Related articles |
---|
Source to source compiling casper@snakemail.hut.fi (Casper Gripenberg) (1994-09-11) |
Re: Source to source compiling eanders+@CMU.EDU (Eric A. Anderson) (1994-09-12) |
Re: Source to source compiling norman@flaubert.bellcore.com (1994-09-14) |
Re: Source to source compiling mabp@bga.com (1994-09-13) |
Re: Source to source compiling f1rederc@iia.org (1994-09-13) |
Re: source to source compiling ghiya@vani.cs.mcgill.ca (1994-09-18) |
Re: Source to source compiling bernecky@eecg.toronto.edu (Robert Bernecky) (1994-09-18) |
Re: Source to source compiling toconnor@vcd.hp.com (1994-09-18) |
Re: Source to source compiling pjj@cs.man.ac.uk (1994-09-16) |
Re: Source to source compiling rgg@tidos.tid.es (1994-09-19) |
[2 later articles] |
Newsgroups: | comp.compilers |
From: | mabp@bga.com (Michael Parkes) |
Keywords: | optimize |
Organization: | Real/Time Communications - Bob Gustwick and Associates |
References: | 94-09-031 |
Date: | Tue, 13 Sep 1994 04:39:16 GMT |
: I am currently working on a source to source compiler. The source
: language is a kind of state langage. A program containes 1-n states. All
: states have the same kind of structure, ie. in the beginning you initilaze
: variables, then comes some code to execute and at last a bunch of
: conditional jumps to other states.
Hum,
I think you are using the wrong tools. My suggestion is to convert
your language into a form of flow graph and then try to de-compile
the flow graph back into source code. This will be pure torture in
C or C++. I suggest you use grammar tools with very good symbolic
manipulation support.
Mike Parkes.
--
Return to the
comp.compilers page.
Search the
comp.compilers archives again.