Skip to content

fix(dart): why the tri-comparison args shape was so large — 3 residual causes (#2309) #2342

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
squid-protocol merged 1 commit into from
Aug 27, 2026
Merged

fix(dart): why the tri-comparison args shape was so large — 3 residual causes (#2309) #2342

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
48 changes: 34 additions & 14 deletions gitgalaxy/core/detector.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3512,20 +3512,30 @@ def _dart_scan_terminator(
continue
elif term_kind == "semi":
end_idx = term_idx + 1 # Bug 2: bodyless constructor
# #2309: count the real parameter list directly instead of
# relying on the `args` regex here -- that regex must never
# accept a bare `;` terminator (it's `.search()`ed over the
# whole block for dart, so accepting `;` would just as
# readily match a real call statement inside some OTHER
# zero-paren declaration's body; see language_standards.py's
# own comment on this same issue for the confirmed false
# positive). `_count_top_level_args` on the signature text
# already known-good (func_start + `_find_balanced_end`
# already validated real, balanced parens to get here) has
# no such ambiguity.
args_count_override = self._count_top_level_args(
safe_code[start_idx:params_end_idx], treat_as_body=False
)
if not has_parens:
# A bodyless, paren-less `;`-terminated declaration -- an abstract
# getter (`Offset get pan;`), `external Type name;` -- has no
# parameter list. `_count_top_level_args` on the bare name text
# ("Offset get pan") finds no `(` to unwrap and returns 1 (the whole
# string as one segment), a phantom argument against tree-sitter's
# correct 0 (the bulk of the small `dart/function/args/agree[none]`
# mismatches -- Flutter's `PointerEvent` getters). #2309-followup.
args_count_override = 0
else:
# #2309: count the real parameter list directly instead of
# relying on the `args` regex here -- that regex must never
# accept a bare `;` terminator (it's `.search()`ed over the
# whole block for dart, so accepting `;` would just as
# readily match a real call statement inside some OTHER
# zero-paren declaration's body; see language_standards.py's
# own comment on this same issue for the confirmed false
# positive). `_count_top_level_args` on the signature text
# already known-good (func_start + `_find_balanced_end`
# already validated real, balanced parens to get here) has
# no such ambiguity.
args_count_override = self._count_top_level_args(
safe_code[start_idx:params_end_idx], treat_as_body=False
)
elif term_kind == "brace":
end_idx = self._find_balanced_end(safe_code, term_idx, opener, closer)
args_sig_end = term_idx + 1
Expand Down Expand Up @@ -5462,6 +5472,16 @@ def _calculate_block_metrics(
# Since func_start already validated this is a function, we can reliably
# fallback to the structural counter.
args_count = self._count_top_level_args(args_search_text)
elif args_search_text is not None and "(" in args_search_text and self.primary_lang_id == "dart":
# #2309-followup: dart's `args` regex can't match an operator overload
# (`bool operator ==(Object other)`, `Widget operator [](int i)`) -- the
# `==`/`[]` sits between the name token and the `(`, so `.search()`
# returns None and the count silently falls to 0 against tree-sitter's
# correct 1. func_start already validated this is a real function; count
# its balanced parameter span structurally. Guarded on a literal `(` so
# a paren-less getter body (`Rect get bounds {`, whose args_search_text
# has none) still correctly reads 0, not 1.
args_count = self._count_top_level_args(args_search_text)
except Exception as e:
self.logger.debug(f"Argument-count regex extraction failed, leaving args_count 0: {e}")

Expand Down
107 changes: 74 additions & 33 deletions tests/extraction/languages/test_dart.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -51,7 +51,7 @@
"Map<String, dynamic> fakeClass = {};",
"void classLikeFunction() {",
"String a = 'class Foo {';",
"print(\"class Foo {\");",
'print("class Foo {");',
"// class Foo {",
"/* class Foo { */",
"/// class Foo {",
Expand All @@ -61,14 +61,17 @@
],
}


@pytest.mark.parametrize("payload,expected", CLASS_START_CASES["valid"])
def test_dart_class_start_valid(payload, expected):
assert_valid_match(DART_RULES["class_start"], payload, expected, "dart.class_start")


