byte-opt.el vs declare

byte-opt.el vs declare

[Sam Steingold]

I thought that byte-opt was intended to set `pure'. `side-effect-free'
&c properties for functions that are defined in C and functions defined
in Lisp should get these from `declare'.

However, I see that, e.g., `xor' uses `declare' but `string-prefix-p' is
listed in byte-opt and does not have a declaration.

What is the official policy?

Thank you.

-- 
Sam Steingold (http://sds.podval.org/) on darwin Ns 10.3.2113
http://childpsy.net http://calmchildstories.com http://steingoldpsychology.com
https://iris.org.il https://www.memritv.org https://ffii.org
Failure is not an option. It comes bundled with your Microsoft product.

Re: byte-opt.el vs declare

[Mattias Engdegård]

26 juli 2022 kl. 21.57 skrev Sam Steingold <sds@gnu.org>:

> I thought that byte-opt was intended to set `pure'. `side-effect-free'
> &c properties for functions that are defined in C and functions defined
> in Lisp should get these from `declare'.

Not official by any means, but I'd suggest `declare` be used for new Lisp functions and set the properties for C functions in byte-opt for now.

There is a considerable legacy, mind, and it's probably not worth changing all of it right away. If there turns out to be a more substantial benefit from doing so then someone working on the byte-compiler will deal with it.