[llesieur99]

What

Match func_L04_003FA7E0, the shared steering and impact-effect update: 4,516
distinct USA code bytes, placed in Barlow and Joba in both USA v1.01 and v2.00.
PAL has no row with this canonical identity.

The current compiler already produces the normal aligned 0x50-byte template copy.
Restoring the helpers' real float, integer and pointer return declarations closes
the remaining camera-store/call scheduling difference. The return contracts were
checked against consuming callers and deliberate return sequences in retail. The
implementation uses canonical external names and the existing per-file profile;
no compiler patch, register pin, synthetic dependency or ad-hoc assembly is added.

The full investigation covered 16 large functions / 68,664 retail bytes. The
other 15 remain nonmatching; they add no matching or functional credit. Retain
their useful corrections and reopening conditions in the existing source/index:

• Restore missing collision arguments, software-double and other return types,
  64-bit render/style parameters, and the loader's heap-pointer contract.
• Reconstruct the waypoint pair's 64-bit sentinel and the actual reset field.
• Give the collision request its complete 40-byte footprint, sharing its tail with
  the later direction vector as retail does. Keep the real three-quadword matrix
  extents. The unexplained request scalar is not filled with an invented value.
• Remove fourteen unreferenced declarations that only changed stack or pseudo
  numbering. Preserve unsigned alpha packing and the measured camera-route layout.
• Record failed loop/address controls without promoting equal sizes or using fake
  rounding barriers. The final countdown loop in func_L00_003C9978 exists in
  retail; it is preserved, not mistaken for dummy code.

The promoted function replaces its marker and old nonmatching copy. No catalogue,
compiler, shared-header, functional-manifest or progress-report changes are included.
All private disassembly, objects and trial evidence remain outside the repository.

Checked

Validation:

• tools/check.py: the new C matches all four carrying USA placements; PAL identity absent.
• tools/build.py all --strict --version us: 28 programs OK, zero C fallbacks.
• tools/build.py all --strict --version us2: 28 programs OK, zero C fallbacks.
• tools/build.py all --strict --version pal: 28 programs OK, zero C fallbacks.
• python3 -m unittest discover -s tests: 339 tests passed.
• tests/check_repository.py and tools/progress.py --check for all three releases passed.
• All 7,568 qualified production-input hashes agree with the host checkout; LF/diff checks passed.
• No native gameplay test or functional capture replay was run; reports were not regenerated.

All 384 applicable candidate placements were checked across the three releases:
four complete placements match, and the remaining 380 do not. Missing canonical
identities do not imply the absence of related gameplay code in another release.

No native gameplay test or functional capture replay was run. Strict production
builds do not qualify the code retained in nonmatching/.

This PR is ready for maintainer review and is intentionally left open.

[bytematchdev]

ByteMatch progress report

SCES_516.07

Code matched	54.15% · 3.09 MB of 5.70 MB
Fuzzy match	54.15%
Functions	6,996 / 8,672
Data matched	0.00%
Units complete	958 / 1,196

No function changes.

SCUS_972.68_v2

Code matched	81.01% · 4.52 MB of 5.58 MB
Fuzzy match	81.01%
Functions	7,595 / 8,634
Data matched	0.00%
Units complete	973 / 1,156

No function changes.

SCUS_972.68

Code matched	84.74% · 4.68 MB of 5.53 MB
Fuzzy match	84.74%
Functions	7,760 / 8,579
Data matched	0.00%
Units complete	924 / 1,018

No function changes.

_{Commit 364ccfe · Ratchet & Clank: Going Commando on ByteMatch}