From: Martin von Zweigbergk <martin.von.zweigbergk@gmail.com>
Date: Sat, 12 Feb 2011 00:55:13 GMT
Subject: Re: [PATCH v2 00/31] refactor rebase
Message-ID: <alpine.DEB.2.00.1102111811300.26684@debian>
In-Reply-To: <7vd3mz33xb.fsf@alter.siamese.dyndns.org>

On Thu, 10 Feb 2011, Junio C Hamano wrote:

> I am not sure if forbidding "-v --continue" adds any value;

If I remember correctly, it was just the simplest possible way to not
have to handle cases like "-i/-p/-m --continue" specially.

> would it be
> too much effort to allow "--continue -v" instead to achieve the same
> degree of consistency between the two?

I think it would be doable. Maybe something like:

 * Check that only the allowable options (-v/-q, -s, -X,
   --[no-]rerere-autoupdate) are passed together with --continue or
   --skip. For simplicity, maybe we can just accept the same set of
   flags with --abort as well, even though only -v/-q would make any
   sense.

 * When reading the saved state, first check for each of the options
   if it is already set. If it is, write it to disk instead, to make
   sure it is remembered across future interruptions. Otherwise, just
   read it as usual.

I think that should be all that's needed. I'll have a look at it when
I get some time.


/Martin