Re: [PATCH v2 00/16] Speeding up DEFVAR_PER_BUFFER

[Eli Zaretskii]

> From: Spencer Baugh <sbaugh@catern.com>
> Cc: emacs-devel@gnu.org, arnold@tdrhq.com, monnier@iro.umontreal.ca,
>  dgutov@yandex.ru, rms@gnu.org
> Date: Mon, 30 Nov 2020 15:11:31 -0500
> 
> > But OTOH, referencing buffer-local values through BVAR is now slower,
> > about twice slower, I guess?  For example, the display code has 90
> > uses of BVAR, and they will now be slower.  I wonder what that means
> > for us: we are making let-binding of local variables much faster, but
> > "punish" the code that just accesses these variables in our inner
> > loops.
> 
> Yes, it's probably a trade-off.

In fact, the current code is slow only when the buffer that's current
at the moment of the let-binding doesn't have a buffer-local value for
the variable.  Because if the current buffer does have a local value,
the let-binding simply changes that single local value.

So in effect we make every access to built-in buffer-local variables
pay for the single use case when the let-binding is done in a buffer
that has no local value.  I'm not sure how to assess this kind of
trade-off.

> My guess is that the extra overhead of BVAR will not be significant,
> but benchmarking is the only way to know for sure.

We should benchmark, but coming up with suitable use cases to make
these measurements is not easy.