Re: [PATCHv2] pull: honor submodule.recurse config option

From: Junio C Hamano <gitster@pobox.com>
Date: Sep 7, 2017, 00:54 UTC
Message-ID: <xmqqbmmn2ur7.fsf@gitster.mtv.corp.google.com>
In-Reply-To: <c3842980-22ef-5a8c-2895-90c48a97ed71@suse.de>

Nicolas Morey-Chaisemartin <NMoreyChaisemartin@suse.de> writes:

Show 6 quoted lines

>> If it is not the latter, perhaps we may want to flip the order of
>> config parsing and option parsing around?  That will allow us to fix
>> the handling of autostash thing to use only one variable, and also
>> fix your patch to do the right thing.
>
> I see what you mean.

Show 8 quoted lines

> It looks like switching the code around works but I think there
> still needs to be 2 variables for autstash for this piece of code:
>
>     if (!opt_rebase && opt_autostash != -1)
>         die(_("--[no-]autostash option is only valid with --rebase."));
>
> The config option should not cause git pull to die when not using
> --rebase, the CLI option should.

Ah, OK. That is a worthwhile observation that needs to be recorded in the log message of a commit that flips the order of option/config parsing.

Thanks.