Use a wrapper around xml.etree.ElementTree.fromstring in python 2.x (#7178)

Attributes aren't unicode objects, so they couldn't be directly used in info_dict fields (for example '--write-description' doesn't work with bytes).

9 files changed, 47 insertions, 17 deletions

diff --git a/youtube_dl/compat.py b/youtube_dl/compat.py
index d103ab9ad..cf10835ca 100644
--- a/youtube_dl/compat.py
+++ b/youtube_dl/compat.py
@@ -14,6 +14,7 @@ import socket
 import subprocess
 import sys
 import itertools
+import xml.etree.ElementTree
 
 
 try:
@@ -212,6 +213,29 @@ try:
 except ImportError:  # Python 2.6
     from xml.parsers.expat import ExpatError as compat_xml_parse_error
 
+if sys.version_info[0] >= 3:
+    compat_etree_fromstring = xml.etree.ElementTree.fromstring
+else:
+    # on python 2.x the the attributes of a node are str objects instead of
+    # unicode
+    etree = xml.etree.ElementTree
+
+    # on 2.6 XML doesn't have a parser argument, function copied from CPython
+    # 2.7 source
+    def _XML(text, parser=None):
+        if not parser:
+            parser = etree.XMLParser(target=etree.TreeBuilder())
+        parser.feed(text)
+        return parser.close()
+
+    def _element_factory(*args, **kwargs):
+        el = etree.Element(*args, **kwargs)
+        for k, v in el.items():
+            el.set(k, v.decode('utf-8'))
+        return el
+
+    def compat_etree_fromstring(text):
+        return _XML(text, parser=etree.XMLParser(target=etree.TreeBuilder(element_factory=_element_factory)))
 
 try:
     from urllib.parse import parse_qs as compat_parse_qs
@@ -507,6 +531,7 @@ __all__ = [
     'compat_chr',
     'compat_cookiejar',
     'compat_cookies',
+    'compat_etree_fromstring',
     'compat_expanduser',
     'compat_get_terminal_size',
     'compat_getenv',
diff --git a/youtube_dl/downloader/f4m.py b/youtube_dl/downloader/f4m.py
index 7f6143954..6170cc155 100644
--- a/youtube_dl/downloader/f4m.py
+++ b/youtube_dl/downloader/f4m.py
@@ -5,10 +5,10 @@ import io
 import itertools
 import os
 import time
-import xml.etree.ElementTree as etree
 
 from .fragment import FragmentFD
 from ..compat import (
+    compat_etree_fromstring,
     compat_urlparse,
     compat_urllib_error,
     compat_urllib_parse_urlparse,
@@ -290,7 +290,7 @@ class F4mFD(FragmentFD):
         man_url = urlh.geturl()
         manifest = urlh.read()
 
-        doc = etree.fromstring(manifest)
+        doc = compat_etree_fromstring(manifest)
         formats = [(int(f.attrib.get('bitrate', -1)), f)
                    for f in self._get_unencrypted_media(doc)]
         if requested_bitrate is None:
diff --git a/youtube_dl/extractor/bbc.py b/youtube_dl/extractor/bbc.py
index 2cdce1eb9..a55a6dbc9 100644
--- a/youtube_dl/extractor/bbc.py
+++ b/youtube_dl/extractor/bbc.py
@@ -2,7 +2,6 @@
 from __future__ import unicode_literals
 
 import re
-import xml.etree.ElementTree
 
 from .common import InfoExtractor
 from ..utils import (
@@ -14,7 +13,10 @@ from ..utils import (
     remove_end,
     unescapeHTML,
 )
-from ..compat import compat_HTTPError
+from ..compat import (
+    compat_etree_fromstring,
+    compat_HTTPError,
+)
 
 
 class BBCCoUkIE(InfoExtractor):
@@ -344,7 +346,7 @@ class BBCCoUkIE(InfoExtractor):
                 url, programme_id, 'Downloading media selection XML')
         except ExtractorError as ee:
             if isinstance(ee.cause, compat_HTTPError) and ee.cause.code == 403:
-                media_selection = xml.etree.ElementTree.fromstring(ee.cause.read().decode('utf-8'))
+                media_selection = compat_etree_fromstring(ee.cause.read().decode('utf-8'))
             else:
                 raise
         return self._process_media_selector(media_selection, programme_id)
diff --git a/youtube_dl/extractor/bilibili.py b/youtube_dl/extractor/bilibili.py
index ecc17ebeb..6c66a1236 100644
--- a/youtube_dl/extractor/bilibili.py
+++ b/youtube_dl/extractor/bilibili.py
@@ -4,9 +4,11 @@ from __future__ import unicode_literals
 import re
 import itertools
 import json
-import xml.etree.ElementTree as ET
 
 from .common import InfoExtractor
+from ..compat import (
+    compat_etree_fromstring,
+)
 from ..utils import (
     int_or_none,
     unified_strdate,
@@ -88,7 +90,7 @@ class BiliBiliIE(InfoExtractor):
         except ValueError:
             pass
 
