Date: Sun, 8 Dec 2002 02:08:13 -0500 (EST) From: Trevor Johnson <trevor@jpj.net> To: Marcel Moolenaar <marcel@xcllnt.net> Cc: cvs-committers@FreeBSD.org, <cvs-all@FreeBSD.org>, <portmgr@FreeBSD.org> Subject: Re: cvs commit: ports/emulators/linux_base Makefile distinfo.alpha distinfo.i386 pkg-plist.i386 Message-ID: <20021208014658.M82142-100000@blues.jpj.net> In-Reply-To: <20021208014146.GA1987@athlon.pn.xcllnt.net>
next in thread | previous in thread | raw e-mail | index | archive | help
Marcel Moolenaar wrote:
> On Sat, Dec 07, 2002 at 08:15:49PM -0500, Trevor Johnson wrote:
> >
> > It sounds as though you plan to implement this in a way which won't create
> > duplicates of all the .alpha.rpm and .i386.rpm files. If so, I'm all for
> > it.
>
> There can never be duplication for the arch specific RPMs because
> *.alpha.rpm and *.i386.rpm are two independent and seperate sets
> of entities.
>
> The following implements what I suggest:
>
> Index: Makefile
> ===================================================================
> RCS file: /home/ncvs/ports/emulators/linux_base/Makefile,v
> retrieving revision 1.70
> diff -u -r1.70 Makefile
> --- Makefile 3 Dec 2002 06:45:56 -0000 1.70
> +++ Makefile 8 Dec 2002 01:39:36 -0000
> @@ -47,7 +47,7 @@
> BUILD_DEPENDS= rpm:${PORTSDIR}/archivers/rpm
>
> ONLY_FOR_ARCHS= alpha i386
> -DIST_SUBDIR= rpm
> +DIST_SUBDIR= rpm-${MACHINE_ARCH}
Presently all the RPMs exist in the rpm/ directory on the distfile
mirrors, CD-ROMs, and in users' private distfile caches. With your patch,
users would have to download a second copy of them all (or manually move
their distfiles around), and mirrors would have to store a second copy of
them all. My rough estimate is that this would take 30 MB of storage
space per architecture per linux_base port--on the order of 120 MB for
someone who works with both Alpha and i386 and with both linux_base 7.1
and linux_base-6. The original problem was that Red Hat changed the time
stamp on one file. I had thought of this solution, but it just seemed too
wasteful.
> > In the meantime, it would be best if the port were restored to a more
> > usable state. IMO its breakage is the greater inconvenience.
>
> Fair enough. I didn't want to apply the above patch because we're
> too close to release to mess with DIST_SUBDIR gratuitously...
Thank you. I read you as saying that it's okay to add back the second
checksums for the time being, until we come up with something better. If
I haven't misunderstood, would you be so kind as to add them back?
--
Trevor Johnson
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe cvs-all" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20021208014658.M82142-100000>