refactor(p1): one source-facts module owns function ranges, callers, declaration tables and language aliases

src/openultrasast/source.py replaces the copies in semantic/functions.py,
plane/tasks/repo_facts.py, learn/features.py, learn/excerpt.py and
plane/tasks/verify.py. Two engines share one sentinel (<module>), one
DECLARATION/KEYWORDS/RESERVED set and one alias map: the parser engine
(ranges, enclosing, bounds, call_sites, callers_of, facts_for) for the
scorer, the overlay and the hunter's annotations; the line-pattern engine
(declared_functions, enclosing_at, function_span, function_bounds, the brace
matcher) that labels, features, roles, excerpts and plane facts were built
on. learn no longer imports plane.tasks. Frozen records that store the old
sentinel are mapped on read (module_function).

Gate: labels rebuilt before and after are identical, 7,921 rows, after
sentinel normalisation (benchmarks/measurements/2026-10-13-source-module-
label-diff/record.json). 490 targeted tests pass; 3,312 collect. Net: 657
lines deleted, 212 added. benchmarks/unseen keeps <global> for its own
frozen manifests.

Co-Authored-By: Claude Fable 5.1 <[email protected]>
This commit is contained in:
Marius Ciepluch
2026-10-13 12:00:37 +02:00
co-authored by Claude Fable 5.1
parent f63703a604
commit 062a7b780e
36 changed files with 1099 additions and 629 deletions
+8 -8
View File
@@ -85,7 +85,7 @@ def php_code(text: str) -> str:
"""`text` with comments, string contents and inline HTML blanked to spaces, newlines kept (line numbers hold).
Enough to find function bodies by brace matching; heredoc/nowdoc bodies are blanked too. Not a parser: a file
that defeats it yields no function, and the site falls back to `<global>`, which the freeze record shows."""
that defeats it yields no function, and the site falls back to `<module>`, which the freeze record shows."""
out, i, n, php = list(text), 0, len(text), False
def blank(start: int, end: int) -> None:
@@ -148,10 +148,10 @@ def php_functions(text: str) -> list[tuple[str, int, int]]:
def enclosing(spans: list[tuple[str, int, int]], line: int, inserted: bool = False) -> str:
"""The innermost named function containing `line`, or `<global>`. For a pure insertion, `line` is the line the
"""The innermost named function containing `line`, or `<module>`. For a pure insertion, `line` is the line the
new code follows, so a function whose closing brace is on that line does not contain it."""
inside = [(last - first, name) for name, first, last in spans if first <= line <= last and not (inserted and line == last)]
return min(inside)[1] if inside else "<global>"
return min(inside)[1] if inside else "<module>"
def old_hunks(repo: Path, vulnerable: str, fixed: str, path: str) -> list[tuple[int, int, bool]]:
@@ -166,14 +166,14 @@ def old_hunks(repo: Path, vulnerable: str, fixed: str, path: str) -> list[tuple[
def changed_functions(repo: Path, vulnerable: str, fixed: str, path: str) -> list[str]:
"""The functions (or `<global>`) of `path` at the vulnerable pin that the fix's hunks touch, in hunk order."""
"""The functions (or `<module>`) of `path` at the vulnerable pin that the fix's hunks touch, in hunk order."""
text = (blob(repo, vulnerable, path) or b"").decode(errors="replace")
spans, names = php_functions(text), []
for first, last, inserted in old_hunks(repo, vulnerable, fixed, path):
for line in range(first, last + 1):
name = enclosing(spans, line, inserted)
# New code added between functions (a new helper) changes no vulnerable-side code: no site.
if name not in names and not (inserted and name == "<global>"):
if name not in names and not (inserted and name == "<module>"):
names.append(name)
return names
@@ -198,12 +198,12 @@ def resolve_sites(repo: Path, case: dict[str, object], changed: list[str]) -> di
def fix_changes_sites(repo: Path, case: dict[str, object], sites: list[str]) -> dict[str, bool]:
"""Per declared site: does the fix's diff change a line inside that function (or the file, for `<global>`)?"""
"""Per declared site: does the fix's diff change a line inside that function (or the file, for `<module>`)?"""
result = {}
for site in sites:
path, _, name = site.partition("::")
hunks = old_hunks(repo, str(case["vulnerable"]), str(case["fixed"]), path)
if name == "<global>" or not hunks:
if name == "<module>" or not hunks:
result[site] = bool(hunks)
continue
spans = php_functions((blob(repo, str(case["vulnerable"]), path) or b"").decode(errors="replace"))
@@ -289,7 +289,7 @@ def verify(case: dict[str, object], cache: Path) -> dict[str, object]:
# Declared `path::function` sites (population v2) are what the verification reads; the first names the
# sink file, and every function declared in that file must appear in it.
sink_file = sites[0].partition("::")[0]
functions = [site.partition("::")[2] for site in sites if site.partition("::")[0] == sink_file and not site.endswith("::<global>")]
functions = [site.partition("::")[2] for site in sites if site.partition("::")[0] == sink_file and not site.endswith("::<module>")]
else:
sink_file, functions = sink_parts(str(case["sink"]))
record["sink_file"], record["sink_functions"] = sink_file, functions