- From: Kang-Hao (Kenny) Lu <kennyluck@csail.mit.edu>
- Date: Tue, 06 Mar 2012 06:47:13 +0800
- To: WWW Style <www-style@w3.org>
- CC: Peter Moulder <peter.moulder@monash.edu>
4.3. Line Thickness: the ‘border-width’ properties The spec now says: # The <length> may not be negative. which appears (at least to me) to say that a conforming UA can optionally support negative border-width, although nowhere in the spec says how such a value should be handled. This legacy sentences seems to originate from css3-box[1], where there is also an informative note saying: [[ Note that the margin, unlike the border and padding, may have a negative thickness. That is one way to make adjacent boxes overlap each other. ]] This, on the other hand, indicates that it is not possible to have negative border width, and hence sightly contradicts the normative "may not" sentence. CSS2.1's "Explicit border widths cannot be negative." is no better. Peter Moulder raised concern about this[2] but there wasn't a response to his concern. Currently tests in CSS2.1 related to negative border width are all marked as invalid presumably because of this sloppy sentence (am I right? From the system I couldn't find the reason why they are marked invalid). All browsers interoperably treat negative border-width as invalid. Opera didn't have testing results for some of the border-top-width-* tests that use negative width (i.e. 012, 023, 034, 056, 067, 078)[3], although these tests all pass in Opera12alpha. In summary, I suggest we just say | Negative values are not allowed. (I am for this change in CSS2.1 too) [1] http://www.w3.org/TR/css3-box/ [2] (the mail this mail replies to) http://lists.w3.org/Archives/Public/www-style/2011Jan/0107 [3] http://test.csswg.org/harness/results/CSS21_DEV/section/8.5.1/ Cheers, Kenny
Received on Monday, 5 March 2012 22:47:43 UTC