Path: utzoo!utgpu!watmath!clyde!ima!compilers-sender
From: think!compass!worley@EDDIE.MIT.EDU (Dale Worley)
Newsgroups: comp.compilers
Subject: Interprocedural dataflow analysis question
Message-ID: <3329@ima.ima.isc.com>
Date: 13 Feb 89 15:11:20 GMT
Sender: compilers-sender@ima.ima.isc.com
Reply-To: think!compass!worley@EDDIE.MIT.EDU (Dale Worley)
Lines: 36
Approved: compilers@ima.UUCP


   From: encore!lsuc!array!len@ai.toronto.edu (Leonard Vanek)

   I am specifically interested in
   obtaining an EXACT (rather than conservative) solution for the values
   (or expressions) killed by a procedure call.

   [Good luck, sounds like the halting problem to me.  -John]

Yeah, it's the halting problem.  Consider (in C):

	int x(int i)
		{
		if (f(i))
			global_variable = 0;
		}

All you have to do is determine if f(i) can ever return true.

A really interesting aspect of interprocedural analysis arises if
functions are first-class values (for instance, in Scheme), so to do
interprocedural analysis you have to first determine which functions
(lambdas) a given call might activate.  Of course, inside a lambda the
potential values of its bound variables depends on which calls might
activate it, and those bound variables might be used as the functions
of calls...  See "Control Flow Analysis in Scheme" by Olin Shivers
(Proceedings SIGPLAN Conference on Prog. Lang. Design and Impl., June
1988).

Dale
[From think!compass!worley@EDDIE.MIT.EDU (Dale Worley)]
--
Send compilers articles to ima!compilers or, in a pinch, to Levine@YALE.EDU
Plausible paths are { decvax | harvard | yale | bbn}!ima
Please send responses to the originator of the message -- I cannot forward
mail accidentally sent back to compilers.  Meta-mail to ima!compilers-request