diskmargin: auto-compute MIMO frequency-grid bounds by default

[baggepinnen]

Motivation

The MIMO diskmargin searched a fixed log-spaced frequency grid l=1e-3 … u=1e3 rad/s (500 points). When the worst-case disk margin lies near or beyond a bound, the reported worst-case frequency lands exactly on the grid edge (e.g. u=1000.0 rad/s) — a grid artifact rather than a true interior minimum. This corrupts derived quantities, in particular the delay margin (phase_margin / ω_crit), where dividing by an artificially-fixed crossover frequency gives a misleading value.

Change

• Default l and u to nothing in diskmargin(L; l, u) and sim_diskmargin(L, σ, l, u).
• When a bound is nothing, derive it from the poles/zeros of L via ControlSystemsBase._bounds_and_features(L, Val{:nyquist}()) — the same heuristic ControlSystems uses to pick frequency ranges for Bode/Nyquist plots, padded around the extreme dynamics with high-frequency headroom and clamped to Nyquist for discrete systems.
• Explicit l/u still override (backward compatible).
• Particle/uncertain systems are already covered by the existing _bounds_and_features specialization in uncertainty_interface.jl.

Only the scalar-l/u path is affected; all AbstractVector-grid callers (e.g. diskmargin(P,C,σ,w)) are unchanged.

Tests

test/test_diskmargin.jl passes locally. The discrete default-grid regression test's reference grid was made dense and Nyquist-relative so the comparison is robust to the now system-derived bounds.

🤖 Generated with Claude Code

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 91.85%. Comparing base (ce8094d) to head (dffd616).

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #151      +/-   ##
==========================================
+ Coverage   91.84%   91.85%   +0.01%     
==========================================
  Files          20       20              
  Lines        3067     3071       +4     
==========================================
+ Hits         2817     2821       +4     
  Misses        250      250              

Flag	Coverage Δ
unittests	91.85% <100.00%> (+0.01%)	⬆️

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

☔ 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.