-        lq_doc = ET.fromstring(lq_page)
+        lq_doc = compat_etree_fromstring(lq_page)
         lq_durls = lq_doc.findall('./durl')
 
         hq_doc = self._download_xml(
diff --git a/youtube_dl/extractor/brightcove.py b/youtube_dl/extractor/brightcove.py
index 4721c2293..1686cdde1 100644
--- a/youtube_dl/extractor/brightcove.py
+++ b/youtube_dl/extractor/brightcove.py
@@ -3,10 +3,10 @@ from __future__ import unicode_literals
 
 import re
 import json
-import xml.etree.ElementTree
 
 from .common import InfoExtractor
 from ..compat import (
+    compat_etree_fromstring,
     compat_parse_qs,
     compat_str,
     compat_urllib_parse,
@@ -119,7 +119,7 @@ class BrightcoveIE(InfoExtractor):
         object_str = fix_xml_ampersands(object_str)
 
         try:
-            object_doc = xml.etree.ElementTree.fromstring(object_str.encode('utf-8'))
+            object_doc = compat_etree_fromstring(object_str.encode('utf-8'))
         except compat_xml_parse_error:
             return
 
diff --git a/youtube_dl/extractor/common.py b/youtube_dl/extractor/common.py
index 10c0d5d1f..52523d7b2 100644
--- a/youtube_dl/extractor/common.py
+++ b/youtube_dl/extractor/common.py
@@ -10,7 +10,6 @@ import re
 import socket
 import sys
 import time
-import xml.etree.ElementTree
 
 from ..compat import (
     compat_cookiejar,
@@ -23,6 +22,7 @@ from ..compat import (
     compat_urllib_request,
     compat_urlparse,
     compat_str,
+    compat_etree_fromstring,
 )
 from ..utils import (
     NO_DEFAULT,
@@ -461,7 +461,7 @@ class InfoExtractor(object):
             return xml_string
         if transform_source:
             xml_string = transform_source(xml_string)
-        return xml.etree.ElementTree.fromstring(xml_string.encode('utf-8'))
+        return compat_etree_fromstring(xml_string.encode('utf-8'))
 
     def _download_json(self, url_or_request, video_id,
                        note='Downloading JSON metadata',
diff --git a/youtube_dl/extractor/crunchyroll.py b/youtube_dl/extractor/crunchyroll.py
index f8ce10111..0c9b8ca02 100644
--- a/youtube_dl/extractor/crunchyroll.py
+++ b/youtube_dl/extractor/crunchyroll.py
@@ -5,12 +5,12 @@ import re
 import json
 import base64
 import zlib
-import xml.etree.ElementTree
 
 from hashlib import sha1
 from math import pow, sqrt, floor
 from .common import InfoExtractor
 from ..compat import (
+    compat_etree_fromstring,
     compat_urllib_parse,
     compat_urllib_parse_unquote,
     compat_urllib_request,
@@ -234,7 +234,7 @@ Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text
         return output
 
     def _extract_subtitles(self, subtitle):
-        sub_root = xml.etree.ElementTree.fromstring(subtitle)
+        sub_root = compat_etree_fromstring(subtitle)
         return [{
             'ext': 'srt',
             'data': self._convert_subtitles_to_srt(sub_root),
diff --git a/youtube_dl/extractor/vevo.py b/youtube_dl/extractor/vevo.py
index c17094f81..4c0de354f 100644
--- a/youtube_dl/extractor/vevo.py
+++ b/youtube_dl/extractor/vevo.py
@@ -1,10 +1,10 @@
 from __future__ import unicode_literals
 
 import re
-import xml.etree.ElementTree
 
 from .common import InfoExtractor
 from ..compat import (
+    compat_etree_fromstring,
     compat_urllib_request,
 )
 from ..utils import (
@@ -97,7 +97,7 @@ class VevoIE(InfoExtractor):
         if last_version['version'] == -1:
             raise ExtractorError('Unable to extract last version of the video')
 
-        renditions = xml.etree.ElementTree.fromstring(last_version['data'])
+        renditions = compat_etree_fromstring(last_version['data'])
         formats = []
         # Already sorted from worst to best quality
         for rend in renditions.findall('rendition'):
@@ -114,7 +114,7 @@ class VevoIE(InfoExtractor):
 
     def _formats_from_smil(self, smil_xml):
         formats = []
-        smil_doc = xml.etree.ElementTree.fromstring(smil_xml.encode('utf-8'))
+        smil_doc = compat_etree_fromstring(smil_xml.encode('utf-8'))
         els = smil_doc.findall('.//{http://www.w3.org/2001/SMIL20/Language}video')
         for el in els:
             src = el.attrib['src']
diff --git a/youtube_dl/utils.py b/youtube_dl/utils.py
index a61e47646..7d846d680 100644
--- a/youtube_dl/utils.py
+++ b/youtube_dl/utils.py
@@ -36,6 +36,7 @@ import zlib
 from .compat import (
     compat_basestring,
     compat_chr,
+    compat_etree_fromstring,
     compat_html_entities,
     compat_http_client,
     compat_kwargs,
@@ -1974,7 +1975,7 @@ def dfxp2srt(dfxp_data):
 
         return out
 
-    dfxp = xml.etree.ElementTree.fromstring(dfxp_data.encode('utf-8'))
+    dfxp = compat_etree_fromstring(dfxp_data.encode('utf-8'))
     out = []
     paras = dfxp.findall(_x('.//ttml:p')) or dfxp.findall(_x('.//ttaf1:p')) or dfxp.findall('.//p')