tee: do not retry on EINTR

[gabrielhnf]

Related to #12469

Writer::write_all was delegating to the standard library's write_all, which unconditionally retries on EINTR.

Replaces the write_all delegation with a manual loop that calls write() directly and propagates EINTR immediately without retrying.

[codspeed-hq]

Merging this PR will not alter performance

✅ 319 untouched benchmarks
⏩ 46 skipped benchmarks¹

---

_{Comparing gabrielhnf:fix-tee-eintr-unconditional-retry (85a6062) with main (55549fa)}

Footnotes

1. 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. ↩

[sylvestre]

it needs comment

[gabrielhnf]

How about adding write_all_with_eintr to https://github.com/uutils/coreutils/blob/main/src/uucore/src/lib/mods/io.rs instead? This PR is duplicating code.

Yeah, might be a good idea. Will close the PR and research the issue

[oech3]

Actually, I was considering to dedup some code with cat's manual loop. cat were not affected.

[gabrielhnf]

Are you planning to tackle this yourself then?

[github-actions]

GNU testsuite comparison:

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

[oech3]

I think there is no compatibility requirement for non-unix. But does EINTR exist on Windows?

[gabrielhnf]

i did not add write_all_no_retry for non-unix...

Sorry, got things a little messy here and accidentally changed that. Just fixed it

[oech3]

Thanks. It is my mistake.

[oech3]

I noticed that there is

  -i, --ignore-interrupts
          ignore interrupt signals (ignored on non-Unix platforms)

. So this solution is still incomplete. done at different place?

[gabrielhnf]

Doesn't seem to be an issue, -i​ sets SIGINT​ to SIG_IGN​ through ignore_interrupts()​ at the start of tee()​, so write()​ shouldn't get EINTR​ from it, no?