Re: On elisp running native

[Andrea Corallo <akrl@sdf.org>]

Eli Zaretskii <eliz@gnu.org> writes:


> That said, it's just a name.  If you feel strongly about it, I won't
> object.

Thanks you for your understanding.

>>   - Is force push allowed in these feature branches?
>
> For such a large feature I'd prefer not to do that, assuming that some
> people will want to track that branch and perhaps contribute changes
> and patches.  Why would force-push be necessary?

Just a curiosity.  It can be used to keep on rebasing the branch fast
forward tho (thing I did till now) but is not good for cooperative
development I agree.

>
> Under feature/, please.
>

Okay, so the branch landed now on emacs.git as feature/native-comp.

I've freshly rebased it on the latest master (I just had to revert
186152ba4 but this is master related).

To test it configure it with --with-nativecomp (this is disabled by
default).  I did some work and the normal build should be compilable and
functional again.

The build system produce elns as side effect of the elc compilation
during the boot for now.  This is because there's no list of
lexical/dynamic scoped files.  Is not the nicest but it works for now.

It bootstrap cleanly on two my machines.  Obviously has to be considered
experimental and I've left comp-speed 0 as default setting.

Bootstrapping at speed 2 should work but has to be considered *meta*
experimental.

The question I wanted to ask is: What's next?

Other than rounding edges should be focus also on more performance
and/or dynamic scope support?

Thanks

  Andrea

-- 
akrl@sdf.org