Old get_module_source API restored, new version moved to ModuleAnalyzer class, tests updated

4 files changed, 81 insertions, 24 deletions

diff --git a/sphinx/pycode/__init__.py b/sphinx/pycode/__init__.py
index 483eed432..2e077a618 100644
--- a/sphinx/pycode/__init__.py
+++ b/sphinx/pycode/__init__.py
@@ -11,8 +11,9 @@
 import re
 from io import StringIO
 from os import path
-from typing import Any, Dict, IO, List, Tuple
+from typing import Any, Dict, IO, List, Tuple, Optional
 from zipfile import ZipFile
+from importlib import import_module
 
 from sphinx.errors import PycodeError
 from sphinx.pycode.parser import Parser
@@ -23,6 +24,55 @@ class ModuleAnalyzer:
     # cache for analyzer objects -- caches both by module and file name
     cache = {}  # type: Dict[Tuple[str, str], Any]
 
+    @staticmethod
+    def get_module_source(modname: str) -> Tuple[Optional[str], Optional[str]]:
+        """Try to find the source code for a module.
+
+        Returns ('filename', 'source'). One of it can be None if
+        no filename or source found
+        """
+        try:
+            mod = import_module(modname)
+        except Exception as err:
+            raise PycodeError('error importing %r' % modname, err)
+        loader = getattr(mod, '__loader__', None)
+        filename = getattr(mod, '__file__', None)
+        if loader and getattr(loader, 'get_source', None):
+            # prefer Native loader, as it respects #coding directive
+            try:  
+                source = loader.get_source(modname)
+                if source:
+                    # no exception and not None - it must be module source
+                    return filename, source
+            except ImportError as err:
+                pass  # Try other "source-mining" methods
+        if filename is None and loader and getattr(loader, 'get_filename', None):
+            # have loader, but no filename
+            try:
+                filename = loader.get_filename(modname)
+            except ImportError as err:
+                raise PycodeError('error getting filename for %r' % modname, err)
+        if filename is None:
+            # all methods for getting filename failed, so raise...
+            raise PycodeError('no source found for module %r' % modname)
+        filename = path.normpath(path.abspath(filename))
+        lfilename = filename.lower()
+        if lfilename.endswith('.pyo') or lfilename.endswith('.pyc'):
+            filename = filename[:-1]
+            if not path.isfile(filename) and path.isfile(filename + 'w'):
+                filename += 'w'
+        elif not (lfilename.endswith('.py') or lfilename.endswith('.pyw')):
+            raise PycodeError('source is not a .py file: %r' % filename)
+        elif ('.egg' + os.path.sep) in filename:
+            pat = '(?<=\\.egg)' + re.escape(os.path.sep)
+            eggpath, _ = re.split(pat, filename, 1)
+            if path.isfile(eggpath):
+                return filename, None
+
+        if not path.isfile(filename):
+            raise PycodeError('source file is not present: %r' % filename)
+        return filename, None
+    
     @classmethod
     def for_string(cls, string: str, modname: str, srcname: str = '<string>'
                    ) -> "ModuleAnalyzer":
@@ -63,7 +113,7 @@ class ModuleAnalyzer:
             return entry
 
         try:
-            filename, source = get_module_source(modname)
+            filename, source = cls.get_module_source(modname)
             if source is not None:
                 obj = cls.for_string(source, modname, filename if filename is not None else '<string>')
             elif filename is not None:
diff --git a/sphinx/util/__init__.py b/sphinx/util/__init__.py
index 81bc51254..19ffec633 100644
--- a/sphinx/util/__init__.py
+++ b/sphinx/util/__init__.py
@@ -25,7 +25,7 @@ from hashlib import md5
 from importlib import import_module
 from os import path
 from time import mktime, strptime
-from typing import Any, Callable, Dict, IO, Iterable, Iterator, List, Pattern, Set, Tuple, Optional
+from typing import Any, Callable, Dict, IO, Iterable, Iterator, List, Pattern, Set, Tuple
 from urllib.parse import urlsplit, urlunsplit, quote_plus, parse_qsl, urlencode
 
 from docutils.utils import relative_path
@@ -265,35 +265,31 @@ def save_traceback(app: "Sphinx") -> str:
     return path
 
 
-def get_module_source(modname: str) -> Tuple[Optional[str], Optional[str]]:
+def get_module_source(modname: str) -> Tuple[str, str]:
     """Try to find the source code for a module.
 
