158043 - <label> should have pointer cursor

Status: RESOLVED FIXED

(Core :: Layout: Form Controls, defect, P2)

Description

[Jesse Ruderman]

<label>s should have a pointer cursor so the user has some hint that clicking
the label does something.  The UI was discussed in bug 49902.  I think this is a
regression, which might have happened at the same time bug 103409 was fixed.

Examples: 
- Checkboxes on http://www.cs.hmc.edu/~jruderma/s/.
- The text "Search the Web" at http://www.msn.com/.

Comment 1

[Jesse Ruderman]

0.9.9 uses a pointer cursor.

Comment 2

[Boris Zbarsky [:bzbarsky]]

patches accepted

Comment 3

[Tom Gilder]

Any reason this couldn't be solved by just adding...

  label {
    cursor: default;
  }

...to html.css? I'd also suggest adding:

  label:active {
    -moz-outline: 1px dotted invert;
  }

...which would give users more visual feedback that they've clicked on the
label. It would also be nice if it outlined the label when the form element
associated with it was active, but that's a seperate issue reported in bug 142898.

Comment 4

[Andrew Smith]

I tried the same in forms.css (which seemed a more appropriate place) and it
worked fine; I just got bogged down trying to install Gerv's patch maker.

It looks like this got missed when the fix for bug 96813 got checked in. From
dbaron's commit comments: "Move implementation of HTML label element entirely
into the content model so that it can have any display type". Part of the commit
removed the last remaining style rules on label elements:

http://bonsai.mozilla.org/cvsview2.cgi?diff_mode=context&whitespace_mode=show&file=forms.css&root=/cvsroot&subdir=mozilla/layout/html/document/src&command=DIFF_FRAMESET&rev1=3.42&rev2=3.43

Comment 5

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

Attachment: patch

Comment 6

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

Taking bug.  The suggested change seems reasonable to me, although it is
somewhat disturbing to be able to select text with a non-I-beam cursor.

Comment 7

[John Keiser (jkeiser)]

Comment on attachment 98736 [details] [diff] [review]
patch

r=jkeiser
Though is there a reason not to also do Tom's suggestion for label:active?  It
seems like it would help out visually.

Comment 8

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

I commented on bug 142898 about label:active -- such a rule wouldn't work
correctly now if the label has child elements -- so I made bug 142898 depend on
the bug about fixing :active, (although we could hack label elements the same
way a elements are hacked, that would probably be bad).  I'd rather deal with
that separately from the regression from 96813.

Comment 9

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 98736 [details] [diff] [review]
patch

sr=bzbarsky... Though I agree that the selection issue is weird... Do we have a
good way of switching the cursor to i-beam when the user starts dragging in the
label?

Comment 10

[Tom Gilder]

Using this would make the cursor an i-beam on select (not sure if I'm in favour 
of this happening, though):

  label {
    cursor: default;
  }

  label:active {
    cursor: auto;
  }

...however, I'm just wondering if this may cause some issues if the element or 
a parent element had been made unselectable?

Comment 11

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

Fix checked in to trunk, 2002-09-16 07:15 PDT.