Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 24 Oct 1996 12:22:09 +0930 (CST)
From:      Michael Smith <msmith@atrad.adelaide.edu.au>
To:        erich@lodgenet.com (Eric L. Hernes)
Cc:        msmith@atrad.adelaide.edu.au, ports@freebsd.org
Subject:   Re: linux_devel port-thing
Message-ID:  <199610240252.MAA04543@genesis.atrad.adelaide.edu.au>
In-Reply-To: <199610231331.IAA08154@jake.lodgenet.com> from "Eric L. Hernes" at Oct 23, 96 08:31:37 am

next in thread | previous in thread | raw e-mail | index | archive | help
Eric L. Hernes stands accused of saying:
> >
> >Ok, so I have this at least partially sorted out; I have an image here
> >of what's supposed to come out at the end, and a 17M compressed
> >tarfile of same.
> 
> Yikes, I guess /compat/linux will have to move
> off my root partition ;-)

Yeah.  I like having /compat, but I stronly feel that it should always 
be a symlink to somewhere else.  (/usr/share/compat is probably the
best place I've heard suggested)

> >What makes the most sense is for this to just get extracted directly over
> >the top of /compat/linux, with nothing other than the 'fetch' and 'extract'
> >targets being relevant.
> 
> `make package' should probably work, or be explicitly disabled
> by NO_PACKAGE.

Aigh.  I can see that exploding pretty badly; pkg_add uses /tmp or /var/tmp,
and lots of people just wouldn't have the 45M that it takes to unpack
it in the sandbox.  The same problem occurs with xemacs, btw.

> >So I guess I should set PREFIX to /compat/linux, and WRKSRC to ${PREFIX}.
> >Unfortunately that means that if I make a Makefile patch like the 
> >linux_lib port does, it'll be written in /compat/linux/, and 'make clean'
> >may have a rather unfortunate effect. 8(
> 
> NO_MTREE should be set too, we don't want a full-blown bsd heirarchy
> in /compat/linux.  I set up the linux_lib port to extract to a staging
> area, with the intent that I may want to edit/patch a few files
> there eventually, or maybe install just pieces of it.  Plus it had
> the benefit that it was more like the rest of the ports, but I didn't
> have 17M to deal with either.

I don't think that staging it would be a good idea.  What I want is
basically "fetch this monster tarball" and "extract this monster into
/compat/linux".  What I'm not so sure about is which parts of the 
port process it should attach to.  

Perhaps :
 fetch :	slurp
 build :	symlink to distfile in WRKSRC
 install :	unpack distfile referenced by symlink

> At one time I thought of having `make fetch' grab the various *.tgz's
> from slackware/redhat/... but finally decided that keeping track of all
> those linux dists *and* the versions of shared libs was way too much. :(

Hmm.  I looked at that too, and it's not really such a bad idea.  I have
a script which unpacks the slackware D distribution and runs the install
scripts, as well as patching ones that need patching before running them.

It is, however, a RRPITA to test.

> eric.

-- 
]] Mike Smith, Software Engineer        msmith@gsoft.com.au.au          [[
]] Genesis Software                     genesis@gsoft.com.au            [[
]] High-speed data acquisition and      (GSM mobile)     0411-222-496   [[
]] realtime instrument control.         (ph)          +61-8-8267-3493   [[
]] Unix hardware collector.             "Where are your PEZ?" The Tick  [[



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199610240252.MAA04543>