[11:16] hi again [11:16] does start on stopping networking [11:17] block networking until the job where it is defined finishes [11:17] because it seems it's not [11:31] pmjdebruijn: yes - as shown in upstart-events(8), "stopping" is a hook type event, which blocks. [11:39] hmmm very odd [11:40] start on (stopping networking) should execute this script before bringing down the networking === Md_ is now known as Md [11:41] oh wait [11:41] networking is a boot only job [11:42] which explains me issue [11:42] networking down is an rc job [11:42] hmmm [11:51] jhunt: thanks [19:50] hello, i am looking for a way to modify environment of upstart itself, so that all children will get certain environment variables. i.e. global export [19:51] either that, or some way to run a job that exports stuff in the environment prior to execution of any job [19:52] i read section on environment in cookbook, and in manual, and doesn't seem to be a way. i'd have to manually change the environment in all job scripts [19:53] which is enormous duplication of course... [19:53] any way to do this? [22:48] vaxerdec: interesting idea... while I look at the source code to see whats possible.. what environment variable do you want to override? [22:56] vaxerdec: far as I see its not really doable in an obvious way [23:26] it's perfect candidate for use of /etc/init/init.conf [23:26] unless there are some other plans for that [23:27] well in particular i want SYSTEMCTL_SKIP_REDIRECT=1 and DBUS_SESSION_BUS_ADDRESS= [23:29] because (1) i have to contend with some systemd infection on this (fedora15) system and (2) i do not use dbus at all for any purpose and don't want it starting up as it seems to want to do, annoyingly [23:30] probably i'd use it for pulseaudio too, another irritation [23:30] although /etc/pulse/client.conf using autospawn=0 might be all i need for the latter case [23:31] vaxerdec: yeah I agree, /etc/init/init.conf would be a good place for global environment variables [23:32] everyone seems to forget that not everyone is using a huge bloated gnome operating system on their machine [23:32] i mean what about people that just want to run twm or something [23:33] at least you can do it with upstart. systemd just puts its hands in everything. [23:33] pulls in all kinds of requirements [23:34] i mean look at this statement in shutdown(8) on a systemd-infected system: [23:34] NOTES This is a legacy command available for compatibility only. [23:34] i mean "shutdown" is a legacy command? [23:35] shutdown? [23:35] next /bin/sh will be "legacy" [23:35] it's going to be unrecognizable... not unix anymore [23:37] it's nice mr pottering wants to make his own operating system but why does everyone else have to be sucked into whatever willy nilly new paradigm and program requirements that he feels like [23:37] vaxerdec: thats why we like upstart .. it does one thing well.. glad you see that. :) [23:39] yes, i'm trying to run my fedora system with it and it's working but i don't know if i'll be successful with fedora16 [23:40] i may have to jump ship to ubuntu, i dare say [23:40] or gentoo or something [23:40] slangasek has been working on making upstart an alternative in debian [23:41] that's great, i hope they don't go the systemd way [23:42] upstart versus systemd is a big split... [23:42] big pain for upstreams too [23:43] well, upstart does have their own too [23:44] native init scripts that is [23:44] i guess upstreams will be forced to make a sysv, an upstart, and a systemd [23:47] Yeah, it would be good if the two shared a common "start things this way" format. :p [23:47] LSB-init wasn't so bad.. this idea that the shell slows things down too much is kind of crazy IMO.. dash is lightning fast.