Re: [PATCH] diff,difftool: Don't use the {0,2} notation in usage strings

[Jeff King <peff@peff.net>]

On Thu, Nov 04, 2010 at 09:26:14PM +0100, Štěpán Němec wrote:

> > [1] If we are on a quest to remove <rev>{0,2}, this is one other spot to
> >     do it.
> 
> That's already included in the patch.

Ah, sorry, I missed that hunk.

> I think this thread already has nothing to do with the purpose of the
> patch, though. I find your suggestion that <commit>{0,2} better conveys
> some information about git-diff's operation modes than [<commit>
> [<commit>]] does (which seems to have been your starting point) rather
> far-fetched.

I find your version slightly harder to read than the original, which is
why I complained. However, I find both way worse than what Jonathan and
I ended up with.

> So AIUI the patch can still be applied and we/you can then work on
> improving the usage string in other ways -- providing that we agree that
> the {M,N} notation should go, of course, which we apparently still
> don't?

My main argument against that would be that if we are planning on
changing it to something totally different right now anyway, your patch
will just end up making textual conflicts for Junio to resolve. :)

> In any case this seems to be less and less my itch to scratch, so I'm
> not sure I'm going to be of much help in this discussion. The thing I
> need to know for the CodingGuidelines patch is whether we're going to
> preserve and encourage the {M,N} usage or not (so please let's arrive to
> some conclusion on that matter -- I waited for several days before
> sending the updated version and nobody chimed in, so I assumed nobody
> other than Mark and Sverre really cared). Although, that's actually
> somewhat orthogonal, too -- we can always re-add the {M,N} subsection
> later.

Yeah, sorry to come in late to the discussion. I missed the other thread
entirely. I think getting rid of {M,N} is fine. In general, your
proposed replacement is better. It's just that in this case it seemed to
be obfuscating a subtle point that the original syntax (as ugly and
incomprehensible as it was) called out.

So yes, go ahead with your coding guidelines patch. I'll summarize what
Jonathan and I discussed with a new patch.

-Peff