fix(ranker): scope a region to the file that defines its handler (M1b 12.4)

A route-registration module registers handlers that other modules define.
app/routes/index.js requires seven handler modules and defines none, yet regions
there were scoped to handleContributionsUpdate and its siblings. The graph rightly
held no such method, so every question in those regions was unresolvable: 138
across six families on that one repository.

The entry record still names the handler. The authorization question is which
middleware guards which handler, and that is cross-file by nature. It now also
records whether the file defines what it names, and a region built from an entry
that does not is scoped to the file instead of to a method the file lacks. The
handler's own module still gets its function-scoped region, so nothing is lost.

My first attempt changed the mapper to stop naming cross-file handlers at all. An
existing control caught it by asserting that a multi-line registration names the
handler its middleware guards, which is exactly the attribution the authorization
family needs. The evidence was against me and the fix belonged one layer up, in
region construction.

1,407 tests passed, nine skipped.

Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
Claude-Session: https://claude.ai/code/session_01EmM1j46qgKMh3zqD1QMwEo
This commit is contained in:
Marius Ciepluch
2026-09-17 14:04:39 +02:00
co-authored by Claude Opus 5
parent ca2c3a8b5f
commit 573b764bbb
4 changed files with 57 additions and 1 deletions
+8
View File
@@ -432,6 +432,14 @@ Preserve all completed task evidence below.
- _Requirements: 2.1, 2.2, 6.3, 8.1_ - _Requirements: 2.1, 2.2, 6.3, 8.1_
- _Verified 2026-09-17:_ 1,406 tests passed, nine skipped; Ruff/format/mypy passed. The cause is recorded rather than guessed: `app/routes/index.js` requires seven handler modules and defines none of them, so a question naming `handleContributionsUpdate` there names a function that file genuinely does not hold and the empty scope is correct. Taint and configuration now separate the two causes, reporting `context_function_unresolved` with the name when a question names a function the file does not define, and `context_scope_empty` only when the file has no methods at all. The driver no longer adds its generic empty-scope gap on top of a query that already explained itself, and the new gap is question-owned so it cannot block another question's comparison. Completion is unchanged by design: a question whose function the graph does not hold stays unresolved, and nothing completes by widening its scope to the whole file. - _Verified 2026-09-17:_ 1,406 tests passed, nine skipped; Ruff/format/mypy passed. The cause is recorded rather than guessed: `app/routes/index.js` requires seven handler modules and defines none of them, so a question naming `handleContributionsUpdate` there names a function that file genuinely does not hold and the empty scope is correct. Taint and configuration now separate the two causes, reporting `context_function_unresolved` with the name when a question names a function the file does not define, and `context_scope_empty` only when the file has no methods at all. The driver no longer adds its generic empty-scope gap on top of a query that already explained itself, and the new gap is question-owned so it cannot block another question's comparison. Completion is unchanged by design: a question whose function the graph does not hold stays unresolved, and nothing completes by widening its scope to the whole file.
- [x] 12.4 Attribute a region to the file that defines its handler
- **The region-attribution half of 12.3.** A route-registration module registers handlers other modules define. `app/routes/index.js` requires seven handler modules and defines none, yet regions there were scoped to `handleContributionsUpdate` and its siblings, so the graph rightly held no such method and every question in those regions was unresolvable: 138 across six families.
- The entry record still names the handler, because the authorization question is which middleware guards which handler and that is cross-file by nature. It now also records whether the file defines what it names, and a region built from an entry that does not is scoped to the file rather than to a method the file lacks.
- _Verified 2026-09-17:_ 1,407 tests passed, nine skipped; Ruff/format/mypy passed. A first attempt changed the mapper to stop naming cross-file handlers and was reverted: an existing control proved the name carries the authorization attribution, so the fix belongs in region construction, not in the mapper. Controls assert the registering file asks about itself with no function-scoped region, while the defining module keeps its function-scoped region.
- _Boundary: Ranker Scope Contract, contributor-scan entry mapping_
- _Depends: 12.3_
- _Requirements: 2.1, 2.2, 8.1_
## Implementation Notes ## Implementation Notes
- Task5.2 measured PMPro timing out after both snapshots were materialized but before change context/graph analysis. Group6 must account for compatible snapshot/discovery reuse as well as graph reuse; identical-tip hits cannot substitute for the later representative changed-code latency gate. No new detector, scope heuristic or incremental CPG mutation is implied. - Task5.2 measured PMPro timing out after both snapshots were materialized but before change context/graph analysis. Group6 must account for compatible snapshot/discovery reuse as well as graph reuse; identical-tip hits cannot substitute for the later representative changed-code latency gate. No new detector, scope heuristic or incremental CPG mutation is implied.
