On 30 Sep 2011, at 15:19, Sanne Grinovero wrote:

Hi all,

the master branch is not compiling since the last commit "b6de38a",

but even before that it's not very clean: I'm seeing many branch
merges, history is an mess.

You must be referring to:

de98c55 - Mircea Markus, 22 hours ago : Merge branch 'master' into t_ISPN-1131_final

166ec14 - Mircea Markus, 3 days ago : SPN-1131 - merge with master

Now there's a reason why merge was used there and not rebase. If you add a new file in a topic branch and modify it in several commits on the same topic branch, then during rebase git sees all these individual commits on the same files as conflicts, which makes rebase a real PTA. Optimistic locking impl had 52 commits, most of them touching same classes.

I'm going to recommend again to use Emmanuel's nice build script
https://gist.github.com/789588 which verifies you're actually testing
the commit and not depending on other state,
and to really avoid using the web interface only to do reviews and
handle pull requests.

I know that's adding a bit of a burden to the reviewer's time, but
that's why the original committer should polish his patch in a logical
sequence and add enough commit comments to ease the reviewer's work as
much as possible; otherwise it seems that more often the reviewer does
more work than the original developer, who is not even taking
responsibility for his code as someone else will check it.

I'm not angry :) just worried the discipline is getting a bit too low
for the very demanding standards we aim for and calling for some more
attention.

+1. My only concern is that during the 10 mins in which I run the tests, my laptop cannot be used for development. Hopefully I'll get a new, more performant one next spring :)