X-Authentication-Warning: delorie.com: mail set sender to geda-user-bounces using -f
X-Recipient: geda-user AT delorie DOT com
Message-ID: <50AD2FED.4040300@xs4all.nl>
Date: Wed, 21 Nov 2012 20:47:57 +0100
From: Bert Timmerman <bert DOT timmerman AT xs4all DOT nl>
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.1.19) Gecko/20110429 Fedora/2.0.14-1.fc13 SeaMonkey/2.0.14
MIME-Version: 1.0
To: geda-user AT delorie DOT com
Subject: Re: [geda-user] branches
References: <CACwWb3DJsO8J_zAfcSLK=_4iviZy05ZiKj1TFMUUSA8nxO5TwA AT mail DOT gmail DOT com> <CAGde_xOFELRBJZ30utu1x+B-y+8Vo+EDGOKPLqyafO+i3ez0QQ AT mail DOT gmail DOT com> <CAC4O8c9Z3zmCeeeVU-=BpYTWv1J+3bBh-YcWyepV3U0nftf=gw AT mail DOT gmail DOT com> <50ACF554 DOT 4010204 AT jump-ing DOT de>
In-Reply-To: <50ACF554.4010204@jump-ing.de>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-Virus-Scanned: by XS4ALL Virus Scanner
Reply-To: geda-user AT delorie DOT com

Markus Hitter wrote:
> Am 20.11.2012 20:14, schrieb Britton Kerin:
>> 3. Bob rebases, which breaks something that he doesn't catch
> > 4. Bob pushes
> > 5. Its a hard mess to sort out since you don't have any
>> working version with Bob's changes in the canonical repository
>
> The idea is to rebase not the master branch, but feature/debug 
> branches and cherry-pick the reviewed parts of debug/feature branches 
> into master until the other branch dissolves into nothing. This way 
> you can bisect in detail on the master branch to find regressions.
>
> Merging tends to expose the problem you describe. A merge is like one 
> big commit with all the changes between two branches. When searching 
> regressions, you can go back on only one of the two branches. What, if 
> both branches work fine, but not the combination/merge of both?
>
>
> my $0.02
> Markus
>
Hi all,

FWIW, I (often) create a patch from a commit in a topic branch and apply 
that patch to master.

This keeps the same hash in both master and the topic branch, while 
cherry-picking gives a different (new) hash for the commit.

Dunno if this breaks git internal stuff though, never had any problems.

Kind regards,

Bert Timmerman.