Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!csd4.milw.wisc.edu!lll-winken!uunet!auspex!guy
From: guy@auspex.auspex.com (Guy Harris)
Newsgroups: comp.windows.x
Subject: Re: bug in XCopyArea() of X11R3 Xsun server
Message-ID: <1652@auspex.auspex.com>
Date: 18 May 89 18:24:44 GMT
References: <8905171605.AA16290@expo.lcs.mit.edu>
Reply-To: guy@auspex.auspex.com (Guy Harris)
Organization: Auspex Systems, Santa Clara
Lines: 24

>Richard suggested putting in ifdef code for sparc's
>to fix a design bug in the sparc architecture.
>In fact he should just complain to his compiler supplier
>for not supporting ANSI C. ANSI C is quite clear
>that the result of a shift of 32 is 0.

The December 7, 1988 ANSI C draft, at least, is quite clear that the
result of a shift of 32 is *undefined*; if you want, I can cite chapter
and verse.  I tend to doubt that this changed in the version of the
standard proposed to X3J11. 

>I would not want to see people defensively coding
>all their shifts in this fashion because one
>machine got it wrong. It will penalize all the
>machines that got it right.

More than one machine "got it wrong"; the 80386 - you know, the one used
by YOUR COMPANY's high-end PS/2 machines - also "got it wrong" (check
out the 80386 Programmer's Reference Manual).  Unless the proposed ANSI
C standard differs from the December 7, 1988 draft, people had damn well
better code defensively. 

BTW, if you use "#if"s, it only penalizes people porting the code, not
the machines that got it right.