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

[Craig Maloney]

Hi David.

Since I'm not a "computer guy" per se, I was wondering if you could 
clear up some more of my confusion.  I don't know much about how 
compilers work, but I would presume that any compiler (eg g++) would
create some internal representation, similar to the .cil file, after
parsing but before compiling.  

Presumably the EDG front end is geared toward producing this information
in a format more suited to being used by the "wrapper generators" in contrast 
with a compiler which produces this information only for itself and with 
the sole goal of compiling an object file?  

What I'm getting at is that I'm confused as to why people wouldn't have
used information from the compiler *itself* to automate the process of
wrapper generation.  Obviously the compiler writer people would have to
collaborate with the wrapper generator people... but there must be technical
obstacles?

Cheers,
Craig 

"David Abrahams" <david.abrahams at rcn.com> wrote in message news:<a5ekpq$n83$1 at bob.news.rcn.net>...
> I've been interested in systems for automatically wrapping C++ and
> generating documentation for a while, so I downloaded the PDT (the parser
> toolkit used by SILOON) and tried it on one of my source files. After lots
> of hacking, I was able to get the EDG front-end component to parse my source
> and generate an intermediate language (.cil) file, but the PDT executable
> which is supposed to interpret the .cil and produce a high-level description
> of the code crashed with an assertion:
> 
> Assertion failed: ptr->assoc_template, file c:\documents and
> settings\bertie\desktop\finalwinstuff\taucpdisp\src\disp-tau\il_display_tau.
> c, line 3047
> 
> Since the PDT doesn't come with source, there was no way for me to debug/fix
> the problem.
> Furthermore, the PDT developers don't seem to want to be reached. Their
> contact page (http://www.cs.uoregon.edu/research/paracomp/pdtoolkit/) says:
> 
>     To contact the developers of PDT, please use the "Comments" section in
> the download page.
> 
> But of course, there is no "Comments" section on the download page. :=(
> 
> -Dave
> 
> "Craig Maloney" <cmaloney at physics.ucsb.edu> wrote in message
> news:3C7A751D.1080703 at physics.ucsb.edu...
> > Hi Craig.
> >
> > Since your site specifically lists g++ 2.95 and NOT 3.x, am I to infer
> > that there will be problems with the new ABI?  -- Or any problems that
> > will, in practice or principle, cause conflicts with g++ 3.0?
> >
> > *Holds breath... and crosses fingers...*
> >
> > --Craig
> >
> > rasmussn at lanl.gov wrote:
> >
> > > On Wednesday, February 20, 2002, at 04:48 PM, David Abrahams wrote:
> > >
> > >>> As I understand it, the SILOON project uses a commercial parser known
>  as
> > >>> PDT from uoregon.  Have you, or has anyone else had any experience
>  with
> > >>> the SILOON/PDT project?  On the face of it, my guess would be that you
> > >>> would be somewhat skeptical of their claims?
> > >>
> > >>
> > >> I haven't had any experience with it. I just found that site. My
> > >> skepticism
> > >> just dropped by a factor of 8. If they are really using the EDG
> > >> front-end as
> > >> they claim to, then I am sure the parser is excellent. I do have to
> > >> wonder
> > >> who is paying for the FE. The issues of dealing with the semantics of
> > >> what
> > >> they've parsed remain, but these guys at least seem to be doing the
> > >> parsing
> > >> part right.
> > >
> > >
> > > It is my understanding that an educational or research license to the
>  EDG
> > > front end is fairly easy to obtain.  The University of Oregon has
>  obtained
> > > the right to distribute the EDG front-end as binaries and does so for
> > > several
> > > platforms in the Program Database Toolkit (PDT).  So for anyone wanting
> > > interface level information from C++ or Fortran 90 source files, the PDT
> > > is an excellent tool.  We use it in several projects.
> > >
> > > See http://www.cs.uoregon.edu/research/paracomp/pdtoolkit/ and
> > > http://www.acl.lanl.gov/siloon/ for more information.
> > >
> > > Craig
> > >
> > >
> >
> >