Re: Minor bug report

From: Junio C Hamano <gitster@pobox.com>
Date: Jun 3, 2015, 06:48 UTC
Message-ID: <CAPc5daU9VdndgkMFNmtiwhVdwyGQgDbFX3rrrsEEEUVewgs-dw@mail.gmail.com>
In-Reply-To: <20150603062005.GA20580@peff.net>

On Tue, Jun 2, 2015 at 11:20 PM, Jeff King <peff@peff.net> wrote:

Show 18 quoted lines

>
> Here are some prior discussions:
>
>   http://thread.gmane.org/gmane.comp.version-control.git/75692
>
>   http://thread.gmane.org/gmane.comp.version-control.git/200504
>
> I just skimmed through them, but I expect the most desirable solution
> would involve:
>
>   1. We still die(), but just improve the error message (so we don't
>      have any regressions for people expecting "git log" to fail).
>
>   2. We use the message only when pointing to an unborn branch, and not
>      on other errors. The simplest way to do this is probably to make an
>      extra call to resolve_ref() in the error code-path.
>
> -Peff

I am kind of surprised after reading these two threads that my take on this issue has changed over time, as my knee-jerk reaction before reading them was the opposite, something along the lines of "This is only immediately after 'git init'; why bother?". Or depending on my mood, that "How stupid do you have to be..." sounds exactly like a message I may advocate for us to send. Perhaps I grew more bitter with age.

Jokes aside, I wonder why we did not pursue that "default to HEAD only when HEAD points at a branch that exists" approach, with the definition of "exists" being "either a file exists under the $GIT_DIR/refs/heads directory with any (including corrupt) contents, or an entry in the packed refs file exists with any (including corrupt) contents". With or without the error exit and warning, that sounds like a very sensible way to set the default, at least to me.