You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
solidcore: a Join whose two rounds both roll wholly from the same face onto the same face beyond fails (StageFailed) #16491
Case: two blends meet at a Join with no edge between them on one side; both lie on face A there. Both have rolled their rail on A over A's smooth far edge onto the same face B, the whole length (the pre-trace rollover of #16149). An example would be the rim of #15767 case 2 with its x = -36 edge split by a vertex.
What happens:
fillet_corner/topology/join.rsrolled_sides only handles a side with exactly one edge between the blends. On a side with no edge between them it leaves the section vertex classified on A.
The join section's ends in fillet_geom/p2.rs then look for a blend whose rails roll on both of the section's faces. Both blends name B, not A, so none matches.
The blend fails with StageFailed ("no blend of the join rolls on the two faces its shared section ends on").
In rolled_sides, also resolve a side with no edge between the blends. If both blends' faces there are the same face A, and both roll their rail on A onto the same face B, the side's face is B.
Mixed sides (one blend rolls, the other doesn't) stay refused.
Add a regression: the arm model of solidcore-ops/tests/tangent_rollover.rs, with the rolled rim edge split into two edges along the same faces. Assert CheckLevel::All and closed-form rails, as the wholly-past tests do.
Written by Claude Code (Anthropic) on behalf of @brett19.
Activity