Merge pull request #183 from mull-project/code-climate

stanislaw · web-flow · commit 8f00c71657a9 · 2021-12-05T15:40:36.000+01:00
CHECK regex match: fix case when there are two+ "CHECK:" in a check line
diff --git a/.coverage b/.coverage
diff --git a/.github/workflows/ci-linux.yml b/.github/workflows/ci-linux.yml
@@ -37,4 +37,4 @@ jobs:
 
     - name: Run tests
       run: |
-        poetry run invoke test
+        poetry run invoke check
diff --git a/.github/workflows/ci-mac.yml b/.github/workflows/ci-mac.yml
@@ -37,4 +37,4 @@ jobs:
 
     - name: Run tests
       run: |
-        poetry run invoke test
+        poetry run invoke check
diff --git a/.github/workflows/ci-windows.yml b/.github/workflows/ci-windows.yml
@@ -50,5 +50,5 @@ jobs:
 
     - name: Run tests
       run: |
-        poetry run invoke test
+        poetry run invoke check
       shell: bash
diff --git a/.gitignore b/.gitignore
@@ -1,3 +1,4 @@
+.coverage
 .idea
 .venv
 __pycache__
diff --git a/filecheck/filecheck.py b/filecheck/filecheck.py
@@ -180,11 +180,6 @@ class CheckResult(Enum):
     CHECK_NOT_WITHOUT_MATCH = 5
 
 
-# Allow check prefixes only at the beginnings of lines or
-# after non-word characters.
-BEFORE_PREFIX = "^(.*?[^\\w-])?"
-
-
 def check_line(
     line, current_check, match_full_lines
 ):  # pylint: disable=too-many-return-statements
@@ -323,6 +318,11 @@ def parse_checks_from_strings(
 
     @staticmethod
     def parse_check(line: str, line_idx, config: Config) -> Optional[Check]:
+        # Allow check prefixes only at the beginnings of lines or
+        # after non-word characters.
+        # /^((?!PART).)*$/ is a negative match.
+        # https://stackoverflow.com/a/6259570/598057
+        before_prefix = f"^(((?!{config.check_prefix}).)*?[^-\\w])?"
         line = line.rstrip()
 
         if not config.strict_whitespace:
@@ -332,23 +332,22 @@ def parse_check(line: str, line_idx, config: Config) -> Optional[Check]:
         strict_whitespace_match = "" if config.strict_mode else " *"
 
         check_regex = (
-            f"{BEFORE_PREFIX}({config.check_prefix}):"
+            f"{before_prefix}({config.check_prefix}):"
             f"{strict_whitespace_match}(.*)"
         )
-
         check_match = re.search(check_regex, line)
         check_type = CheckType.CHECK
         if not check_match:
             check_regex = (
-                f"{BEFORE_PREFIX}({config.check_prefix}-NEXT):"
+                f"{before_prefix}({config.check_prefix}-NEXT):"
                 f"{strict_whitespace_match}(.*)"
             )
             check_match = re.search(check_regex, line)
             check_type = CheckType.CHECK_NEXT
 
         if check_match:
-            check_keyword = check_match.group(2)
-            check_expression = check_match.group(3)
+            check_keyword = check_match.group(3)
+            check_expression = check_match.group(4)
             if not config.strict_mode:
                 check_expression = check_expression.strip(" ")
 
@@ -386,20 +385,20 @@ def parse_check(line: str, line_idx, config: Config) -> Optional[Check]:
                 expression=check_expression,
                 source_line=line,
                 check_line_idx=line_idx,
-                start_index=check_match.start(3),
+                start_index=check_match.start(4),
             )
             return check
 
         check_not_regex = (
-            f"{BEFORE_PREFIX}({config.check_prefix}-NOT):"
+            f"{before_prefix}({config.check_prefix}-NOT):"
             f"{strict_whitespace_match}(.*)"
         )
         check_match = re.search(check_not_regex, line)
         if check_match:
             match_type = MatchType.SUBSTRING
 
-            check_keyword = check_match.group(2)
-            check_expression = check_match.group(3)
+            check_keyword = check_match.group(3)
+            check_expression = check_match.group(4)
             if not config.strict_mode:
                 check_expression = check_expression.strip(" ")
 
@@ -416,14 +415,14 @@ def parse_check(line: str, line_idx, config: Config) -> Optional[Check]:
                 expression=check_expression,
                 source_line=line,
                 check_line_idx=line_idx,
-                start_index=check_match.start(3),
+                start_index=check_match.start(4),
             )
             return check
 
-        check_empty_regex = f"{BEFORE_PREFIX}({config.check_prefix}-EMPTY):"
+        check_empty_regex = f"{before_prefix}({config.check_prefix}-EMPTY):"
         check_match = re.search(check_empty_regex, line)
         if check_match:
-            check_keyword = check_match.group(2)
+            check_keyword = check_match.group(3)
 
             check = Check(
                 check_type=CheckType.CHECK_EMPTY,
@@ -432,7 +431,7 @@ def parse_check(line: str, line_idx, config: Config) -> Optional[Check]:
                 expression=None,
                 source_line=line,
                 check_line_idx=line_idx,
-                start_index=check_match.start(2),
+                start_index=check_match.start(3),
             )
             return check
 
diff --git a/tasks.py b/tasks.py
@@ -187,12 +187,22 @@ def test_filecheck_py_using_filecheck_py_tester(context, focus=None):
     run_lit_tests(context, filecheck_exec, filecheck_tester_exec, focus, False)
 
