Paul Mossman wrote:

[...]

> I used 'git apply' because there was one change to make.  (Not in
> alarm-codes.sql.)  
> 
> In hindsight, 'git am' and 'git commit -a --amend' would have avoided
> the problem.
> 

I *always* use git am - in this way I have a commit (or a series of
commits) that clearly correspond to submitted patches. Also git is merging
(and not just applying) the patch for me which let's me use:

  git mergetool

if there are any conflicts.

I also commit all my changes on top of the patch: adding comments for
myself - in this way I have a record of everything that I had to change.
Once I am ready to commit to mainline I use:

  git rebase -i

to squash and reorder the commits in the way that would make sense for
someone looking at how the feature was implemented (and not someone
interested in who did what).

I can alway look at the original history of commits (including original
comment and my comments) by checking the branch reflog:

  git log -g <branch>

As clearly visible from subversion history all that does not stop me from
making stupid mistakes but it did save my skin more than once.
Damian


_______________________________________________
sipx-dev mailing list [email protected]
List Archive: http://list.sipfoundry.org/archive/sipx-dev
Unsubscribe: http://list.sipfoundry.org/mailman/listinfo/sipx-dev
sipXecs IP PBX -- http://www.sipfoundry.org/

Reply via email to