Re: [PATCH] hash: pacify gcc -Wsuggest-attribute=const

[Paul Eggert]

On 2023-05-19 05:12, Bruno Haible wrote:
> These are global data flow properties, that "gcc -flto" can well exploit to
> produce better code. But a function's declaration must remain valid
> independently of the global data flow in the surrounding program.
>
> > So I guess _GL_ATTRIBUTE_CONST is fine there.
> No, it's not fine. You proposed a patch to gnulib, not to coreutils.

Yes, and besides, it's not helpful for GCC to issue -Wsuggest-attribute 
diagnostics for static functions. If GCC has already figured out that 
the function is pure or const or whatever then that's good enough: GCC 
shouldn't badger the programmer to complicate the program to record 
something that GCC can easily calculate for itself.

This sounds like GCC bug 85734, which has been marked as fixed. 
Evidently it's rearing its ugly head again. We should fix the bug rather 
than pollute the code with compiler pacifications.

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85734