Julian Foad <julianfoad_at_btopenworld.com> writes:
> A "reintegrate" merge will never try to delete "locally modified"
> items because it requires an unmodified target tree. It will never try
> to delete "unversioned" items because it requires the source tree to
> be already caught up with the latest versioned state of the target
> tree.
>
> So is the "force" argument appropriate for this API? It seems to me
> that it is redundant, and certainly it shouldn't be there if that's
> true.
I think you're right, though a merge could result in
formerly-versioned things in the target becoming unversioned (because
they've been removed on the branch). There's no need to protect that
with a flag, though -- it's not dangerous, because unversioning
something won't cause the working file to be removed, and we already
know it's not modified, so we're not losing a diff.
> --- subversion/include/svn_client.h (revision 29422)
> +++ subversion/include/svn_client.h (working copy)
> @@ -2517,7 +2517,8 @@
>
>
> /**
> - * Perform a reintegration merge of @a source into @target_wc_path.
> + * Perform a reintegration merge of @a source at @a peg_revision
> + * into @target_wc_path.
> * @a target_wc_path must be a single-revision, @c svn_depth_infinity,
> * pristine, unswitched working copy -- in other words, it must
> * reflect a single revision tree, the "target". The mergeinfo on @a
+1
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe_at_subversion.tigris.org
For additional commands, e-mail: dev-help_at_subversion.tigris.org
Received on 2008-02-27 23:18:03 CET