From: Stefan Sperling <stsp@stsp.name>
Subject: Re: fix regress of diff.git
To: Omar Polo <op@omarpolo.com>
Cc: gameoftrees@openbsd.org
Date: Tue, 2 Aug 2022 13:11:49 +0200

On Mon, Aug 01, 2022 at 07:29:53PM +0200, Omar Polo wrote:
> as per subject, this fixes most of the issues with the regress suite in
> diff.git.  test 015 still fails because of a bug in patch(1).  I've sent
> a diff to tech@: with that applied the regress suite should be all fine.
> 
> The diff is not really a pleasure to read.  I've adjusted after manual
> inspection the expected output so that it's equal to the one produced by
> the tool.  Comparing the diffs produced is moot, the same diff can be
> expressed in different ways, but i think it's useful to keep it to avoid
> surprises in the future.
> 
> ok?

Fine with me. ok.

Thanks for looking into this!
I never got around to fixing these tests on OpenBSD, and got regress
passing was good enough for me.

I think we always had issues with the diffs tests between platforms.
The tests were already passing on Linux as far as I remember.
Though perhaps they broke on Linux as well while more changes were
made to the diff.git code? Not sure.