Path: utzoo!utgpu!news-server.csri.toronto.edu!clyde.concordia.ca!uunet!wuarchive!brutus.cs.uiuc.edu!apple!sun-barr!ccut!wnoc-tyo-news!scslwide!wsgw!wsserva!naga2!nagamatu
From: nagamatu@sm.sony.co.jp (Tatsuo Nagamatu)
Newsgroups: comp.sys.mips
Subject: Re: mips cc bug
Message-ID: <NAGAMATU.90May29145128@naga2.sm.sony.co.jp>
Date: 29 May 90 05:51:28 GMT
References: <1990May28.152151.5506@maytag.waterloo.edu>
Sender: news@sm.sony.co.jp
Organization: Sony Workstation Div.
Lines: 43
In-reply-to: gamiddon@maytag.waterloo.edu's message of 28 May 90 15:21:51 GMT


In article <1990May28.152151.5506@maytag.waterloo.edu>
	gamiddon@maytag.waterloo.edu (Guy Middleton) writes:
>This code works everywhere except on my MIPS machine:

I simplyfy the posted source code as follows:

	% cat a.c
	main ()
	{
	    int x;
	
	    switch (x) {
		case ((unsigned)0x10 << 2):
		    printf ("blah\n");
		    break;
	    }
	}

Ccom can produce the debug message of building tree structure like
this:

	% /usr/lib/cmplrs/cc/ccom -Xb a.c

In the debug message, we can see the followings.

	2001031760) CAST, unsigned, 16, 14
	    2001031600) NAME, 0, -1, unsigned, 16, 14
	    2001032050) SCONV, unsigned, 0, 14
		2001031670) ICON, 16, 16384, int, 0, 4

With this message, MIPS C compiler can't pre-calculation of the
constant casting. SCONV for constant is not nessesary.

	2001031760) CAST, unsigned, 16, 14
	    2001031600) NAME, 0, -1, unsigned, 16, 14
	    2001031670) ICON, 16, 16384, unsigned, 0, 4
                                         ~~~~~~~~
Is this bug is included in version 2.1?

--
  $@%=%K!<(J($@3t(J)$@%9!<%Q!<%^%$%/%m;v6HK\It%o!<%/%9%F!<%7%g%s;v6HIt(J
	$@1J>>(J $@N5IW(J	nagamatu@sm.sony.co.jp