Emacs Devel: Design and Architecture

2024-01-09 00:19:05

This is a collection of links pulled together from some interesting discussions on Devel related to making Emacs more multi-threaded.

With various apologies to each of the authors I've parapharsing and who's words I abstract from their context, I can only say: parsing devel is hard.

Eli

main design features of Emacs which we found restrictive (Eli) - https://lists.gnu.org/archive/html/emacs-devel/2023-09/msg00803.html
[to move a line vertically the display engine must calc metrics of each character traversed] (Eli) - https://lists.gnu.org/archive/html/emacs-devel/2023-09/msg01056.html-
the few calls into Lisp we do in redisplay [is a seperate problem] from being able to run several async Lisp threads concurrently (Eli) - https://lists.gnu.org/archive/html/emacs-devel/2023-09/msg01121.html
[concerning adding global locks, more than slowness] I don't even know how to identify all of the changes this will incur (Eli) - https://lists.gnu.org/archive/html/emacs-devel/2023-09/msg01286.html

RMS

[it would be a mistake to add Javascript evaluation to Emacs] (Richard) - https://lists.gnu.org/archive/html/emacs-devel/2023-09/msg00963.html
I would not object to using keyword arguments for functions [that are otherwise too] complex and cumbersome to use, and not used often. (Richard) - https://lists.gnu.org/archive/html/emacs-devel/2023-09/msg00814.html

And Many Others

What about automatically debouncing slow :eval constructs in the mode line? (Ihor) - https://lists.gnu.org/archive/html/emacs-devel/2023-09/msg01190.html ** Assuming that's feasable [] why not give users more [] control? (Drew Adams) - https://lists.gnu.org/archive/html/emacs-devel/2023-09/msg01257.html
a 60 FPS (or higher) redraw rate would [is] desired (Adam Porter) - https://lists.gnu.org/archive/html/emacs-devel/2023-09/msg01142.html
I there that there is at least one way to addres long lines using asyncronous redisplay (Ihor) - https://lists.gnu.org/archive/html/emacs-devel/2023-09/msg01160.html
[We can't warn from redisply because that would redisplay and we don't support recursive redisplay] (Eli) - https://lists.gnu.org/archive/html/emacs-devel/2023-09/msg01245.html
for a parallel redisplay [] a persistent piece tree [would be good] (Gerd Möllmann) - https://lists.gnu.org/archive/html/emacs-devel/2023-09/msg00969.html
[Idiots for Mor CL] (various authors) - https://lists.gnu.org/archive/html/emacs-devel/2023-09/msg00678.html
it would, in principle, be possible to write an SBCL core that impelements all of the Emacs primitives [shows sample shim code] (Eric S. Raymond) - https://lists.gnu.org/archive/html/emacs-devel/2023-09/msg00769.html