]> git.madduck.net Git - etc/vim.git/commitdiff

madduck's git repository

Every one of the projects in this repository is available at the canonical URL git://git.madduck.net/madduck/pub/<projectpath> — see each project's metadata for the exact URL.

All patches and comments are welcome. Please squash your changes to logical commits before using git-format-patch and git-send-email to patches@git.madduck.net. If you'd read over the Git project's submission guidelines and adhered to them, I'd be especially grateful.

SSH access, as well as push access can be individually arranged.

If you use my repositories frequently, consider adding the following snippet to ~/.gitconfig and using the third clone URL listed for each project:

[url "git://git.madduck.net/madduck/"]
  insteadOf = madduck:

Add black-primer unittests (#1426)
authorCooper Lees <cooper@fb.com>
Sun, 17 May 2020 19:18:49 +0000 (12:18 -0700)
committerGitHub <noreply@github.com>
Sun, 17 May 2020 19:18:49 +0000 (12:18 -0700)
* Add black-primer unittests

- Get this tool covered with some decent unittests for all unittests wins
- Have a CLI and lib test class
- Import it from `test_black.py` so we always run tests
- Revert typing asyncio.Queue as Queue[str] so we can work in 3.6
- **mypy**: Until black > 3.6 disallow_any_generics=False for primer code

Test:
- Run tests: `coverage run tests/test_primer.py` or `coverage run -m unittest`
```
(b) cooper-mbp1:black cooper$ coverage report
Name                      Stmts   Miss  Cover
---------------------------------------------
src/black_primer/cli.py      49      8    84%
src/black_primer/lib.py     148     28    81%
tests/test_primer.py        114      1    99%
---------------------------------------------
TOTAL                       311     37    88%
```

* Use ProactorEventLoop for Windows + fix false path for Linux

* Set Windows to use ProactorEventLoop in  to benefit all callers

* sys.platform seems to not having the loop applied - So type ignore and use platform.system() gate

* Have each test loop correctly set to ProactorEventLoop on Windows for < 3.8 too

mypy.ini
src/black_primer/cli.py
src/black_primer/lib.py
tests/test_black.py
tests/test_primer.py [new file with mode: 0644]

index a484d6638e356e3fae12a8c5ebc4be62e92f5458..295bab4e30288d3aae369d841f3c652d441f56d0 100644 (file)
--- a/mypy.ini
+++ b/mypy.ini
@@ -21,7 +21,8 @@ strict_optional=True
 warn_no_return=True
 warn_redundant_casts=True
 warn_unused_ignores=True
-disallow_any_generics=True
+# Until we're not supporting 3.6 primer needs this
+disallow_any_generics=False
 
 # The following are off by default.  Flip them on if you feel
 # adventurous.
index f7d3321a9ea44e7631f8f44d66007fcf7a9e476e..83357a9f26f6a05052f64fd0e95cdf85e24bd4d3 100644 (file)
@@ -63,7 +63,7 @@ async def async_main(
             LOG.debug(f"Removing {work_path}")
             rmtree(work_path)
 
-    return -1
+    return -2
 
 
 @click.command(context_settings={"help_option_names": ["-h", "--help"]})
@@ -131,5 +131,5 @@ def main(ctx: click.core.Context, **kwargs: Any) -> None:
         loop.close()
 
 
-if __name__ == "__main__":
+if __name__ == "__main__":  # pragma: nocover
     main()
index be0618ca5b2f53d3fe5c12a5f9ef51256951a14e..4480f27ee1e1eb65c01823771ab083935a3b368a 100644 (file)
@@ -1,12 +1,11 @@
 #!/usr/bin/env python3
 
-# Module '__future__' has no attribute 'annotations'
-from __future__ import annotations  # type: ignore
-
 import asyncio
 import json
 import logging
+import sys
 from pathlib import Path
+from platform import system
 from shutil import rmtree, which
 from subprocess import CalledProcessError
 from sys import version_info
@@ -16,9 +15,20 @@ from urllib.parse import urlparse
 import click
 
 
+WINDOWS = system() == "Windows"
+BLACK_BINARY = "black.exe" if WINDOWS else "black"
+GIT_BIANRY = "git.exe" if WINDOWS else "git"
 LOG = logging.getLogger(__name__)
 
 
+# Windows needs a ProactorEventLoop if you want to exec subprocesses
+# Startng 3.8 this is the default - Can remove when black >= 3.8
+# mypy only respects sys.platform if directly in the evaluation
+# # https://mypy.readthedocs.io/en/latest/common_issues.html#python-version-and-system-platform-checks  # noqa: B950
+if sys.platform == "win32":
+    asyncio.set_event_loop(asyncio.ProactorEventLoop())
+
+
 class Results(NamedTuple):
     stats: Dict[str, int] = {}
     failed_projects: Dict[str, CalledProcessError] = {}
@@ -53,7 +63,7 @@ async def _gen_check_output(
     return (stdout, stderr)
 
 
-async def analyze_results(project_count: int, results: Results) -> int:
+def analyze_results(project_count: int, results: Results) -> int:
     failed_pct = round(((results.stats["failed"] / project_count) * 100), 2)
     success_pct = round(((results.stats["success"] / project_count) * 100), 2)
 
@@ -95,8 +105,8 @@ async def black_run(
     repo_path: Path, project_config: Dict[str, Any], results: Results
 ) -> None:
     """Run black and record failures"""
-    cmd = [str(which("black"))]
-    if project_config["cli_arguments"]:
+    cmd = [str(which(BLACK_BINARY))]
+    if "cli_arguments" in project_config and project_config["cli_arguments"]:
         cmd.extend(*project_config["cli_arguments"])
     cmd.extend(["--check", "--diff", "."])
 
@@ -106,12 +116,27 @@ async def black_run(
         results.stats["failed"] += 1
         LOG.error(f"Running black for {repo_path} timed out ({cmd})")
     except CalledProcessError as cpe:
-        # TODO: This might need to be tuned and made smarter for higher signal
-        if not project_config["expect_formatting_changes"] and cpe.returncode == 1:
-            results.stats["failed"] += 1
-            results.failed_projects[repo_path.name] = cpe
+        # TODO: Tune for smarter for higher signal
+        # If any other reutrn value than 1 we raise - can disable project in config
+        if cpe.returncode == 1:
+            if not project_config["expect_formatting_changes"]:
+                results.stats["failed"] += 1
+                results.failed_projects[repo_path.name] = cpe
+            else:
+                results.stats["success"] += 1
             return
 
+        LOG.error(f"Unkown error with {repo_path}")
+        raise
+
+    # If we get here and expect formatting changes something is up
+    if project_config["expect_formatting_changes"]:
+        results.stats["failed"] += 1
+        results.failed_projects[repo_path.name] = CalledProcessError(
+            0, cmd, b"Expected formatting changes but didn't get any!", b""
+        )
+        return
+
     results.stats["success"] += 1
 
 
@@ -123,7 +148,7 @@ async def git_checkout_or_rebase(
     depth: int = 1,
 ) -> Optional[Path]:
     """git Clone project or rebase"""
-    git_bin = str(which("git"))
+    git_bin = str(which(GIT_BIANRY))
     if not git_bin:
         LOG.error("No git binary found")
         return None
@@ -151,7 +176,7 @@ async def git_checkout_or_rebase(
 
 async def load_projects_queue(
     config_path: Path,
-) -> Tuple[Dict[str, Any], asyncio.Queue[str]]:
+) -> Tuple[Dict[str, Any], asyncio.Queue]:
     """Load project config and fill queue with all the project names"""
     with config_path.open("r") as cfp:
         config = json.load(cfp)
@@ -159,7 +184,7 @@ async def load_projects_queue(
     # TODO: Offer more options here
     # e.g. Run on X random packages or specific sub list etc.
     project_names = sorted(config["projects"].keys())
-    queue: asyncio.Queue[str] = asyncio.Queue(maxsize=len(project_names))
+    queue: asyncio.Queue = asyncio.Queue(maxsize=len(project_names))
     for project in project_names:
         await queue.put(project)
 
@@ -169,7 +194,7 @@ async def load_projects_queue(
 async def project_runner(
     idx: int,
     config: Dict[str, Any],
-    queue: asyncio.Queue[str],
+    queue: asyncio.Queue,
     work_path: Path,
     results: Results,
     long_checkouts: bool = False,
@@ -243,7 +268,7 @@ async def process_queue(
     config, queue = await load_projects_queue(Path(config_file))
     project_count = queue.qsize()
     LOG.info(f"{project_count} projects to run black over")
-    if not project_count:
+    if project_count < 1:
         return -1
 
