Merge pull request #979 from sin-ha/new-copyright-style

Added new copyright style with spdx and string c as per #948
fsfe · Jul 23, 2024 · 37f342d · 37f342d
2 parents fb847af + 0e9bd7d
commit 37f342d
Show file tree

Hide file tree

Showing 6 changed files with 120 additions and 15 deletions.
diff --git a/changelog.d/added/new-prefixes.md b/changelog.d/added/new-prefixes.md
@@ -0,0 +1,2 @@
+- Added new copyright prefixes `spdx-string`, `spdx-string-c`, and
+  `spdx-string-symbol`. (#979)
diff --git a/changelog.d/fixed/prefix-merge.md b/changelog.d/fixed/prefix-merge.md
@@ -0,0 +1,2 @@
+- `reuse annotate --merge-copyrights` now works more reliably with copyright
+  prefixes. This still needs some work, though. (#979)
diff --git a/docs/man/reuse-annotate.rst b/docs/man/reuse-annotate.rst
@@ -84,13 +84,16 @@ Other options
 
   .. code-block::
 
-    spdx:           SPDX-FileCopyrightText: <year> <statement>
-    spdx-c:         SPDX-FileCopyrightText: (C) <year> <statement>
-    spdx-symbol:    SPDX-FileCopyrightText: © <year> <statement>
-    string:         Copyright <year> <statement>
-    string-c:       Copyright (C) <year> <statement>
-    string-symbol:  Copyright © <year> <statement>
-    symbol:         © <year> <statement>
+    spdx:               SPDX-FileCopyrightText: <year> <statement>
+    spdx-c:             SPDX-FileCopyrightText: (C) <year> <statement>
+    spdx-symbol:        SPDX-FileCopyrightText: © <year> <statement>
+    spdx-string:        SPDX-FileCopyrightText: Copyright <year> <statement>
+    spdx-string-c:      SPDX-FileCopyrightText: Copyright (C) <year> <statement>
+    spdx-string-symbol: SPDX-FileCopyrightText: Copyright © <year> <statement>
+    string:             Copyright <year> <statement>
+    string-c:           Copyright (C) <year> <statement>
+    string-symbol:      Copyright © <year> <statement>
+    symbol:             © <year> <statement>
 
 .. option:: -t, --template TEMPLATE
 

diff --git a/src/reuse/_util.py b/src/reuse/_util.py
@@ -116,12 +116,13 @@
 
 _COPYRIGHT_PATTERNS = [
     re.compile(
-        r"(?P<copyright>(?P<prefix>SPDX-(File|Snippet)CopyrightText:)\s+"
+        r"(?P<copyright>(?P<prefix>SPDX-(File|Snippet)CopyrightText:"
+        r"(\s(\([Cc]\)|©|Copyright(\s(©|\([Cc]\)))?))?)\s+"
         r"((?P<year>\d{4} ?- ?\d{4}|\d{4}),?\s+)?"
         r"(?P<statement>.*?))" + _END_PATTERN
     ),
     re.compile(
-        r"(?P<copyright>(?P<prefix>Copyright(\s?\([cC]\))?)\s+"
+        r"(?P<copyright>(?P<prefix>Copyright(\s(\([Cc]\)|©))?)\s+"
         r"((?P<year>\d{4} ?- ?\d{4}|\d{4}),?\s+)?"
         r"(?P<statement>.*?))" + _END_PATTERN
     ),
@@ -134,6 +135,9 @@
 _COPYRIGHT_PREFIXES = {
     "spdx": "SPDX-FileCopyrightText:",
     "spdx-c": "SPDX-FileCopyrightText: (C)",
+    "spdx-string-c": "SPDX-FileCopyrightText: Copyright (C)",
+    "spdx-string": "SPDX-FileCopyrightText: Copyright",
+    "spdx-string-symbol": "SPDX-FileCopyrightText: Copyright ©",
     "spdx-symbol": "SPDX-FileCopyrightText: ©",
     "string": "Copyright",
     "string-c": "Copyright (C)",
@@ -313,6 +317,7 @@ def merge_copyright_lines(copyright_lines: Set[str]) -> Set[str]:
                         "prefix": match.groupdict()["prefix"],
                     }
                 )
+                break
 
     copyright_out = set()
     for line_info in copyright_in:

diff --git a/tests/test_main_annotate_merge.py b/tests/test_main_annotate_merge.py
@@ -8,6 +8,7 @@
 from inspect import cleandoc
 
 from reuse._main import main
+from reuse._util import _COPYRIGHT_PREFIXES
 
 # pylint: disable=unused-argument
 
@@ -191,4 +192,69 @@ def test_annotate_merge_copyrights_no_year_in_existing(
     )
 
