Cutlery/calibre
1
0
Fork 0
mirror of https://github.com/kovidgoyal/calibre.git synced 2025-06-23 15:30:45 -04:00
Code Issues Packages Projects Releases Wiki Activity

Fix #2630 (lxml.etree.XMLSyntaxError: error parsing attribute name, line 5, column 1340)

Browse Source
This commit is contained in:
Kovid Goyal 2009-06-17 12:40:59 -07:00
parent 4c586ff916
commit d0be2b91a9
1 changed files with 9 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
src/calibre/ebooks/oeb/reader.py
View File

@ -6,7 +6,7 @@ from __future__ import with_statement
__license__ = 'GPL v3' __license__ = 'GPL v3'
__copyright__ = '2008, Marshall T. Vandegrift <llasram@gmail.com>' __copyright__ = '2008, Marshall T. Vandegrift <llasram@gmail.com>'
import sys, os, uuid, copy import sys, os, uuid, copy, re
from itertools import izip from itertools import izip
from urlparse import urldefrag, urlparse from urlparse import urldefrag, urlparse
from urllib import unquote as urlunquote from urllib import unquote as urlunquote
@ -107,8 +107,14 @@ class OEBReader(object):
except etree.XMLSyntaxError: except etree.XMLSyntaxError:
repl = lambda m: ENTITYDEFS.get(m.group(1), m.group(0)) repl = lambda m: ENTITYDEFS.get(m.group(1), m.group(0))
data = ENTITY_RE.sub(repl, data) data = ENTITY_RE.sub(repl, data)
opf = etree.fromstring(data) try:
self.logger.warn('OPF contains invalid HTML named entities') opf = etree.fromstring(data)
self.logger.warn('OPF contains invalid HTML named entities')
except etree.XMLSyntaxError:
data = re.sub(r'(?is)<tours>.+</tours>', '', data)
self.logger.warn('OPF contains invalid tours section')
opf = etree.fromstring(data)
ns = namespace(opf.tag) ns = namespace(opf.tag)
if ns not in ('', OPF1_NS, OPF2_NS): if ns not in ('', OPF1_NS, OPF2_NS):
raise OEBError('Invalid namespace %r for OPF document' % ns) raise OEBError('Invalid namespace %r for OPF document' % ns)