Rietveld Code Review Tool
Help | Bug tracker | Discussion group | Source code | Sign in
(40)

Side by Side Diff: translate/storage/cpo.py

Issue 81: Fix CPO memory leak SVN Base: https://translate.svn.sourceforge.net/svnroot/translate/src/trunk/
Patch Set: Fix addlocation method Created 1 year, 3 months ago
Use n/p to move between diff chunks; N/P to move between comments. Please Sign in to add in-line comments.
Jump to:
View unified diff
OLDNEW
1 #!/usr/bin/env python 1 #!/usr/bin/env python
2 # -*- coding: utf-8 -*- 2 # -*- coding: utf-8 -*-
3 # 3 #
4 # Copyright 2002-2007 Zuza Software Foundation 4 # Copyright 2002-2007 Zuza Software Foundation
5 # 5 #
6 # This file is part of translate. 6 # This file is part of translate.
7 # 7 #
8 # translate is free software; you can redistribute it and/or modify 8 # translate is free software; you can redistribute it and/or modify
9 # it under the terms of the GNU General Public License as published by 9 # it under the terms of the GNU General Public License as published by
10 # the Free Software Foundation; either version 2 of the License, or 10 # the Free Software Foundation; either version 2 of the License, or
11 # (at your option) any later version. 11 # (at your option) any later version.
12 # 12 #
13 # translate is distributed in the hope that it will be useful, 13 # translate is distributed in the hope that it will be useful,
14 # but WITHOUT ANY WARRANTY; without even the implied warranty of 14 # but WITHOUT ANY WARRANTY; without even the implied warranty of
15 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the 15 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 # GNU General Public License for more details. 16 # GNU General Public License for more details.
17 # 17 #
18 # You should have received a copy of the GNU General Public License 18 # You should have received a copy of the GNU General Public License
19 # along with translate; if not, write to the Free Software 19 # along with translate; if not, write to the Free Software
20 # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA 20 # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
21 21
22 """Classes that hold units of .po files (pounit) or entire files (pofile). 22 """Classes that hold units of .po files (pounit) or entire files (pofile).
23 23
24 Gettext-style .po (or .pot) files are used in translations for KDE, GNOME and 24 Gettext-style .po (or .pot) files are used in translations for KDE, GNOME and
25 many other projects. 25 many other projects.
26 26
27 This uses libgettextpo from the gettext package. Any version before 0.17 will 27 This uses libgettextpo from the gettext package. Any version before 0.17 will
28 at least cause some subtle bugs or may not work at all. Developers might want 28 at least cause some subtle bugs or may not work at all. Developers might want
29 to have a look at gettext-tools/libgettextpo/gettext-po.h from the gettext 29 to have a look at gettext-tools/libgettextpo/gettext-po.h from the gettext
30 package for the public API of the library. 30 package for the public API of the library.
31 """ 31 """
32 32
33 from translate.misc.multistring import multistring 33 from translate.misc.multistring import multistring
34 from translate.storage import pocommon 34 from translate.storage import pocommon
35 from translate.misc import quote 35 from translate.misc import quote
36 from translate.lang import data 36 from translate.lang import data
37 from ctypes import * 37 from ctypes import *
38 import ctypes.util 38 import ctypes.util
39 try: 39 try:
40 import cStringIO as StringIO 40 import cStringIO as StringIO
41 except ImportError: 41 except ImportError:
42 import StringIO 42 import StringIO
43 import os 43 import os
44 import pypo 44 import pypo
45 import re 45 import re
46 import sys 46 import sys
47 import tempfile 47 import tempfile
48 48
49 lsep = " " 49 lsep = " "
50 """Seperator for #: entries""" 50 """Seperator for #: entries"""
51 51
52 STRING = c_char_p 52 STRING = c_char_p
53 53
54 # Structures 54 # Structures
55 class po_message(Structure): 55 class po_message(Structure):
56 _fields_ = [] 56 _fields_ = []
57 57
58 # Function prototypes 58 # Function prototypes
59 xerror_prototype = CFUNCTYPE(None, c_int, POINTER(po_message), STRING, c_uint, c _uint, c_int, STRING) 59 xerror_prototype = CFUNCTYPE(None, c_int, POINTER(po_message), STRING, c_uint, c _uint, c_int, STRING)
60 xerror2_prototype = CFUNCTYPE(None, c_int, POINTER(po_message), STRING, c_uint, c_uint, c_int, STRING, POINTER(po_message), STRING, c_uint, c_uint, c_int, STRIN G) 60 xerror2_prototype = CFUNCTYPE(None, c_int, POINTER(po_message), STRING, c_uint, c_uint, c_int, STRING, POINTER(po_message), STRING, c_uint, c_uint, c_int, STRIN G)
61 61
62 62
63 # Structures (error handler) 63 # Structures (error handler)
64 class po_xerror_handler(Structure): 64 class po_xerror_handler(Structure):
