On Fri, 24.12.10 09:04, fykc...@gmail.com (fykc...@gmail.com) wrote: > Hi Lennart, > > Thank you for the reply. Still get one question: what does the comment > "Proper handling of /dev/console would be cool" mean?
/dev/console is an alias name for some other tty and the kernel offers no nice way to determine where /dev/console actually points to. Due to that we currently fallback to the safe choice for TERM (i.e. vt100-nav) even if /dev/console might actually point to the much more powerful VT system where "linux" might be a much better choice as TERM. Kay has posted a kernel patch which allows easy detection of the device /dev/console really points to. As it looks right now this will enter the kernel soon. As soon as that in place we can update default_term_for_tty() to handle /dev/console a bit smarter. > I found exec_spawn(src/execute.c) will call default_term_for_tty: > > our_env[n_env++] = strdup(default_term_for_tty(tty_path(context))) > > If I didn't specify TTYPath in .service, then tty_path will return > "/dev/console", which causes default_term_for_tty always to return > "TERM=vt100-nav". Yes, we default to the safe choice of TERM=vt100-nav in case we cannot really determine what kind of tty we are really connected to. That is the only logic that makes sense. Now, as it turns out vt100-nav is not the safe choice we thought it was, since it seems to break for you. Lennart -- Lennart Poettering - Red Hat, Inc. _______________________________________________ systemd-devel mailing list systemd-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/systemd-devel
