timers: fix unsafe array iteration

[RaisinTen]

Fixes: #37222

[nodejs-github-bot]

CI: https://ci.nodejs.org/job/node-test-pull-request/35937/

[RaisinTen]

Benchmark CI for timers: https://ci.nodejs.org/view/Node.js%20benchmark/job/benchmark-node-micro-benchmarks/930/
Using SafeArrayIterator has resulted in a significant perf regression:

Benchmark results

19:56:10 timers/immediate.jstype='breadth4' n=5000000                         *     -4.65 %       ±3.97%  ±5.30%  ±6.92%
19:56:10 timers/immediate.jstype='depth1' n=5000000                           *      1.93 %       ±1.86%  ±2.49%  ±3.26%
19:56:10 timers/timers-breadth-args.jsn=1000000                             ***    -24.02 %       ±5.82%  ±7.78% ±10.19%

[nodejs-github-bot]

CI: https://ci.nodejs.org/job/node-test-pull-request/35940/

[RaisinTen]

New Benchmark CI (using ReflectApply):
https://ci.nodejs.org/view/Node.js%20benchmark/job/benchmark-node-micro-benchmarks/932/

Benchmark results

                                                            confidence improvement accuracy (*)    (**)   (***)
timers/immediate.jstype='breadth1' n=5000000                                0.50 %       ±3.18%  ±4.23%  ±5.51%
timers/immediate.jstype='breadth4' n=5000000                               -1.48 %       ±4.55%  ±6.06%  ±7.89%
timers/immediate.jstype='breadth' n=5000000                                -1.31 %       ±3.90%  ±5.18%  ±6.75%
timers/immediate.jstype='clear' n=5000000                                   1.78 %       ±3.34%  ±4.45%  ±5.80%
timers/immediate.jstype='depth1' n=5000000                                  1.06 %       ±2.53%  ±3.37%  ±4.40%
timers/immediate.jstype='depth' n=5000000                                  -0.27 %       ±2.51%  ±3.35%  ±4.37%
timers/set-immediate-breadth-args.jsn=5000000                              -2.31 %       ±3.44%  ±4.57%  ±5.95%
timers/set-immediate-breadth.jsn=10000000                                  -1.88 %       ±4.10%  ±5.46%  ±7.11%
timers/set-immediate-depth-args.jsn=5000000                                 0.42 %       ±1.57%  ±2.09%  ±2.72%
timers/timers-breadth-args.jsn=1000000                                      1.50 %       ±6.38%  ±8.48% ±11.04%
timers/timers-breadth.jsn=5000000                                           1.18 %       ±3.17%  ±4.22%  ±5.50%
timers/timers-cancel-pooled.jsn=5000000                                     8.24 %      ±11.36% ±15.12% ±19.69%
timers/timers-cancel-unpooled.jsdirection='end' n=1000000                  -1.42 %       ±8.14% ±10.83% ±14.10%
timers/timers-cancel-unpooled.jsdirection='start' n=1000000                -2.38 %       ±5.38%  ±7.16%  ±9.32%
timers/timers-depth.jsn=1000                                                0.48 %       ±0.78%  ±1.04%  ±1.35%
timers/timers-insert-pooled.jsn=5000000                                     0.65 %       ±3.05%  ±4.06%  ±5.29%
timers/timers-insert-unpooled.jsdirection='end' n=1000000                  -2.16 %       ±3.85%  ±5.12%  ±6.66%
timers/timers-insert-unpooled.jsdirection='start' n=1000000                 1.73 %       ±5.02%  ±6.68%  ±8.70%
timers/timers-timeout-nexttick.jsn=50000                                   -0.24 %       ±2.17%  ±2.89%  ±3.76%
timers/timers-timeout-nexttick.jsn=5000000                                  0.25 %       ±6.66%  ±8.87% ±11.55%
timers/timers-timeout-pooled.jsn=10000000                                  -0.62 %       ±6.87%  ±9.15% ±11.91%
timers/timers-timeout-unpooled.jsn=1000000                           *    -12.70 %       ±9.87% ±13.19% ±17.28%

Be aware that when doing many comparisons the risk of a false-positive
result increases. In this case, there are 22 comparisons, you can thus
expect the following amount of false-positive results:
  1.10 false positives, when considering a   5% risk acceptance (*, **, ***),
  0.22 false positives, when considering a   1% risk acceptance (**, ***),
  0.02 false positives, when considering a 0.1% risk acceptance (***)

[aduh95]

Benchmark results are a bit weird, I kicked off a new benchmark CI to see if it's just a flakiness.
https://ci.nodejs.org/view/Node.js%20benchmark/job/benchmark-node-micro-benchmarks/934/

[Fishrock123]

I think Reflect probably causes more overhead than just doing a more manual check.

