Re: [exim-dev] Git unknotting

Top Page
Delete this message
Reply to this message
Author: Nigel Metheringham
Date:  
To: exim-dev
Subject: Re: [exim-dev] Git unknotting
Github won't let you push --force unless you unprotect the branch first...

OK, both repos have the right version now, and the sync process between
them is working again. It could have been much worse - if there had
been a subsequent series of changes pushed to different sides it would
have been fun to sort out.

    Nigel.


> Heiko Schlittermann <mailto:hs@schlittermann.de>
> 7 April 2016 at 12:03
> Hi,
>
> Nigel Metheringham <nigel@???> (Do 07 Apr 2016 12:55:04 CEST):
>> Hi,
>>
>> I've just had to unknot git a little bit - the HEAD commit on the github
>> and hummus respositories were different versions of the same commit.
>>
>> I presume that Heiko had amended a commit that was previously pushed to
>> one of the repos, and the periodic process that syncs the two repos
>> could not determine which was the correct HEAD to use.
>
> Yes, it was my fault.
>
>     commit
>     push
>     commit --amend
>     push --force

>
>> Please could you check that I have ended up with the correct commit
>> (03f68c52bd7d014304b03ad6394343350ea71b2c) as the top one, as both repos
>> now use that commit.
> …
>> The difference between them was whether 0 or FALSE was used in the
>> changed line.
>
> Almost.
>
>     dns_init(0, 0)
> vs
>     dns_init(FALSE, FALSE, FALSE); 

>
> The latter being the right one (03f68c52bd7d014304b03ad6394343350ea71b2c)
>
> Sorry for the mess.
>
>     Best regards from Dresden/Germany
>     Viele Grüße aus Dresden
>     Heiko Schlittermann
> Nigel Metheringham <mailto:nigel@dotdot.it>
> 7 April 2016 at 11:55
> Hi,

>
> I've just had to unknot git a little bit - the HEAD commit on the
> github and hummus respositories were different versions of the same
> commit.
>
> I presume that Heiko had amended a commit that was previously pushed
> to one of the repos, and the periodic process that syncs the two repos
> could not determine which was the correct HEAD to use.
>
> Please could you check that I have ended up with the correct commit
> (03f68c52bd7d014304b03ad6394343350ea71b2c) as the top one, as both
> repos now use that commit.
>
> The two commits involved are:-
>
> * 03f68c52bd7d014304b03ad6394343350ea71b2c
> * c71d2667807bacf3cf54c05c79fa12453bb30679
>
> The log entry for both was:-
> Author: Heiko Schlittermann (HS12-RIPE) <hs@???>
> Date: Thu Mar 24 16:01:15 2016 +0100
>
>     Init the resolver even in -be mode.

>
>     This gives us consistent behaviour if we use the resolver in -be
> and in
>     -bh, and probably other modes. (Perl uses the resolver library linked
>     into Exim and Perl's Socket6::getaddrinfo() returned different results
>     in -be and -bh mode.)

>
> The difference between them was whether 0 or FALSE was used in the
> changed line.
>
> Cheers
>     Nigel.


--

[ Nigel Metheringham ------------------------------ nigel@??? ] 
[                 Ellipsis Intangible Technologies                  ]