feat(examples): add ikine_XX timing columns to ik_speed.py

[petercorke]

Summary

Extends examples/ik_speed.py from a single ik_XX-only benchmark to three timing columns per method:

• ik_XX — the existing fast, C-only solver (unchanged)
• ikine_XX, C++ ETS — the pure-Python ikine_XX solver, with fkine/jacob0 evaluated via the normal C++-accelerated ETS
• ikine_XX, pure Python ETS — the same ikine_XX solver, but with the ETS forced onto its pure-Python fallback (both the solver loop and every fkine/jacob0 call interpreted) -- what a pure-Python wheel / Pyodide build actually experiences

The pure-Python column uses 1,000 samples instead of 10,000 -- it's orders of magnitude slower per solve (confirmed by a small dry run: LM Sugihara alone projects to ~28 minutes at 1,000 samples with this script's existing k values, which I left unchanged).

Forcing the pure-Python ETS path required figuring out the right lever: robot.ets() returns a cached/memoized instance, so a fresh robot.ets() call doesn't help. Instead, roboticstoolbox.ets.fknm._C_AVAILABLE is patched to False and the existing ets's _fknm_stale flag is set, forcing its lazily-built C++ handle to rebuild as None on next access -- confirmed directly that ets._fknm becomes None and ikine_LM still works correctly against it.

Test plan

• Dry run at reduced scale (200 / 20 samples) -- all three columns populate correctly, numbers are sensible (ik_XX fastest, ikine_XX/C++ in between, ikine_XX/pure-Python orders of magnitude slower)
• Confirmed ets._fknm is None after the patch (via direct repr check) and reverts cleanly afterward
• py_compile clean
• Full 1,000-sample run not executed in this session (would take ~35-40 min) -- left for a real benchmark pass since the dry run already validates correctness

🤖 Generated with Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 0.00%. Comparing base (07a60f4) to head (f153011).
⚠️ Report is 3 commits behind head on main.

Additional details and impacted files

@@          Coverage Diff          @@
##            main    #654   +/-   ##
=====================================
  Coverage   0.00%   0.00%           
=====================================
  Files        143     143           
  Lines      14027   14050   +23     
=====================================
- Misses     14027   14050   +23     

☔ 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.