17 files changed, 143 insertions, 38 deletions
diff --git a/Tools/scripts/byext.py b/Tools/scripts/byext.py
index b79ff37..736a441 100755
--- a/Tools/scripts/byext.py
+++ b/Tools/scripts/byext.py
@@ -25,7 +25,7 @@ class Stats:
         self.addstats("<dir>", "dirs", 1)
         try:
             names = os.listdir(dir)
-        except os.error as err:
+        except OSError as err:
             sys.stderr.write("Can't list %s: %s\n" % (dir, err))
             self.addstats("<dir>", "unlistable", 1)
             return
diff --git a/Tools/scripts/byteyears.py b/Tools/scripts/byteyears.py
index 490b37f..f58c346 100755
--- a/Tools/scripts/byteyears.py
+++ b/Tools/scripts/byteyears.py
@@ -43,7 +43,7 @@ def main():
     for filename in sys.argv[1:]:
         try:
             st = statfunc(filename)
-        except os.error as msg:
+        except OSError as msg:
             sys.stderr.write("can't stat %r: %r\n" % (filename, msg))
             status = 1
             st = ()
diff --git a/Tools/scripts/checkpyc.py b/Tools/scripts/checkpyc.py
index d4fdce2..009f85b 100755
--- a/Tools/scripts/checkpyc.py
+++ b/Tools/scripts/checkpyc.py
@@ -24,7 +24,7 @@ def main():
     for dirname in sys.path:
         try:
             names = os.listdir(dirname)
-        except os.error:
+        except OSError:
             print('Cannot list directory', repr(dirname))
             continue
         if not silent:
@@ -34,7 +34,7 @@ def main():
                 name = os.path.join(dirname, name)
                 try:
                     st = os.stat(name)
-                except os.error:
+                except OSError:
                     print('Cannot stat', repr(name))
                     continue
                 if verbose:
diff --git a/Tools/scripts/copytime.py b/Tools/scripts/copytime.py
index e0220b5..715683f 100755
--- a/Tools/scripts/copytime.py
+++ b/Tools/scripts/copytime.py
@@ -13,12 +13,12 @@ def main():
     file1, file2 = sys.argv[1], sys.argv[2]
     try:
         stat1 = os.stat(file1)
-    except os.error:
+    except OSError:
         sys.stderr.write(file1 + ': cannot stat\n')
         sys.exit(1)
     try:
         os.utime(file2, (stat1[ST_ATIME], stat1[ST_MTIME]))
-    except os.error:
+    except OSError:
         sys.stderr.write(file2 + ': cannot change time\n')
         sys.exit(2)
 
diff --git a/Tools/scripts/finddiv.py b/Tools/scripts/finddiv.py
index f24a702..a705f56 100755
--- a/Tools/scripts/finddiv.py
+++ b/Tools/scripts/finddiv.py
@@ -70,7 +70,7 @@ def process(filename, listnames):
 def processdir(dir, listnames):
     try:
         names = os.listdir(dir)
-    except os.error as msg:
+    except OSError as msg:
         sys.stderr.write("Can't list directory: %s\n" % dir)
         return 1
     files = []
diff --git a/Tools/scripts/findlinksto.py b/Tools/scripts/findlinksto.py
index b4c09ef..b924f27 100755
--- a/Tools/scripts/findlinksto.py
+++ b/Tools/scripts/findlinksto.py
@@ -36,7 +36,7 @@ def visit(prog, dirname, names):
             linkto = os.readlink(name)
             if prog.search(linkto) is not None:
                 print(name, '->', linkto)
-        except os.error:
+        except OSError:
             pass
 
 if __name__ == '__main__':
diff --git a/Tools/scripts/fixcid.py b/Tools/scripts/fixcid.py
index 87e2a09..1e4c428 100755
--- a/Tools/scripts/fixcid.py
+++ b/Tools/scripts/fixcid.py
@@ -97,7 +97,7 @@ def recursedown(dirname):
     bad = 0
     try:
         names = os.listdir(dirname)
-    except os.error as msg:
+    except OSError as msg:
         err(dirname + ': cannot list directory: ' + str(msg) + '\n')
         return 1
     names.sort()
@@ -175,17 +175,17 @@ def fix(filename):
     try:
         statbuf = os.stat(filename)
         os.chmod(tempname, statbuf[ST_MODE] & 0o7777)
-    except os.error as msg:
+    except OSError as msg:
         err(tempname + ': warning: chmod failed (' + str(msg) + ')\n')
     # Then make a backup of the original file as filename~
     try:
         os.rename(filename, filename + '~')
-    except os.error as msg:
+    except OSError as msg:
         err(filename + ': warning: backup failed (' + str(msg) + ')\n')
     # Now move the temp file to the original file
     try:
         os.rename(tempname, filename)
-    except os.error as msg:
