Re: vc-mode permissions problems on NT

[Andre Spiegel <spiegel@gnu.org>]

On Tue, 2003-09-09 at 17:08, David Abrahams wrote:

> I normally have CVSREAD set in my environment so that files are
> checked out read-only by default.  After building emacs from the CVS
> HEAD as of a few days ago, `C-x v v' on an *un-edited* write-protected
> file under source control yields the following messages:
> 
>     File is edited but read-only; making it writable
>     File is under version-control; use C-x v v to check in/out

I think a quick workaround might be to set vc-stay-local to nil, and
maybe also vc-cvs-stay-local to nil.  This will cause VC not to rely on
its timestamp heuristics, and do a real "cvs status" during C-x v v.

I cannot reproduce your actual problem over here, but I have some
guesses.

Is this with a local repository, or a remote one?  Does the problem
happen for every file, even when it's freshly checked out from CVS? 
Before you do anything with the file in Emacs, do the timestamp in
CVS/Entries and the file's modification time agree?  When you visit the
file in Emacs (without doing anything yet), does the modeline show the
file as edited ("CVS:x.xx") or up-to-date ("CVS-x.xx")?

I have a suspicion that this might be related to the libc problem we've
recently seen, where file's modification times are incorrectly set to
1970-01-01.  If you see anything along these lines, please also let me
know.