Turns the hand-maintained 5.0.7D mech specification spreadsheet into something
generated from Content/, so the numbers come from the data the game loads
rather than being transcribed. 65 stock chassis, 55 columns, every non-weapon
column populated.
Standard library only. An .xlsx is a zip of XML and only sheet1.xml is
rewritten, so the template's formatting and drawings survive untouched.
Two findings worth recording, both in the README:
The Armor block's per-zone values in .subsystems are TONS, not multipliers.
MechLab.cpp GetCurrentMechArmorData does tonnage = points / points_per, so
installed points = tons x points-per-ton, with the rates in
Subsystems/Armor.data (Standard 32, Ferro 38, Reflective/Reactive 30,
Solarian 60). Battlemaster = 13.0 x 32 = 416 points. Cross-checked against
MechLab.cpp:796 armor_bar = (TotalArmor/535)*100: 50 of 65 chassis land within
5 points of their stored ArmorRating. An earlier pass summed MaxArmorValue from
.damage instead, which gives armour CAPACITY (1024 for the Battlemaster), not
what is fitted.
The Gladiator mounts six Clan Medium Pulse Lasers in its right arm, so the
template's three arm slots were silently dropping three weapons. Maximum usage
was measured per location across all 65 chassis before widening anything: only
the right arm overflowed. Right_Arm_4/5/6 inserted, trailing columns shifted
+3. The generator now warns by name if any location overflows again.
Judgement calls, all documented:
- columns 10-15 are "installed in the stock loadout", read from .subsystems,
not the CanLoad* flags in .data. The template sample marks the Battlemaster
n for ECM/AMS/LAMS/Beagle even though it can load all four.
- column 7 follows the data (Standard) over the sample (Ferro-Fibrous).
- rating bars are pulled from .instance rather than recomputed. MechLab
recalculates them live while editing, so stored and computed can drift;
15 chassis diverge by more than 5 points, Fafnir worst at 38.
verify-mech-specs.py runs five checks. The load-bearing one calibrates the
generated Battlemaster row against the template's hand-made sample: those 11
weapon cells are an independent statement of the loadout, and they match
exactly, which is the evidence that .subsystems is being read correctly.
Co-authored-by: Claude Opus 5 (Anthropic) <noreply@anthropic.com>
Co-authored-by: GitHub Copilot <copilot@github.com>
653265c928