Re: Ignoring whitespace and CR/LF when checking into repository

[Mark D. Baushke]

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

Antony Paul <address@hidden> writes:

> What I mean by a server is something like pserver. My understanding
> about CVS is that it requires client only(downloaded from cvshome.org
> ) to create a repository and do checkout and checkin. I tried this in
> windows.
>     The repository that created this problem have no server running.

In a client/server environment, the triggers are all run on the server
in a temporary directory where your modified files have been copied.
This is true regardless of use of :ext: or :server: or :pserver: or
:gserver: or :kserver: as the server-side of the connection.

The same cvs executable may be configured and compiled to be a
client-only, server-only or non-client-server executable. In the typical
case, the default *NIX configuration enables both client and server
operations with the same executable. The default WindowsNT configuration
enables client and local operations, but does not provide the
server-side.

CVSNT has both client and server configurations for Windows boxen as
well as other host targets such as GNU/Linux.

        -- Mark

> rgds
> Antony Paul 
> 
> 
> On Fri, 19 Nov 2004 08:55:43 -0800, Mark D. Baushke <address@hidden> wrote:
> > -----BEGIN PGP SIGNED MESSAGE-----
> > Hash: SHA1
> > 
> > Antony Paul <address@hidden> writes:
> > 
> > > Does making files binary using -kb compromises any other
> > > functionality. Some developers want CVS to merge changes
> > > automatically. If the file is treated as binary will it merge ?.
> > 
> > No.
> > 
> > > Does any one have an example of a commitinfo filter which can identify
> > > ASCII files ?
> > 
> > That depends on what you mean... the commitinfo script could look at the
> > output of `cvs -n status filename` to see if the user has specified -kb
> > or not.
> > 
> > It might also be possible to use the output of `file filename` on *NIX
> > servers for new files that are being added to see if the user is doing
> > something unexpected. Of, you could write your own custom tool that
> > looked at the file, but do understand that the file copy on the server
> > side may already have been 'corrupted' in transit if the -kb switch was
> > not specified in the first place and the file really was a binary that
> > the client is not properly treating as a binary.
> > 
> > > Using commitinfo does requires a server ?.
> > 
> > The commitinfo trigger is run on the server.
> > 
> >         -- Mark
> > -----BEGIN PGP SIGNATURE-----
> > Version: GnuPG v1.2.3 (FreeBSD)
> > 
> > iD8DBQFBniWO3x41pRYZE/gRAh0RAKC65AzFeBAlSbbEFisr2fx3dVGLSgCgs9qL
> > zi/DFkjsCkZVxbuBWkTLSRQ=
> > =GuRj
> > -----END PGP SIGNATURE-----
> >
> 
> 
> _______________________________________________
> Info-cvs mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/info-cvs
> 
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (FreeBSD)

iD8DBQFBnxV03x41pRYZE/gRAo/GAKCWJLCK/QnXyTBOXoCg3Wt+2PBeywCggd9r
ELORr76Suja+BjJqiOr/CuU=
=rVOc
-----END PGP SIGNATURE-----