     LOG.debug(f"Using {workers} parallel workers to run black")
@@ -258,4 +283,8 @@ async def process_queue(
     )
 
     LOG.info("Analyzing results")
-    return await analyze_results(project_count, results)
+    return analyze_results(project_count, results)
+
+
+if __name__ == "__main__":  # pragma: nocover
+    raise NotImplementedError("lib is a library, funnily enough.")
index cb765aff227fba6de25a2304a81e9cbd538c7d28..1b282b8d41134f8d6feda6867e67f2a7b1ebdff9 100644 (file)
@@ -31,6 +31,10 @@ else:
 
 from pathspec import PathSpec
 
+# Import other test classes
+from .test_primer import PrimerCLITests  # noqa: F401
+
+
 ff = partial(black.format_file_in_place, mode=black.FileMode(), fast=True)
 fs = partial(black.format_str, mode=black.FileMode())
 THIS_FILE = Path(__file__)
diff --git a/tests/test_primer.py b/tests/test_primer.py
new file mode 100644 (file)
index 0000000..da1465f
--- /dev/null
@@ -0,0 +1,207 @@
+#!/usr/bin/env python3
+
+import asyncio
+import sys
+import unittest
+from contextlib import contextmanager
+from copy import deepcopy
+from io import StringIO
+from os import getpid
+from pathlib import Path
+from platform import system
+from subprocess import CalledProcessError
+from tempfile import TemporaryDirectory, gettempdir
+from typing import Any, Callable, Generator, Iterator, Tuple
+from unittest.mock import Mock, patch
+
+from click.testing import CliRunner
+
+from black_primer import cli, lib
+
+
+EXPECTED_ANALYSIS_OUTPUT = """\
+-- primer results 📊 --
+
+68 / 69 succeeded (98.55%) ✅
+1 / 69 FAILED (1.45%) 💩
+ - 0 projects Disabled by config
+ - 0 projects skipped due to Python Version
+ - 0 skipped due to long checkout
+
+Failed Projects:
+
+## black:
+ - Returned 69
+ - stdout:
+black didn't work
+
+"""
+FAKE_PROJECT_CONFIG = {
+    "cli_arguments": ["--unittest"],
+    "expect_formatting_changes": False,
+    "git_clone_url": "https://github.com/psf/black.git",
+}
+
+
+@contextmanager
+def capture_stdout(command: Callable, *args: Any, **kwargs: Any) -> Generator:
+    old_stdout, sys.stdout = sys.stdout, StringIO()
+    try:
+        command(*args, **kwargs)
+        sys.stdout.seek(0)
+        yield sys.stdout.read()
+    finally:
+        sys.stdout = old_stdout
+
+
+@contextmanager
+def event_loop() -> Iterator[None]:
+    policy = asyncio.get_event_loop_policy()
+    loop = policy.new_event_loop()
+    asyncio.set_event_loop(loop)
+    if sys.platform == "win32":
+        asyncio.set_event_loop(asyncio.ProactorEventLoop())
+    try:
+        yield
+    finally:
+        loop.close()
+
+
+async def raise_subprocess_error(*args: Any, **kwargs: Any) -> None:
+    raise CalledProcessError(1, ["unittest", "error"], b"", b"")
+
+
+async def return_false(*args: Any, **kwargs: Any) -> bool:
+    return False
+
+
+async def return_subproccess_output(*args: Any, **kwargs: Any) -> Tuple[bytes, bytes]:
+    return (b"stdout", b"stderr")
+
+
+async def return_zero(*args: Any, **kwargs: Any) -> int:
+    return 0
+
+
+class PrimerLibTests(unittest.TestCase):
+    def test_analyze_results(self) -> None:
+        fake_results = lib.Results(
+            {
+                "disabled": 0,
+                "failed": 1,
+                "skipped_long_checkout": 0,
+                "success": 68,
+                "wrong_py_ver": 0,
+            },
+            {"black": CalledProcessError(69, ["black"], b"black didn't work", b"")},
+        )
+        with capture_stdout(lib.analyze_results, 69, fake_results) as analyze_stdout:
+            self.assertEqual(EXPECTED_ANALYSIS_OUTPUT, analyze_stdout)
+
+    @event_loop()
+    def test_black_run(self) -> None:
+        """Pretend run black to ensure we cater for all scenarios"""
+        loop = asyncio.get_event_loop()
+        repo_path = Path(gettempdir())
+        project_config = deepcopy(FAKE_PROJECT_CONFIG)
+        results = lib.Results({"failed": 0, "success": 0}, {})
+
+        # Test a successful black run
+        with patch("black_primer.lib._gen_check_output", return_subproccess_output):
