stabilize geometric inverse cdf and clean lint warnings

[day01]

Summary

• Stabilize Geometric::inverse_cdf with a closed-form implementation to avoid platform-dependent bisection drift (Windows/no_std failure).
• Remove unused testing_boiler imports that caused -Dwarnings failures in coverage/nightly.

Testing

• cargo fmt -- --check
• cargo clippy --all-targets (remaining warnings only in kernel tests: uninlined_format_args, unchanged)
• (to run in CI) cargo test --no-default-features -F rand --target x86_64-pc-windows-gnu or windows-latest and cargo +nightly llvm-cov --no-report nextest

[codecov]

Codecov Report

❌ Patch coverage is 90.90909% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 94.40%. Comparing base (5da3470) to head (a09585c).
⚠️ Report is 10 commits behind head on master.

Files with missing lines	Patch %	Lines
src/distribution/geometric.rs	90.90%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #364      +/-   ##
==========================================
- Coverage   94.99%   94.40%   -0.60%     
==========================================
  Files          61       59       -2     
  Lines       13615    12963     -652     
==========================================
- Hits        12934    12238     -696     
- Misses        681      725      +44     

☔ 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.

[day01]

@YeungOnion added tests and some lint fixes

[YeungOnion]

Partially merged via cherry-pick rebase onto master (736e412, 72d6630) to preserve author attribution for the testing import cleanup and CI fixes. The inverse_cdf stabilization (bad330b) was not used — the distribution was instead addressed with a different implementation (514a9b9) with explicit panic bounds and num_traits compatibility. All tests pass.

[YeungOnion]

I sat on #343 so long that your changes functionally overlapped. I added in some other changes to more clearly denote the failure modes, #366 identifies some need for how we handle failure modes which has some approach in #377 I kept the safe comparisons to 1.0 let me know if I missed another theme.