Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 31 Jan 2002 17:04:32 -0500
From:      Garance A Drosihn <drosih@rpi.edu>
To:        nate@yogotech.com (Nate Williams), "Matthew D. Fuller" <fullermd@over-yonder.net>
Cc:        arch@FreeBSD.ORG
Subject:   Re: Adding support for a global src tree serial number
Message-ID:  <p05101401b87f6aaa81ab@[128.113.24.47]>
In-Reply-To: <15449.45750.572076.480900@caddis.yogotech.com>
References:  <3C5944A4.4927F812@mindspring.com> <80628.1012484102@axl.seasidesoftware.co.za> <15449.30438.698921.182380@caddis.yogotech.com> <20020131173702.J77899@genius.tao.org.uk> <20020131183321.GA59544@gattaca.yadt.co.uk> <20020131184230.D84715@genius.tao.org.uk> <20020131150720.A33201@over-yonder.net> <15449.45750.572076.480900@caddis.yogotech.com>

next in thread | previous in thread | raw e-mail | index | archive | help
At 2:10 PM -0700 1/31/02, Nate Williams wrote:
>  > > The problem is where to put the serial number, and how to
>  > > distribute it.  It needs to be maintained in a ,v file for cvs
>  > > to be able to check it out.  That adds additional complexity.
>  >
>  > Note also the additional trouble of underhandedly hand-crafting a
>  > ,v-alike file will break checking out a revision other than the
>  > absolute latest on the branch, since you'll end up with either no
>  > serial, or still the latest serial, depending on how the file is
>  > crafted.
>
>In general, that would only be a problem with folks that have the
>CVS tree who have updated their CVS tree but have not updated their
>/usr/src tree.

Would it work for me?  I have a local cvs tree which I update using
cvsup.  It exists in a separate partition, /usr/cvs

The machine is a dual-boot machine (-stable and -current).  Both systems
mount the same /usr/cvs partition, and use it to update their respective
/usr/src's.  How does each /usr/src get the correct serial-number, if the
serial-number file is not in cvs?

I use this same tactic on both my machines (at work and at home), so I
can track multiple OS images and yet only have to contact a cvsup server
once per machine.  Sometimes those multiple images are not separate
branches, but just separate time-stamps in the same branch.

>In general, I don't consider this to be a bad thing, since otherwise any
>solution you had would have to keep a complete running history of all
>the revisions, which is IMO unacceptable.  (JDP has also stated his
>dislike for such a file.)

I agree that having the serial-number file in CVS is distasteful, due to
how often it will be changed, but it seems to me that is the only way to
have things work right for people who are using a local cvs repository.
/usr/src has to end up with the right serial number for that /usr/src,
and not the most recent one stuffed into the cvs repository.

However, I am not a cvs wizard, so maybe it is possible to track the
serial numbers without having a serial-number file under cvs.  I don't
know, but I wanted to describe how I use cvsup so people could consider
how the serial-number would need to work in my situation.

-- 
Garance Alistair Drosehn            =   gad@eclipse.acs.rpi.edu
Senior Systems Programmer           or  gad@freebsd.org
Rensselaer Polytechnic Institute    or  drosih@rpi.edu

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




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?p05101401b87f6aaa81ab>