Path: utzoo!dptcdc!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!ucbvax!pasteur!cory.Berkeley.EDU!talvola
From: talvola@cory.Berkeley.EDU (Erik Talvola)
Newsgroups: comp.lang.c
Subject: Microsoft C v5.1 bug
Message-ID: <12502@pasteur.Berkeley.EDU>
Date: 16 Apr 89 21:25:12 GMT
Sender: news@pasteur.Berkeley.EDU
Reply-To: talvola@cory.Berkeley.EDU (Erik Talvola)
Distribution: usa
Organization: University of California, Berkeley
Lines: 40

I noticed that someone else posted a Microsoft C bug, so I thought I'd
post my favorite one.  I haven't yet sent this to Microsoft, because all
my registration stuff is at home (while I'm at school...), but maybe
someone from Microsoft will see it (unlikely).  

The program:
---------------------------------------------------------
main(int argc, char **argv)
{
   unsigned char workbm[16][16];
   unsigned char colors[16][16];
   unsigned char pic[8][4][8][2];
   int   i,j,k;

   pic[i][(j>=8)*2][j % 8][0] += workbm[k][j];
   pic[i][(j>=8)*2][j % 8][1] = colors[k][j];
}
---------------------------------------------------------
Yes, it looks hideous, but look at the result of a compilation with
Compact or Large model...

C:\UC> cl -AC editbug.c
Microsoft (R) C Optimizing Compiler Version 5.10
Copyright (c) Microsoft Corp 1984, 1985, 1986, 1987, 1988. All rights reserved.

editbug.c
editbug.c(8) : fatal error C1001: Internal Compiler Error
		(compiler file '@(#)code.c:1.46', line 393)
		Contact Microsoft Technical Support

----------------------------------------------------------

The problem does not exist under Quick C, only Microsoft C, and only appears
when Compact or Large model is used.  Has anyone else ever had Microsoft C
give them an Internal Compiler Error?

--
Erik Talvola               | "It's just what we need... a colossal negative 
talvola@cory.berkeley.edu  | space wedgie of great power coming right at us
..!ucbvax!cory!talvola    | at warp speed." -- Star Drek