feat(bench): --report-json / --report-junit for benchmark results

[Chemaclass]

Problem

bashunit bench is a second-class citizen. Its entire option set is:

-e/--env, -f/--filter, -s/--simple, --detailed, -vvv, --skip-env-file, -l/--login, --no-color, -h

No reports of any kind. Results are printed by bashunit::benchmark::print_results (src/benchmark/results.sh:19) and then discarded, so a benchmark run in CI leaves nothing behind: no artifact to store, nothing to chart over time, nothing for a later run to compare against. Meanwhile bashunit test has JUnit, TAP, JSON, HTML and GHA reporters.

This blocks the obvious next step of failing a build on a performance regression, which is the whole point of running benchmarks in CI.

Proposal

--report-json <file> Machine-readable benchmark results
--report-junit <file> JUnit XML (so CI shows benchmarks alongside tests)

JSON shape, at minimum, per benchmark function: file, function, human-readable name, revs, its, per-iteration timings, average, min, max, median, and the @max_ms threshold with its pass/fail verdict. Plus a run-level block: total duration, timestamp, bashunit version, and enough host context (OS, bash version) to make two runs comparable.

Where to change

• src/benchmark/results.sh:11 add_result — the collection point.
• src/benchmark/run.sh:5 run_function, src/benchmark/annotations.sh for the @max_ms verdict.
• Reuse src/reports/ writers where the shape allows; do not copy-paste a fourth parser (refactor(coverage,reports): share three copy-pasted parsers #971 consolidated three).
• src/main/bench.sh for the flags plus --help, and both completion scripts.

Acceptance criteria

• bashunit bench --report-json out.json writes valid JSON (verify with jq)
• Every benchmark function appears exactly once with its timings and its revs/its
• The @max_ms threshold and verdict are present when the annotation is set
• Run-level metadata included: timestamp, total duration, bashunit version, bash version, OS
• --report-junit produces XML that a CI test reporter accepts
• Numbers in the report match what the console prints
• Unwritable path fails fast (A missing --env bootstrap or an unwritable report path fails silently and exits 0 #875)
• Locale-independent decimal formatting (env LC_ALL=C awk, the 7 test failures on Bash 5.3 macOS (nix-shell) #912 workaround)
• bashunit bench --help and both completion scripts updated (parity test)
• Tests in tests/unit/benchmark/ and tests/benchmark/

Repo checklist (agent)

• TDD: RED → GREEN → REFACTOR. Write the failing test first.
• Bash 3.0+ only: no printf -v, no += append, no declare -A, no [[ ]], no ${var,,}, no &>>, no ${arr[-1]}. Expanding a possibly-empty array under set -u needs ${arr[@]+"${arr[@]}"}.
• A new CLI flag must be wired in all of these or a parity test fails:
  • parse in src/main/test.sh (report-style flags need export -n, see src/main/test.sh:188-196 for why)
  • validate the value in bashunit::main::validate_config_or_exit (src/main/validate.sh:60) — unvalidated input used to run the wrong thing and exit 0 (Unknown options are silently ignored: a typo'd flag runs a different suite and exits 0 #871, --jobs with a non-integer value hangs on Bash 3.x and is silently ignored on Bash 4.3+ #873)
  • default + env var in src/config/env.sh and a documented line in .env.example
  • --help text in the same block it belongs to
  • both completions/bashunit.bash and completions/_bashunit (anti-drift test feat(cli): bash and zsh completion scripts with an anti-drift test #778 fails otherwise)
• Gates: make sa, make lint, ./bashunit tests/, ./bashunit --parallel tests/. Never run shfmt -w.
• Docs: update docs/command-line.md. Editing docs/assertions.md invalidates the bashunit doc acceptance snapshot — regenerate it.
• CHANGELOG.md: add one line under ## Unreleased.
• Fixtures under tests/acceptance/fixtures/ must not end in *test.sh.
• One issue = one PR.

[Chemaclass]

Done in #1072 (merged).