W3C home > Mailing lists > Public > public-xml-processing-model-wg@w3.org > July 2007

Re: p:map

From: Norman Walsh <ndw@nwalsh.com>
Date: Wed, 18 Jul 2007 12:48:46 -0400
To: public-xml-processing-model-wg@w3.org
Message-ID: <87tzs1mzdd.fsf@nwalsh.com>
/ Innovimax SARL <innovimax@gmail.com> was heard to say:
| I think we should at least say that the resolution of name would be
| consistent through all the pipeline (I say the resolution, not the
| result of the get)
|
| That means :
|
| If I have a "toto.xml" in the xinclude and a "toto.xml" in the in a
| p:document and a "toto.xml" in the doc() function of my xslt what can
| I expect ?

You can expect the implementation to tell you what to expect, hence
"implementation-defined" in my earlier text (as opposed to
implementation-dependent).

But I don't think the spec can say what to expect in all cases for all
implementations.

I can't imagine an implementation going out of its way to give you
different answers anyway. But maybe the XSLT step doesn't have a
resolver, who knows.

All of this is of course the easy case. The hard cases is when there's
an XSLT step in the pipeline that generates a document that has the
base URI "toto.xml". But I don't think we're going to solve that in V1.

                                        Be seeing you,
                                          norm

-- 
Norman Walsh <ndw@nwalsh.com> | The important thing is not what the
http://nwalsh.com/            | author, or any artist, had in mind to
                              | begin with but at what point he decided
                              | to stop.--D. W. Harding

Received on Wednesday, 18 July 2007 16:48:54 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:21:53 GMT