Re: [PATCH 2/2] diff: teach diff to read gitattribute diff-algorithm

From: Jeff King <peff@peff.net>
Date: Feb 11, 2023, 02:04 UTC
Message-ID: <Y+b3opa1xgnGM+VN@coredump.intra.peff.net>
In-Reply-To: <230209.86lel7xi8l.gmgdl@evledraar.gmail.com>

On Thu, Feb 09, 2023 at 10:41:33AM +0100, Ævar Arnfjörð Bjarmason wrote:

> > If you run with more than 1 run, are your numbers even repeatable?
> 
> Yes, but tl;dr it's diff.colorMoved=true, sorry, see below.

Wow, that's really slow. I was slightly surprised, because I also use colorMoved. But I set it to "plain", which is way faster:

  $ time git diff --color-moved=default v2.0.0 v2.28.0 >/dev/null
  real	0m18.492s
  user	0m18.411s
  sys	0m0.081s

  $ time git diff --color-moved=plain v2.0.0 v2.28.0 >/dev/null
  real	0m0.942s
  user	0m0.841s
  sys	0m0.101s

I didn't dig into why, but it's possible there's some low-hanging fruit in the zebra/block code.

I also have a mild feeling of deja vu that we may have discussed this before, but a quick search in the archive didn't yield anything. So I'll leave it for somebody to investigate further if they're interested.

-Peff