LJ not correctly parsing <link... > tags.

Mon Jul 4 08:47:55 PDT 2005

Brad Fitzpatrick wrote:
> ...
> 
> It also doesn't resolve relative URLs, but I'm not sure that it should or
> not.  Should we just say full URLs are required?  Either way it needs to
> be documented in the spec.

I was almost onto writing a complaint about that. Yes, urls _MUST_
be absolute, because there are half a dozen rules for constructing
full URLs from relative paths in HTML:

- <a href="./standard/relative/path/"
- <base target="...
- <meta http-equiv="Content-Location" and the real HTTP header
- <html xml:base="...
- <a href="//server.name/..."
- <a href="http:../../subdir/./././f.htm"

If the spec selectively only allowed for the standard case, it
wouldn't look to me as actually being all too HTML-compliant.
OTH it's not too much asked from users to provide absolute URLs,
while at the same time this simplifies consumers considerably.