[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Emacs design and architecture. How about copy-on-write?

From: Emanuel Berg
Subject: Re: Emacs design and architecture. How about copy-on-write?
Date: Fri, 22 Sep 2023 18:22:41 +0200
User-agent: Gnus/5.13 (Gnus v5.13)

Ihor Radchenko wrote:

>>> Sorry, but I am completely lost. Cannot understand what
>>> you are trying to propose.
>> We can't build a new solution filled with exceptions so it
>> won't break existing programs that were programmed for
>> another solution.
> Eli's point is that the new concurrency framework must allow
> the existing code to run without modifying it to fit
> concurrency. Not necessarily fast (maybe with some
> interlocking), but at least without breaking.
> Otherwise, making actual use of concurrency will require
> rewriting a lot of Elisp, in addition to C-level support,
> which is too much effort - we will have to fight many
> non-obvious problems any time we want to call an "old" Elisp
> function from purposely-written async thread.

Multi-threaded concurrent access to global variables that will
not break programs that were written using global variables
with the then correct assumption no one else would touch them?

But it is still possible, just the meaning of `setq', `let'
etc will have to be changed so that they now are referred to
the lock mechanism and also actually create local variables
transparently which are subsequently used.

underground experts united

reply via email to

[Prev in Thread] Current Thread [Next in Thread]