hashsum: remove sha3sum binaries (or determine upstream)

[oech3]

uutil's sha3sum is clearly not a drop-in for https://man.archlinux.org/man/sha3sum.1.en

Please remove.

[sylvestre]

i am not sure i want to do that, please hold

[oech3]

Multicall feat for this should be droppen at least for. cksum -a sha3sum might still useful.

[sylvestre]

maybe but please don't jump on the gun :)

[oech3]

Optionally mask them at make might milder than removing suddenly.

i am not sure i want to do that,

Also someone might hope making drop-in for sha3sums at here.

[collinfunk]

When I added SHA-3 support to GNU cksum, I looked a bit at sha3sum commands and what digest lengths they used by default [1].

The man page you link is a different sha3sum than what is used in Debain and Fedora. The one used there is a Perl script to interact with the Digest::SHA3 module [2].

If you wanted to write a program compatible with sha3sum you would have to choose which one you want to be compatible with.

A nice side effect of consolidating everything into cksum is that we don't need to worry about that.

[1] https://lists.gnu.org/archive/html/coreutils/2025-08/msg00087.html
[2] https://manpages.debian.org/unstable/libdigest-sha3-perl/sha3sum.1p.en.html

[oech3]

How about starting from removing (multicall) b3sum which is not in GNU coreutils?

[oech3]

It seems 5 utils are broken at main(#9108) and 0.3.0(#9106) while cksum -a is fine. How about removing multicall logic for them instead of fixing them ?

[oech3]

@collinfunk Is GNU planning to add blake3b/b3sum via cksum -a? I don't know why uu supports it, but I think it should be removed if GNU does not.