From: Brent (brent@eng.sun.com)
Date: 11/05/99-01:59:35 AM Z
Message-ID: <38228E67.83662EB4@eng.sun.com> Date: Thu, 04 Nov 1999 23:59:35 -0800 From: Brent <brent@eng.sun.com> Subject: Re: minor versioning revisited I like Mike Eisler's rules for minor versioning. However, I'd prefer not to see the minor version required in the COMPOUND4args. I think it's simpler as an attribute: minor_version The minor version number of the protocol. The first version is number zero. The client can query the server's minor version anytime with a GETATTR. Most likely on the first call (at mount time) along with lots of other server information. The client could use a VERIFY operator to abort a compound op if the server's minor version is too small. For instance, if the server was to use a compound op that required version 2 then it could prepend a "VERIFY minor_version 2" (I'm assuming the server would evaluate this attribute comparison as an inequality rather than as a strict match on equal). There's an advantage for client's that use only minor version 0. They don't need to query minor_version or VERIFY it - nor do they need to put the version number in the COMPOUND4args. Brent
This archive was generated by hypermail 2.1.2 : 03/04/05-01:47:51 AM Z CST