Date: Wed, 4 Sep 1996 13:45:30 -0700 (MST) From: Terry Lambert <terry@lambert.org> To: nate@mt.sri.com (Nate Williams) Cc: terry@lambert.org, rkw@dataplex.net, current@freebsd.org Subject: Re: Latest Current build failure Message-ID: <199609042045.NAA07361@phaeton.artisoft.com> In-Reply-To: <199609041904.NAA01512@rocky.mt.sri.com> from "Nate Williams" at Sep 4, 96 01:04:25 pm
next in thread | previous in thread | raw e-mail | index | archive | help
> As stated, it's simply not a problem worth messing with. (BTW, early > versions of CVS used a single 'write-lock' scheme, but after testing it > was considered overkill hence the 'new' scheme which used directory-lock > schemes.) A recent discussion of this point went on in devel-cvs. You and I both know that a directory lock schema fails when there is implied context spanning multiple directories. For instance, when a manifest constant is changed in a header file, but used in multiple source files in multiple directories. The problem with tree inconsistency at present is that it is possible to copy the tree while it is in transit from an inconsistent to a consistent state after starting in a consistent state. This assumes that there is policy in place to guarantee an initial consistent state (a supposition you seem unwilling to grant, given the material you chose to quote/not-quote in your response). Other than process failing to accommodate the possibility of human error, the is the only failure mode that can result in "hel, -current doesn't build!" mail messages. Note the use of the word "can". Would you argue that, because a race condition does not frequently cause problems, the race condition should not be eliminated? Before you answer, note that you are condemned by your own words from the argument that it would burden developers -- if the problem doesn't happen, then the conflict condition that would "burden" them will not occur. It is only a failure-mode saftey protocol. Regards, Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199609042045.NAA07361>