Re: (no subject)

[Stefan Monnier]

> This bug has been around (partially) for at least several weeks though - I
> have been procrastinating re sending it off.

> When visiting a file that is under Subversion control I get the following in
> the *Messages* buffer and whilst the file is loaded into Emacs, it doesn't
> become visible until I switch buffers to it: 

> Loading vc-svn...done
> Loading vc...done
> vc-do-command: Running svn...FAILED (status 1)

Can you try to set debug-on-error (or maybe debug-on-signal) and get
a backtrace?  Also, what does "svn status -v" say when you run
it on the command line.  Supposingly, it should give some kind of error
message and return a status code of 1.

> But what does appear to be a "new" aspect of this bug is that when I modify
> the buffer and attempt to save the file, I get the same error message and
> the write does not happen i.e. Emacs will nolonger allow me to save the
> file. (well, I think this is "new" - it is quite possible that in the past I
> have plugged in the svn repository and made it available prior to attempting
> to save the file - see below)

Getting a backtrace for this case would be very helpful.
For the other case I think I know where the problem lies, but for this one,
I have no idea.

> The issue that is exposing the "bug" here is that the Subversion repository
> is on a USB Flash Disk and is not actually plugged into the PC when Emacs
> access to the file is attempted. Of course, when the repository is
> "available" then the problem does not arise, file access works as it should.

But Emacs always uses "svn status -v" which tells SVN to work locally
without contacting the repository.  Isn't it a bug in SVN, then?


        Stefan