The xpath: /root/node[condition]
works as expected, however:
The xpath normalize-space(root/node[condition])
always results in Chinese characters.
Interestingly, this applies for some functions I try to apply (substring-before, substring-after, normalize=space), but not all. The 'count' function works as expected.
This leads me to believe this might be some sort of encoding bug, but I'd just be guessing.