Avoid repeated conversions when shifting Calendar.ISO datetimes

[preciz]

Assisted-by: Codex:GPT-6

~1.5x faster and uses less memory.

Combine seconds and microseconds into one shift after applying months,
preserving existing precision semantics. Inline duration arithmetic and
set precision upfront to tail-call the shift helper.

Bench:

# Run with: bin/elixir bench.exs
Mix.install([:benchee])

path = "lib/elixir/lib/calendar/iso.ex"

for {ref, module} <- [{"main", Old}, {"HEAD", New}] do
  {source, 0} = System.cmd("git", ["show", "#{ref}:#{path}"], cd: __DIR__)

  source
  |> String.replace("defmodule Calendar.ISO do", "defmodule #{inspect(module)} do")
  |> Code.compile_string(path)
end

inputs = %{
  "months only, clamp to leap day" => [month: 1],
  "seconds only, cross day" => [second: 86_401],
  "microseconds only" => [microsecond: {500_001, 6}],
  "seconds and microseconds" => [second: 86_401, microsecond: {500_001, 6}],
  "all units" => [
    year: 1,
    month: 1,
    week: 2,
    day: 3,
    hour: 4,
    minute: 5,
    second: 6,
    microsecond: {7, 6}
  ]
}

Benchee.run(
  %{
    "main" => &Old.shift_naive_datetime(2024, 1, 31, 23, 59, 59, {120_000, 3}, &1),
    "HEAD" => &New.shift_naive_datetime(2024, 1, 31, 23, 59, 59, {120_000, 3}, &1)
  },
  inputs: Map.new(inputs, fn {name, units} -> {name, Duration.new!(units)} end),
  pre_check: :all_same,
  warmup: 1,
  time: 3,
  memory_time: 1
)

Results:

Operating System: Linux
CPU Information: AMD Ryzen 7 8845HS w
Number of Available Cores: 16
Available memory: 54.72 GB
Elixir 1.21.0-dev
Erlang 29.0.5
JIT enabled: true

Benchmark suite executing with the following configuration:
warmup: 1 s
time: 3 s
memory time: 1 s
reduction time: 0 ns
parallel: 1
inputs: all units, microseconds only, months only, clamp to leap day, seconds and microseconds, seconds only, cross day
Estimated total run time: 50 s
Excluding outliers: false

##### With input all units #####
Name           ips        average  deviation         median         99th %
HEAD        4.46 M      224.11 ns  ±2511.99%         200 ns         370 ns
main        2.31 M      432.57 ns  ±1539.64%         400 ns         591 ns

Comparison: 
HEAD        4.46 M
main        2.31 M - 1.93x slower +208.46 ns

Memory usage statistics:

Name    Memory usage
HEAD         0.53 KB
main         1.05 KB - 1.97x memory usage +0.52 KB

**All measurements for memory usage were the same**

##### With input microseconds only #####
Name           ips        average  deviation         median         99th %
HEAD        4.67 M      214.06 ns  ±3867.57%         180 ns         341 ns
main        3.91 M      255.70 ns  ±4816.90%         200 ns         371 ns

Comparison: 
HEAD        4.67 M
main        3.91 M - 1.19x slower +41.64 ns

Memory usage statistics:

Name    Memory usage
HEAD           512 B
main           552 B - 1.08x memory usage +40 B

**All measurements for memory usage were the same**

##### With input months only, clamp to leap day #####
Name           ips        average  deviation         median         99th %
HEAD       11.81 M       84.67 ns  ±7808.55%          60 ns         100 ns
main        6.96 M      143.75 ns  ±5285.02%         110 ns         250 ns

Comparison: 
HEAD       11.81 M
main        6.96 M - 1.70x slower +59.08 ns

Memory usage statistics:

Name    Memory usage
HEAD           152 B
main           312 B - 2.05x memory usage +160 B

**All measurements for memory usage were the same**

##### With input seconds and microseconds #####
Name           ips        average  deviation         median         99th %
HEAD        4.89 M      204.67 ns  ±3842.69%         170 ns         330 ns
main        2.45 M      408.15 ns  ±3616.15%         331 ns         711 ns

Comparison: 
HEAD        4.89 M
main        2.45 M - 1.99x slower +203.48 ns

Memory usage statistics:

Name    Memory usage
HEAD           512 B
main           944 B - 1.84x memory usage +432 B

**All measurements for memory usage were the same**

##### With input seconds only, cross day #####
Name           ips        average  deviation         median         99th %
HEAD        4.83 M      206.98 ns  ±3756.77%         170 ns         330 ns
main        3.90 M      256.49 ns  ±4007.68%         201 ns         391 ns

Comparison: 
HEAD        4.83 M
main        3.90 M - 1.24x slower +49.50 ns

Memory usage statistics:

Name    Memory usage
HEAD           536 B
main           576 B - 1.07x memory usage +40 B

**All measurements for memory usage were the same**

[preciz]

nesting in parentheses here allows the compiler to better optimize

[josevalim]

💚 💙 💜 💛 ❤️