seq: Buffer writes to stdout

[drinkcat]

Use a BufWriter to wrap stdout: reduces the numbers of system calls, improves performance drastically (2x in some cases).

Fixes part of #7482, see this comment: #7482 (comment)

Thanks @mjguzik!

For floating point prints, we are now faster than GNU:

taskset -c 0 hyperfine -L seq seq,./seq-main,target/release/seq "{seq} 0 0.1 100000"
Benchmark 1: seq 0 0.1 100000
  Time (mean ± σ):     161.6 ms ±   0.3 ms    [User: 160.8 ms, System: 0.6 ms]
  Range (min … max):   161.2 ms … 162.4 ms    18 runs
 
Benchmark 2: ./seq-main 0 0.1 100000
  Time (mean ± σ):     282.7 ms ±   5.0 ms    [User: 221.0 ms, System: 60.0 ms]
  Range (min … max):   279.7 ms … 296.2 ms    10 runs

Benchmark 3: target/release/seq 0 0.1 100000
  Time (mean ± σ):     143.8 ms ±   0.3 ms    [User: 143.0 ms, System: 0.6 ms]
  Range (min … max):   143.2 ms … 144.4 ms    20 runs
 
Summary
  target/release/seq 0 0.1 100000 ran
    1.12 ± 0.00 times faster than seq 0 0.1 100000
    1.97 ± 0.03 times faster than ./seq-main 0 0.1 100000

[github-actions]

GNU testsuite comparison:

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

[sylvestre]

please add this bench in https://github.com/uutils/coreutils/blob/main/src/uu/seq/BENCHMARKING.md
thanks

[drinkcat]

please add this bench in https://github.com/uutils/coreutils/blob/main/src/uu/seq/BENCHMARKING.md thanks

Sure, done. I also stole some of the content I had written for #7458 (will rebase that MR on top of this one, I assume this one will be merged first).

[sylvestre]

i don't think we need to document the results
just the commands

[drinkcat]

Dropped.

I saw some results in some other BENCHMARKING.md so I just put them there... Might be nice to codify what's expected in those documentation files (maybe in CONTRIBUTING.md?). Not a big deal.

[sylvestre]

sure, don't hesitate to make a proposition :)

[github-actions]

GNU testsuite comparison:

GNU test failed: timeout/timeout.log. timeout/timeout.log is passing on 'main'. Maybe you have to rebase?
Congrats! The gnu test misc/stdbuf.log is no longer failing!

[github-actions]

GNU testsuite comparison:

Congrats! The gnu test timeout/timeout.log is no longer failing!