Make lyx_pot.py python3 compatible

Now it produces the same output if running under python3 (tested with 3.4.2)
or python2 (tested with 2.7.9). python3 always uses unicode strings
internally, so we have to specify the file encoding on opening a file, such
that strings can be converted from and to the file encoding on reading and
writing. Using the io module for file io ensures that the behaviour is the
same for python2 and python3. For python2 we also have to mark string literals
as unicode strings by using the u prefix (which is a noop in python3).

Many thanks to José for review and pointing out all the details.

diff --git a/po/lyx_pot.py b/po/lyx_pot.py
index 432a126056886743bf16d8eace43eb8b59f0fddf..500185fdba6a38a5f3c49bf74c4226c117d67dc4 100755 (executable)
--- a/po/lyx_pot.py
+++ b/po/lyx_pot.py
@@ -19,6 +19,7 @@
 from __future__ import print_function
 
 import sys, os, re, getopt
+import io
 
 def relativePath(path, base):
     '''return relative path from top source dir'''
@@ -37,13 +38,13 @@ def writeString(outfile, infile, basefile, lineno, string):
     string = string.replace('\\', '\\\\').replace('"', '')
     if string == "":
         return
-    print('#: %s:%d\nmsgid "%s"\nmsgstr ""\n' % \
+    print(u'#: %s:%d\nmsgid "%s"\nmsgstr ""\n' % \
         (relativePath(infile, basefile), lineno, string), file=outfile)
 
 
 def ui_l10n(input_files, output, base):
     '''Generate pot file from lib/ui/*'''
-    output = open(output, 'w')
+    output = io.open(output, 'w', encoding='utf_8')
     Submenu = re.compile(r'^[^#]*Submenu\s+"([^"]*)"', re.IGNORECASE)
     Popupmenu = re.compile(r'^[^#]*PopupMenu\s+"[^"]+"\s+"([^"]*)"', re.IGNORECASE)
     IconPalette = re.compile(r'^[^#]*IconPalette\s+"[^"]+"\s+"([^"]*)"', re.IGNORECASE)
@@ -51,7 +52,7 @@ def ui_l10n(input_files, output, base):
     Item = re.compile(r'[^#]*Item\s+"([^"]*)"', re.IGNORECASE)
     TableInsert = re.compile(r'[^#]*TableInsert\s+"([^"]*)"', re.IGNORECASE)
     for src in input_files:
-        input = open(src)
+        input = io.open(src, encoding='utf_8')
         for lineno, line in enumerate(input.readlines()):
             if Submenu.match(line):
                 (string,) = Submenu.match(line).groups()
@@ -70,7 +71,7 @@ def ui_l10n(input_files, output, base):
                 continue
             string = string.replace('"', '')
             if string != "":
-                print('#: %s:%d\nmsgid "%s"\nmsgstr ""\n' % \
+                print(u'#: %s:%d\nmsgid "%s"\nmsgstr ""\n' % \
                     (relativePath(src, base), lineno+1, string), file=output)
         input.close()
     output.close()
@@ -125,7 +126,7 @@ def layouts_l10n(input_files, output, base, layouttranslations):
 
         # read old translations if available
         try:
-            input = open(output)
+            input = io.open(output, encoding='utf_8')
             lang = ''
             for line in input.readlines():
                 res = Comment.search(line)
@@ -147,8 +148,8 @@ def layouts_l10n(input_files, output, base, layouttranslations):
                     continue
                 res = KeyValPair.search(line)
                 if res and lang != '':
-                    key = res.group(1).decode('utf-8')
-                    val = res.group(2).decode('utf-8')
+                    key = res.group(1)
+                    val = res.group(2)
                     key = key.replace('\\"', '"').replace('\\\\', '\\')
                     val = val.replace('\\"', '"').replace('\\\\', '\\')
                     oldtrans[lang][key] = val
@@ -165,7 +166,7 @@ def layouts_l10n(input_files, output, base, layouttranslations):
         if 'wa' in languages:
             languages.remove('wa')
 
-    out = open(output, 'w')
+    out = io.open(output, 'w', encoding='utf_8')
     for src in input_files:
         readingDescription = False
         readingI18nPreamble = False
@@ -178,7 +179,7 @@ def layouts_l10n(input_files, output, base, layouttranslations):
         descStartLine = -1
         descLines = []
         lineno = 0
-        for line in open(src).readlines():
+        for line in io.open(src, encoding='utf_8').readlines():
             lineno += 1
             res = ClassDescription.search(line)
             if res != None:
@@ -381,7 +382,7 @@ def layouts_l10n(input_files, output, base, layouttranslations):
 
         ContextRe = re.compile(r'(.*)(\[\[.*\]\])')
 
