"drop-in" DOM replacement for minidom?

[Uche Ogbuji]

I think you may have thought there was animus in my message, but I assure you 
there was none.

> Quoting Uche Ogbuji (uche at ogbuji.net):
> > 
> > My response is usually more like "what are you using XML for a single
> > 30MB file for?"
> 
> I don't think you responded when I asked. :)

I would have liked to.  I'm on so many newsgroups that I pretty mych stagger 
under the load.  My approach to Usenet right now is to do a "python xml" 
search on Google/Deja every Friday and answer what I can from what comes up.  
I'd like to be in closer touch, but such is life.


> http://groups.google.com/groups?th=bf5c3ace18a4e2b6&seekm=mailman.1053714039.24283.python-list%40python.org
> 
> > I've long maintained that when working with XML, modest document sizes
> > is very important, regardless of what tools you're using.
> 
> I think I agree, though some are better suited for large documents
> than others.

True, though since in my experience it is not a linear curve under *any* 
processor, you're most likely to find out that one processor can handle a 30MB 
document, but not a 33MB document.  I've eaten all brands of dog food, not 
just my own, and I haven't found any that have let me be truly cavalier about 
document sizes.  Such is the curse of XML.


> > > My friend also had some more challenging problems. He was running on a
> > > DEC Alpha, I think under Digital Unix, and as a consequence 4Suite had
> > > byte-ordering problems.
> > 
> > 4Suite used to have byte-ordering problems, originally reported under
> > Solaris 9, and also affecting some Mac OS X users.   Those are fixed
> > now.
> 
> As of May this year, 4Suite's cDomlette was having problems for these
> users under DEC Alpha running Digital Unix, when dealing with Unicode.
> I don't know whether those problems have been fixed since;

Nor do I.  The "when dealing with Unicode" is a twist I haven't come across, 
that I remember.


> I only know
> they existed. I've personally never had a problem with cDomlette. If
> you would like, I can put you in touch with the user directly and he
> may be able to do some tests for you. Just to follow up, I'll be
> asking him if he's checked recently, see if he can reproduce it.

That would be helpful.  Thanks.


> > > However, I believe that if you can get 4Suite to compile,
> > 
> > Which I hardly expect to be a problem.
> 
> Not intended as a slight on 4Suite's stability;

I didn't take it that way.  Just being matter-of-fact that compiling shouldn't 
be a problem for folks on most platforms.  I would think that the decision 
point for selection of cDomlette lies on some other branch.


> I use it regularly and
> find it quite solid. This was more in reference to the current
> disclaimer that 4Suite may not always use cDomlette for all platforms;
> sometimes it falls back to a pure-python solution.

Right.


> > Your memories must be from long ago :-)  That API is how it's been for
> > a while.
> 
> I think I last used cDomlette last fall or late last summer. At the
> time, I did not know how to invoke cDomlette directly, but only
> stumbled across a newspost, when looking for ways to tighten-up parse
> performance. The post -- probably from a mailing list -- referred me
> to cDomlette, but I had to go read source code to figure out how to
> use it.

Yeah.  I think my Akara page dates from last Fall and before that cDomlette 
docs were non-existent.  Sorry my effort came late to save you the pain of 
code speleology ('specially in C).


-- 
Uche Ogbuji
Personal:  uche at ogbuji.net                http://uche.ogbuji.net
Work:      uche.ogbuji at fourthought.com	  http://Fourthought.com