fix calls to create_oebbook, mark several strings as unicode

This commit is contained in:
ldolse 2011-01-19 00:41:59 +08:00
parent 80065cb443
commit b4c5cd0122
4 changed files with 6 additions and 6 deletions

View File

@ -75,7 +75,7 @@ class CHMInput(InputFormatPlugin):
def _create_oebbook(self, hhcpath, basedir, opts, log, mi):
from calibre.ebooks.conversion.plumber import create_oebbook
from calibre.ebooks.oeb.base import DirContainer
oeb = create_oebbook(log, None, opts, self,
oeb = create_oebbook(log, None, opts,
encoding=opts.input_encoding, populate=False)
self.oeb = oeb

View File

@ -299,9 +299,9 @@ class HeuristicProcessor(object):
supports a range of html markup and text files
'''
# define the pieces of the regex
lookahead = "(?<=.{"+str(length)+"}([a-zäëïöüàèìòùáćéíóńśúâêîôûçąężıãõñæøþðßě,:)\IA\u00DF]|(?<!\&\w{4});))" # (?<!\&\w{4});) is a semicolon not part of an entity
em_en_lookahead = "(?<=.{"+str(length)+"}[\u2013\u2014])"
soft_hyphen = "\xad"
lookahead = "(?<=.{"+str(length)+u"}([a-zäëïöüàèìòùáćéíóńśúâêîôûçąężıãõñæøþðßě,:)\IA\u00DF]|(?<!\&\w{4});))" # (?<!\&\w{4});) is a semicolon not part of an entity
em_en_lookahead = "(?<=.{"+str(length)+u"}[\u2013\u2014])"
soft_hyphen = u"\xad"
line_ending = "\s*</(span|[iubp]|div)>\s*(</(span|[iubp]|div)>)?"
blanklines = "\s*(?P<up2threeblanks><(p|span|div)[^>]*>\s*(<(p|span|div)[^>]*>\s*</(span|p|div)>\s*)</(span|p|div)>\s*){0,3}\s*"
line_opening = "<(span|[iubp]|div)[^>]*>\s*(<(span|[iubp]|div)[^>]*>)?\s*"

View File

@ -295,7 +295,7 @@ class HTMLInput(InputFormatPlugin):
return oeb
from calibre.ebooks.conversion.plumber import create_oebbook
return create_oebbook(log, stream.name, opts, self,
return create_oebbook(log, stream.name, opts,
encoding=opts.input_encoding)
def is_case_sensitive(self, path):

View File

@ -41,7 +41,7 @@ class SNBInput(InputFormatPlugin):
raise ValueError("Invalid SNB file")
log.debug("Handle meta data ...")
from calibre.ebooks.conversion.plumber import create_oebbook
oeb = create_oebbook(log, None, options, self,
oeb = create_oebbook(log, None, options,
encoding=options.input_encoding, populate=False)
meta = snbFile.GetFileStream('snbf/book.snbf')
if meta != None: