Merge pull request #984 from carmenbianca/convert-dep5-asterisks

Make sure single asterisks become double asterisks in convert-dep5

src/reuse/convert_dep5.py

@@ -4,6 +4,7 @@
 
 """Logic to convert a .reuse/dep5 file to a REUSE.toml file."""
 
+import re
 import sys
 from argparse import ArgumentParser, Namespace
 from gettext import gettext as _
@@ -15,6 +16,15 @@
 from .global_licensing import REUSE_TOML_VERSION, ReuseDep5
 from .project import Project
 
+_SINGLE_ASTERISK_PATTERN = re.compile(r"(?<!\*)\*(?!\*)")
+
+
+def _convert_asterisk(path: str) -> str:
+    """This solves a semantics difference. A singular asterisk is semantically
+    identical to a double asterisk in REUSE.toml.
+    """
+    return _SINGLE_ASTERISK_PATTERN.sub("**", path)
+
 
 def toml_from_dep5(dep5: Copyright) -> str:
     """Given a Copyright object, return an equivalent REUSE.toml string."""
@@ -27,6 +37,7 @@ def toml_from_dep5(dep5: Copyright) -> str:
         if len(copyrights) == 1:
             copyrights = copyrights[0]
         paths: Union[str, List[str]] = list(paragraph.files)
+        paths = [_convert_asterisk(path) for path in paths]
         if len(paths) == 1:
             paths = paths[0]
         annotations.append(

tests/test_convert_dep5.py

@@ -44,6 +44,38 @@ def test_toml_from_dep5_single_file():
     assert toml_from_dep5(Copyright(text)) == expected
 
 
+def test_toml_from_dep5_asterisks():
+    """Single asterisks get converted to double asterisks. Double asterisks get
+    left alone.
+    """
+    text = StringIO(
+        cleandoc_nl(
+            """
+            Format: https://www.debian.org/doc/packaging-manuals/copyright-format/1.0/
+            Upstream-Name: test
+            Upstream-Contact: test
+            Source: test
+
+            Files: */**/***
+            Copyright: 2018 Jane Doe
+            License: MIT
+            """
+        )
+    )
+    expected = cleandoc_nl(
+        """
+        version = 1
+
+        [[annotations]]
+        path = "**/**/***"
+        precedence = "aggregate"
+        SPDX-FileCopyrightText = "2018 Jane Doe"
+        SPDX-License-Identifier = "MIT"
+        """
+    )
+    assert toml_from_dep5(Copyright(text)) == expected
+
+
 def test_toml_from_dep5_multiple_files_in_paragraph():
     """Correctly convert a DEP5 file with a more files in a paragraph."""
     text = StringIO(
@@ -65,7 +97,7 @@ def test_toml_from_dep5_multiple_files_in_paragraph():
         version = 1
 
         [[annotations]]
-        path = ["hello.txt", "foo*.txt"]
+        path = ["hello.txt", "foo**.txt"]
         precedence = "aggregate"
         SPDX-FileCopyrightText = "2018 Jane Doe"
         SPDX-License-Identifier = "MIT"

tests/test_main.py

@@ -625,7 +625,7 @@ def test_convert_dep5(fake_repository_dep5, stringio):
         version = 1
 
         [[annotations]]
-        path = "doc/*"
+        path = "doc/**"
         precedence = "aggregate"
         SPDX-FileCopyrightText = "2017 Jane Doe"
         SPDX-License-Identifier = "CC0-1.0"