Fixed the stop root-finding problem for transmissive stop surfaces

[jacobdparker]

Third PR in the stop-solver stack — stacked on #169 (the diff shown here is relative to that branch).

Problem

The reversed trace in _calc_rayfunction_stops propagates the solved stop rays back to the object by re-applying every surface from the stop backward. That only works when the stop is a mirror: reflection is an involution, so applying the stop surface to its own forward output turns the rays around exactly. A transmissive stop that modifies the rays — a transmission grating or a Fresnel zone plate — is not an involution: its diffraction was applied a second time instead of undone, corrupting the computed field extents by an order of magnitude (~20× for the FZP tutorial, #153).

(For the record: a true time-reversed trace through transmissive rulings would require negating the diffraction order, while reflective rulings are time-reversal symmetric with the same order, because the sign(direction · normal) factor in incident_effective flips along with the reflection. This PR sidesteps reversed traces through transmissive elements entirely instead.)

Changes

• New _stop_is_involutory predicate (mirrors and no-op vacuum surfaces qualify; surfaces with rulings or refraction do not).
• For non-involutory stops, the stop's own diffraction/refraction is included inside the root-finding problem (the solved variable becomes the pre-stop ray), and the stop's material and rulings are stripped on the reversed leg so only its geometry participates.
• _ray_error now takes explicit propagators / transformation_last arguments so callers control which surfaces are inside the solve.

Tests

Adds TestSequentialSystemFZP: a telescope whose only element is a transmissive Fresnel zone plate (HolographicRulingSpacing) acting as the pupil stop, asserting the recovered field half-angle equals arctan(sensor half-width / focal length) to 1e-3 deg.

🤖 Generated with Claude Code

[codecov]

Codecov Report

❌ Patch coverage is 96.66667% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 99.37%. Comparing base (40c7ea5) to head (df9358d).

Files with missing lines	Patch %	Lines
optika/systems/_sequential.py	94.11%	1 Missing ⚠️

Additional details and impacted files

@@                  Coverage Diff                   @@
##           fix/object-field-stop     #170   +/-   ##
======================================================
  Coverage                  99.36%   99.37%           
======================================================
  Files                        116      116           
  Lines                       6024     6053   +29     
======================================================
+ Hits                        5986     6015   +29     
  Misses                        38       38           

Flag	Coverage Δ
unittests	99.37% <96.66%> (+<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.