Date: Sun, 9 Jul 1995 19:21:36 -0700 (PDT) From: "Rodney W. Grimes" <rgrimes@gndrsh.aac.dev.com> To: peter@haywire.DIALix.COM (Peter Wemm) Cc: FreeBSD-current@FreeBSD.Org (FreeBSD current), jkh@FreeBSD.Org, davidg@FreeBSD.Org Subject: Re: Version numbers of the different branches? Message-ID: <199507100221.TAA09904@gndrsh.aac.dev.com> In-Reply-To: <Pine.SV4.3.91.950710074134.8698B-100000@haywire.DIALix.COM> from "Peter Wemm" at Jul 10, 95 07:58:28 am
next in thread | previous in thread | raw e-mail | index | archive | help
[cc: trimmed to -current, jkh, davidg added as it will impact them
on the branches and need an ack from both of them before doing this]
>
> On Sun, 9 Jul 1995, Rodney W. Grimes wrote:
> > > > > How about making it either "2.2-BUILT-xxxxxx" or "2.X-BUILT-xxxxx" or
> > > > > even "2.X-CURRENT-xxxxxx".
> > > > >
> > > >
> > > > We should revert to something like it was before e.g. 2.2-Development.
> > > > Putting the build date is not really informative because it gives
> > > > no insight about the date of the files themselves...
> > > >
> > >
> > > You can tell that the files are older than a particular date, that
> > > is very useful.
> > >
> > > ("BUILT-950703" ? Hey, He hasn't got the "ls -s panic" patch that
> > > got comitted on july 5th !)
> >
> > That may be true, but you surely can not say that he has all patches
> > applied up to 950703 by that string, as he may have built it from
> > 950301 sources...
> >
> > Any way I would like to have this small bit of confusing mess cleaned
> > up and propose the following changes based on my understanding of
...
> > Comments?
>
> I'd certainly vote for something like that!
Thats 3 folks I have affirmation from, and no nack's so it is slated to
be done some time real soon now. (Oh, joy, 3 way branch commits, hope
no one is planning major work on newvers.sh :-)).
Jordan, I will need your help with the make release stuff to make sure
I get this right!
> How about, as part of the sup scanning routines, the routine stamps in
> the X.Y-DEVELOPMENT-yymmdd entrys into the freefall's
> /usr/src/sys/conf/newvers.sh right before the supscan?
Why don't we walk a little before we try to run. I prefer to get what
I have outlined implemented and working correctly, with some testing
done before we try to go to far with it. I will keep this in the back
of my mind as a wizzy bang feature I should make sure I don't preclude
by design.
> Another different option might be to get the cron job to touch newvers.sh
> immediately before the supscan, and get newvers.sh to read the timestamp
> on itself while running..
I am not sure that all distribution mechanisms correctly propogate all
time stamps :-(. But will keep this in mind for a future possiblility.
> IMHO, the date of the source release is useful, but the build date is
> potentially quite misleading.
Agreed, not only ``potentially'', it has infact caused problems between
people trying to diagnose a problem.
> Oh, and dont forget, there's X.Y-ALPHA, -BETA and friends.. :-)
Forget them I bloddy well, they are simple cvs tag's and are not branches,
I don't plan to change newvers.sh every time I tag the tree. Besides
10 seconds after -ALPHA the tag needs to revert to either -DEVELOPEMENT
or -STABLE. Besides only releng's can cut official ALPHA/BETA/RELEASE
code so this will be handled by the ``make release SNAP= | RELEASE=''
in /usr/src/release.
--
Rod Grimes rgrimes@gndrsh.aac.dev.com
Accurate Automation Company Reliable computers for FreeBSD
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199507100221.TAA09904>