fix(scenarios): preserve scenario labels for unclosed comments

Author: kvlonge

src/sqlbuild/compiler/compile/_helpers/scenarios/core.py

@@ -60,7 +60,7 @@ def extract_sql_scenario_ctes(*, sql: str, file_label: str) -> CompileSqlScenari
 def extract_sql_scenario_expected_model_names(*, sql: str, file_label: str) -> tuple[str, ...]:
     """Extract explicit expected-model relationships without inspecting CTE bodies."""
 
-    start: int = _skip_ignorable(sql=sql, start=0)
+    start: int = _skip_ignorable(sql=sql, start=0, context_label=_CONTEXT)
     if _try_consume_keyword(sql=sql, start=start, keyword=SQL_WITH_KEYWORD) is None:
         return ()
     ctes: tuple[CompileSqlScenarioCte, ...] = extract_top_level_ctes_with_scanner(

src/sqlbuild/compiler/compile/_helpers/sql_tests/core.py

@@ -215,16 +215,18 @@ def extract_top_level_ctes_with_scanner[CteT](
     """Scan top-level `WITH` CTEs followed by the ceremonial `SELECT 1` of a test file."""
 
     with_end: int | None = _try_consume_keyword(
-        sql=sql, start=_skip_ignorable(sql=sql, start=0), keyword=SQL_WITH_KEYWORD
+        sql=sql,
+        start=_skip_ignorable(sql=sql, start=0, context_label=context_label),
+        keyword=SQL_WITH_KEYWORD,
     )
     if with_end is None:
         raise CompileInputError(
             f"{context_label} '{file_label}' must declare {with_requirement} before `SELECT 1`"
         )
-    index: int = _skip_ignorable(sql=sql, start=with_end)
+    index: int = _skip_ignorable(sql=sql, start=with_end, context_label=context_label)
     recursive_end: int | None = _try_consume_keyword(sql=sql, start=index, keyword="RECURSIVE")
     if recursive_end is not None:
-        index = _skip_ignorable(sql=sql, start=recursive_end)
+        index = _skip_ignorable(sql=sql, start=recursive_end, context_label=context_label)
 
     ctes: list[CteT] = []
     seen_cte_names: set[str] = set()
@@ -238,18 +240,18 @@ def extract_top_level_ctes_with_scanner[CteT](
             )
         seen_cte_names.add(cte_name)
 
-        index = _skip_ignorable(sql=sql, start=index)
+        index = _skip_ignorable(sql=sql, start=index, context_label=context_label)
         if index < len(sql) and sql[index] == SQL_OPEN_PAREN_TOKEN:
             index = find_matching_paren(sql=sql, open_paren_index=index, context=context_label) + 1
-            index = _skip_ignorable(sql=sql, start=index)
+            index = _skip_ignorable(sql=sql, start=index, context_label=context_label)
         index = _consume_keyword(
             sql=sql,
             start=index,
             keyword="AS",
             file_label=file_label,
             context_label=context_label,
         )
-        index = _skip_ignorable(sql=sql, start=index)
+        index = _skip_ignorable(sql=sql, start=index, context_label=context_label)
         if index >= len(sql) or sql[index] != SQL_OPEN_PAREN_TOKEN:
             raise CompileInputError(
                 f"{context_label} '{file_label}' CTE '{cte_name}' must use AS (...)"
@@ -259,9 +261,9 @@ def extract_top_level_ctes_with_scanner[CteT](
             sql=sql, open_paren_index=index, context=context_label
         )
         ctes.append(cte_type(name=cte_name, sql_body=sql[cte_body_start:cte_body_end].strip()))
-        index = _skip_ignorable(sql=sql, start=cte_body_end + 1)
+        index = _skip_ignorable(sql=sql, start=cte_body_end + 1, context_label=context_label)
         if index < len(sql) and sql[index] == SQL_ARGUMENT_SEPARATOR_TOKEN:
-            index = _skip_ignorable(sql=sql, start=index + 1)
+            index = _skip_ignorable(sql=sql, start=index + 1, context_label=context_label)
             continue
         break
 
@@ -857,30 +859,30 @@ def _validate_ceremonial_select(
     file_label: str,
     context_label: str = _CONTEXT,
 ) -> None:
-    if _is_ceremonial_select_statement(sql=sql, start=start):
+    if _is_ceremonial_select_statement(sql=sql, start=start, context_label=context_label):
         return
     raise CompileInputError(
         f"{context_label} '{file_label}' must end with a ceremonial top-level `SELECT 1` "
         "after its CTEs"
     )
 
