[10:12] <ion_> Pretty. And it only takes a kilobyte. http://pouet.net/prod.php?which=32194
[10:15] <Keybuk> what is it?
[10:16] <ion_> An intro. There’s a link to a video capture on the page.
[10:19] <ion_> http://heh.fi/heroes/heroes-scrape/pipeline.png
[10:20] <Keybuk> heh
[12:03]  * Keybuk finally discovers the fuck-up
[12:03] <Keybuk> parse_job () will free the old job with the same name
[12:03] <Keybuk> yet so will conf_item_destroy ()
[12:03] <Keybuk> the difference being that parse_job () frees the job without freeing the ConfItem for it
[12:03] <Keybuk> which means that job_change_state () must do the same!
[12:04]  * Jc2k is somewhat baffled but nods and smiles
[12:04] <Keybuk> this bug must have always been present, the newer code happens to tickle it more
[12:04]  * Keybuk tries to think how you'd replicate it
[12:07] <Keybuk> job of the same name in two different directories
[12:07] <Keybuk> then reload
[12:07] <Keybuk> that'd cause older code to core dump
[13:38]  * Keybuk frowns
[13:39]  * Jc2k bounces
[14:12]  * Keybuk almost hacks it
[14:26] <Keybuk> hurrah
[14:27] <Jc2k> :O
[14:27] <Jc2k> victoreh?
[14:30] <Keybuk> not quiet
[14:31] <Jc2k> :(
[14:31] <Keybuk> I'm tempted to ignore the problem and move on
[14:31] <Keybuk> since it's in code I intend to completely rewrite
[14:31] <Jc2k> sounds like a plan :O
[15:12] <Keybuk> odd, I made the bug go away
[15:12] <Keybuk> la la la
[15:38] <Keybuk>  woo, only test_job left to fix
[16:41] <Keybuk> all 5,300 lines of it :-/
[16:42] <Jc2k> woah
[16:42] <Jc2k> large test then?
[16:42] <Keybuk> yeah
[18:52] <kelsin> Hey, is there a way to specify a start script/ stop script and then set the PID directly for jobs? Basically have a service that comes with a normal "script start/stop/status" as the way to start it, didn't want to have to work a lot with it if possible
[18:53] <kelsin> actually might have answered my own question by finding the wait command :) I'll ask again if this doesn't work