@pytest.mark.parametrize("payload", CLASS_START_CASES["invalid"])
def test_dart_class_start_invalid(payload):
assert_invalid_no_match(DART_RULES["class_start"], payload, "dart.class_start")


@pytest.mark.parametrize("payload", CLASS_START_CASES["xfail_invalid"])
@pytest.mark.xfail(reason="String/comment lookalikes lack AST block shielding", strict=True)
def test_dart_class_start_xfail_invalid(payload):
Expand Down Expand Up @@ -107,7 +110,7 @@ def test_dart_class_start_xfail_invalid(payload):
"var x = functionStart;",
"typedef IntFunc = int Function();",
"var myFunc = () {};",
"String s = \"Future<int> foo() { \";",
'String s = "Future<int> foo() { ";',
"// void main() {",
# Issue #1417: Return-type shield (should not wander across `) {` and match body calls)
" T? result,\n ) {\n Navigator.of(context).popUntilWithResult<T>(predicate, result);",
Expand Down Expand Up @@ -150,30 +153,38 @@ def test_dart_class_start_xfail_invalid(payload):
],
}


@pytest.mark.parametrize("payload,expected", FUNC_START_CASES["valid"])
def test_dart_func_start_valid(payload, expected):
assert_valid_match(DART_RULES["func_start"], payload, expected, "dart.func_start")


@pytest.mark.parametrize("payload", FUNC_START_CASES["invalid"])
def test_dart_func_start_invalid(payload):
assert_invalid_no_match(DART_RULES["func_start"], payload, "dart.func_start")


@pytest.mark.parametrize("payload", FUNC_START_CASES["xfail_invalid"])
@pytest.mark.xfail(reason="String/comment lookalikes lack AST block shielding", strict=True)
def test_dart_func_start_xfail_invalid(payload):
assert_invalid_no_match(DART_RULES["func_start"], payload, "dart.func_start")


def test_dart_slicer_bug_2_and_4():
from gitgalaxy.core.detector import StructuralExtractor
from gitgalaxy.standards.language_standards import LANGUAGE_DEFINITIONS

splicer = StructuralExtractor("dart", LANGUAGE_DEFINITIONS)

# Bug 2: colon-initializer, no-braced-body constructor
bug2_code = "LabeledGlobalKey(this._debugLabel) : super.constructor();\nvoid nextMethod() {}"
blocks, _ = splicer._slice_by_braces(bug2_code, "dart", DART_RULES, 0, {})
assert len(blocks) == 2, "Bug 2: Should find exactly 2 functions (the constructor and nextMethod)"
assert blocks[0]["name"] == "LabeledGlobalKey"
assert bug2_code[blocks[0]["start_idx"]:blocks[0]["end_idx"]].strip() == "LabeledGlobalKey(this._debugLabel) : super.constructor();"
assert (
bug2_code[blocks[0]["start_idx"] : blocks[0]["end_idx"]].strip()
== "LabeledGlobalKey(this._debugLabel) : super.constructor();"
)
assert blocks[1]["name"] == "nextMethod"

# Bug 4: multi-line bare call-site invocation used as a list-literal element
Expand Down Expand Up @@ -210,13 +221,7 @@ def test_dart_slicer_multi_initializer_constructor():
assert "Foo.raw" in names, "multi-initializer bodyless constructor must still be found"
assert "nextMethod" in names

bodied_multi_init = (
"class Foo {\n"
" Foo(this.a) : b = a, assert(a != null) {\n"
" print(a);\n"
" }\n"
"}\n"
)
bodied_multi_init = "class Foo {\n Foo(this.a) : b = a, assert(a != null) {\n print(a);\n }\n}\n"
blocks, _ = splicer._slice_by_braces(bodied_multi_init, "dart", DART_RULES, 0, {})
assert [b["name"] for b in blocks] == ["Foo"], "multi-initializer constructor WITH a body must still be found"

