Re: typing the input from within xslt

>>>>> "Owen" == Owen Rees <owen.rees@xxxxxx> writes:

    Owen> --On Tuesday, November 04, 2008 15:14:49 +0000 Colin Paul Adams wrote:
    >> I agree with both those statements.
    >> 
    >> It is still non-conformant, and I don't think you should do
    >> this.

    Owen> As far as I have been able to discover, RFC1738 is still the
    Owen> one that contains the current definition of the file URL
    Owen> scheme.

Yes, i think so too.

    Owen> The syntax explicitly allows both "?" and "=" in 'fsegment'

    Owen> from which fpath' is formed. The file scheme does not have a
    Owen> query string because what looks like a query string
    Owen> according to the generic URI syntax is actually part of the
    Owen> file path.

Ouch! That means you can have URIs that contradict the generic syntax:

file:///top_level_directory?perhaps/name.xml

In which case I was wrong (maybe).

It goes on later to say:

   A file URL takes the form:

       file://<host>/<path>

   where <host> is the fully qualified domain name of the system on
   which the <path> is accessible, and <path> is a hierarchical
   directory path of the form <directory>/<directory>/.../<name>.

which might be interpreted to mean that ? and # are allowed as part of
the component names, and NOT to indicate query parameters and a
fragment id.

in which case I wasn't wrong.
This is surely the case for the ftp scheme, since the optional
;type=<filetype> comes at the end, so I think this is the intended representation.

    Owen> There is scope for some chaos there ... 

and almost everywhere with file URIs. I remember there has been at
least one attempt to just document existing practise, which proved to
be too difficult.

    Owen> (how many relative URI resolvers have the correct special case for file
    Owen> scheme?)

Close to none, I should think.
-- 
Colin Adams
Preston Lancashire