-        print('''# This file has been automatically generated by po/lyx_pot.py.
+        print(u'''# This file has been automatically generated by po/lyx_pot.py.
 # PLEASE MODIFY ONLY THE LAGUAGES HAVING NO .po FILE! If you want to regenerate
 # this file from the translations, run `make ../lib/layouttranslations' in po.
 # Python polib library is needed for building the output file.
@@ -389,7 +390,7 @@ def layouts_l10n(input_files, output, base, layouttranslations):
 # This file should remain fixed during minor LyX releases.
 # For more comments see README.localization file.''', file=out)
         for lang in languages:
-            print('\nTranslation %s' % lang, file=out)
+            print(u'\nTranslation %s' % lang, file=out)
             if lang in list(oldtrans.keys()):
                 trans = oldtrans[lang]
             else:
@@ -415,8 +416,7 @@ def layouts_l10n(input_files, output, base, layouttranslations):
                     if res != None:
                         val = res.group(1)
                     key = key.replace('\\', '\\\\').replace('"', '\\"')
-                    print('\t"%s" "%s"' % \
-                             (key.encode('utf-8'), val.encode('utf-8')), file=out)
+                    print(u'\t"%s" "%s"' % (key, val), file=out)
                 # also print untranslated entries to help translators
                 elif not lang in oldlanguages:
                     key = key.replace('\\', '\\\\').replace('"', '\\"')
@@ -425,20 +425,19 @@ def layouts_l10n(input_files, output, base, layouttranslations):
                         val = res.group(1)
                     else:
                         val = key
-                    print('\t"%s" "%s"' % \
-                             (key.encode('utf-8'), val.encode('utf-8')), file=out)
-            print('End', file=out)
+                    print(u'\t"%s" "%s"' % (key, val), file=out)
+            print(u'End', file=out)
 
     out.close()
 
 
 def qt4_l10n(input_files, output, base):
     '''Generate pot file from src/frontends/qt4/ui/*.ui'''
-    output = open(output, 'w')
+    output = io.open(output, 'w', encoding='utf_8')
     pat = re.compile(r'\s*<string>(.*)</string>')
     prop = re.compile(r'\s*<property.*name.*=.*shortcut')
     for src in input_files:
-        input = open(src)
+        input = io.open(src, encoding='utf_8')
         skipNextLine = False
         for lineno, line in enumerate(input.readlines()):
             # skip the line after <property name=shortcut>
@@ -455,7 +454,7 @@ def qt4_l10n(input_files, output, base):
                 string = string.replace('&lt;', '<').replace('&gt;', '>')
                 string = string.replace('\\', '\\\\').replace('"', r'\"')
                 string = string.replace('&#x0a;', r'\n')
-                print('#: %s:%d\nmsgid "%s"\nmsgstr ""\n' % \
+                print(u'#: %s:%d\nmsgid "%s"\nmsgstr ""\n' % \
                     (relativePath(src, base), lineno+1, string), file=output)
         input.close()
     output.close()
@@ -463,47 +462,47 @@ def qt4_l10n(input_files, output, base):
 
 def languages_l10n(input_files, output, base):
     '''Generate pot file from lib/languages'''
-    out = open(output, 'w')
+    out = io.open(output, 'w', encoding='utf_8')
     GuiName = re.compile(r'^[^#]*GuiName\s+(.*)', re.IGNORECASE)
-    
+
     for src in input_files:
         descStartLine = -1
         descLines = []
         lineno = 0
-        for line in open(src).readlines():
+        for line in io.open(src, encoding='utf_8').readlines():
             lineno += 1
             res = GuiName.search(line)
             if res != None:
                 string = res.group(1)
                 writeString(out, src, base, lineno, string)
                 continue
-               
+
     out.close()
 
 
 def latexfonts_l10n(input_files, output, base):
     '''Generate pot file from lib/latexfonts'''
-    out = open(output, 'w')
+    out = io.open(output, 'w', encoding='utf_8')
     GuiName = re.compile(r'^[^#]*GuiName\s+(.*)', re.IGNORECASE)
-    
+
     for src in input_files:
         descStartLine = -1
         descLines = []
         lineno = 0
-        for line in open(src).readlines():
+        for line in io.open(src, encoding='utf_8').readlines():
             lineno += 1
             res = GuiName.search(line)
             if res != None:
                 string = res.group(1)
                 writeString(out, src, base, lineno, string)
                 continue
-               
+
     out.close()
 
