From: Omar Polo <op@omarpolo.com>
Subject: Re: gitwrap: don't wait(), just execl()
To: Stefan Sperling <stsp@stsp.name>
Cc: gameoftrees@openbsd.org
Date: Tue, 28 Mar 2023 16:32:16 +0200

On 2023/03/28 13:30:24 +0200, Stefan Sperling <stsp@stsp.name> wrote:
> On Tue, Mar 28, 2023 at 12:48:45PM +0200, Omar Polo wrote:
> > As a semplification, since we're going to exec(), we could directly do
> > so without fork() + wait() in the main process.  So far I've only
> > tested it from the shell (and hit an issue in
> > got_serve_parse_command()) but it's easy to see that git-*-pack or
> > gotsh are still correctly executed.
> > 
> > (we loose the _exit(127) but that was dead code anyway.)
> > 
> > ok?
> 
> Ah, so it was the "proc" pledge promise getting in the way of
> just running exec()?

Nope, if you re-add the proc pledge promise it still works fine, as
expected.  No idea why you were getting a SIGABRT before, maybe there
was typo... I've looked again at the previous diffs on the ML and they
look fine.