mirror/BFM-decomp
1
0
Fork 0
mirror of https://github.com/Druthulu/BFM-decomp synced 2026-09-30 23:37:39 -04:00
Code Issues Packages Projects Releases Wiki Activity

fix(phase-30): extract_unit was blind to the asm-label ALIAS definition form (the 137 "no matched unit")

Browse Source 
The h_seq sweep's 137 "no matched unit for func" skips were ONE exemplar x 137
same-address members, not 137 distinct failures: func_8016191C @ ov_SC01_077,
band=mid, non-jr, 24 ins, ALL 137 members still INCLUDE_ASM stubs => 3,288 ins
left on the floor by a tool lookup miss (R35 again).

ROOT CAUSE (and it refutes the SESSION-27 checkpoint's own diagnosis, R14): the
exemplar is NOT "banked in a different binary". It is banked in ov_SC01_077 under
the SS37/SS73 asm-label alias — the byte-true body conflicts with the fleet-canonical
decl on BOTH SS73 axes (return void vs s32 AND params void*/s32 vs int/unsigned),
so it was banked zero-touch as
    int aF8016191C(int, unsigned int) __asm__("func_8016191C");
extract_unit only ever matched a definition head literally NAMED func_<ADDR>, so
it returned None, and every caller reads None as "not matched".

FIX (tool-only, T0 blast radius):
- _alias_decl_for(): resolve `<ident>(...) __asm__("func_<ADDR>");` -> <ident>.
- extract_unit(): accept the alias identifier as the definition head, re-derived
  PER FILE so one file's alias can never leak into the next.
- carry the alias DECLARATION into the unit (without it the sibling emits the
  symbol aF8016191C and the body never lands at func_<ADDR>); the backscan now
  walks PAST the alias line so the fn's own preceding externs are carried exactly
  as for a plain definition, with a start<=alias_ln<=end guard against double-emit.
- R32: an alias decl with no findable definition now REFUSES LOUDLY instead of
  falling through to _macro_unit and reporting "not matched" — the silent-skip
  class this fix exists to delete.

Rejected the alternative the banking agent suggested (widen engine_core.h's decl
void->s32, drop the alias): it addresses only SS73's RETURN axis while decl and body
also disagree on PARAMS, and it is a T2 fleet-shared edit where the alias is T0.

