This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Linkage of GPLed GCC to Closed Source via XML or Perl
- From: James Michael DuPont <mdupont777 at yahoo dot com>
- To: Robert Dewar <dewar at gnat dot com>, bryce at waitaki dot otago dot ac dot nz, dnovillo at acm dot org
- Cc: Joe dot Buck at synopsys dot com, gcc at gnu dot org, mdupont777 at yahoo dot com
- Date: Sun, 3 Mar 2002 23:27:09 -0800 (PST)
- Subject: Re: Linkage of GPLed GCC to Closed Source via XML or Perl
--- Robert Dewar <dewar@gnat.com> wrote:
> Unparsing capabilities are enormously useful, >both
or debugging the compiler and for the >programmer
understanding enerated code ).
Here are some more supporting facts:
The cbreeze
(http://www.cs.utexas.edu/users/lin/cbz/)
and sage++
(http://www.extreme.indiana.edu/sage/overview.html)
projects both have such deparsers tools.
Sage has a class model for the ast that allows you to
contruct optimised code for parallel computing.The
OpenC++ also allows for construction of code on the
ast level and then uparsing it.
The Sage compiler supports restructuring and
de-parsing. This sounds like the AST optimiser.
This will eliminate the need to extract the code out
into XML/DB before using it.
With that unparsing tool, most of the meta-programming
and introspection code that I am interested in can be
handled.
The only thing is, for the end user of a introspection
system you will need to present them an interface into
the AST.
Would be interesting to look into a set of functions
that allow for user-specified ast transformations
during the compiler phase?
Unparsing the data, or dumping the data in code that
will create the data is a great feature of perl. Java
has tools to allows even serialization in XML using
introspection.
mike
=====
James Michael DuPont
__________________________________________________
Do You Yahoo!?
Yahoo! Sports - sign up for Fantasy Baseball
http://sports.yahoo.com