[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#673207: /etc/default/rcS is perhaps a better place for FANCYTTY variable



Hi Stanislav, and thanks for your bugreport, 
(CC'ing the sysvinit maintainers),

Le jeudi, 17 mai 2012 00.19:38, vous avez écrit :
> Perhaps, /etc/default/rcS could be a better place for setting FANCYTTY
> variable, because this place is documented (man rcS) and also there is
> a variable VERBOSE that has a similar purpose and is defined in this
> file. The file /etc/default/rcS also exists by default on the newly
> installed Debian, and, thus, changing the way how the messages are
> displayed can be less confusing for users.
> 
> Evidently, if you decide to use /etc/default/rcS for this setting,
> this file should *not* be simply sourced from /lib/lsb/init-functions.
> Instead, you may use something like
> 
> FANCYTTY=$([ -e /etc/default/rcS ] && . /etc/default/rcS && echo $FANCYTTY)

After an initial perplexity, I think I now quite like this idea. From the rcS 
manpage: "rcS - variables that affect the behavior of boot scripts", so it 
would be a good fit for /etc/default/rcS.

Given that /etc/default/rcS is managed by initscripts (from src:sysvinit), I'm 
hereby asking the sysvinit maintainers: what do you think?

That said, my non-ordered and open-ended list of concerns is:

a) migrating FANCYTTY away from non-conffile /etc/lsb-base-logging.sh;
b) coping with the fact that FANCYTTY is a variable that works reversely (see 
log_use_fancy_tty function from /lib/lsb/init-functions): when /not/ defined, 
the functions use "fancy output" if possible; when defined to whatever 
negative (actually not in 1|Y|yes|true ), the functions avoid using the "fancy 
output").
c) what might I have forgotten?

Cheers,

OdyX



Reply to: