re: read the defaulted weapon stats out of live guest memory

Canary backs the whole guest address space with one shared-memory file, so the
game's RAM is readable from the host while it runs -- no debugger, no emulator
patch. That turns the parked Route-B question (fields the IDXD omits because
they sit at their default) from "unreachable" into a table lookup.

gmem.py maps guest VAs into /dev/shm/xenia_memory_* via Xenia's fixed table and
searches only the allocated extents, so a full-RAM scan is ~0.2 s.

weapon_runtime.py finds the parsed objects by scanning for their vtable, then
*solves* the struct layout instead of guessing it: every (field, offset,
encoding) triple is scored against the disc records, and a binding is accepted
only with zero contradictions -- and marked confirmed only when >=10 records
agree on >=3 distinct values, because a field whose samples are all one number
matches any offset holding that constant.

Result: Weapon (0xc0) and Shell (0x200) mapped, 4393 values the disc does not
carry, for all 126 weapons at 5 % save progress. Cross-checks hold -- the
Arsenal DATA SHEET read 4 for wep_05/wep_60 Max. Lock Ons and the memory read
agrees; the in-flight HUD's 06000/00300 match LoadingCount; all 252 objects are
byte-identical across a screen change, so this is definition data, not state.

Two findings fell out: `IsCharging = Yes` switches LoadingCount and
TriggerShotCount to float32 (it partitions the 8 float-encoded records exactly),
and two .tbl entries declare Shell_TCAF_Ship_AAGun with conflicting Power -- the
runtime keeps the one that omits it.

Where the defaulted values *come from* (the IDXD's undecoded binary node region
vs code defaults) is not settled here; they vary per weapon, so they are not one
constructor constant.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

