Merge e2b0217 into 2cd99d8

README.md

@@ -55,15 +55,16 @@ You can list the available options by running `reformat-gherkin --help`.
 ```text
 Usage: reformat-gherkin [OPTIONS] [SRC]...
 
-  Reformat the given Gherkin files and all files in the given directories
-  recursively.
+  Reformat the given SRC files and all .feature files in SRC folders. If -
+  is passed as a file, reformat stdin and print the result to stdout.
 
 Options:
   --check                         Don't write the files back, just return the
                                   status. Return code 0 means nothing would
                                   change. Return code 1 means some files would
                                   be reformatted. Return code 123 means there
                                   was an internal error.
+
   -a, --alignment [left|right]    Specify the alignment of step keywords
                                   (Given, When, Then,...). If specified, all
                                   statements after step keywords are left-
@@ -72,18 +73,23 @@ Options:
                                   default, step keywords are left-aligned, and
                                   there is a single space between the step
                                   keyword and the statement.
+
   -n, --newline [LF|CRLF]         Specify the line separators when formatting
                                   files inplace. If not specified, line
                                   separators are preserved.
+
   --fast / --safe                 If --fast given, skip the sanity checks of
                                   file contents. [default: --safe]
+
   --single-line-tags / --multi-line-tags
                                   If --single-line-tags given, output
                                   consecutive tags on one line. If --multi-
                                   line-tags given, output one tag per line.
                                   [default: --single-line-tags]
+
   --tab-width INTEGER             Specify the number of spaces per
                                   indentation-level. [default: 2]
+
   --use-tabs                      Indent lines with tabs instead of spaces.
   --config FILE                   Read configuration from FILE.
   --version                       Show the version and exit.
@@ -93,6 +99,7 @@ Options:
 Reformat-gherkin is a well-behaved Unix-style command-line tool:
 
 - it does nothing if no sources are passed to it;
+- it will read from standard input and write to standard output if - is used as the filename;
 - it only outputs messages to users on standard error;
 - it exits with code 0 unless an internal error occurred (or --check was used).
 

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "reformat-gherkin"
-version = "2.0.2"
+version = "2.0.3"
 description = "Formatter for Gherkin language"
 readme = "README.md"
 authors = ["Duc-Minh Phan <alephvn@gmail.com>"]

reformat_gherkin/cli.py

@@ -23,7 +23,12 @@
     "src",
     nargs=-1,
     type=click.Path(
-        exists=True, file_okay=True, dir_okay=True, readable=True, resolve_path=True
+        exists=True,
+        file_okay=True,
+        dir_okay=True,
+        readable=True,
+        resolve_path=True,
+        allow_dash=True,
     ),
     is_eager=True,
 )
@@ -108,7 +113,7 @@ def main(
     config: Optional[str],
 ) -> None:
     """
-    Reformat the given Gherkin files and all files in the given directories recursively.
+    Reformat the given SRC files and all .feature files in SRC folders. If - is passed as a file, reformat stdin and print the result to stdout.
     """
     if config:
         out(f"Using configuration from {config}.", bold=False, fg="blue")

reformat_gherkin/core.py

@@ -1,6 +1,8 @@
+import sys
 import traceback
+from io import TextIOWrapper
 from pathlib import Path
-from typing import Iterator, Set, Tuple
+from typing import BinaryIO, Iterable, Iterator, Set, Tuple, Union
 
 from .ast_node import GherkinDocument
 from .errors import (
@@ -15,14 +17,14 @@
 from .options import NewlineMode, Options, WriteBackMode
 from .parser import parse
 from .report import Report
-from .utils import decode_bytes, diff, dump_to_file, err
+from .utils import decode_stream, diff, dump_to_file, err, open_stream_or_path
 
 REPORT_URL = "https://github.com/ducminh-phan/reformat-gherkin/issues"
 
 NEWLINE_FROM_OPTION = {NewlineMode.CRLF: "\r\n", NewlineMode.LF: "\n"}
 
 
