Re: [css-variables] Initial value of variables

Why not treat un-initialized variables as a null value? Empty string is
just a poor-man's null that I think will come back to bite us. We've just
added a null initial value for variables to Sass and it really cleaned up a
lot of edge-case scenarios like the one mentioned here.

Chris

On Wed, May 30, 2012 at 11:51 PM, François REMY
<fremycompany_pub@yahoo.fr>wrote:

> I'm fine with it but I don't understand why an empty variable is invalid.
>
> Background: black $url;
>
> This declaration is valid if we replace the 'variable' by zero token.
> ------------------------------
> De : Tab Atkins Jr.
> Envoyé : 31/05/2012 03:13
> À : www-style list
> Objet : [css-variables] Initial value of variables
>
>
> The initial value of a variable is always invalid, regardless of how
> you try to use it - you need to explicitly provide a variable property
> to get a valid variable.  Thus, you don't actually *need* an initial
> value, since it will never be used.  However, operating under the
> assumption that I needed to provide *something*, I've defined the
> initial value to be the keyword "invalid", except it's also always
> invalid.
>
> Divya recently reported that she found this boggling.  I agree it's
> not great.  Here's an example from the spec:
>
> <div>
>   <p>Sample text</p>
> </div>
> <style>
> p { var-foo: invalid; }
> div,p { font-family: $foo; }
> </style>
>
> In this example, the "foo" variable is an invalid variable at the time
> the DIV element references it, because the ‘var-foo’ property still
> has its initial value. This causes the DIV's ‘font-family’ property to
> compute to the initial value for ‘font-family’.
>
> On the other hand, the P element defines an explicit value for the
> ‘var-foo’ property. Its ‘font-family’ property thus references a font
> named "invalid".
>
> In a similar vein, it means that assigning a property's value to
> itself via JS, which should be a no-op, flips the variable to being
> valid with the value "invalid":
>
> el.style.varFoo = el.style.varFoo
>
>
> Is it acceptable to instead just define that the initial value of a
> var property is nothing?  That is, if you ask for it to be serialized,
> you'll get "" out of JS.  You can't actually use this value, since the
> variable is guaranteed to be invalid, and you can't actually *create*
> this value manually either, since the <value> production has a + in
> it.  This would avoid the weirdness discussed above.
>
> ~TJ
>
>