[8094] in Athena Bugs
decmips 7.3M: dosread
daemon@ATHENA.MIT.EDU (Shamu the Chiller Whale)
Tue Sep 3 20:26:33 1991
To: bugs@ATHENA.MIT.EDU, cfyi@ATHENA.MIT.EDU
Date: Tue, 03 Sep 91 20:26:44 EDT
From: Shamu the Chiller Whale <krishna@ATHENA.MIT.EDU>
System name: hydrox-cookie
Type and version: KN01 7.3M (1 update(s) to same version)
Display type: PMAX-MFB
What were you trying to do?
Dosread in a file with the -a (ascii read) option.
What's wrong:
This section of code in /source/bsd-4.3/common/ibm/dosread.c needs to be
changed, reflecting what I believe to be a change in IBM PC DOS, or at least
the way files are stored these days. Now instead of a ^Z for eof, there's a ^D
at the beginning and end of a file, I think. Anyone want to verify this?
{
static next = 0;
register char *end;
register char *p = buff;
register int c;
if (!aflg)
return (read(fileno(file), buff, length));
for (end = buff + length; p < end;)
{
if (c = next)
{
*p++ = c;
next = 0;
continue;
} /* handle worst case */
c = getc(file);
if (c == EOF)
{
if (p != buff)
*p++ = CTL_Z; /* not quite correct but close */
break;
}
if (c == NL)
{
next = c;
c = CR;
}
*p++ = c;
}
return (p - buff);
}
What should have happened:
It should have removed any non-supposed-to-be-there ascii characters.
It missed the beginning and trailing ^D.
Please describe any relevant documentation references:
