Re: working copy corruption when network connection lost

Karl Fogel <kfogel@newton.ch.collab.net> writes:

> Philip Martin <philip@codematters.co.uk> writes:
> > That said I don't really see what the ref_count achieves. It goes up
> > when subdirectories or files are opened, but these will be closed and
> > the ref_count decremented before the directory is closed. Is there
> > ever a case when close_directory calls decrement_ref_count and the
> > ref_count does not equal 1? Something to do with postfix text deltas
> > perhaps?
>
> [I think I'm just confirming Ben's confirmation of your guess...]
>
> Yes, it's for postfix text deltas. Those are the one exception to the
> "nestiness" of the editor call sequence -- you can have a directory
> closed before all the files in it are closed. This is what's actually
> going on over the RA layer, so it seemed best to have the editor
> reflect it, even if it costs some simplicity. Thus, we have to do ref
> counting. :-)

What determines when postfix text deltas are sent? The new diff
editors don't do reference counting, perhaps they should. They seem to
work at the moment, but then I don't appear to be getting postfix text
deltas.

-- 
Philip
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org