*BSD News Article 12320


Return to BSD News archive

Newsgroups: comp.os.386bsd.bugs
Path: sserve!manuel.anu.edu.au!munnari.oz.au!constellation!convex!convex!cs.utexas.edu!wupost!newsfeed.rice.edu!rice!news.Rice.edu!rich
From: rich@Rice.edu (Richard Murphey)
Subject: Re: cvs 1.3 bugfix
In-Reply-To: bde@runx.oz.au's message of 5 Mar 93 06:26:11 GMT
Message-ID: <RICH.93Mar5215806@omicron.Rice.edu>
Sender: news@rice.edu (News)
Reply-To: Rich-Murphey@Rice.edu
Organization: Department of Electrical and Computer Engineering, Rice
	University
References: <RICH.93Mar3114046@omicron.Rice.edu>
	<SOMMERFELD.93Mar3180839@blatt.apollo.hp.com>
	<1993Mar5.062611.26942@runx.oz.au>
Date: Sat, 6 Mar 1993 03:58:06 GMT
Lines: 18

In article <1993Mar5.062611.26942@runx.oz.au> bde@runx.oz.au (Bruce Evans) writes:
   In article <SOMMERFELD.93Mar3180839@blatt.apollo.hp.com> sommerfeld@apollo.hp.com (Bill Sommerfeld) writes:

   >I think this is a bug in the sigismember macro, not in cvs....

   sigismember(), sigaddset() and sigdelset() all fail to check their signal
   number arg.

   The bug was recently fixed for linux by not defining the bad macros, but
   the 386BSD library doesn't have functions behind the macros (probably
   another bug) so this doesn't work.
   -- 
   Bruce Evans  bde%runx.oz.au@ips.oz.au

OK, I'm hearing conflicting interpretations of 3.3.3.3 and 3.3.3.4 in
POSIX 1003.1.  The standard states that *if* an error is detected
errno will be set.  It does not say that invalid arguments must be
detected.  I'm not at all sure who's right.  Rich