On Mon, 02 Apr 2007, Daniel Rall wrote:
> On Mon, 02 Apr 2007, Peter Lundblad wrote:
...
> > > --- trunk/subversion/libsvn_client/merge.c (original)
> > > +++ trunk/subversion/libsvn_client/merge.c Fri Mar 30 19:41:09 2007
...
> > > +svn_error_t *
> > > +svn_client_get_mergeinfo(apr_hash_t **mergeinfo,
...
> > > + SVN_ERR(svn_client__open_ra_session_internal(&ra_session, entry->url,
> > > + NULL, NULL, NULL, FALSE,
> > > + TRUE, ctx, pool));
> >
> > Seems a bit unnecesary to open an RA session if this just retrieves the
> > merge info for BASE or WORKING.
>
> If the merge info for BASE is inherited, it will need to be retrieved
> from the repository.
>
> If the merge info for BASE or WORKING stored locally in the WC, the
> opening an RA session is indeed unnecessary. I will refactor
> get_wc_or_repos_merge_info() to open the RA session on demand, if one
> was not provided.
This was done in r24347.
- application/pgp-signature attachment: stored
Received on Mon Apr 2 23:24:11 2007