NetBSD Problem Report #15920

Received: (qmail 26792 invoked from network); 15 Mar 2002 00:21:23 -0000
Message-Id: <20020315002216.6077868ED2@hiro.dyn.dhs.org>
Date: Fri, 15 Mar 2002 01:22:16 +0100 (CET)
From: wiz@netbsd.org
Reply-To: wiz@netbsd.org
To: gnats-bugs@gnats.netbsd.org
Subject: curses vs. ncurses: ERR/OK definition mismatch
X-Send-Pr-Version: 3.95

>Number:         15920
>Category:       lib
>Synopsis:       curses vs. ncurses: ERR/OK definition mismatch
>Confidential:   no
>Severity:       serious
>Priority:       high
>Responsible:    blymn
>State:          closed
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Fri Mar 15 00:22:00 +0000 2002
>Closed-Date:    Tue Apr 01 18:31:16 +0000 2003
>Last-Modified:  Tue Mar 23 21:36:00 +0000 2004
>Originator:     Thomas Klausner
>Release:        NetBSD 1.5ZB
>Organization:
>Environment:


System: NetBSD hiro 1.5ZB NetBSD 1.5ZB (HIRO) #0: Wed Mar 13 01:43:46 CET 2002 wiz@hiro:/archive/cvs/src/sys/arch/i386/compile/HIRO i386
Architecture: i386
Machine: i386
>Description:
In ncurses-5.2's ncurses.h, ERR is defined to be -1, and OK 0.
In NetBSD's curses.h, ERR is 0, and OK is 1.
This will probably lead to compatibility problems.
>How-To-Repeat:
Compare the two header files.
>Fix:
Change ERR and OK in curses.h to -1 and 0 respectively, and audit
the curses code on hidden assumptions?

>Release-Note:
>Audit-Trail:
Responsible-Changed-From-To: lib-bug-people->blymn 
Responsible-Changed-By: wiz 
Responsible-Changed-When: Thu Mar 14 16:22:34 PST 2002 
Responsible-Changed-Why:  
blymn said he'd take a look at this. 

From: matthew green <mrg@eterna.com.au>
To: wiz@netbsd.org
Cc: gnats-bugs@gnats.netbsd.org
Subject: re: lib/15920: curses vs. ncurses: ERR/OK definition mismatch 
Date: Fri, 15 Mar 2002 12:11:17 +1100



    System: NetBSD hiro 1.5ZB NetBSD 1.5ZB (HIRO) #0: Wed Mar 13 01:43:46 CET 2002 wiz@hiro:/archive/cvs/src/sys/arch/i386/compile/HIRO i386
    Architecture: i386
    Machine: i386
    >Description:
    In ncurses-5.2's ncurses.h, ERR is defined to be -1, and OK 0.
    In NetBSD's curses.h, ERR is 0, and OK is 1.
    This will probably lead to compatibility problems.
    >How-To-Repeat:
    Compare the two header files.
    >Fix:
    Change ERR and OK in curses.h to -1 and 0 respectively, and audit
    the curses code on hidden assumptions?


 i don't see why this is a problem and changing it changes our
 ABI yet again?  people shouldn't link code with ncurses.h and
 curses.h together, that's going to blow up more than just this.

From: Julian Coleman <jdc@coris.demon.co.uk>
To: wiz@netbsd.org
Cc: gnats-bugs@netbsd.org
Subject: Re: lib/15920: curses vs. ncurses: ERR/OK definition mismatch
Date: Tue, 1 Apr 2003 18:38:01 +0100

 > >Fix:
 > Change ERR and OK in curses.h to -1 and 0 respectively, and audit
 > the curses code on hidden assumptions?

 As MRG mentioned, I'm not sure that this is a problem.  The specification
 notes that functions return ERR and OK and that these are defined in
 curses.h.  No mention is made of explicit values for them.  As long as a
 program uses curses.h, everything should be fine.

 Can the PR be closed and this looked at if there is a problem with a package?

 Thanks,

 J

 -- 
                     My other computer also runs NetBSD
                           http://www.netbsd.org/

From: Thomas Klausner <wiz@netbsd.org>
To: Julian Coleman <jdc@coris.demon.co.uk>
Cc: gnats-bugs@netbsd.org
Subject: Re: lib/15920: curses vs. ncurses: ERR/OK definition mismatch
Date: Tue, 1 Apr 2003 19:49:02 +0200

 On Tue, Apr 01, 2003 at 06:38:01PM +0100, Julian Coleman wrote:
 > As MRG mentioned, I'm not sure that this is a problem.  The specification
 > notes that functions return ERR and OK and that these are defined in
 > curses.h.  No mention is made of explicit values for them.  As long as a
 > program uses curses.h, everything should be fine.
 > 
 > Can the PR be closed and this looked at if there is a problem with a package?

 It was prompted by a problem in pkgsrc/games/frotz, but I think this has
 been fixed in its sources now.

 Brett asked me to send a PR when I mentioned the difference.

 It is fine with me to close it.
  Thomas

From: Julian Coleman <jdc@coris.demon.co.uk>
To: Thomas Klausner <wiz@netbsd.org>
Cc: gnats-bugs@netbsd.org
Subject: Re: lib/15920: curses vs. ncurses: ERR/OK definition mismatch
Date: Tue, 1 Apr 2003 19:29:24 +0100

 > It was prompted by a problem in pkgsrc/games/frotz, but I think this has
 > been fixed in its sources now.

 Ah.  I missed that.

 > It is fine with me to close it.

 Great!  Thanks,

 J

 (bug team hat on ;-)
 -- 
                     My other computer also runs NetBSD
                           http://www.netbsd.org/
State-Changed-From-To: open->closed 
State-Changed-By: jdc 
State-Changed-When: Tue Apr 1 10:30:48 PST 2003 
State-Changed-Why:  
Problem fixed in package. 

From: Julian Coleman <jdc@netbsd.org>
To: gnats-bugs@gnats.NetBSD.org
Cc:  
Subject: pr/15920 CVS commit: src/lib/libcurses
Date: Tue, 23 Mar 2004 21:35:29 +0000 (UTC)

 Module Name:	src
 Committed By:	jdc
 Date:		Tue Mar 23 21:35:29 UTC 2004

 Modified Files:
 	src/lib/libcurses: curses.h

 Log Message:
 Change the values of ERR and OK.
 This makes our getch() and related functions conform to SUSv2.
 Also fixes PR lib/15920.

 XXX: This causes an API incompatability, but comes less than 27 hours after
      the libcurses major number was incremented, so should have minimal impact.


 To generate a diff of this commit:
 cvs rdiff -r1.84 -r1.85 src/lib/libcurses/curses.h

 Please note that diffs are not public domain; they are subject to the
 copyright notices on the relevant files.

>Unformatted:

Home	PR Database Search