src/util/xmlpool/gen_xmlpool.py

   1
   2 #
   3 # Usage:
   4 #     gen_xmlpool.py /path/to/t_option.h localedir lang lang lang ...
   5 #
   6 # For each given language, this script expects to find a .mo file at
   7 # `{localedir}/{language}/LC_MESSAGES/options.mo`.
   8 #
   9
  10 from __future__ import print_function
  11
  12 import sys
  13 import gettext
  14 import re
  15
  16 # Path to t_options.h
  17 template_header_path = sys.argv[1]
  18
  19 localedir = sys.argv[2]
  20
  21 # List of supported languages
  22 languages = sys.argv[3:]
  23
  24 # Escape special characters in C strings
  25 def escapeCString (s):
  26     escapeSeqs = {'\a' : '\\a', '\b' : '\\b', '\f' : '\\f', '\n' : '\\n',
  27                   '\r' : '\\r', '\t' : '\\t', '\v' : '\\v', '\\' : '\\\\'}
  28     # " -> '' is a hack. Quotes (") aren't possible in XML attributes.
  29     # Better use Unicode characters for typographic quotes in option
  30     # descriptions and translations.
  31     i = 0
  32     r = ''
  33     while i < len(s):
  34         # Special case: escape double quote with \u201c or \u201d, depending
  35         # on whether it's an open or close quote. This is needed because plain
  36         # double quotes are not possible in XML attributes.
  37         if s[i] == '"':
  38             if i == len(s)-1 or s[i+1].isspace():
  39                 # close quote
  40                 q = u'\u201c'
  41             else:
  42                 # open quote
  43                 q = u'\u201d'
  44             r = r + q
  45         elif s[i] in escapeSeqs:
  46             r = r + escapeSeqs[s[i]]
  47         else:
  48             r = r + s[i]
  49         i = i + 1
  50     return r
  51
  52 # Expand escape sequences in C strings (needed for gettext lookup)
  53 def expandCString (s):
  54     escapeSeqs = {'a' : '\a', 'b' : '\b', 'f' : '\f', 'n' : '\n',
  55                   'r' : '\r', 't' : '\t', 'v' : '\v',
  56                   '"' : '"', '\\' : '\\'}
  57     i = 0
  58     escape = False
  59     hexa = False
  60     octa = False
  61     num = 0
  62     digits = 0
  63     r = ''
  64     while i < len(s):
  65         if not escape:
  66             if s[i] == '\\':
  67                 escape = True
  68             else:
  69                 r = r + s[i]
  70         elif hexa:
  71             if (s[i] >= '0' and s[i] <= '9') or \
  72                (s[i] >= 'a' and s[i] <= 'f') or \
  73                (s[i] >= 'A' and s[i] <= 'F'):
  74                 num = num * 16 + int(s[i],16)
  75                 digits = digits + 1
  76             else:
  77                 digits = 2
  78             if digits >= 2:
  79                 hexa = False
  80                 escape = False
  81                 r = r + chr(num)
  82         elif octa:
  83             if s[i] >= '0' and s[i] <= '7':
  84                 num = num * 8 + int(s[i],8)
  85                 digits = digits + 1
  86             else:
  87                 digits = 3
  88             if digits >= 3:
  89                 octa = False
  90                 escape = False
  91                 r = r + chr(num)
  92         else:
  93             if s[i] in escapeSeqs:
  94                 r = r + escapeSeqs[s[i]]
  95                 escape = False
  96             elif s[i] >= '0' and s[i] <= '7':
  97                 octa = True
  98                 num = int(s[i],8)
  99                 if num <= 3:
 100                     digits = 1
 101                 else:
 102                     digits = 2
 103             elif s[i] == 'x' or s[i] == 'X':
 104                 hexa = True
 105                 num = 0
 106                 digits = 0
 107             else:
 108                 r = r + s[i]
 109                 escape = False
 110         i = i + 1
 111     return r
 112
 113 # Expand matches. The first match is always a DESC or DESC_BEGIN match.
 114 # Subsequent matches are ENUM matches.
