From maintainers-request at octave dot org Tue Mar 21 10:51:27 2006
Subject: Re: octave on biowulf
From: John Swensen <jpswensen at comcast dot net>
To: "John W. Eaton" <jwe at bevo dot che dot wisc dot edu>
CC: maintainers at octave dot org
Date: Tue, 21 Mar 2006 09:49:49 -0700

John W. Eaton wrote:
> On 21-Mar-2006, Tom Holroyd (NIH/NIMH) [E] wrote:
>
> | As long as the CVS version number is different from the version on the website, that should avoid any future confusion.  Calling the CVS version 2.9.6 should be fine, and you don't have to increment the CVS version number each time, so long as there is no way to get 2.9.6 any place else.
>
> I don't think that will solve the problem because someone who installs
> the CVS version could easily mistake what they have for 2.9.6 when
> they really have some CVS version between 2.9.5 and 2.9.6.  So if they
> report a bug after the 2.9.6 snapshot is available, we don't really
> know what version they have.  It is essentially the same problem as
> now, but the version number is relative to a future snapshot instead
> of the previous one.  At least if we tag the CVS version with "-cvs",
> we will know to ask more about the date it was checked out/updated/built.
>
> jwe
>
>
>
>   
Just as an aside, Subversion has a solution for this.  I don't remember
the exact reason why you guys didn't want to move to subversion, but
since it has atomic commits, a revision number is given to each and
every snapshot of the repository.  I can't remember the command off the
top of my head, but that revision number can be easily retrieved from
the subversion metadata.

-John Swensen