[antlr-interest] Re: Seperating Grammar and Actions..

[mzukowski at yci.com]

CIL's approach is not for translation so much as for a certain set of
transformations, particularly those transformations which would not ever be
propagated back into the original source code.  Their application CCured,
which does pointer analysis and boxing, is the type of thing that it's good
at.

You're right, there is a whole class of translation problems that would not
be amenable to that approach.

Monty

-----Original Message-----
From: micheal_jor <open.zone at virgin.net> [mailto:open.zone at virgin.net]
Sent: Monday, January 13, 2003 4:24 PM
To: antlr-interest at yahoogroups.com
Subject: [antlr-interest] Re: Seperating Grammar and Actions..


--- In antlr-interest at yahoogroups.com, mzukowski at y... wrote:
> Getting into the deeper issue of AST design, I highly recommend 
studying the
> CIL project.  http://manju.cs.berkeley.edu/cil/.  They parse C and 
GCC and
> MSVC into a subset of C.  All loops are reduced to one form, etc.

The extent of the C->CIL transforms does seem a bit extreme in many 
places (just flicked thru the info on the site).

I'd be worried that any code that results from any source-to-source 
translator (e. C-to-C++ or C-to-C#) built on this would be just 
as "alien" as comparing some of the C-source and CIL-output fragments.

Am I missing some bigger picture here?

Micheal


 

Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/ 


 

Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/