Re: [xml] more questions on catalogs

On Wed, Feb 22, 2006 at 06:19:37PM -0800, Rick Jones wrote:

  Hey you can't blame me if other OSes are slow to adopt this. 

Well, actually, I am trying to :)  Libxml2 is a multi-platform package.  As 
such, it seems (to me) that if it has a default/hardcoded place where it is 
expecting to see the catalog, it should create that place when the libxml2 
software is installed rather than asking the OS install to do it, or some 
other package.

  Well /etc/ is kind of touchy.

Regardless, while I feel a little funny about it, I'm going to charge 
boldly forth with this added to my Makefile:

# a boatload of assumptions here, but this should create the 
# with our entry in it. we really should provide some overrides in the
# configure script one of these days
        if [ ! -d /etc/xml ]; then mkdir /etc/xml; chmod 555 /etc/xml; fi
        if [ ! -f /etc/xml/catalog ]; then xmlcatalog --create --noout 
        xmlcatalog --noout --add "system" 
"file://$(netperfdir)/netperf_docs.dtd" /etc/xml/catalog

where $(netperfdir) typically ends-up as /usr/local/share/netperf

I found that xmlcatalog --add will not add duplicate entries - thanks for 
that, it made that a bit simpler.

  Fine by me, now it will push the problem onto people managing the integration
of your code within the OS, which is good, the more people having to deal with
this the more pressure it will put onto OS designers to consider /etc/xml case.

 Using a canonical URL for the System identifier is definitely a good 
But portability (of the document) requires a catalog, and portability of 
catalog system is unfortunately not there. I think I did my share to try to
solve this in the area I could have an impact on, but sorry there is 
I can do to get changes in HP-UX or Windows !

All the more reason to have libxml2 installation create the default catalog.

As quixotic as I am, I don't want to flog a dead horse or annoy the libxml2 
gods :) so I'll just drop the matter, having had more than my say on the 
topic :)


BTW, the typscript with the 2.6.23 compiation warnings tripped over the 40K 
limit (it was about 48K) so is sitting waiting for moderator action.

  Okay I will catch that later !


Daniel Veillard      | Red Hat
veillard redhat com  | libxml GNOME XML XSLT toolkit | Rpmfind RPM search engine

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]