checked 
in until a patch is accepted for the other bug. I expect that by the time that 
happens, I will value the r= I put in this bug.

Comment 16

[Daniel Glazman (:glazou) (not active in Mozilla any more)]

Reassigning to myself

Comment 17

[Kathleen :Brade]

r=brade

Comment 18

[Daniel Glazman (:glazou) (not active in Mozilla any more)]

checked in :-) Yay !!!!

Comment 19

[Kathleen :Brade]

verified with optimized linux mozilla build from 8am today.