Turn on solidcore's new fixes at c9a06dc7c

Two new UpstreamFixes flags, neither shown broken. On the rig on
solidcore every case keeps its status and no case reaches C++. Six
volumes move, all because of fillet_surface_plane_curve_on_plane_equation
(each goes back when that flag alone is turned off), by at most 3.8e-8
relative. Where the case has a closed form, the shift moves it closer:
the 851/1113 arc chamfers go from 1.7e-4..3.4e-4 off to 1.4e-5..2.2e-5
off (#1113's full wall height from 3.3e-4 to 1.9e-5), and the pipes
touching at a point go from 3.7e-4 to 3.1e-4 off.
fillet_extended_rail_end_at_side_edge_crossing changes no case on the
rig. The rig's total time stays the same (974 s with the flags off,
977 s with them on). HELD_BACK stays empty.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SdcgEu4Lr3moCCisdZShKz

crates/smlib/src/core.rs

@@ -363,6 +363,8 @@ pub(crate) fn fixes() -> UpstreamFixes {
     f.is_topological_vertex_rejects_cusps_and_tangent_errors = true;
     f.cone_create_canonical_preserves_inside_out_and_checks_height = true;
     f.point_class_gap_3d_propagates_no_drop_error = true;
+    f.fillet_surface_plane_curve_on_plane_equation = true;
+    f.fillet_extended_rail_end_at_side_edge_crossing = true;
     f
 }