touch: use write-open + futimens on Unix to trigger IN_CLOSE_WRITE by mattsu2020 · Pull Request #9813 · uutils/coreutils

mattsu2020 · 2025-12-24T09:48:13Z

Summary

Make touch open regular files for write and call futimens on the fd (Unix only). This matches GNU touch behavior and ensures inotify emits IN_CLOSE_WRITE when timestamps are updated.

Motivation

Some inotify-based reloaders watch only IN_CLOSE_WRITE. With the current path-based timestamp update, touch emits only IN_ATTRIB, so those watchers miss the change. This caused infinite loops in downstream tests (e.g., 0 A.D. hotload tests on Ubuntu Questing).

Changes

On Unix, attempt OpenOptions::new().write(true) and call futimens on the fd for regular files.
Fallback to the existing set_file_times path-based update when open/futimens fails or the target is not a regular file.
Keep existing symlink handling intact.

Testing

cargo test -p uu_touch

Notes

Behavior is Unix-only and preserves current semantics on non-Unix platforms.

related
#9812

src/uu/touch/src/touch.rs

sylvestre · 2025-12-24T09:57:16Z

is it possible to add tests ? thanks

mattsu2020 · 2025-12-25T10:20:52Z

is it possible to add tests ? thanks

add

sylvestre · 2025-12-29T12:31:30Z

sorry, it needs to be rebased

github-actions · 2026-01-05T11:28:39Z

GNU testsuite comparison:

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

github-actions · 2026-01-10T06:31:00Z

GNU testsuite comparison:

Congrats! The gnu test tests/stty/bad-speed is now passing!

codspeed-hq · 2026-01-19T12:16:43Z

Merging this PR will improve performance by 3.42%

⚡ 1 improved benchmark
✅ 16 untouched benchmarks
⏩ 311 skipped benchmarks¹

Performance Changes

	Mode	Benchmark	`BASE`	`HEAD`	Efficiency
⚡	Simulation	`cp_large_file[16]`	390.9 µs	377.9 µs	+3.42%

_{Comparing mattsu2020:bug_analystic (d0822d6) with main (de5c827)}

311 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. ↩

github-actions · 2026-01-19T12:36:38Z

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)

sylvestre · 2026-02-10T21:27:44Z

src/uu/touch/src/touch.rs

+    let mtime_spec = TimeSpec::new(mtime.unix_seconds() as _, mtime.nanoseconds() as _);
+
+    futimens(&file, &atime_spec, &mtime_spec)
+        .map_err(|err| std::io::Error::from_raw_os_error(err as i32))


he conversion from nix::Error to std::io::Error is incorrect

sylvestre · 2026-02-10T21:28:13Z

src/uu/touch/src/touch.rs

+    #[cfg(unix)]
+    {
+        // Open write-only and use futimens to trigger IN_CLOSE_WRITE on Linux.
+        if !is_stdout && try_futimens_via_write_fd(path, atime, mtime).is_ok() {


Silent error swallowing with is_ok() loses information about why futimens failed.

sylvestre · 2026-02-10T21:28:40Z

src/uu/touch/src/touch.rs

+/// access and modification times on the open FD (not by path), which also
+/// triggers `IN_CLOSE_WRITE` on Linux when the FD is closed.
+fn try_futimens_via_write_fd(path: &Path, atime: FileTime, mtime: FileTime) -> std::io::Result<()> {
+    let metadata = fs::metadata(path)?;


Unnecessary metadata check before open. Opens file twice (metadata + open). Should check metadata after opening via file.metadata() to avoid TOCTOU and reduce syscalls.

sylvestre · 2026-02-10T21:28:56Z

src/uu/touch/src/touch.rs

+fn try_futimens_via_write_fd(path: &Path, atime: FileTime, mtime: FileTime) -> std::io::Result<()> {
+    let metadata = fs::metadata(path)?;
+    if !metadata.is_file() {
+        return Err(Error::other("not a regular file"));


please use translate!()

sylvestre · 2026-02-10T21:29:21Z

src/uu/touch/src/touch.rs

+    }
+
+    let file = OpenOptions::new().write(true).open(path)?;
+    let atime_spec = TimeSpec::new(atime.unix_seconds() as _, atime.nanoseconds() as _);


Unsafe type cast 'as _' for nanoseconds. Should explicitly cast to i32: atime.nanoseconds() as i32

github-actions · 2026-02-10T21:39:33Z

GNU testsuite comparison:

Congrats! The gnu test tests/pr/bounded-memory is no longer failing!

sylvestre · 2026-02-10T21:41:24Z

src/uu/touch/src/touch.rs

        uu_app,
    };