65 _fields_ = [('xerror', xerror_prototype), 65 _fields_ = [('xerror', xerror_prototype),
66 ('xerror2', xerror2_prototype)] 66 ('xerror2', xerror2_prototype)]
67 67
68 class po_error_handler(Structure): 68 class po_error_handler(Structure):
69 _fields_ = [ 69 _fields_ = [
70 ('error', CFUNCTYPE(None, c_int, c_int, STRING)), 70 ('error', CFUNCTYPE(None, c_int, c_int, STRING)),
71 ('error_at_line', CFUNCTYPE(None, c_int, c_int, STRING, c_uint, STRING)), 71 ('error_at_line', CFUNCTYPE(None, c_int, c_int, STRING, c_uint, STRING)),
72 ('multiline_warning', CFUNCTYPE(None, STRING, STRING)), 72 ('multiline_warning', CFUNCTYPE(None, STRING, STRING)),
73 ('multiline_error', CFUNCTYPE(None, STRING, STRING)), 73 ('multiline_error', CFUNCTYPE(None, STRING, STRING)),
74 ] 74 ]
75 75
76 # Callback functions for po_xerror_handler 76 # Callback functions for po_xerror_handler
77 def xerror_cb(severity, message, filename, lineno, column, multilint_p, message_ text): 77 def xerror_cb(severity, message, filename, lineno, column, multilint_p, message_ text):
78 print >> sys.stderr, "xerror_cb", severity, message, filename, lineno, colum n, multilint_p, message_text 78 print >> sys.stderr, "xerror_cb", severity, message, filename, lineno, colum n, multilint_p, message_text
79 if severity >= 1: 79 if severity >= 1:
80 raise ValueError(message_text) 80 raise ValueError(message_text)
81 81
82 def xerror2_cb(severity, message1, filename1, lineno1, column1, multiline_p1, me ssage_text1, message2, filename2, lineno2, column2, multiline_p2, message_text2) : 82 def xerror2_cb(severity, message1, filename1, lineno1, column1, multiline_p1, me ssage_text1, message2, filename2, lineno2, column2, multiline_p2, message_text2) :
83 print >> sys.stderr, "xerror2_cb", severity, message1, filename1, lineno1, c olumn1, multiline_p1, message_text1, message2, filename2, lineno2, column2, mult iline_p2, message_text2 83 print >> sys.stderr, "xerror2_cb", severity, message1, filename1, lineno1, c olumn1, multiline_p1, message_text1, message2, filename2, lineno2, column2, mult iline_p2, message_text2
84 if severity >= 1: 84 if severity >= 1:
85 raise ValueError(message_text1) 85 raise ValueError(message_text1)
86 86
87 87
88 88
89 # Load libgettextpo 89 # Load libgettextpo
90 gpo = None 90 gpo = None
91 # 'gettextpo' is recognised on Unix, while only 'libgettextpo' is recognised on 91 # 'gettextpo' is recognised on Unix, while only 'libgettextpo' is recognised on
92 # windows. Therefore we test both. 92 # windows. Therefore we test both.
93 names = ['gettextpo', 'libgettextpo'] 93 names = ['gettextpo', 'libgettextpo']
94 for name in names: 94 for name in names:
95 lib_location = ctypes.util.find_library(name) 95 lib_location = ctypes.util.find_library(name)
96 if lib_location: 96 if lib_location:
97 gpo = cdll.LoadLibrary(lib_location) 97 gpo = cdll.LoadLibrary(lib_location)
98 if gpo: 98 if gpo:
99 break 99 break
100 else: 100 else:
101 # Now we are getting desperate, so let's guess a unix type DLL that might 101 # Now we are getting desperate, so let's guess a unix type DLL that might
102 # be in LD_LIBRARY_PATH or loaded with LD_PRELOAD 102 # be in LD_LIBRARY_PATH or loaded with LD_PRELOAD
103 try: 103 try:
104 gpo = cdll.LoadLibrary('libgettextpo.so') 104 gpo = cdll.LoadLibrary('libgettextpo.so')
105 except OSError, e: 105 except OSError, e:
106 raise ImportError("gettext PO library not found") 106 raise ImportError("gettext PO library not found")
107 107
108 # Setup return and paramater types 108 # Setup return and paramater types
109 # File access 109 # File access
110 gpo.po_file_read_v3.argtypes = [STRING, POINTER(po_xerror_handler)] 110 gpo.po_file_read_v3.argtypes = [STRING, POINTER(po_xerror_handler)]
111 gpo.po_file_write_v2.argtypes = [c_int, STRING, POINTER(po_xerror_handler)] 111 gpo.po_file_write_v2.argtypes = [c_int, STRING, POINTER(po_xerror_handler)]
112 gpo.po_file_write_v2.retype = c_int 112 gpo.po_file_write_v2.retype = c_int
113 113
114 # Header 114 # Header
115 gpo.po_file_domain_header.restype = STRING 115 gpo.po_file_domain_header.restype = STRING
116 gpo.po_header_field.restype = STRING 116 gpo.po_header_field.restype = STRING
117 gpo.po_header_field.argtypes = [STRING, STRING] 117 gpo.po_header_field.argtypes = [STRING, STRING]
118 118
119 # Locations (filepos) 119 # Locations (filepos)
120 gpo.po_filepos_file.restype = STRING 120 gpo.po_filepos_file.restype = STRING
121 gpo.po_message_filepos.restype = c_int 121 gpo.po_message_filepos.restype = c_int
122 gpo.po_message_filepos.argtypes = [c_int, c_int] 122 gpo.po_message_filepos.argtypes = [c_int, c_int]
123 gpo.po_message_add_filepos.argtypes = [c_int, STRING, c_int] 123 gpo.po_message_add_filepos.argtypes = [c_int, STRING, c_size_t]
dwaynebailey 2008/08/29 13:19:25 I assume this solves the -1 issue? Crafty!
