Fix unicode string conversion issue in Python 2 (#16168)
* Fix unicode string conversion issue in Python 2 This fix tries to address the issue raised in 16149 where the unicode string conversion with Python 2 does not match the behavior with Python 3. The issue was that in Python 3, TensorFlow tries to do a unicode conversion in UTF8 while in Python 2 the default conversion was used. This fix addresses the issue so that behaviors of TensorFlow with Python 2 and Python 3 match. This fix fixes 16149. Signed-off-by:Yong Tang <yong.tang.github@outlook.com> * Format py_func.cc with clang-foramt -i --style=Google Signed-off-by:
Yong Tang <yong.tang.github@outlook.com> * Add test case for unicode string conversion issue in Python 2 Signed-off-by:
Yong Tang <yong.tang.github@outlook.com> * Update NULL -> `nullptr` Signed-off-by:
Yong Tang <yong.tang.github@outlook.com>
Loading
Please sign in to comment