[XML-SIG] Error 404 and xml.dom.ext.reader

[Alexandre]

On Wed, Aug 14, 2002 at 04:42:04PM -0600, Uche Ogbuji wrote:
> > On Tue, Aug 13, 2002 at 09:38:51AM +0200, Alexandre wrote:
> > > On Mon, Aug 12, 2002 at 05:50:06PM +0200, Martin v. Loewis wrote:
> > > > Mike Olson <Mike.Olson@fourthought.com> writes:
> > > > 
> > > > > I think it is a good idea.
> > > > 
> > > > I agree. Alexandre, would you like to implement that change?
> > > 
> > > I'll give it a try today. 
> > 
> > OK, I've got something up and running, which runs nicely with
> > regrtest.py. I'll be committing it very soon. 
> > 
> > The changes are mainly on the line of changing urllib.urlopen to
> > urllib2.urlopen, and urllib.basejoin to urlparse.urljoin. 
> 
> Hmm.  Here be dragons.

;o)

> There are subtle but important differences between the behavior of  
> urllib.basejoin and  urlparse.urljoin, especially with file URLs in Windows.  
> I don't remember the full details off-head but I think Tom Passim and Mike 
> Olson had detailed messages on this in the past.

If anyone has an approximative date for this one, I'd be glad to go and
have a look in the archives.
 
> I would suggest keeping urllib around just for basejoin, since it seems we've 
> reached, after a lot of tinkering,  a balance where the fewest number of 
> platform users scream  :-)

Sounds fair. I'll do the changes today and commit them. 
 
Alexandre Fayolle
-- 
LOGILAB, Paris (France).
http://www.logilab.com   http://www.logilab.fr  http://www.logilab.org
Narval, the first software agent available as free software (GPL).