decoder-bundle-minimization/work/package_all.py

#!/usr/bin/env python3
"""Package every B6 binary variant, run the end-to-end chain proof on the candidates,
and emit REPORT.md + bundle67.json from the measured artifacts."""
from __future__ import annotations

import json
import pathlib
import subprocess

B6 = pathlib.Path("/root/hutter67/swarm67/B6-bundle")
OUT = B6 / "out"
WORK = B6 / "work"

# variant -> (binary path, short description, decode-proof status source)
VARIANTS = {
    "orig_static_p1": (WORK / "paq8px.orig", "P1's shipped static binary, unmodified (baseline payload)"),
    "dyn_Os": (OUT / "build/dyn_Os", "same source, -Os -flto=2, dynamic, gc-sections, stripped"),
    "dyn_Os_hard": (OUT / "build/dyn_Os_hard", "-Os -flto=2 dynamic + hard size flags (no-plt/stack-protector off)"),
    "dyn_Os_flto1": (OUT / "build/dyn_Os_flto1", "-Os -flto=1 dynamic"),
    "dyn_Os_noflto": (OUT / "build/dyn_Os_noflto", "-Os dynamic, no LTO"),
    "dyn_O3": (OUT / "build/dyn_O3", "P1 flags verbatim (-O3 -flto) except dynamic link instead of -static"),
    "dyn_O2": (OUT / "build/dyn_O2", "-O2 -flto=2 dynamic"),
    "dyn_Oz_clang": (OUT / "build/dyn_Oz_clang", "clang++-17 -Oz dynamic"),
    "static_Os": (OUT / "build/static_Os", "-Os static (self-contained: no host libstdc++/libc needed)"),
}
PROVEN_IN_E2E = ["dyn_Os", "dyn_O3", "static_Os"]


def run(cmd, **kw):
    return subprocess.run(cmd, capture_output=True, text=True, **kw)


def package(label: str, binary: pathlib.Path) -> dict | None:
    if not binary.exists():
        return None
    d = OUT / f"bundle_{label}"
    j = OUT / f"bundle_{label}.json"
    r = run(["python3", str(WORK / "make_bundle.py"), "--binary", str(binary), "--out", str(d),
             "--label", label, "--json", str(j)])
    if r.returncode != 0:
        print("package failed", label, r.stdout[-400:], r.stderr[-400:])
        return None
    return json.loads(j.read_text())


def e2e(label: str) -> dict | None:
    r = run(["bash", str(WORK / "e2e_bundle.sh"), label, str(OUT / f"bundle_{label}")])
    p = B6 / "test" / f"e2e_{label}.json"
    if not p.exists():
        print("e2e failed", label, r.stdout[-300:], r.stderr[-300:])
        return None
    return json.loads(p.read_text())


def main() -> int:
    table = {}
    for label, (binary, desc) in VARIANTS.items():
        doc = package(label, binary)
        if doc is None:
            print(f"[skip] {label}: packaging failed or binary missing")
            continue
        doc["description"] = desc
        doc["raw_binary_bytes"] = binary.stat().st_size
        if label in PROVEN_IN_E2E:
            ev = e2e(label)
            doc["e2e"] = ev
            doc["decode_proof"] = bool(ev and ev["all_match"])
        table[label] = doc
        print(f"[ok] {label}: bundle={doc['bundle_bytes']} proven={doc.get('decode_proof')}")
    (OUT / "variant_table.json").write_text(json.dumps(table, indent=1))
    return 0


if __name__ == "__main__":
    raise SystemExit(main())