seq: Directly write separator string, instead of using format

[drinkcat]

Doing stdout.write_all(separator.as_bytes())? is quite a bit faster than using format to do the same operation: write!(stdout, "{separator}")?.

This speeds up by about 10% on simple cases.

We do the same for the terminator even though this has no measurable performance impact.

---

Integer increments (no point comparing with coreutils on that one, it's much faster):

$ cargo build -r -p uu_seq && taskset -c 0 hyperfine --warmup 3 -L seq target/release/seq,./seq-main "{seq} 1 1000000"
    Finished `release` profile [optimized] target(s) in 0.12s
Benchmark 1: target/release/seq 1 1000000
  Time (mean ± σ):     162.9 ms ±   2.4 ms    [User: 161.5 ms, System: 0.9 ms]
  Range (min … max):   160.4 ms … 167.7 ms    17 runs
 
Benchmark 2: ./seq-main 1 1000000
  Time (mean ± σ):     182.2 ms ±   4.6 ms    [User: 179.9 ms, System: 0.8 ms]
  Range (min … max):   178.6 ms … 198.5 ms    16 runs
  
Summary
  target/release/seq 1 1000000 ran
    1.12 ± 0.03 times faster than ./seq-main 1 1000000

Float increments, we were already faster than coreutils, we gain another 10%:

$ cargo build -r -p uu_seq && taskset -c 0 hyperfine --warmup 3 -L seq target/release/seq,./seq-main,seq "{seq} 1 1.1 1000000"
    Finished `release` profile [optimized] target(s) in 0.12s
Benchmark 1: target/release/seq 1 1.1 1000000
  Time (mean ± σ):     222.0 ms ±   1.8 ms    [User: 220.9 ms, System: 0.6 ms]
  Range (min … max):   219.9 ms … 225.0 ms    13 runs
 
Benchmark 2: ./seq-main 1 1.1 1000000
  Time (mean ± σ):     246.4 ms ±   5.9 ms    [User: 243.5 ms, System: 0.8 ms]
  Range (min … max):   241.3 ms … 262.7 ms    11 runs
 
Benchmark 3: seq 1 1.1 1000000
  Time (mean ± σ):     274.2 ms ±   6.3 ms    [User: 271.1 ms, System: 0.9 ms]
  Range (min … max):   267.7 ms … 290.0 ms    10 runs
 
Summary
  target/release/seq 1 1.1 1000000 ran
    1.11 ± 0.03 times faster than ./seq-main 1 1.1 1000000
    1.23 ± 0.03 times faster than seq 1 1.1 1000000

[drinkcat]

(probably a good idea to grep the rest of the repo for that pattern...)

[github-actions]

GNU testsuite comparison:

Congrats! The gnu test tests/tail/inotify-dir-recreate is no longer failing!
Congrats! The gnu test tests/timeout/timeout is no longer failing!

[sylvestre]

wahou, i wasn't expecting such difference

[drinkcat]

Oh, I also tried to move that test out of the loop. That makes no difference.