Use of CVS in the context of Unix server, unix based developers and Windows (VC++) based developers.

[Arcin Bozkurt]

Hi everyone.
The problem at hand uses the following tools:

Unix server running cvs 1.11.1p1
Linux developers using cvs 1.11.1p1 and using Emacs to edit code (latest
emacs)
Windows developers using WinCvs1.2 for CVS interface and VC++ to edit code.

The problem is the following: (as much as I could figure out)

Everything was fine, when everyone was working in Unix.
When windows developers started changing code and committing them, the
checkout on Linux and the diff on Linux started showing ^M characters for
the lines updated. The diff in this context is using the PCL-CVS in Emacs.
(I am also using the NT Emacs, the behaviour is the same)

In a simple scenario:
Windows codes are committed.
WinCVS checkout.
Open emacs and do cvs-status on the directory (PCL-CVS).
  - At this point Emacs seems to update the files and convert the linefeed
characters. The buffer is shown as a (Unix) encoded buffer and then WinCvs
starts showing these files as "Modified".


This is a combination of a lot of things. Any one any ideas on how to solve
this issue? I experimented a bit with the : "Checkout text files with the
Unix LF (0xa)" options but could not get a consistent behaviour that will
solve my problem...

Any suggestions are greatly appreciated.

-arcin