Expand Down Expand Up @@ -244,6 +249,38 @@ def test_dart_slicer_long_parameter_list():
assert "normalMethod" in names


def test_dart_slicer_args_getters_and_operators_2309():
"""
#2309-followup: two remaining `dart/function/args/agree[none]` causes on GitGalaxy's
side -- a paren-less getter (`Offset get pan;`) was counted as 1 arg (the bare name
text, with no `(` to unwrap, collapses to a single segment), and an operator overload
(`bool operator ==(Object other)`) was counted as 0 because the `args` regex can't
match a name token separated from its `(` by `==`/`[]`.
"""
from gitgalaxy.core.detector import StructuralExtractor
from gitgalaxy.standards.language_standards import LANGUAGE_DEFINITIONS

splicer = StructuralExtractor("dart", LANGUAGE_DEFINITIONS)
code = (
"class Vec {\n"
" Offset get pan;\n" # abstract getter -> 0 args
" double get scale => _scale;\n" # arrow getter -> 0 args
" bool operator ==(Object other) {\n return true;\n }\n" # -> 1 arg
" Vec operator +(Vec other) => _add(other);\n" # -> 1 arg
" Widget operator [](int index) {\n return _at(index);\n }\n" # -> 1 arg
" void move(double dx, double dy) {}\n" # regression guard -> 2 args
"}\n"
)
blocks, _ = splicer._slice_by_braces(code, "dart", DART_RULES, 0, {})
args_by_name = {b["name"]: b.get("args") for b in blocks}
assert args_by_name.get("pan") == 0, args_by_name
assert args_by_name.get("scale") == 0, args_by_name
assert args_by_name.get("operator==") == 1, args_by_name
assert args_by_name.get("operator+") == 1, args_by_name
assert args_by_name.get("operator[]") == 1, args_by_name
assert args_by_name.get("move") == 2, args_by_name # normal method unaffected


def test_dart_slicer_bug_5_closure_invocation():
# Regression case for GitHub issue #1624: closure-arguments in calls were wrongly identified as functions.
from gitgalaxy.core.detector import StructuralExtractor
Expand All @@ -252,15 +289,7 @@ def test_dart_slicer_bug_5_closure_invocation():
splicer = StructuralExtractor("dart", LANGUAGE_DEFINITIONS)

# empty-args closure invocation shape
code1 = (
"class Foo {\n"
" void nextMethod() {\n"
" setState(() {\n"
" x = 1;\n"
" });\n"
" }\n"
"}\n"
)
code1 = "class Foo {\n void nextMethod() {\n setState(() {\n x = 1;\n });\n }\n}\n"
blocks, _ = splicer._slice_by_braces(code1, "dart", DART_RULES, 0, {})
names = [b["name"] for b in blocks]
assert "setState" not in names, "Bug 5: Should exclude empty-args closure invocation"
Expand All @@ -283,13 +312,7 @@ def test_dart_slicer_bug_5_closure_invocation():
assert "run" in names