-def find_sources(src: Tuple[str]) -> Set[Path]:
+def find_sources(src: Iterable[str]) -> Set[Path]:
     sources: Set[Path] = set()
 
     for s in src:
@@ -39,25 +41,48 @@ def find_sources(src: Tuple[str]) -> Set[Path]:
 
 
 def reformat(src: Tuple[str], report: Report, *, options: Options):
-    sources = find_sources(src)
+    use_stdin = "-" in src
+    sources = find_sources(filter((lambda it: it != "-"), src))
 
-    if not sources:
+    if not sources and not use_stdin:
         raise EmptySources
 
+    if use_stdin:
+        changed = reformat_stdin(options=options)
+        report.done("stdin", changed)
+
     for path in sources:
         try:
             changed = reformat_single_file(path, options=options)
-            report.done(path, changed)
+            report.done(str(path), changed)
         except Exception as e:
             report.failed(path, str(e))
 
 
-# noinspection PyTypeChecker
+def reformat_stdin(*, options: Options) -> bool:
+    output = sys.stdout.buffer if options.write_back == WriteBackMode.INPLACE else None
+    return reformat_stream_or_path(
+        sys.stdin.buffer, output, force_write=True, options=options
+    )
+
+
 def reformat_single_file(path: Path, *, options: Options) -> bool:
-    with open(path, "rb") as buf:
-        src_contents, encoding, existing_newline = decode_bytes(buf.read())
+    out_path = path if options.write_back == WriteBackMode.INPLACE else None
+    return reformat_stream_or_path(path, out_path, options=options)
+
+
+def reformat_stream_or_path(
+    in_stream_or_path: Union[BinaryIO, Path],
+    out_stream_or_path: Union[None, BinaryIO, Path],
+    *,
+    force_write: bool = False,
+    options: Options,
+) -> bool:
+    with open_stream_or_path(in_stream_or_path, "rb") as in_stream:
+        src_contents, encoding, existing_newline = decode_stream(in_stream)
 
     newline = NEWLINE_FROM_OPTION.get(options.newline, existing_newline)
+    newline_changed = newline != existing_newline
 
     content_changed = True
     try:
@@ -66,16 +91,18 @@ def reformat_single_file(path: Path, *, options: Options) -> bool:
         content_changed = False
         dst_contents = src_contents
 
-    # We reformat the file if either the content is changed, or the line separators
-    # need to be changed.
-    if not content_changed and newline == existing_newline:
-        return False
+    will_write = force_write or content_changed or newline_changed
 
-    if options.write_back == WriteBackMode.INPLACE:
-        with open(path, "w", encoding=encoding, newline=newline) as f:
-            f.write(dst_contents)
+    if will_write and out_stream_or_path is not None:
+        with open_stream_or_path(out_stream_or_path, "wb") as out_stream:
+            tiow = TextIOWrapper(out_stream, encoding=encoding, newline=newline)
+            tiow.write(dst_contents)
+            # Ensures that the underlying stream is not closed when the
+            # TextIOWrapper is garbage collected. We don't want to close a
+            # stream that was passed to us.
+            tiow.detach()
 
-    return True
+    return content_changed or newline_changed
 
 
 def format_file_contents(src_contents: str, *, options: Options) -> str:

reformat_gherkin/report.py

@@ -15,7 +15,7 @@ class Report:
     same_count: int = 0
     failure_count: int = 0
 
-    def done(self, path: Path, changed: bool) -> None:
+    def done(self, path: str, changed: bool) -> None:
         """Increment the counter for successful reformatting. Write out a message."""
         if changed:
             reformatted = "Would reformat" if self.check else "Reformatted"

reformat_gherkin/utils.py

@@ -3,8 +3,10 @@
 import re
 import tempfile
 import tokenize
+from contextlib import nullcontext
 from functools import lru_cache, partial
-from typing import Tuple
+from pathlib import Path
+from typing import IO, AnyStr, BinaryIO, Tuple, Union
 
 import click
 from wcwidth import wcswidth
@@ -61,21 +63,22 @@ def remove_trailing_spaces(string: str) -> str:
     return "\n".join(line.rstrip() for line in lines)
 
 
-def decode_bytes(src: bytes) -> Tuple[str, str, str]:
+def decode_stream(src: BinaryIO) -> Tuple[str, str, str]:
     """
     Return a tuple of (decoded_contents, encoding, newline).
 
     `newline` is either CRLF or LF but `decoded_contents` is decoded with
     universal newlines (i.e. only contains LF).
