- From: Tab Atkins Jr. <jackalmage@gmail.com>
- Date: Fri, 18 Dec 2015 10:47:22 -0800
- To: Simon Pieters <simonp@opera.com>
- Cc: Philip Walton <philip@philipwalton.com>, www-style list <www-style@w3.org>
On Fri, Dec 18, 2015 at 3:05 AM, Simon Pieters <simonp@opera.com> wrote: > On Fri, 13 Nov 2015 18:33:28 +0100, Tab Atkins Jr. <jackalmage@gmail.com> > wrote: >> As Boris says, FF's behavior is correct. `-var(...)` is a function >> named -var, which is not a variable reference. You can't negate >> functions like that *anywhere* due to the way ident parsing works; >> it's not specific to variables. > > But it seems to me you're only likely to want it for var(). Since this > appears to be a pattern people use in the preprocessor world, should we just > support it? It wouldn't be difficult to define a new -var(x) function as > calc(-1 * var(x)), right? Too much of a hack? Slippery slope? I feel like it's too much of a hack. I don't like reinterpreting syntax like this. It implies, for instance, that `+var(...)` is also valid, and presumably a no-op. And as Chris says, authors will expect other numeric functions, if they exist, to support this as well. ~TJ
Received on Friday, 18 December 2015 18:48:08 UTC