Remove C printf formatters for spell checking

Author: flashcode

ChangeLog.asciidoc

@@ -4,6 +4,10 @@
 :lang: en
 
 
+== Version 2.9 (under dev)
+
+* remove C printf formatters for spell checking
+
 == Version 2.8 (2014-12-07)
 
 * fix read of fuzzy flag

msgcheck/po.py

@@ -40,15 +40,7 @@
 except ImportError:
     pass
 
-
-def count_lines(string):
-    """
-    Count the number of lines in a string or translation.
-    """
-    count = len(string.split('\n'))
-    if count > 1 and string.endswith('\n'):
-        count -= 1
-    return count
+from . utils import count_lines, replace_formatters
 
 
 # pylint: disable=too-few-public-methods
@@ -119,7 +111,7 @@ class PoMessage(object):
     """
 
     # pylint: disable=too-many-arguments
-    def __init__(self, filename, line, msg, charset, fuzzy):
+    def __init__(self, filename, line, msg, charset, fuzzy, fmt):
         """Build a PO message."""
         self.filename = filename
         self.line = line
@@ -144,6 +136,7 @@ def __init__(self, filename, line, msg, charset, fuzzy):
         else:
             self.messages.append((msg.get('msgid', ''), msg.get('msgstr', '')))
         self.fuzzy = fuzzy
+        self.fmt = fmt
 
     def check_lines(self):
         """
@@ -271,7 +264,10 @@ def check_spelling(self, spelling, checkers):
         for mid, mstr in self.messages:
             if not mid or not mstr:
                 continue
-            checkers[0].set_text(mstr if spelling == 'str' else mid)
+            text = mstr if spelling == 'str' else mid
+            if self.fmt:
+                text = replace_formatters(text, ' ', self.fmt)
+            checkers[0].set_text(text)
             misspelled = []
             for err in checkers[0]:
                 misspelled_word = True
@@ -302,7 +298,7 @@ def __init__(self, filename):
         }
         self.msgs = []
 
-    def _add_message(self, numline_msgid, msgfuzzy, msg):
+    def _add_message(self, numline_msgid, fuzzy, fmt, msg):
         """
         Add a message from PO file in list of messages.
         """
@@ -319,15 +315,16 @@ def _add_message(self, numline_msgid, msgfuzzy, msg):
             if match:
                 self.props['charset'] = match.group(1)
         self.msgs.append(PoMessage(self.filename, numline_msgid, msg,
-                                   self.props['charset'], msgfuzzy))
+                                   self.props['charset'], fuzzy, fmt))
 
     def read(self):
         """
         Read messages in PO file.
         """
         self.msgs = []
-        (numline, numline_msgid) = (0, 0)
-        (fuzzy, msgfuzzy) = (False, False)
+        numline, numline_msgid = (0, 0)
+        fuzzy, msgfuzzy = (False, False)
+        fmt, msgfmt = (None, None)
         msg = {}
         msgcurrent = ''
         with open(self.filename, 'r') as po_file:
@@ -338,6 +335,9 @@ def read(self):
                     continue
                 if line.startswith('#,'):
                     fuzzy = 'fuzzy' in line
+                    match = re.search(r'([a-z-]+)-format', line, re.IGNORECASE)
+                    fmt = match.group(1) if match else None
+                if line.startswith('#'):
                     continue
                 if line.startswith('msg'):
                     match = re.match(
@@ -351,16 +351,20 @@ def read(self):
                             if oldmsgcurrent.startswith('msgstr'):
                                 self._add_message(numline_msgid,
                                                   msgfuzzy,
+                                                  msgfmt,
                                                   msg)
                             msgfuzzy = fuzzy
                             fuzzy = False
+                            msgfmt = fmt
+                            fmt = None
                             msg = {}
                             numline_msgid = numline
                 if msgcurrent and line.startswith('"'):
                     msg[msgcurrent] = msg.get(msgcurrent, '') + line[1:-1]
             if msgcurrent.startswith('msgstr'):
                 self._add_message(numline_msgid,
                                   msgfuzzy,
+                                  msgfmt,
                                   msg)
 
     def compile(self):