124 124
125 # Message (get methods) 125 # Message (get methods)
126 gpo.po_message_comments.restype = STRING 126 gpo.po_message_comments.restype = STRING
127 gpo.po_message_extracted_comments.restype = STRING 127 gpo.po_message_extracted_comments.restype = STRING
128 gpo.po_message_prev_msgctxt.restype = STRING 128 gpo.po_message_prev_msgctxt.restype = STRING
129 gpo.po_message_prev_msgid.restype = STRING 129 gpo.po_message_prev_msgid.restype = STRING
130 gpo.po_message_prev_msgid_plural.restype = STRING 130 gpo.po_message_prev_msgid_plural.restype = STRING
131 gpo.po_message_is_format.restype = c_int 131 gpo.po_message_is_format.restype = c_int
132 gpo.po_message_msgctxt.restype = STRING 132 gpo.po_message_msgctxt.restype = STRING
133 gpo.po_message_msgid.restype = STRING 133 gpo.po_message_msgid.restype = STRING
134 gpo.po_message_msgid_plural.restype = STRING 134 gpo.po_message_msgid_plural.restype = STRING
135 gpo.po_message_msgstr.restype = STRING 135 gpo.po_message_msgstr.restype = STRING
136 gpo.po_message_msgstr_plural.restype = STRING 136 gpo.po_message_msgstr_plural.restype = STRING
137 137
138 # Message (set methods) 138 # Message (set methods)
139 gpo.po_message_set_comments.argtypes = [c_int, STRING] 139 gpo.po_message_set_comments.argtypes = [c_int, STRING]
140 gpo.po_message_set_extracted_comments.argtypes = [c_int, STRING] 140 gpo.po_message_set_extracted_comments.argtypes = [c_int, STRING]
141 gpo.po_message_set_fuzzy.argtypes = [c_int, c_int] 141 gpo.po_message_set_fuzzy.argtypes = [c_int, c_int]
142 gpo.po_message_set_msgctxt.argtypes = [c_int, STRING] 142 gpo.po_message_set_msgctxt.argtypes = [c_int, STRING]
143 gpo.po_message_set_format.argtypes = [c_int, STRING, c_int]
dwaynebailey 2008/08/29 13:19:25 Not sure what this has to do with addlocations? Y
143 144
144 # Setup the po_xerror_handler 145 # Setup the po_xerror_handler
145 xerror_handler = po_xerror_handler() 146 xerror_handler = po_xerror_handler()
146 xerror_handler.xerror = xerror_prototype(xerror_cb) 147 xerror_handler.xerror = xerror_prototype(xerror_cb)
147 xerror_handler.xerror2 = xerror2_prototype(xerror2_cb) 148 xerror_handler.xerror2 = xerror2_prototype(xerror2_cb)
148 149
149 def escapeforpo(text): 150 def escapeforpo(text):
150 return pypo.escapeforpo(text) 151 return pypo.escapeforpo(text)
151 152
152 def quoteforpo(text): 153 def quoteforpo(text):
153 return pypo.quoteforpo(text) 154 return pypo.quoteforpo(text)
154 155
155 def unquotefrompo(postr, joinwithlinebreak=False): 156 def unquotefrompo(postr, joinwithlinebreak=False):
156 return pypo.unquotefrompo(postr, joinwithlinebreak) 157 return pypo.unquotefrompo(postr, joinwithlinebreak)
157 158
158 def encodingToUse(encoding): 159 def encodingToUse(encoding):
159 return pypo.encodingToUse(encoding) 160 return pypo.encodingToUse(encoding)
160 161
161 def get_libgettextpo_version(): 162 def get_libgettextpo_version():
162 """Returns the libgettextpo version 163 """Returns the libgettextpo version
163 164
164 @return: a three-value tuple containing the libgettextpo version in the 165 @return: a three-value tuple containing the libgettextpo version in the
165 following format: 166 following format:
166 (major version, minor version, subminor version) 167 (major version, minor version, subminor version)
167 """ 168 """
168 libversion = c_long.in_dll(gpo, 'libgettextpo_version') 169 libversion = c_long.in_dll(gpo, 'libgettextpo_version')
169 major = libversion.value >> 16 170 major = libversion.value >> 16
170 minor = libversion.value >> 8 171 minor = libversion.value >> 8
171 subminor = libversion.value - (major << 16) - (minor << 8) 172 subminor = libversion.value - (major << 16) - (minor << 8)
172 return major, minor, subminor 173 return major, minor, subminor
173 174
174 175
175 class pounit(pocommon.pounit): 176 class pounit(pocommon.pounit):
176 def __init__(self, source=None, encoding='utf-8', gpo_message=None): 177 def __init__(self, source=None, encoding='utf-8', gpo_message=None):
177 self._encoding = encoding 178 self._encoding = encoding
178 if not gpo_message: 179 if not gpo_message:
179 self._gpo_message = gpo.po_message_create() 180 self._gpo_message = gpo.po_message_create()
180 if source or source == "": 181 if source or source == "":
181 self.source = source 182 self.source = source
182 self.target = "" 183 self.target = ""
183 elif gpo_message: 184 elif gpo_message:
184 self._gpo_message = gpo_message 185 self._gpo_message = gpo_message
185 186
186 def setmsgidcomment(self, msgidcomment): 187 def setmsgidcomment(self, msgidcomment):
187 if msgidcomment: 188 if msgidcomment:
188 newsource = "_: " + msgidcomment + "\n" + self.source 189 newsource = "_: " + msgidcomment + "\n" + self.source
189 self.source = newsource 190 self.source = newsource
190 msgidcomment = property(None, setmsgidcomment) 191 msgidcomment = property(None, setmsgidcomment)
191 192
192 def setmsgid_plural(self, msgid_plural): 193 def setmsgid_plural(self, msgid_plural):
(...skipping 131 matching lines...) Show 10 above Show 10 below
324 else: 325 else:
325 raise ValueError("Comment type not valid") 326 raise ValueError("Comment type not valid")
326 327
327 if comments and get_libgettextpo_version() < (0, 17, 0): 328 if comments and get_libgettextpo_version() < (0, 17, 0):
328 comments = "\n".join([line.strip() for line in comments.split("\n")] ) 329 comments = "\n".join([line.strip() for line in comments.split("\n")] )
329 # Let's drop the last newline 330 # Let's drop the last newline
330 return comments[:-1].decode(self._encoding) 331 return comments[:-1].decode(self._encoding)
331 332
332 def addnote(self, text, origin=None, position="append"): 333 def addnote(self, text, origin=None, position="append"):
333 # ignore empty strings and strings without non-space characters 334 # ignore empty strings and strings without non-space characters
