Merge pull request #698 from jsa34/gherkin-official-parser

Implement the official gherkin parser

Author: youtux

.github/workflows/main.yml

@@ -13,27 +13,27 @@ jobs:
         include:
           - python-version: "3.8"
             toxfactor: py3.8
-            ignore-typecheck-outcome: true
+            ignore-typecheck-outcome: false
             ignore-test-outcome: false
           - python-version: "3.9"
             toxfactor: py3.9
-            ignore-typecheck-outcome: true
+            ignore-typecheck-outcome: false
             ignore-test-outcome: false
           - python-version: "3.10"
             toxfactor: py3.10
-            ignore-typecheck-outcome: true
+            ignore-typecheck-outcome: false
             ignore-test-outcome: false
           - python-version: "3.11"
             toxfactor: py3.11
-            ignore-typecheck-outcome: true
+            ignore-typecheck-outcome: false
             ignore-test-outcome: false
           - python-version: "3.12"
             toxfactor: py3.12
-            ignore-typecheck-outcome: true
+            ignore-typecheck-outcome: false
             ignore-test-outcome: false
           - python-version: "3.13-dev"
             toxfactor: py3.13
-            ignore-typecheck-outcome: true
+            ignore-typecheck-outcome: false
             ignore-test-outcome: false
 
     steps:
@@ -47,7 +47,7 @@ jobs:
 
     - name: Install poetry
       run: |
-        python -m pip install poetry==1.8.2
+        python -m pip install poetry==1.8.3
 
     - name: Configure poetry
       run: |
@@ -81,10 +81,10 @@ jobs:
         coverage combine
         coverage xml
 
-    - uses: codecov/codecov-action@v3
+    - uses: codecov/codecov-action@v4
       with:
         # Explicitly using the token to avoid Codecov rate limit errors
         # See https://community.codecov.com/t/upload-issues-unable-to-locate-build-via-github-actions-api/3954
         token: ${{ secrets.CODECOV_TOKEN }}
-        fail_ci_if_error: true
+        fail_ci_if_error: false
         verbose: true # optional (default = false)

CHANGES.rst

@@ -3,7 +3,10 @@ Changelog
 
 Unreleased
 ----------
-
+- Use `gherkin-official` parser to replace custom parsing logic.
+- Multiline steps must now always use triple-quotes for the additional lines.
+- All feature files must now use the keyword `Feature:` to be considered valid.
+- Tags can no longer have spaces (e.g. "@tag one" "@tag two" are no longer valid).
 
 7.3.0
 ----------

poetry.lock

@@ -42,6 +42,7 @@ parse-type = "*"
 pytest = ">=6.2.0"
 typing-extensions = "*"
 packaging = "*"
+gherkin-official = "^29.0.0"
 
 [tool.poetry.group.dev.dependencies]
 tox = ">=4.11.3"
@@ -93,6 +94,7 @@ python_version = "3.8"
 warn_return_any = true
 warn_unused_configs = true
 files = "src/pytest_bdd/**/*.py"
+disallow_untyped_defs = true
 
 [[tool.mypy.overrides]]
 module = ["parse", "parse_type"]

pyproject.toml

@@ -1,4 +1,5 @@
 [pytest]
 testpaths = tests
 filterwarnings =
-    error
+    # only ignore errors from the pytest_bdd package
+    error:::(src)?\.pytest_bdd.*

pytest.ini

@@ -34,12 +34,12 @@ def inject_fixture(request: FixtureRequest, arg: str, value: Any) -> None:
         # if there was already one registered, so we need to force its value
         # to the one we want to inject.
         fixture_def = request._get_active_fixturedef(arg)
-        fixture_def.cached_result = (value, None, None)
+        fixture_def.cached_result = (value, None, None)  # type: ignore
 
 else:
 
     def getfixturedefs(fixturemanager: FixtureManager, fixturename: str, node: Node) -> Sequence[FixtureDef] | None:
-        return fixturemanager.getfixturedefs(fixturename, node.nodeid)
+        return fixturemanager.getfixturedefs(fixturename, node.nodeid)  # type: ignore
 
     def inject_fixture(request: FixtureRequest, arg: str, value: Any) -> None:
         """Inject fixture into pytest fixture request.
@@ -49,7 +49,7 @@ def inject_fixture(request: FixtureRequest, arg: str, value: Any) -> None:
         :param value: argument value
         """
         fd = FixtureDef(
-            fixturemanager=request._fixturemanager,
+            fixturemanager=request._fixturemanager,  # type: ignore
             baseid=None,
             argname=arg,
             func=lambda: value,

src/pytest_bdd/compat.py

@@ -35,14 +35,14 @@ def configure(config: Config) -> None:
     cucumber_json_path = config.option.cucumber_json_path
     # prevent opening json log on worker nodes (xdist)
     if cucumber_json_path and not hasattr(config, "workerinput"):
-        config._bddcucumberjson = LogBDDCucumberJSON(cucumber_json_path)
-        config.pluginmanager.register(config._bddcucumberjson)
+        config._bddcucumberjson = LogBDDCucumberJSON(cucumber_json_path)  # type: ignore[attr-defined]
+        config.pluginmanager.register(config._bddcucumberjson)  # type: ignore[attr-defined]
 
 
 def unconfigure(config: Config) -> None:
-    xml = getattr(config, "_bddcucumberjson", None)
+    xml = getattr(config, "_bddcucumberjson", None)  # type: ignore[attr-defined]
     if xml is not None:
-        del config._bddcucumberjson
+        del config._bddcucumberjson  # type: ignore[attr-defined]
         config.pluginmanager.unregister(xml)
 
