Scripting C++ -- Boost.Python vs CORBA vs ???

Craig Maloney cmaloney at physics.ucsb.edu
Wed Feb 27 11:20:04 EST 2002


"David Abrahams" <david.abrahams at rcn.com> wrote in message news:<a5h9qn$a3h$1 at bob.news.rcn.net>...
> "Craig Maloney" <cmaloney at physics.ucsb.edu> wrote in message
> news:3C7BCCA5.604 at physics.ucsb.edu...
> 
> Another promising but incomplete entry, last time I looked (it still
> couldn't represent the full C++ language in its XML). The mailing list
> archives indicate that's still the case, though things may be changing:
> http://public.kitware.com/pipermail/gccxml/2002-February/000029.html
> 
> > This is what I would have guessed would be the "right" way to go about
> > things (having known nothing about the EDG front-end).  Presumably gcc
> > gets the parsing right if it can compile ;)
> 
> Even the EDG parser has bugs. However, if your code compiles with GCC 3.0,
> gccxml should be able to handle it (note that GCC is up to 3.0.3 and 3.1
> will be released soon, so things have improved further since gccxml was
> developed).
> 
> -Dave

My apologies to the group for my ignorance in compiler issues.
It was unclear to me before that the EDG front end *was* used
in commercial compilers.  *grins sheepishly* I had assumed that
it was a tool constructed for independent reasons (e.g. writing
automatic wrapper generators).

This is what I meant when I said that gccxml was the "right" way 
to do the wrapper generation.  Now that my misconceptions have 
been cleared up, I retract my statement about gccxml being the
"right way" as opposed to the PDT's use of the EDG front end.

As Dave points out, since I'm compiling with gcc anyway, and not,
say KAI or Intel, then I'm at a *slight* advantage using gccxml,
all other things being equal.  But now it seems more like a wash 
to me.

Also -- does anyone have any news or thoughts about gccxml or
something else like it being incorporated into the gcc proper?
It seems that since gcc tries to be a "compiler collection" it might
be nice to include things like automatic wrapper generation for
multi-language support?  Is this on the horizon?

Cheers,
Craig



More information about the Python-list mailing list