Factor out all_files into its own module covered_files

Signed-off-by: Carmen Bianca BAKKER <[email protected]>

src/reuse/covered_files.py

@@ -0,0 +1,120 @@
+# SPDX-FileCopyrightText: 2017 Free Software Foundation Europe e.V. <https://fsfe.org>
+#
+# SPDX-License-Identifier: GPL-3.0-or-later
+
+"""The REUSE Specification has a concept called Covered Files; files which must
+contain licensing information. Some files in a project are not Covered Files,
+and thus needn't contain licensing information. This module contains all that
+logic.
+"""
+
+import contextlib
+import logging
+import os
+from pathlib import Path
+from typing import Generator, Optional
+
+from . import (
+    _IGNORE_DIR_PATTERNS,
+    _IGNORE_FILE_PATTERNS,
+    _IGNORE_MESON_PARENT_DIR_PATTERNS,
+)
+from ._util import StrPath
+from .vcs import VCSStrategy
+
+_LOGGER = logging.getLogger(__name__)
+
+
+def is_path_ignored(
+    path: Path,
+    include_submodules: bool = False,
+    include_meson_subprojects: bool = False,
+    vcs_strategy: Optional[VCSStrategy] = None,
+) -> bool:
+    """Is *path* ignored by some mechanism?"""
+    # pylint: disable=too-many-return-statements,too-many-branches
+    name = path.name
+    parent_parts = path.parent.parts
+    parent_dir = parent_parts[-1] if len(parent_parts) > 0 else ""
+
+    if path.is_symlink():
+        _LOGGER.debug("skipping symlink '%s'", path)
+        return True
+
+    if path.is_file():
+        for pattern in _IGNORE_FILE_PATTERNS:
+            if pattern.match(name):
+                return True
+        # Suppressing this error because I simply don't want to deal
+        # with that here.
+        with contextlib.suppress(OSError):
+            if path.stat().st_size == 0:
+                _LOGGER.debug("skipping 0-sized file '%s'", path)
+                return True
+
+    elif path.is_dir():
+        for pattern in _IGNORE_DIR_PATTERNS:
+            if pattern.match(name):
+                return True
+        if not include_meson_subprojects:
+            for pattern in _IGNORE_MESON_PARENT_DIR_PATTERNS:
+                if pattern.match(parent_dir):
+                    _LOGGER.info(
+                        "ignoring '%s' because it is a Meson subproject", path
+                    )
+                    return True
+        if (
+            not include_submodules
+            and vcs_strategy
+            and vcs_strategy.is_submodule(path)
+        ):
+            _LOGGER.info("ignoring '%s' because it is a submodule", path)
+            return True
+
+    if vcs_strategy and vcs_strategy.is_ignored(path):
+        return True
+
+    return False
+
+
+def all_files(
+    directory: StrPath,
+    include_submodules: bool = False,
+    include_meson_subprojects: bool = False,
+    vcs_strategy: Optional[VCSStrategy] = None,
+) -> Generator[Path, None, None]:
+    """Yield all Covered Files in *directory* and its subdirectories according
+    to the REUSE Specification.
+    """
+    directory = Path(directory)
+
+    for root_str, dirs, files in os.walk(directory):
+        root = Path(root_str)
+        _LOGGER.debug("currently walking in '%s'", root)
+
+        # Don't walk ignored directories
+        for dir_ in list(dirs):
+            the_dir = root / dir_
+            if is_path_ignored(
+                the_dir,
+                include_submodules=include_submodules,
+                include_meson_subprojects=include_meson_subprojects,
+                vcs_strategy=vcs_strategy,
+            ):
+                _LOGGER.debug("ignoring '%s'", the_dir)
+                dirs.remove(dir_)
+
+        # Filter files.
+        for file_ in files:
+            the_file = root / file_
+            if is_path_ignored(
+                the_file,
+                include_submodules=include_submodules,
+                include_meson_subprojects=include_meson_subprojects,
+                vcs_strategy=vcs_strategy,
+            ):
+                _LOGGER.debug("ignoring '%s'", the_file)
+                continue
+
+            _LOGGER.debug("yielding '%s'", the_file)
+            yield the_file

