Correct handling of \ as \par for old RTF

src/calibre/ebooks/rtf/input.py

@@ -312,7 +312,6 @@ class RTFInput(InputFormatPlugin):
         try:
             xml = self.generate_xml(stream.name)
         except RtfInvalidCodeException, e:
-            raise
             raise ValueError(_('This RTF file has a feature calibre does not '
             'support. Convert it to HTML first and then try it.\n%s')%e)
 

src/calibre/ebooks/rtf2xml/ParseRtf.py

@@ -226,10 +226,6 @@ class ParseRtf:
         try:
             return_value = process_tokens_obj.process_tokens()
         except InvalidRtfException, msg:
-            try:
-                os.remove(self.__temp_file)
-            except OSError:
-                pass
             #Check to see if the file is correctly encoded
             encode_obj = default_encoding.DefaultEncoding(
             in_file = self.__temp_file,
@@ -244,11 +240,16 @@ class ParseRtf:
             enc = encode_obj.get_codepage()
             if enc != 'mac_roman':
                 enc = 'cp' + enc
+            msg = 'Exception in token processing'
             if check_encoding_obj.check_encoding(self.__file, enc):
                 file_name = self.__file if isinstance(self.__file, str) \
                                     else self.__file.encode('utf-8')
                 msg = 'File %s does not appear to be correctly encoded.\n' % file_name
-                raise InvalidRtfException, msg
+            try:
+                os.remove(self.__temp_file)
+            except OSError:
+                pass
+            raise InvalidRtfException, msg
         delete_info_obj = delete_info.DeleteInfo(
             in_file = self.__temp_file,
             copy = self.__copy,

src/calibre/ebooks/rtf2xml/default_encoding.py

@@ -3,6 +3,7 @@
 #   copyright 2002 Paul Henry Tremblay                                  #
 #                                                                       #
 #########################################################################
+
 '''
 Codepages as to RTF 1.9.1:
     437	United States IBM

src/calibre/ebooks/rtf2xml/process_tokens.py

@@ -70,7 +70,7 @@ class ProcessTokens:
         ';'                  :	('mc', ';', self.ms_sub_func),
         # this must be wrong
         '-'                  :	('mc', '-', self.ms_sub_func),
-        'line'               :  ('mi', 'hardline-break', self.hardline_func), #calibre
+        'line'               :  ('mi', 'hardline-break', self.direct_conv_func), #calibre
         # misc => ml
         '*'                  :	('ml', 'asterisk__', self.default_func),
         ':'                  :	('ml', 'colon_____', self.default_func),
@@ -605,7 +605,7 @@ class ProcessTokens:
     def ms_sub_func(self, pre, token, num):
         return 'tx<mc<__________<%s\n' % token
 
-    def hardline_func(self, pre, token, num):
+    def direct_conv_func(self, pre, token, num):
         return 'mi<tg<empty_____<%s\n' % token
 
     def default_func(self, pre, token, num):

src/calibre/ebooks/rtf2xml/tokenize.py

@@ -27,11 +27,13 @@ class Tokenize:
             bug_handler,
             copy = None,
             run_level = 1,
-    ):
+            # out_file = None,
+        ):
         self.__file = in_file
         self.__bug_handler = bug_handler
         self.__copy = copy
         self.__write_to = tempfile.mktemp()
+        # self.__out_file = out_file
         self.__compile_expressions()
         #variables
         self.__uc_char = 0
@@ -113,6 +115,8 @@ class Tokenize:
 
     def __sub_reg_split(self,input_file):
         input_file = self.__replace_spchar.mreplace(input_file)
+        # this is for older RTF
+        input_file = self.__par_exp.sub('\n\\par \n', input_file)
         input_file = self.__ms_hex_exp.sub("\\mshex0\g<1> ", input_file)
         input_file = self.__utf_ud.sub("\\{\\uc0 \g<1>\\}", input_file)
         #remove \n in bin data
@@ -153,8 +157,6 @@ class Tokenize:
             # put a backslash in front of to eliminate special cases and
             # make processing easier
             "}": "\\}",
-            # this is for older RTF
-            r'\\$': '\\par ',
             }
         self.__replace_spchar = MReplace(SIMPLE_RPL)
         #add ;? in case of char following \u
@@ -168,10 +170,12 @@ class Tokenize:
         #why keep backslash whereas \is replaced before?
         #remove \n from endline char
         self.__splitexp = re.compile(r"(\\[{}]|\n|\\[^\s\\{}&]+(?:[ \t\r\f\v])?)")
+        #this is for old RTF
+        self.__par_exp = re.compile(r'\\\n+')
+        # self.__par_exp = re.compile(r'\\$')
         #self.__bin_exp = re.compile(r"\\bin(-?\d{1,8}) {0,1}")
         #self.__utf_exp = re.compile(r"^\\u(-?\d{3,6})")
         #self.__splitexp = re.compile(r"(\\[\\{}]|{|}|\n|\\[^\s\\{}&]+(?:\s)?)")
-        #self.__par_exp = re.compile(r'\\$')
         #self.__remove_line = re.compile(r'\n+')
         #self.__mixed_exp = re.compile(r"(\\[a-zA-Z]+\d+)(\D+)")
         ##self.num_exp = re.compile(r"(\*|:|[a-zA-Z]+)(.*)")
@@ -199,7 +203,24 @@ class Tokenize:
         copy_obj = copy.Copy(bug_handler = self.__bug_handler)
         if self.__copy:
             copy_obj.copy_file(self.__write_to, "tokenize.data")
+        # if self.__out_file:
+            # self.__file = self.__out_file
         copy_obj.rename(self.__write_to, self.__file)
         os.remove(self.__write_to)
         
         #self.__special_tokens = [ '_', '~', "'", '{', '}' ]
+
+# import sys
+# def main(args=sys.argv):
+    # if len(args) < 1:
+        # print 'No file'
+        # return
+    # file = 'data_tokens.txt'
+    # if len(args) == 3:
+        # file = args[2]
+    # to = Tokenize(args[1], Exception, out_file = file)
+    # to.tokenize()
+
+
+# if __name__ == '__main__':
+    # sys.exit(main())