 
 def external_l10n(input_files, output, base):
     '''Generate pot file from lib/external_templates'''
-    output = open(output, 'w')
+    output = io.open(output, 'w', encoding='utf_8')
     Template = re.compile(r'^Template\s+(.*)', re.IGNORECASE)
     GuiName = re.compile(r'\s*GuiName\s+(.*)', re.IGNORECASE)
     HelpTextStart = re.compile(r'\s*HelpText\s', re.IGNORECASE)
@@ -511,7 +510,7 @@ def external_l10n(input_files, output, base):
     HelpTextEnd = re.compile(r'\s*HelpTextEnd\s', re.IGNORECASE)
     i = -1
     for src in input_files:
-        input = open(src)
+        input = io.open(src, encoding='utf_8')
         inHelp = False
         hadHelp = False
         prev_help_string = ''
@@ -523,7 +522,7 @@ def external_l10n(input_files, output, base):
             elif inHelp:
                 if HelpTextEnd.match(line):
                     if hadHelp:
-                        print('\nmsgstr ""\n', file=output)
+                        print(u'\nmsgstr ""\n', file=output)
                     inHelp = False
                     hadHelp = False
                     prev_help_string = ''
@@ -531,11 +530,11 @@ def external_l10n(input_files, output, base):
                     (help_string,) = HelpTextSection.match(line).groups()
                     help_string = help_string.replace('"', '')
                     if help_string != "" and prev_help_string == '':
-                        print('#: %s:%d\nmsgid ""\n"%s\\n"' % \
+                        print(u'#: %s:%d\nmsgid ""\n"%s\\n"' % \
                             (relativePath(src, base), lineno+1, help_string), file=output)
                         hadHelp = True
                     elif help_string != "":
-                        print('"%s\\n"' % help_string, file=output)
+                        print(u'"%s\\n"' % help_string, file=output)
                     prev_help_string = help_string
             elif HelpTextStart.match(line):
                 inHelp = True
@@ -544,7 +543,7 @@ def external_l10n(input_files, output, base):
                 continue
             string = string.replace('"', '')
             if string != "" and not inHelp:
-                print('#: %s:%d\nmsgid "%s"\nmsgstr ""\n' % \
+                print(u'#: %s:%d\nmsgid "%s"\nmsgstr ""\n' % \
                     (relativePath(src, base), lineno+1, string), file=output)
         input.close()
     output.close()
@@ -552,10 +551,10 @@ def external_l10n(input_files, output, base):
 
 def formats_l10n(input_files, output, base):
     '''Generate pot file from configure.py'''
-    output = open(output, 'w')
+    output = io.open(output, 'w', encoding='utf_8')
     GuiName = re.compile(r'.*\\Format\s+\S+\s+\S+\s+"([^"]*)"\s+(\S*)\s+.*', re.IGNORECASE)
     GuiName2 = re.compile(r'.*\\Format\s+\S+\s+\S+\s+([^"]\S+)\s+(\S*)\s+.*', re.IGNORECASE)
-    input = open(input_files[0])
+    input = io.open(input_files[0], encoding='utf_8')
     for lineno, line in enumerate(input.readlines()):
         label = ""
         labelsc = ""
@@ -571,10 +570,10 @@ def formats_l10n(input_files, output, base):
         if shortcut != "":
             labelsc = label + "|" + shortcut
         if label != "":
-            print('#: %s:%d\nmsgid "%s"\nmsgstr ""\n' % \
+            print(u'#: %s:%d\nmsgid "%s"\nmsgstr ""\n' % \
                 (relativePath(input_files[0], base), lineno+1, label), file=output)
         if labelsc != "":
-            print('#: %s:%d\nmsgid "%s"\nmsgstr ""\n' % \
+            print(u'#: %s:%d\nmsgid "%s"\nmsgstr ""\n' % \
                 (relativePath(input_files[0], base), lineno+1, labelsc), file=output)
     input.close()
     output.close()
@@ -582,16 +581,16 @@ def formats_l10n(input_files, output, base):
 
 def encodings_l10n(input_files, output, base):
     '''Generate pot file from lib/encodings'''
-    output = open(output, 'w')
+    output = io.open(output, 'w', encoding='utf_8')
     # assuming only one encodings file
     #                 Encoding utf8      utf8    "Unicode (utf8)" UTF-8    variable inputenc
     reg = re.compile('Encoding [\w-]+\s+[\w-]+\s+"([\w \-\(\)]+)"\s+[\w-]+\s+(fixed|variable|variableunsafe)\s+\w+.*')
-    input = open(input_files[0])
+    input = io.open(input_files[0], encoding='utf_8')
     for lineno, line in enumerate(input.readlines()):
         if not line.startswith('Encoding'):
             continue
         if reg.match(line):
-            print('#: %s:%d\nmsgid "%s"\nmsgstr ""\n' % \
+            print(u'#: %s:%d\nmsgid "%s"\nmsgstr ""\n' % \
                 (relativePath(input_files[0], base), lineno+1, reg.match(line).groups()[0]), file=output)
         else:
             print("Error: Unable to handle line:")
@@ -644,7 +643,7 @@ if __name__ == '__main__':
         elif opt in ['-t', '--type']:
             input_type = value
         elif opt in ['-s', '--src_file']:
-            input_files = [f.strip() for f in open(value)]
+            input_files = [f.strip() for f in io.open(value, encoding='utf_8')]
 
     if input_type not in ['ui', 'layouts', 'layouttranslations', 'qt4', 'languages', 'latexfonts', 'encodings', 'external', 'formats'] or output is None:
         print('Wrong input type or output filename.')