Convert CLDR scripts to Python 3

The convertion is moslty done using 2to3 script with manual cleanup
afterwards.

Task-number: QTBUG-83488
Pick-to: 6.2
Change-Id: I4d33b04e7269c55a83ff2deb876a23a78a89f39d
Reviewed-by: Cristian Maureira-Fredes <cristian.maureira-fredes@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>

util/locale_database/cldr.py

@@ -153,7 +153,7 @@ class CldrReader (object):
 
     def __parseTags(self, locale):
         tags = self.__splitLocale(locale)
-        language = tags.next()
+        language = next(tags)
         script = territory = variant = ''
         try:
             script, territory, variant = tags
@@ -171,10 +171,10 @@ class CldrReader (object):
         single tag (i.e. contains no underscores).  Always yields 1 or
         4 values, never 2 or 3."""
         tags = iter(name.split('_'))
-        yield tags.next() # Language
+        yield next(tags) # Language
 
         try:
-            tag = tags.next()
+            tag = next(tags)
         except StopIteration:
             return
 
@@ -182,7 +182,7 @@ class CldrReader (object):
         if len(tag) == 4 and tag[0].isupper() and tag[1:].islower():
             yield tag
             try:
-                tag = tags.next()
+                tag = next(tags)
             except StopIteration:
                 tag = ''
         else:
@@ -192,7 +192,7 @@ class CldrReader (object):
         if tag and tag.isupper() or tag.isdigit():
             yield tag
             try:
-                tag = tags.next()
+                tag = next(tags)
             except StopIteration:
                 tag = ''
         else:
@@ -726,7 +726,7 @@ enumdata.py (keeping the old name as an alias):
             except (KeyError, ValueError, TypeError):
                 pass
             else:
-                if key not in seen or not elt.attributes.has_key('alt'):
+                if key not in seen or 'alt' not in elt.attributes:
                     yield key, value
                     seen.add(key)
 

util/locale_database/cldr2qlocalexml.py

@@ -1,4 +1,4 @@
-#!/usr/bin/env python2
+#!/usr/bin/env python3
 # coding=utf8
 #############################################################################
 ##
@@ -100,10 +100,6 @@ def main(args, out, err):
         usage(name, err, 'Too many arguments - excess: ' + ' '.join(args))
         return 1
 
-    if emit.encoding != 'UTF-8' or (emit.encoding is None and sys.getdefaultencoding() != 'UTF-8'):
-        reload(sys) # Weirdly, this gets a richer sys module than the plain import got us !
-        sys.setdefaultencoding('UTF-8')
-
     # TODO - command line options to tune choice of grumble and whitter:
     reader = CldrReader(root, err.write, err.write)
     writer = QLocaleXmlWriter(emit.write)

util/locale_database/cldr2qtimezone.py

@@ -1,4 +1,4 @@
-#!/usr/bin/env python2
+#!/usr/bin/env python3
 #############################################################################
 ##
 ## Copyright (C) 2020 The Qt Company Ltd.

util/locale_database/dateconverter.py

@@ -60,7 +60,7 @@ def _convert_pattern(pattern):
         "v" : "t", "vv" : "t", "vvv" : "t", "vvvv" : "t", # timezone
         "V" : "t", "VV" : "t", "VVV" : "t", "VVVV" : "t"  # timezone
     }
-    if qt_patterns.has_key(pattern):
+    if pattern in qt_patterns:
         return qt_patterns[pattern]
     for r,v in qt_regexps.items():
         pattern = re.sub(r, v, pattern)

util/locale_database/ldml.py

@@ -124,7 +124,7 @@ class Node (object):
         one."""
         seq = self.findAllChildren(tag)
         try:
-            node = seq.next()
+            node = next(seq)
         except StopIteration:
             raise Error('No child found where one was expected', tag)
         for it in seq:
@@ -197,7 +197,7 @@ class Supplement (XmlScanner):
                                 for e in elts):
             if elt.attributes:
                 yield (elt.nodeName,
-                       dict((k, v if isinstance(v, basestring) else v.nodeValue)
+                       dict((k, v if isinstance(v, str) else v.nodeValue)
                             for k, v in elt.attributes.items()))
 
