Re: New GC concept

[Daniel Colascione]

On 6/21/21 6:00 AM, Fejfighter wrote:
> Not wanting to see this drop, I found a little time to bring the 
> changes up to emacs master with a few notes:
>
> 1) now sits on top of a4fb5811f (Do not attempt to write .elc....)


Awesome. Thanks!


> 2) The C will compile with a couple of warnings as the native comp 
> cases are currently not handled.
> 3) There's dead code, commented code and other atrocities on top of 
> the current wip

Yep. There were also plenty of atrocities in the original. :-)

> 3) It segfaults during the pdumper step in the build, there is not 
> immediately obvious reason for this, but I suspect a move or free 
> occurs and it's not tracked
> 4) I have ignored the comments for large_vector and large_vector_meta 
> for now, so the meta is kept with the vector struct.

Yeah. That's probably a minor optimization, but we should get around to 
completing it. The next big chunk of work is actually implementing 
concurrent marking.

> code is up here: 
> https://github.com/fejfighter/emacs/tree/feature/newgc 
> <https://github.com/fejfighter/emacs/tree/feature/newgc>
> I'm hoping now that it's a little closer to master and compiling 
> someone might have a little more luck with the segfault issue I have 
> been facing but I will keep try in the mean time,

FWIW, I find rr [1] to be exceptionally useful for diagnosing segfaults 
like this.


[1] https://rr-project.org/