+    #[cfg(unix)]


Redundant import - std::fs is already imported unconditionally at line 26

github-actions · 2026-02-11T12:17:32Z

GNU testsuite comparison:

GNU test failed: tests/touch/not-owner. tests/touch/not-owner is passing on 'main'. Maybe you have to rebase?

github-actions · 2026-02-14T12:11:48Z

GNU testsuite comparison:

GNU test failed: tests/factor/t10. tests/factor/t10 is passing on 'main'. Maybe you have to rebase?
GNU test failed: tests/factor/t34. tests/factor/t34 is passing on 'main'. Maybe you have to rebase?
GNU test failed: tests/touch/not-owner. tests/touch/not-owner is passing on 'main'. Maybe you have to rebase?
Skipping an intermittent issue tests/tail/follow-name (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)
Congrats! The gnu test tests/cut/cut-huge-range is now passing!
Congrats! The gnu test tests/rm/many-dir-entries-vs-OOM is now passing!

github-actions · 2026-02-18T09:07:19Z

GNU testsuite comparison:

GNU test failed: tests/rm/isatty. tests/rm/isatty is passing on 'main'. Maybe you have to rebase?
GNU test failed: tests/touch/not-owner. tests/touch/not-owner is passing on 'main'. Maybe you have to rebase?
Congrats! The gnu test tests/cut/bounded-memory is no longer failing!
Note: The gnu test tests/cp/link-heap is now being skipped but was previously passing.

github-actions · 2026-02-18T09:19:55Z

GNU testsuite comparison:

Congrats! The gnu test tests/cut/bounded-memory is no longer failing!
Note: The gnu test tests/rm/many-dir-entries-vs-OOM is now being skipped but was previously passing.
Note: The gnu test tests/unexpand/bounded-memory is now being skipped but was previously passing.

On Unix, try futimens through a write-opened file descriptor to trigger IN_CLOSE_WRITE semantics, and fall back to set_file_times when needed. Also adds a unit test for the futimens path and wires unix-only dependencies.

github-actions · 2026-02-21T18:59:37Z

GNU testsuite comparison:

GNU test failed: tests/cut/bounded-memory. tests/cut/bounded-memory is passing on 'main'. Maybe you have to rebase?
Skip an intermittent issue tests/pr/bounded-memory (fails in this run but passes in the 'main' branch)
Skipping an intermittent issue tests/tail/follow-name (passes in this run but fails in the 'main' branch)
Congrats! The gnu test tests/date/date-locale-hour is no longer failing!

sylvestre reviewed Dec 24, 2025

View reviewed changes

src/uu/touch/src/touch.rs Outdated Show resolved Hide resolved

sylvestre reviewed Dec 24, 2025

View reviewed changes

src/uu/touch/src/touch.rs Show resolved Hide resolved

mattsu2020 requested a review from sylvestre January 5, 2026 10:16

sylvestre reviewed Feb 10, 2026

View reviewed changes

mattsu2020 force-pushed the bug_analystic branch from d741ba8 to 8e88196 Compare February 18, 2026 09:11

touch: set file times via futimens on a write fd

d0822d6

On Unix, try futimens through a write-opened file descriptor to trigger IN_CLOSE_WRITE semantics, and fall back to set_file_times when needed. Also adds a unit test for the futimens path and wires unix-only dependencies.

sylvestre force-pushed the bug_analystic branch from 8e88196 to d0822d6 Compare February 21, 2026 18:51

Uh oh!

Comments

Conversation

mattsu2020 commented Dec 24, 2025

Summary

Motivation

Changes

Testing

Notes

Uh oh!

Uh oh!

Uh oh!

sylvestre commented Dec 24, 2025

Uh oh!

mattsu2020 commented Dec 25, 2025

Uh oh!

sylvestre commented Dec 29, 2025

Uh oh!

github-actions bot commented Jan 5, 2026

Uh oh!

github-actions bot commented Jan 10, 2026

Uh oh!

codspeed-hq bot commented Jan 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Merging this PR will improve performance by 3.42%

Performance Changes

Footnotes

Uh oh!

github-actions bot commented Jan 19, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

github-actions bot commented Feb 10, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

github-actions bot commented Feb 11, 2026

Uh oh!

github-actions bot commented Feb 14, 2026

Uh oh!

github-actions bot commented Feb 18, 2026

Uh oh!

github-actions bot commented Feb 18, 2026

Uh oh!

github-actions bot commented Feb 21, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

codspeed-hq bot commented Jan 19, 2026 •

edited

Loading