cmlenz@162: # -*- coding: utf-8 -*-
cmlenz@162: #
cmlenz@162: # Copyright (C) 2007 Edgewall Software
cmlenz@162: # All rights reserved.
cmlenz@162: #
cmlenz@162: # This software is licensed as described in the file COPYING, which
cmlenz@162: # you should have received as part of this distribution. The terms
cmlenz@162: # are also available at http://babel.edgewall.org/wiki/License.
cmlenz@162: #
cmlenz@162: # This software consists of voluntary contributions made by many
cmlenz@162: # individuals. For the exact contribution history, see the revision
cmlenz@162: # history and logs, available at http://babel.edgewall.org/log/.
cmlenz@162: 
cmlenz@162: """Writing of files in the ``gettext`` MO (machine object) format.
cmlenz@162: 
cmlenz@162: :see: `The Format of MO Files
cmlenz@162:        <http://www.gnu.org/software/gettext/manual/gettext.html#MO-Files>`_
cmlenz@162: """
cmlenz@162: 
cmlenz@162: import array
cmlenz@162: import struct
cmlenz@162: 
cmlenz@163: __all__ = ['write_mo']
cmlenz@163: __docformat__ = 'restructuredtext en'
cmlenz@163: 
cmlenz@162: def write_mo(fileobj, catalog, use_fuzzy=False):
cmlenz@162:     """Write a catalog to the specified file-like object using the GNU MO file
cmlenz@162:     format.
cmlenz@162:     
cmlenz@162:     >>> from babel.messages import Catalog
cmlenz@162:     >>> from gettext import GNUTranslations
cmlenz@162:     >>> from StringIO import StringIO
cmlenz@162:     
cmlenz@162:     >>> catalog = Catalog(locale='en_US')
cmlenz@162:     >>> catalog.add('foo', 'Voh')
cmlenz@162:     >>> catalog.add((u'bar', u'baz'), (u'Bahr', u'Batz'))
cmlenz@162:     >>> catalog.add('fuz', 'Futz', flags=['fuzzy'])
cmlenz@162:     >>> buf = StringIO()
cmlenz@162:     
cmlenz@162:     >>> write_mo(buf, catalog)
cmlenz@162:     >>> buf.seek(0)
cmlenz@162:     >>> translations = GNUTranslations(fp=buf)
cmlenz@162:     >>> translations.ugettext('foo')
cmlenz@162:     u'Voh'
cmlenz@162:     >>> translations.ungettext('bar', 'baz', 1)
cmlenz@162:     u'Bahr'
cmlenz@162:     >>> translations.ungettext('bar', 'baz', 2)
cmlenz@162:     u'Batz'
cmlenz@162:     >>> translations.ugettext('fuz')
cmlenz@162:     u'fuz'
cmlenz@162:     
cmlenz@162:     :param fileobj: the file-like object to write to
cmlenz@162:     :param catalog: the `Catalog` instance
cmlenz@162:     :param use_fuzzy: whether translations marked as "fuzzy" should be included
cmlenz@162:                       in the output
cmlenz@162:     """
cmlenz@162:     messages = list(catalog)
cmlenz@162:     if not use_fuzzy:
cmlenz@162:         messages[1:] = [m for m in messages[1:] if not m.fuzzy]
cmlenz@162:     messages.sort(lambda x,y: cmp(x.id, y.id))
cmlenz@162: 
cmlenz@162:     ids = strs = ''
cmlenz@162:     offsets = []
cmlenz@162: 
cmlenz@162:     for message in messages:
cmlenz@162:         # For each string, we need size and file offset.  Each string is NUL
cmlenz@162:         # terminated; the NUL does not count into the size.
cmlenz@162:         if message.pluralizable:
cmlenz@162:             msgid = '\x00'.join([
cmlenz@162:                 msgid.encode(catalog.charset) for msgid in message.id
cmlenz@162:             ])
cmlenz@162:             msgstr = '\x00'.join([
cmlenz@162:                 msgstr.encode(catalog.charset) for msgstr in message.string
cmlenz@162:             ])
cmlenz@162:         else:
cmlenz@162:             msgid = message.id.encode(catalog.charset)
cmlenz@162:             msgstr = message.string.encode(catalog.charset)
cmlenz@162:         offsets.append((len(ids), len(msgid), len(strs), len(msgstr)))
cmlenz@162:         ids += msgid + '\x00'
cmlenz@162:         strs += msgstr + '\x00'
cmlenz@162: 
cmlenz@162:     # The header is 7 32-bit unsigned integers.  We don't use hash tables, so
cmlenz@162:     # the keys start right after the index tables.
cmlenz@162:     keystart = 7 * 4 + 16 * len(messages)
cmlenz@162:     valuestart = keystart + len(ids)
cmlenz@162: 
cmlenz@162:     # The string table first has the list of keys, then the list of values.
cmlenz@162:     # Each entry has first the size of the string, then the file offset.
cmlenz@162:     koffsets = []
cmlenz@162:     voffsets = []
cmlenz@162:     for o1, l1, o2, l2 in offsets:
cmlenz@162:         koffsets += [l1, o1 + keystart]
cmlenz@162:         voffsets += [l2, o2 + valuestart]
cmlenz@162:     offsets = koffsets + voffsets
cmlenz@162: 
cmlenz@162:     fileobj.write(struct.pack('Iiiiiii',
cmlenz@162:         0x950412deL,                # magic
cmlenz@162:         0,                          # version
cmlenz@162:         len(messages),              # number of entries
cmlenz@162:         7 * 4,                      # start of key index
cmlenz@162:         7 * 4 + len(messages) * 8,  # start of value index
cmlenz@162:         0, 0                        # size and offset of hash table
cmlenz@162:     ) + array.array("i", offsets).tostring() + ids + strs)