HEADS UP: New bsd.*.mk changes

Joe Marcus Clarke marcus at FreeBSD.org
Tue Jan 20 10:22:00 PST 2004


On Tue, 2004-01-20 at 13:16, Oliver Eikemeier wrote:
> Joe Marcus Clarke wrote:
> 
> > On Tue, 2004-01-20 at 13:03, Oliver Eikemeier wrote:
> > 
> >>Adam Weinberger wrote:
> >>
> >>
> >>>>>(01.20.2004 @ 1248 PST): Joe Marcus Clarke said, in 3.9K: <<
> >>>>>Can we at least have
> >>>>>
> >>>>>OPTIONSNAME?=${PKGNAMEPREFIX}${PORTNAME}
> >>>>>
> >>>>>here? It makes sense for all the localized ports, perl, ruby, python, linux
> >>>>>and others.
> >>>>
> >>>>That makes sense...maybe having the whole trifecta is a good idea, too:
> >>>>${PKGNAMEPREFIX}${PORTNAME}${PKGNAMESUFFIX}
> >>>>
> >>>>And what about UNIQUENAME instead of OPTIONSNAME?  This way, it could be
> >>>>used for other purposes.
> >>>>
> >>>>
> >>>>>end of "Re: HEADS UP: New bsd.*.mk changes" from Joe Marcus Clarke <<
> >>>
> >>>
> >>>If all you're going for is a unique identifier for each port, this is
> >>>guaranteed unique:
> >>>
> >>>${.CURDIR:S@${PORTSDIR}@@:S@/@_ at g}
> >>>
> >>>x11_gnome2, etc.
> >>
> >>Sorry to bring it up again, LATEST_LINK is designed to do what we want, it
> >>has just a stupid name...
> >>
> >>Look at
> >>  <http://bento.freebsd.org/errorlogs/i386-packages-5-latest/Latest/>
> >>all packages in one directory, no collisions, version independend. It even
> >>survives renaming / category changes.
> > 
> > Except not every port creates a LATEST_LINK.
> 
> Not every port creates PKGLATESTFILE (${PKGLATESTREPOSITORY}/${LATEST_LINK}${PKG_SUFX}).
> Every port sets the variable LATEST_LINK (by default to
> ${PKGBASE}=${PKGNAMEPREFIX}${PORTNAME}${PKGNAMESUFFIX}), and ports that need to
> override it with something suitable (openldap, apache). Only the ports that build
> no packages (like the jdk ports) don't care and have the same LATEST_LINK.
> 
> I totally agree that nobody want to use a variable called 'LATEST_LINK' used for that
> purpose, but, hey, it's just a name. If we use something like
> OPTIONSNAME?=${LATEST_LINK}
> we have to use an workaround only for the port that build no packages, the rest should
> be automagically right. All the work of thinking of an unique name has already been done
> for a lot of ports.

Yes, sorry I've missed some of these points.  I'm trying to balance real
work and this discussion.  You're right that we could just "borrow"
LATEST_LINK for the options purpose, but it will require your
re-ordering patch that will require testing on bento.

So, the way I see it, we could keep things the way they are and wait for
the next experimental build cycle, or commit a UNIQUENAME patch that is
temporary until LATEST_LINK can be evaluated.  Once that gets decided,
we can PR the patch we want tested.

Joe

-- 
Joe Marcus Clarke
FreeBSD GNOME Team	::	marcus at FreeBSD.org
gnome at FreeBSD.org
FreeNode / #freebsd-gnome
http://www.FreeBSD.org/gnome

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 187 bytes
Desc: This is a digitally signed message part
Url : http://lists.freebsd.org/pipermail/freebsd-ports/attachments/20040120/0262dab4/attachment.bin


More information about the freebsd-ports mailing list