From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.8) Gecko/20020301 Description of problem: Enscript does setlocale only for LC_MESSAGES, not for LC_CTYPE. That fails with recent glibc versions which tries to convert to the selected character set on-the-fly. The result is non-ASCII character replaced by question marks. For some reason I don't understand, this appears to happen only to strings coming from libc, not enscript's own strings. Version-Release number of selected component (if applicable):1.6.1-16.2 How reproducible: Always Steps to Reproduce: 1.env LANG=fr_FR enscript /etc/passwd -o /apa Actual Results: enscript: Ne peut crier un fichier de sortie "/apa": Permission non accord?e. (In case it get's mangled in the processing, the middle character in the fourth word is an e-acute.) Expected Results: The question mark which is before the e at the end should (also) have been an e-acute. Additional info: I've mentioned this to the developer Markku Rossi <mtr>, so we can hope it will be fixed in some future version of enscript. In the mean time, maybe the simple patch I encolse could be used. I don't dare to do a setlocale(LC_CTYPE,...) since enscript uses ctype functions and I haven't analysed it enough to know if it would be affected. But using the bind_textdomain_codeset() ought to be safe. If this were to be ported to the distribution, one would of course have to adjust the configuration, similar to what I suggested in bug 49182.
Created attachment 48725 [details] Patch to bind the codeset for libc strings.
Fixed package is enscript-1.6.1-18. Thanks.