+    except OSError as msg:
         err(filename + ': rename failed (' + str(msg) + ')\n')
         return 1
     # Return success
diff --git a/Tools/scripts/ftpmirror.py b/Tools/scripts/ftpmirror.py
index 9e8be1d..a1b683a 100755
--- a/Tools/scripts/ftpmirror.py
+++ b/Tools/scripts/ftpmirror.py
@@ -108,7 +108,7 @@ def mirrorsubdir(f, localdir):
         if verbose: print('Creating local directory', repr(localdir))
         try:
             makedir(localdir)
-        except os.error as msg:
+        except OSError as msg:
             print("Failed to establish local directory", repr(localdir))
             return
     infofilename = os.path.join(localdir, '.mirrorinfo')
@@ -183,7 +183,7 @@ def mirrorsubdir(f, localdir):
                 continue
         try:
             os.unlink(tempname)
-        except os.error:
+        except OSError:
             pass
         if mode[0] == 'l':
             if verbose:
@@ -218,11 +218,11 @@ def mirrorsubdir(f, localdir):
                 fp1.close()
         try:
             os.unlink(fullname)
-        except os.error:
+        except OSError:
             pass            # Ignore the error
         try:
             os.rename(tempname, fullname)
-        except os.error as msg:
+        except OSError as msg:
             print("Can't rename %r to %r: %s" % (tempname, fullname, msg))
             continue
         info[filename] = infostuff
@@ -255,7 +255,7 @@ def mirrorsubdir(f, localdir):
     try:
         if not localdir: names = os.listdir(os.curdir)
         else: names = os.listdir(localdir)
-    except os.error:
+    except OSError:
         names = []
     for name in names:
         if name[0] == '.' or name in info or name in subdirs:
@@ -312,7 +312,7 @@ def remove(fullname):
     if os.path.isdir(fullname) and not os.path.islink(fullname):
         try:
             names = os.listdir(fullname)
-        except os.error:
+        except OSError:
             names = []
         ok = 1
         for name in names:
@@ -322,13 +322,13 @@ def remove(fullname):
             return 0
         try:
             os.rmdir(fullname)
-        except os.error as msg:
+        except OSError as msg:
             print("Can't remove local directory %r: %s" % (fullname, msg))
             return 0
     else:
         try:
             os.unlink(fullname)
-        except os.error as msg:
+        except OSError as msg:
             print("Can't remove local file %r: %s" % (fullname, msg))
             return 0
     return 1
@@ -386,7 +386,7 @@ def writedict(dict, filename):
     backup = os.path.join(dir, fname + '~')
     try:
         os.unlink(backup)
-    except os.error:
+    except OSError:
         pass
     fp = open(tempname, 'w')
     fp.write('{\n')
@@ -396,7 +396,7 @@ def writedict(dict, filename):
     fp.close()
     try:
         os.rename(filename, backup)
-    except os.error:
+    except OSError:
         pass
     os.rename(tempname, filename)
 
diff --git a/Tools/scripts/linktree.py b/Tools/scripts/linktree.py
index 982f480..e83f198 100755
--- a/Tools/scripts/linktree.py
+++ b/Tools/scripts/linktree.py
@@ -32,13 +32,13 @@ def main():
         return 1
     try:
         os.mkdir(newtree, 0o777)
-    except os.error as msg:
+    except OSError as msg:
         print(newtree + ': cannot mkdir:', msg)
         return 1
     linkname = os.path.join(newtree, link)
     try:
         os.symlink(os.path.join(os.pardir, oldtree), linkname)
-    except os.error as msg:
+    except OSError as msg:
         if not link_may_fail:
             print(linkname + ': cannot symlink:', msg)
             return 1
@@ -51,7 +51,7 @@ def linknames(old, new, link):
     if debug: print('linknames', (old, new, link))
     try:
         names = os.listdir(old)
-    except os.error as msg:
+    except OSError as msg:
         print(old + ': warning: cannot listdir:', msg)
         return
     for name in names:
diff --git a/Tools/scripts/parse_html5_entities.py b/Tools/scripts/parse_html5_entities.py
new file mode 100644
index 0000000..c011328
--- /dev/null
+++ b/Tools/scripts/parse_html5_entities.py
@@ -0,0 +1,105 @@
+#!/usr/bin/env python3
+"""
+Utility for parsing HTML5 entity definitions available from:
+
+    http://dev.w3.org/html5/spec/entities.json
+
+Written by Ezio Melotti and Iuliia Proskurnia.
