Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!apple!brutus.cs.uiuc.edu!ginosko!cs.utexas.edu!uunet!mcsun!ukc!slxsys!jclark!jjc
From: jjc@jclark.UUCP (James Clark)
Newsgroups: comp.lang.c++
Subject: cfront 2.0 bug
Message-ID: <JJC.89Oct11115213@yquem.jclark.UUCP>
Date: 11 Oct 89 10:52:13 GMT
Sender: news@jclark.uucp
Distribution: comp
Organization: None
Lines: 27

cfront 2.0 (on a Sun 4) generates incorrect (but legal) C code for the
following:

inline int and(int x, int y)
{
  if (x) {
    if (y)
      return 1;
  }
  return 0;
}

extern "C" int abort(...);

main()
{
  int m = 1, n = 1;
  if (!and(m, n))
    abort();
}

Is there anybody at AT&T who is interested in receiving bug reports?
I have a couple more (not as bad as this one). Is anybody keeping a
list of cfront 2.0 bugs that aren't covered in the release notes?

James Clark
jjc@jclark.uucp