-    Returns ('filename', 'source'). One of it can be None if
-    no filename or source found
+    Can return ('file', 'filename') in which case the source is in the given
+    file, or ('string', 'source') which which case the source is the string.
     """
     try:
         mod = import_module(modname)
     except Exception as err:
         raise PycodeError('error importing %r' % modname, err)
-    loader = getattr(mod, '__loader__', None)
     filename = getattr(mod, '__file__', None)
-    if loader and getattr(loader, 'get_source', None):
-        # prefer Native loader, as it respects #coding directive
-        try:  
-            source = loader.get_source(modname)
-            if source:
-                # no exception and not None - it must be module source
-                return filename, source
-        except ImportError as err:
-            pass  # Try other "source-mining" methods
-    if filename is None and loader and getattr(loader, 'get_filename', None):
-        # have loader, but no filename
+    loader = getattr(mod, '__loader__', None)
+    if loader and getattr(loader, 'get_filename', None):
         try:
             filename = loader.get_filename(modname)
-        except ImportError as err:
-            raise PycodeError('error getting filename for %r' % modname, err)
+        except Exception as err:
+            raise PycodeError('error getting filename for %r' % filename, err)
+    if filename is None and loader:
+        try:
+            filename = loader.get_source(modname)
+            if filename:
+                return 'string', filename
+        except Exception as err:
+            raise PycodeError('error getting source for %r' % modname, err)
     if filename is None:
-        # all methods for getting filename failed, so raise...
         raise PycodeError('no source found for module %r' % modname)
     filename = path.normpath(path.abspath(filename))
     lfilename = filename.lower()
@@ -307,11 +303,11 @@ def get_module_source(modname: str) -> Tuple[Optional[str], Optional[str]]:
         pat = '(?<=\\.egg)' + re.escape(os.path.sep)
         eggpath, _ = re.split(pat, filename, 1)
         if path.isfile(eggpath):
-            return filename, None
+            return 'file', filename
 
     if not path.isfile(filename):
         raise PycodeError('source file is not present: %r' % filename)
-    return filename, None
+    return 'file', filename
 
 
 def get_full_modname(modname: str, attribute: str) -> str:
diff --git a/tests/test_pycode.py b/tests/test_pycode.py
index cd039070c..be61d9efb 100644
--- a/tests/test_pycode.py
+++ b/tests/test_pycode.py
@@ -10,12 +10,23 @@
 
 import os
 import sys
+import pytest
 
 import sphinx
 from sphinx.pycode import ModuleAnalyzer
+from sphinx.errors import PycodeError
 
 SPHINX_MODULE_PATH = os.path.splitext(sphinx.__file__)[0] + '.py'
 
+def test_ModuleAnalyzer_get_module_source():
+    assert ModuleAnalyzer.get_module_source('sphinx') == (sphinx.__file__, sphinx.__loader__.get_source('sphinx'))
+
+    # failed to obtain source information from builtin modules
+    with pytest.raises(PycodeError):
+        ModuleAnalyzer.get_module_source('builtins')
+    with pytest.raises(PycodeError):
+        ModuleAnalyzer.get_module_source('itertools')
+        
 
 def test_ModuleAnalyzer_for_string():
     analyzer = ModuleAnalyzer.for_string('print("Hello world")', 'module_name')
diff --git a/tests/test_util.py b/tests/test_util.py
index 4f9317df2..44a41dca1 100644
--- a/tests/test_util.py
+++ b/tests/test_util.py
@@ -62,7 +62,7 @@ def test_display_chunk():
 
 
 def test_get_module_source():
-    assert get_module_source('sphinx') == (sphinx.__file__, sphinx.__loader__.get_source('sphinx'))
+    assert get_module_source('sphinx') == ('file', sphinx.__file__)
 
     # failed to obtain source information from builtin modules
     with pytest.raises(PycodeError):