[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: Tue, 19 Sep 2023 21:21:11 +0200
User-agent: Gnus/5.13 (Gnus v5.13)

Ihor Radchenko wrote:

> Because implementation details are tricky - a lot of Elisp
> internal machinery is relying upon modifying global symbol
> objects, having certain global C variables assigned to the
> "right" value, and buffer objects having the right
> buffer-local values. This particular issue has been
> discussed in details in
> https://yhetil.org/emacs-devel/871qhnr4ty.fsf@localhost/

Yeah, but instead of adopting the lock mechanism to take into
account a possibly huge amount of such cases the lock
mechanism should be solid and work the same way for everyone
and everything.

If that breaks code that relies on the previous solution, that
will have to be fixed.

Because if the lock mechanism has to take into account tons of
weird cases and situations written for another solution, it
won't be good. And new code - what solution should it be
written for? The old or the new solution? Or the new
solution's exceptions not to break legacy code?

Oh, no. 2 wrongs don't make 1 right. Solid foundation
bottom-up approach, things incompatible with the sound
solution? No exceptions added to the sound solution, instead
fix them one at a time.

No shortcuts to the top!

underground experts united

reply via email to

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