--Signature=_Mon__26_Apr_2004_11_06_53_-0500_/Ke7eEizi4mTmmn=
Content-Type: text/plain; charset=US-ASCII
Content-Disposition: inline
Content-Transfer-Encoding: 7bit

* Julio M. Merino Vidal, Mon, 26 Apr 2004 17:16:30 +0200:
> [snip]
> 
> The problem is that bsd.buildlink3.mk contains the following chunk of code:
> 
> .  if !empty(BUILDLINK_DEPMETHOD.${_pkg_}:Mfull)
> _BLNK_DEPMETHOD.${_pkg_}=       _BLNK_ADD_TO.DEPENDS
> _BLNK_RECMETHOD.${_pkg_}=       _BLNK_ADD_TO.RECOMMENDED
> .  elif !empty(BUILDLINK_DEPMETHOD.${_pkg_}:Mbuild)
> _BLNK_DEPMETHOD.${_pkg_}=       _BLNK_ADD_TO.BUILD_DEPENDS
> .  endif
> 
> As you can see, _BLNK_RECMETHOD.${_pkg_} is not defined in the second part of
> the conditional, so the conditional in line 243 fails later.  (I tried to
> define the variable to the empty string, but no luck).
> 
> I'm not sure about the following patch... it adds another check, and avoids
> doing that check in case we are using a full dependency; is it ok?
> Maybe I should just check if the variable is defined in line 243 before
> checking its value?

IMHO, it would be a more consistent way to define
_BLNK_DEPMETHOD.${_pkg_} in every case.  See the patch I posted last
week.

http://mail-index.netbsd.org/tech-pkg/2004/04/20/0025.html

-- 
Min Sik Kim

--Signature=_Mon__26_Apr_2004_11_06_53_-0500_/Ke7eEizi4mTmmn=
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (NetBSD)

iQEVAwUBQI0znfQsKR/Zi7L+AQKOwgf/UjWOp/+WH+7sA0L556+gQQT1H0ugM7Y2
4YMpav+CzS1rc49rZf5ums1UpnLZl+0K/LYyudXs1f+MHZSushkQqXYWRUM96rIl
mCqAyaFXe2FB5vZDqYXQGRkOvsKhJOJ9QAFKeJnEHnyGPdTREDr2F36Gtk9vo2yO
d50HyfEHmXLdMnqLcUvVTHyX7CXkTsiO211iQixaL4CNgnhzq8dRPxNrjrAraw3C
fcIDfaYjKlOjrbzbNrKT61I0ZtPJEFsm86tXA6Iai7G99sBZJdLVdpCOXx1M9SSO
Yoym2gYintP25hoH1XYajmvLZxxEoSH3bFc7TKiOO7ROaj810B2xqg==
=2QJC
-----END PGP SIGNATURE-----

--Signature=_Mon__26_Apr_2004_11_06_53_-0500_/Ke7eEizi4mTmmn=--