cksum/hashsum: Merge common logic under a common crate #10142

RenjiSann · 2026-01-09T12:46:51Z

This MR introduces a new crate uu_checksum_common, which declare

It lies on top of #10141

Architecture explanation

Starting from now, each binary has its own crate in src/uu : cksum, but also standalone tools like md5sum, b2sum, etc...

All of them depend on uu_checksum_common, which lives under src/uu/checksum_common, which itself depends on uucore::checksum.

As for who does what:

Top-level binaries contain the program entrypoint. They process arguments that are not common to all binaries, for example
- cksum handles --untagged, which doesn't exist for standalones
- b2sum handles --length where md5sum doesn't allow it
Then, the top-level binaries call a common main function : uu_checksum_common::checksum_main
checksum_main handles the common argument processing (verbose flags, --zero, etc), and calls uucore::checksum::validation or uucore::checksum::compute accordingly.

Note that all common cli error messages live under uu_checksum_common, so a bit of piping was needed in the locale setup.

Reason behind this architecture

The main goal is really to avoid code duplication as much as possible.

The symlink approach which was looked into several times, didn't feel robust enough to account for CLI differences.

Since most of the flags and the clap code was very similar, I wanted to move it in uucore::checksum::cli at first.
However, I felt like having CLI handling living under uucore was not a very good idea, since it would mix it with logic, and I wanted uucore to stay free from clap.

So I took the same approach as what's happening in base32|64|nc, where all the common code is defined in base32 which is then re-imported in base64 and basenc, only here I created a dedicated crate.

oech3 · 2026-01-09T13:21:05Z

Can we split this PR much more e.g. cli.rs?. diff would too big.

RenjiSann · 2026-01-09T13:55:28Z

Can we split this PR much more e.g. cli.rs?. diff would too big.

Some of the current diff is actually part of #10141,
I could split the introduction of uu_checksum_common, and the transition of the executables to it, but I think it would harm the review process to split them since they are very related

github-actions · 2026-01-09T14:40:10Z

GNU testsuite comparison:

Skip an intermittent issue tests/shuf/shuf-reservoir (fails in this run but passes in the 'main' branch)
Skip an intermittent issue tests/sort/sort-stale-thread-mem (fails in this run but passes in the 'main' branch)

github-actions · 2026-01-09T15:20:00Z

GNU testsuite comparison:

Skipping an intermittent issue tests/shuf/shuf-reservoir (passes in this run but fails in the 'main' branch)
Skipping an intermittent issue tests/sort/sort-stale-thread-mem (passes in this run but fails in the 'main' branch)

github-actions · 2026-01-09T15:45:15Z

GNU testsuite comparison:

Skipping an intermittent issue tests/shuf/shuf-reservoir (passes in this run but fails in the 'main' branch)
Skipping an intermittent issue tests/sort/sort-stale-thread-mem (passes in this run but fails in the 'main' branch)

github-actions · 2026-01-09T17:24:06Z

GNU testsuite comparison:

Skipping an intermittent issue tests/shuf/shuf-reservoir (passes in this run but fails in the 'main' branch)
Skipping an intermittent issue tests/sort/sort-stale-thread-mem (passes in this run but fails in the 'main' branch)

github-actions · 2026-01-09T18:04:01Z

GNU testsuite comparison:

Skipping an intermittent issue tests/shuf/shuf-reservoir (passes in this run but fails in the 'main' branch)
Skipping an intermittent issue tests/sort/sort-stale-thread-mem (passes in this run but fails in the 'main' branch)
Skipping an intermittent issue tests/timeout/timeout (passes in this run but fails in the 'main' branch)

github-actions · 2026-01-09T23:52:21Z

GNU testsuite comparison:

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

oech3 · 2026-01-10T11:24:36Z

When I measured binary size, I have
main make MULTICALL=y PROFILE=release-fast UTILS="cksum hashsum":
2577384 byte
This PR+main make MULTICALL=y PROFILE=release-fast UTILS="cksum md5sum ...":
2557320 byte
Symlinkable cksum: make MULTICALL=y PROFILE=release-fast UTILS="cksum":
2546184 byte

Also I don't see any reason to avoid having

#!/bin/sh
exec cksum --untagged -a ALGO "$@"

with --debug and --untagged extension.

oech3 · 2026-01-10T14:53:45Z

https://github.com/uutils/coreutils/actions/runs/20868639121/job/59976371446?pr=10142#step:13:1125 still assuming hashsum.
Maybe, we can discard hashsum after it was fixed?

oech3 · 2026-01-10T14:57:16Z

tests/by-util/test_sha512sum.rs

+    new_ucmd!()
+        .arg("--tag")
+        .arg("--text")
+        .arg("--md5")


--md5 should not be existing at anywhere.

