Re: [PATCH] For Real - Fixed pluralization in diff reports

From: Jeff King <peff@peff.net>
Date: Aug 1, 2011, 18:38 UTC
Message-ID: <20110801183812.GA16438@sigill.intra.peff.net>
In-Reply-To: <CAGdFq_j0Hr4LMaFC3irS8rhv2N0ePz_t8xpNTDCA-arC-E2UDA@mail.gmail.com>

On Mon, Aug 01, 2011 at 08:32:51PM +0200, Sverre Rabbelier wrote:

Show 10 quoted lines

> On Mon, Aug 1, 2011 at 20:27, Jon Forrest <nobozo@gmail.com> wrote:
> > I entirely agree. My point is only that the various tests
> > that expect the current behavior will have to be changed
> > whether the implementation of correct plurals uses my
> > inferior method or the way more correct i18n method.
> 
> Wouldn't it be a nice hack if we just solved problem through i18n
> then? Have all the plumbing see the current wording, but through i18n
> change it to something grammatically correct for the porcelain.
> Probably not possible, but a nice daydream :).

I thought there was still some question of whether this text was something that should be script-parseable. If it is, then it shouldn't be i18n'd at all, nor should we lightly change the format with pluralization magic. And if it isn't, then we should definitely go the full i18n route. So in either case, the original patch isn't appropriate.

I don't have a strong opinion myself. I tend to lean towards i18n-ing it, because any scripts should be using --numstat to parse, anyway. OTOH, as Junio pointed out, we are matching the output of much older tools, so pre-git scripts might be written to read the --stat format. I've never seen such a script, and I have no idea how many there really are.

-Peff