v0.6.0

[HH-MWB]

Description

Pre-release fixes for v0.6.0: correct callback order in Timer.__enter__, make the CPU-bound block scenario stable, and simplify the BDD step layout (remove utils.py, centralize shared steps).

Summary of Change

1. Library: invoke on_start before recording timestamps (timerun.py)

• In Timer.__enter__, on_start(measurement) is now invoked before recording perf_counter_ns() and process_time_ns().
• Ensures the measured interval does not include callback work and reflects the intended start of timing.

2. Tests: fix flaky CPU-bound scenario

• Scenario: Renamed from "yields wall and CPU time close together" to "yields meaningful wall and CPU time."
• Assertion: Removed the "CPU time is close to wall time" step that failed under load when the process was scheduled out (wall > CPU by more than the buffer). The scenario now only asserts wall time ≥ 10,000,000 ns and CPU time within the configured buffer of 10,000,000 ns.
• Step: New step "the measurement's wall time duration is at least X nanoseconds" in block_timing_steps.py for this case; other scenarios still use the shared "within buffer" step from common_steps.py.
• CPU-bound step: The "measure operation using with" step now uses a process-time spin (process_time_ns()) for CPU-bound operations instead of the old wall-time spin from utils.py, so the scenario actually exercises CPU time.

3. Test structure and docs

• Removed features/steps/utils.py: Inlined or moved helpers into common_steps.py. BUFFER_NS and the "within buffer" / "which measurement" logic live in common_steps.py; sleep/spin in step files use time.sleep / time.process_time_ns() directly.
• common_steps.py: Clarified as "steps shared by more than one feature." Added _measurement(context, which) and parameterized steps so one implementation supports "the measurement's…" and "the first/second/outer/inner measurement's…" (wall time within buffer, metadata key). Minor comment and structure cleanup; CPU lower slack inlined as 1_000_000 ns.
• block_timing_steps.py: Uses common_steps for shared steps; only block-specific steps and the new "at least" wall step remain here. Sleep/spin via time only.
• function_timing_steps.py: Replaced sleep_wall_at_least with time.sleep(...); no dependency on utils.py. Light comment/formatting cleanup.
• measurement_steps.py: Empty-metadata assertion simplified to assert context.measurement.metadata == {}.
• CONTRIBUTING.md: Prerequisites list now includes pip and Make; clone URL uses YOUR_USERNAME; feature layout description updated (no utils.py; shared steps in common_steps.py).
• docs/recipes/analyze-results.md: Wording updated to "wrapped callable's measurements" for clarity.

Checklist

• I have run make lint test and both pass
• I have updated the documentation accordingly (if applicable)

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 100.00%. Comparing base (cb6c5c9) to head (e08ef17).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff            @@
##              main        #7   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files            1         1           
  Lines          107       107           
=========================================
  Hits           107       107           

Flag	Coverage Δ
python3.10	100.00% <100.00%> (ø)
python3.11	100.00% <100.00%> (ø)
python3.12	100.00% <100.00%> (ø)
python3.13	100.00% <100.00%> (ø)
python3.14	100.00% <100.00%> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.