Date: Sat, 25 Aug 2012 16:24:02 +0200 From: Steffen "Daode" Nurpmeso <sdaoden@gmail.com> To: freebsd-bugs@freebsd.org Cc: Mark Linimon <linimon@lonesome.com>, Eitan Adler <lists@eitanadler.com> Subject: Re: bin/170651: On 9.0-RELEASE#0 and master sh(1) gobbles high bit at first Message-ID: <5038e002.QK5/RZfJHkeD9ynWYPmTG13O@dietcurd.wild-life.local> In-Reply-To: <CAF6rxg=S32P_E0CnOh4jFPFHhT5WNB1QsYJ5gdD68mJ91YTokQ@mail.gmail.com> References: <201208192010.q7JKAB5F095228@freefall.freebsd.org> <50321e67.a9HQfbVt/xoOsHvAHBd/iFcc@dietcurd.wild-life.local> <CAF6rxg=S32P_E0CnOh4jFPFHhT5WNB1QsYJ5gdD68mJ91YTokQ@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Yes thanks, Mark Linimon told me almost the same in a private mail, but since it seems all those indeed ended up on the list, i guess it was a confidential mismatch ; I've finally found the solution to *my* problem: i did not uncomment my .login_conf and so the initial locale setting was C. Since all the five setlocale() calls that happen before the first prompt appears are actually performed before the sh(1) initialization scripts are loaded any environmental change on LANG and LC_* is not yet incorporated at that time. Stupid me, i only had to copy over the config file from my FreeBSD 5.3 config file repo, but forgot it. (First new in eight years, so i stepped slowly instead of doing a bunch cp(1).) So this is why my patch worked -- libedits map_init_nls() is evaluated again and the isprint(3) therein is true in the then updated locale. But this is not an issue with FreeBSDs sh(1) as this is supposed to work that way -- is it? I.e., prioritization of login.conf(5) against .profile (/.shrc). But anyway this is a different topic, here i was simply too stupid to get that right. So this PR can be closed by someone who can, and i'm rather sorry for the noise. Miau. --steffen
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?5038e002.QK5/RZfJHkeD9ynWYPmTG13O>