- From: Zack Weinberg <zweinberg@mozilla.com>
- Date: Tue, 11 May 2010 17:46:00 -0700
- To: robert@ocallahan.org
- Cc: "Tab Atkins Jr." <jackalmage@gmail.com>, Alex Mogilevsky <alexmog@microsoft.com>, Adam Del Vecchio <adam.delvecchio@go-techo.com>, "www-style@w3.org" <www-style@w3.org>
"Robert O'Callahan" <robert@ocallahan.org> wrote: > On Wed, May 12, 2010 at 12:12 PM, Robert O'Callahan > <robert@ocallahan.org>wrote: > > > That's actually a spec change. And you still have an equivalent > > problem when you collapse 20px with 2fl, as it turns out. At least > > some set of cases reduce to the problem of solving for F an > > equation of the form sum_over_i( max(m_i, k_i F) ) = N > > It's piecewise-linear, but increasing and continuous in F > > > > Well, assuming we don't allow negative flex values. I certainly hope > that's true! TeX allows infinitely stretchable negative glue (more or less equivalent to negative flex) and it is quite useful under some circumstances, although its box model is different enough that I won't claim it'd definitely be useful in CSS. Speaking of glue, though, I'm pretty sure people *would* find use for a CSS equivalent of "2pt plus4pt minus1pt" (i.e. "try to make this 2pt wide, but you can stretch it up to 6pt or squash it down to 1pt if that makes things fit better"). And I kinda think we should just bite the bullet and require a linear constraint solver. zw
Received on Wednesday, 12 May 2010 00:46:38 UTC