It seems EDITOR variable is unaccessible from some initd scripts (unless the rc_env_allow="EDITOR" is set in /etc/rc.conf) Builds could be affected: uwsgi, spawn-fcgi Reproducible: Always Steps to reproduce: 1. emerge www-servers/uwsgi 2. /etc/init.d/uwsgi start As the result, error messages shown lack completeness, e.g. for uwsgi-1.2.5 : $ /etc/init.d/uwsgi start * Caching service dependencies ... [ ok ] * You are not supposed to run this script directly unless you * want to run in Emperor mode. In that case please set the UWSGI_EMPEROR_PATH. * Otherwise create a symlink for the uwsgi application you want to run as well as * a copy of the configuration file and modify it appropriately like so... * ln -s uwsgi /etc/init.d/uwsgi.trac * cp /etc/conf.d/uwsgi /etc/conf.d/uwsgi.trac * /etc/conf.d/uwsgi.trac * ERROR: uwsgi failed to start Line before last line should read (if EDITOR == vim): * vim /etc/conf.d/uwsgi.trac Problematic source line is (/etc/init.d/uwsgi): eerror " `basename "${EDITOR}"` /etc/conf.d/uwsgi.trac"
It shouldn't bother with that kind of hand-holding at all. People generally know their (preferred) $EDITOR so there's no point in reminding them.
Yes, it sounds like minor problem, but nevertheless, it required me to look at script's source to understand what a heck the suggested sequence of commands was trying to express (first time uwsgi launcher, myself): * cp /etc/conf.d/uwsgi /etc/conf.d/uwsgi.trac * /etc/conf.d/uwsgi.trac
Maybe just hard-code "nano" as the editor of the masses. As jer suggests, anybody who knows what they are doing can make the translation.
You're right guys, thanks Andrius for pointing this out. + 27 Aug 2012; Ultrabug <ultrabug@gentoo.org> files/uwsgi.initd, + files/uwsgi.initd-r1: + hard-code EDITOR in init scripts wrt #432606. +