 class LocaleScanner (object):
@@ -312,7 +312,7 @@ class LocaleScanner (object):
         except Error:
             money = self.find(xpath)
         money = self.__currencyFormats(money, plus, minus)
-        yield 'currencyFormat', money.next()
+        yield 'currencyFormat', next(money)
         neg = ''
         for it in money:
             assert not neg, 'There should be at most one more pattern'

util/locale_database/localetools.py

@@ -40,8 +40,8 @@ Classes:
 import os
 import tempfile
 
-class Error (StandardError):
+class Error (Exception):
-    __upinit = StandardError.__init__
+    __upinit = Exception.__init__
     def __init__(self, msg, *args):
         self.__upinit(msg, *args)
         self.message = msg

util/locale_database/qlocalexml.py

@@ -36,14 +36,14 @@ Provides classes:
 Support:
   Spacer -- provides control over indentation of the output.
 """
-from __future__ import print_function
+
 from xml.sax.saxutils import escape
 
 from localetools import Error
 
 # Tools used by Locale:
 def camel(seq):
-    yield seq.next()
+    yield next(seq)
     for word in seq:
         yield word.capitalize()
 
@@ -58,7 +58,7 @@ def startCount(c, text): # strspn
     """First index in text where it doesn't have a character in c"""
     assert text and text[0] in c
     try:
-        return (j for j, d in enumerate(text) if d not in c).next()
+        return next((j for j, d in enumerate(text) if d not in c))
     except StopIteration:
         return len(text)
 
@@ -166,7 +166,7 @@ class QLocaleXmlReader (object):
 
     def languageIndices(self, locales):
         index = 0
-        for key, value in self.languages.iteritems():
+        for key, value in self.languages.items():
             i, count = 0, locales.count(key)
             if count > 0:
                 i = index
@@ -360,9 +360,7 @@ class QLocaleXmlWriter (object):
         self.__openTag('locale')
         self.__writeLocale(Locale.C(calendars), calendars)
         self.__closeTag('locale')
-        keys = locales.keys()
+        for key in sorted(locales.keys()):
-        keys.sort()
-        for key in keys:
             self.__openTag('locale')
             self.__writeLocale(locales[key], calendars)
             self.__closeTag('locale')
@@ -403,7 +401,7 @@ class QLocaleXmlWriter (object):
 
     def __enumTable(self, tag, table):
         self.__openTag(tag + 'List')
-        for key, value in table.iteritems():
+        for key, value in table.items():
             self.__openTag(tag)
             self.inTag('name', value[0])
             self.inTag('id', key)
@@ -545,7 +543,7 @@ class Locale (object):
                 '_'.join((k, cal))
                 for k in self.propsMonthDay('months')
                 for cal in calendars):
-            write(key, escape(get(key)).encode('utf-8'))
+            write(key, escape(get(key)))
 
         write('groupSizes', ';'.join(str(x) for x in get('groupSizes')))
         for key in ('currencyDigits', 'currencyRounding'):

util/locale_database/qlocalexml2cpp.py

@@ -1,4 +1,4 @@
-#!/usr/bin/env python2
+#!/usr/bin/env python3
 #############################################################################
 ##
 ## Copyright (C) 2021 The Qt Company Ltd.
@@ -162,8 +162,7 @@ class LocaleDataWriter (LocaleSourceEditor):
         def keyLikely(entry):
             have = entry[1] # Numeric id triple
             return have[0] or huge, have[2] or huge, have[1] or huge # language, region, script
-        likely = list(likely) # Turn generator into list so we can sort it
+        likely = sorted(likely, key=keyLikely)
-        likely.sort(key=keyLikely)
 
         i = 0
         self.writer.write('static const QLocaleId likely_subtags[] = {\n')