AST to XML

["Jeremias A. Sauceda" <pi314@iastate.edu>]

Hello,

I am a research assistant at Iowa State's Software Reengineering Lab.
Last semester I created a program to convert Edison Design Group's
Intermediate Language files (an Abstract Syntax Tree) into XML.
Unfortunately their C/C++ front end does not support many gcc
extensions.  I would like to use GCC's front end instead, since it
obviously support gcc extensions and is open source.  I desperately need
gcc extensions since I am analyzing the Linux Kernel, which is very
biased to gcc.

There are two reasons why EDG's front end was chosen.  The abstract
syntax tree it creates is a complete representation of the source code,
no optimizations are done.  We received the full source code (after
signing an NDA).  We need a pristine representation of the source code
since we are doing high level analysis of code; for automatic MPI
parallelization, slicing, etc.

By glancing through tree.h and tree.def I think that most of the
information I need is present except for line and column numbers.  I
need at least the block start and end points, preferably those of
statements, optimally of all expression terminals.  Is this already in
the tree?  What do I have to do to add it?

Ultimately I would like to create an XML backend to gcc.  So instead of
an object file I get an XML AST.  What do you think?

Thanks,

Jeremias Sauceda