Re: [RFC/PATCH] Additional fast forward strategies.

From: Junio C Hamano <gitster@pobox.com>
Date: Mar 11, 2008, 06:34 UTC
Message-ID: <7v8x0pepl8.fsf@gitster.siamese.dyndns.org>
In-Reply-To: <402c10cd0803102004x7892f01cledfef6f277fba26a@mail.gmail.com>

"Sverre Hvammen Johansen" <hvammen@gmail.com> writes:

>>From 779142c6b22f032c877f550733e1bccc99aca761 Mon Sep 17 00:00:00 2001
> From: Sverre Hvammen Johansen <hvammen@gmail.com>
> Date: Sun, 9 Mar 2008 21:16:03 -0800
> Subject: [PATCH] Additional fast forward strategies.

Ditto..

> New fast forward strategies, common, fork, path, and same
> is introduced.  These new fast forward strategies allows
> additional work flows.

Is it sufficient to say "allows additional work flows" here? I only see what they do, and do not see why these are needed/useful anywhere in the proposed commit log message nor documentation.

You need to illustrate the motivation after the above three-line paragraph and before the description of what each of them does below. "Suppose your history looks like this, and you and your group's workflow is like this. Then you would want to only allow these but want to forbid those for such and such reasons." Then list what the "FF options" do. After that, you can conclude "By using these FF options this way, you can achieve that control like this..."

Then people might be convinced that "these allow additional work flows". What you wrote is insufficient, I am afraid. At least to me.

> +	if test $fast_forward = same
> +	then
> +		for b in "$@"

'in "$@"' is always redundant after "for X".

More seriously, why is the user forced to give more than one "same" thing to begin with? The user can give a single thing and be done with it.