re_match_program: the whole-program LTCG build and check; progress counts its exact matches

harbor/reverse.py

@@ -531,3 +531,38 @@ def run_toolkit(config, key, argv, json_to_file=False, timeout=900):
             return json.loads(text)
         except ValueError:
             raise HarborError(f"{key} failed: " + (result.stdout + result.stderr).strip()[-800:]) from None
+
+
+def match_program(config, targets=None, decomp=None):
+    """Build every readable function into one /GL + /LTCG image the way the game was built (stand-in callers, stubs
+    for the rest, globals at their original addresses) and check the targets against the original bytes. Slow: a
+    build takes 1.5 to 9 minutes depending on load."""
+    import subprocess
+    import sys
+    import tempfile
+    config = config or {}
+    script = config.get("ltcg")
+    if not script or not Path(script).exists():
+        raise HarborError("this project has no ltcg script in its [reverse] table")
+    wanted = [t.strip() for t in re.split(r"[,\s]+", str(targets or "")) if t.strip()]
+    extra = ["--decomp", str(decomp)] if decomp else []
+    run = lambda argv, timeout: subprocess.run([sys.executable, "-B", str(script), *argv], capture_output=True,
+                                               text=True, encoding="utf-8", errors="replace", timeout=timeout,
+                                               cwd=str(Path(script).parent),
+                                               creationflags=getattr(subprocess, "CREATE_NO_WINDOW", 0))
+    built = run(["build", *extra], 1800)
+    if built.returncode != 0:
+        raise HarborError("the whole-program build failed: " + (built.stdout + built.stderr).strip()[-800:])
+    with tempfile.TemporaryDirectory(dir=config.get("debugger_temp")) as tmp:
+        out = Path(tmp) / "check.json"
+        checked = run(["check", *extra, "--json", str(out)] + (["--only", ",".join(wanted)] if wanted else []), 900)
+        if not out.exists():
+            raise HarborError("the whole-program check failed: " + (checked.stdout + checked.stderr).strip()[-800:])
+        rows = json.loads(out.read_text(encoding="utf-8"))
+    counts = {}
+    for r in rows:
+        counts[r.get("status")] = counts.get(r.get("status"), 0) + 1
+    keep = ("address", "name", "status", "first_difference", "similarity", "registers_in", "ours_registers_in",
+            "ret_pops", "ours_ret_pops", "relocation_problems", "note")
+    return {"counts": counts, "results": [{k: r[k] for k in keep if k in r} for r in rows] if wanted else
+            [{k: r[k] for k in keep if k in r} for r in rows if r.get("status") != "differs"][:200]}