Re: [PATCH 7/7] push: document --lockref

[Johannes Sixt <j6t@kdbg.org>]

Am 12.07.2013 00:14, schrieb Junio C Hamano:
> Johannes Sixt <j6t@kdbg.org> writes:
> 
>> Again: Why not just define +refspec as the way to achieve this check?
> 
> What justification do we have to break existing people's
> configuration that says something like:
> 
> 	[remote "ko"]
> 		url = kernel.org:/pub/scm/git/git.git
>                 push = master
>                 push = next
>                 push = +pu
>                 push = maint
> 
> by adding a _new_ requirement they may not be able to satisify?
> Notice that the above is a typical "push only" publishing point,
> where you do not need any remote tracking branches.

Why would it break? When you do not specify --lockref, there is no
change whatsoever.

To achieve any safety at all for these push-only refs, you have to be
very explicit by saying --lockref=pu:$myoldpu
--lockref=master:$myoldmaster etc, and what is wrong if in this case
--lockref semantics are applied, but only pu is allowed to be no-ff?

> I am not opposed if your proposal were to introduce a new syntax
> element that calls for this new feature, e.g.
> 
> 	[remote "ko"]
> 		url = kernel.org:/pub/scm/git/git.git
>                 push = *pu
>                 fetch = +refs/heads/*:refs/remotes/ko/*
> 
> but changing what "+" means to something new will simply not fly.

I still do not see why we need two different kinds of ways to spell the
same strong kind of override (--force and +refspec) under the presence
of --lockref, and why we need a third one (--allow-no-ff) to give a
weaker kind of override (that makes sense only when --lockref was given
in the first place).

-- Hannes