C++ Front End for S-S Translator?

greened@eecs.umich.edu (David Greene)
25 Jan 1997 22:15:27 -0500

          From comp.compilers

Related articles
C++ Front End for S-S Translator? greened@eecs.umich.edu (1997-01-25)
Re: C++ Front End for S-S Translator? jlilley@empathy.com (John Lilley) (1997-01-26)
| List of all articles for this month |
From: greened@eecs.umich.edu (David Greene)
Newsgroups: comp.compilers,comp.compilers.tools.pccts
Date: 25 Jan 1997 22:15:27 -0500
Organization: University of Michigan EECS
Summary: Looking for a C++ front end suitable for a source-source translator
Keywords: C++, C, translator, question

I'm looking for a C (and eventually C++) front end that generates
information suitable for doing a source-source translation. Ideally,
the information would be in an AST or similar form. It would also be
nice if type information were included (class members as well as the
declaration information), but if a symbol table is used, I can
probably extract the information.


I've been working with the C++ grammar packaged with PCCTS (with the
additions by John Lilley), but the symbol table provided does not
carry much information and there is no semantic checking, error
reporting, etc. If I can avoid work that someone else has already
done, I would be very happy indeed.


So far, I've found pointers to Eli (at the University of Colorado
Boulder), cppp (at Brown University), Sage++ (at Indiana University)
and EDG's C++ Front End. I've looked at lcc and gcc, but their
intermediate forms are too low-level (if's converted to branches,
etc.).


It seems that cppp or EDG's product would be the best bets, but we
can't justify spending money right now, so EDG is probably out. Does
cppp carry most of the information in the source (including class
declarations, definitions, etc.), and if not, is it easy to modify?
Will Eli or Sage++ provide a (relatively) straightforward way to
generate a front end suitable for source-source translation?


If anyone has a PCCTS C++ grammar (especially the packaged NeXT
version) with the appropriate symbol table and semantic checks (though
I could add those), that would be great, as I've already annotated
most of the packaged grammar with the AST constructors (well, for most
of the C subset, anyway).


Thanks for any help!


                                                                                            -David
--


Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.