[06:00] <maro_> Keybuk: regarding the script syntax, have you considered using a make-like syntax instead?
[06:00] <thom> oh please no
[06:03] <Keybuk> maro_: I think I share thom's response here ;)
[06:03] <Keybuk> DEPENDENCIES := foo
[06:03] <Keybuk> DEPENDENCIES += bar
[06:03] <Keybuk> start::
[06:03] <Keybuk>    some script
[06:03] <thom> i couldn't imagine anything more hideous, personally. but you could have some fun
[06:03] <Keybuk> .EXEC: /sbin/some --daemon
[06:04] <Keybuk> actually, that'd be illegal, you'd need \t before "some script" :)
[06:04] <Keybuk> thom: X. M. L.
[06:05] <thom> vpath %.event \ /etc/event.d/ ; %.event: ...
[06:05] <thom> Keybuk: true
[06:06] <maro_> Keybuk: no, that wasn't what I meant :)
[06:06] <maro_> it was only for the "start script"/"end script" stanza
[06:06] <Keybuk> maro_: elaborate
[06:07] <maro_> I was initially confused with the "start script" wording since I thought it meant "start of script" (versus "before running the command")
[06:08] <maro_> Keybuk: http://borkware.net/~mark/upstart-service.txt
[06:12] <maro_> (refresh)
[06:12] <Keybuk> the trouble there is the requirement for indentation and physical tabs
[06:14] <maro_> python fans will love it, though ;)
[06:16] <maro_> my guess is that most distros/users will be consistent in their indentation anyway, and enforcing it will make 3rd party supplied event scripts more consistent
[06:17] <maro_> anyway, that's just my POV, you're the code monkey :)
[06:19] <maro_> btw, the 'while' thing is mainly to avoid "start on this-up", "stop on this-down"
[06:20] <_ion> I'm not very fond of the pythonish way of indentation.
[06:21] <thom> can we please not have a pointless indentation debate
[06:21] <maro_> yeah :)
[06:21] <thom> the problem with make like indentation is that some poor user will always cock it up and then get really, _really_ confused
[06:24] <maro_> my main liking about it is that it's more human/language-friendly to use a colon, and it saves a line for each script - screw the indentation :)
[06:25] <maro_> (the only argument I can see for that is that the people who edit event scripts are likely to already be familiar with make syntax)
[06:27] <thom> maro_: i don't think that's particularly likely; think about the poor sysadmin who needs to knock up an event script quickly
[06:30] <maro_> what about the poor sysadmin that though "start script" was the syntax for beginning any script block since it's stopped with "end script"? (e.g. me ;))
[06:31] <Keybuk> there's a more interesting problem that "start script" implies the script should start the daemon
[06:31] <maro_> anyway, screw the make syntax part, what do you think about "while"/"when"?
[06:31] <Keybuk> it's planned to be renamed to pre-start and post-start anyway, which should make things clearer
[06:32] <maro_> Keybuk: true :)
[06:32] <Keybuk> while/when ?
[06:32] <maro_> Keybuk: the first non-commented line in the file
[06:35] <maro_> also allowing "functions" to be either script blocks or binaries ("run" in that file)
[06:35] <Keybuk> "while X is Y" doesn't make sense ... events are one-shot occurrences
[06:36] <maro_> Keybuk: hal would run "while dbus is running"
[06:37] <maro_> i.e. respawn if it dies, and stop when dbus is no longer running
[06:37] <Keybuk> what is "dbus" in this context?
[06:37] <maro_> system bus is available :)
[06:38] <Keybuk> but what is that?
[06:38] <Keybuk> it's not an event
[06:38] <maro_> the start/stop of it is, but yeah, point taken
[06:39] <Keybuk> start on job-started dbus
[06:39] <Keybuk> stop on job-stopped dbus
[06:39] <Keybuk> start on job-failed self
[06:39] <Keybuk> accomplishes much the same thing
[06:40] <Keybuk> though not quite, interestingly
[06:40] <maro_> and which one is nicer to read/understand? :)
[06:40] <Keybuk> maro_: I'd say yours is easier to read, but mine is easier to understand
[06:41] <maro_> I really like the "start on job-failed self" thing instead of exec/respawn :)
[06:42] <x4NothinKx> hi
[06:42] <x4NothinKx> is mahangu there?
[06:43] <maro_> how does "when storage is added" look like?
[06:43] <maro_> "on storage-added"?
[06:43] <Keybuk> define "added"
[06:43] <Keybuk> you have block-device-added, or path-mounted, etc.
[06:44] <Keybuk> depending what you name those events
[06:44] <maro_> that's set by the storage event (and could be anything)
[06:44] <maro_> are all those files in /etc/event.d? :/
[06:46] <_ion> Huh?
[06:46] <Keybuk> no, those are events
[06:48] <maro_> ah, right, "arbitrary strings" - I need coffee :)
[06:48] <x4NothinKx> could someone tell me when mahangu is regular in the channel?
[06:48] <Keybuk> also see https://wiki.ubuntu.com/UpstartDesignChanges
[06:48] <Keybuk> that details where we're thinking of going with them
[06:49] <x4NothinKx> i want to do the documentation and german translation for upstart
[06:50] <maro_> looking at it, thanks :)
[06:51] <Keybuk> x4NothinKx: https://launchpad.net/products/upstart/main/+pots/upstart/de/+translate
[07:03] <maro_> Keybuk: btw, I noticed that initctl list doesn't remove moved files
[07:03] <maro_> (in current upstart, obviously)
[07:04] <maro_> s/moved/re&/
[07:05] <Keybuk> right
[07:05] <Keybuk> known bug
[07:05] <maro_> ok, as long as it's considered a bug :)
[07:05] <Keybuk> or moved files, actually ;)
[07:06] <maro_> or both :)
[07:06] <Keybuk> yup, it's actually just a lack of implementation
[07:06] <Keybuk> e.g. should a job be stopped if you remove its file? :)
[07:06] <maro_> hmm, good question
[07:07] <maro_> I'll avoid thinking about that, might get a headache
[07:07] <_ion> IMO definitely yes (unless it's only renamed)
[07:07] <maro_> _ion: in which case... should it be restarted?
[07:08] <_ion> I don't see a reason to restart it.
[07:08] <maro_> nah, true, initctl should just list the new name
[07:08] <_ion> Just change the job's name in memory.
[07:08] <Keybuk> they're not restarted when modified
[07:09] <Keybuk> so in theory they probably shouldn't be when renamed
[07:09] <_ion> Yep.
[07:09] <maro_> agree completely
[07:10] <maro_> I'd hate to upgrade initscripts if they were ;)
[07:12] <maro_> Keybuk: I also get a ton of hal noise on my console even when using "console null" - I guess hal's helpers are to blame?
[07:14] <Keybuk> yes
[07:19] <maro_> heh, hal just had pid 666, wonder why
[07:23] <_ion> Because it's an integer between 2 and kernel.pid_max
[07:23] <Artanicus> heh
[07:37] <x4NothinKx> i do the translation now
[07:37] <x4NothinKx> but i think rosetta is not good
[07:38] <x4NothinKx> its very difficult to find matching vocabulary
[07:38] <x4NothinKx> because the most words are without a context
[07:38] <x4NothinKx> thats a puzzle
[07:39] <Keybuk> I'm afraid I know little about translation, or the difficulty involved
[07:44] <x4NothinKx> some words should be kept in german, like "killed" or so. But some should'nt. I need to test the translation intensively.