Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!ucbvax!husc6!rice!sun-spots-request
From: stanonik@nprdc.navy.mil (Ron Stanonik)
Newsgroups: comp.sys.sun
Subject: fscanf bug?
Keywords: SunOS
Message-ID: <326@brazos.Rice.edu>
Date: 20 Jul 89 07:00:00 GMT
Sender: root@rice.edu
Organization: Sun-Spots
Lines: 42
Approved: Sun-Spots@rice.edu
X-Sun-Spots-Digest: Volume 8, Issue 80, message 11 of 15

Or maybe it's a feature?  But it was news to me.

I believe we've encountered an fscanf bug in sunos4.0.1 on sun4's, sun3's,
and sun386i's.

The apparent bug is that given a control string of "%c", fscanf returns 0,
rather than EOF on encountering the end of file.

Here's a sample program:
#include <stdio.h>

main()
{
	int i;
	char c;

	for (i = 5; i > 0; i--) {
		printf("cnt=%d\n", fscanf(stdin, "%c", &c));
		printf("%x\n", c);
	}
}

Here's a sample file:
ab

Here's the result:
arctic[244] t < a
cnt=1
31
cnt=1
33
cnt=1
a
cnt=0
a
cnt=0
a

Back in sunos3.5 days, the 0 cnt's would be -1 (ie, EOF), as expected.

Ron Stanonik
stanonik@nprdc.navy.mil