should show a message when pushing(overwriting?) tags

https://answers.edge.launchpad.net/bzr/+question/71544 asks for this.

Hello,

This would be a useful feature to have. I was about to file a new bug
when I read about this. I feel this applies to push as well as pull.

Below is the report I had already composed to be for filing as a bug:

While pushing/pulling tags bzr gives no indication
of it.

I applied tags to a local branch and pushed it to
our server using sftp. I get the "No revisions to push"
message. So I was not sure if the tags were pushed.

So for another branch I needed to tag I applied tags
directly to the repo on the server so that the tags can be
seen by everyone. On doing a pull locally I see the
"No revisions to pull" message but the tags command
shows that tags have been pulled.

[acpstable]% bzr pull
Using saved parent location: sftp://<email address hidden>/home/user0/repo/acpstable/
No revisions to pull
[acpstable]% bzr tags
v0.90.0 1455
[acpstable]%

It seems bzr is doing the right thing but as the user
does not get an indication of it its can be confusing.

It would be good if bzr can show some indication that
even though revisions are not pulled some meta data.
"Metadata updated", "Tags updated" or something.

This behaviour of silently pushing tags lead to some confusion on the Emacs development list...so please add some little display.

I decided to hack this a bit. I have the following working.

[foo]% bzr tag abc0
Created tag abc0.
[foo]% ~/src/bzr.dev/164450-push-pull-tags/bzr --no-plugins push ../bar
No new revisions to push.
Pushed 1 tags.
[foo]% ~/src/bzr.dev/164450-push-pull-tags/bzr --no-plugins push ../bar
No new revisions to push.
No tags to push.
[foo]%

.. but I am not too familiar with the way push/pull works, so I would
appreciate any help in understanding the existing design.

This is what I put together (push only right now). Updated methods are
branch.BzrBranch._basic_push, branch.BranchPushResult.report, and
tag.BasicTags._reconcile_tags.

Does this approach seem reasonable?
Would a similar approach apply to pull?

[bzrlib]% bzr cdiff -r -3
=== modified file 'bzrlib/branch.py'
--- bzrlib/branch.py 2010-01-12 02:48:41 +0000
+++ bzrlib/branch.py 2010-02-20 04:23:23 +0000
@@ -2328,7 +2328,7 @@
             target.update_revisions(self, stop_revision,
                 overwrite=overwrite, graph=graph)
         if self._push_should_merge_tags():
- result.tag_conflicts = self.tags.merge_to(target.tags,
+ result.tag_conflicts, result.copied_tag_count = self.tags.merge_to(target.tags,
                 overwrite)
         result.new_revno, result.new_revid = target.last_revision_info()
         return result
@@ -2831,6 +2831,10 @@
             note('No new revisions to push.')
         else:
             note('Pushed up to revision %d.' % self.new_revno)
+ if self.copied_tag_count == 0:
+ note('No tags to push.')
+ else:
+ note('Pushed %d tags.' % self.copied_tag_count)
         self._show_tag_conficts(to_file)

=== modified file 'bzrlib/tag.py'
--- bzrlib/tag.py 2009-06-10 03:56:49 +0000
+++ bzrlib/tag.py 2010-02-20 03:44:48 +0000
@@ -208,13 +208,13 @@
         to_tags.branch.lock_write()
         try:
             dest_dict = to_tags.get_tag_dict()
- result, conflicts = self._reconcile_tags(source_dict, dest_dict,
- overwrite)
+ result, conflicts, copied_tag_count = self._reconcile_tags(source_dict,
+ dest_dict, overwrite)
             if result != dest_dict:
                 to_tags._set_tag_dict(result)
         finally:
             to_tags.branch.unlock()
- return conflicts
+ return conflicts, copied_tag_count

     def rename_revisions(self, rename_map):
         """Rename revisions in this tags dictionary.
@@ -241,14 +241,18 @@
         """
         conflicts = []
         result = dict(dest_dict) # copy
+ copied_tag_count = 0
         for name, target in source_dict.items():
- if name not in result or overwrite:
+ name_not_in_result = name not in result
+ if name_not_in_result:
+ copied_tag_count += 1
+ if name_not_in_result or overwrite:
                 result[name] = target
             elif result[name] == target:
                 pass
             else:
                 conflicts.append((name, target, result[name]))
- return result, conflicts
+ return result, conflicts, co...

Setting to high, this can be very confusing.