Good catch, thanks

build.rs

oech3 · 2026-01-10T15:49:04Z

#10141 was merged, but diff is still big...

sylvestre · 2026-01-10T16:09:36Z

Starting from now, each binary has its own crate in src/uu

why ?

what is the impact on build time ?

oech3 · 2026-01-10T16:17:42Z

main: time make PROFILE=release MULTICALL=y UTILS="cksum hashsum": 1m 28s
patched: time make PROFILE=release MULTICALL=y UTILS="cksum md5sum ...": 1m 13s

main: time make PROFILE=release UTILS="cksum hashsum": 37.70s
patched: time make PROFILE=release UTILS="cksum md5sum ...": 2m 29s

oech3 · 2026-01-11T05:25:06Z

please remove https://github.com/uutils/coreutils/actions/runs/20888696488/job/60016203582?pr=10142#step:13:1125 too.

About build time, I don't mind. If distributor care about it, they should build multicall-binary (thought it affests to our CI...).

RenjiSann requested a review from sylvestre January 9, 2026 12:46

checksum: Rework the OutputFormat decision

da5da3b

RenjiSann force-pushed the peron/checksum-common-cli-post-reverse branch from 28e41e6 to ca32bc7 Compare January 9, 2026 12:49

RenjiSann added 2 commits January 9, 2026 15:12

introduce checksum_common

57df915

cksum: transition to checksum_common

559ecf9

RenjiSann force-pushed the peron/checksum-common-cli-post-reverse branch 2 times, most recently from b0e3843 to b640d07 Compare January 9, 2026 14:18

RenjiSann force-pushed the peron/checksum-common-cli-post-reverse branch from b640d07 to 9562744 Compare January 9, 2026 14:59

RenjiSann mentioned this pull request Jan 9, 2026

cksum, hashsum: Create a package to unify logic between cksum and standalone *sum binaries #10082

Closed

RenjiSann force-pushed the peron/checksum-common-cli-post-reverse branch from 9562744 to 8824943 Compare January 9, 2026 15:31

RenjiSann force-pushed the peron/checksum-common-cli-post-reverse branch 2 times, most recently from c449994 to e2a9486 Compare January 9, 2026 17:09

RenjiSann force-pushed the peron/checksum-common-cli-post-reverse branch from e2a9486 to a642bd6 Compare January 9, 2026 17:49

RenjiSann added 6 commits January 10, 2026 00:30

md5sum: introduce standalone binary

233c683

b2sum: introduce standalone binary

525df40

sha1sum: introduce standalone binary

bb06123

sha224sum: introduce standalone binary

2ab7680

sha256sum: introduce standalone binary

988541c

sha384sum: introduce standalone binary

662e7b5

RenjiSann force-pushed the peron/checksum-common-cli-post-reverse branch from a642bd6 to 959640e Compare January 9, 2026 23:30

RenjiSann marked this pull request as ready for review January 10, 2026 13:45

oech3 reviewed Jan 10, 2026

View reviewed changes

build.rs Show resolved Hide resolved

sha512sum: introduce standalone binary

fc4ea59

RenjiSann force-pushed the peron/checksum-common-cli-post-reverse branch from 959640e to fc4ea59 Compare January 11, 2026 03:20

Uh oh!

cksum/hashsum: Merge common logic under a common crate #10142

Are you sure you want to change the base?

cksum/hashsum: Merge common logic under a common crate #10142

Uh oh!

Conversation

RenjiSann commented Jan 9, 2026

Architecture explanation

Reason behind this architecture

Uh oh!

oech3 commented Jan 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

RenjiSann commented Jan 9, 2026

Uh oh!

github-actions bot commented Jan 9, 2026

Uh oh!

github-actions bot commented Jan 9, 2026

Uh oh!

github-actions bot commented Jan 9, 2026

Uh oh!

github-actions bot commented Jan 9, 2026

Uh oh!

github-actions bot commented Jan 9, 2026

Uh oh!

github-actions bot commented Jan 9, 2026

Uh oh!

oech3 commented Jan 10, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

oech3 commented Jan 10, 2026

Uh oh!

oech3 Jan 10, 2026

Choose a reason for hiding this comment

Uh oh!

RenjiSann Jan 11, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

oech3 commented Jan 10, 2026

Uh oh!

sylvestre commented Jan 10, 2026

Uh oh!

oech3 commented Jan 10, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

oech3 commented Jan 11, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

oech3 commented Jan 9, 2026 •

edited

Loading

oech3 commented Jan 10, 2026 •

edited

Loading

oech3 commented Jan 10, 2026 •

edited

Loading