Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 23 May 1996 22:32:55 +0800
From:      Peter Wemm <peter@spinner.DIALix.COM>
To:        Poul-Henning Kamp <phk@critter.tfs.com>
Cc:        "Gary Palmer" <gpalmer@FreeBSD.ORG>, hackers@FreeBSD.ORG, peter@FreeBSD.ORG
Subject:   Re: src/gnu 
Message-ID:  <199605231432.WAA03506@spinner.DIALix.COM>
In-Reply-To: Your message of "Wed, 22 May 1996 20:00:34 GMT." <779.832795234@critter.tfs.com> 

next in thread | previous in thread | raw e-mail | index | archive | help
>> Comments? What does our current cvs meister thing?
>> 
>> (The reason behind the move is the fact that the import wasn't done
>>  properly in the first place, and if I remember, the tools were just
>>  brought over wholesale from the 1.x source tree rather than being
>>  seperately imported so that they could be mainained).
>
>Isn't the fundamental problem that nobody maintains them, rather than
>how they're maintained ?

Yes, that is more the problem IMHO.  I dont think we have a lot of spare 
time/energy to sink into re-porting stuff to a new tree..  However, if we 
came up with a workable plan, we could import new stuff that way.  Rod and 
I did a bit of work on a strategy for this before, I think it can still 
work, even though it's a bit of a pain to set up the first time.

The other main problem is that CVS is really quite limited in it's vendor 
branch area when faced with a vendor source that moves/renames/etc a fair 
bit.

This would not be a problem if we used something like P3 which handles 
renames while hardly raising an eyebrow (ie: you "import" the code in it's 
natural file layout and adapt the "view" of it so that it appears in a 
4.4-style bmake'able tree.)

(and before anybody asks "What's P3?", see http://www.p3.com/ and note the 
special (and extremely generous) FreeBSD arrangement)

>Apart from that, wouldn't it make more sense to use the "ports" paradigm
>on them instead of checking in all the files ?  In theory we should be
>running pretty much footool-9.3.tar.gz with a few patches, so why not
>speed up our CVS-tree by actually not checking in the thousands of files
>but rather access them by reference via the tar file ?

That might be OK, but there's a few gotchas:
1: The sup users would be getting multi-megabyte files in a single lump
2: The files would have to be extracted, patched and then compiled as part 
of the build process... That'd add a fair bit of overhead to the disk 
space requirements (ie: gcc would start at a 20MB or so .tar file, then 
we'd extract 20MB of files out of it and compile half of them.)

>That would also make it easier to show that we do the "GPL" thing...

I am not too concerned with that really, as we pretty much comply with the 
intent of the requirements..  We are only obligated to offer the sources 
somehow, which includes over the internet (which we do), and the changes 
we make are also visible on the internet (Thanks to the CvsWeb interface 
that somebody (who's name escapes me) wrote) that allows access to the 
logs and histories etc.  It just so happens that we also include a copy of 
the source on the CD's as well, and the files that we have modified are 
maked with the time it was modified (ie: the $Id$ line.  The GPL does not 
require is to keep a history of changes in the file, nor a description of 
the change, just the time of the last change in our hands.  IMHO, that's 
completely stupid as it's the most useless requirement, and especially 
since the majority of the GPL'ed stuff these days has ChangeLog files 
which do not comply with the letter of the GPL conditions).

I'd better stop now before I start really ranting about the GPL.  For a 
system that's meant to encourage "freedom of programming", all it's done 
is bitterly divide the non-commercial programming world into two groups, 
neither of which can freely use the other's source code..  So, the end 
result is that the non-commercial world has to reinvent each other's work 
which is an utter shame and only helps the proprietary companies get 
further ahead.

>--
>Poul-Henning Kamp           | phk@FreeBSD.ORG       FreeBSD Core-team.
>http://www.freebsd.org/~phk | phk@login.dknet.dk    Private mailbox.
>whois: [PHK]                | phk@ref.tfs.com       TRW Financial 
Systems, Inc
>.
>Future will arrive by its own means, progress not so.

Cheers,
-Peter





Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199605231432.WAA03506>