Merge pull request #1100 from MITx/fix/brian/nonascii-html-xmodule

Fix/brian/nonascii html xmodule
This commit is contained in:
David Ormsbee
2012-12-06 15:33:21 -08:00
37 changed files with 136 additions and 151 deletions

View File

@@ -149,14 +149,14 @@ class ErrorDescriptor(JSONEditingDescriptor):
'''
try:
xml = etree.fromstring(self.definition['data']['contents'])
return etree.tostring(xml)
return etree.tostring(xml, encoding='unicode')
except etree.XMLSyntaxError:
# still not valid.
root = etree.Element('error')
root.text = self.definition['data']['contents']
err_node = etree.SubElement(root, 'error_msg')
err_node.text = self.definition['data']['error_msg']
return etree.tostring(root)
return etree.tostring(root, encoding='unicode')
class NonStaffErrorDescriptor(ErrorDescriptor):

View File

@@ -7,15 +7,14 @@ from lxml import etree
from lxml.html import rewrite_links
from path import path
from .x_module import XModule
from pkg_resources import resource_string
from .xml_module import XmlDescriptor, name_to_pathname
from .editing_module import EditingDescriptor
from .stringify import stringify_children
from .html_checker import check_html
from xmodule.modulestore import Location
from xmodule.contentstore.content import XASSET_SRCREF_PREFIX, StaticContent
from xmodule.editing_module import EditingDescriptor
from xmodule.html_checker import check_html
from xmodule.modulestore import Location
from xmodule.stringify import stringify_children
from xmodule.x_module import XModule
from xmodule.xml_module import XmlDescriptor, name_to_pathname
log = logging.getLogger("mitx.courseware")
@@ -123,7 +122,7 @@ class HtmlDescriptor(XmlDescriptor, EditingDescriptor):
try:
with system.resources_fs.open(filepath) as file:
html = file.read()
html = file.read().decode('utf-8')
# Log a warning if we can't parse the file, but don't error
if not check_html(html):
msg = "Couldn't parse html in {0}.".format(filepath)
@@ -164,7 +163,7 @@ class HtmlDescriptor(XmlDescriptor, EditingDescriptor):
resource_fs.makedir(os.path.dirname(filepath), allow_recreate=True)
with resource_fs.open(filepath, 'w') as file:
file.write(self.definition['data'])
file.write(self.definition['data'].encode('utf-8'))
# write out the relative name
relname = path(pathname).basename()

View File

@@ -152,7 +152,7 @@ class ImportSystem(XMLParsingSystem, MakoDescriptorSystem):
make_name_unique(xml_data)
descriptor = XModuleDescriptor.load_from_xml(
etree.tostring(xml_data), self, self.org,
etree.tostring(xml_data, encoding='unicode'), self, self.org,
self.course, xmlstore.default_class)
except Exception as err:
print err, self.load_error_modules
@@ -436,7 +436,7 @@ class XMLModuleStore(ModuleStoreBase):
self.load_error_modules,
)
course_descriptor = system.process_xml(etree.tostring(course_data))
course_descriptor = system.process_xml(etree.tostring(course_data, encoding='unicode'))
# NOTE: The descriptors end up loading somewhat bottom up, which
# breaks metadata inheritance via get_children(). Instead

View File

@@ -13,7 +13,7 @@ class RawDescriptor(XmlDescriptor, XMLEditingDescriptor):
"""
@classmethod
def definition_from_xml(cls, xml_object, system):
return {'data': etree.tostring(xml_object, pretty_print=True)}
return {'data': etree.tostring(xml_object, pretty_print=True,encoding='unicode')}
def definition_to_xml(self, resource_fs):
try:

View File

@@ -124,7 +124,7 @@ class SequenceDescriptor(MakoModuleDescriptor, XmlDescriptor):
children = []
for child in xml_object:
try:
children.append(system.process_xml(etree.tostring(child)).location.url())
children.append(system.process_xml(etree.tostring(child, encoding='unicode')).location.url())
except:
log.exception("Unable to load child when parsing Sequence. Continuing...")
continue

View File

@@ -22,7 +22,7 @@ def stringify_children(node):
# next element.
parts = [node.text]
for c in node.getchildren():
parts.append(etree.tostring(c, with_tail=True))
parts.append(etree.tostring(c, with_tail=True, encoding='unicode'))
# filter removes possible Nones in texts and tails
return ''.join(filter(None, parts))
return u''.join(filter(None, parts))

View File

@@ -58,7 +58,7 @@ class CustomTagDescriptor(RawDescriptor):
params = dict(xmltree.items())
with system.resources_fs.open('custom_tags/{name}'
.format(name=template_name)) as template:
return Template(template.read()).render(**params)
return Template(template.read().decode('utf-8')).render(**params)
def __init__(self, system, definition, **kwargs):

View File

@@ -1,19 +1,20 @@
from xmodule.x_module import (XModuleDescriptor, policy_key)
from xmodule.modulestore import Location
from lxml import etree
import json
import copy
import logging
import traceback
from collections import namedtuple
from fs.errors import ResourceNotFoundError
import os
import sys
from collections import namedtuple
from lxml import etree
from xmodule.x_module import (XModuleDescriptor, policy_key)
from xmodule.modulestore import Location
log = logging.getLogger(__name__)
# assume all XML files are persisted as utf-8.
edx_xml_parser = etree.XMLParser(dtd_validation=False, load_dtd=False,
remove_comments=True, remove_blank_text=True)
remove_comments=True, remove_blank_text=True,
encoding='utf-8')
def name_to_pathname(name):
"""
@@ -366,7 +367,7 @@ class XmlDescriptor(XModuleDescriptor):
filepath = self.__class__._format_filepath(self.category, url_path)
resource_fs.makedir(os.path.dirname(filepath), allow_recreate=True)
with resource_fs.open(filepath, 'w') as file:
file.write(etree.tostring(xml_object, pretty_print=True))
file.write(etree.tostring(xml_object, pretty_print=True, encoding='utf-8'))
# And return just a pointer with the category and filename.
record_object = etree.Element(self.category)
@@ -381,7 +382,7 @@ class XmlDescriptor(XModuleDescriptor):
record_object.set('org', self.location.org)
record_object.set('course', self.location.course)
return etree.tostring(record_object, pretty_print=True)
return etree.tostring(record_object, pretty_print=True, encoding='utf-8')
def definition_to_xml(self, resource_fs):
"""