Re: [PATCH v2] ls-files: add pathspec matching for submodules

From: Junio C Hamano <gitster@pobox.com>
Date: Sep 21, 2016, 17:12 UTC
Message-ID: <xmqqmvj19nyp.fsf@gitster.mtv.corp.google.com>
In-Reply-To: <CAKoko1qS0+DgnMeNnjayEK3sWnvpuiS4oRDBSR=6s8i4okQ_Hw@mail.gmail.com>

Brandon Williams <bmwill@google.com> writes:

> On a similar but slightly different note.  In general do we want
> the pathspec '??b' to match against the sib/ directory and
> subsequently have ls-files print all entries inside of the sib/
> directory?  (this is in the non-recursive case)

I'd need to find time to dig a bit of history before I can give a firm opinion on this, but here is a knee-jerk version of my reaction.

 * A pathspec element that matches literally to a directory causes
   itself and everything underneath the directory match that
   element, e.g. "sib" would be considered a match.

 * Otherwise, a pathspec that matches with the whole path as a
   pattern matches the path, e.g. "??b" would match "sib" itself,
   but not "sib/file".  Note that "??b*" would match "sib" and
   "sib/file" because the pattern match is without FNM_PATNAME
   unless ':(glob)' magic is in effect.

Historically, some commands treated a pathspec as purely a prefix match (i.e. the former) and did not use _any_ pattern matching, while other commands did both of the above two (e.g. compare ls-tree and ls-files). I thought we were slowly moving towards unifying them, but apparently 'git log -- "D?cumentation"' does not show anything close to what 'git log -- Documentation' gives us even in today's Git.

Probably we want to change it at some point so that a pattern that matches one leading directory would cause everything underneath to match, e.g. "??b" would include "sib/file" just because "sib" would.