Re: [PATCH v2 3/3] rev-parse: match ^{<type>} case-insensitively

From: Junio C Hamano <gitster@pobox.com>
Date: Mar 27, 2017, 05:39 UTC
Message-ID: <xmqqk27b2s8p.fsf@gitster.mtv.corp.google.com>
In-Reply-To: <20170327025818.pbm4nwqrfolyto7g@sigill.intra.peff.net>

Jeff King <peff@peff.net> writes:

> FWIW, I cannot see us ever adding TREE (or Tree) as a separate type.
> It's too confusing for no gain. We'd call it "tree2" or something more
> obvious.

In case it was not clear, I didn't mean to say I _want_ to leave that door open. Well, I cannot imagine it was unclear, as I said I do not at all mind declaring that all object names will be lowercase to allow us freely downcase what we got at the UI level.

> So I don't mind closing that door, but I'm not sure if a partial
> conversion (where some commands are case-insensitive but others aren't
> yet) might not leave us in a more confusing place.

Exactly.

> I dunno. I guess I have never wanted to type "^{Tree}" in the first
> place, so I do not personally see the _benefit_. Which makes it easy to
> see even small negatives as a net loss.

As to the potential _benefit_, I do not see much either myself, but we already are seeing somebody cared enough to throw us a patch, so to some people there are clearly perceived benefit. I do not think closing the door for typenames that are not lowercase is a negative change at all.

I just wanted the patch to make it clear that it is making such a system-wide design decision and casting it in stone. Which includes that "cat-file <type>" and "hash-object -t <type>" get the same case-insensitivity update and probably writing that design decision down somewhere in the documentation, perhaps in the glossary where we talk about the "object type".