Xref: utzoo gnu.gcc.bug:2171 comp.unix.xenix:10825
Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uwm.edu!zaphod.mps.ohio-state.edu!sunybcs!dsinc!netnews.upenn.edu!vax1.cc.lehigh.edu!vlsi3b15!lehi3b15!jsnyder
From: jsnyder@lehi3b15.csee.Lehigh.EDU (Jay Snyder)
Newsgroups: gnu.gcc.bug,comp.unix.xenix
Subject: GCC Bug under SCO Xenix/386
Message-ID: <864@lehi3b15.csee.Lehigh.EDU>
Date: 29 Mar 90 08:02:04 GMT
Followup-To: gnu.gcc.bug
Organization: CSEE Dept. Lehigh University, Bethlehem, PA
Lines: 39

The following is a piece of code that I use for a primitive password encoding
scheme in my BBS software:

encodepw(pw,name)
char *pw,*name;
{
char i,ii;
ii=0x15;i=0;
while (*pw)
  {
  if (i>strlen(name)) i=0;
  *pw=*pw^(0x1f & name[i++])^(i++);
  *pw=*pw-5;
  if (*pw<'!') *pw=*pw+32;
  if (*pw==':') *pw='$';
  ++pw;
  }
}

here is the output from input: name="Username", passwd="password"

output from cc code:  "_bstyvoj"
output from gcc code: "`atszupi"


note that the gcc code is make each character of the output 1 less
numerically.

Anybody have any suggestions.

Also, I compiled the same code on an AT&T 3B15 running SYSV R3.1.1
and the results were the same as from cc on SCO.

My conclusion is that gcc has a bug.
-- 
====================================================
Jay A. Snyder       jsnyder@lehi3b15.csee.lehigh.edu
CSEE Dept.	    jsnyder%gdx@lehi3b15.csee.lehigh.edu
Lehigh University