Add src/data/hwid_keys.rs — the 34-row HWID key/edition table and the 5-row
alternate-edition fallback table, transcribed verbatim from HWID_Activation.cmd
(:hwiddata / :hwidfallback, git history f34d025). Generated programmatically
from the source rows (split on the batch's collapsed `_` delimiters, `%f%`
splitter stripped) so no key was hand-typed; a test shape-checks every product
key (5x5 groups) and asserts the row counts.
Wire it into hwid.rs via resolve_key(sku, edition): prefer the working primary
key, else the alternate-edition fallback (the script's :dk_checksku selection
for the common single-branch case; multi-branch SKUs like EnterpriseS=125 are
noted for version refinement).
53 tests pass (36 mas + 17 libtsforge), clippy clean, offline.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Re-architect the Microsoft Activation Scripts from ~46k lines of Windows
batch into an idiomatic Rust crate (MAS/). Not a line-by-line translation:
typed enums + Result flow replace goto/errorlevel/global-set, and a single
`Spp` trait is the only seam between portable logic and Windows internals.
- Portable core (compiles + 20 tests pass, zero deps): model (real SPP
GUIDs / WMI class names / LicenseStatus codes), typed errors, CLI menu,
Check_Activation_Status read path, Online KMS orchestration with host
fallback, stub backend.
- Windows backend (cfg(windows) + `winapi` feature): WMI/COM via the
`windows` crate, all unsafe/VARIANT handling encapsulated behind `Spp`.
Written from the documented SPP contract; not compiled on the Linux port
host.
- HWID/KMS38/Ohook/TSforge scaffolded against analyzed specs (in module
docs) — unported paths return a typed Unsupported error rather than
falsely reporting activation. Roadmap in MAS/README.md.
The original .cmd scripts remain recoverable in history (f34d025, MAS 3.12).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
RustyProgrammerCmd