Path: utzoo!mnetor!uunet!lll-winken!lll-tis!ames!pasteur!agate!ig!uwmcsd1!bbn!rochester!pt.cs.cmu.edu!sei!sei.cmu.edu!firth
From: firth@sei.cmu.edu (Robert Firth)
Newsgroups: comp.lang.c
Subject: Re: Toy compilers (was Re: gotos)
Message-ID: <5320@aw.sei.cmu.edu>
Date: 3 May 88 19:16:28 GMT
References: <1988Apr8.183815.3187@utzoo.uucp> <449@goofy.megatest.UUCP> <1988Apr24.012322.3951@utzoo.uucp> <504@wsccs.UUCP> <1080@maynard.BSW.COM>
Sender: netnews@sei.cmu.edu
Reply-To: firth@bd.sei.cmu.edu.UUCP (Robert Firth)
Organization: Carnegie-Mellon University, SEI, Pgh, Pa
Lines: 19

In article <1080@maynard.BSW.COM> campbell@maynard.UUCP (Larry Campbell) writes:
>P.S. -	I just compiled a "while (1)" loop on my system, which has one
>	of the wimpier C compilers around.  It generated the right code
>	(i.e., no test instructions, just a jump to the top of the loop).

Same here.  I compiled the BCPL equivalent

	WHILE TRUE DO
	{
	  Thing
	}

and got out

    LA4:
	code for Thing
	BRB LA4

This is so obvious a micro optimisation I find it hard to believe any
compiler won't do it.