From 5806de1ebe998ccba452c6aa6824d9b6ada8f4ca Mon Sep 17 00:00:00 2001
From: "Jason R. Coombs" <jaraco@jaraco.com>
Date: Mon, 29 Mar 2021 20:43:53 -0400
Subject: [PATCH 1/4] Unify section parser behavior. Removes reliance on regex
 (ref #292). Fixes #295.

---
 importlib_metadata/__init__.py | 72 ++++++++++++++--------------------
 1 file changed, 29 insertions(+), 43 deletions(-)

diff --git a/importlib_metadata/__init__.py b/importlib_metadata/__init__.py
index c83e3e9d..b545e1c2 100644
--- a/importlib_metadata/__init__.py
+++ b/importlib_metadata/__init__.py
@@ -65,20 +65,25 @@ class Sectioned:
     A simple entry point config parser for performance
 
     >>> res = Sectioned.get_sections(Sectioned._sample)
-    >>> sec, values = next(res)
+    >>> sec, pair = next(res)
     >>> sec
     'sec1'
-    >>> [(key, value) for key, value in values]
-    [('a', '1'), ('b', '2')]
-    >>> sec, values = next(res)
+    >>> tuple(pair)
+    ('a', '1')
+    >>> sec, pair = next(res)
+    >>> tuple(pair)
+    ('b', '2')
+    >>> sec, pair = next(res)
     >>> sec
     'sec2'
-    >>> [(key, value) for key, value in values]
-    [('a', '2')]
+    >>> tuple(pair)
+    ('a', '2')
     >>> list(res)
     []
     """
 
+    Pair = collections.namedtuple('Pair', 'name value')
+
     _sample = textwrap.dedent(
         """
         [sec1]
@@ -91,25 +96,25 @@ class Sectioned:
         """
     ).lstrip()
 
-    def __init__(self):
-        self.section = None
-
-    def __call__(self, line):
-        if line.startswith('[') and line.endswith(']'):
-            # new section
-            self.section = line.strip('[]')
-            return
-        return self.section
-
     @classmethod
     def get_sections(cls, text):
-        lines = filter(cls.valid, map(str.strip, text.splitlines()))
         return (
-            (section, map(cls.parse_value, values))
-            for section, values in itertools.groupby(lines, cls())
-            if section is not None
+            (section.name, cls.parse_value(section.value))
+            for section in cls.read(text, filter_=cls.valid)
+            if section.name is not None
         )
 
+    @staticmethod
+    def read(text, filter_=None):
+        lines = filter(filter_, map(str.strip, text.splitlines()))
+        name = None
+        for value in lines:
+            section_match = value.startswith('[') and value.endswith(']')
+            if section_match:
+                name = value.strip('[]')
+                continue
+            yield Sectioned.Pair(name, value)
+
     @staticmethod
     def valid(line):
         return line and not line.startswith('#')
@@ -256,8 +261,7 @@ def _from_text(cls, text):
     def _parse_groups(text):
         return (
             (name, value, section)
-            for section, values in Sectioned.get_sections(text)
-            for name, value in values
+            for section, (name, value) in Sectioned.get_sections(text)
         )
 
 
@@ -573,24 +577,7 @@ def _read_egg_info_reqs(self):
 
     @classmethod
     def _deps_from_requires_text(cls, source):
-        section_pairs = cls._read_sections(source.splitlines())
-        sections = {
-            section: list(map(operator.itemgetter('line'), results))
-            for section, results in itertools.groupby(
-                section_pairs, operator.itemgetter('section')
-            )
-        }
-        return cls._convert_egg_info_reqs_to_simple_reqs(sections)
-
-    @staticmethod
-    def _read_sections(lines):
-        section = None
-        for line in filter(None, lines):
-            section_match = re.match(r'\[(.*)\]$', line)
-            if section_match:
-                section = section_match.group(1)
-                continue
-            yield locals()
+        return cls._convert_egg_info_reqs_to_simple_reqs(Sectioned.read(source))
 
     @staticmethod
     def _convert_egg_info_reqs_to_simple_reqs(sections):
@@ -615,9 +602,8 @@ def parse_condition(section):
             conditions = list(filter(None, [markers, make_condition(extra)]))
             return '; ' + ' and '.join(conditions) if conditions else ''
 
-        for section, deps in sections.items():
-            for dep in deps:
-                yield dep + parse_condition(section)
+        for section in sections:
+            yield section.value + parse_condition(section.name)
 
 
 class DistributionFinder(MetaPathFinder):

From bf6a3b13b3eec27e48abed353761a29676d6a0ff Mon Sep 17 00:00:00 2001
From: "Jason R. Coombs" <jaraco@jaraco.com>
Date: Mon, 29 Mar 2021 21:11:53 -0400
Subject: [PATCH 2/4] Use Pair in other places and extract it to _collections.