334 if (not text) or (not text.strip()): 335 if (not text) or (not text.strip()):
335 return 336 return
336 text = data.forceunicode(text) 337 text = data.forceunicode(text)
337 oldnotes = self.getnotes(origin) 338 oldnotes = self.getnotes(origin)
338 newnotes = None 339 newnotes = None
339 if oldnotes: 340 if oldnotes:
340 if position == "append": 341 if position == "append":
341 newnotes = oldnotes + "\n" + text 342 newnotes = oldnotes + "\n" + text
342 elif position == "merge": 343 elif position == "merge":
343 if oldnotes != text: 344 if oldnotes != text:
344 oldnoteslist = oldnotes.split("\n") 345 oldnoteslist = oldnotes.split("\n")
345 for newline in text.split("\n"): 346 for newline in text.split("\n"):
346 newline = newline.rstrip() 347 newline = newline.rstrip()
347 # avoid duplicate comment lines (this might cause some p roblems) 348 # avoid duplicate comment lines (this might cause some p roblems)
348 if newline not in oldnotes or len(newline) < 5: 349 if newline not in oldnotes or len(newline) < 5:
349 oldnoteslist.append(newline) 350 oldnoteslist.append(newline)
350 newnotes = "\n".join(oldnoteslist) 351 newnotes = "\n".join(oldnoteslist)
351 else: 352 else:
352 newnotes = text + '\n' + oldnotes 353 newnotes = text + '\n' + oldnotes
353 else: 354 else:
354 newnotes = "\n".join([line.rstrip() for line in text.split("\n")]) 355 newnotes = "\n".join([line.rstrip() for line in text.split("\n")])
355 356
356 if newnotes: 357 if newnotes:
357 newlines = [] 358 newlines = []
358 needs_space = get_libgettextpo_version() < (0, 17, 0) 359 needs_space = get_libgettextpo_version() < (0, 17, 0)
359 for line in newnotes.split("\n"): 360 for line in newnotes.split("\n"):
360 if line and needs_space: 361 if line and needs_space:
361 newlines.append(" " + line) 362 newlines.append(" " + line)
362 else: 363 else:
363 newlines.append(line) 364 newlines.append(line)
364 newnotes = "\n".join(newlines) 365 newnotes = "\n".join(newlines)
365 if origin in ["programmer", "developer", "source code"]: 366 if origin in ["programmer", "developer", "source code"]:
366 gpo.po_message_set_extracted_comments(self._gpo_message, newnote s) 367 gpo.po_message_set_extracted_comments(self._gpo_message, newnote s)
367 else: 368 else:
368 gpo.po_message_set_comments(self._gpo_message, newnotes) 369 gpo.po_message_set_comments(self._gpo_message, newnotes)
369 370
370 def removenotes(self): 371 def removenotes(self):
371 gpo.po_message_set_comments(self._gpo_message, "") 372 gpo.po_message_set_comments(self._gpo_message, "")
372 373
373 def copy(self): 374 def copy(self):
374 newpo = self.__class__() 375 """Returns a copy of the this unit that can be used independently.
375 newpo._gpo_message = self._gpo_message 376 """
376 return newpo 377 unit = pounit()
378 unit.source = self.source
379 unit.target = self.target
380 unit.markfuzzy(self.isfuzzy())
381 if self.isobsolete():
382 unit.makeobsolete()
383 for location in self.getlocations():
384 unit.addlocation(location)
385 notes = []
386 origins = ['translator', 'developer', 'programmer', 'source code']
dwaynebailey 2008/08/29 13:19:25 Maybe call them note_types
387 for origin in origins:
388 note = self.getnotes(origin=origin)
389 if note in notes:
390 continue
391 else:
392 unit.addnote(note, origin=origin)
393 notes.append(note)
394 context = gpo.po_message_msgctxt(self._gpo_message)
395 if context:
396 unit.setcontext(context)
397 comments = self._extract_msgidcomments()
398 if comments:
399 unit.setmsgidcomment(comments)
400 for typecomment in self.gettypecomments():
401 unit.addtypecomment(typecomment)
402 return unit
dwaynebailey 2008/08/29 13:19:25 Generally I think this part looks good
377 403
378 def merge(self, otherpo, overwrite=False, comments=True, authoritative=False ): 404 def merge(self, otherpo, overwrite=False, comments=True, authoritative=False ):
379 """Merges the otherpo (with the same msgid) into this one. 405 """Merges the otherpo (with the same msgid) into this one.
380 406
381 Overwrite non-blank self.msgstr only if overwrite is True 407 Overwrite non-blank self.msgstr only if overwrite is True
382 merge comments only if comments is True 408 merge comments only if comments is True
383 409
384 """ 410 """
385 411
386 if not isinstance(otherpo, pounit): 412 if not isinstance(otherpo, pounit):
387 super(pounit, self).merge(otherpo, overwrite, comments) 413 super(pounit, self).merge(otherpo, overwrite, comments)
388 return 414 return
389 if comments: 415 if comments:
390 self.addnote(otherpo.getnotes("translator"), origin="translator", po sition="merge") 416 self.addnote(otherpo.getnotes("translator"), origin="translator", po sition="merge")
391 # FIXME mergelists(self.typecomments, otherpo.typecomments) 417 # FIXME mergelists(self.typecomments, otherpo.typecomments)
392 if not authoritative: 418 if not authoritative:
393 # We don't bring across otherpo.automaticcomments as we consider ourself 419 # We don't bring across otherpo.automaticcomments as we consider ourself
394 # to be the the authority. Same applies to otherpo.msgidcomment s 420 # to be the the authority. Same applies to otherpo.msgidcomment s
395 self.addnote(otherpo.getnotes("developer"), origin="developer", position="merge") 421 self.addnote(otherpo.getnotes("developer"), origin="developer", position="merge")
396 self.msgidcomment = otherpo._extract_msgidcomments() or None 422 self.msgidcomment = otherpo._extract_msgidcomments() or None
397 self.addlocations(otherpo.getlocations()) 423 self.addlocations(otherpo.getlocations())
398 if not self.istranslated() or overwrite: 424 if not self.istranslated() or overwrite:
399 # Remove kde-style comments from the translation (if any). 425 # Remove kde-style comments from the translation (if any).