# No regression: named-parameter constructor must still be found
code3 = (
"class Baz {\n"
" const Baz.raw({required this.a}) : b = a;\n"
" void foo(int a, {int x = 1}) {\n"
" }\n"
"}\n"
)
code3 = "class Baz {\n const Baz.raw({required this.a}) : b = a;\n void foo(int a, {int x = 1}) {\n }\n}\n"
blocks, _ = splicer._slice_by_braces(code3, "dart", DART_RULES, 0, {})
names = [b["name"] for b in blocks]
assert "Baz.raw" in names, "No regression: named-parameter bodyless constructor must still be found"
Expand All @@ -307,12 +330,24 @@ def test_dart_slicer_bug_5_closure_invocation():
("TargetFunc({required int a, String b = 'default'}) {", "({required int a, String b = 'default'})"),
("TargetFunc([int a = 1, int b = 2]) {", "([int a = 1, int b = 2])"),
("TargetFunc([List<int> x = const [1, 2, 3]]) {", "([List<int> x = const [1, 2, 3]])"),
("TargetFunc({Map<String, dynamic> config = const {'key': 'value'}}) {", "({Map<String, dynamic> config = const {'key': 'value'}})",),
("TargetFunc(Map<String, List<Map<int, String>>> crazyNested) {", "(Map<String, List<Map<int, String>>> crazyNested)"),
("TargetFunc(void Function(int, String) cb, {required bool Function() test}) {", "(void Function(int, String) cb, {required bool Function() test})"),
(
"TargetFunc({Map<String, dynamic> config = const {'key': 'value'}}) {",
"({Map<String, dynamic> config = const {'key': 'value'}})",
),
(
"TargetFunc(Map<String, List<Map<int, String>>> crazyNested) {",
"(Map<String, List<Map<int, String>>> crazyNested)",
),
(
"TargetFunc(void Function(int, String) cb, {required bool Function() test}) {",
"(void Function(int, String) cb, {required bool Function() test})",
),
("TargetFunc((int, String) recordArg) {", "((int, String) recordArg)"),
("TargetFunc({({int x, int y}) point}) {", "({({int x, int y}) point})"),
("TargetFunc(\n int a,\n {\n required String b,\n }\n) {", "(\n int a,\n {\n required String b,\n }\n)"),
(
"TargetFunc(\n int a,\n {\n required String b,\n }\n) {",
"(\n int a,\n {\n required String b,\n }\n)",
),
("TargetFunc( int a , \n String b ) {", "( int a , \n String b )"),
],
"invalid": [
Expand All @@ -321,23 +356,26 @@ def test_dart_slicer_bug_5_closure_invocation():
"/* (int a) */",
"String str = '(nested(parens))';",
],
"xfail_invalid": [
],
"xfail_invalid": [],
}


@pytest.mark.parametrize("payload,expected", ARGS_CASES["valid"])
def test_dart_args_valid(payload, expected):
assert_valid_match(DART_RULES["args"], payload, expected, "dart.args")


@pytest.mark.parametrize("payload", ARGS_CASES["invalid"])
def test_dart_args_invalid(payload):
assert_invalid_no_match(DART_RULES["args"], payload, "dart.args")


@pytest.mark.parametrize("payload", ARGS_CASES["xfail_invalid"])
@pytest.mark.xfail(reason="String/comment lookalikes lack AST block shielding", strict=True)
def test_dart_args_xfail_invalid(payload):
assert_invalid_no_match(DART_RULES["args"], payload, "dart.args")


# -------------------------------------------------------------------------
# 4. DEPENDENCY CAPTURE RULES
# -------------------------------------------------------------------------
Expand All @@ -355,7 +393,7 @@ def test_dart_args_xfail_invalid(payload):
("export 'src/internal.dart' show InternalClass;", "src/internal.dart"),
("import \n 'package:multiline/multiline.dart'\n as ml;", "package:multiline/multiline.dart"),
("import \t 'package:tab/tab.dart';", "package:tab/tab.dart"),
("import \"package:double_quotes/double_quotes.dart\";", "package:double_quotes/double_quotes.dart"),
('import "package:double_quotes/double_quotes.dart";', "package:double_quotes/double_quotes.dart"),
("part 'foo.g.dart';", "foo.g.dart"),
("part of 'foo.dart';", "foo.dart"),
("part of my_library_name;", "my_library_name"),
Expand All @@ -371,14 +409,17 @@ def test_dart_args_xfail_invalid(payload):
],
}


@pytest.mark.parametrize("payload,expected", DEPENDENCY_CAPTURE_CASES["valid"])
def test_dart_dependency_capture_valid(payload, expected):
assert_valid_dependency_match(DART_RULES["_dependency_capture"], payload, expected, "dart._dependency_capture")


@pytest.mark.parametrize("payload", DEPENDENCY_CAPTURE_CASES["invalid"])
def test_dart_dependency_capture_invalid(payload):
assert_invalid_no_match(DART_RULES["_dependency_capture"], payload, "dart._dependency_capture")


@pytest.mark.parametrize("payload", DEPENDENCY_CAPTURE_CASES["xfail_invalid"])
@pytest.mark.xfail(reason="String/comment lookalikes lack AST block shielding", strict=True)
def test_dart_dependency_capture_xfail_invalid(payload):
Expand Down
Loading
Loading