 
-@task(lint, test_unit)
-def test(context, focus=None):
+@task
+def test_integration(context, focus=None):
     test_filecheck_llvm(context, focus)
     test_filecheck_py_using_file_check_llvm_tester(context, focus)
 
 
+@task(test_unit, test_integration)
+def test(_):
+    pass
+
+
+@task(lint, test)
+def check(_):
+    pass
+
+
 @task
 def clean(context):
     find_command = one_line_command(
diff --git a/tests/integration/tests/check_commands/CHECK/one_string/positive-match/05-CHECK-two-times-matches-first/sample.itest b/tests/integration/tests/check_commands/CHECK/one_string/positive-match/05-CHECK-two-times-matches-first/sample.itest
diff --git a/tests/unit/test_check_parser.py b/tests/unit/test_check_parser.py
@@ -0,0 +1,102 @@
+from filecheck.filecheck import CheckParser, Config, MatchType, CheckType
+
+
+def create_default_config(check_file):
+    return Config(
+        check_file=check_file,
+        match_full_lines=False,
+        strict_whitespace=False,
+        check_prefix=None,
+        implicit_check_not=None,
+        dump_input=None,
+    )
+
+
+def test_01_most_basic():
+    config = create_default_config("fake.filecheck")
+    input_string = "CHECK: String 1"
+    check = CheckParser.parse_check(input_string, 0, config)
+    assert check.check_type == CheckType.CHECK
+    assert check.match_type == MatchType.SUBSTRING
+    assert check.check_keyword == "CHECK"
+    assert check.expression == "String 1"
+    assert check.source_line == input_string
+    assert check.check_line_idx == 0
+    assert check.start_index == 7
+
+
+def test_02_two_check_words():
+    config = create_default_config("fake.filecheck")
+    input_string = "CHECK: CHECK: String 1"
+    check = CheckParser.parse_check(input_string, 0, config)
+    assert check.match_type == MatchType.SUBSTRING
+    assert check.check_keyword == "CHECK"
+    assert check.expression == "CHECK: String 1"
+    assert check.source_line == input_string
+    assert check.check_line_idx == 0
+    assert check.start_index == 7
+
+
+def test_10_regex_basic():
+    config = create_default_config("fake.filecheck")
+    input_string = "CHECK: {{String 1}}"
+    check = CheckParser.parse_check(input_string, 0, config)
+    assert check.match_type == MatchType.REGEX
+    assert check.check_keyword == "CHECK"
+    assert check.expression == "String 1"
+    assert check.source_line == input_string
+    assert check.check_line_idx == 0
+    assert check.start_index == 7
+
+
+def test_11_regex_two_check_words():
+    config = create_default_config("fake.filecheck")
+    input_string = "CHECK: CHECK: {{String 1}}"
+    check = CheckParser.parse_check(input_string, 0, config)
+    assert check.match_type == MatchType.REGEX
+    assert check.check_keyword == "CHECK"
+    assert check.source_line == input_string
+    assert check.check_line_idx == 0
+    assert check.start_index == 7
+
+    # The behavior seems to be different across Python 3.6 - 3.9.
+    assert (
+        check.expression == r"CHECK\:\ String 1"
+        or check.expression == r"CHECK:\ String 1"
+    )
+
+
+def test_90_check_two_times_no_offset():
+    config = create_default_config("fake.filecheck")
+    input_string = (
+        "CHECK:{{^.*filecheck.check:1:(9|10): "
+        "error: CHECK: expected string not found in input$}}"
+    )
+    check = CheckParser.parse_check(input_string, 0, config)
+    assert check.match_type == MatchType.REGEX
+    assert check.expression == (
+        "^.*filecheck.check:1:(9|10): "
+        "error: CHECK: expected string not found in input$"
+    )
+    assert check.check_keyword == "CHECK"
+    assert check.source_line == input_string
+    assert check.check_line_idx == 0
+    assert check.start_index == 6
+
+
+def test_91_check_two_times_with_offset():
+    config = create_default_config("fake.filecheck")
+    input_string = (
+        "; CHECK:{{^.*filecheck.check:1:(9|10): "
+        "error: CHECK: expected string not found in input$}}"
+    )
+    check = CheckParser.parse_check(input_string, 0, config)
+    assert check.match_type == MatchType.REGEX
+    assert check.expression == (
+        "^.*filecheck.check:1:(9|10): "
+        "error: CHECK: expected string not found in input$"
+    )
+    assert check.check_keyword == "CHECK"
+    assert check.source_line == input_string
+    assert check.check_line_idx == 0
+    assert check.start_index == 8

Original file line number	Diff line number	Diff line change
`@@ -1,3 +1,4 @@`
	`1`	`+.coverage`
`1`	`2`	`.idea`
`2`	`3`	`.venv`
`3`	`4`	`__pycache__`