I'm using org-mode to write a paper and export it to a pdf. It compiles just
fine.
However, when opening the experted pdf (with evince or okular) and searching
for a word, the output from the search function is a list of words with
apparently incorrect character encoding.
I get the same result when compiling directly the exported tex file. However,
if I remove the line:
\usepackage[T1]{fontenc}
the compiled pdf becomes correctly interpreted by the search function. According to the
documentation of the org-article latex class, "There is no facility to disable
loading fontenc".