"pushed 0 revisions" incorrect

[Andrew Cowie]

Output from a bzr push-and-update run:

Updating saved location: bzr+ssh://centauri.lhr.operationaldynamics.com/export/web/com/operationaldynamics/research/bzr/java-gnome/mainline/
Pushed 0 revisions (0 mainline) to revno 373
/home/andrew/.bazaar/plugins/push_and_update/push_and_update.py:118: DeprecationWarning: bzrlib.transport.split_url was deprecated in version 0.90.
  (scheme, user, password, host,
running "ssh centauri.lhr.operationaldynamics.com bzr update /export/web/com/operationaldynamics/research/bzr/java-gnome/mainline/"
+N src/bindings/org/gnome/glib/package.html
 M .aspell.en.pws
 M Makefile
 M src/bindings/org/gnome/gtk/FileChooser.java
 M src/bindings/org/gnome/gtk/FileChooserButton.java
 M src/bindings/org/gnome/gtk/FileChooserDialog.java
 M tests/bindings/org/gnome/gtk/TestCaseGtk.java
 M tests/bindings/org/gnome/gtk/ValidateFileChoosing.java
 M tests/screenshots/Harness.java
 M tests/screenshots/org/gnome/gtk/SnapshotFileChooserDialog.java
All changes applied successfully.
Updated to revision 376.

"Pushed 0 revisions"? That's very strange... and from a user point of view, wrong.

AfC

[Martin Pool (mbp)]

bzr itself used to show this, and I believe we corrected it by returning a Result object from this method and printing it appropriately. If this only occurs with the plugin, then the plugin probably needs to be updated.

[Andrew Cowie (afcowie)]

[Yes, it only occurs when I use bzr push-and-update. I only just gave it a try. Ordinarily doing bzr push && ssh hostname bzr update very/long/path is what I do, and it doesn't report that strange message]

[John A Meinel (jameinel)]

A couple things...

It sounds like the "push" part of "push-and-update" had nothing to do. (Hence 0 revisions pushed). The "update" part did.

We got rid of the "N revisions pushed" message in 'bzr push' because it had times where it would go backwards. However, I really missed it, because it gave me an idea of how much changed, rather than just telling me where I ended up.

Would "copied 0 revisions" or something like that be better? Should that part just be suppressed when it is 0.

I'm not really sure what you consider a valid resolution to this bug.

This specific output tells you how many total revisions (including merges) which the old 'bzr push' message did not. The old value is the "(0 mainline)".

[Andrew Cowie (afcowie)]

Took a while to comment because I wanted to wait until I had some new code that I could guaruntee hadn't preivously made its way to the remote server.

push-and-update *definitely* said "pushed 0 revisions" when I just had 4 new revisions to publish.

4 != 0. This really *is* a bug, sorry.

As for what it should say, I'd kinda go for the "if push alone doesn't say anything, then this probably doesn't need to" but that's up to you.

AfC