REGRESSION - Track the LayerWidths length scales in diiid_slayer_n1

Adds delta_s_abs, delta_norm and delta_visco to the SLAYER case. The two
algebraic scales are built from rs/S/P_perp alone, so they pin the layer
normalization independently of the del_s Riccati solve -- a drift in either
localizes the cause to the parameter build rather than the ODE.

The case previously tracked the SLAYER inputs and the extracted roots but
nothing from the LayerWidths group, so the new struct fields had no coverage.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PSrf6JCViFfVzqzkQ66o6b

Author: logan-nc

regression-harness/cases/diiid_slayer_n1.toml

@@ -132,6 +132,33 @@ label = "SLAYER no_root flags [2/1,3/1,4/1]"
 noise_threshold = 0
 order = 34
 
+# Resistive layer widths. delta_s_abs is the Riccati layer thickness; delta_norm and
+# delta_visco are algebraic scales derived from rs/S/P_perp, so they pin the layer
+# normalization independently of the ODE solve.
+[quantities.slayer_delta_s_abs]
+h5path = "Tearing/LayerWidths/delta_s_abs"
+type = "real_vector"
+extract = "all_real"
+label = "SLAYER |delta_s| [m]"
+noise_threshold = 1e-12
+order = 40
+
+[quantities.slayer_delta_norm]
+h5path = "Tearing/LayerWidths/delta_norm"
+type = "real_vector"
+extract = "all_real"
+label = "SLAYER delta_norm [m]"
+noise_threshold = 1e-12
+order = 41
+
+[quantities.slayer_delta_visco]
+h5path = "Tearing/LayerWidths/delta_visco"
+type = "real_vector"
+extract = "all_real"
+label = "SLAYER delta_visco [m]"
+noise_threshold = 1e-12
+order = 42
+
 # Settings (catches accidental config drift)
 [quantities.slayer_enabled]
 h5path = "Tearing/enabled"