Pádraig Brady <[email protected]> writes:

> BTW it's tempting to use tagged format in the above to simplify things for 
> users,
> and also avoid FAILED messages for mixed checksum input. I.e.:
>
>  Here are the SHA256 and SHA3-256 checksums:
>
>    SHA256 (coreutils-9.9.35-cf973.tar.gz) = 
> tWYsM2przwPj1p9c1T0noDxRKTFixSWmZX5tzPZ6cW8=
>    SHA3-256 (coreutils-9.9.35-cf973.tar.gz) = 
> FgV2UuDSv+U3Ud9fdSA5mWRGQYZ9CAKx3wYCHiTm3Ls=
>    SHA256 (coreutils-9.9.35-cf973.tar.xz) = 
> BkhqCaxeKIT51WtP73fuja5ekZYb4DDhPzh88ux2glo=
>    SHA3-256 (coreutils-9.9.35-cf973.tar.xz) = 
> bD/kFhX/O1uQ9ciHUQygPa8YyotTBrmsV5Do+49yAV8=
>
>  Verify the base64 SHA256 checksums with cksum --check
>  from coreutils-9.2 or OpenBSD's cksum since 2007.
>  Also verify the base64 SHA3-256 checksums with cksum --check from 
> coreutils-9.9.

Yes, Colin, please use base64 encoding here, at least for SHA3-256.
There is no legacy here, right?  Anyone who want to verify SHA3-256's
ought to have a tool that supports base64 encoded checksums.

I'm also +1 on tagged format since it handle empty leading SPC better
than untagged, wich makes it easier to paste the lines from the
announcement e-mail into a file without having lots of editing:

Default, tagged format:

jas@kaka:~$ echo foo > README
jas@kaka:~$ (echo -n "   "; cksum -a sha256 README) > README.SHA256   
jas@kaka:~$ cksum -c README.SHA256 && echo ok
README: OK
ok

Untagged format:

jas@kaka:~$ (echo -n "   "; cksum --untagged -a sha256 README) > README.SHA256
jas@kaka:~$ cksum -c README.SHA256 && echo ok
cksum: README.SHA256: no properly formatted checksum lines found
jas@kaka:~$ 

Surprisingly, using a text editor is not a universal ability, and I've
had this issue come up occasionally.

/Simon

Attachment: signature.asc
Description: PGP signature

Reply via email to