Path: utzoo!attcan!uunet!munnari.oz.au!mel.dit.csiro.au!yarra!pta!teti!teslab!andrew
From: andrew@teslab.lab.OZ (Andrew Phillips)
Newsgroups: comp.sys.amiga.tech
Subject: Re: feof does not work: HELP!
Summary: Aztec C feof() is OK
Message-ID: <1164@teslab.lab.OZ>
Date: 10 Dec 90 02:53:34 GMT
References: <4413@ruuinf.cs.ruu.nl>
Reply-To: andrew@teslab.lab.oz.au (Andrew Phillips)
Organization: Technology Evaluation Section, L.A.B., Sydney
Lines: 37

In article <4413@ruuinf.cs.ruu.nl> ptavoly@cs.ruu.nl (Peter Tavoly) writes:
>   40  while (!feof(in))        /* <- this is where the trouble starts :( */
>   41   {

>   43    while ((ch=fgetc (in))!='@') /* read until separator encountered */
>   44     putchar (ch);               /* and output characters one by one */

>   46    fgets (line [y],74,in);      /* read rest of line for later use */
>   47   if (feof(in)) return;

>   49   } /* while */

>Now the problem is that AztecC 5.0a somehow doesn't recognize the 'feof' ...

Your trouble is that feof() does not return 1 until after you try
reading PAST the end of file.  So feof(in) is not true until after
you read the first character of the next line (assuming you have a
line-feed at the end of the last line of your file).

>... however the program keeps printing the character 255
>(an y with an umlaut on top) endlessly.

>I could try to look for a character 255 but this is not a nice way, and
>suppose that this character appears in your text, (as in Queensryche..:).
>In case of binaries it is even more likely to appear.

The character 255 comes from fgetc() returning -1 (fgetc returns an
INT) to indicate eof (or error) and this being converted to a char
when assigned to ch.  It would be safer to make ch an int and check
for ch being -1 after each call to fgetc().

So I would say that Aztec C feof() is OK, although I haven't bothered
to test it.

Andrew.
-- 
Andrew Phillips (andrew@teslab.lab.oz.au) Phone +61 (Aust) 2 (Sydney) 289 8712