Re: Bug#38708: eq vs eql in byte-compiled code

[Stefan Monnier]

> In other words, making bignums/flonums slower may seem to have no bad
> effects, until suddenly it has.

Agreed for flonums, but not for bignums: the bignum-replacements in
Emacs<27 were pretty inefficient, so I think we can use hash-consed
bignums without incurring a performance cost relative to Emacs-26.

IOW hash-cons'd bignums may be slower than "the best bignums we can
have", but they're still faster than the "the best bignums we used to
have".


        Stefan