Re: wysiwig TeX in dia

[James Michael DuPont <mdupont777 yahoo com>]

--- Cyrille Chepelov <cyrille@chepelov.org> wrote:
> Le Thu, Jun 20, 2002, à 02:06:25AM -0700, James
> Michael DuPont a écrit:
> 
> 
> > To be honest, I have not tried CVS on XML files. 
> I
> > mean not the DIFF. Do you mean for DIFFING the dia
> > output? Do you want to extend CVS proper or
> transform
> > the XML into a DIFF freindly output?
> 
> For the specific case of dia files into a CVS
> archive; been there, done
> that. 
What have you done?

>As far as CVS is concernet, these files could
> have been MS Word files,
> it would have done no more good. I think it would be
> wonderful if cvs was
> aware of the concept of XML files; if it was able to
> uncompress them on the
> fly, so that an uncompressed file would be deemed
> identical to the same
> file, compressed. 
Yes that would be interesting. But a makefile that
would transform them into diffable XML and a cvs
ignore would be a good start.
> 
> Bringing the reasoning to its end, in fact it's the
> current diff/patch tool
> which ought to know about XML files (and then use a
> suitable and humanly
> readable (a.k.a "Torvalds-editable") representation
> of the difference
> between two XML trees).
Yes, I have to say that that is a bit too big for me
right now. If you want me to look into it for DIA
specifically I will, but I am not ready to do anything
on that scale yet.

> 
> > Fair enough. I think the right place is in LIBXML,
> to
> > add in some opaque in-memory pointer system and
> > callbacks.. I will look into that first. 
> 
> Good luck dealing with the libxml guys. 
Well, it would be a patch for this project to begin
with, but yes a general persistance layer on the
libxml level with versioning and diffing of objects.

I dont want to fight too many front here. 
I have too many already!

>But yes, I
> think the solution to
> your feature request may be achieved more
> efficiently by breaking the
> problem into using a database back-end for libxml on
> one hand, and
> maintaining the associativity between the libxml
> tree and dia's diagram tree
> on the other hand. 
Yes and the ability to diff them in memory!

>The first part of the solution
> may very well prove itself
> very useful to other projects.

Great that we agree!

mike

=====
James Michael DuPont

__________________________________________________
Do You Yahoo!?
Yahoo! - Official partner of 2002 FIFA World Cup
http://fifaworldcup.yahoo.com