- From: Tab Atkins Jr. <jackalmage@gmail.com>
- Date: Tue, 4 Mar 2014 16:43:27 -0800
- To: Matt Rakow <marakow@microsoft.com>
- Cc: "www-style@w3.org" <www-style@w3.org>
On Tue, Mar 4, 2014 at 4:37 PM, Matt Rakow <marakow@microsoft.com> wrote: > Thanks, agreed that snap-list as a function isn't necessary since it's just wrapping a list of lengths. I've updated the spec to just use the list of lengths instead of a snap-list function. However, based on the guidelines that Fantasai forwarded [1] I'm wondering if these should be comma-separated (per "Lists of parallel items are comma-separated"). Hm, yeah, you're probably right. I think it should use commas. > The "repeat" syntax you suggest below seems difficult to interpret at face-value to me. Rather than suggesting a single interval is jointly established by the three tokens, it looks more like 3 orthogonal arguments (e.g. drawing parallels to the shorthand for "background"). Is there some value that already establishes this syntax as precedence? Otherwise maybe something like one of these? > > repeat(<length>? <length>) > <length>? repeat(<length>) Yeah, using a repeat() function works fine too: <length># | [ <length># ,]? repeat(<length>) | elements I keep allowing one or more lengths preceding the repetition, just in case you want to set up more than one initial points before the repetition kicks in. That seems reasonable to me. > (Also, I think a lot of this will change again as I incorporate the feedback to merge x/y into a single property with coordinate pairs) Yeah, probably. ^_^ ~TJ
Received on Wednesday, 5 March 2014 00:44:14 UTC