- From: Tab Atkins Jr. <jackalmage@gmail.com>
- Date: Wed, 17 Jun 2009 12:40:59 -0500
- To: Boris Zbarsky <bzbarsky@mit.edu>
- Cc: Brad Kemper <brad.kemper@gmail.com>, www-style@w3.org
On Wed, Jun 17, 2009 at 12:27 PM, Boris Zbarsky<bzbarsky@mit.edu> wrote: > Brad Kemper wrote: >> >> I dislike having more and more different display values, especially one >> that is almost, but not quite, identical to another. I would prefer a new >> value for 'clear' that affected rows instead of floats. >> .my_table_structure > div { display: table-cell } >> .my_table_structure > div:nth-of-type(4n+1) { clear:row; } > > This doesn't seem that unreasonable, though it raises the issue of what to > do for non-table elements when "clear:row" is set. Presumably it'd just > have no effect. This would be my assumption as well. >> Perhaps this sort of clearing could even be applied to an element in the >> contents of a table cell, to force its nearest table-cell ancestor into a >> new row. > > That's much much harder to implement than just allowing it on the table > cell, for what it's worth. I _think_ the table-cell case could be done in a > few hours with very little performance cost; the other would take a great > deal more time; I can't even speculate as to whether it could be done > without performance costs at this point. > > Are there obvious use cases for it? I can't think of any off the top of my head that couldn't be done equally well by signaling the clearing on the table-cell. In any case, this could always be done with the child-matcher pseudoclass when that finally makes it in. I'm betting perf would be similar. ~TJ
Received on Wednesday, 17 June 2009 17:41:56 UTC