sha3-224sum, sha3-256sum, sha3-384sum, sha3-512sum should not ask --bits

[mykhal]

$ cargo build
…
$ cd target/debug
$ ./coreutils | grep sha3
    sha256sum, sha3-224sum, sha3-256sum, sha3-384sum, sha3-512sum, sha384sum, sha3sum,
$ ln coreutils sha3sum
$ ln coreutils sha3-512sum

$ ./sha3sum /dev/null
sha3sum: --bits required for SHA3
$ ./sha3-512sum /dev/null
sha3-512sum: --bits required for SHA3

I'm not sure if sha3sum should even exist (like there is no sha2sum; but I know it already exists in some implementations/packages, being sha3-224sum).

But sha3-224sum, sha3-256sum, sha3-384sum, sha3-512sum (and probably some other utils based on Keccak family hashes) should not require any more (mandatory) args.

(I'm rather new to Rust, so not providing a fix at the moment.)

[oech3]

They don't have proper reference(upstream). See #8996 (comment) .

[mykhal]

@oech3 you mean it's not clear what length should be used ?

Providing one does not seem to help either:

$ ./sha3sum --bits 512 /dev/null
sha3sum: --bits required for SHA3

$ ./sha3-512sum --bits 512 /dev/null
error: unexpected argument '--bits' found

  tip: to pass '--bits' as a value, use '-- --bits'

Usage: sha3-512sum-usage

For more information, try '--help'.

[oech3]

No. This project is not considering https://codeberg.org/maandree/sha3sum as a upstream.
ALso I have different err from main branch #9108

[mykhal]

When using one (and the one more related to this issue), in Arch linux (uutils-coreutils 0.3.0-2), I get:

$ uu-sha3-512sum 
uu-sha3-512sum: --bits required for SHA3

[oech3]

Manbe, fixed (but still causing another bug) by 8b7286a ?

[mykhal]

(the sha3sum itself mention in the issue description was intended just as a side note; not sure if the wording was optimal; this is about sha3-XXXsum)

[RenjiSann]

That's my bad.
I'm working on removing the hashsum utility entirely, as there is no reason for us to maintain both our implementation of cksum AND hashsum, which initially existed to handle separately the algorithm-specific binaries.

[oech3]

So I think we can close this and #9108 as duplications with #8996 .

[mykhal]

This issue is based on existence of these utils, it would be rather pointless/invalid instead of a duplicate.

(And I have an opinion on existence of these useful utils, but it probably does not belong here.)

[oech3]

I have an opinion on existence of these useful utils

Please use cksum -a instead. They are deprecated and not compat with other tools with same binary names. Please consider leaving something at #8996 if you have different opinion.