Re: implementing basic metadata in cvs using tags

[Greg A. Woods]

[ On Sunday, October 14, 2001 at 16:09:36 (+1000), Matthew Herrmann wrote: ]
> Subject: implementing basic metadata in cvs using tags
>
> why not have registered programs on the client side which mangle permissions
> and so on into a sort of UUencoded string, put into a file's tag, which is
> then read by the same program and reapplied when the file is checked out?

Well, as Paul said there are problems with your proposal.

On major problem is the clutter it would cause.  Paul mentioned some of
the practical issues that would result.

One of the issues with tags is that they cannot really be used to
refer to more than one revision (at least not unless all the revisions
in question would happen to exist on a branch).

There's also a practical limit to the length of a tag (though none
explicitly imposed by rcsfile(5)).

> 3. Is this useful? It seems to me to solve the whole metadata problem -- you
> can store file permissions for different platforms and other random
> information that may be required.

Keep in mind that there's really only one bit of permissions information
that is even potentially of any use in a shared revision tracking tool,
and that's the 'execute' bit.  Nothing else makes any sense in this
context, especially not on Unix.  Any other bits are the responsibility
of the build system -- they make no sense whatsoever in the context of
the source code files managed by the revision tracking tool, no matter
what form the "code" takes.  You cannot think clearly about this issue
unless you always consider the larger SCM process.

-- 
                                                        Greg A. Woods

+1 416 218-0098      VE3TCP      <address@hidden>     <address@hidden>
Planix, Inc. <address@hidden>;   Secrets of the Weird <address@hidden>