On Thursday 27 November 2008, Andrei Pelinescu-Onciul wrote:
Author: Henning Westerholt henning.westerholt@1und1.de Committer: Henning Westerholt henning.westerholt@1und1.de Date: Thu Nov 27 14:00:30 2008 +0100
Merge commit 'origin/henning/trie'
- commit 'origin/henning/trie': add trie to library directory add SHM_MEM_ERROR and PKG_MEM_ERROR logging macros to mem.h
Aaaaaaahrg, this git is killing me. :-/ I just want to start a new branch from the master with:
It's ok, you didn't do anything wrong :-)
git push --dry-run --verbose origin origin:refs/heads/henning/trie_mods Pushing to ssh://git.sip-router.org/sip-router To ssh://git.sip-router.org/sip-router
- [new branch] origin/HEAD -> henning/trie_mods
This looked ok, and trie_mods branch is now created, but now i've made probably another mistake. Andrei, can you perhaps reset the master again? Why i got the same log message now? Also an issue with my git installation? I'll tomorrow update my system, this can not continue this way. :-(
No, there isn't any problem, the commit is on your trie_mods branch not on master. You see the commit message because of the email script we use and not because a mistake in the merge.
Hi Andrei,
i understand, thanks for the clarification.
The email script tries to send mail only for new commits, but if it doesn't find anything new (like in this case), it sends the top of the branch commit anyway as a failsafe (it doesn't have special support for detecting newly created branches). You should read this message as: henning/trie_mods created with head == displayed commit :-) We might extends the mail script in the future to send a better message in cases as above if people find the current behaviour too confusing (but for now I consider this quite low priority).
Yes, its not a problem if one knows about it.
For now our email script does not support:
- new branch created messages
- non fast forward updates of a branch
- branch deletion
For all of the above a commit message with the current head of the branch will be sent (for the delete it would be the old branch head and the branch name will have DELETED appended to it). Should I make the script public somewhere? Is anybody interesting in extending it?
Would be cool to have a better detection of the situation i just run into, but i fear that i lack the git skills at the moment to do this job.
Cheers,
Henning