LaTeX builder fails with non-ascii characters in production list token names
Since version 1.0.7, or changeset 3211:70f2984f86d7, Sphinx LaTeX writer generates hypertargets in the production list. However when a production list contains non-ascii unicode characters in its token name, LaTeX builder fails with UnicodeEncodeError see attached traceback log.
It seems non-ascii unicode characters are acceptable to reference commands, \ref, \hyperref, \cite, \label, \bibitem. Therefore I propose a removal of str() from idescape in sphinx/writer/latex.py to evade this problem. If it is not acceptable, a removal of non-ascii unicode characters from ids may be required.