Date: Wed, 21 Jan 2004 13:15:00 -0500 From: Joe Marcus Clarke <marcus@FreeBSD.org> To: Oliver Eikemeier <eikemeier@fillmore-labs.com> Cc: Kris Kennaway <kris@obsecurity.org> Subject: Re: OPTIONSFILE path [HEADS UP: New bsd.*.mk changes] Message-ID: <1074708900.768.70.camel@gyros> In-Reply-To: <400EBAF2.5040400@fillmore-labs.com> References: <1074617147.757.16.camel@gyros> <20040120171315.GH94636@FreeBSD.org> <1074619795.757.43.camel@gyros> <400D68A1.4030501@fillmore-labs.com> <1074620919.757.51.camel@gyros> <20040120175136.GC3365@toxic.magnesium.net> <400D6D8F.7010708@fillmore-labs.com> <1074621945.757.63.camel@gyros> <400D7077.30009@fillmore-labs.com> <1074622845.757.69.camel@gyros> <20040120184123.GI94636@FreeBSD.org> <1074637133.757.128.camel@gyros> <400EA9E4.3040004@fillmore-labs.com> <1074704419.768.41.camel@gyros> <400EBAF2.5040400@fillmore-labs.com>
next in thread | previous in thread | raw e-mail | index | archive | help
--=-tVRqm4mZaDEV1vzMteCr Content-Type: text/plain Content-Transfer-Encoding: quoted-printable On Wed, 2004-01-21 at 12:46, Oliver Eikemeier wrote: > Joe Marcus Clarke wrote: >=20 > > On Wed, 2004-01-21 at 11:33, Oliver Eikemeier wrote: > >=20 > > [I think we're all on the same page at this point.] >=20 > Maybe I have to restate that I'm extremely happy that we finally have > started ports option handling.=20 Me, too. >=20 > No, I wanted to show that we have contradicting premises: >=20 > a.) the PKGNAMESUFFIX may be dependend on user settable options I agree. I was just making a point that there are ports (esp. for the -client/-server case, where a slave port is used as the trigger). >=20 > b.) the set of saved options to choose may be dependant on the PKGNAMESUF= FIX > (even though it might not be determined by user settable options) >=20 > This means that in case a) we can't use the PKGNAMESUFFIX to determine th= e > set of saved options, and if we can't do this we have to work around this > for case b.), e.g. set LATEST_LINK or UNIQUENAME for *every* port that > has options depending on the PKGNAMESUFFIX. >=20 > This is more dangerous as it seems: If one port doesn't use OPTIONS and h= as > a portname of apache, it will nevertheless get the options from > /var/db/ports/apache/options. >=20 > Say for example you do a test installation of apache2 (which we assume su= pport > OPTIONS) and decide to downgrade to apache13 (which we assume doesn't sup= port > OPTIONS), then the file /var/db/ports/apache/options is sourced neverthel= ess, > and you get all saved options from apache2 for apache13. There is no 'Hey= , I'm > OPTIONS-aware and want saved options, and I am willing to care for a uniq= ue name', > you get the options when a 'suitable' file is found. >=20 > Did you expect that? Yes, I counted on it :-). Just kidding, yes this is a problem...but, it can be solved if the porter adjusts UNIQUENAME accordingly. However, I'm not arguing with you: we need a better way. I'm just offering a workaround with what we have in the ports tree _now_. And, as you say below, adjusting UNIQUENAME is akin to having to add CONFLICTS. Please send-pr your patch when it's ready. Thanks. Joe --=20 Joe Marcus Clarke FreeBSD GNOME Team :: marcus@FreeBSD.org gnome@FreeBSD.org FreeNode / #freebsd-gnome http://www.FreeBSD.org/gnome --=-tVRqm4mZaDEV1vzMteCr Content-Type: application/pgp-signature; name=signature.asc Content-Description: This is a digitally signed message part -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.4 (FreeBSD) iD8DBQBADsGkb2iPiv4Uz4cRAgj9AJ9xgU6H0mbRbY50LhBkm49SktX40wCgltIi Mx9aCZjOlMEBWcw+ZMFhDYU= =WBq6 -----END PGP SIGNATURE----- --=-tVRqm4mZaDEV1vzMteCr--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1074708900.768.70.camel>