Re: [boost] boost and XML

From: Phil Nash (phil.nash.lists_at_[hidden])
Date: 2002-06-18 04:06:06

Thanks for getting back, Dietmar,

> Unfortunately, I have made real progress:

Did you mean, "have not"? I take it from your following comments that you
did.

> All I did since then was to
> tweak some stuff while making it compile with other standard C++
> libraries (mostly my own implementation). From a functionality point of
> view the latest version is the one I made available in the files section.
>
> > I for one would be very interested in it - and I know others would.
>
> Unless you need fancy DTD stuff you can actually use it: It should
> work just fine. My only trouble was in the area of this good-for-nothing
> DTD parsing, in particular with the nasty lot of exceptions to the rule.
> Even DTD parsing is probably just fine - except if you have errors in
> your DTD in which case these may go undetected.

Well, the XML savvy code I am writing at the moment does not need any "fancy
DTD stuff" ... yet!
We will probably be adding schema validation at some point, though - are
they supported yet?

> > I have just written a wrapper for the xerces library to std c++-size it
> > (and give it iterators).
>
> I think that xerces could be used as an implementation underneath the
> iterator interface I'm envisioning: the SAX parsers apparently have
> functions which parse the next item. These functions are used between
> the callback calls to do the actual parsing. I haven't looked into this in
> detail, however: I just found functions like this when I looked through
> other XML implementations.

If I understand what you are suggesting here correctly, I believe that is
what I intended to down a little down the line. So far I have just wrapped
the Xerces DOM hierarchy, but it does seem like a logical step to use the
Xerces SAX parser at the back end (which I'm fairly sure Xerces DOM does
itself - not looked at the code yet).

> > I might even be persuaded to shared the workload of your effort...
>
> This would be great! However, I have to admit that the code is rather
> messy which I claim is due to the fact that XML, in particular XML's
> DTD, is messy (raw XML is pretty simple). Since I'm actually using my
> XML parser in an internal project and since there seems to be some
> interest in the approach I should really bring this stuff into a clean
state.

Ok, well my offer still stands, although I can't guarantee how much time I
can spend on it. I've been interested in this for a while because I dislike
using Xerces, which is pretty much de facto, in projects that are otherwise
more in the standard c++ style. I know that following the DOM and SAX
interfaces closely do not lead to a strong standard C++ style, but there is
still room my alternative iterator-based interface (maybe using the BGL - I
keep meaning to look into that).
Let me know, probably offlist) if you get a chance to pick this up again -
or if there is anything you particularly feel I could look at in the
meantime. When I get some time I'll have a closer look at what you have
already done (working 24/7 at the moment :-( ).

Regards

[)o
IhIL..

P.S. If you contact me off list, drop the word "lists" from my email
address - giving: phil.nash_at_squareseven.com.