Re: [PATCH] doc: ls-tree paths do not support wildcards

From: Junio C Hamano <gitster@pobox.com>
Date: May 28, 2020, 22:21 UTC
Message-ID: <xmqqy2pb1z0x.fsf@gitster.c.googlers.com>
In-Reply-To: <20200528215107.GA1265681@coredump.intra.peff.net>

Jeff King <peff@peff.net> writes:

> So I think for now we ought to explain the situation a bit more clearly:
> leave this language as-is, but add a new section describing what
> patterns we do support.

Thanks; you said a lot better than I could ;-)

Show 19 quoted lines

> In the long run it would be nice to actually match regular pathspecs.
> That would be a backwards-incompatibility, which I think is why nobody
> has pursued it further (and ls-tree is meant to be plumbing that should
> stay consistent, so we need to be extra careful). So we'd need a
> transition plan. Perhaps:
>
>   1. Deprecate the current behavior in the documentation and release
>      notes, encouraging people who want literal matching to use
>      --literal-pathspecs or the ":(literal)" magic. AFAICT we've
>      supported these since at least 2013 for this command, so it should
>      be safe to use unconditionally.
>
>   2. Add a new option, "--use-pathspecs" or similar, that switches the
>      matching code to use match_pathspec(). That lets people use the new
>      feature immediately if they want to.
>
>   3. When --use-pathspecs is not in use, warn to stderr about any
>      wildcard characters in the input. That reinforces the deprecation
>      notice in (1) and is likely to get more people's attention.

Hmph, if we are serious about deprecation and migration, I would image that in stage #1, we should do this check already. When "--literal-pathspecs" is NOT in use, if a pathspec would change its meaning if not taken literally (e.g. has glob letters, begins with :-magic introducer, etc.), we warn and do so from the very beginning of the migration process.

>   4. After several releases, flip the default to --use-pathspecs,
>      leaving --no-use-pathspecs as an escape hatch for people who still
>      haven't switched their scripts.

Wouldn't --literal-pathspecs be the accepted escape hatch that will always be accepted, even after --use-pathspecs becomes a no-op?

>   5. After several more releases, eventually remove the old-style
>      matching (perhaps leaving --use-pathspecs as a noop).

> To be honest, that may be more careful than we absolutely need to be.

Yeah, there seem to be some room for optimization, but I think the key steps are about right if we wanted to do this migration.

Thanks.