Greg Stein <gst...@gmail.com> writes:

> Rather than explicit acquire/release, I think we're trying to use
> svn_wc__call_with_write_lock(). That will ensure that we don't leave
> write locks around on an error exit.

I've used svn_wc__call_with_write_lock in the past but I feel that the
baton interface is cumbersome.  Since the commit code already used
explicit error tracking I decided simply to continue to use it.

-- 
Philip

Reply via email to