 
-def _is_ceremonial_select_statement(*, sql: str, start: int) -> bool:
-    index: int = _skip_ignorable(sql=sql, start=start)
+def _is_ceremonial_select_statement(*, sql: str, start: int, context_label: str = _CONTEXT) -> bool:
+    index: int = _skip_ignorable(sql=sql, start=start, context_label=context_label)
     select_end: int | None = _try_consume_keyword(sql=sql, start=index, keyword="SELECT")
     if select_end is None:
         return False
-    index = _skip_ignorable(sql=sql, start=select_end)
+    index = _skip_ignorable(sql=sql, start=select_end, context_label=context_label)
     if index >= len(sql) or sql[index] != SQL_CEREMONIAL_SELECT_VALUE:
         return False
-    index = _skip_ignorable(sql=sql, start=index + 1)
-    return _is_statement_end(sql=sql, start=index)
+    index = _skip_ignorable(sql=sql, start=index + 1, context_label=context_label)
+    return _is_statement_end(sql=sql, start=index, context_label=context_label)
 
 
-def _is_statement_end(*, sql: str, start: int) -> bool:
+def _is_statement_end(*, sql: str, start: int, context_label: str = _CONTEXT) -> bool:
     index: int = start
     if index < len(sql) and sql[index] == SQL_STATEMENT_TERMINATOR_TOKEN:
-        index = _skip_ignorable(sql=sql, start=index + 1)
+        index = _skip_ignorable(sql=sql, start=index + 1, context_label=context_label)
     return index == len(sql)
 
 
@@ -1097,7 +1099,7 @@ def _read_identifier(
     return sql[start:index], index
 
 
-def _skip_ignorable(*, sql: str, start: int) -> int:
+def _skip_ignorable(*, sql: str, start: int, context_label: str = _CONTEXT) -> int:
     index: int = start
     while index < len(sql):
         if sql[index].isspace():
@@ -1107,7 +1109,7 @@ def _skip_ignorable(*, sql: str, start: int) -> int:
             index = skip_line_comment(sql=sql, start=index)
             continue
         if sql.startswith("/*", index):
-            index = skip_block_comment(sql=sql, start=index, context=_CONTEXT)
+            index = skip_block_comment(sql=sql, start=index, context=context_label)
             continue
         return index
     return index

tests/e2e/src/sqlbuild/cli/commands/main/scenario/test_scenario.py

@@ -42,6 +42,36 @@
 )
 
 
+@pytest.mark.parametrize(
+    "test_case",
+    (
+        ScenarioCliE2ETestCase(
+            description="unclosed scenario comment names the scenario",
+            command=("--no-color", "scenario", "test"),
+            expected_exit_code=1,
+            expected_stderr_fragments=("SQL scenario contains an unclosed block comment",),
+        ),
+    ),
+    ids=lambda case: case.description,
+)
+def test_given_unclosed_scenario_comment_when_running_cli_then_error_names_scenario(
+    test_case: ScenarioCliE2ETestCase,
+    tmp_path: Path,
+) -> None:
+    repo_files: dict[str, str] = build_scenario_project_files()
+    repo_files["tests/scenarios/order_totals_pass.sql"] = "SCENARIO ();\nWITH /* unfinished"
+    project_dir: Path = prepare_inline_project(
+        tmp_path=tmp_path, project_name="scenario_comment", repo_files=repo_files
+    )
+    result: subprocess.CompletedProcess[str] = run_sqb(
+        command=test_case.command, project_dir=project_dir
+    )
+
+    assert result.returncode == test_case.expected_exit_code
+    for fragment in test_case.expected_stderr_fragments:
+        assert fragment in result.stderr
+
+
 @pytest.mark.parametrize(
     "test_case",
     (

tests/unit/src/sqlbuild/compiler/compile/_helpers/test_scenarios.py

@@ -457,6 +457,16 @@ def test_given_invalid_scenario_source_refs_when_building_inputs_then_it_raises_
                 "one __expected__<model> or __assert__<assertion> CTE before `SELECT 1`"
             ),
         ),
+        CteScannerMessageTestCase(
+            description="unclosed comment before a fixture uses scenario wording",
+            sql="WITH /* unfinished",
+            expected_message="SQL scenario contains an unclosed block comment",
+        ),
+        CteScannerMessageTestCase(
+            description="unclosed trailing comment uses scenario wording",
+            sql="WITH __expected__orders AS (SELECT 1 AS order_id) SELECT 1; /* unfinished",
+            expected_message="SQL scenario contains an unclosed block comment",
+        ),
         CteScannerMessageTestCase(
             description="scenario with an invalid cte name uses scenario wording",
             sql="WITH 1orders AS (SELECT 1) SELECT 1",
@@ -505,6 +515,16 @@ def test_given_malformed_scenario_ctes_when_extracting_then_messages_name_the_sc
             sql="WITH __source__raw_orders (SELECT 1 AS order_id) SELECT 1",
             expected_message="SQL scenario 'tests/scenarios/orders.sql' expected keyword AS",
         ),
+        CteScannerMessageTestCase(
+            description="expected model scan unclosed comment uses scenario wording",
+            sql="WITH /* unfinished",
+            expected_message="SQL scenario contains an unclosed block comment",
+        ),
+        CteScannerMessageTestCase(
+            description="expected model scan leading unclosed comment uses scenario wording",
+            sql="/* unfinished",
+            expected_message="SQL scenario contains an unclosed block comment",
+        ),
         CteScannerMessageTestCase(
             description="expected model scan without a target uses scenario wording",
             sql=(