fix(gmsh): force preview-quality settings when GMSH fallback path is taken

Fine production settings (0.03mm/0.05rad) with GMSH → CharacteristicLengthMax
0.45mm → OOM kill on large assemblies even with per-solid iteration.

GMSH quality is algorithmic (conforming seams, no fan triangles) — a denser
mesh provides no extra UV-unwrap benefit. Cap GMSH tessellation at preview
settings (0.1mm/0.1rad) in the production fallback path.

Normal path (GMSH + valid geometry GLB) is unaffected — continues to reuse
_geometry.glb directly without re-tessellating.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

backend/app/domains/pipeline/tasks/export_glb.py

@@ -283,20 +283,30 @@ def generate_gltf_production_task(self, cad_file_id: str, product_id: str | None
             "info",
         )
     else:
-        # No usable cache: run tessellation from STEP
+        # No usable cache: run tessellation from STEP.
+        # When GMSH is selected, force preview-quality settings (0.1mm / 0.1rad) even here.
+        # Fine production settings (e.g. 0.03mm) combined with GMSH OOM-kill on large assemblies
+        # because CharacteristicLengthMax becomes too small. GMSH quality is algorithmic
+        # (conforming seams) not density-based — a denser GMSH mesh adds no UV-unwrap benefit.
+        if tessellation_engine == "gmsh":
+            eff_linear = float(sys_settings.get("gltf_preview_linear_deflection", "0.1"))
+            eff_angular = float(sys_settings.get("gltf_preview_angular_deflection", "0.1"))
+        else:
+            eff_linear = prod_linear
+            eff_angular = prod_angular
         occ_cmd = [
             python_bin, str(occ_script),
             "--step_path", str(step_path),
             "--output_path", str(prod_geom_glb),
-            "--linear_deflection", str(prod_linear),
+            "--linear_deflection", str(eff_linear),
-            "--angular_deflection", str(prod_angular),
+            "--angular_deflection", str(eff_angular),
             "--sharp_threshold", str(sharp_threshold),
             "--tessellation_engine", tessellation_engine,
         ]
         log_task_event(
             self.request.id,
-            f"Tessellating STEP at production quality ({tessellation_engine}, "
+            f"Tessellating STEP for production ({tessellation_engine}, "
-            f"linear={prod_linear}mm, angular={prod_angular}rad)",
+            f"linear={eff_linear}mm, angular={eff_angular}rad)",
             "info",
         )
         try: