Rename FallBlockBfs to daObjKm2_Fall_Block_c and consolidate it into one TU

[andrewboudreau]

Independent of the #2130 chain — this branches off main.

FallBlockBfs, Bowser in the Fire Sea's falling block, moves to the spelling the
cartridge's own RTTI carries, and its five per-function objects collapse into one genuine
translation unit under the snake_case scheme in notes/tu-naming-and-swallowers.md sec 1
(tools/tu_names.py).

coined name	cartridge name	new TU	module / .text
FallBlockBfs	daObjKm2_Fall_Block_c	src/actors/d_a_obj_km2_fall_block.cpp	ov045 0x02111d48..0x02111e60

Why the rename is load-bearing, not cosmetic

A coined class name reaches no symbol home. Its _ZTI/_ZTS/_ZTV records are emitted
under a mangling the config carries nowhere, so romdata_check can never word-compare
them against the cartridge and the TU cannot declare a verifiable compiler_only_output
(tools/tubuild.py apply_compiler_only_policy). Renamed, this TU is its class's
key-function TU and claims the whole RTTI group as retail data.

Measured, not assumed, in extracted/overlays/overlay_0045.bin (ov045 base 0x021111a0,
the .text start: on line 1 of config/arm9/overlays/ov045/delinks.txt): the vtable
object's preamble at 0x021130ec is [offset-to-top 0, 0x021130a0], and
_ZTI21daObjKm2_Fall_Block_c at 0x021130a0 reads [0x0209a764, 0x021130b8, 0x02114874]
— _ZTVN3abi20__si_class_type_infoE, a typeinfo name at 0x021130b8 whose bytes are the
string "21daObjKm2_Fall_Block_c", and _ZTI16daObjFallBlock_c, so the cartridge states
the direct base too. That base pointer lands in ov015, and coo_rows independently
resolved _ZTI16daObjFallBlock_c to ov015 0x02114874 — the same address, from the other
direction.

The trap this branch deliberately reverses

The destructor moves inline into the header: written out-of-line, mwccarm emits D0
before D1 — the reverse of the cartridge's order, which objisolate refuses for the
whole TU — plus a third D2 with no ROM home.

That only works while some other virtual keeps the TU as the key-function TU. Both slot
methods override daObjFallBlock_c's pure virtual s32 InitResources() = 0 /
CleanupResources() = 0 pair, so they were already virtual; they now carry the base's
s32 return type and are declared plainly, and CleanupResources — declared first — is
the key function.

notes/plan-cpp-conversion-queue.md sec 5 held this exact pair up as the exemplar of
declaring vtable-slot methods non-virtual on purpose, so no one-function file became the
key-function TU and emitted _ZTV. That was right while the class was spread over five
files. With one TU there is exactly one place for the group to land, and the class stays
unverifiable until it does. That section is rewritten to say so rather than left
contradicting the tree.

The bodies are the recovered ones, not the generator's shadow

tu_promote's starting point spelled all five bodies as extern-C free functions over
void *self, and its factory stored the leaf vtable twice. The ROM stores the base's
vptr and then the leaf's — the base sub-object is constructed first. The bodies here are
the recovered C++ ones from the deleted per-function sources. The factory's second store
is now &_ZTV21daObjKm2_Fall_Block_c[2], because this TU defines that vtable and mwccarm's
symbol addresses the object start rather than the slot array.

Verification

• tools/rombuild.py -j16 — 106/106 exact, 100.000000% of compared bytes, mismatching 0
• ROM data from source: 570 — up one. _ZTV21daObjKm2_Fall_Block_c is a record no
  source proved before.
• romdata_check per-symbol, on a raw (non-isolated) compile:
  _ZTI21daObjKm2_Fall_Block_c VERIFIED 12 bytes, _ZTS21daObjKm2_Fall_Block_c
  VERIFIED 24 bytes, _ZTV21daObjKm2_Fall_Block_c PARTIAL, 128 of a 140-byte
  extent.
  The PARTIAL is an extent artifact, not a mismatch, and I read the cartridge to be sure:
  the table is 32 slots, 0x021130f4..0x02113174, and all 32 words are emitted and match.
  The next symbol is data_ov045_02113180, in .bss, so the config's extent runs 12 bytes
  past the table into the zero padding ahead of it; slots 32–34 read 0x00000000 in the
  ROM. Pre-existing config, not introduced here.
• 13 RTTI symbols licensed as deadstrip-data, every canonical address checked against
  symbols.txt.
• tiers_ratchet.py --check PASS 2591/2591 · eligible 11096/11211 ·
  check_src_tu_compiles 98/98 · check_header_offsets spans 0x34c ·
  check_dead_references, layout_check, port_refcheck clean
• No config_tu/ tree exists for ov045, so the dangling-shadow defect an agent review
  found on the ov019 promotion cannot reach this branch. Checked, not assumed.

The rename commit is separate, and was reviewed for the tool's blind spots

545ba599d is the rename alone. tools/class_rename.py --with-derived did 22 edits and 8
renames; four of its known blind spots were present and are corrected in that commit —
including tools/test_tubuild.py, which looked the manifest entry up by the id the rename
changed and would have failed the suite closed, and the manifest's linkerOutput, a
verbatim captured mwldarm transcript the tool rewrote into output the linker never
produced. Blind spot (d) — ROM-true names asserted at addresses symbols.txt contradicts
— is absent: all 7 changed symbols/actor_renames.tsv rows were cross-checked and
every address agrees.

🤖 Generated with Claude Code

https://claude.ai/code/session_01FsXWgEM3Gtbqs6zEdwLMYV

[tangos-validator]

✅ PR validation — Passed

Committed merge passes; 5 contributor credit reassignment(s) noted, not a blocker.

Full merge validation

Check	Result
Committed test merge	yes
Byte-verified functions	10,704 / 11,347 (94.33%, -4)
Byte-verified code bytes	1,993,536 / 2,211,124 (90.16%, -200)
Claimed, not byte-verified	513 functions, 112,144 bytes (+4)
Perfect source moves	0 R100
Enrolled ranges (delinks complete)	11,059 functions, 2,053,148 bytes (92.86%, +0) -- differs from byte-verified by +355
Contributor credit	0 added, 5 changed, 0 lost
Relocation check	23 checked; 23 VERIFIED
Port reference check	405 checked; 0 stale
Module fidelity	106/106 exact; 100.000000% compared bytes
Code linked from verified source	11,088 functions, 2,067,148 bytes (93.49%)
Module bytes from source	2,067,148 / 3,049,600 (67.8%); 811,492 (26.6%) are data no delink entry reaches
ROM data reproduced from source	576 symbol(s) exact, 164 partial, 5 differ

Contributor credit moved (5)

Function	Source	Before	After
ov045:0x02111d48	src/_ZN12FallBlockBfsD1Ev.cpp -> src/actors/d_a_obj_km2_fall_block.cpp	tangosdev	github-actions[bot]
ov045:0x02111d98	src/_ZN12FallBlockBfsD0Ev.cpp -> src/actors/d_a_obj_km2_fall_block.cpp	tangosdev	github-actions[bot]
ov045:0x02111dfc	src/_ZN12FallBlockBfs16CleanupResourcesEv.cpp -> src/actors/d_a_obj_km2_fall_block.cpp	lunavyqo	github-actions[bot]
ov045:0x02111e10	src/_ZN12FallBlockBfs13InitResourcesEv.cpp -> src/actors/d_a_obj_km2_fall_block.cpp	lunavyqo	github-actions[bot]
ov045:0x02111e24	src/FallBlockBfs_Spawn.c -> src/actors/d_a_obj_km2_fall_block.cpp	tangosdev	github-actions[bot]

Byte-verified means the range carries complete in a delinks.txt, so the ROM build compiled it and compared it to the cartridge. The 513 claimed functions have a src/ file named after the symbol with no NONMATCHING banner, and nothing compiles them -- dsd fills their addresses with the ROM's own bytes. Both together are the 11,217 this project calls matched.

Warnings: contributor attribution changed, not a blocker (5 changed, 0 lost -- src/actors/d_a_obj_km2_fall_block.cpp: tangosdev -> github-actions[bot]; src/actors/d_a_obj_km2_fall_block.cpp: tangosdev -> github-actions[bot]; src/actors/d_a_obj_km2_fall_block.cpp: lunavyqo -> github-actions[bot]; +2 more); 5 address range(s) left the byte-verified set while enrolled totals held steady: ov045:0x02111d48-0x02111d98, ov045:0x02111d98-0x02111dfc, ov045:0x02111dfc-0x02111e10, ov045:0x02111e10-0x02111e24, ov045:0x02111e24-0x02111e60; 4 more function(s) now claim a match that nothing compiles; enroll them in a delinks.txt to have the ROM build check them.

Per-file link-check detail

All 16 changed file(s) compile to the ROM byte-for-byte with correct relocation targets.

File	Symbol	Result	Slots checked
src/_ZN11FallBlockWf13InitResourcesEv.cpp	_ZN11FallBlockWf13InitResourcesEv	✅ verified	1
src/_ZN11FallBlockWf16CleanupResourcesEv.cpp	_ZN11FallBlockWf16CleanupResourcesEv	✅ verified	1
src/_ZN11FallBlockWfD0Ev.cpp	_ZN11FallBlockWfD0Ev	✅ verified	1
src/_ZN11FallBlockWfD1Ev.cpp	_ZN11FallBlockWfD1Ev	✅ verified	1
src/_ZN16daObjFallBlock_c15OnHitByMegaCharER6Player.cpp	_ZN16daObjFallBlock_c15OnHitByMegaCharER6Player	✅ verified	1
src/_ZN16daObjFallBlock_c4KillEv.cpp	_ZN16daObjFallBlock_c4KillEv	✅ verified	1
src/_ZN16daObjFallBlock_c6RenderEv.cpp	_ZN16daObjFallBlock_c6RenderEv	✅ verified	1
src/_ZN16daObjFallBlock_c8BehaviorEv.cpp	_ZN16daObjFallBlock_c8BehaviorEv	✅ verified	1
src/_ZN16daObjFallBlock_cD0Ev.cpp	_ZN16daObjFallBlock_cD0Ev	✅ verified	1
src/_ZN16daObjFallBlock_cD1Ev.cpp	_ZN16daObjFallBlock_cD1Ev	✅ verified	1
src/actors/FallBlockBbh/_ZN12FallBlockBbh13InitResourcesEv.cpp	_ZN12FallBlockBbh13InitResourcesEv	✅ verified	1
src/actors/FallBlockBbh/_ZN12FallBlockBbh16CleanupResourcesEv.cpp	_ZN12FallBlockBbh16CleanupResourcesEv	✅ verified	1
src/actors/FallBlockBbh/_ZN12FallBlockBbhD0Ev.cpp	_ZN12FallBlockBbhD0Ev	✅ verified	1
src/actors/FallBlockBbh/_ZN12FallBlockBbhD1Ev.cpp	_ZN12FallBlockBbhD1Ev	✅ verified	1
src/actors/d_a_obj_km2_fall_block.cpp	_ZN21daObjKm2_Fall_Block_cD1Ev + _ZN21daObjKm2_Fall_Block_cD0Ev + _ZN21daObjKm2_Fall_Block_c16CleanupResourcesEv + _ZN21daObjKm2_Fall_Block_c13InitResourcesEv + daObjKm2_Fall_Block_c_Spawn	✅ verified	5
src/actors/daObjFl_Fall_Block_c.cpp	_ZN20daObjFl_Fall_Block_cD1Ev + _ZN20daObjFl_Fall_Block_cD0Ev + _ZN20daObjFl_Fall_Block_c16CleanupResourcesEv + _ZN20daObjFl_Fall_Block_c13InitResourcesEv	✅ verified	4

_{The private worker commits a test merge, builds the stock ROM profile, compares every executable module, measures matched and source-built code, checks contributor lineage, and verifies affected relocations. The mod profile is opt-in and is not part of this merge gate.}