Specification changes about myth

[mikemanger]

#39 touched on the same issue with the color() function - myth got lucky that time, however, you can't just hope that the specification doesn't change (or even get dropped). The draft specification warnings are there for a reason!

[ianstormtaylor]

wow the spec authors are really making it hard to argue against the "css version of these features is super ugly" crowd :p

"very unlikely" seems like an overstatement though. we can add a more prominent "draft" warning, but Myth's entire goal is to let us make a tradeoff in a bit of stability (CSS is literally 100% backwards compat) for years of waiting for features that authors have been asking for forever. Sass and LESS are also not 100% backwards compat, they have their own deprecation notices, so it doesn't seem like an unreasonable, unprecedented stance to me

waiting for things to be so solid they are guaranteed not to change is what we want to avoid. sometimes the specs will change, but i don't think it's unreasonable in this case to get caught by a draft that's about to land ASAP in the latest Firefox in it's (now) old form, especially since a bunch of the replies to that thread mention how it's super late to be making the change

whatever the decision ends up being here, you're right it looks like the -- prefix, we could even backwards compat so that both variants are usable until a major version bump comes along. and even then people could use Myth as a Rework plugin alongside the old plugin if they absolutely need the backwards support

[mikemanger]

I think in this case "very unlikely" is an understatement - your site will look terrible if you used var- to define key elements (same goes for calc() which is widely implemented but could get axed from the spec)

I understand wanting to implement features but W3C can do big changes to seemingly stable specs (flexbox comes to mind). A simple tag or key saying if it is WR, CR or PR would probably make things clearer and not interfere with the current site design too much, could maybe incorporate it with the current link to the specification (if you can't tell I'm not a designer 👀).

caught by a draft that's about to land ASAP in the latest Firefox in it's (now) old form

I might be wrong on this but as far as I can tell Firefox 29 has the variables feature marked as 'enabled for non-release builds only' so you will need to enable a flag in the final release next month (I think this is the new preferred way of doing vendor prefixes).

[sjonnet19]

I've already submitted a pull request with the new sytax.

#51

[ianstormtaylor]

Thanks @sjonnet19 !