+
+"""
+
+import os
+import sys
+import json
+from urllib.request import urlopen
+from html.entities import html5
+
+entities_url = 'http://dev.w3.org/html5/spec/entities.json'
+
+def get_json(url):
+    """Download the json file from the url and returns a decoded object."""
+    with urlopen(url) as f:
+        data = f.read().decode('utf-8')
+    return json.loads(data)
+
+def create_dict(entities):
+    """Create the html5 dict from the decoded json object."""
+    new_html5 = {}
+    for name, value in entities.items():
+        new_html5[name.lstrip('&')] = value['characters']
+    return new_html5
+
+def compare_dicts(old, new):
+    """Compare the old and new dicts and print the differences."""
+    added = new.keys() - old.keys()
+    if added:
+        print('{} entitie(s) have been added:'.format(len(added)))
+        for name in sorted(added):
+            print('  {!r}: {!r}'.format(name, new[name]))
+    removed = old.keys() - new.keys()
+    if removed:
+        print('{} entitie(s) have been removed:'.format(len(removed)))
+        for name in sorted(removed):
+            print('  {!r}: {!r}'.format(name, old[name]))
+    changed = set()
+    for name in (old.keys() & new.keys()):
+        if old[name] != new[name]:
+            changed.add((name, old[name], new[name]))
+    if changed:
+        print('{} entitie(s) have been modified:'.format(len(changed)))
+        for item in sorted(changed):
+            print('  {!r}: {!r} -> {!r}'.format(*item))
+
+def write_items(entities, file=sys.stdout):
+    """Write the items of the dictionary in the specified file."""
+    # The keys in the generated dictionary should be sorted
+    # in a case-insensitive way, however, when two keys are equal,
+    # the uppercase version should come first so that the result
+    # looks like: ['Aacute', 'aacute', 'Aacute;', 'aacute;', ...]
+    # To do this we first sort in a case-sensitive way (so all the
+    # uppercase chars come first) and then sort with key=str.lower.
+    # Since the sorting is stable the uppercase keys will eventually
+    # be before their equivalent lowercase version.
+    keys = sorted(entities.keys())
+    keys = sorted(keys, key=str.lower)
+    print('html5 = {', file=file)
+    for name in keys:
+        print('    {!r}: {!a},'.format(name, entities[name]), file=file)
+    print('}', file=file)
+
+
+if __name__ == '__main__':
+    # without args print a diff between html.entities.html5 and new_html5
+    # with --create print the new html5 dict
+    # with --patch patch the Lib/html/entities.py file
+    new_html5 = create_dict(get_json(entities_url))
+    if '--create' in sys.argv:
+        print('# map the HTML5 named character references to the '
+              'equivalent Unicode character(s)')
+        print('# Generated by {}.  Do not edit manually.'.format(__file__))
+        write_items(new_html5)
+    elif '--patch' in sys.argv:
+        fname = 'Lib/html/entities.py'
+        temp_fname = fname + '.temp'
+        with open(fname) as f1, open(temp_fname, 'w') as f2:
+            skip = False
+            for line in f1:
+                if line.startswith('html5 = {'):
+                    write_items(new_html5, file=f2)
+                    skip = True
+                    continue
+                if skip:
+                    # skip the old items until the }
+                    if line.startswith('}'):
+                        skip = False
+                    continue
+                f2.write(line)
+        os.remove(fname)
+        os.rename(temp_fname, fname)
+    else:
+        if html5 == new_html5:
+            print('The current dictionary is updated.')