400 if self._extract_msgidcomments(otherpo.target): 426 if self._extract_msgidcomments(otherpo.target):
401 otherpo.target = otherpo.target.replace('_: ' + otherpo._extract _msgidcomments()+ '\n', '') 427 otherpo.target = otherpo.target.replace('_: ' + otherpo._extract _msgidcomments()+ '\n', '')
402 self.target = otherpo.target 428 self.target = otherpo.target
403 if self.source != otherpo.source: 429 if self.source != otherpo.source:
404 self.markfuzzy() 430 self.markfuzzy()
405 else: 431 else:
406 self.markfuzzy(otherpo.isfuzzy()) 432 self.markfuzzy(otherpo.isfuzzy())
407 elif not otherpo.istranslated(): 433 elif not otherpo.istranslated():
408 if self.source != otherpo.source: 434 if self.source != otherpo.source:
409 self.markfuzzy() 435 self.markfuzzy()
410 else: 436 else:
411 if self.target != otherpo.target: 437 if self.target != otherpo.target:
412 self.markfuzzy() 438 self.markfuzzy()
413 439
414 def isheader(self): 440 def isheader(self):
415 #return self.source == u"" and self.target != u"" 441 #return self.source == u"" and self.target != u""
416 # we really want to make sure that there is no msgidcomment or msgctxt 442 # we really want to make sure that there is no msgidcomment or msgctxt
417 return self.getid() == "" and len(self.target) > 0 443 return self.getid() == "" and len(self.target) > 0
418 444
419 def isblank(self): 445 def isblank(self):
420 return len(self.source) == 0 and len(self.target) == 0 446 return len(self.source) == 0 and len(self.target) == 0
421 447
422 def hastypecomment(self, typecomment): 448 def hastypecomment(self, typecomment):
423 return gpo.po_message_is_format(self._gpo_message, typecomment) 449 return gpo.po_message_is_format(self._gpo_message, typecomment)
450
451 def gettypecomments(self):
452 """Returns the type comments of a unit, if any.
453
454 @return: a list containing the type comments of the unit
455 """
456 #XXX: brute force
457 formats = [
458 "c", "objc", "sh", "python", "lisp", "elisp", "librep", "scheme",
459 "smalltalk", "java", "csharp", "awk", "object-pascal", "ycp", "tcl",
460 "perl", "perl-brace", "php", "gcc-internal", "qt", "kde", "boost",
461 ]
dwaynebailey 2008/08/29 13:19:25 I thought we could get a list of valid formats fro
462 comments = []
463 for format in formats:
464 format += '-format'
465 if self.hastypecomment(format):
466 comments.append(format)
467 return comments
468
469 def addtypecomment(self, typecomment):
470 """Adds a type comment to the unit
471
472 @oaram typecomment: the type comment to be added, for example
473 "c-format".
474 """
475 gpo.po_message_set_format(self._gpo_message, typecomment, 1)
424 476
425 def hasmarkedcomment(self, commentmarker): 477 def hasmarkedcomment(self, commentmarker):
426 commentmarker = "(%s)" % commentmarker 478 commentmarker = "(%s)" % commentmarker
427 for comment in self.getnotes("translator").split("\n"): 479 for comment in self.getnotes("translator").split("\n"):
428 if comment.startswith(commentmarker): 480 if comment.startswith(commentmarker):
429 return True 481 return True
430 return False 482 return False
431 483
432 def istranslated(self): 484 def istranslated(self):
433 return super(pounit, self).istranslated() and not self.isobsolete() 485 return super(pounit, self).istranslated() and not self.isobsolete()
434 486
435 def istranslatable(self): 487 def istranslatable(self):
436 return not (self.isheader() or self.isblank() or self.isobsolete()) 488 return not (self.isheader() or self.isblank() or self.isobsolete())
437 489
438 def isfuzzy(self): 490 def isfuzzy(self):
439 return gpo.po_message_is_fuzzy(self._gpo_message) 491 return gpo.po_message_is_fuzzy(self._gpo_message)
440 492
441 def markfuzzy(self, present=True): 493 def markfuzzy(self, present=True):
442 gpo.po_message_set_fuzzy(self._gpo_message, present) 494 gpo.po_message_set_fuzzy(self._gpo_message, present)
443 495
444 def isreview(self): 496 def isreview(self):
445 return self.hasmarkedcomment("review") or self.hasmarkedcomment("pofilte r") 497 return self.hasmarkedcomment("review") or self.hasmarkedcomment("pofilte r")
446 498
447 def isobsolete(self): 499 def isobsolete(self):
448 return gpo.po_message_is_obsolete(self._gpo_message) 500 return gpo.po_message_is_obsolete(self._gpo_message)
449 501
450 def makeobsolete(self): 502 def makeobsolete(self):
451 # FIXME: libgettexpo currently does not reset other data, we probably wa nt to do that 503 # FIXME: libgettexpo currently does not reset other data, we probably wa nt to do that
452 # but a better solution would be for libgettextpo to output correct data on serialisation 504 # but a better solution would be for libgettextpo to output correct data on serialisation
453 gpo.po_message_set_obsolete(self._gpo_message, True) 505 gpo.po_message_set_obsolete(self._gpo_message, True)
454 506
455 def resurrect(self): 507 def resurrect(self):
456 gpo.po_message_set_obsolete(self._gpo_message, False) 508 gpo.po_message_set_obsolete(self._gpo_message, False)
457 509
458 def hasplural(self): 510 def hasplural(self):
459 return gpo.po_message_msgid_plural(self._gpo_message) is not None 511 return gpo.po_message_msgid_plural(self._gpo_message) is not None
460 512
461 def _extract_msgidcomments(self, text=None): 513 def _extract_msgidcomments(self, text=None):
462 """Extract KDE style msgid comments from the unit. 514 """Extract KDE style msgid comments from the unit.