 

src/pytest_bdd/cucumber_json.py

@@ -15,10 +15,6 @@ class ScenarioNotFound(ScenarioValidationError):
     """Scenario Not Found."""
 
 
-class ExamplesNotValidError(ScenarioValidationError):
-    """Example table is not valid."""
-
-
 class StepDefinitionNotFoundError(Exception):
     """Step definition not found."""
 
@@ -27,11 +23,39 @@ class NoScenariosFound(Exception):
     """No scenarios found."""
 
 
-class FeatureError(Exception):
-    """Feature parse error."""
+class GherkinParseError(Exception):
+    """Base class for all Gherkin parsing errors."""
 
-    message = "{0}.\nLine number: {1}.\nLine: {2}.\nFile: {3}"
+    def __init__(self, message: str, line: int, line_content: str, filename: str) -> None:
+        super().__init__(message)
+        self.message = message
+        self.line = line
+        self.line_content = line_content
+        self.filename = filename
 
     def __str__(self) -> str:
-        """String representation."""
-        return self.message.format(*self.args)
+        return f"{self.message}\nLine number: {self.line}\nLine: {self.line_content}\nFile: {self.filename}"
+
+
+class FeatureError(GherkinParseError):
+    pass
+
+
+class BackgroundError(GherkinParseError):
+    pass
+
+
+class ScenarioError(GherkinParseError):
+    pass
+
+
+class StepError(GherkinParseError):
+    pass
+
+
+class RuleError(GherkinParseError):
+    pass
+
+
+class TokenError(GherkinParseError):
+    pass

src/pytest_bdd/exceptions.py

@@ -29,7 +29,7 @@
 import glob
 import os.path
 
-from .parser import Feature, parse_feature
+from .parser import Feature, FeatureParser
 
 # Global features dictionary
 features: dict[str, Feature] = {}
@@ -52,30 +52,29 @@ def get_feature(base_path: str, filename: str, encoding: str = "utf-8") -> Featu
     full_name = os.path.abspath(os.path.join(base_path, filename))
     feature = features.get(full_name)
     if not feature:
-        feature = parse_feature(base_path, filename, encoding=encoding)
+        feature = FeatureParser(base_path, filename, encoding).parse()
         features[full_name] = feature
     return feature
 
 
-def get_features(paths: list[str], **kwargs) -> list[Feature]:
+def get_features(paths: list[str], encoding: str = "utf-8") -> list[Feature]:
     """Get features for given paths.
 
     :param list paths: `list` of paths (file or dirs)
 
     :return: `list` of `Feature` objects.
     """
     seen_names = set()
-    features = []
+    _features = []
     for path in paths:
         if path not in seen_names:
             seen_names.add(path)
             if os.path.isdir(path):
-                features.extend(
-                    get_features(glob.iglob(os.path.join(path, "**", "*.feature"), recursive=True), **kwargs)
-                )
+                file_paths = list(glob.iglob(os.path.join(path, "**", "*.feature"), recursive=True))
+                _features.extend(get_features(file_paths, encoding=encoding))
             else:
                 base, name = os.path.split(path)
-                feature = get_feature(base, name, **kwargs)
-                features.append(feature)
-    features.sort(key=lambda feature: feature.name or feature.filename)
-    return features
+                feature = get_feature(base, name, encoding=encoding)
+                _features.append(feature)
+    _features.sort(key=lambda _feature: _feature.name or _feature.filename)
+    return _features

src/pytest_bdd/feature.py

@@ -7,7 +7,7 @@
 from typing import TYPE_CHECKING, cast
 
 from _pytest._io import TerminalWriter
-from mako.lookup import TemplateLookup
+from mako.lookup import TemplateLookup  # type: ignore
 
 from .compat import getfixturedefs
 from .feature import get_features
@@ -181,11 +181,11 @@ def _show_missing_code_main(config: Config, session: Session) -> None:
     features, scenarios, steps = parse_feature_files(config.option.features)
 
     for item in session.items:
-        if scenario := getattr(item.obj, "__scenario__", None):
+        if scenario := getattr(item.obj, "__scenario__", None):  # type: ignore
             if scenario in scenarios:
                 scenarios.remove(scenario)
             for step in scenario.steps:
-                if _find_step_fixturedef(fm, item, step=step):
+                if _find_step_fixturedef(fm, item, step=step):  # type: ignore
                     try:
                         steps.remove(step)
                     except ValueError: