Re: sqlite3

[Qiantan Hong]

> Indeed.  BTW, there's another approach which, instead of looking at the
> higher-level API looks at the lower level steps (like "change this slot
> to that value" where a value can be a reference), so adding an element
> to a doubly linked list would record a set of 4 slot changes.
I considered this approach very early on and have probably mentioned it
in the thread… let me see… 
https://lists.gnu.org/archive/html/emacs-devel/2021-12/msg00513.html
The problem is that how we let user inform us about those changes.
It either require significant change to the VM, or have to ask user
to use variants of all mutation operation explicitly.


Best,
Qiantan