improve encoding handling for `setup.cfg`

Support the same mechanism as for Python sources for declaring
the encoding to be used when reading `setup.cfg` (see PEP 263),
and return the results of reading it as Unicode.

Fix #1062 and #1136.

setuptools/__init__.py

@@ -4,9 +4,12 @@ import os
 import functools
 import distutils.core
 import distutils.filelist
+import re
+from distutils.errors import DistutilsOptionError
 from distutils.util import convert_path
 from fnmatch import fnmatchcase
 
+from setuptools.extern.six import string_types
 from setuptools.extern.six.moves import filter, map
 
 import setuptools.version
@@ -127,6 +130,37 @@ class Command(_Command):
         _Command.__init__(self, dist)
         vars(self).update(kw)
 
+    def _ensure_stringlike(self, option, what, default=None):
+        val = getattr(self, option)
+        if val is None:
+            setattr(self, option, default)
+            return default
+        elif not isinstance(val, string_types):
+            raise DistutilsOptionError("'%s' must be a %s (got `%s`)"
+                                       % (option, what, val))
+        return val
+
+    def ensure_string_list(self, option):
+        r"""Ensure that 'option' is a list of strings.  If 'option' is
+        currently a string, we split it either on /,\s*/ or /\s+/, so
+        "foo bar baz", "foo,bar,baz", and "foo,   bar baz" all become
+        ["foo", "bar", "baz"].
+        """
+        val = getattr(self, option)
+        if val is None:
+            return
+        elif isinstance(val, string_types):
+            setattr(self, option, re.split(r',\s*|\s+', val))
+        else:
+            if isinstance(val, list):
+                ok = all(isinstance(v, string_types) for v in val)
+            else:
+                ok = False
+            if not ok:
+                raise DistutilsOptionError(
+                      "'%s' must be a list of strings (got %r)"
+                      % (option, val))
+
     def reinitialize_command(self, command, reinit_subcommands=0, **kw):
         cmd = _Command.reinitialize_command(self, command, reinit_subcommands)
         vars(cmd).update(kw)

setuptools/dist.py

@@ -432,7 +432,7 @@ class Distribution(Distribution_parse_config_files, _Distribution):
         and loads configuration.
 
         """
-        _Distribution.parse_config_files(self, filenames=filenames)
+        Distribution_parse_config_files.parse_config_files(self, filenames=filenames)
 
         parse_configuration(self, self.command_options)
         self._finalize_requires()

setuptools/py36compat.py

+import io
+import re
 import sys
 from distutils.errors import DistutilsOptionError
 from distutils.util import strtobool
 from distutils.debug import DEBUG
+from setuptools.extern import six
+
+
+CODING_RE = re.compile(br'^[ \t\f]*#.*?coding[:=][ \t]*([-\w.]+)')
+
+def detect_encoding(fp):
+    first_line = fp.readline()
+    fp.seek(0)
+    m = CODING_RE.match(first_line)
+    if m is None:
+        return None
+    return m.group(1).decode('ascii')
 
 
 class Distribution_parse_config_files:
@@ -13,10 +27,10 @@ class Distribution_parse_config_files:
     as implemented in distutils.
     """
     def parse_config_files(self, filenames=None):
-        from configparser import ConfigParser
+        from setuptools.extern.six.moves.configparser import ConfigParser
 
         # Ignore install directory options if we have a venv
-        if sys.prefix != sys.base_prefix:
+        if six.PY3 and sys.prefix != sys.base_prefix:
             ignore_options = [
                 'install-base', 'install-platbase', 'install-lib',
                 'install-platlib', 'install-purelib', 'install-headers',
@@ -33,11 +47,16 @@ class Distribution_parse_config_files:
         if DEBUG:
             self.announce("Distribution.parse_config_files():")
 
-        parser = ConfigParser(interpolation=None)
+        parser = ConfigParser()
         for filename in filenames:
-            if DEBUG:
-                self.announce("  reading %s" % filename)
-            parser.read(filename)
+            with io.open(filename, 'rb') as fp:
+                encoding = detect_encoding(fp)
+                if DEBUG:
+                    self.announce("  reading %s [%s]" % (
+                        filename, encoding or 'locale')
+                    )
+                reader = io.TextIOWrapper(fp, encoding=encoding)
+                (parser.read_file if six.PY3 else parser.readfp)(reader)
             for section in parser.sections():
                 options = parser.options(section)
                 opt_dict = self.get_option_dict(section)
@@ -69,12 +88,6 @@ class Distribution_parse_config_files:
                     raise DistutilsOptionError(msg)
 
