Tools/i18n/msgfmt.py

   1 #! /usr/bin/env python
   2
   3 # Written by Martin v. Löwis <loewis@informatik.hu-berlin.de>
   4
   5 """Generate binary message catalog from textual translation description.
   6
   7 This program converts a textual Uniforum-style message catalog (.po file) into
   8 a binary GNU catalog (.mo file).  This is essentially the same function as the
   9 GNU msgfmt program, however, it is a simpler implementation.
  10
  11 Usage: msgfmt.py [OPTIONS] filename.po
  12
  13 Options:
  14     -h
  15     --help
  16         Print this message and exit.
  17
  18     -V
  19     --version
  20         Display version information and exit.
  21
  22 """
  23
  24 import sys
  25 import getopt
  26 import struct
  27 import array
  28
  29 __version__ = "1.0"
  30
  31 MESSAGES = {}
  32
  33
  34 \f
  35 def usage(code, msg=''):
  36     print >> sys.stderr, __doc__
  37     if msg:
  38         print >> sys.stderr, msg
  39     sys.exit(code)
  40
  41
  42 \f
  43 def add(id, str, fuzzy):
  44     "Add a non-fuzzy translation to the dictionary."
  45     global MESSAGES
  46     if not fuzzy and str:
  47         MESSAGES[id] = str
  48
  49
  50 \f
  51 def generate():
  52     "Return the generated output."
  53     global MESSAGES
  54     keys = MESSAGES.keys()
  55     # the keys are sorted in the .mo file
  56     keys.sort()
  57     offsets = []
  58     ids = strs = ''
  59     for id in keys:
  60         # For each string, we need size and file offset.  Each string is NUL
  61         # terminated; the NUL does not count into the size.
  62         offsets.append((len(ids), len(id), len(strs), len(MESSAGES[id])))
  63         ids += id + '\0'
  64         strs += MESSAGES[id] + '\0'
  65     output = ''
  66     # The header is 7 32-bit unsigned integers.  We don't use hash tables, so
  67     # the keys start right after the index tables.
  68     # translated string.
  69     keystart = 7*4+16*len(keys)
  70     # and the values start after the keys
  71     valuestart = keystart + len(ids)
  72     koffsets = []
  73     voffsets = []
  74     # The string table first has the list of keys, then the list of values.
  75     # Each entry has first the size of the string, then the file offset.
  76     for o1, l1, o2, l2 in offsets:
  77         koffsets += [l1, o1+keystart]
  78         voffsets += [l2, o2+valuestart]
  79     offsets = koffsets + voffsets
  80     output = struct.pack("iiiiiii",
  81                          0x950412de,        # Magic
  82                          0,                 # Version
  83                          len(keys),         # # of entries
  84                          7*4,               # start of key index
  85                          7*4+len(keys)*8,   # start of value index
  86                          0, 0)              # size and offset of hash table
  87     output += array.array("i", offsets).tostring()
  88     output += ids
  89     output += strs
  90     return output
  91
  92
  93 \f
  94 def make(filename):
  95     ID = 1
  96     STR = 2
  97
  98     # Compute .mo name from .po name
  99     if filename.endswith('.po'):
 100         infile = filename
 101         outfile = filename[:-2] + 'mo'
 102     else:
 103         infile = filename + '.po'
 104         outfile = filename + '.mo'
 105     try:
 106         lines = open(infile).readlines()
 107     except IOError, msg:
 108         print >> sys.stderr, msg
 109         sys.exit(1)
 110
 111     section = None
 112     fuzzy = 0
 113
 114     # Parse the catalog
 115     lno = 0
 116     for l in lines:
 117         lno += 1
 118         # If we get a comment line after a msgstr, this is a new entry
 119         if l[0] == '#' and section == STR:
 120             add(msgid, msgstr, fuzzy)
 121             section = None
 122             fuzzy = 0
 123         # Record a fuzzy mark
 124         if l[:2] == '#,' and l.find('fuzzy'):
 125             fuzzy = 1
 126         # Skip comments
 127         if l[0] == '#':
 128             continue
 129         # Now we are in a msgid section, output previous section
 130         if l.startswith('msgid'):
 131             if section == STR:
 132                 add(msgid, msgstr, fuzzy)
 133             section = ID
 134             l = l[5:]
 135             msgid = msgstr = ''
 136         # Now we are in a msgstr section
 137         elif l.startswith('msgstr'):
 138             section = STR
 139             l = l[6:]
 140         # Skip empty lines
 141         l = l.strip()
 142         if not l:
 143             continue
 144         # XXX: Does this always follow Python escape semantics?
 145         l = eval(l)
 146         if section == ID:
 147             msgid += l
 148         elif section == STR:
 149             msgstr += l
 150         else:
 151             print >> sys.stderr, 'Syntax error on %s:%d' % (infile, lno), \
 152                   'before:'
 153             print >> sys.stderr, l
 154             sys.exit(1)
 155     # Add last entry
 156     if section == STR:
 157         add(msgid, msgstr, fuzzy)
 158
 159     # Compute output
 160     output = generate()
 161
 162     # Save output
 163     try:
 164         open(outfile,"wb").write(output)
 165     except IOError,msg:
 166         print >> sys.stderr, msg
 167
 168
 169 \f
 170 def main():
 171     try:
 172         opts, args = getopt.getopt(sys.argv[1:], 'hV', ['help','version'])
 173     except getopt.error, msg:
 174         usage(1, msg)
 175
 176     # parse options
 177     for opt, arg in opts:
 178         if opt in ('-h', '--help'):
 179             usage(0)
 180         elif opt in ('-V', '--version'):
 181             print >> sys.stderr, "msgfmt.py", __version__
 182             sys.exit(0)
 183     # do it
 184     if not args:
 185         print >> sys.stderr, 'No input file given'
 186         print >> sys.stderr, "Try `msgfmt --help' for more information."
 187         return
 188
 189     for filename in args:
 190         make(filename)
 191
 192
 193 if __name__ == '__main__':
 194     main()