Re: advisory locks patch

[Derek Robert Price]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Paul Gelderblom (ptok) wrote:

>>I've completely rewritten the patch.  I really didn't like overloading
>>and parsing M responses (user messages) from the server when we have
>>control of the client/server protocol.  The basic functionality is still
>>the same, except that now failed edits will not block an entire command
>>but will only cause a single file to be skipped with a warning.
>>
>
>Does this re-write of the protocol also imply incompatibility with other
>clients
>that implement the "Noel Yap" version of the patch?
>That might mean that we loose compatibility with cvsnt-based clients.
>(including
>wincvs and TortoiseCvs).


No.  Other clients using the original method will still work.  The new
client, however, will only work with a new server.

The advantage of the new protocol is that there is no race condition
involved.  With the original patch, two users who attempted to edit a
file at approximately the same time might both succeed.

It would be possible to have the new code fall back on the "editors"
method when the server did not claim to be capable of an "edit"
request, but I wasn't sure of the value of that and did not code it.

Cheers,

Derek

- --
                *8^)

Email: derek@ximbiot.com

Get CVS support at <http://ximbiot.com>!
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFBWHE6LD1OTBfyMaQRApnZAJ44LJNPDHV/9CDlSKES+TBdvNC6dgCgq3t+
Ke/pCnePyn6se8H+SHQz6PE=
=CaBK
-----END PGP SIGNATURE-----