Re: move some test-tools to 'unstable plumbing' built-ins

From: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Date: Nov 8, 2021, 20:19 UTC
Message-ID: <211108.86zgqee6rm.gmgdl@evledraar.gmail.com>
In-Reply-To: <xmqq7ddiv4oi.fsf@gitster.g>

On Mon, Nov 08 2021, Junio C Hamano wrote:

Show 12 quoted lines

> Taylor Blau <me@ttaylorr.com> writes:
>
>> In my experience I *rarely* rely on test-helpers when debugging wedged
>> repositories, and much more often end up either in gdb, or in an
>> anonymized copy of the repository on a different server. I would imagine
>> that others have similar experiences.
>>
>> So unless we had a much more compelling reason to have the test helpers
>> more readily available, I do not think that the risk our users will
>> begin to depend on these unstable tools is worth taking.
>
> OK.  It sounds like a sensible argument against such a change.

It's an argument against not flipping "make installing them be optional" flag on by default, but we could otherwise move some of t/helper to builtin/, which would help by encouraging us to write at least boilerplate docs for them.

Git developers & similar parties could then set them to be installed for ad-hoc debugging.

Whatever anyone things on that, just on Taylor's "begin to depend on"...

I really don't buy the argument that there's no amount of warnings in our documentation that we can include which would give us future license to willy-nilly change certain things.

If that is being argued then that seems to categorically exclude certain other things, e.g. including "scalar" in-tree at all, because if we can't trust users to read the warnings about it being "contrib-y"...