+            loop.run_until_complete(lib.black_run(repo_path, project_config, results))
+        self.assertEqual(1, results.stats["success"])
+        self.assertFalse(results.failed_projects)
+
+        # Test a fail based on expecting formatting changes but not getting any
+        project_config["expect_formatting_changes"] = True
+        results = lib.Results({"failed": 0, "success": 0}, {})
+        with patch("black_primer.lib._gen_check_output", return_subproccess_output):
+            loop.run_until_complete(lib.black_run(repo_path, project_config, results))
+        self.assertEqual(1, results.stats["failed"])
+        self.assertTrue(results.failed_projects)
+
+        # Test a fail based on returning 1 and not expecting formatting changes
+        project_config["expect_formatting_changes"] = False
+        results = lib.Results({"failed": 0, "success": 0}, {})
+        with patch("black_primer.lib._gen_check_output", raise_subprocess_error):
+            loop.run_until_complete(lib.black_run(repo_path, project_config, results))
+        self.assertEqual(1, results.stats["failed"])
+        self.assertTrue(results.failed_projects)
+
+    @event_loop()
+    def test_gen_check_output(self) -> None:
+        loop = asyncio.get_event_loop()
+        stdout, stderr = loop.run_until_complete(
+            lib._gen_check_output([lib.BLACK_BINARY, "--help"])
+        )
+        self.assertTrue("The uncompromising code formatter" in stdout.decode("utf8"))
+        self.assertEqual(None, stderr)
+
+        # TODO: Add a test to see failure works on Windows
+        if lib.WINDOWS:
+            return
+
+        false_bin = "/usr/bin/false" if system() == "Darwin" else "/bin/false"
+        with self.assertRaises(CalledProcessError):
+            loop.run_until_complete(lib._gen_check_output([false_bin]))
+
+        with self.assertRaises(asyncio.TimeoutError):
+            loop.run_until_complete(
+                lib._gen_check_output(["/bin/sleep", "2"], timeout=0.1)
+            )
+
+    @event_loop()
+    def test_git_checkout_or_rebase(self) -> None:
+        loop = asyncio.get_event_loop()
+        project_config = deepcopy(FAKE_PROJECT_CONFIG)
+        work_path = Path(gettempdir())
+
+        expected_repo_path = work_path / "black"
+        with patch("black_primer.lib._gen_check_output", return_subproccess_output):
+            returned_repo_path = loop.run_until_complete(
+                lib.git_checkout_or_rebase(work_path, project_config)
+            )
+        self.assertEqual(expected_repo_path, returned_repo_path)
+
+    @patch("sys.stdout", new_callable=StringIO)
+    @event_loop()
+    def test_process_queue(self, mock_stdout: Mock) -> None:
+        loop = asyncio.get_event_loop()
+        config_path = Path(lib.__file__).parent / "primer.json"
+        with patch("black_primer.lib.git_checkout_or_rebase", return_false):
+            with TemporaryDirectory() as td:
+                return_val = loop.run_until_complete(
+                    lib.process_queue(str(config_path), td, 2)
+                )
+                self.assertEqual(0, return_val)
+
+
+class PrimerCLITests(unittest.TestCase):
+    @event_loop()
+    def test_async_main(self) -> None:
+        loop = asyncio.get_event_loop()
+        work_dir = Path(gettempdir()) / f"primer_ut_{getpid()}"
+        args = {
+            "config": "/config",
+            "debug": False,
+            "keep": False,
+            "long_checkouts": False,
+            "rebase": False,
+            "workdir": str(work_dir),
+            "workers": 69,
+        }
+        with patch("black_primer.cli.lib.process_queue", return_zero):
+            return_val = loop.run_until_complete(cli.async_main(**args))
+            self.assertEqual(0, return_val)
+
+    def test_handle_debug(self) -> None:
+        self.assertTrue(cli._handle_debug(None, None, True))
+
+    def test_help_output(self) -> None:
+        runner = CliRunner()
+        result = runner.invoke(cli.main, ["--help"])
+        self.assertEqual(result.exit_code, 0)
+
+
+if __name__ == "__main__":
+    unittest.main()