> Am 09.11.2021 um 19:12 schrieb Aaron Jensen <aaronjensen@gmail.com>:
> 
> On Tue, Nov 9, 2021 at 12:05 PM Eli Zaretskii <eliz@gnu.org> wrote:
>> 
>>> From: Aaron Jensen <aaronjensen@gmail.com>
>>> Date: Tue, 9 Nov 2021 10:57:11 -0500
>>> Cc: Emacs developers <emacs-devel@gnu.org>, Eli Zaretskii <eliz@gnu.org>, Alan Third <alan@idiocy.org>,
>>>      YAMAMOTO Mitsuharu <mituharu@math.s.chiba-u.ac.jp>
>>> On Tue, Nov 9, 2021 at 9:46 AM Philipp <p.stephani2@gmail.com> wrote:
>>>> 
>>>> I guess either way is fine.  I'd slightly lean towards enabling it whenever it's available (i.e. remove the condition on DARWIN_OS):
>>> 
>>> I'm good with that, I only added it because I believe in an earlier
>>> message in the thread that was called for.
>> 
>> Didn't you want to install this on the release branch?  I thought you
>> did, and that's why I asked for a patch that only affected macOS: on
>> the release branch we will not install such a significant change that
>> affects platforms which don't absolutely have to have it.
> 
> Not me, but I think that's a good idea. The Darwin only patch should
> install on emacs-28, so maybe we could install that there and the
> all-platform patch on master?

For now I did the second part.  There's a small slowdown on GNU/Linux (because posix_spawn performs a few additional safety checks), and a large speedup on macOS, see the attached plot.