bug#4451: 23.1; EOL problems with vc-diff and cygwin

[Reiner Steib <reiner.steib@gmx.de>]

Eli Zaretskii wrote:

> What I'd like to see is where in Emacs sources we examine the output
> we get from Diff, and where and why we err as to what EOL format
> should be used for decoding that output.

`vc-coding-system-for-diff' is called 2 times when I do `C-x v ='

1)
   vc-coding-system-for-diff("c:/Users/x123456/tmp/check-out/K3.xml")
   vc-diff-internal(t (CVS ("c:/Users/x123456/tmp/check-out/K3.xml"))  
nil nil t)
   vc-diff(nil t)
   call-interactively(vc-diff nil nil)

2)
   vc-coding-system-for-diff("c:/Users/x123456/tmp/check-out/K3.xml")
   vc-cvs-diff(("c:/Users/x123456/tmp/check-out/K3.xml") nil nil "*vc-diff*")
   apply(vc-cvs-diff (("c:/Users/x123456/tmp/check-out/K3.xml") nil  
nil "*vc-diff*"))
   vc-call-backend(CVS diff ("c:/Users/x123456/tmp/check-out/K3.xml")  
nil nil "*vc-diff*")
   vc-diff-internal(t (CVS ("c:/Users/x123456/tmp/check-out/K3.xml"))  
nil nil t)
   vc-diff(nil t)
   call-interactively(vc-diff nil nil)

> One possibility for this mistake might be that Diff produces
> inconsistent EOL format in its output, for example if Diff or its VC
> front-end outputs some headers that have Unix EOLs and then the actual
> diffs with DOS EOLs.

The repository file (K3.xml,v) has Unix EOLs.  But if I do a fresh
checkout, I get a file K3.xml with DOS EOLs (I think this is the usual
behavoir of the Windows cvs binaries[1] for text files unless you
specify the switch "-ko").  However, in my workflow I overwrite the file
with a Unix EOL file (exported from some application), do modifications,
diffs and check it in.

Probably this conversion is also the reason that "cvs diff --binary"
outputs "^M^M$" for the old file and "^M$" for the new file.  "cvs diff"
outputs consitent DOS EOLs (both diff markers and the text):

$ cvs diff SK3.xml | cat --show-all | grep -F -v '^M' | wc -l


> Another possibility is that somewhere along the chain of processing
> the output, we force EOL conversion to be Unix-style, instead of
> detecting EOLs dynamically, or maybe even forcing it to DOS (if we
> have clear evidence for doing the latter).

`vc-coding-system-for-diff' returns `utf-8-unix' in both calls.


Stefan Monnier wrote:
> Could it be that the RCS files accessed this way get an accidental
> LF->CRLF conversion done by the network-file-system?
> Seems pretty unlikely.  But could you try and copy (part of) the
> repository to a local directory and try the operation again, just to
> rule out any funny business from this side?

I copied the repository file to a local drive and also used local drive
working copy (drive c:).  But that doesn't make any difference.

Bye, Reiner

[1] $ cvs --version

Concurrent Versions System (CVS) 1.11.22 (client)

Copyright (C) 2006 Free Software Foundation, Inc.