msgcheck/utils.py

@@ -0,0 +1,83 @@
+# -*- coding: utf-8 -*-
+#
+# Copyright (C) 2009-2015 Sébastien Helleu <flashcode@flashtux.org>
+#
+# This file is part of msgcheck.
+#
+# Msgcheck is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 3 of the License, or
+# (at your option) any later version.
+#
+# Msgcheck is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with msgcheck.  If not, see <http://www.gnu.org/licenses/>.
+#
+
+"""
+Some utility functions for msgcheck.
+"""
+
+from __future__ import print_function
+
+
+# TODO: add support for other languages
+STR_FORMATTERS = {
+    'c': ('\\', '%', '#- +\'I.0123456789hlLqjzt', 'diouxXeEfFgGaAcsCSpnm'),
+}
+
+def count_lines(string):
+    """Count the number of lines in a string or translation."""
+    count = len(string.split('\n'))
+    if count > 1 and string.endswith('\n'):
+        count -= 1
+    return count
+
+
+def replace_formatters(string, replace, fmt):
+    """
+    Replace formatters (like "%s" or "%03d") with a replacement string.
+    """
+    if fmt not in STR_FORMATTERS:
+        return string
+    formatters = STR_FORMATTERS[fmt]
+    formatter, escape = (False, False)
+    strformat = []
+    result = []
+
+    for char in string:
+        if formatter:
+            if char == formatters[1]:
+                result.append(char)
+                formatter = False
+            elif char in formatters[2]:
+                strformat.append(char)
+            elif char in formatters[3]:
+                result.append(replace)
+                formatter = False
+            else:
+                strformat.append(char)
+                result += strformat
+                formatter = False
+        elif escape:
+            result.append(formatters[0])
+            result.append(char)
+            escape = False
+        elif char == formatters[0]:
+            escape = True
+        elif char == formatters[1]:
+            formatter = True
+            strformat = [char]
+        else:
+            result.append(char)
+
+    if escape:  # unterminated escaped char?
+        result.append(formatters[0])
+    elif formatter:  # unterminated formatter?
+        result.append(replace)
+
+    return ''.join(result)

tests/fr_spelling_id.po

@@ -35,8 +35,9 @@ msgstr ""
 "Content-Transfer-Encoding: 8bit\n"
 "Plural-Forms: nplurals=2; plural=(n > 1);\n"
 
-msgid "Thsi is a test.\n"
-msgstr "Ceci est un test.\n"
+#, c-format
+msgid "%.3fThsi is a test.\n"
+msgstr "%.3fCeci est un test.\n"
 
 msgid "Test 1"
 msgstr "Test 1"

tests/fr_spelling_str.po

@@ -35,8 +35,13 @@ msgstr ""
 "Content-Transfer-Encoding: 8bit\n"
 "Plural-Forms: nplurals=2; plural=(n > 1);\n"
 
-msgid "This is a test.\n"
-msgstr "Ceci est un test.\n"
+#, c-format
+msgid "%.3fThis is a test.\n"
+msgstr "%.3fCeci est un test.\n"
+
+#, c-format
+msgid "%.3fThis is another test.\n"
+msgstr "%.3fCecX est un autre test.\n"
 
 msgid "Test 1"
 msgstr "aabbcc"

tests/test_msgcheck.py

@@ -26,6 +26,7 @@
 import unittest
 
 from msgcheck.po import PoFile, PoCheck
+from msgcheck.utils import replace_formatters
 
 
 def local_path(filename):
@@ -95,6 +96,21 @@ def test_checks_fuzzy(self):
         # the file has 11 errors (with the fuzzy string)
         self.assertEquals(len(result[0][1]), 11)
 
+    def test_replace_formatters(self):
+        """Test removal of formatters in a string."""
+        self.assertEquals(replace_formatters('%', '', 'c'), '')
+        self.assertEquals(replace_formatters('\\', '', 'c'), '\\')
+        self.assertEquals(replace_formatters('%s', ' ', 'c'), ' ')
+        self.assertEquals(replace_formatters('%.02f', ' ', 'c'), ' ')
+        self.assertEquals(replace_formatters('%!%s%!', '', 'c'), '%!%!')
+        self.assertEquals(replace_formatters('%.02!', ' ', 'c'), '%.02!')
+        self.assertEquals(
+            replace_formatters('%.3fThis is a %stest', ' ', 'c'),
+            ' This is a  test')
+        self.assertEquals(
+            replace_formatters('%.3fTest%s%d%%%.03f%luhere% s', '', 'c'),
+            'Test%here')
+
     def test_spelling_id(self):
         """Test spelling on source messages (English) of gettext files."""
         po_check = PoCheck()
@@ -104,7 +120,7 @@ def test_spelling_id(self):
         # be sure we have 1 file in result
         self.assertEquals(len(result), 1)
 
-        # the file has 2 spelling errors: words "Thsi" and "errro"
+        # the file has 2 spelling errors: "Thsi" and "errro"
         errors = result[0][1]
         self.assertEquals(len(errors), 2)
         for i, word in enumerate(('Thsi', 'errro')):
@@ -123,10 +139,10 @@ def test_spelling_str(self):
         # be sure we have 2 files in result
         self.assertEquals(len(result), 2)
 
-        # first file has 2 spelling errors: words "aabbcc" and "xxyyzz"
+        # first file has 3 spelling errors: "CecX", "aabbcc" and "xxyyzz"
         errors = result[0][1]
-        self.assertEquals(len(errors), 2)
-        for i, word in enumerate(('aabbcc', 'xxyyzz')):
+        self.assertEquals(len(errors), 3)
+        for i, word in enumerate(('CecX', 'aabbcc', 'xxyyzz')):
             self.assertEquals(errors[i].idmsg, 'spelling-str')
             self.assertTrue(type(errors[i].message) is list)
             self.assertEquals(len(errors[i].message), 1)