fix: make SKIP_UTILS correctly filter hashsum utilities

[ChrisX101010]

Fixes #8654

Description

SKIP_UTILS does not skip part of hashsum

Problem

The make install SKIP_UTILS="sha3-224sum sha3-256sum ..." command was not properly skipping hashsum utilities during installation, which was a barrier to packaging uutils as a drop-in replacement for coreutils.

Root Cause

The install logic only filtered from the INSTALLEES list, but hashsum utilities are provided by the multicall binary and needed to be filtered from the full utility list obtained via coreutils --list.

Solution

Replaced the old foreach loop with new logic that:

1. Gets all available utilities from coreutils --list
2. Properly filters against SKIP_UTILS using grep -qw
3. Provides informative messages during installation
4. Maintains backward compatibility

Testing

Verified the fix works correctly with:

• ✅ sha3-224sum and sha3-256sum are correctly skipped
• ✅ Other hashsum utilities (sha3-384sum, sha3-512sum, sha3sum) are correctly installed
• ✅ Regular utilities continue to work normally
• ✅ Multiple skip patterns work correctly
• ✅ Edge cases handled properly

Changes Made

• Modified GNUmakefile install target for multicall binaries
• Removed old foreach loop logic
• Added new filtering logic that uses coreutils --list
• Added informative installation messages

Closes #8654

[oech3]

I think filterling symlink is not a solution if binary size is kept.

[ChrisX101010]

I think filterling symlink is not a solution if binary size is kept.

What do you propose as a different solution and I will make changes? Thank you for the feedback. Appreciate it.

[oech3]

Maybe, this is more complex issue and should be solved at same time with solving #4414

[ChrisX101010]

Maybe, this is more complex issue and should be solved at same time with solving #4414

Thank you, brother. I will research deeply into this issue then and the one you referenced me throughout the week. 🙂

[github-actions]

GNU testsuite comparison:

Skip an intermittent issue tests/misc/tee (fails in this run but passes in the 'main' branch)
Skipping an intermittent issue tests/timeout/timeout (passes in this run but fails in the 'main' branch)

[oech3]

I don't know how many people needs small indivisual (non-multicall) binaries. But I think hasusums are treated as same as other commands.

[ChrisX101010]

I don't know how many people needs small indivisual (non-multicall) binaries. But I think hasusums are treated as same as other commands.

I saw someone mentioned an issue to remove multicall as a whole, so should I focus on this issue first then move on to this one? 🙂

[oech3]

Maybe, yes. It should also fix some bugs automatically.
@sylvestre Do you think that multicall hashsum binary still needed?