463 515
464 @rtype: String 516 @rtype: String
465 @return: Returns the extracted msgidcomments found in this unit's msgid. 517 @return: Returns the extracted msgidcomments found in this unit's msgid.
466 518
467 """ 519 """
468 520
469 if not text: 521 if not text:
470 text = gpo.po_message_msgid(self._gpo_message) 522 text = gpo.po_message_msgid(self._gpo_message)
471 if text: 523 if text:
472 msgidcomment = re.search("_: (.*)\n", text) 524 msgidcomment = re.search("_: (.*)\n", text)
473 if msgidcomment: 525 if msgidcomment:
474 return msgidcomment.group(1).decode(self._encoding) 526 return msgidcomment.group(1).decode(self._encoding)
475 return u"" 527 return u""
476 528
477 def __str__(self): 529 def __str__(self):
478 pf = pofile() 530 pf = pofile()
479 pf.addunit(self) 531 unit = self.copy()
532 pf.addunit(unit)
480 return str(pf) 533 return str(pf)
481 534
482 def getlocations(self): 535 def getlocations(self):
483 locations = [] 536 locations = []
484 i = 0 537 i = 0
485 location = gpo.po_message_filepos(self._gpo_message, i) 538 location = gpo.po_message_filepos(self._gpo_message, i)
486 while location: 539 while location:
487 locname = gpo.po_filepos_file(location) 540 locname = gpo.po_filepos_file(location)
488 locline = gpo.po_filepos_start_line(location) 541 locline = gpo.po_filepos_start_line(location)
489 if locline == -1: 542 if locline == -1:
490 locstring = locname 543 locstring = locname
491 else: 544 else:
492 locstring = locname + ":" + str(locline) 545 locstring = locname + ":" + str(locline)
493 locations.append(locstring) 546 locations.append(locstring)
494 i += 1 547 i += 1
495 location = gpo.po_message_filepos(self._gpo_message, i) 548 location = gpo.po_message_filepos(self._gpo_message, i)
496 return locations 549 return locations
497 550
498 def addlocation(self, location): 551 def addlocation(self, location):
499 for loc in location.split(): 552 for loc in location.split():
500 parts = loc.split(":") 553 parts = loc.split(":")
501 file = parts[0] 554 file = parts[0]
502 if len(parts) == 2: 555 if len(parts) == 2:
503 line = int(parts[1]) 556 line = int(parts[1])
504 else: 557 else:
505 line = -1 558 # libgettextpo requires a (size_t)(-1) if no file
559 # location is available
560 line = c_size_t(-1)
506 gpo.po_message_add_filepos(self._gpo_message, file, line) 561 gpo.po_message_add_filepos(self._gpo_message, file, line)
562
563 def setcontext(self, context):
564 """Sets the context message"""
565 gpo.po_message_set_msgctxt(self._gpo_message, context)
dwaynebailey 2008/08/29 13:19:25 Again one piece at a time. If you can separate ou
507 566
508 def getcontext(self): 567 def getcontext(self):
509 msgctxt = gpo.po_message_msgctxt(self._gpo_message) 568 msgctxt = gpo.po_message_msgctxt(self._gpo_message)
510 msgidcomment = self._extract_msgidcomments() 569 msgidcomment = self._extract_msgidcomments()
511 if msgctxt: 570 if msgctxt:
512 return msgctxt + msgidcomment 571 return msgctxt + msgidcomment
513 else: 572 else:
514 return msgidcomment 573 return msgidcomment
515 574
516 class pofile(pocommon.pofile): 575 class pofile(pocommon.pofile):
517 UnitClass = pounit 576 UnitClass = pounit
518 def __init__(self, inputfile=None, encoding=None, unitclass=pounit): 577 def __init__(self, inputfile=None, encoding=None, unitclass=pounit):
519 self.UnitClass = unitclass 578 self.UnitClass = unitclass
520 pocommon.pofile.__init__(self, unitclass=unitclass) 579 pocommon.pofile.__init__(self, unitclass=unitclass)
521 self._gpo_memory_file = None 580 self._gpo_memory_file = None
522 self._gpo_message_iterator = None 581 self._gpo_message_iterator = None
523 self._encoding = encodingToUse(encoding) 582 self._encoding = encodingToUse(encoding)
524 if inputfile is not None: 583 if inputfile is not None:
525 self.parse(inputfile) 584 self.parse(inputfile)
526 else: 585 else:
527 self._gpo_memory_file = gpo.po_file_create() 586 self._gpo_memory_file = gpo.po_file_create()
528 self._gpo_message_iterator = gpo.po_message_iterator(self._gpo_memor y_file, None) 587 self._gpo_message_iterator = gpo.po_message_iterator(self._gpo_memor y_file, None)
529 588
530 def addunit(self, unit): 589 def addunit(self, unit):
531 gpo.po_message_insert(self._gpo_message_iterator, unit._gpo_message) 590 unitcopy = unit.copy()
532 self.units.append(unit) 591 gpo.po_message_insert(self._gpo_message_iterator, unitcopy._gpo_message)
592 self.units.append(unitcopy)
593
594 def addsourceunit(self, source):
595 self.addunit(self.UnitClass(source))
596 unit = self.findunit(source)
597 return unit
533 598
534 def removeduplicates(self, duplicatestyle="merge"): 599 def removeduplicates(self, duplicatestyle="merge"):
535 """make sure each msgid is unique ; merge comments etc from duplicates i nto original""" 600 """make sure each msgid is unique ; merge comments etc from duplicates i nto original"""
536 msgiddict = {} 601 msgiddict = {}
537 uniqueunits = [] 602 uniqueunits = []
538 # we sometimes need to keep track of what has been marked 603 # we sometimes need to keep track of what has been marked
539 # TODO: this is using a list as the pos aren't hashable, but this is slo w... 604 # TODO: this is using a list as the pos aren't hashable, but this is slo w...
