Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 1 Dec 1998 11:10:47 -0800 (PST)
From:      Archie Cobbs <archie@whistle.com>
To:        imp@village.org (Warner Losh)
Cc:        obrien@NUXI.com, committers@FreeBSD.ORG
Subject:   Re: ports tree FROZEN now
Message-ID:  <199812011910.LAA27894@bubba.whistle.com>
In-Reply-To: <199812010833.BAA01703@harmony.village.org> from Warner Losh at "Dec 1, 98 01:33:32 am"

next in thread | previous in thread | raw e-mail | index | archive | help
Warner Losh writes:
> One suggestion that I've not seen yet:
> 
> Make a .cvs-lock file, check it in.  CVS would then look for this file
> in ., then .., etc up to the root of the tree when you want to commit
> something.  If the file is present, then it will splat its contents to
> stdout and ask "In light of the above, are you really sure you want to
> commit?"  For the ports freeze, the file would look like:
> 
> 	WE ARE IN PORTS FREEZE NOW.  ALL CHANGES MUST BE APPROVED BY
> 	ASAMI-SAN.  COMMIT PRIVS WILL BE REVOKED IF YOU COMMIT NOW.
> 
> 	In light of the above, are you really sure you want to commit?
> 
> I don't know about you, but I'd think twice about answering yes :-)
> 
> Basically, it is an advisory lock rather than a strict lock.  cvs
> commit has a -f option, which could be used by the release engineer to
> bypass this message (or an environment variable CVS_RELEASE_ENGINEER
> that would bypass this).
> 
> Sure, I can still check stuff into the tree, but at least I'd get a
> warning, or I'd get rained on for aughting to have known better if I'm
> silly enough to have CVS_RELEASE_ENGINEER in my environment and not
> pay extra special close attention to -current or whatever.

This system seems to make *a lot* more sense than:

 - Satoshi broadcasts an email which, because the world is not
   perfect, a couple of people miss or forget or whatever, and then

 - Several people check-in code when they're not supposed to

 - Satoshi has to yell at all of them

 - Same people read the email with Satoshi yeling at them

 - Same people un-check-in they code they checked-in before

 - Satoshi has to make sure the check-in's didn't screw anything
   up in the mean time

Which is what we witnessed in the past few days..

-Archie

___________________________________________________________________________
Archie Cobbs   *   Whistle Communications, Inc.  *   http://www.whistle.com

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



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199812011910.LAA27894>