fix: lock CPU power history swap

[davidberenstein1957]

Fixes a lost-update race on CPU._power_history.

What changed

CPU._power_history is appended to by the 1 Hz monitor scheduler thread (CPU.monitor_power) and drained by the measure_power_secs scheduler thread (CPU.total_power) — two distinct PeriodicScheduler timer threads (codecarbon/emissions_tracker.py:345-352). The read-then-rebind in total_power was not atomic, so any sample appended between the list comprehension and self._power_history = [] went into the discarded list.

• Added a threading.Lock to CPU, held only for the O(1) list swap. _get_power_from_cpus() is called outside the lock in both paths, so a slow backend (e.g. the IntelPowerGadget subprocess) never blocks the monitor thread.
• Removed the unreachable if not power_history_in_W: branch — a sample is unconditionally appended before the average, so the list is never empty.

Behaviour is otherwise unchanged, including the extra synchronous sample taken at measurement time.

Why

In cpu_load mode the reported cpu_power is the mean of the buffered samples. The samples dropped in the swap window are exactly those taken while a slow measurement is in flight, so the resulting error is small but systematic and silent.

How it was verified

New test tests/test_cpu_load.py::TestCPULoad::test_cpu_total_power_keeps_samples_added_while_draining makes the race deterministic (a list subclass that fires monitor_power() at iteration exhaustion, i.e. inside the lost-update window) and asserts the injected sample survives. It fails on master and passes with this change. uv run pytest tests/test_cpu_load.py -q → 9 passed.

Note: uv run task format reformats a large number of unrelated files on the current tree, so only the two touched files are included here.

Closes#1315

🤖 Generated with Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 91.48%. Comparing base (3ec31a0) to head (f72f432).

Additional details and impacted files

@@ Coverage Diff @@## master #1335 +/- ##
==========================================
+ Coverage 91.43% 91.48% +0.04% 
==========================================
Files 49 49 Lines 5057 5060 +3 ==========================================
+ Hits 4624 4629 +5 + Misses 433 431 -2 

☔ View full report in Codecov by Harness.
📢 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.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.