Second phase of refactoring for runpy, pkgutil, pydoc, and setuptools

to share common PEP 302 support code, as described here:

http://mail.python.org/pipermail/python-dev/2006-April/063724.html

pydoc now supports PEP 302 importers, by way of utility functions in
pkgutil, such as 'walk_packages()'.  It will properly document
modules that are in zip files, and is backward compatible to Python
2.3 (setuptools installs for Python <2.5 will bundle it so pydoc
doesn't break when used with eggs.)

What has not changed is that pydoc command line options do not support
zip paths or other importer paths, and the webserver index does not
support sys.meta_path.  Those are probably okay as limitations.

Tasks remaining: write docs and Misc/NEWS for pkgutil/pydoc changes,
and update setuptools to use pkgutil wherever possible, then add it
to the stdlib.

1 files changed, 73 insertions, 109 deletions

diff --git a/Lib/pydoc.py b/Lib/pydoc.py
index ee45643..ff6e7ca 100755
--- a/Lib/pydoc.py
+++ b/Lib/pydoc.py
@@ -52,10 +52,16 @@ Richard Chamberlain, for the first implementation of textdoc.
 #     the current directory is changed with os.chdir(), an incorrect
 #     path will be displayed.
 
-import sys, imp, os, re, types, inspect, __builtin__
+import sys, imp, os, re, types, inspect, __builtin__, pkgutil
 from repr import Repr
 from string import expandtabs, find, join, lower, split, strip, rfind, rstrip
-from collections import deque
+try:
+    from collections import deque
+except ImportError:
+    # Python 2.3 compatibility
+    class deque(list):
+        def popleft(self):
+            return self.pop(0)
 
 # --------------------------------------------------------- common routines
 
@@ -182,6 +188,23 @@ def ispackage(path):
                 return True
     return False
 
+def source_synopsis(file):
+    line = file.readline()
+    while line[:1] == '#' or not strip(line):
+        line = file.readline()
+        if not line: break
+    line = strip(line)
+    if line[:4] == 'r"""': line = line[1:]
+    if line[:3] == '"""':
+        line = line[3:]
+        if line[-1:] == '\\': line = line[:-1]
+        while not strip(line):
+            line = file.readline()
+            if not line: break
+        result = strip(split(line, '"""')[0])
+    else: result = None
+    return result
+
 def synopsis(filename, cache={}):
     """Get the one-line summary out of a module file."""
     mtime = os.stat(filename).st_mtime
@@ -196,24 +219,11 @@ def synopsis(filename, cache={}):
         if info and 'b' in info[2]: # binary modules have to be imported
             try: module = imp.load_module('__temp__', file, filename, info[1:])
             except: return None
-            result = split(module.__doc__ or '', '\n')[0]
+            result = (module.__doc__ or '').splitlines()[0]
             del sys.modules['__temp__']
         else: # text modules can be directly examined
-            line = file.readline()
-            while line[:1] == '#' or not strip(line):
-                line = file.readline()
-                if not line: break
-            line = strip(line)
-            if line[:4] == 'r"""': line = line[1:]
-            if line[:3] == '"""':
-                line = line[3:]
-                if line[-1:] == '\\': line = line[:-1]
-                while not strip(line):
-                    line = file.readline()
-                    if not line: break
-                result = strip(split(line, '"""')[0])
-            else: result = None
-        file.close()
+            result = source_synopsis(file)
+            file.close()
         cache[filename] = (mtime, result)
     return result
 
@@ -643,16 +653,8 @@ class HTMLDoc(Doc):
 
         if hasattr(object, '__path__'):
             modpkgs = []
-            modnames = []
-            for file in os.listdir(object.__path__[0]):
-                path = os.path.join(object.__path__[0], file)
-                modname = inspect.getmodulename(file)
-                if modname != '__init__':
-                    if modname and modname not in modnames:
-                        modpkgs.append((modname, name, 0, 0))
-                        modnames.append(modname)
-                    elif ispackage(path):
-                        modpkgs.append((file, name, 1, 0))
+            for importer, modname, ispkg in pkgutil.iter_modules(object.__path__):
+                modpkgs.append((modname, name, ispkg, 0))
             modpkgs.sort()
             contents = self.multicolumn(modpkgs, self.modpkglink)
             result = result + self.bigsection(
@@ -796,7 +798,10 @@ class HTMLDoc(Doc):
             tag += ':<br>\n'
 
             # Sort attrs by name.
-            attrs.sort(key=lambda t: t[0])
+            try:
+                attrs.sort(key=lambda t: t[0])
+            except TypeError:
+                attrs.sort(lambda t1, t2: cmp(t1[0], t2[0]))    # 2.3 compat
 
             # Pump out the attrs, segregated by kind.
             attrs = spill('Methods %s' % tag, attrs,
@@ -914,25 +919,9 @@ class HTMLDoc(Doc):
         """Generate an HTML index for a directory of modules."""
         modpkgs = []
         if shadowed is None: shadowed = {}
-        seen = {}
-        files = os.listdir(dir)
-
-        def found(name, ispackage,
-                  modpkgs=modpkgs, shadowed=shadowed, seen=seen):
-            if name not in seen:
-                modpkgs.append((name, '', ispackage, name in shadowed))
-                seen[name] = 1
-                shadowed[name] = 1
-
-        # Package spam/__init__.py takes precedence over module spam.py.
-        for file in files:
-            path = os.path.join(dir, file)
-            if ispackage(path): found(file, 1)
-        for file in files:
-            path = os.path.join(dir, file)
-            if os.path.isfile(path):
-                modname = inspect.getmodulename(file)
-                if modname: found(modname, 0)
+        for importer, name, ispkg in pkgutil.iter_modules([dir]):
+            modpkgs.append((name, '', ispkg, name in shadowed))
+            shadowed[name] = 1
 