tools/re-capture/gmem.py

@@ -0,0 +1,162 @@
+#!/usr/bin/env python3
+"""Read the *live* guest memory of a running Xenia Canary.
+
+Canary backs the whole guest address space with a single shared-memory file,
+`/dev/shm/xenia_memory_<pid-ish>`, so the guest's RAM is readable from the host
+with no debugger, no emulator patch and no pause: open the file, seek, read.
+
+The file is one flat image of Xenia's *physical* backing store; guest virtual
+addresses map into it through Xenia's fixed table (memory.cc `map_info`).
+`va_to_off()` implements that table, so callers work in guest VAs.
+
+Sub-commands
+    find  <pattern>        search every allocated extent, print guest VAs
+    read  <va> [len]       hexdump guest memory
+    words <va> [n]         dump n big-endian u32 / f32 pairs
+
+`<pattern>` is a python literal-ish string: plain text, or `hex:0011aabb`.
+Numbers accept 0x form. The scan uses SEEK_DATA so the ~4.6 GB of sparse holes
+cost nothing.
+"""
+
+import os
+import re
+import sys
+import struct
+
+# (guest_va_lo, guest_va_hi_inclusive, file_offset_of_lo) — Xenia memory.cc.
+MAP = [
+    (0x00000000, 0x3FFFFFFF, 0x00000000),
+    (0x40000000, 0x7EFFFFFF, 0x40000000),
+    (0x7F000000, 0x7F0FFFFF, 0x00000000),
+    (0x7F100000, 0x7FFFFFFF, 0x00100000),
+    (0x80000000, 0x8FFFFFFF, 0x80000000),
+    (0x90000000, 0x9FFFFFFF, 0x80000000),
+    (0xA0000000, 0xBFFFFFFF, 0x100000000),
+    (0xC0000000, 0xDFFFFFFF, 0x100000000),
+    (0xE0000000, 0xFFFFFFFF, 0x100000000),
+]
+
+
+def va_to_off(va):
+    for lo, hi, base in MAP:
+        if lo <= va <= hi:
+            return base + (va - lo)
+    raise ValueError(f"va {va:#x} outside the guest map")
+
+
+def off_to_vas(off):
+    """All guest VAs that alias this file offset (the map is many-to-one)."""
+    out = []
+    for lo, hi, base in MAP:
+        if base <= off <= base + (hi - lo):
+            out.append(lo + (off - base))
+    return out
+
+
+def primary_va(off):
+    """The most useful VA for an offset: physical 0xA0000000+ / xex 0x80000000+."""
+    vas = off_to_vas(off)
+    return vas[0] if vas else None
+
+
+def mem_path():
+    if len(sys.argv) > 1 and sys.argv[1].startswith("/dev/shm/"):
+        return sys.argv.pop(1)
+    cands = [f"/dev/shm/{n}" for n in os.listdir("/dev/shm") if n.startswith("xenia_memory_")]
+    if not cands:
+        sys.exit("no /dev/shm/xenia_memory_* — is Canary running?")
+    if len(cands) > 1:
+        sys.exit(f"several memory files, pass one explicitly: {cands}")
+    return cands[0]
+
+
+def extents(fd, size):
+    """Yield (start, end) of the file's allocated (non-hole) ranges."""
+    pos = 0
+    while pos < size:
+        try:
+            data = os.lseek(fd, pos, os.SEEK_DATA)
+        except OSError:
+            return
+        try:
+            hole = os.lseek(fd, data, os.SEEK_HOLE)
+        except OSError:
+            hole = size
+        yield (data, hole)
+        pos = hole
+
+
+def parse_pattern(s):
+    if s.startswith("hex:"):
+        return bytes.fromhex(s[4:])
+    return s.encode("latin-1")
+
+
+def cmd_find(f, size, args):
+    pat = parse_pattern(args[0])
+    limit = int(args[1]) if len(args) > 1 else 64
+    hits = 0
+    CHUNK = 1 << 24
+    for start, end in extents(f.fileno(), size):
+        pos = start
+        carry = b""
+        carry_at = start
+        while pos < end:
+            f.seek(pos)
+            buf = f.read(min(CHUNK, end - pos))
+            if not buf:
+                break
+            blob = carry + buf
+            base = carry_at
+            for m in re.finditer(re.escape(pat), blob):
+                off = base + m.start()
+                va = primary_va(off)
+                print(f"{off:#013x}  va {va:#010x}" if va is not None else f"{off:#013x}  va ?")
+                hits += 1
+                if hits >= limit:
+                    return
+            keep = len(pat) - 1
+            carry = blob[-keep:] if keep else b""
+            carry_at = base + len(blob) - len(carry)
+            pos += len(buf)
+    if hits == 0:
+        print("(no hits)", file=sys.stderr)
+
+
+def cmd_read(f, size, args):
+    va = int(args[0], 0)
+    n = int(args[1], 0) if len(args) > 1 else 256
+    f.seek(va_to_off(va))
+    data = f.read(n)
+    for i in range(0, len(data), 16):
+        row = data[i : i + 16]
+        hexs = " ".join(f"{b:02x}" for b in row)
+        txt = "".join(chr(b) if 0x20 <= b < 0x7F else "." for b in row)
+        print(f"{va + i:08x}  {hexs:<47}  |{txt}|")
+
+
+def cmd_words(f, size, args):
+    va = int(args[0], 0)
+    n = int(args[1], 0) if len(args) > 1 else 32
+    f.seek(va_to_off(va))
+    data = f.read(n * 4)
+    for i in range(0, len(data) - 3, 4):
+        (u,) = struct.unpack_from(">I", data, i)
+        (fl,) = struct.unpack_from(">f", data, i)
+        fs = f"{fl:.6g}" if -1e30 < fl < 1e30 else ""
+        print(f"{va + i:08x} +{i:04x}  {u:#010x}  {u:>12}  {fs}")
+
+
+def main():
+    path = mem_path()
+    if len(sys.argv) < 2:
+        sys.exit(__doc__)
+    cmd, args = sys.argv[1], sys.argv[2:]
+    size = os.path.getsize(path)
+    with open(path, "rb") as f:
+        {"find": cmd_find, "read": cmd_read, "words": cmd_words}[cmd](f, size, args)
+
+
+if __name__ == "__main__":
+    main()