Re: [css2.1] Comments within url() tokens

"Linss, Peter" <peter.linss@hp.com> wrote:
[...]
> This literally means that comments can appear anywhere (except within
> a token)
[...]
> Where this gets interesting is that both 4.1.1 and Appendix G.2 define
> "url(http://example.com/)" as a single token, which means that
> comments are not allowed within.

Right.  Everything I said presumed that, as url(...) is one token, the
only way comments could be allowed within was if the lexical rules
explicitly said so.

> I know that Gecko tokenizes this as
> three separate tokens: "url(", "http://example.com/", and ")"; a
> FUNCTION token, a URL token and a symbol token.

Yes.  And, at present, it allows comments in between those three tokens
under some circumstances, but not all the things that could be read as
comments are.  I'm pretty sure the current behavior is not desirable, 
but before I go touching it at all I want to be sure what we ought to
be doing.

I'm attaching a probe of the behavior of comments within url().  I can
only easily test with Gecko (3.0 and trunk), Linux Opera 9.52, and
Windows IE 7.0.6001.18000.  Their behavior is:

Gecko: Comments are allowed on either side of a quoted URL.  Comments
are also allowed before an unquoted URL, but only if there is a space
between the comment and the URL.

Opera: No comments are allowed within the parentheses.

IE7: Comments are allowed except where they follow an unquoted URL with
no intervening spaces.

The HTML file relies on two external files, allowed.gif and
forbidden.gif (also attached) - these are just green and red single
pixels.  I don't use data: URLs because I am curious what happens in
old browsers that don't support that.

In the test, cells where the background color is green are cases where
comments are allowed (that is, the uri-token quoted resolves to
"allowed.gif"); if the background is any other color, comments are not
allowed in that case.  Some cases will be red and others white - all
browsers seem to make a distinction between "this is nonsense, I'm
falling back to the less specific rule" (red) and "that uri-token
resolves to a file that doesn't exist, so you get no background image
at all" (white).

> It could be relatively safe to allow comments within a url() function
> so long as there is whitespace between the comment and the URL part
> (unless the URL is quoted), something like:
> 
> {U}{R}{L}"("({w}|{comment})*{string}({w}|{comment})*")"
> {return URI;}
> {U}{R}{L}"("{w}({comment}{s})*{url}({s}{comment})*{w}")"
> {return URI;}

I'm not eager to match IE in this case; I'd be fine with your proposal
or with allowing comments only in quoted URLs; I would rather not
forbid comments in conjunction with quoted URLs.

zw