Path: utzoo!utgpu!news-server.csri.toronto.edu!clyde.concordia.ca!uunet!cs.utexas.edu!usc!snorkelwacker!spdcc!ima!esegue!compilers-sender
From: preston@titan.rice.edu (Preston Briggs)
Newsgroups: comp.compilers
Subject: Re: Intermediate Representation
Keywords: code, optimize, design
Message-ID: <1990Aug12.212723.1247@rice.edu>
Date: 13 Aug 90 01:57:41 GMT
References: <1990Aug09.180627.18848@esegue.segue.boston.ma.us> <1990Aug9.225859.10175@rice.edu> <1990Aug11.203048.848@mintaka.lcs.mit.edu>
Sender: compilers-sender@esegue.segue.boston.ma.us
Reply-To: preston@titan.rice.edu (Preston Briggs)
Organization: Rice University, Houston
Lines: 23
Approved: compilers@esegue.segue.boston.ma.us

In article <1990Aug11.203048.848@mintaka.lcs.mit.edu> jouvelot@brokaw.lcs.mit.edu (Pierre Jouvelot) writes:

>With the AST approach, the cleanest way to deal with this overabundance of
>control structures is to define a single "generic" loop construct; all the
>control structures can then be trivially desugared (e.g., by the parser)
>into your own version of loop.

I don't see the trivial desuguring by the parser, particularly with an IF -
GOTO form of the loop.

I can see discovering the entire CFG and then doing interval analysis to
discover a nice tree-like grouping of nested control structures a la Sharir
(or others), but you're still punting on irreducible flow graphs.

However, thanks for the interesting post on your design.  It's fun to see
how other people approach problems.

-- 
Preston Briggs				looking for the great leap forward
preston@titan.rice.edu
-- 
Send compilers articles to compilers@esegue.segue.boston.ma.us
{spdcc | ima | lotus| world}!esegue.  Meta-mail to compilers-request@esegue.