src/reuse/project.py

@@ -29,13 +29,7 @@
 
 from binaryornot.check import is_binary
 
-from . import (
-    _IGNORE_DIR_PATTERNS,
-    _IGNORE_FILE_PATTERNS,
-    _IGNORE_MESON_PARENT_DIR_PATTERNS,
-    IdentifierNotFound,
-    ReuseInfo,
-)
+from . import IdentifierNotFound, ReuseInfo
 from ._licenses import EXCEPTION_MAP, LICENSE_MAP
 from ._util import (
     _LICENSEREF_PATTERN,
@@ -44,6 +38,7 @@
     relative_from_root,
     reuse_info_of_file,
 )
+from .covered_files import all_files
 from .global_licensing import (
     GlobalLicensing,
     NestedReuseTOML,
@@ -431,98 +426,3 @@ def _detect_vcs_strategy(cls, root: StrPath) -> VCSStrategy:
             ).format(root)
         )
         return VCSStrategyNone(root)
-
-
-def is_path_ignored(
-    path: Path,
-    include_submodules: bool = False,
-    include_meson_subprojects: bool = False,
-    vcs_strategy: Optional[VCSStrategy] = None,
-) -> bool:
-    """Is *path* ignored by some mechanism?"""
-    # pylint: disable=too-many-return-statements,too-many-branches
-    name = path.name
-    parent_parts = path.parent.parts
-    parent_dir = parent_parts[-1] if len(parent_parts) > 0 else ""
-
-    if path.is_symlink():
-        _LOGGER.debug("skipping symlink '%s'", path)
-        return True
-
-    if path.is_file():
-        for pattern in _IGNORE_FILE_PATTERNS:
-            if pattern.match(name):
-                return True
-        # Suppressing this error because I simply don't want to deal
-        # with that here.
-        with contextlib.suppress(OSError):
-            if path.stat().st_size == 0:
-                _LOGGER.debug("skipping 0-sized file '%s'", path)
-                return True
-
-    elif path.is_dir():
-        for pattern in _IGNORE_DIR_PATTERNS:
-            if pattern.match(name):
-                return True
-        if not include_meson_subprojects:
-            for pattern in _IGNORE_MESON_PARENT_DIR_PATTERNS:
-                if pattern.match(parent_dir):
-                    _LOGGER.info(
-                        "ignoring '%s' because it is a Meson subproject", path
-                    )
-                    return True
-        if (
-            not include_submodules
-            and vcs_strategy
-            and vcs_strategy.is_submodule(path)
-        ):
-            _LOGGER.info("ignoring '%s' because it is a submodule", path)
-            return True
-
-    if vcs_strategy and vcs_strategy.is_ignored(path):
-        return True
-
-    return False
-
-
-def all_files(
-    directory: StrPath,
-    include_submodules: bool = False,
-    include_meson_subprojects: bool = False,
-    vcs_strategy: Optional[VCSStrategy] = None,
-) -> Generator[Path, None, None]:
-    """Yield all Covered Files in *directory* and its subdirectories according
-    to the REUSE Specification.
-    """
-    directory = Path(directory)
-
-    for root_str, dirs, files in os.walk(directory):
-        root = Path(root_str)
-        _LOGGER.debug("currently walking in '%s'", root)
-
-        # Don't walk ignored directories
-        for dir_ in list(dirs):
-            the_dir = root / dir_
-            if is_path_ignored(
-                the_dir,
-                include_submodules=include_submodules,
-                include_meson_subprojects=include_meson_subprojects,
-                vcs_strategy=vcs_strategy,
-            ):
-                _LOGGER.debug("ignoring '%s'", the_dir)
-                dirs.remove(dir_)
-
-        # Filter files.
-        for file_ in files:
-            the_file = root / file_
-            if is_path_ignored(
-                the_file,
-                include_submodules=include_submodules,
-                include_meson_subprojects=include_meson_subprojects,
-                vcs_strategy=vcs_strategy,
-            ):
-                _LOGGER.debug("ignoring '%s'", the_file)
-                continue
-
-            _LOGGER.debug("yielding '%s'", the_file)
-            yield the_file

tests/conftest.py

@@ -58,6 +58,8 @@
     sys.implementation.name != "cpython", reason="only CPython supported"
 )
 git = pytest.mark.skipif(not GIT_EXE, reason="requires git")
+hg = pytest.mark.skipif(not HG_EXE, reason="requires mercurial")
+pijul = pytest.mark.skipif(not PIJUL_EXE, reason="requires pijul")
 no_root = pytest.mark.xfail(is_root, reason="fails when user is root")
 posix = pytest.mark.skipif(not is_posix, reason="Windows not supported")