+        else:
+            compare_dicts(html5, new_html5)
+            print('Run "./python {0} --patch" to update Lib/html/entities.html '
+                  'or "./python {0} --create" to see the generated ' 'dictionary.'.format(__file__))
diff --git a/Tools/scripts/pathfix.py b/Tools/scripts/pathfix.py
index dd08e0a..ae15561 100755
--- a/Tools/scripts/pathfix.py
+++ b/Tools/scripts/pathfix.py
@@ -73,7 +73,7 @@ def recursedown(dirname):
     bad = 0
     try:
         names = os.listdir(dirname)
-    except os.error as msg:
+    except OSError as msg:
         err('%s: cannot list directory: %r\n' % (dirname, msg))
         return 1
     names.sort()
@@ -131,24 +131,24 @@ def fix(filename):
         mtime = statbuf.st_mtime
         atime = statbuf.st_atime
         os.chmod(tempname, statbuf[ST_MODE] & 0o7777)
-    except os.error as msg:
+    except OSError as msg:
         err('%s: warning: chmod failed (%r)\n' % (tempname, msg))
     # Then make a backup of the original file as filename~
     try:
         os.rename(filename, filename + '~')
-    except os.error as msg:
+    except OSError as msg:
         err('%s: warning: backup failed (%r)\n' % (filename, msg))
     # Now move the temp file to the original file
     try:
         os.rename(tempname, filename)
-    except os.error as msg:
+    except OSError as msg:
         err('%s: rename failed (%r)\n' % (filename, msg))
         return 1
     if preserve_timestamps:
         if atime and mtime:
             try:
                 os.utime(filename, (atime, mtime))
-            except os.error as msg:
+            except OSError as msg:
                 err('%s: reset of timestamp failed (%r)\n' % (filename, msg))
                 return 1
     # Return succes
diff --git a/Tools/scripts/pindent.py b/Tools/scripts/pindent.py
index 2872dc0..3333420 100755
--- a/Tools/scripts/pindent.py
+++ b/Tools/scripts/pindent.py
@@ -376,13 +376,13 @@ def make_backup(filename):
     if os.path.lexists(backup):
         try:
             os.remove(backup)
-        except os.error:
+        except OSError:
             print("Can't remove backup %r" % (backup,), file=sys.stderr)
         # end try
     # end if
     try:
         os.rename(filename, backup)
-    except os.error:
+    except OSError:
         print("Can't rename %r to %r" % (filename, backup), file=sys.stderr)
     # end try
 # end def make_backup
diff --git a/Tools/scripts/reindent.py b/Tools/scripts/reindent.py
index 4a916ea..18424de 100755
--- a/Tools/scripts/reindent.py
+++ b/Tools/scripts/reindent.py
@@ -52,8 +52,8 @@ verbose = False
 recurse = False
 dryrun = False
 makebackup = True
+# A specified newline to be used in the output (set by --newline option)
 spec_newline = None
-"""A specified newline to be used in the output (set by --newline option)"""
 
 
 def usage(msg=None):
diff --git a/Tools/scripts/treesync.py b/Tools/scripts/treesync.py
index b2649c4..652d394 100755
--- a/Tools/scripts/treesync.py
+++ b/Tools/scripts/treesync.py
@@ -78,7 +78,7 @@ def process(slave, master):
         print("creating slave directory", slave)
         try:
             os.mkdir(slave)
-        except os.error as msg:
+        except OSError as msg:
             print("can't make slave directory", slave, ":", msg)
             return
         else:
diff --git a/Tools/scripts/untabify.py b/Tools/scripts/untabify.py
index 4b67c15..861c83c 100755
--- a/Tools/scripts/untabify.py
+++ b/Tools/scripts/untabify.py
@@ -39,11 +39,11 @@ def process(filename, tabsize, verbose=True):
     backup = filename + "~"
     try:
         os.unlink(backup)
-    except os.error:
+    except OSError:
         pass
     try:
         os.rename(filename, backup)
-    except os.error:
+    except OSError:
         pass
     with open(filename, "w", encoding=encoding) as f:
         f.write(newtext)
diff --git a/Tools/scripts/which.py b/Tools/scripts/which.py
index 4fc37a0..df54ce0 100755
--- a/Tools/scripts/which.py
+++ b/Tools/scripts/which.py
@@ -29,7 +29,7 @@ def main():
             filename = os.path.join(dir, prog)
             try:
                 st = os.stat(filename)
-            except os.error:
+            except OSError:
                 continue
             if not S_ISREG(st[ST_MODE]):
                 msg(filename + ': not a disk file')
diff --git a/Tools/scripts/win_add2path.py b/Tools/scripts/win_add2path.py
index 9259b44..c85bea5 100644
--- a/Tools/scripts/win_add2path.py
+++ b/Tools/scripts/win_add2path.py
@@ -30,7 +30,7 @@ def modify():
     with winreg.CreateKey(HKCU, ENV) as key:
         try:
             envpath = winreg.QueryValueEx(key, PATH)[0]
-        except WindowsError:
+        except OSError:
             envpath = DEFAULT
 
         paths = [envpath]