Issues

• OffsetSurface upper-half derivatives overwrite named slots and use uninitialized mixed locals (WR-162 family)

  bug
  tangency-labs/solidcore#2717 · Filter by author altvnk opened on Nov 4, 2026

• A cloned Brep writes no surfaces: Surface::clone drops the owner, and the writer keeps only face-owned surfaces

  tangency-labs/solidcore#2715 · Filter by author aapolipponen opened on Nov 4, 2026
  1

• blend: a variable round whose B-spline radius law is a single point fails with SM_ERR_NULL_POINTER in the solver

  tangency-labs/solidcore#2706 · Filter by author bherbruck opened on Nov 4, 2026

• Fillet: with only fillet_gn_section_generator, a variable box round from 4 to 0 fails at its radius-4 end

  tangency-labs/solidcore#2704 · Filter by author bherbruck opened on Nov 4, 2026

• Fillet: a variable round 1e-9 wide at its end builds a near-singular surface with no fixes

  tangency-labs/solidcore#2703 · Filter by author bherbruck opened on Nov 4, 2026

• Twisted-disc top round leaves vertex gaps up to 5.6e-5 (tolerance 2e-5) at 1.10

  tangency-labs/solidcore#2701 · Filter by author brett19 opened on Nov 4, 2026

• Twisted-disc top round at 1.9586 runs 877 s and fails at crossing_rails with a bare 1001

  tangency-labs/solidcore#2700 · Filter by author brett19 opened on Nov 4, 2026

• Twisted-disc side rounds miss 2 mm by up to 1.8e-3 between traced sections

  tangency-labs/solidcore#2699 · Filter by author brett19 opened on Nov 4, 2026
  1

• Help wanted: G2 (curvature-continuous) setback corner fill, certified convex

  enhancementhelp wantedpriority: high
  tangency-labs/solidcore#2697 · Filter by author jyviko opened on Nov 4, 2026

• Faces with many inner loops: three O(n²) costs in building and checking them

  tangency-labs/solidcore#2696 · Filter by author bherbruck opened on Nov 4, 2026

• blend: build two rounds meeting at a knife-edge tip by cutting each against the other along a ridge

  tangency-labs/solidcore#2695 · Filter by author bherbruck opened on Nov 4, 2026

• Fillet: the G^n section certificate refines one 3-edge fillet into 1,234 spans (29 s)

  tangency-labs/solidcore#2690 · Filter by author bherbruck opened on Nov 4, 2026

• A plain cylinder read from STEP fails Level 2 in 6 places

  bug
  tangency-labs/solidcore#2684 · Filter by author jyviko opened on Nov 4, 2026

• Control-net monotonic check: in closed mode the last step is never compared with the closing step

  bug
  tangency-labs/solidcore#2683 · Filter by author jyviko opened on Nov 4, 2026

• UnStitch checks how many faces moved, not which ones (#2676 follow-up)

  bug
  tangency-labs/solidcore#2682 · Filter by author jyviko opened on Nov 4, 2026

• Curve property: the minimum-radius search finds a boundary minimum from one end of the domain but not the other

  tangency-labs/solidcore#2670 · Filter by author MarkusReichart opened on Nov 4, 2026
  1

• Curve property: the minimum-radius objective has no answer at a zero-speed point, though its limit is in closed form

  tangency-labs/solidcore#2669 · Filter by author MarkusReichart opened on Nov 4, 2026
  1

• A rounded box's top rim takes ~10x longer than the C++ reference (second fillet of a tangent chain, r 3)

  tangency-labs/solidcore#2667 · Filter by author kaltenbrunnermarkus opened on Nov 4, 2026
  1

• Solidcore-only blend fix: a blend run on past a smooth vertex that leaves the part is not capped (fillet_blend_corner, WR-697 addendum)

  tangency-labs/solidcore#2664 · Filter by author brett19 opened on Nov 4, 2026

• Blend End: a half-selected symmetric rim removes 2.57e-4 mm3 more than half its round (sphere_plane)

  tangency-labs/solidcore#2660 · Filter by author MarkusReichart opened on Nov 4, 2026

• Regressions SolidCraft sees on solidcore 5d2635d36: #2364, #2480, #2369, #2382

  bug
  tangency-labs/solidcore#2656 · Filter by author jyviko opened on Nov 4, 2026
  2

• solidcore-convert: bound the nesting depth of curves and surfaces it converts, and return an error past it

  tangency-labs/solidcore#2655 · Filter by author jyviko opened on Nov 4, 2026
  1

• Fillet: a rail tangent to a face's edge fails the round with 1001 at simple_coincidence_checker (nearest) (#2648's twisted disc)

  tangency-labs/solidcore#2652 · Filter by author brett19 opened on Nov 4, 2026
  4

• Automated first-pass merge: skip stacked PRs, open reviews and PRs marked as waiting

  tangency-labs/solidcore#2651 · Filter by author jyviko opened on Nov 4, 2026

• perf: band_cuts_the_part rebuilds every face's surface cache for every section of every rounded edge: a 2 mm round of a twisted part's top outline runs 13.5 min, then fails with 1001

  tangency-labs/solidcore#2648 · Filter by author brett19 opened on Nov 4, 2026
  3