G-1 records the standing lesson for this document: an interface document should contain no statement a reader will take as fact when it is not yet one. Three remained in it. Section 2 chain diagram read DECIDES against CT 101, which decides nothing and sets no header. Section 5 routing table listed /m/stl as members. Section 5 stated flatly that export is gated today. Section 8 said the opposite of all three, further down, where an outside implementer would reach it second.
None of the three is now removed or softened. The scheme stays and each says which it is - where the line falls once enforced, not where it falls now.
README: deploy/ added to the layout, absent since 12 SEP. The reading order no longer contradicts PROCESS section 8 and HANDOFF section 0, and points at DIVERGENCES.md. The AGPL section 13 obligation is marked not met, stated in the document that states the obligation.
ACCEPTANCE section 7 records that its correction finally landed in HANDOFF section 7, and why parking a correction in a second document is a bad pattern: it sat here for three weeks while the wrong numbers stayed where people read first.
Applied by anchored patcher. Suite 643 passed, oracle intact. Documentation only.
Reference implementation of the strap-beam generators at revision 8.0.0, kept
so the acceptance oracle can be regenerated. Not a live target; the running
application has no OpenSCAD dependency.
The oracle holds 123 frozen cases, 113 accepted and 10 rejected, produced by
OpenSCAD 2021.01 with BOSL2 at 92d697c2. The ten rejections are part of the
contract: a port that accepts them is wrong.
tests/test_oracle.py specifies the port API and was written before the port,
so the interface follows from what must be verified rather than what is
convenient to implement. Proven by adversarial stub: a build() that rejects
everything passes all 10 rejection tests and fails all 226 acceptance tests.
TheRON
civicus-build