[PATCH 0/3] handling pathspec magic with --follow

From: Jeff King <peff@peff.net>
Date: Jun 1, 2023, 17:37 UTC
Message-ID: <20230601173724.GA4158369@coredump.intra.peff.net>
In-Reply-To: <51012e79-c6ce-41dd-9755-960e065214ef@app.fastmail.com>

On Sat, May 27, 2023 at 09:00:27PM +0200, Jim Pryor wrote:

> Confirmed I have log.follow = true, and also GIT_NOGLOB_PATHSPECS set,
> thus the explicit `(glob)` magic.

OK, all of that makes sense. The same problem hits ":(icase)" and other pathspec magic, too. So talking about :(glob) too much is kind of a red herring, anyway.

> I'd be happy with --follow not working with the wildcard pathspecs,
> but with the crash avoided.

Here's a series that I think should improve things for you. For an explicit --follow, it finds the forbidden magic sooner (which gives us a nice error message instead of the BUG()), and for log.follow, it avoids follow-mode when forbidden pathspecs are used rather than enabling it and then complaining.

The interesting bits are in the third patch. The other two are preparatory.

  [1/3]: pathspec: factor out magic-to-name function
  [2/3]: diff: factor out --follow pathspec check
  [3/3]: diff: detect pathspec magic not supported by --follow

 builtin/log.c  |  2 +-
 diff.c         | 29 +++++++++++++++++++++++++++--
 diff.h         |  7 +++++++
 pathspec.c     | 19 ++++++++++++-------
 pathspec.h     |  8 ++++++++
 t/t4202-log.sh | 15 +++++++++++++++
 6 files changed, 70 insertions(+), 10 deletions(-)