nx_gatequality.nx

nx_gatequality.nx -- HOW MUCH OF AN ORGAN ITS GATE ACTUALLY EXERCISES. ===== WHY THIS EXISTS ============================================ nx_gatescan answers "does anything test this file". That question has a blind spot large enough to have hidden a real defect in this tree: a gate touching four of an organ's twelve functions counts as a gate, so the organ reads as covered. That is exactly what happened in the food-chemistry lane. A neighbouring gate imported the ingredient catalog and exercised si_name, si_density_mg_per_ml, si_capsule_fill_mg and si_cost_is_indicative -- the packaging and cost helpers. si_class and si_is_lawful_dietary, the two functions deciding whether a substance may LAWFULLY BE SOLD, were exercised by nothing at all, and the scan reported the file as gated. ★COVERAGE ACCUMULATES AROUND WHATEVER A NEIGHBOUR HAPPENED TO NEED, so the most consequential function in a file is frequently the least checked. Existence is the wrong question. This organ asks the other one. ===== THE MEASUREMENT ============================================ 1. From each organ, extract its top-level `func` definitions. 2. Find the gates attributed to it (same rule as nx_gatescan: the organ stem followed by an underscore, ending _test.nx or _gate.nx). 3. In each gate, look for a real reference to each function name. 4. Coverage = referenced / defined, unioned across all of that organ's gates -- an organ tested by three files is covered by their union. ★★COMMENTS ARE STRIPPED BEFORE MATCHING, AND THAT IS NOT A DETAIL. Gates in this tree carry long explanatory headers that NAME the functions they discuss. Matching raw text would count a function as exercised because its own gate mentioned it in prose -- and the worse the coverage, the more likely someone wrote a comment about it. The measurement would have been anti-correlated with the truth. ★MATCHING IS TOKEN-EXACT. A substring match would count si_class as covered by a reference to si_class_of, which is how a coverage tool starts reporting numbers that are reassuring and false.