540 markedpos = [] 605 markedpos = []
541 def addcomment(thepo): 606 def addcomment(thepo):
542 thepo.msgidcomment = " ".join(thepo.getlocations()) 607 thepo.msgidcomment = " ".join(thepo.getlocations())
543 markedpos.append(thepo) 608 markedpos.append(thepo)
544 for thepo in self.units: 609 for thepo in self.units:
545 if thepo.isheader(): 610 if thepo.isheader():
546 uniqueunits.append(thepo) 611 uniqueunits.append(thepo)
547 continue 612 continue
548 if duplicatestyle.startswith("msgid_comment"): 613 if duplicatestyle.startswith("msgid_comment"):
549 msgid = thepo._extract_msgidcomments() + thepo.source 614 msgid = thepo._extract_msgidcomments() + thepo.source
550 else: 615 else:
551 msgid = thepo.source 616 msgid = thepo.source
552 if duplicatestyle == "msgid_comment_all": 617 if duplicatestyle == "msgid_comment_all":
553 addcomment(thepo) 618 addcomment(thepo)
554 uniqueunits.append(thepo) 619 uniqueunits.append(thepo)
555 elif msgid in msgiddict: 620 elif msgid in msgiddict:
556 if duplicatestyle == "merge": 621 if duplicatestyle == "merge":
557 if msgid: 622 if msgid:
558 msgiddict[msgid].merge(thepo) 623 msgiddict[msgid].merge(thepo)
559 else: 624 else:
560 addcomment(thepo) 625 addcomment(thepo)
561 uniqueunits.append(thepo) 626 uniqueunits.append(thepo)
562 elif duplicatestyle == "keep": 627 elif duplicatestyle == "keep":
563 uniqueunits.append(thepo) 628 uniqueunits.append(thepo)
564 elif duplicatestyle == "msgid_comment": 629 elif duplicatestyle == "msgid_comment":
565 origpo = msgiddict[msgid] 630 origpo = msgiddict[msgid]
566 if origpo not in markedpos: 631 if origpo not in markedpos:
567 addcomment(origpo) 632 addcomment(origpo)
568 addcomment(thepo) 633 addcomment(thepo)
569 uniqueunits.append(thepo) 634 uniqueunits.append(thepo)
570 elif duplicatestyle == "msgctxt": 635 elif duplicatestyle == "msgctxt":
571 origpo = msgiddict[msgid] 636 origpo = msgiddict[msgid]
572 if origpo not in markedpos: 637 if origpo not in markedpos:
573 gpo.po_message_set_msgctxt(origpo._gpo_message, " ".join (origpo.getlocations())) 638 gpo.po_message_set_msgctxt(origpo._gpo_message, " ".join (origpo.getlocations()))
574 markedpos.append(thepo) 639 markedpos.append(thepo)
575 gpo.po_message_set_msgctxt(thepo._gpo_message, " ".join(thep o.getlocations())) 640 gpo.po_message_set_msgctxt(thepo._gpo_message, " ".join(thep o.getlocations()))
576 uniqueunits.append(thepo) 641 uniqueunits.append(thepo)
577 else: 642 else:
578 if not msgid and duplicatestyle != "keep": 643 if not msgid and duplicatestyle != "keep":
579 addcomment(thepo) 644 addcomment(thepo)
580 msgiddict[msgid] = thepo 645 msgiddict[msgid] = thepo
581 uniqueunits.append(thepo) 646 uniqueunits.append(thepo)
582 new_gpo_memory_file = gpo.po_file_create() 647 new_gpo_memory_file = gpo.po_file_create()
583 new_gpo_message_iterator = gpo.po_message_iterator(new_gpo_memory_file, None) 648 new_gpo_message_iterator = gpo.po_message_iterator(new_gpo_memory_file, None)
584 for unit in uniqueunits: 649 for unit in uniqueunits:
585 gpo.po_message_insert(new_gpo_message_iterator, unit._gpo_message) 650 gpo.po_message_insert(new_gpo_message_iterator, unit._gpo_message)
586 gpo.po_message_iterator_free(self._gpo_message_iterator) 651 gpo.po_message_iterator_free(self._gpo_message_iterator)
587 self._gpo_message_iterator = new_gpo_message_iterator 652 self._gpo_message_iterator = new_gpo_message_iterator
588 self._gpo_memory_file = new_gpo_memory_file 653 self._gpo_memory_file = new_gpo_memory_file
589 self.units = uniqueunits 654 self.units = uniqueunits
590 655
591 def __str__(self): 656 def __str__(self):
592 def obsolete_workaround(): 657 def obsolete_workaround():
593 # Remove all items that are not output by msgmerge when a unit is ob solete. This is a work 658 # Remove all items that are not output by msgmerge when a unit is ob solete. This is a work
594 # around for bug in libgettextpo 659 # around for bug in libgettextpo
595 # FIXME Do version test in case they fix this bug 660 # FIXME Do version test in case they fix this bug
596 for unit in self.units: 661 for unit in self.units:
597 if unit.isobsolete(): 662 if unit.isobsolete():
598 gpo.po_message_set_extracted_comments(unit._gpo_message, "") 663 gpo.po_message_set_extracted_comments(unit._gpo_message, "")
599 location = gpo.po_message_filepos(unit._gpo_message, 0) 664 location = gpo.po_message_filepos(unit._gpo_message, 0)
600 while location: 665 while location:
601 gpo.po_message_remove_filepos(unit._gpo_message, 0) 666 gpo.po_message_remove_filepos(unit._gpo_message, 0)
602 location = gpo.po_message_filepos(unit._gpo_message, 0) 667 location = gpo.po_message_filepos(unit._gpo_message, 0)
603 outputstring = "" 668 outputstring = ""
604 if self._gpo_memory_file: 669 if self._gpo_memory_file:
605 obsolete_workaround() 670 obsolete_workaround()
606 f = tempfile.NamedTemporaryFile(prefix='translate', suffix='.po') 671 f = tempfile.NamedTemporaryFile(prefix='translate', suffix='.po')
607 self._gpo_memory_file = gpo.po_file_write_v2(self._gpo_memory_file, f.name, xerror_handler) 672 self._gpo_memory_file = gpo.po_file_write_v2(self._gpo_memory_file, f.name, xerror_handler)
608 f.seek(0) 673 f.seek(0)
609 outputstring = f.read() 674 outputstring = f.read()
610 f.close() 675 f.close()
611 return outputstring 676 return outputstring
612 677
613 def isempty(self): 678 def isempty(self):
614 """Returns True if the object doesn't contain any translation units.""" 679 """Returns True if the object doesn't contain any translation units."""
