Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 1 Apr 1999 12:47:08 -0500 (EST)
From:      Chuck Robey <chuckr@mat.net>
To:        Richard Wackerbarth <rkw@dataplex.net>
Cc:        Satoshi - the Wraith - Asami <asami@FreeBSD.ORG>, jkh@zippy.cdrom.com, current@FreeBSD.ORG
Subject:   Re: /var/db/pkg/.mkversion
Message-ID:  <Pine.BSF.4.10.9904011237370.498-100000@picnic.mat.net>
In-Reply-To: <Pine.BSF.4.05.9904010703520.3443-100000@nomad.dataplex.net>

next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 1 Apr 1999, Richard Wackerbarth wrote:

> > Richard, don't forget that having /usr/src isn't required to build
> > ports. 
> 
> I don't think that I did forget. I explicitly reference that situation.
> 
> The real solution (re the TARGET) has to depend on something that the USER
> sets for a particular "run". As such, it should be an environment variable
> which defaults to the HOST value. 
> 
> At the same time, the ports have to consider the purpose for which they
> need the identification. For example, "fetch -A" is a HOST/TOOLSET
> situation. But selecting the set of sysctls to compile into the
> code is a TARGET question. Nobody said cross compilation is easy :-)

Well, seeing as FreeBSD doesn't have control over the software involved,
and nearly all the software involved is broken for the kind of
portability you're talking about, this seems unreasonable.  It's one
thing for ports to adapt a particular piece of software to run on
FreeBSD hosts, it's quite a different thing to make it cross-build, any
to any.  That would convert doing a port from being a reasonably short
job, to being (for each port) a major rewrite project.  Seeing the
extreme fluidity of many ports, and the large number of ports, this
sounds like a very unreal expectation.

Your statement "Nobody said cross compilation is easy :-)", well, the
work involved is huge, and the demand ... huh, you're the ONLY one to
demand it.  OK, maybe even "demand" is wrong, but you get the idea, that
it's a lot of work for something that isn't seen as necessary.

If you were talking about the FreeBSD sources themselves, everything I
said is out the window, because we *do* have total control of those, and
the change rate is under our purview.  For ports, with that outside our
control, and no established need, it seems excessive.

Go adapt one of the gnome things, see how long it takes you to do *one*
port with the features you seem to be requiring.  If it's less than what
I envision, perhaps you're getting what you want over badly, and seeing
a port done as you want would be instructive to the rest of us.  I think
you're going to spend a *lot* of time doing that one port.

----------------------------+-----------------------------------------------
Chuck Robey                 | Interests include any kind of voice or data 
chuckr@glue.umd.edu         | communications topic, C programming, and Unix.
213 Lakeside Drive Apt T-1  |
Greenbelt, MD 20770         | I run picnic (FreeBSD-current)
(301) 220-2114              | and jaunt (Solaris7).
----------------------------+-----------------------------------------------






To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.4.10.9904011237370.498-100000>