From owner-freebsd-ports  Mon May 31  6: 6: 1 1999
Delivered-To: freebsd-ports@freebsd.org
Received: from labinfo.iet.unipi.it (labinfo.iet.unipi.it [131.114.9.5])
	by hub.freebsd.org (Postfix) with SMTP
	id 0CFB91501B; Mon, 31 May 1999 06:05:55 -0700 (PDT)
	(envelope-from luigi@labinfo.iet.unipi.it)
Received: from localhost (luigi@localhost) by labinfo.iet.unipi.it (8.6.5/8.6.5) id MAA18732; Mon, 31 May 1999 12:39:03 +0200
From: Luigi Rizzo <luigi@labinfo.iet.unipi.it>
Message-Id: <199905311039.MAA18732@labinfo.iet.unipi.it>
Subject: Re: a two-level port system?  (fwd)
To: mladavac@metropolitan.at (Ladavac Marino)
Date: Mon, 31 May 1999 12:39:02 +0200 (MET DST)
Cc: taavi@uninet.ee, mladavac@metropolitan.at, nick.hibma@jrc.it,
	freebsd-ports@FreeBSD.ORG, freebsd-hackers@FreeBSD.ORG
In-Reply-To: <55586E7391ACD211B9730000C1100276179630@r-lmh-wi-100.corpnet.at> from "Ladavac Marino" at May 31, 99 02:43:30 pm
X-Mailer: ELM [version 2.4 PL23]
Content-Type: text
Content-Length: 2466      
Sender: owner-freebsd-ports@FreeBSD.ORG
Precedence: bulk
X-Loop: FreeBSD.org

> > With one big file it is next to impossible to build version 1.1.1 of
> > one
> > port and 1.1.2 of another.
> > 
> > With current model i can check out specific branch for
> > all files/ports separately.
> > 
> 	[ML]  You have a point there :)
> 	Of course, you could check out the 1.1.1 version of the big
> file, build your port, and then the 1.1.2 version of the big file, build
> the other one, but this would be probably rather slow--some testing will
> be needed.
> 
> 	Another possibility is one file per port, thus keeping the stuff
> in more manageable chunks.  For this I don't even have to write
> anything--shar will do (something better than shar, something that keeps
> the file entries alphabetized and thus guarantees minimal diffs would be
> good, though).
> 
> 	I hope we all agree that a reduction in file/directory count is
> desirable.

in fact i think the biggest problem, performancewise, is the presence
of multiple subdirs per port.
I'd be happy if we could build a backward compatible method that (in
order of importance)

  1) allows short "files" such as those in pkg/ (with perhaps the
     exception of PLIST and files/md5 for other reasons) to be
     stored as Makefile variables instead of external files
     (the backward mechanism would be to look at the file if a
     matching name is not found;).
     pkg/PLIST and files/md5, if really needed, could be moved to
     the main directory.
     This would remove one dir (pkg/) plus 3-4 files in 100% of the ports,
     plus another dir (files/) in perhaps 90% of the cases.

  2) keeps patches in the main directory instead of a separate subdir.

  3) pieces from files/ are also moved into the main directory with
     an adequate prefix (e.g. new-foo.c) to be stripped at install time
     (it is my understanding that in most cases the copy from files/ to
     the right place is done by explicit commands in the Makefile,
     right ?)
     This is trickier and maybe not worthwhile...

comments ?

	luigi
-----------------------------------+-------------------------------------
  Luigi RIZZO, luigi@iet.unipi.it  . Dip. di Ing. dell'Informazione
  http://www.iet.unipi.it/~luigi/  . Universita` di Pisa
  TEL/FAX: +39-050-568.533/522     . via Diotisalvi 2, 56126 PISA (Italy)

		  http://www.iet.unipi.it/~luigi/ngc99/
====  First International Workshop on Networked Group Communication  ====
-----------------------------------+-------------------------------------


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