615 if len(self.units) == 0: 680 if len(self.units) == 0:
616 return True 681 return True
617 # Skip the first unit if it is a header. 682 # Skip the first unit if it is a header.
618 if self.units[0].isheader(): 683 if self.units[0].isheader():
619 units = self.units[1:] 684 units = self.units[1:]
620 else: 685 else:
621 units = self.units 686 units = self.units
622 687
623 for unit in units: 688 for unit in units:
624 if not unit.isblank() and not unit.isobsolete(): 689 if not unit.isblank() and not unit.isobsolete():
625 return False 690 return False
626 return True 691 return True
627 692
628 def parse(self, input): 693 def parse(self, input):
629 if hasattr(input, 'name'): 694 if hasattr(input, 'name'):
630 self.filename = input.name 695 self.filename = input.name
631 elif not getattr(self, 'filename', ''): 696 elif not getattr(self, 'filename', ''):
632 self.filename = '' 697 self.filename = ''
633 698
634 if hasattr(input, "read"): 699 if hasattr(input, "read"):
635 posrc = input.read() 700 posrc = input.read()
636 input.close() 701 input.close()
637 input = posrc 702 input = posrc
638 703
639 needtmpfile = not os.path.isfile(input) 704 needtmpfile = not os.path.isfile(input)
640 if needtmpfile: 705 if needtmpfile:
641 # This is not a file - we write the string to a temporary file 706 # This is not a file - we write the string to a temporary file
642 fd, fname = tempfile.mkstemp(prefix='translate', suffix='.po') 707 fd, fname = tempfile.mkstemp(prefix='translate', suffix='.po')
643 os.write(fd, input) 708 os.write(fd, input)
644 input = fname 709 input = fname
645 os.close(fd) 710 os.close(fd)
646 711
647 self._gpo_memory_file = gpo.po_file_read_v3(input, xerror_handler) 712 self._gpo_memory_file = gpo.po_file_read_v3(input, xerror_handler)
648 if self._gpo_memory_file is None: 713 if self._gpo_memory_file is None:
649 print >> sys.stderr, "Error:" 714 print >> sys.stderr, "Error:"
650 715
651 if needtmpfile: 716 if needtmpfile:
652 os.remove(input) 717 os.remove(input)
653 718
654 # Handle xerrors here 719 # Handle xerrors here
655 self._header = gpo.po_file_domain_header(self._gpo_memory_file, None) 720 self._header = gpo.po_file_domain_header(self._gpo_memory_file, None)
656 if self._header: 721 if self._header:
657 charset = gpo.po_header_field(self._header, "Content-Type") 722 charset = gpo.po_header_field(self._header, "Content-Type")
658 if charset: 723 if charset:
659 charset = re.search("charset=([^\\s]+)", charset).group(1) 724 charset = re.search("charset=([^\\s]+)", charset).group(1)
660 self._encoding = encodingToUse(charset) 725 self._encoding = encodingToUse(charset)
661 self._gpo_message_iterator = gpo.po_message_iterator(self._gpo_memory_fi le, None) 726 self._gpo_message_iterator = gpo.po_message_iterator(self._gpo_memory_fi le, None)
662 newmessage = gpo.po_next_message(self._gpo_message_iterator) 727 newmessage = gpo.po_next_message(self._gpo_message_iterator)
663 while newmessage: 728 while newmessage:
664 newunit = pounit(gpo_message=newmessage) 729 newunit = pounit(gpo_message=newmessage)
665 self.units.append(newunit) 730 self.units.append(newunit)
666 newmessage = gpo.po_next_message(self._gpo_message_iterator) 731 newmessage = gpo.po_next_message(self._gpo_message_iterator)
667 self._free_iterator()
668 732
669 def __del__(self): 733 def __del__(self):
670 # We currently disable this while we still get segmentation faults. 734 # We currently disable this while we still get segmentation faults.
671 # Note that this is definitely leaking memory because of this. 735 # Note that this is definitely leaking memory because of this.
672 return
673 self._free_iterator() 736 self._free_iterator()
674 if self._gpo_memory_file is not None: 737 if self._gpo_memory_file is not None:
675 gpo.po_file_free(self._gpo_memory_file) 738 gpo.po_file_free(self._gpo_memory_file)
676 self._gpo_memory_file = None 739 self._gpo_memory_file = None
677 740
678 def _free_iterator(self): 741 def _free_iterator(self):
679 # We currently disable this while we still get segmentation faults. 742 # We currently disable this while we still get segmentation faults.
680 # Note that this is definitely leaking memory because of this. 743 # Note that this is definitely leaking memory because of this.
681 return
682 if self._gpo_message_iterator is not None: 744 if self._gpo_message_iterator is not None:
683 gpo.po_message_iterator_free(self._gpo_message_iterator) 745 gpo.po_message_iterator_free(self._gpo_message_iterator)
684 self._gpo_message_iterator = None 746 self._gpo_message_iterator = None
OLDNEW

Powered by Google App Engine
This is Rietveld r159