+14 -1
View File
@@ -74,6 +74,15 @@ class EntryPointRecord:
conditions: list[str] conditions: list[str]
provenance: str provenance: str
rationale: str rationale: str
# Does the file this record belongs to actually DEFINE the handler it names?
#
# A route-registration module registers handlers that live elsewhere: NodeGoat's `index.js`
# requires seven handler modules and defines none. The record still names the handler, because
# the authorization question is about which middleware guards which handler and that is
# inherently cross-file. But a REGION scoped to a function its file does not hold is a question
# the graph cannot answer, and 138 questions across six families were unresolvable for exactly
# that reason on one repository. The region builder uses this to scope such an entry to the file.
defines_handler: bool = True
# Did the access level come from a DECLARATION, or from the absence of one? # Did the access level come from a DECLARATION, or from the absence of one?
# #
# The distinction decides whether "public" can be trusted. An OpenAPI operation with no `security` block # The distinction decides whether "public" can be trusted. An OpenAPI operation with no `security` block
@@ -424,18 +433,20 @@ def _js_entry_points(target: FileTarget, text: str) -> list[EntryPointRecord]:
handler, middleware = _js_registration(stripped, bounds) handler, middleware = _js_registration(stripped, bounds)
access, evidence = _js_route_access(middleware, stripped) access, evidence = _js_route_access(middleware, stripped)
start, end = bounds.get(handler or "", (number, number)) start, end = bounds.get(handler or "", (number, number))
named = handler or _js_handler_name(stripped)
records.append( records.append(
_entry( _entry(
target, target,
start, start,
end, end,
handler or _js_handler_name(stripped), named,
"http_handler", "http_handler",
"route", "route",
access, access,
"http_request", "http_request",
evidence, evidence,
_line_conditions(stripped), _line_conditions(stripped),
defines_handler=bool(named) and named in bounds,
) )
) )
if "process.argv" in stripped: if "process.argv" in stripped:
@@ -762,6 +773,7 @@ def _entry(
access_evidence: list[str], access_evidence: list[str],
conditions: list[str], conditions: list[str],
access_declared: bool = False, access_declared: bool = False,
defines_handler: bool = True,
) -> EntryPointRecord: ) -> EntryPointRecord:
return EntryPointRecord( return EntryPointRecord(
path=target.path, path=target.path,
@@ -773,6 +785,7 @@ def _entry(
access_level=access_level, access_level=access_level,
trust_boundary=trust_boundary, trust_boundary=trust_boundary,
access_evidence=access_evidence, access_evidence=access_evidence,
defines_handler=defines_handler,
conditions=conditions, conditions=conditions,
provenance=f"entrypoint:{kind}:{access_level}", provenance=f"entrypoint:{kind}:{access_level}",
rationale=f"{kind} surface classified as {access_level} at {trust_boundary}", rationale=f"{kind} surface classified as {access_level} at {trust_boundary}",
+7
View File
@@ -99,6 +99,13 @@ def regions_for(entries: Sequence[object], targets: Sequence[object], *, shipped
if not families: if not families:
continue continue
function = getattr(entry, "function_name", None) or None function = getattr(entry, "function_name", None) or None
if function is not None and not getattr(entry, "defines_handler", True):
# The entry names a handler this file registers but does not define, which is what a
# route-registration module does. Scoping a region to it asks the graph about a method
# the file does not hold, and every question in that region is then unresolvable. The
# handler's own module gets its own regions, so the route is scoped to the file here and
# the record keeps the name for the authorization question, which is cross-file by nature.
function = None
module_body = function in _MODULE_MARKERS module_body = function in _MODULE_MARKERS
if module_body: if module_body:
# A module body is not a handler, whatever the mapper called it, so it is offered the families a # A module body is not a handler, whatever the mapper called it, so it is offered the families a
+28
View File
@@ -272,3 +272,31 @@ def test_an_inferred_public_endpoint_keeps_the_obligation() -> None:
handler = next(r for r in regions if r.function == "handler") handler = next(r for r in regions if r.function == "handler")
assert "access_control" in handler.families assert "access_control" in handler.families
def test_a_route_registered_but_not_defined_here_is_scoped_to_its_file():
"""12.4: a registration module names handlers other modules define; the graph holds none of them."""
from types import SimpleNamespace
from openultrasast.model.regions import regions_for
target = SimpleNamespace(path="app/routes/index.js", language="javascript")
registered = SimpleNamespace(
path="app/routes/index.js",
function_name="handleContributionsUpdate",
access_level="public",
defines_handler=False,
)
defined = SimpleNamespace(
path="app/routes/contributions.js",
function_name="handleContributionsUpdate",
access_level="public",
defines_handler=True,
)
other = SimpleNamespace(path="app/routes/contributions.js", language="javascript")
scoped = {(region.path, region.function) for region in regions_for([registered, defined], [target, other])}
# The registering file asks about itself, not about a method it does not hold.
assert ("app/routes/index.js", None) in scoped
assert not any(path == "app/routes/index.js" and function for path, function in scoped)
# The module that defines it still gets the function-scoped region.
assert ("app/routes/contributions.js", "handleContributionsUpdate") in scoped