#!/usr/bin/env python3
"""Emit include/generated/sots_addresses.h for sots-engine from ghidra/addresses.json.
Only facts (name, RVA, convention, prototype, status) — never decompiler text."""
import glob, json, os, subprocess, sys, datetime
root = os.path.join(os.path.dirname(__file__), "..")
j = json.load(open(os.path.join(root, "ghidra", "addresses.json")))
base = int(j["image_base"], 16)

# Per-lane fragments. addresses.json is a single shared file, so concurrent lanes kept
# sweeping each other's in-flight entries into the wrong commit. A lane may instead write
# ghidra/addresses.d/<lane>.json ({"entries": [...]}) — its own file, no shared-line edits.
# Fragments are merged here in sorted order; a duplicate name is an error, not a silent
# last-wins, because two lanes disagreeing about an address is exactly what we must not paper over.
# Two collisions are possible and BOTH are errors:
#   same name, different address -> two lanes disagree about a fact.
#   same address, different name -> two lanes fork the vocabulary. This one is worse, because
#     nothing downstream notices: the header just grows a synonym and later readers cannot tell
#     that `RNG_NextInt` and `RNG_NextIntInclusive` are one function. A lane caught exactly that
#     by hand on the campaign's most-used RNG primitive; the check below is so nobody has to.
def _key(e):
    return ("offset", e["offset"].lower()) if "offset" in e else ("addr", e["addr"].lower())

seen = {e["name"]: "addresses.json" for e in j["entries"]}
# Offsets are only meaningful per owning struct, so they are not globally unique — key addresses only.
by_addr = {_key(e)[1]: (e["name"], "addresses.json") for e in j["entries"] if "addr" in e}
for frag_path in sorted(glob.glob(os.path.join(root, "ghidra", "addresses.d", "*.json"))):
    frag_name = os.path.basename(frag_path)
    for e in json.load(open(frag_path))["entries"]:
        if e["name"] in seen:
            sys.exit(f"duplicate address entry {e['name']!r}: in {seen[e['name']]} and {frag_name}")
        if "addr" in e:
            a = e["addr"].lower()
            if a in by_addr:
                other, where = by_addr[a]
                sys.exit(f"address {e['addr']} named twice: {other!r} in {where}, "
                         f"{e['name']!r} in {frag_name} — pick one name and record the agreement")
            by_addr[a] = (e["name"], frag_name)
        seen[e["name"]] = frag_name
        j["entries"].append(e)
try: rev = subprocess.check_output(["git","-C",root,"rev-parse","--short","HEAD"]).decode().strip()
except Exception: rev = "unknown"
out = [ "// GENERATED — do not edit. Facts about Sword of the Stars.exe (GOG 1.8.1).",
        f"// Source: sots-re ghidra/addresses.json @ {rev}, generated {datetime.date.today()} by tools/gen_addresses.py",
        "// Runtime address = (uintptr_t)GetModuleHandle(NULL) + RVA (the exe is ASLR-relocated).",
        "#pragma once", "#include <cstdint>", "", "namespace sots::addr {", f"constexpr uint32_t IMAGE_BASE = 0x{base:08x};", "" ]
for e in j["entries"]:
    # Two entry shapes: an "addr" (absolute VA -> RVA) or an "offset" (a field offset inside a
    # struct, emitted verbatim). Offsets are facts about layout, not load addresses.
    val = int(e["offset"], 16) if "offset" in e else int(e["addr"], 16) - base
    out.append(f"// {e['convention']:8s} {e['prototype']}  [{e['status']}]")
    out.append(f"constexpr uint32_t {e['name']} = 0x{val:08x};")
out += ["", "} // namespace sots::addr", ""]
dest = sys.argv[1] if len(sys.argv) > 1 else os.path.join(root, "ghidra", "generated", "sots_addresses.h")
os.makedirs(os.path.dirname(dest), exist_ok=True); open(dest, "w").write("\n".join(out)); print("wrote", dest, f"({len(j['entries'])} entries)")