cksum: Bring back blake3 algorithm by RenjiSann · Pull Request #11316 · uutils/coreutils

RenjiSann · 2026-03-13T15:07:18Z

github-actions · 2026-03-13T15:16:25Z

GNU testsuite comparison:

Skip an intermittent issue tests/date/date-locale-hour (fails in this run but passes in the 'main' branch)
Skipping an intermittent issue tests/pr/bounded-memory (passes in this run but fails in the 'main' branch)
Skipping an intermittent issue tests/tty/tty-eof (passes in this run but fails in the 'main' branch)
Note: The gnu test tests/cut/bounded-memory is now being skipped but was previously passing.
Note: The gnu test tests/tail/pipe-f is now being skipped but was previously passing.

src/uucore/src/lib/features/checksum/mod.rs

codspeed-hq · 2026-03-13T16:18:03Z

Merging this PR will degrade performance by 99.88%

❌ 2 regressed benchmarks
✅ 298 untouched benchmarks
⏩ 46 skipped benchmarks¹

⚠️ Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	`BASE`	`HEAD`	Efficiency
❌	Simulation	`cksum_blake3`	116.3 µs	93,152.5 µs	-99.88%
❌	Memory	`cksum_blake3`	33.9 KB	43.5 KB	-21.95%

_{Comparing RenjiSann:cksum-blake3 (f9c8370) with main (d830e84)}

46 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

oech3 · 2026-03-13T16:18:57Z

CodSpeed is invalid since it is based on a bunch of clap error.

github-actions · 2026-03-13T16:26:06Z

GNU testsuite comparison:

Skipping an intermittent issue tests/pr/bounded-memory (passes in this run but fails in the 'main' branch)
Skipping an intermittent issue tests/tty/tty-eof (passes in this run but fails in the 'main' branch)

src/uucore/src/lib/features/sum.rs

github-actions · 2026-03-16T13:46:32Z

GNU testsuite comparison:

Skip an intermittent issue tests/date/resolution (fails in this run but passes in the 'main' branch)
Skipping an intermittent issue tests/date/date-locale-hour (passes in this run but fails in the 'main' branch)
Skipping an intermittent issue tests/tail/inotify-dir-recreate (passes in this run but fails in the 'main' branch)
Note: The gnu test tests/tail/pipe-f is now being skipped but was previously passing.
Congrats! The gnu test tests/printf/printf-surprise is now passing!
Note: The gnu test tests/seq/seq-epipe was skipped on 'main' but is now failing.

github-actions · 2026-03-16T19:21:23Z

GNU testsuite comparison:

GNU test failed: tests/rm/isatty. tests/rm/isatty is passing on 'main'. Maybe you have to rebase?
Skipping an intermittent issue tests/date/date-locale-hour (passes in this run but fails in the 'main' branch)
Congrats! The gnu test tests/expand/bounded-memory is now passing!

github-actions · 2026-03-16T19:29:05Z

GNU testsuite comparison:

GNU test failed: tests/rm/isatty. tests/rm/isatty is passing on 'main'. Maybe you have to rebase?
Skipping an intermittent issue tests/date/date-locale-hour (passes in this run but fails in the 'main' branch)
Congrats! The gnu test tests/expand/bounded-memory is now passing!

AldanTanneo

If the length of blake3 is stored in bytes, it should validate that the input bit size is a multiple of 8, like in the blake2b implementation.

oech3 · 2026-03-20T15:00:39Z

Any length support is coming from https://github.com/BLAKE3-team/BLAKE3

AldanTanneo · 2026-03-20T15:09:11Z

Any length support is coming from https://github.com/BLAKE3-team/BLAKE3

Your point?

oech3 · 2026-03-20T15:14:28Z

The existing b3sum support -l 7. So we should not validate to 8N bytes.

AldanTanneo · 2026-03-20T15:16:38Z

The existing b3sum support -l 7. So we should not validate to 8N bytes.

Yes but the existing b3sum takes its input in bytes, not in bits, which goes against coreutils convention. The coreutils version should take the input in bits, and then validate that it's a multiple of 8 (like blake2b).

oech3 · 2026-03-20T15:22:19Z

We should not divergent interfaces. So we should be same with the b3sum.

AldanTanneo · 2026-03-20T15:27:39Z

We should not divergent interfaces. So we should be same with the b3sum.

So matching the interface of a non GNU utility is more important that following the standard conventions for coreutils (always taking the length in bits)?

oech3 · 2026-03-20T15:29:48Z

Yes, Some people might have shebang compat wrapper (even it is very bad idea).

oech3 · 2026-03-20T15:32:22Z

To be honestly, I think we should just remove b3sum extenstons. We should not be interested in such wars.

RenjiSann · 2026-03-20T16:09:44Z

To be honestly, I think we should just remove b3sum extenstons. We should not be interested in such wars.

Ok so to explain my reasoning here:

IIUC, GNU is willing to support blake3 at some point, so it makes sense that we support it, either now or once GNU does. @pixelb you may want to add your 2 cents here.
The GNU convention is to interpret lengths as bits, not bytes, so if we want to be consistent with our own interface, bits is the way to go, even if it's incompatible with the existing b3sum.
If we ever want to relax the "multiple of 8" rule to support half-bytes digests like we do for SHAKE*, we will have backwards compatibility for free, whereas if we switch from bytes to bits, we will definitely break stuff WE were responsible of.

