- From: Ojan Vafai <ojan@chromium.org>
- Date: Mon, 29 Aug 2011 10:50:54 -0700
- To: "Tab Atkins Jr." <jackalmage@gmail.com>
- Cc: www-style@w3.org, Tony Chang <tony@chromium.org>
- Message-ID: <CANMdWTsAH+PXc5R5obUGv_AtjbdZD--aQPPNxUhqQ1QhjWTDUA@mail.gmail.com>
On Fri, Aug 26, 2011 at 4:30 PM, Tab Atkins Jr. <jackalmage@gmail.com>wrote: > On Fri, Aug 26, 2011 at 4:09 PM, Ojan Vafai <ojan@chromium.org> wrote: > > The old flex spec required natural numbers for box-ordinal-group and also > > box-ordinal-group is 0 indexed instead of 1 indexed. This lets you avoid > > doing an initial pass before you start laying out. You can layout in > > document order until you hit a non-zero box-ordinal-group. > > The new spec says flex-order is an "integer" and that it defaults to 1. > We > > should either require the flex-order be a positive integer or default it > to > > 0 and require it to be non-negative. > > The change from <number> to <integer> and from default to 0 to 1 were > both accidental. They're changed in the latest ED. > > I'd be okay with limiting it to non-negative numbers, but that means > you can't move things earlier in the order without changing the > flex-order of *everything*. You already have to do passes over all > the children for earlier layout stages; does this actually save us > anything? > I think you are probably right here. As per your email today about the new layout algorithm, we should be able to combine steps 1 and 2 (i.e. collect the flex-order of each child as we compute its hypothetical size). In which case, allowing negative numbers seems worthwhile for the use-case you mention. > (Note - I'm in the middle of a major rewrite of the spec to integrate > the latest decisions and changes. Don't expect it to be stable until > sometime next week. I should probably add a note to that effect in > the spec...) > > ~TJ >
Received on Monday, 29 August 2011 17:51:39 UTC