Re: pkg/45046: mk/* minix support

[Alistair Crooks <agc%pkgsrc.org@localhost>]

The following reply was made to PR pkg/45046; it has been noted by GNATS.

From: Alistair Crooks <agc%pkgsrc.org@localhost>
To: gnats-bugs%NetBSD.org@localhost
Cc: pkg-manager%NetBSD.org@localhost, gnats-admin%NetBSD.org@localhost, 
pkgsrc-bugs%NetBSD.org@localhost,
        tcort%minix3.org@localhost
Subject: Re: pkg/45046: mk/* minix support
Date: Tue, 2 Aug 2011 07:50:01 +0200

 Thanks for the patch, Thomas.
 
 On Tue, Jul 19, 2011 at 01:30:07PM +0000, Thomas Cort wrote:
 > The following reply was made to PR pkg/45046; it has been noted by GNATS.
 > 
 > From: Thomas Cort <tcort%minix3.org@localhost>
 > To: gnats-bugs%netbsd.org@localhost
 > Cc: 
 > Subject: Re: pkg/45046: mk/* minix support
 > Date: Tue, 19 Jul 2011 09:29:08 -0400
 > 
 >  After submitting the original patch, Minix got a new C library and headers.
 >  The C library and headers were ported from NetBSD. There are changes to
 >  CPPFLAGS to add some definitions needed for the new headers. LDFLAGS is
 >  set to include -lcompat_minix and -lminlib, two compatibility libraries
 >  that go with our new libc. We also fixed the path to rmdir. An updated
 >  patch follows.
 >  
 >  Patch Summary:
 >  
 >   * We had to do a workaround in bsd.pkg.mk and bsd.prefs.mk for X11.
 >     We don't have modular xorg ported yet and we don't have a truly
 >     native (outside of pkgsrc) version of X11. We're using X11R6.8.2
 >     in a package we maintain in our pkgsrc tree (minix/x11).
 
 I'd really be inclined to special-case your X11_TYPE then, to be
 "minix3_native" or similar.  We have moved away from checking for
 specific operating system types in our main files (and especially
 bsd.pkg.mk), because it does not scale.  It may be the case that other
 operating systems are also going through the X11 bring-up process, and
 having a way to bring them into pkgsrc in an easy manner would be
 beneficial for them.
  
 >   * We changed a few lines in buildlink3/imake-check to avoid piping
 >     the output because imake segfaults on our platform when reading
 >     from STDIN.
 
 I'm just catching up now, but I'm not sure if this shouldn't be covered
 by a check for X11_TYPE being native_minix3 or whatever?
   
 Thanks,
 Alistair