For all these reasons, I think we should interpret the length in bits for blake3 rather than in bytes.

oech3 · 2026-03-20T16:21:18Z

If we ever want to relax the "multiple of 8" rule to support half-bytes digests like we do for SHAKE*, we will have backwards compatibility for free, whereas if we switch from bytes to bits, we will definitely break stuff WE were responsible of.

I missed the case...

RenjiSann · 2026-03-20T18:41:56Z

I missed the case...

Sorry I don't understand what you mean. What do you mean ?

oech3 · 2026-03-20T18:51:56Z

sorry. Please forget the comment which is not important.

AldanTanneo · 2026-03-20T23:00:10Z

src/uucore/src/lib/features/checksum/mod.rs

            (ak::Sha1, _) => Ok(Self::Sha1),
-            (ak::Blake3, _) => Ok(Self::Blake3),

+            (ak::Blake3, l) => Ok(Self::Blake3(l)),


the length is not validated here, but it is for the blake2b algo

I've addressed that in #11437

RenjiSann · 2026-03-21T12:38:42Z

Benchmark test fails because I re-enabled one.
GNUTests seems to fail everywhere, so I assume that's not the changes in this PR

This comment was marked as resolved.

Sign in to view

xtqqczze reviewed Mar 13, 2026

View reviewed changes

src/uucore/src/lib/features/checksum/mod.rs Outdated Show resolved Hide resolved

xtqqczze reviewed Mar 13, 2026

View reviewed changes

src/uucore/src/lib/features/checksum/mod.rs Outdated Show resolved Hide resolved

RenjiSann force-pushed the cksum-blake3 branch 4 times, most recently from b2d43e2 to 3387ee5 Compare March 13, 2026 16:16

xtqqczze reviewed Mar 13, 2026

View reviewed changes

src/uucore/src/lib/features/sum.rs Outdated Show resolved Hide resolved

RenjiSann force-pushed the cksum-blake3 branch 2 times, most recently from 72790dd to 45b28c6 Compare March 16, 2026 13:35

RenjiSann force-pushed the cksum-blake3 branch from 45b28c6 to 509d857 Compare March 16, 2026 19:11

RenjiSann marked this pull request as ready for review March 17, 2026 15:49

RenjiSann force-pushed the cksum-blake3 branch from 509d857 to c673bba Compare March 17, 2026 15:49

AldanTanneo reviewed Mar 20, 2026

View reviewed changes

RenjiSann force-pushed the cksum-blake3 branch from c673bba to bd7dddb Compare March 20, 2026 15:28

RenjiSann force-pushed the cksum-blake3 branch from bd7dddb to 105237e Compare March 20, 2026 16:19

RenjiSann force-pushed the cksum-blake3 branch from 105237e to bf69679 Compare March 20, 2026 18:39

RenjiSann force-pushed the cksum-blake3 branch from bf69679 to 25915da Compare March 20, 2026 18:44

AldanTanneo reviewed Mar 20, 2026

View reviewed changes

RenjiSann added 3 commits March 21, 2026 03:58

cksum: Add support for blake3

f21913d

test(cksum): Test for cksum -a blake3

b7eb2b9

bench(cksum): Re-enable blake3 benchmarks

f9c8370

RenjiSann force-pushed the cksum-blake3 branch from 25915da to f9c8370 Compare March 21, 2026 02:58

AldanTanneo approved these changes Mar 21, 2026

View reviewed changes

RenjiSann merged commit f0805ff into uutils:main Mar 21, 2026
161 of 163 checks passed

Uh oh!

Conversation

RenjiSann commented Mar 13, 2026

Uh oh!

github-actions bot commented Mar 13, 2026

Uh oh!

This comment was marked as resolved.

Uh oh!

Uh oh!

codspeed-hq bot commented Mar 13, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Merging this PR will degrade performance by 99.88%

Performance Changes

Footnotes

Uh oh!

oech3 commented Mar 13, 2026

Uh oh!

github-actions bot commented Mar 13, 2026

Uh oh!

Uh oh!

github-actions bot commented Mar 16, 2026

Uh oh!

github-actions bot commented Mar 16, 2026

Uh oh!

github-actions bot commented Mar 16, 2026

Uh oh!

AldanTanneo left a comment

Choose a reason for hiding this comment

Uh oh!

oech3 commented Mar 20, 2026

Uh oh!

AldanTanneo commented Mar 20, 2026

Uh oh!

oech3 commented Mar 20, 2026

Uh oh!

AldanTanneo commented Mar 20, 2026

Uh oh!

oech3 commented Mar 20, 2026

Uh oh!

AldanTanneo commented Mar 20, 2026

Uh oh!

oech3 commented Mar 20, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

oech3 commented Mar 20, 2026

Uh oh!

RenjiSann commented Mar 20, 2026

Uh oh!

oech3 commented Mar 20, 2026

Uh oh!

RenjiSann commented Mar 20, 2026

Uh oh!

oech3 commented Mar 20, 2026

Uh oh!

AldanTanneo Mar 20, 2026

Choose a reason for hiding this comment

Uh oh!

RenjiSann Mar 21, 2026

Choose a reason for hiding this comment

Uh oh!

RenjiSann commented Mar 21, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

codspeed-hq bot commented Mar 13, 2026 •

edited

Loading

oech3 commented Mar 20, 2026 •

edited

Loading