Re: [translations] lost committishes [WAS Re: Doc-it: revision of snippets in LM (please backport to 2.14)]

[Francisco Vila <paconet.org@xxxxxxxxx>]

Federico,

Here I try to explain some things in more detail.

2011/7/22 Federico Bruni <fedelogy@xxxxxxxxx>:
> Il giorno mer, 20/07/2011 alle 00.41 +0200, Federico Bruni ha scritto:
>> I've added the corrections of the proofreader and fixed the committish
>> of .texidoc files (which was a fatal bad object, probably because of
>> the
>> troubles we had in applying the previous patch).
>
> I think I was wrong here, because it's happened again.
> The problem is that Francisco has probably changed some mistake in my
> patch and then committed with a new commit.

No, I didn't, but continue reading.

> At least this is my guess,
> because after pulling I have double commits plus a merge:
>
> commit f42b010530f58c3817f4d4a985d76b2181b1f1df
> Merge: e11d8e7 c3b519f
> Author: Federico Bruni <fedelogy@xxxxxxxxx>
> Date:   Fri Jul 22 17:41:37 2011 +0200
>
>    Merge branch 'lilypond/translation' of git://git.sv.gnu.org/lilypond
> into lilypond/translation

This merge suggests that you had made and commited changes to your
branch before pulling.  Pulling involves a merge, in general.

> commit c3b519f0dd5ff0f8ccfc9a39ed1fe8df8b43741c
> Author: Federico Bruni <fedelogy@xxxxxxxxx>
> Date:   Wed Jul 20 00:17:06 2011 +0200
>
>    Doc-it: fix translations in snippets. Run makelsr.

This comes from your patch 0001 which I applied and pushed immediately.

> commit 4726227b76f1016098c10d0b20ef2062f931c26b
> Author: Federico Bruni <fedelogy@xxxxxxxxx>
> Date:   Wed Jul 20 00:25:44 2011 +0200
>
>    Doc-it: fix translation of snippets in LM

Also, this is what I applied and pushed directly from your patch 0002.

But looking at the content, these lines

-%% Translation of GIT committish: 4077120c18ac1dc490501b3d7d5886bc93e61a42
+%% Translation of GIT committish: 514674cb00c18629242dfcde0c1a4976758adc56

in every changed file indicate that you put a non-existant committish
on your files, in other words you "invented" the ID 514674cb00c :-)

In fact I have invented committishises many times, that only means
that I took them from a version that lived only in my hard disk,
unpublished, and everything looked fine until I reset --hard to a
version in Savannah, making those invented versions to vanish.

> commit e11d8e736858097eb4e400bfdb7f7593db25de1f
> Author: Federico Bruni <fedelogy@xxxxxxxxx>
> Date:   Wed Jul 20 00:25:44 2011 +0200
>
>    Doc-it: fix translation of snippets in LM

This could be your local commit previous to pulling.  Probably you
produced 0002 from here.  Why is it different from 4726227b, I don't
know.  Maybe I applied 0002, then 0001, and you had 0002 on top
instead.  You had X-A-B, you pulled X-B-A, the result is

                     A---B
                    /         \
               X---B---A---M

Where M is the merge commit produced by your pull.

e11d8e73 does not exist in Savannah, see

  http://git.savannah.gnu.org/gitweb/?p=lilypond.git;a=commit;h=e11d8e7368

All four commits A, B, B, A have different version IDs as you can see.

> commit 514674cb00c18629242dfcde0c1a4976758adc56
> Author: Federico Bruni <fedelogy@xxxxxxxxx>
> Date:   Wed Jul 20 00:17:06 2011 +0200
>
>    Doc-it: fix translations in snippets. Run makelsr.

Same as previous one.

In addition, your patches did apply cleanly on the
lilypond/translation branch, which does not mean that they could also
be applied cleanly on the stable/2.14 branch, in fact they couldn't
and I am sure that Carl had to resolve conflicts to get them applied.
The difference was an unequal count of blank lines in the files.

HTH
-- 
Francisco Vila. Badajoz (Spain)
www.paconet.org , www.csmbadajoz.com