---
 importlib_metadata/__init__.py     | 36 +++++++++++++-----------------
 importlib_metadata/_collections.py |  6 +++++
 2 files changed, 21 insertions(+), 21 deletions(-)

diff --git a/importlib_metadata/__init__.py b/importlib_metadata/__init__.py
index b545e1c2..5abc1e9e 100644
--- a/importlib_metadata/__init__.py
+++ b/importlib_metadata/__init__.py
@@ -15,7 +15,7 @@
 import contextlib
 import collections
 
-from ._collections import FreezableDefaultDict
+from ._collections import FreezableDefaultDict, Pair
 from ._compat import (
     NullFinder,
     Protocol,
@@ -64,26 +64,24 @@ class Sectioned:
     """
     A simple entry point config parser for performance
 
-    >>> res = Sectioned.get_sections(Sectioned._sample)
-    >>> sec, pair = next(res)
-    >>> sec
+    >>> res = Sectioned.section_pairs(Sectioned._sample)
+    >>> item = next(res)
+    >>> item.name
     'sec1'
-    >>> tuple(pair)
+    >>> tuple(item.value)
     ('a', '1')
-    >>> sec, pair = next(res)
-    >>> tuple(pair)
+    >>> item = next(res)
+    >>> tuple(item.value)
     ('b', '2')
-    >>> sec, pair = next(res)
-    >>> sec
+    >>> item = next(res)
+    >>> item.name
     'sec2'
-    >>> tuple(pair)
+    >>> tuple(item.value)
     ('a', '2')
     >>> list(res)
     []
     """
 
-    Pair = collections.namedtuple('Pair', 'name value')
-
     _sample = textwrap.dedent(
         """
         [sec1]
@@ -97,9 +95,9 @@ class Sectioned:
     ).lstrip()
 
     @classmethod
-    def get_sections(cls, text):
+    def section_pairs(cls, text):
         return (
-            (section.name, cls.parse_value(section.value))
+            section._replace(value=Pair.parse(section.value))
             for section in cls.read(text, filter_=cls.valid)
             if section.name is not None
         )
@@ -113,16 +111,12 @@ def read(text, filter_=None):
             if section_match:
                 name = value.strip('[]')
                 continue
-            yield Sectioned.Pair(name, value)
+            yield Pair(name, value)
 
     @staticmethod
     def valid(line):
         return line and not line.startswith('#')
 
-    @staticmethod
-    def parse_value(line):
-        return map(str.strip, line.split("=", 1))
-
 
 class EntryPoint(
     PyPy_repr, collections.namedtuple('EntryPointBase', 'name value group')
@@ -260,8 +254,8 @@ def _from_text(cls, text):
     @staticmethod
     def _parse_groups(text):
         return (
-            (name, value, section)
-            for section, (name, value) in Sectioned.get_sections(text)
+            (item.value.name, item.value.value, item.name)
+            for item in Sectioned.section_pairs(text)
         )
 
 
diff --git a/importlib_metadata/_collections.py b/importlib_metadata/_collections.py
index 6aa17c84..cf0954e1 100644
--- a/importlib_metadata/_collections.py
+++ b/importlib_metadata/_collections.py
@@ -22,3 +22,9 @@ def __missing__(self, key):
 
     def freeze(self):
         self._frozen = lambda key: self.default_factory()
+
+
+class Pair(collections.namedtuple('Pair', 'name value')):
+    @classmethod
+    def parse(cls, text):
+        return cls(*map(str.strip, text.split("=", 1)))

From c891d5065be9948bc593808884b85c3ad364dbfd Mon Sep 17 00:00:00 2001
From: "Jason R. Coombs" <jaraco@jaraco.com>
Date: Mon, 29 Mar 2021 21:26:51 -0400
Subject: [PATCH 3/4] Expand and simplify Sectioned doctest.

---
 importlib_metadata/__init__.py | 19 +++++++++++++------
 1 file changed, 13 insertions(+), 6 deletions(-)

diff --git a/importlib_metadata/__init__.py b/importlib_metadata/__init__.py
index 5abc1e9e..7c5eb2c7 100644
--- a/importlib_metadata/__init__.py
+++ b/importlib_metadata/__init__.py
@@ -64,20 +64,27 @@ class Sectioned:
     """
     A simple entry point config parser for performance
 
+    >>> for item in Sectioned.read(Sectioned._sample):
+    ...     print(item)
+    Pair(name='sec1', value='# comments ignored')
+    Pair(name='sec1', value='a = 1')
+    Pair(name='sec1', value='b = 2')
+    Pair(name='sec2', value='a = 2')
+
     >>> res = Sectioned.section_pairs(Sectioned._sample)
     >>> item = next(res)
     >>> item.name
     'sec1'
-    >>> tuple(item.value)
-    ('a', '1')
+    >>> item.value
+    Pair(name='a', value='1')
     >>> item = next(res)
-    >>> tuple(item.value)
-    ('b', '2')
+    >>> item.value
+    Pair(name='b', value='2')
     >>> item = next(res)
     >>> item.name
     'sec2'
-    >>> tuple(item.value)
-    ('a', '2')
+    >>> item.value
+    Pair(name='a', value='2')
     >>> list(res)
     []
     """

From 244fc482479925b75851194989518d97061a1b28 Mon Sep 17 00:00:00 2001
From: "Jason R. Coombs" <jaraco@jaraco.com>
Date: Mon, 29 Mar 2021 21:36:36 -0400
Subject: [PATCH 4/4] Update changelog.

---
 CHANGES.rst | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/CHANGES.rst b/CHANGES.rst
index 8427a258..696da526 100644
--- a/CHANGES.rst
+++ b/CHANGES.rst
@@ -1,3 +1,8 @@
+v3.10.0
+=======
+
+* #295: Internal refactoring to unify section parsing logic.
+
 v3.9.1
 ======