VERIFIED: unit extracts with externs + exactly one asm label; remap_hseq produces a
sibling draft; rtu_match 3/3 MATCH (24 ins) in the real TU (ov_SC01_000/001/004).
make tools-health RC=0. The whole-binary byte-gate remains the sole arbiter (G3/P9).
This commit is contained in:
Drew T
2026-07-31 07:47:47 -06:00
parent 703b8e6863
commit e99c823cb4
1 changed files with 55 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tools/family_remap.py
+55 -3
View File
@@ -669,13 +669,52 @@ def _foreign_defs(unit_text, addr):
return out
def _alias_decl_for(lines, addr):
"""(alias_ident, decl_line_index) when the function is defined under the §37/§73 ASM-LABEL ALIAS
form, else (None, None).
A body whose byte-true signature conflicts with the fleet-canonical decl (BOTH §73 axes at once —
return AND params) is banked zero-touch by giving the definition a different C identifier and
binding the emitted SYMBOL with a GNU asm label:
extern void func_8016191C(void *a0, s32 a1); <- the fleet canon, in engine_core.h
int aF8016191C(int param_1, unsigned int param_2) __asm__("func_8016191C");
int aF8016191C(int param_1, unsigned int param_2) { ... }
`extract_unit` matched only a definition head literally NAMED func_<ADDR>, so it was structurally
blind to this form and returned None — which every caller reads as "not matched". MEASURED
(Phase 30): that blindness was the WHOLE of the h_seq sweep's 137 "no matched unit" skips — one
exemplar (func_8016191C @ ov_SC01_077) × 137 same-address members, 3,288 ins, every member still
an INCLUDE_ASM stub and otherwise sweep-ready. The tool, not the compiler (R35).
The alias DECLARATION must travel with the unit: without it the sibling TU emits the symbol
`aF8016191C` and the function never lands at func_<ADDR>."""
rx = re.compile(rf'^\s*[A-Za-z_][\w \*]*?\b([A-Za-z_]\w*)\s*\([^;]*\)\s*'
rf'__asm__\s*\(\s*"func_{addr:08X}"\s*\)\s*;', re.I)
for j, ln in enumerate(lines):
m = rx.match(ln)
if m and m.group(1).lower() != f"func_{addr:08x}":
return m.group(1), j
return None, None
def extract_unit(ov, addr):
"""the matched inline def + its contiguous preceding extern/blank/comment lines, from the overlay src.
func_<addr> names are UPPERCASE-hex in src (func_8013DBE4); match case-insensitively to be safe."""
pat = re.compile(rf'^\s*[A-Za-z_][\w \*]*\bfunc_{addr:08X}\s*\(', re.I)
func_<addr> names are UPPERCASE-hex in src (func_8013DBE4); match case-insensitively to be safe.
Also resolves the §37/§73 ASM-LABEL ALIAS definition form (see `_alias_decl_for`): the accepted
definition name becomes the alias identifier, and the alias declaration line is carried into the
unit so the sibling still emits the func_<ADDR> symbol."""
plain = re.compile(rf'^\s*[A-Za-z_][\w \*]*\bfunc_{addr:08X}\s*\(', re.I)
for cf in sorted(glob.glob(f"src/{ov}/{ov}*.c")):
lines = open(cf).read().split("\n")
alias_ident, alias_ln = _alias_decl_for(lines, addr)
# RE-DERIVE per file: `pat` must never leak an earlier file's alias into a later one.
pat = (re.compile(rf'^\s*[A-Za-z_][\w \*]*\b{re.escape(alias_ident)}\s*\(', re.I)
if alias_ident else plain)
for i, ln in enumerate(lines):
if alias_ident and i == alias_ln:
continue # the alias DECLARATION is not the definition
# A DECLARATION ends in `;` — but m2c writes them with a trailing comment
# (`M2C_UNK func_80178D40(s32, s32); /* extern */`), so the raw line ends in `*/` and the
# old `endswith(";")` guard let it through as a DEFINITION. The forward brace-scan then ran
@@ -698,7 +737,7 @@ def extract_unit(ov, addr):
# those belong to the FILE, not to the first function under them. Absorbing one makes
# the unit drag ~140 unrelated externs into every sibling — several naming types the
# sibling's TU lacks — and the whole family gate-fails (Phase 26 session 6).
while j >= 0 and (lines[j].strip() == "" or
while j >= 0 and (lines[j].strip() == "" or j == alias_ln or
lines[j].lstrip().startswith(("extern", "//", "/*", "*", "typedef"))):
if _DECL_LAYER_END.search(lines[j]):
break
@@ -724,6 +763,12 @@ def extract_unit(ov, addr):
end = k
break
unit_text = "\n".join(lines[start:end + 1])
if alias_ident and not (start <= alias_ln <= end):
# The backscan walks PAST the alias line (so the fn's own preceding externs are
# carried, exactly as for a plain definition) and normally sweeps it into
# [start:end]. If some layout kept it out, carry it explicitly and FIRST — the asm
# label is what binds this body to the func_<ADDR> symbol in the sibling TU.
unit_text = lines[alias_ln].strip() + "\n" + unit_text
extra = _foreign_defs(unit_text, addr)
if extra:
# R32: a unit that defines a function OTHER than its target cannot be templated —
@@ -737,6 +782,13 @@ def extract_unit(ov, addr):
if macros:
unit_text = "\n".join(macros) + "\n" + unit_text
return unit_text, cf
if alias_ident:
# R32: we PROVED the function is defined here (the asm label binds the symbol) but could
# not find the alias identifier's definition. Falling through to _macro_unit would report
# "not matched" — the silent skip this whole fix exists to delete. Refuse loudly instead.
print(f"[extract_unit] func_{addr:08X} in {ov}: {cf} declares the asm-label alias "
f"`{alias_ident}` but no definition of it was found — refusing (R32).")
return None, None
return _macro_unit(addr)