Re: [xml] Support for Python

[Dave Kuhlman <dkuhlman cutter rexx com>]

On Mon, Jan 28, 2002 at 03:35:34PM -0500, Daniel Veillard wrote:
>  On Thu, Jan 24, 2002 at 04:57:39PM -0800, Dave Kuhlman wrote:
>  > I've been thinking about ways that libxml (and libxslt) can be used
>  > to provide XML support for the Python programming language.
>
>    Right, me too :-)
>
>
>  >    We should consider providing support in the following areas:
>  >   
>  >      * Support for the DOM interface built on libxml (or gdome?).
>
>     I would split it into:
>       - Support for the tree interface built on libxml
>       - Support for the DOM2 api built on gdome2

Good idea.  Divide the problem into managable parts.

In two or three more days, I hope to make available what I've done
in the way of SWIG wrappers for the definitions in tree.h.  Then,
we can decide whether that is a worthwhile starting point or not.

>  >      * Support for the SAX interface built on libxml.
>  >      * Support for XSLT built on libxslt. (Possibly a discussion for the
>  >        libxslt list.)
>
>
>    Yep, let's keep XSLt separate for a bit.

Sounds good.

[snip]

>    SAX support should be close to trivial. I expect something similar
>  to your current interface but also allowing a compatible use with
>  the xmllib and sgmlop interface where the callbacks are just
>      - close
>      - getmethodname
>      - data
>      - start
>
>      - end
>      - handle_entityref

I'll look into this.  I believe that the interface is defined (more
or less) in PyXmlversion/doc/xml-ref.txt.  I'm also assuming (until
proven wrong) that pyexpat implements this interface and that I can
discover it by studying the implementation of pyexpat.

I wonder, is there a minimal interface that each driver must
implement?  I believe that we will want to implement that basic
interface _plus_ a few extensions, e.g. (1) request validation, (2)
turn individual event handlers on an off during the parse.  I'll
try to look into this, too.

>     Should be trivial to be able to handle both and would allow very easilly
>  migration of existing code,
>
>  >    Creating a parser driver for PyXML built on libxml seems like a very
>  >    good idea. There are several benefits to be gained from doing so:
>
>     Since I don't know PyXML, I will abstain from commenting on this ATM.
>  seems this should be trivially implementable with just glue python code
>  on top of the sAx interface.

My understanding is that other support in PyXML (e.g. DOM) is built
on top of the SAX drivers.  For example, when you build a DOM tree,
a SAX driver is used (although it seems a bit tricky to control
which driver is selected).


>  >   Additional Notes
>
>     Isn't Python2 internationalization layer based on UTF16 strings ?
>  libxml/libxslt uses UTF8 . Is there any gain of trying to follow the
>  Python2 conventions ? Is there any risk by staying with UTF8 strings
>  seens as usual python strings ?
>
>     Converting to/from UTF16 all the time would be a killer.
>
>
>  Daniel
>
>  --
>  Daniel Veillard      | Red Hat Network https://rhn.redhat.com/
>  veillard redhat com  | libxml Gnome XML XSLT toolkit  http://xmlsoft.org/
>  http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/



-- 
Dave Kuhlman
dkuhlman rexx com
http://www.rexx.com/~dkuhlman