Re: The Emacs master is much slower than the emacs-27 branch.

[Eli Zaretskii]

> Date: Sat, 05 Dec 2020 15:17:34 +0000
> From: Gregory Heytings <ghe@sdf.org>
> cc: rudalics@gmx.at, stephen.berman@gmx.net, emacs-devel@gnu.org, acm@muc.de
> 
> >> I'd guess that Eli will merge b9b32074cecb0e0d20cf527eadacbb0b7f5cced9 
> >> into emacs-27 very soon after the Emacs 27.2 release
> >
> > No, this is unlikely to happen.  The emacs-27 branch should only get 
> > safe bugfixes at this time.
> 
> I did not know that an Emacs 27.3 release was not planned.

I didn't say it wasn't.  (We won't consider whether 27.3 should be
released until some time after releasing 27.2, based on the bugs and
regressions that will still be reported against the emacs-27 branch.)
However, whether there will be Emacs 27.3 or not, the emacs-27 branch,
like any other release branch, should only get changes that fix bugs
and aren't too risky (i.e. cannot cause additional bugs/issues).

> In that case, the emacs-27 branch will forever remain slower with
> -Og and -O0 (at least on GNU/Linux 64-bit platforms with GCC).

Yes.  And that is not so bad, given that your timings indicate Emacs
27.1 compiled with optimizations is better than Emacs 25 and only
slightly slower than Emacs 26.