feat(cli): --repeat <n> to hunt flaky tests before they reach CI

[Chemaclass]

Problem

--retry mitigates flakiness once it has already burned a CI run. There is no way to go looking for it: to prove a test is deterministic, you have to run the suite N times by hand and diff the outputs.

Related: #1012 classifies a test that only passed on retry as flaky — that tells you a test was flaky after the fact; --repeat lets you find out before merging.

Proposal

--repeat <n>      Run each selected test n times (default 1); the test fails if any iteration fails

• Applies per test, inside the runner, so --repeat 50 --filter flaky_candidate is the natural "hammer this one test" invocation.
• A test is reported once, with the aggregate outcome. On failure, the message names the failing iteration: failed on iteration 7 of 50.
• Counters count the test once, not n times.
• Composes with --parallel (each worker repeats its own test) and with --retry (repeat is the outer loop, retry the inner one) — document the interaction explicitly.
• --repeat 0 is a usage error, not a silent no-op.

Where to change

• src/runner/exec.sh:312 — the retry while :; loop is the natural place; repeat wraps it.
• src/main/test.sh, bashunit::main::require_non_negative_int_or_exit (src/main/validate.sh:19) for validation, src/config/env.sh (BASHUNIT_REPEAT), .env.example.

Acceptance criteria

• --repeat 3 runs each test body three times
• A test that fails on any iteration is reported failed, with the iteration number in the message
• The test appears once in the summary and in every report
• Assertion counts are not multiplied by the repeat factor (or the chosen semantics are documented and tested)
• --repeat 1 is byte-identical to no flag
• --repeat abc and --repeat -1 exit non-zero with a usage error (--jobs with a non-integer value hangs on Bash 3.x and is silently ignored on Bash 4.3+ #873 class of bug)
• Interaction with --retry is documented and covered by a test
• Works under --parallel
• Per-test hooks (set_up / tear_down) run once per iteration; set_up_before_script does not
• Acceptance test in tests/acceptance/

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 #1041 (merged).