Path: utzoo!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!caen!news.cs.indiana.edu!arizona.edu!arizona!dave
From: dave@cs.arizona.edu (Dave P. Schaumann)
Newsgroups: comp.sys.amiga.programmer
Subject: Manx/Lattice divergant behaviour
Message-ID: <1001@caslon.cs.arizona.edu>
Date: 4 Mar 91 00:04:25 GMT
Organization: U of Arizona CS Dept, Tucson
Lines: 41
Weather: sunny and warm!


I've come across a curious incompatability between Manx and Lattice, and
I'm hoping that someone can clear the air for me.

Through the course of events, a friend and I both compiled the following code.
I with Manx 3.6a, he with a Lattice compiler (sorry, don't know the version):

/* Please excuse any typos; I typed this from memory.
*/
#include <stdio.h>

main() {

  FILE *my_win ;
  int c ;

  if( (my_win = fopen( "CON:10/10/100/200/C window", "a+" )) == NULL ) {
    puts( "Can't open window!" ) ; exit(1) ;
    }

  fprintf( my_win, "Some text\nAnd more text\n" ) ;
  while( (c = fgetc(my_win)) != EOF ) putc( c ) ;

  fclose( my_win ) ; puts( "Successful!" ) ; exit(0) ;
}

As you can see, this is a very straight forward, simple program that opens
a window, prints some text, then enters a loop that reads characters from
the program window and prints them to the shell window.

My version of this works fine.  The version compiled with Lattice sees EOF
at fgetc() the first time it is executed, and exits immediately, whatever
the input is.

My question is: is the Lattice compiler broken/deficient, or is this an
example of code with undefined behaviour?
-- 
		Dave Schaumann		dave@cs.arizona.edu
'Dog Gang'!  Where do they get off calling us the 'Dog Gang'?  I'm beginning to
think the party's over.  I'm beginning to think maybe we don't need a dog.  Or
maybe we need a *new* dog.  Or maybe we need a *cat*! - Amazing Stories