 115 #
 116 # DESC, DESC_BEGIN format: \1 \2=<lang> \3 \4=gettext(" \5=<text> \6=") \7
 117 # ENUM format:             \1 \2=gettext(" \3=<text> \4=") \5
 118 def expandMatches (matches, translations, end=None):
 119     assert len(matches) > 0
 120     nTranslations = len(translations)
 121     i = 0
 122     # Expand the description+enums for all translations
 123     for lang,trans in translations:
 124         i = i + 1
 125         # Make sure that all but the last line of a simple description
 126         # are extended with a backslash.
 127         suffix = ''
 128         if len(matches) == 1 and i < len(translations) and \
 129                not matches[0].expand (r'\7').endswith('\\'):
 130             suffix = ' \\'
 131         # Expand the description line. Need to use ugettext in order to allow
 132         # non-ascii unicode chars in the original English descriptions.
 133         text = escapeCString (trans.ugettext (unicode (expandCString (
 134             matches[0].expand (r'\5')), "utf-8"))).encode("utf-8")
 135         print(matches[0].expand (r'\1' + lang + r'\3"' + text + r'"\7') + suffix)
 136         # Expand any subsequent enum lines
 137         for match in matches[1:]:
 138             text = escapeCString (trans.ugettext (unicode (expandCString (
 139                 match.expand (r'\3')), "utf-8"))).encode("utf-8")
 140             print(match.expand (r'\1"' + text + r'"\5'))
 141
 142         # Expand description end
 143         if end:
 144             print(end, end='')
 145
 146 # Compile a list of translation classes to all supported languages.
 147 # The first translation is always a NullTranslations.
 148 translations = [("en", gettext.NullTranslations())]
 149 for lang in languages:
 150     try:
 151         trans = gettext.translation ("options", localedir, [lang])
 152     except IOError:
 153         sys.stderr.write ("Warning: language '%s' not found.\n" % lang)
 154         continue
 155     translations.append ((lang, trans))
 156
 157 # Regular expressions:
 158 reLibintl_h  = re.compile (r'#\s*include\s*<libintl.h>')
 159 reDESC       = re.compile (r'(\s*DRI_CONF_DESC\s*\(\s*)([a-z]+)(\s*,\s*)(gettext\s*\(\s*")(.*)("\s*\))(\s*\)[ \t]*\\?)$')
 160 reDESC_BEGIN = re.compile (r'(\s*DRI_CONF_DESC_BEGIN\s*\(\s*)([a-z]+)(\s*,\s*)(gettext\s*\(\s*")(.*)("\s*\))(\s*\)[ \t]*\\?)$')
 161 reENUM       = re.compile (r'(\s*DRI_CONF_ENUM\s*\([^,]+,\s*)(gettext\s*\(\s*")(.*)("\s*\))(\s*\)[ \t]*\\?)$')
 162 reDESC_END   = re.compile (r'\s*DRI_CONF_DESC_END')
 163
 164 # Print a header
 165 print("/***********************************************************************\n" \
 166 " ***        THIS FILE IS GENERATED AUTOMATICALLY. DON'T EDIT!        ***\n" \
 167 " ***********************************************************************/")
 168
 169 # Process the options template and generate options.h with all
 170 # translations.
 171 template = file (template_header_path, "r")
 172 descMatches = []
 173 for line in template:
 174     if len(descMatches) > 0:
 175         matchENUM     = reENUM    .match (line)
 176         matchDESC_END = reDESC_END.match (line)
 177         if matchENUM:
 178             descMatches.append (matchENUM)
 179         elif matchDESC_END:
 180             expandMatches (descMatches, translations, line)
 181             descMatches = []
 182         else:
 183             sys.stderr.write (
 184                 "Warning: unexpected line inside description dropped:\n%s\n" \
 185                 % line)
 186         continue
 187     if reLibintl_h.search (line):
 188         # Ignore (comment out) #include <libintl.h>
 189         print("/* %s * commented out by gen_xmlpool.py */" % line)
 190         continue
 191     matchDESC       = reDESC      .match (line)
 192     matchDESC_BEGIN = reDESC_BEGIN.match (line)
 193     if matchDESC:
 194         assert len(descMatches) == 0
 195         expandMatches ([matchDESC], translations)
 196     elif matchDESC_BEGIN:
 197         assert len(descMatches) == 0
 198         descMatches = [matchDESC_BEGIN]
 199     else:
 200         print(line, end='')
 201
 202 if len(descMatches) > 0:
 203     sys.stderr.write ("Warning: unterminated description at end of file.\n")
 204     expandMatches (descMatches, translations)