[aduh95]

Benchmark CI results look good 🎉

                                                            confidence improvement accuracy (*)    (**)   (***)
timers/immediate.jstype='breadth1' n=5000000                                1.00 %       ±4.06%  ±5.41%  ±7.04%
timers/immediate.jstype='breadth4' n=5000000                               -1.00 %       ±4.15%  ±5.53%  ±7.21%
timers/immediate.jstype='breadth' n=5000000                                 3.12 %       ±4.05%  ±5.38%  ±7.01%
timers/immediate.jstype='clear' n=5000000                                  -0.04 %       ±2.15%  ±2.86%  ±3.72%
timers/immediate.jstype='depth1' n=5000000                                 -0.03 %       ±1.74%  ±2.31%  ±3.02%
timers/immediate.jstype='depth' n=5000000                                   0.75 %       ±2.22%  ±2.95%  ±3.85%
timers/set-immediate-breadth-args.jsn=5000000                               0.19 %       ±4.62%  ±6.16%  ±8.03%
timers/set-immediate-breadth.jsn=10000000                            *      4.00 %       ±3.89%  ±5.18%  ±6.74%
timers/set-immediate-depth-args.jsn=5000000                          *      2.03 %       ±1.80%  ±2.40%  ±3.14%
timers/timers-breadth-args.jsn=1000000                                     -5.37 %       ±5.51%  ±7.33%  ±9.54%
timers/timers-breadth.jsn=5000000                                           0.81 %       ±3.38%  ±4.50%  ±5.86%
timers/timers-cancel-pooled.jsn=5000000                                     1.91 %      ±10.69% ±14.22% ±18.51%
timers/timers-cancel-unpooled.jsdirection='end' n=1000000                  -0.92 %       ±5.95%  ±7.91% ±10.30%
timers/timers-cancel-unpooled.jsdirection='start' n=1000000                 0.18 %       ±6.67%  ±8.89% ±11.58%
timers/timers-depth.jsn=1000                                               -0.13 %       ±0.65%  ±0.87%  ±1.14%
timers/timers-insert-pooled.jsn=5000000                                    -0.54 %       ±2.80%  ±3.74%  ±4.88%
timers/timers-insert-unpooled.jsdirection='end' n=1000000                   1.44 %       ±4.47%  ±5.95%  ±7.74%
timers/timers-insert-unpooled.jsdirection='start' n=1000000                 0.55 %       ±5.13%  ±6.82%  ±8.88%
timers/timers-timeout-nexttick.jsn=50000                                    0.14 %       ±2.20%  ±2.93%  ±3.83%
timers/timers-timeout-nexttick.jsn=5000000                                  0.13 %       ±3.60%  ±4.80%  ±6.25%
timers/timers-timeout-pooled.jsn=10000000                            *      8.82 %       ±8.47% ±11.29% ±14.72%
timers/timers-timeout-unpooled.jsn=1000000                                  0.62 %      ±13.21% ±17.58% ±22.88%

Be aware that when doing many comparisons the risk of a false-positive
result increases. In this case, there are 22 comparisons, you can thus
expect the following amount of false-positive results:
  1.10 false positives, when considering a   5% risk acceptance (*, **, ***),
  0.22 false positives, when considering a   1% risk acceptance (**, ***),
  0.02 false positives, when considering a 0.1% risk acceptance (***)

[nodejs-github-bot]

CI: https://ci.nodejs.org/job/node-test-pull-request/35949/

[benjamingr]

Ping @nodejs/timers

[Trott]

Landed in c0e66e3