 
-if sys.version_info < (3,):
-    # Python 2 behavior is sufficient
-    class Distribution_parse_config_files:
-        pass
-
-
 if False:
     # When updated behavior is available upstream,
     # disable override here.

setuptools/tests/test_config.py

+# -*- coding: UTF-8 -*-
+from __future__ import unicode_literals
+
 import contextlib
 import pytest
 from distutils.errors import DistutilsOptionError, DistutilsFileError
 from setuptools.dist import Distribution
 from setuptools.config import ConfigHandler, read_configuration
 from setuptools.extern.six.moves.configparser import InterpolationMissingOptionError
+from setuptools.tests import is_ascii
 
 
 class ErrConfigHandler(ConfigHandler):
@@ -17,7 +21,7 @@ def make_package_dir(name, base_dir):
     return dir_package, init_file
 
 
-def fake_env(tmpdir, setup_cfg, setup_py=None):
+def fake_env(tmpdir, setup_cfg, setup_py=None, encoding='ascii'):
 
     if setup_py is None:
         setup_py = (
@@ -27,7 +31,7 @@ def fake_env(tmpdir, setup_cfg, setup_py=None):
 
     tmpdir.join('setup.py').write(setup_py)
     config = tmpdir.join('setup.cfg')
-    config.write(setup_cfg)
+    config.write(setup_cfg.encode(encoding), mode='wb')
 
     package_dir, init_file = make_package_dir('fake_package', tmpdir)
 
@@ -317,6 +321,63 @@ class TestMetadata:
             with get_dist(tmpdir):
                 pass
 
+    skip_if_not_ascii = pytest.mark.skipif(not is_ascii, reason='Test not supported with this locale')
+
+    @skip_if_not_ascii
+    def test_non_ascii_1(self, tmpdir):
+        fake_env(
+            tmpdir,
+            '[metadata]\n'
+            'description = éàïôñ\n',
+            encoding='utf-8'
+        )
+        with pytest.raises(UnicodeDecodeError):
+            with get_dist(tmpdir):
+                pass
+
+    def test_non_ascii_2(self, tmpdir):
+        fake_env(
+            tmpdir,
+            '# -*- coding: invalid\n'
+        )
+        with pytest.raises(LookupError):
+            with get_dist(tmpdir):
+                pass
+
+    def test_non_ascii_3(self, tmpdir):
+        fake_env(
+            tmpdir,
+            '\n'
+            '# -*- coding: invalid\n'
+        )
+        with get_dist(tmpdir):
+            pass
+
+    @skip_if_not_ascii
+    def test_non_ascii_4(self, tmpdir):
+        fake_env(
+            tmpdir,
+            '# -*- coding: utf-8\n'
+            '[metadata]\n'
+            'description = éàïôñ\n',
+            encoding='utf-8'
+        )
+        with get_dist(tmpdir) as dist:
+            assert dist.metadata.description == 'éàïôñ'
+
+    @skip_if_not_ascii
+    def test_non_ascii_5(self, tmpdir):
+        fake_env(
+            tmpdir,
+            '# vim: set fileencoding=iso-8859-15 :\n'
+            '[metadata]\n'
+            'description = éàïôñ\n',
+            encoding='iso-8859-15'
+        )
+        with get_dist(tmpdir) as dist:
+            assert dist.metadata.description == 'éàïôñ'
+
+
 class TestOptions:
 
     def test_basic(self, tmpdir):

setuptools/tests/test_egg_info.py

@@ -497,3 +497,49 @@ class TestEggInfo(object):
         # expect exactly one result
         result, = results
         return result
+
+    EGG_INFO_TESTS = (
+        # Check for issue #1136: invalid string type when
+        # reading declarative `setup.cfg` under Python 2.
+        {
+            'setup.py': DALS(
+                """
+                from setuptools import setup
+                setup(
+                    name="foo",
+                )
+                """),
+            'setup.cfg': DALS(
+                """
+                [options]
+                package_dir =
+                    = src
+                """),
+            'src': {},
+        },
+        # Check Unicode can be used in `setup.py` under Python 2.
+        {
+            'setup.py': DALS(
+                """
+                # -*- coding: utf-8 -*-
+                from __future__ import unicode_literals
+                from setuptools import setup, find_packages
+                setup(
+                    name="foo",
+                    package_dir={'': 'src'},
+                )
+                """),
+            'src': {},
+        }
+    )
+
+    @pytest.mark.parametrize('package_files', EGG_INFO_TESTS)
+    def test_egg_info(self, tmpdir_cwd, env, package_files):
+        """
+        """
+        build_files(package_files)
+        code, data = environment.run_setup_py(
+            cmd=['egg_info'],
+            data_stream=1,
+        )
+        assert not code, data