--vkogqOf2sHV7VnPd
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Wed, Nov 28, 2001 at 12:26:44AM -0500, James Chacon wrote:
> I'd have to look at the exact package but some of the sub makes do not re=
ad
> in bsd.pkg.mk but do read in the system .mk files.

Is there a good reason for that? That kind of breaks the whole
"avoid damaging package builds with #ifndef BSD_PKG_MK" idea...

I thought I'd also seen this problem with a couple other packages,
so I figured it was a general "pkgsrc mk got broke". But perhaps
not.

> Don't set USETOOLS=3Dyes...The build system will set it for you by defaul=
t.
> By setting it explictly to yes you are forcing the toolchain to be used
> by anything which may happen to use the system make files.

Yeah, I saw the conversation about that today (er, yesterday, in my
TZ), but I haven't even attempted to do a build since I read it,
so I hadn't bothered to change it.  (I'm aware of that, though.)

Note that commenting only USETOOLS in my /etc/mk.conf does not fix
the package building problem; DESTDIR also gets in the way. So, just
the fact that the new toolchain's gotten past needing USE* set
doesn't make this go away.

> Ok...you can set it like that but the behavior you're seeing is what it's
> supposed to do in that case.

Huh? Well, sure, using the nbmake wrapper, but that behavior isn't
supposed to happen within pkgsrc, as I understood things. Shouldn't
packages that don't source bsd.pkg.mk at least set sanity values
like this? (If so, I'll do a send-pr on apla.)

--=20
       ~ g r @ eclipsed.net

--vkogqOf2sHV7VnPd
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (NetBSD)
Comment: For info see http://www.gnupg.org

iEYEARECAAYFAjwEiZsACgkQ9ehacAz5CRoYdACeKAoSDgV09Zc62mnu+kIKFzVW
HmIAn16HPCUGkHClNWJ7m85/V7yrV4wr
=KgrV
-----END PGP SIGNATURE-----

--vkogqOf2sHV7VnPd--