     """
-    srcbuf = io.BytesIO(src)
-    encoding, lines = tokenize.detect_encoding(srcbuf.readline)
+    # in case the source is not seekable, read into memory now
+    src = io.BytesIO(src.read())
+    encoding, lines = tokenize.detect_encoding(src.readline)
     if not lines:
         return "", encoding, "\n"
 
     newline = "\r\n" if b"\r\n" == lines[0][-2:] else "\n"
-    srcbuf.seek(0)
-    with io.TextIOWrapper(srcbuf, encoding) as tiow:
+    src.seek(0)
+    with io.TextIOWrapper(src, encoding) as tiow:
         return tiow.read(), encoding, newline
 
 
@@ -98,3 +101,10 @@ def get_display_width(text: str) -> int:
     if width < 0:
         width = len(text)
     return width
+
+
+def open_stream_or_path(stream_or_path: Union[IO[AnyStr], Path], mode: str):
+    if isinstance(stream_or_path, Path):
+        return open(stream_or_path, mode)
+    else:
+        return nullcontext(stream_or_path)

reformat_gherkin/version.py

@@ -1 +1 @@
-__version__ = "2.0.2"
+__version__ = "2.0.3"

tests/conftest.py

@@ -111,4 +111,4 @@ def fin():
 
 @pytest.fixture
 def runner():
-    return CliRunner()
+    return CliRunner(mix_stderr=False)

tests/helpers.py

@@ -23,6 +23,25 @@ def make_options(
     )
 
 
+def options_to_cli_args(options):
+    return [
+        "" if options.write_back == WriteBackMode.INPLACE else "--check",
+        ""
+        if options.step_keyword_alignment == AlignmentMode.NONE
+        else f"--alignment {options.step_keyword_alignment.value}",
+        ""
+        if options.newline == NewlineMode.KEEP
+        else f"--newline {options.newline.value}",
+        "--fast" if options.fast else "--safe",
+        "--single-line-tags"
+        if options.tag_line_mode == TagLineMode.SINGLELINE
+        else "--multi-line-tags",
+        "--use-tabs"
+        if options.indent == "\t"
+        else f"--tab-width {len(options.indent)}",
+    ]
+
+
 OPTIONS = [
     make_options(step_keyword_alignment=alignment_mode)
     for alignment_mode in AlignmentMode

tests/test_cli.py

@@ -1,27 +1,55 @@
 from reformat_gherkin.cli import main
 
+from .helpers import options_to_cli_args
+
 
 def test_cli_success(runner, sources):
     result = runner.invoke(main, [*sources(contain_invalid=False)])
 
+    assert len(result.stdout) == 0
     assert result.exit_code == 0
 
 
+def test_cli_stdin_success(runner, valid_contents):
+    for content, expected, options in valid_contents(
+        with_expected=True, with_options=True
+    ):
+        args = options_to_cli_args(options) + ["-"]
+        args.remove("--check")
+        args = " ".join(args)
+        result = runner.invoke(main, args=args, input=content)
+
+        assert result.stdout == expected
+        assert result.exit_code == 0
+
+
 def test_cli_check(runner, sources):
     result = runner.invoke(main, [*sources(contain_invalid=False), "--check"])
 
+    assert len(result.stdout) == 0
     assert result.exit_code == 1
 
 
+def test_cli_stdin_check(runner, valid_contents):
+    for content in valid_contents():
+        args = " ".join(["-", "--check"])
+        result = runner.invoke(main, args=args, input=content)
+
+        assert len(result.stdout) == 0
+        assert result.exit_code == 1
+
+
 def test_cli_failed(runner, sources):
     result = runner.invoke(main, [*sources(), "--check"])
 
+    assert len(result.stdout) == 0
     assert result.exit_code == 123
 
 
 def test_cli_empty_sources(runner):
     result = runner.invoke(main)
 
+    assert len(result.stdout) == 0
     assert result.exit_code == 0
 
 
@@ -30,8 +58,9 @@ def test_cli_check_with_valid_config(runner, sources):
         main, [*sources(contain_invalid=False, with_config_file=True)]
     )
 
+    assert len(result.stdout) == 0
     assert result.exit_code == 1
-    assert result.stdout.startswith("Using configuration from")
+    assert result.stderr.startswith("Using configuration from")
 
 
 def test_cli_check_with_invalid_config(runner, sources):
@@ -40,8 +69,9 @@ def test_cli_check_with_invalid_config(runner, sources):
         [*sources(contain_invalid=False, with_config_file=True, valid_config=False)],
     )
 
+    assert len(result.stdout) == 0
     assert result.exit_code == 1
-    assert result.stdout.startswith("Error: Could not open file")
+    assert result.stderr.startswith("Error: Could not open file")
 
 
 def test_cli_check_with_empty_config(runner, sources):
@@ -50,4 +80,5 @@ def test_cli_check_with_empty_config(runner, sources):
         [*sources(contain_invalid=False, with_config_file=True, empty_config=True)],
     )
 
+    assert len(result.stdout) == 0
     assert result.exit_code == 0