The reason found:
the real encoding of pages is windows-1254, but encoding detected by 'chardet' module is iso-8859-2.
and there are not encoding declaration in the http header.
Solved:
Update to version 1.20, and set the variable 'TRUST_ENCODING_IN_HEADER_OR_META' to True, keep the variable 'ALWAYS_CHAR_DETECT' is False (default value).
PS:
I can use link 'http://rss.hurriyet.com.tr/rss.aspx?sectionId=9' to capture page, no need to replaced by yahoo pipe. please remember uncheck the switch 'isfulltext' before you add this link to custom rss.
|