[Dune-devel] [Dune-Commit] [branch] dune-istl - 'master' updated

Tue Jun 9 18:59:58 CEST 2015

Hi,
first of all thanks to Ansgar and Jö for their explanations.

After trying this out on an example I meanwhile understood
this myself. Since the merge commit of the master->feature
has master as one of it's parents, a subsequent feature->master
merge is fast forward. Because the pointer can just be moved
along the history.

While I understand it now, I looks really awkward to me
and I'm tempted to propose that we introduce the policy
to always do a manual non-fast-foward merge in this situations,
because it makes tracking the nonlinear history much easier:

A normal 'git log' will show you the commits from all parent
branches. Instead of this you ofter really want

  git log --first-parent

which shows the merge commits itself but excludes individual
commits coming with these merges. However in the situation
described below it will show you the history branch from
feature/FS1622 and not the one from master which is really nasty.

Best,
Carsten

Am 09.06.2015 um 18:35 schrieb Jö Fahlke:
> Am Tue,  9. Jun 2015, 18:05:32 +0200 schrieb Carsten Gräser:
>> Date: Tue, 09 Jun 2015 18:05:32 +0200
>> From: Carsten Gräser <graeser at mi.fu-berlin.de>
>> To: dune-devel at dune-project.org
>> Subject: Re: [Dune-devel] [Dune-Commit] [branch] dune-istl - 'master'
>>  updated
>>
>> Dear all,
>> I'm just trying to understand a bit of git history cited below.
>> I'd really appreciate if someone with deeper git insight could
>> explain what's going on here:
>>
>> All the commits from 0e3ddd4 to 6fdf16c went originally
>> into the feature branch 'feature/FS1622' with the same
>> revision numbers. After that the following happened:
>>
>> According to the history the following happened:
>>
>> 1) 0e3ddd4..6fdf16c: These commits where pushed into 'feature/FS1622'
>> 2) 1246f: Merge 'feature/FS1557' into 'master'
>> 3) 287b0: Merge 'master' into 'feature/FS1622'
>> 4) All commits from 'feature/FS1622' went into master
>>    _with the same revision numbers_ (see below)
>>
>> I really don't understand how 4) could happen, or
>> in other words, what action generated the e-mail
>> below. I can see two options:
>>
>> A)Someone did a local rebase of 'feature/FS1622' on top
>>   of master and pushed this. But then the revision numbers
>>   of those commits would differ for 'master' and 'feature/FS1622'
>>
>> B)Someone merged 'feature/FS1622' into 'master'. Then we
>>   would normally see a corresponding merge commit in 'master'
>>   instead of these individual commits. The only possibility seems
>>   to be a fast-forward merge that would keep these commits.
>>   But how can this be if 'master' changed meanwhile in 2) ?
>>
>> I'm waiting for enlightenment.
> 
> I think 4) is B) with a fast-forward (and consequently no new merge commit,
> master is just updated to a descendent commit).  But why do you think 2) is
> happening meanwhile?  If 1)..4) happen in their numerical order, that should
> explain everything.
> 
> Regards,
> Jö.
> 
>> Best,
>> Carsten
>>
>>
>> Am 12.05.2015 um 11:42 schrieb Andreas Dedner:
>>> Andreas Dedner <dedner at dune-project.org> updated branch 'master' (1246f56 -> 287b0ae)
>>> at Tue May 12 11:42:16 2015 +0200
>>>
>>> Browsable log: http://cgit.dune-project.org/repositories/dune-istl/log/?h=master&qt=range&q=1246f567674ca1b0242e6ba02d67be3a89adfecb..287b0ae90a8b8200c0614a66b64a4d24a7a2b53b
>>> Complete diff: http://cgit.dune-project.org/repositories/dune-istl/diff/?id=287b0ae90a8b8200c0614a66b64a4d24a7a2b53b&id2=1246f567674ca1b0242e6ba02d67be3a89adfecb
>>>
>>> ======================================================================
>>>
>>> via      0e3ddd4 http://cgit.dune-project.org/repositories/dune-istl/commit/?id=0e3ddd4c219d42ffe424b8b8896b1e6883badd7c
>>>              [bugfix] "properly" from field_type to int
>>> via      bb5854d http://cgit.dune-project.org/repositories/dune-istl/commit/?id=bb5854d6212860e354a42307e532fda5156ceb59
>>>              [bugfix,paamg] update norm to return a real number, even when working with complex
>>> via      e271e4d http://cgit.dune-project.org/repositories/dune-istl/commit/?id=e271e4da6792f45dbe8c1edc25d188d8d226f1ea
>>>              [paamg] generalize the Diagonal norm for complex values
>>> via      beaa066 http://cgit.dune-project.org/repositories/dune-istl/commit/?id=beaa0661568de62d908cbdb4c4c626ba433129f9
>>>              [bugfix] add missing typedef
>>> via      dcf4195 http://cgit.dune-project.org/repositories/dune-istl/commit/?id=dcf419597b209543ea1aece7c2ccc0949e101c8f
>>>              [bugfix] fix typo in csgn
>>> via      6fdf16c http://cgit.dune-project.org/repositories/dune-istl/commit/?id=6fdf16cb76a5e20df624233ef91543fb904c2802
>>>              fix some minor issues with amg and complex extended the amgtest to also use field_type
>>>              = complex by switching the definition of XREAL
>>> via      287b0ae http://cgit.dune-project.org/repositories/dune-istl/commit/?id=287b0ae90a8b8200c0614a66b64a4d24a7a2b53b
>>>              Merge branch 'master' into feature/FS1622
>>>
>>> _______________________________________________
>>> Dune-Commit mailing list
>>> Dune-Commit at dune-project.org
>>> http://lists.dune-project.org/mailman/listinfo/dune-commit
>>>
>>
>>
> 
> 
> 
>> _______________________________________________
>> Dune-devel mailing list
>> Dune-devel at dune-project.org
>> http://lists.dune-project.org/mailman/listinfo/dune-devel
> 
> 

-- 
Prof. Dr. Carsten Gräser
Freie Universität Berlin
Institut für Mathematik
Arnimallee 6
14195 Berlin, Germany
phone: +49 30 838 72637
fax  : +49 30 838 472637
email: graeser at mi.fu-berlin.de
URL  : http://page.mi.fu-berlin.de/graeser

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 473 bytes
Desc: OpenPGP digital signature
URL: <https://lists.dune-project.org/pipermail/dune-devel/attachments/20150609/99391c1d/attachment.sig>