Re: [PATCH] config: Introduce GIT_CONFIG_NOGLOBAL

From: Junio C Hamano <gitster@pobox.com>
Date: Apr 8, 2021, 23:43 UTC
Message-ID: <xmqq35w0qr5f.fsf@gitster.g>
In-Reply-To: <YG+PLROZbSo32F3k@coredump.intra.peff.net>

Jeff King <peff@peff.net> writes:

Show 12 quoted lines

>>  (3) when set to any other string, it is taken as a filename that
>>      has the global configuration.  Unlike $HOME/.gitconfig or
>>      $XDG_HOME/git/config, it is an error if the named file does not
>>      exist (this is to catch misconfiguration).
>> 
>> And once this is accepted by users and established as a pattern, we
>> could migrate GIT_CONFIG_NOSYSTEM to GIT_CONFIG_SYSTEM=/dev/null
>
> That seems pretty reasonable. I'm on the fence on your (3). Conceivably
> somebody could want to override the baked-in defaults without being sure
> the file is present. But I'm not sure how useful that would be in
> practice.

I was also on the fence.

If your plan is to create $HOME/.alternate-config and point at it by setting GIT_CONFIG_GLOBAL=$HOME/.alternate-config, there are two places you can make typo. You may write a file with a wrong name. You may export a variable with a wrong name.

Show 6 quoted lines

> Some other things to consider:
>
>   - does setting GIT_CONFIG_GLOBAL override both the $HOME and
>     $XDG_CONFIG_HOME? If the plan is to override them, that makes sense.
>     But we do usually read from both of them, so conceivably you might
>     want to override just one? That's probably over-engineering, though.

I viewed this to be working at the more conceptual "here is the file to read 'system' (or 'per-user') stuff from" level, and not at the level of the individual file, as I consider that it is a mere implementation detail that 'per-user' may read from multiple files.

Show 9 quoted lines

>   - if we have config for "read from this file instead of the system
>     config" and "read from this instead of the user-level config", then
>     I wonder if people will want "read this instead of the repo config".
>     We have resisted having GIT_CONFIG mean that for many years, because
>     I think it gets awkward in some cases (e.g., we'd still want to read
>     it for core.repositoryformatversion, etc). I'm OK with drawing the
>     line there and saying it's not a support feature, but we should be
>     prepared to explain it to users (in the docs or at least in the
>     commit message adding the system/global override variables).

We may have to bite the bullet and make an official catalog of really "structurely fundamental" configuration variables that must appear in the per-repository file, and a mechanism to enforce that by always reading these variables from "$GIT_DIR/config" and always ignoring appearances of them in any other places.

That alone would probably be a good thing to do regardless of the GIT_CONFIG_NOGLOBAL issue, as I suspect you may be able to wreak havoc by adding random configuration like [extension] in $HOME/.gitconfig ;-)

With that, it would make sense to allow overriding the per-repo configuration in a similar way, only as a mechanism to override configuration variables that are about "preferences" (as opposed to the structurally fundamental ones).