         modpkgs.sort()
         contents = self.multicolumn(modpkgs, self.modpkglink)
@@ -1059,14 +1048,12 @@ class TextDoc(Doc):
 
         if hasattr(object, '__path__'):
             modpkgs = []
-            for file in os.listdir(object.__path__[0]):
-                path = os.path.join(object.__path__[0], file)
-                modname = inspect.getmodulename(file)
-                if modname != '__init__':
-                    if modname and modname not in modpkgs:
-                        modpkgs.append(modname)
-                    elif ispackage(path):
-                        modpkgs.append(file + ' (package)')
+            for importer, modname, ispkg in pkgutil.iter_modules(object.__path__):
+                if ispkg:
+                    modpkgs.append(modname + ' (package)')
+                else:
+                    modpkgs.append(modname)
+
             modpkgs.sort()
             result = result + self.section(
                 'PACKAGE CONTENTS', join(modpkgs, '\n'))
@@ -1490,20 +1477,9 @@ def writedoc(thing, forceload=0):
 def writedocs(dir, pkgpath='', done=None):
     """Write out HTML documentation for all modules in a directory tree."""
     if done is None: done = {}
-    for file in os.listdir(dir):
-        path = os.path.join(dir, file)
-        if ispackage(path):
-            writedocs(path, pkgpath + file + '.', done)
-        elif os.path.isfile(path):
-            modname = inspect.getmodulename(path)
-            if modname:
-                if modname == '__init__':
-                    modname = pkgpath[:-1] # remove trailing period
-                else:
-                    modname = pkgpath + modname
-                if modname not in done:
-                    done[modname] = 1
-                    writedoc(modname)
+    for importer, modname, ispkg in pkgutil.walk_packages([dir], pkgpath):
+        writedoc(modname)
+    return
 
 class Helper:
     keywords = {
@@ -1830,30 +1806,9 @@ class Scanner:
             self.state.append((child, self.children(child)))
         return child
 
-class ModuleScanner(Scanner):
-    """An interruptible scanner that searches module synopses."""
-    def __init__(self):
-        roots = map(lambda dir: (dir, ''), pathdirs())
-        Scanner.__init__(self, roots, self.submodules, self.isnewpackage)
-        self.inodes = map(lambda (dir, pkg): os.stat(dir).st_ino, roots)
-
-    def submodules(self, (dir, package)):
-        children = []
-        for file in os.listdir(dir):
-            path = os.path.join(dir, file)
-            if ispackage(path):
-                children.append((path, package + (package and '.') + file))
-            else:
-                children.append((path, package))
-        children.sort() # so that spam.py comes before spam.pyc or spam.pyo
-        return children
 
-    def isnewpackage(self, (dir, package)):
-        inode = os.path.exists(dir) and os.stat(dir).st_ino
-        if not (os.path.islink(dir) and inode in self.inodes):
-            self.inodes.append(inode) # detect circular symbolic links
-            return ispackage(dir)
-        return False
+class ModuleScanner:
+    """An interruptible scanner that searches module synopses."""
 
     def run(self, callback, key=None, completer=None):
         if key: key = lower(key)
@@ -1870,22 +1825,31 @@ class ModuleScanner(Scanner):
                     if find(lower(modname + ' - ' + desc), key) >= 0:
                         callback(None, modname, desc)
 
-        while not self.quit:
-            node = self.next()
-            if not node: break
-            path, package = node
-            modname = inspect.getmodulename(path)
-            if os.path.isfile(path) and modname:
-                modname = package + (package and '.') + modname
-                if not modname in seen:
-                    seen[modname] = 1 # if we see spam.py, skip spam.pyc
-                    if key is None:
-                        callback(path, modname, '')
+        for importer, modname, ispkg in pkgutil.walk_packages():
+            if self.quit:
+                break
+            if key is None:
+                callback(None, modname, '')
+            else:
+                loader = importer.find_module(modname)
+                if hasattr(loader,'get_source'):
+                    import StringIO
+                    desc = source_synopsis(
+                        StringIO.StringIO(loader.get_source(modname))
+                    ) or ''
+                    if hasattr(loader,'get_filename'):
+                        path = loader.get_filename(modname)
                     else:
-                        desc = synopsis(path) or ''
-                        if find(lower(modname + ' - ' + desc), key) >= 0:
-                            callback(path, modname, desc)
-        if completer: completer()
+                        path = None
+                else:
+                    module = loader.load_module(modname)
+                    desc = (module.__doc__ or '').splitlines()[0]
+                    path = getattr(module,'__file__',None)
+                if find(lower(modname + ' - ' + desc), key) >= 0:
+                    callback(path, modname, desc)
+
+        if completer:
+            completer()
 
 def apropos(key):
     """Print all the one-line module summaries that contain a substring."""
@@ -1950,7 +1914,7 @@ def serve(port, callback=None, completer=None):
                     'Built-in Modules', '#ffffff', '#ee77aa', contents)]
 
                 seen = {}
-                for dir in pathdirs():
+                for dir in sys.path:
                     indices.append(html.index(dir, seen))
                 contents = heading + join(indices) + '''<p align=right>
 <font color="#909090" face="helvetica, arial"><strong>