Re: Should native compilation be enabled by default?

[Eli Zaretskii]

> From: Richard Stallman <rms@gnu.org>
> Cc: eliz@gnu.org, emacs-devel@gnu.org
> Date: Thu, 07 Mar 2024 21:29:51 -0500
> 
>   > Anyway as a data point: AFAIK major distros decided since a while to
>   > ship with native compilation as (I assume) they deemed it beneficial for
>   > their users.
> 
> I ssppose so.  But the fact that they think so is not a convincing
> argument for us.
> 
> If we poll the users, we would have a basis to conclude that those
> distros' choice is wise, or that it was hasty and unwise.  It is
> better to make sure of this than to follow the herd.

We in effect polled the users already: we have the experience of 2
major Emacs releases under our belt that provided native compilation
as an opt-in feature.  Based on that experience, we know that most
people do use native compilation, and many even compile all the Lisp
files ahead of time (instead of relying on JIT compilation when a Lisp
file is first loaded into Emacs).  And the fact that distros decided
to ship Emacs with native compilation is another indication that users
like it, and are not bothered with the few downsides, especially if
they don't have to build the *.eln files themselves.

We could decide that this evidence is not enough for us if we had some
basis for thinking there are problems lurking under the surface that
we should consider.  But we have no such evidence, so I don't see why
we should try digging deeper for signs of problems which are at this
point purely imaginary, at least from where I stand (and that includes
my own personal experience).