Re: [PATCH v6 0/8] push: update remote tags only with force

Hi there,

I was just working on improving git-remote-helper.txt by documenting how remote helper can signal error conditions to git. This lead me to notice a (to me) surprising change in behavior between master and next that I traced back to this patch series.

Specifically:

On 30.11.2012, at 02:41, Chris Rorvick wrote:

This patch series originated in response to the following thread:

 http://thread.gmane.org/gmane.comp.version-control.git/208354

I made some adjustments based on Junio's last round of feedback
including a new patch reworking the "push rules" comment in remote.c.
Also refined some of the log messages--nothing major.  Finally, took a
stab at putting something together for the release notes, see below.

From the discussion in that gmane thread and from the commits in this series, I had the impression that it should mostly affect pushing tags. However, this is not the case: It also changes messages upon regular push "conflicts. Consider this test script:

#!/bin/sh -ex
git init repo_orig
cd repo_orig
echo a > a
git add a
git commit -m a
cd ..

git clone repo_orig repo_clone

cd repo_orig
echo b > b
git add b
git commit -m b
cd ..

cd repo_clone
echo B > b
git add b
git commit -m B
git push


With git 1.8.1, I get this message:

 ! [rejected]        master -> master (non-fast-forward)
error: failed to push some refs to '/Users/mhorn/Projekte/foreign/gitifyhg/bugs/git-push-conflict/repo_orig'
hint: Updates were rejected because the tip of your current branch is behind
hint: its remote counterpart. Merge the remote changes (e.g. 'git pull')
hint: before pushing again.
hint: See the 'Note about fast-forwards' in 'git push --help' for details.



But with next, I get this:


 ! [rejected]        master -> master (already exists)
error: failed to push some refs to '/Users/mhorn/Projekte/foreign/gitifyhg/bugs/git-push-conflict/repo_orig'
hint: Updates were rejected because the destination reference already exists
hint: in the remote.


This looks like a regression to me. No tags were involve, and the new message is very confusing if not outright wrong -- at least in my mind, but perhaps I am missing a way to interpret it "correctly" ? What am I missing?


Cheers,
Max