diff options
author | Sergey M․ <dstftw@gmail.com> | 2015-07-14 22:36:30 +0600 |
---|---|---|
committer | Sergey M․ <dstftw@gmail.com> | 2015-07-14 22:36:30 +0600 |
commit | f8da79f828637757889f3f35d7adfa9aabbfc721 (patch) | |
tree | c95160001699c868f5157300e6f4839bab11bff2 /youtube_dl/extractor/common.py | |
parent | 9750e7d70eed92a6b05637465698cdd30e87a44c (diff) |
[extractor/common] Improve _form_hidden_inputs and rename to _hidden_inputs
Diffstat (limited to 'youtube_dl/extractor/common.py')
-rw-r--r-- | youtube_dl/extractor/common.py | 15 |
1 files changed, 11 insertions, 4 deletions
diff --git a/youtube_dl/extractor/common.py b/youtube_dl/extractor/common.py index 82f5de2d8..315fe4a72 100644 --- a/youtube_dl/extractor/common.py +++ b/youtube_dl/extractor/common.py @@ -706,10 +706,17 @@ class InfoExtractor(object): 'twitter card player') @staticmethod - def _form_hidden_inputs(html): - return dict(re.findall( - r'<input\s+type="hidden"\s+name="([^"]+)"\s+(?:id="[^"]+"\s+)?value="([^"]*)"', - html)) + def _hidden_inputs(html): + return dict([ + (input.group('name'), input.group('value')) for input in re.finditer( + r'''(?x) + <input\s+ + type=(?P<q_hidden>["\'])hidden(?P=q_hidden)\s+ + name=(?P<q_name>["\'])(?P<name>.+?)(?P=q_name)\s+ + (?:id=(?P<q_id>["\']).+?(?P=q_id)\s+)? + value=(?P<q_value>["\'])(?P<value>.*?)(?P=q_value) + ''', html) + ]) def _sort_formats(self, formats, field_preference=None): if not formats: |