Partitioned all 131 Weapon.ID records against the two data routes by which a
weapon is ever named - a field name in an EnumWeapon record, and a Turret_NNN
record's WeaponID value:
mounted not mounted
declared 105 22
not declared 0 4
105 + 22 + 0 + 4 = 131, with no unexplained cell. Both marginals re-derive
unit-substructure-records' own numbers independently: mounted 105, and
22 + 4 = its "26 weapons no turret mounts".
The empty cell is the result: nothing is mounted without being declared, 0 of
131, so declaration is the superset of mounting. The four undeclared and
unmounted records are therefore reachable by neither route. Each appears
exactly once per pak copy (x6) and only as a Weapon record's ID value; the live
control Weapon_TCAF_DeltaSaber_Beam appears x18 as an EnumWeapon field name and
x228 as a Turret_NNN.WeaponID. That places Weapon_TCAF_DeltaSaber_Laser_Ttrl,
the weapon this pass set out to find, and unifies all four as orphan datasheet
rows.
Recorded as powerless rather than as evidence: the image holds 0 exact strings
for the four, but also 0 for the live control and 0 for Weapon_NULL. Only three
Weapon_-prefixed strings exist at all and 0 contain Ttrl - the executable names
no weapon, so the negative does not discriminate.
New: sweeping EnumWeapon RECORDS rather than the three declared tables gives 130
distinct field names, not 127. The extra three come from a fourth record, pak
key 305595319, resolved as EnumWeapon_TestS01.tbl, carrying Weapon_TestS01_Gun/
_Laser/_Missile. It dangles at both ends - no stage declares it, and its three
weapons are the only declared names with no Weapon.ID record. 130 - 3 = 127
restores the previous pass's union.
All fifteen artefacts byte-identical.
074c0296af