test: replace custom cfg with TEST_SKIP environment variable

[KSXGitHub]

Summary

This PR replaces the custom --cfg pdu_test_skip_* mechanism with a simpler TEST_SKIP environment variable for skipping tests at runtime. This avoids unnecessary recompilation when tests need to be skipped based on runtime conditions.

Key Changes

• test.sh: Added support for TEST_SKIP environment variable that accepts space-separated test names and passes them to cargo test via --skip flags
• tests/cli_errors.rs: Removed #[cfg_attr(pdu_test_skip_fs_errors, ignore)] and updated the error hint to use TEST_SKIP='fs_errors' instead of RUSTFLAGS='--cfg pdu_test_skip_fs_errors'
• Cargo.toml: Removed the [lints.rust] section that declared the pdu_test_skip_fs_errors cfg as expected
• CONTRIBUTING.md: Updated test skipping guidance to recommend TEST_SKIP for runtime environment conditions and removed the example using custom cfg flags
• Documentation: Updated CONTRIBUTING.md, CLAUDE.md, AGENTS.md, and shared AI instructions to reference TEST_SKIP instead of RUSTFLAGS and custom cfg flags

Implementation Details

The test.sh script now:

1. Reads the TEST_SKIP environment variable as a space-separated list
2. Converts each test name into a --skip argument
3. Passes these arguments to cargo test only when skip arguments are present

This approach allows tests to remain compiled on all configurations (catching type errors early) while being skipped at runtime based on environment conditions, without requiring recompilation.

https://claude.ai/code/session_01UBzLwHNmqEC2SafFKjEqYz

[github-actions]

Performance Regression Reports

commit: c4a33e3

--quantity=apparent-size --max-depth=1 --min-ratio=0.01

Command	Mean [ms]	Min [ms]	Max [ms]	Relative
pdu	120.1 ± 13.3	95.1	132.3	1.25 ± 0.14
pdu-0.20.0	95.8 ± 0.4	95.1	97.0	1.00

Logs

Benchmark 1: pdu
  Time (mean ± σ):     120.1 ms ±  13.3 ms    [User: 88.6 ms, System: 286.5 ms]
  Range (min … max):    95.1 ms … 132.3 ms    23 runs
 
  Warning: Statistical outliers were detected. Consider re-running this benchmark on a quiet system without any interferences from other programs. It might help to use the '--warmup' or '--prepare' options.
 
Benchmark 2: pdu-0.20.0
  Time (mean ± σ):      95.8 ms ±   0.4 ms    [User: 93.0 ms, System: 280.8 ms]
  Range (min … max):    95.1 ms …  97.0 ms    31 runs
 
Summary
  pdu-0.20.0 ran
    1.25 ± 0.14 times faster than pdu

JSON

{
  "results": [
    {
      "command": "pdu",
      "mean": 0.12006047361217392,
      "stddev": 0.013343557065973555,
      "median": 0.12705781496000002,
      "user": 0.08860871739130434,
      "system": 0.28654005217391304,
      "min": 0.09510904396,
      "max": 0.13226357696000002,
      "times": [
        0.12774852796000002,
        0.12691368096000002,
        0.12205733196,
        0.12771803096,
        0.12698449196000003,
        0.12590222896000003,
        0.13226357696000002,
        0.12772239096000002,
        0.12743405396000002,
        0.12705781496000002,
        0.12684034196000002,
        0.12713915596000003,
        0.12754833296,
        0.12750079696000002,
        0.12771375496,
        0.12733985696000003,
        0.12743210996,
        0.11984200396,
        0.09606498796,
        0.09659742696,
        0.09530366496,
        0.09515728596,
        0.09510904396
      ],
      "exit_codes": [
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0
      ]
    },
    {
      "command": "pdu-0.20.0",
      "mean": 0.0958133803470968,
      "stddev": 0.00044423652271469643,
      "median": 0.09582443196,
      "user": 0.09303498387096773,
      "system": 0.28079410967741936,
      "min": 0.09505560496,
      "max": 0.09696005896,
      "times": [
        0.09518470396,
        0.09599715196,
        0.09564958696,
        0.09586240196,
        0.09598966796,
        0.09640742096,
        0.09532143996,
        0.09547397796,
        0.09531694296,
        0.09532619696,
        0.09600397496,
        0.09598114196,
        0.09696005896,
        0.09505560496,
        0.09534498196,
        0.09585327496,
        0.09608002596,
        0.09582443196,
        0.09615227996,
        0.09627631996,
        0.09563300696,
        0.09543804396,
        0.09588356896,
        0.09559013596,
        0.09546652696,
        0.09611445896,
        0.09582058696,
        0.09662026896,
        0.09549784596,
        0.09561676596,
        0.09647199396
      ],
      "exit_codes": [
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0
      ]
    }
  ]
}

[Copilot]

Pull request overview

This PR migrates the repo’s “skip certain integration tests” workflow from compile-time custom --cfg pdu_test_skip_* flags to a runtime TEST_SKIP environment variable, aiming to avoid recompilation when skipping environment-dependent tests.

Changes:

• Add TEST_SKIP handling to test.sh, translating names into libtest --skip arguments.
• Update the fs_errors test hint to reference TEST_SKIP and remove the custom cfg-based ignore.
• Remove the Cargo.toml unexpected_cfgs allowance and update developer/AI docs to the new mechanism.

Reviewed changes

Copilot reviewed 9 out of 9 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
tests/cli_errors.rs	Removes cfg-based ignore and updates the root-run hint to use TEST_SKIP.
test.sh	Parses TEST_SKIP and conditionally appends libtest --skip args to cargo test.
template/ai-instructions/shared.md	Updates guidance to reference TEST_SKIP instead of custom cfg flags.
README.md	Documents the new TEST_SKIP environment variable for test.sh.
Cargo.toml	Removes unexpected_cfgs configuration for the deleted custom cfg.
CONTRIBUTING.md	Updates test-skipping guidance to prefer TEST_SKIP for runtime-environment skips.
CLAUDE.md	Syncs AI instructions to reference TEST_SKIP.
AGENTS.md	Syncs AI instructions to reference TEST_SKIP.
.github/copilot-instructions.md	Syncs AI instructions to reference TEST_SKIP.

[Copilot]

Pull request overview

Copilot reviewed 9 out of 9 changed files in this pull request and generated 1 comment.

[Copilot]

Pull request overview

Copilot reviewed 9 out of 9 changed files in this pull request and generated 2 comments.