 
+def test_annotate_merge_copyrights_all_prefixes(
+    fake_repository, stringio, mock_date_today
+):
+    """Test that merging works for all copyright prefixes."""
+    # TODO: there should probably also be a test for mixing copyright prefixes,
+    # but this behaviour is really unpredictable to me at the moment, and the
+    # whole copyright-line-as-string thing needs overhauling.
+    simple_file = fake_repository / "foo.py"
+    for copyright_prefix, copyright_string in _COPYRIGHT_PREFIXES.items():
+        simple_file.write_text("pass")
+        result = main(
+            [
+                "annotate",
+                "--year",
+                "2016",
+                "--license",
+                "GPL-3.0-or-later",
+                "--copyright",
+                "Jane Doe",
+                "--copyright-style",
+                copyright_prefix,
+                "--merge-copyrights",
+                "foo.py",
+            ],
+            out=stringio,
+        )
+        assert result == 0
+        assert simple_file.read_text(encoding="utf-8") == cleandoc(
+            f"""
+                # {copyright_string} 2016 Jane Doe
+                #
+                # SPDX-License-Identifier: GPL-3.0-or-later
+
+                pass
+                """
+        )
+
+        result = main(
+            [
+                "annotate",
+                "--year",
+                "2018",
+                "--license",
+                "GPL-3.0-or-later",
+                "--copyright",
+                "Jane Doe",
+                "--copyright-style",
+                copyright_prefix,
+                "--merge-copyrights",
+                "foo.py",
+            ],
+            out=stringio,
+        )
+        assert result == 0
+        assert simple_file.read_text(encoding="utf-8") == cleandoc(
+            f"""
+                # {copyright_string} 2016 - 2018 Jane Doe
+                #
+                # SPDX-License-Identifier: GPL-3.0-or-later
+
+                pass
+                """
+        )
+
+
 # REUSE-IgnoreEnd
diff --git a/tests/test_util.py b/tests/test_util.py
@@ -124,12 +124,15 @@ def test_extract_copyright_variations():
     """There are multiple ways to declare copyright. All should be detected."""
     text = cleandoc(
         """
-        SPDX-FileCopyrightText: 2019 Jane Doe
-        SPDX-FileCopyrightText: © 2019 Jane Doe
-        © 2019 Jane Doe
-        Copyright © 2019 Jane Doe
-        Copyright 2019 Jane Doe
-        Copyright (C) 2019 Jane Doe
+        SPDX-FileCopyrightText: 2019 spdx
+        SPDX-FileCopyrightText: (C) 2019 spdx-c
+        SPDX-FileCopyrightText: © 2019 spdx-symbol
+        SPDX-FileCopyrightText: Copyright (C) 2019 spdx-string-c
+        SPDX-FileCopyrightText: Copyright © 2019 spdx-string-symbol
+        Copyright 2019 string
+        Copyright (C) 2019 string-c
+        Copyright © 2019 string-symbol
+        © 2019 symbol
         """
     )
 
@@ -428,6 +431,30 @@ def test_make_copyright_line_prefix_string_c_year():
     assert statement == "Copyright (C) 2019 hello"
 
 
+def test_make_copyright_line_prefix_spdx_string_c_year():
+    """Given a simple statement, prefix and a year, make it a copyright line."""
+    statement = _util.make_copyright_line(
+        "hello", year=2019, copyright_prefix="spdx-string-c"
+    )
+    assert statement == "SPDX-FileCopyrightText: Copyright (C) 2019 hello"
+
+
+def test_make_copyright_line_prefix_spdx_string_year():
+    """Given a simple statement, prefix and a year, make it a copyright line."""
+    statement = _util.make_copyright_line(
+        "hello", year=2019, copyright_prefix="spdx-string"
+    )
+    assert statement == "SPDX-FileCopyrightText: Copyright 2019 hello"
+
+
+def test_make_copyright_line_prefix_spdx_string_symbol_year():
+    """Given a simple statement, prefix and a year, make it a copyright line."""
+    statement = _util.make_copyright_line(
+        "hello", year=2019, copyright_prefix="spdx-string-symbol"
+    )
+    assert statement == "SPDX-FileCopyrightText: Copyright © 2019 hello"
+
+
 def test_make_copyright_line_prefix_string_symbol_year():
     """Given a simple statement, prefix and a year, make it a copyright line."""
     statement = _util.make_copyright_line(