Re: flatten-merge history

2014-10-27 Thread Andreas Schwab
Henning Moll  writes:

> Just a final question: Is it possible to keep the GIT_COMMITTER_DATE in
> all those rebases?

If you want that you need to work with git filter-branch.

Andreas.

-- 
Andreas Schwab, sch...@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html


Re: flatten-merge history

2014-10-26 Thread Henning Moll

Am 26.10.2014 um 20:02 schrieb Christian Couder:
Yeah. Henning, when interactively rebasing, in our editor, you should 
have something like: pick A pick P pick B pick Q pick C pick D pick R 
pick E which should work without any conflict. And then you can rebase 
the b1 and b2 branches on the resulting branch.


Thank you all very much. It worked!

I didn't know that i could also introduce new commits in the interactive 
rebase list. Just for the record, in case others need to do such a task:


intial situation:

P - - - Q - - - - - R   <-extern

A - - - B - - - C - D - - - E   <-master
 \   \
  M ...   \ <-b1
   \
W ...   <-b2

On master:
$ git rebase -i --root

(editor opens showing

pick A
pick B
pick C
pick D
pick E

modify this to

pick A
pick P
pick B
pick Q
pick C
pick D
pick R
pick E

save, quit. Now, master is already fixed. Now fix the branches:

$ git rebase --onto B' B b1
$ git rebase --onto D' D b2

Just a final question: Is it possible to keep the GIT_COMMITTER_DATE in 
all those rebases?


Thanks again
Henning
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html


Re: flatten-merge history

2014-10-26 Thread Christian Couder
On Sun, Oct 26, 2014 at 4:19 PM, Andreas Schwab  wrote:
> Henning Moll  writes:
>
>> 1. For P, A is the nearest prior commit on 'master'
>> 2. on master: git rebase -i A^
>> 3. change A from pick to edit. save. quit
>> 4. git merge P
>> 5. git rebase --continue
>>
>> From the perspective of 'master' this worked. But as all of the commits
>> have been rewritten, the branches b1 and b2 no longer refer to
>> 'master'. Branch b2, for example, still branches off at B and not B'.
>
> You only rebased master, so b1 and b2 were unchanged.  If you want to
> change b1 and b2 you have to rebase them as well.

Yeah. Henning, when interactively rebasing, in our editor, you should
have something like:

pick A
pick P
pick B
pick Q
pick C
pick D
pick R
pick E

which should work without any conflict.
And then you can rebase the b1 and b2 branches on the resulting branch.

Best,
Christian.
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html


Re: flatten-merge history

2014-10-26 Thread Andreas Schwab
Henning Moll  writes:

> 1. For P, A is the nearest prior commit on 'master'
> 2. on master: git rebase -i A^
> 3. change A from pick to edit. save. quit
> 4. git merge P
> 5. git rebase --continue
>
> From the perspective of 'master' this worked. But as all of the commits
> have been rewritten, the branches b1 and b2 no longer refer to
> 'master'. Branch b2, for example, still branches off at B and not B'.

You only rebased master, so b1 and b2 were unchanged.  If you want to
change b1 and b2 you have to rebase them as well.

Andreas.

-- 
Andreas Schwab, sch...@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html


Re: flatten-merge history

2014-10-26 Thread Henning Moll

Am 26.10.2014 um 06:46 schrieb Christian Couder:

Hi,Could you tell us why interactive rebase did not work?


First of all i have to admit that i did not i-rebase-merge, but 
i-rebase-cherrypick. Now i tried it with i-rebase-merge and it didn't 
work either.
Here's my workflow, which i repeated for all commits of 'extern' (in 
chronological order'). For simplicity let's assume that there is a 
parent commit for 'A'


1. For P, A is the nearest prior commit on 'master'
2. on master: git rebase -i A^
3. change A from pick to edit. save. quit
4. git merge P
5. git rebase --continue

From the perspective of 'master' this worked. But as all of the commits 
have been rewritten, the branches b1 and b2 no longer refer to 'master'. 
Branch b2, for example, still branches off at B and not B'. So after 
this action, my history looks like this:



P - - - Q - - - - - R   <-extern
  \
A'- m - B'- - - C'- D'- - - E'  <-master

A - - - B - - - C - D
 \   \
  M ...   \ <-b1
   \
W ...   <-b2

A and A' are nearly the same: they only differ in the new 
GIT_COMITTER_DATE.

Now this is what i try to get (after only applying P):



P - - - Q - - - - - R   <-extern
\
A'- m - B'- - - C'- D'- - - E'  <-master
 \   \
  M'...   \ <-b1
   \
W'...   <-b2

Regards
Henning
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html


Re: flatten-merge history

2014-10-25 Thread Christian Couder
Hi,

On Sat, Oct 25, 2014 at 2:31 PM, Henning Moll  wrote:
> Hi,
>
> suppose the following history
>
> P - - - Q - - - - - R   <-extern
>
> A -- - B - - - C - D - - - E   <-master
>   \   \
>   M ...   \ <-b1
>   \
>   W ...   <-b2
>
>
> Note that master and extern do not have a common parent. Both histories are
> 'distinct', they do not share common files, so there can't be any merge
> conflicts. What i want to achieve is this history:
>
> P - - - Q - - - - - R   <-extern
>
> A -P'- B'- Q'- C'- D'- R'- E'  <-master
> \   \
> M'...   \ <-b1
> \
> W'...   <-b2
>
> The two histories should be merged in chronological order.
> So while master reflects P-Q-R, b2 should only reflect P-Q and b1 should
> only reflect P.
>
> All my current attempts (surgery with git replace or interactive rebase
> combined with merging) were not successfull.

Could you tell us why interactive rebase did not work?
If there can't be any merge conflict between both histories, it should
have worked without asking you to resolve any conflict.

Best,
Christian.
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html


flatten-merge history

2014-10-25 Thread Henning Moll

Hi,

suppose the following history

P - - - Q - - - - - R   <-extern

A -- - B - - - C - D - - - E   <-master
 \   \
  M ...   \ <-b1
   \
W ...   <-b2


Note that master and extern do not have a common parent. Both histories 
are 'distinct', they do not share common files, so there can't be any 
merge conflicts. What i want to achieve is this history:


P - - - Q - - - - - R   <-extern

A -P'- B'- Q'- C'- D'- R'- E'  <-master
 \   \
  M'...   \ <-b1
   \
W'...   <-b2

The two histories should be merged in chronological order.
So while master reflects P-Q-R, b2 should only reflect P-Q and b1 should 
only reflect P.


All my current attempts (surgery with git replace or interactive rebase 
combined with merging) were not successfull.


Any ideas?

Best regards
Henning


--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html