Use bats-assert library functions in BATS test suite

[Copilot]

As suggested by @casperklein, replace manual shell assertions with bats-assert library functions for cleaner tests and better failure output (expected vs. actual diffs on failure).

Changes

• test/run.sh: Install bats-support and bats-assert alongside bats-core, pinned to specific versions (v0.3.0 / v2.1.0) and overridable via BATS_SUPPORT_VERSION / BATS_ASSERT_VERSION
• test/test_suite.bats: Load the libraries and replace manual assertions throughout:
  • [ "$status" -eq 0 ] → assert_success
  • [[ "$output" == *"..."* ]] → assert_output --partial "..."
  • [ "$output" = "..." ] → assert_output "..."
  • In the shutdown test (logs captured to a variable rather than via run), pipe through run echo "$logs" to make assert_output --partial usable

Before / After

# Before
run docker exec "$CONTAINER_DEFAULT" pihole-FTL -vv
[ "$status" -eq 0 ]
[[ "$output" == *"Version:"* ]]

# After
run docker exec "$CONTAINER_DEFAULT" pihole-FTL -vv
assert_success
assert_output --partial "Version:"

[T-o-m-H-u]

The changelog suggests this to be the first AI generated code in pi-hole–sort of, since this only affects code test.

Does the project plan on writing an AI code policy, to make the use of such code and its review process transparent to users?

[yubiuser]

Does the project plan on writing an AI code policy, to make the use of such code and its review process transparent to users?

Sounds reasonable.

[PromoFaux]

I don't suppose I've ever really considered it. Can you point to some other projects that implement such a policy?

[T-o-m-H-u]

@PromoFaux Sure, there are plenty. The arguments for and against AI generated code are rather well hashed out already. I link to the Linux policy as an example, as others are to be found. There is probably a better place for discussion, than this PR.

Linux AI policy