Re: [PATCH] init-db: support --import to add all files and commit right after init

From: Jeff King <peff@peff.net>
Date: Mar 25, 2009, 04:49 UTC
Message-ID: <20090325044909.GA21311@coredump.intra.peff.net>
In-Reply-To: <fcaeb9bf0903242132k71e78219xfda5553854df4a2c@mail.gmail.com>

On Wed, Mar 25, 2009 at 03:32:17PM +1100, Nguyen Thai Ngoc Duy wrote:

Show 5 quoted lines

> I tend to do "commit -q -m i", not a big deal for two more keystrokes.
> My initial version allowed "git init /path/to/workdir" too so it might
> be ambiguous for optional message. So all vote for "Initial commit"?
> Dynamic message seems overkill because most of the cases, this is a
> throw away repository.

Initial commit is fine with me. I plan to use this mostly for throw-away repositories. I was thinking it might be useful for this:

  tar xzf package-0.1.tar.gz
  cd package-0.1
  git init --import

so that later you remember it came from package-0.1 (when you are looking at package 0.2). OTOH, if we take the information from the directory, then the fact that it came from 0.1 is by definition already in the directory name. So it is pretty pointless unless you later rename the directory to just "package".

Show 5 quoted lines

> > I seem to recall that we were phasing out "--long-option <arg>" at some
> > point, and that all long-options should use "--long-option=". But maybe
> > I am mis-remembering.
> 
> I did not know too. Any reference?

I was, in fact, mis-remembering. The problem is not about long versus short options, but rather about options which have an optional value. See "Separating argument from the option" in "git help cli".

However, what we are discussing above _is_ making it an optional value, in which case "--import <arg>" would have no value for "--import".

-Peff