=== st3 [n=st3@81-208-31-210.ip.fastwebnet.it] has joined #upstart | ||
=== j_ack [n=rudi@p508da10d.dip0.t-ipconnect.de] has joined #upstart | ||
=== Amaranth [n=travis@ubuntu/member/Amaranth] has joined #upstart | ||
=== juergbi [n=juerg@80-219-17-127.dclient.hispeed.ch] has joined #upstart | ||
=== Keybuk [n=scott@quest.netsplit.com] has joined #upstart | ||
=== Amaranth [n=travis@ubuntu/member/Amaranth] has joined #upstart | ||
AlexExtreme | hmm, in ubuntu, does usplash get started by the initramfs or is it started by the init scripts? | 09:58 |
---|---|---|
_ion | It's started ASAP from initramfs. | 09:59 |
AlexExtreme | thought so | 09:59 |
AlexExtreme | btw _ion, did you get anywhere with your event based mounting ideas? | 10:03 |
_ion | I haven't yet got around to implementing any of that. | 10:04 |
AlexExtreme | k | 10:04 |
=== Md [i=md@freenode/staff/md] has joined #upstart | ||
=== Amaranth [n=travis@ubuntu/member/Amaranth] has joined #upstart | ||
=== wasabi [n=wasabi@ubuntu/member/wasabi] has joined #upstart | ||
=== tale [n=tale@207.235.54.1] has joined #upstart | ||
=== j_ack [n=rudi@p508D8586.dip0.t-ipconnect.de] has joined #upstart | ||
=== Starting logfile irclogs/upstart.log | ||
=== Starting logfile irclogs/upstart.log | ||
=== ubuntulog [i=ubuntulo@ubuntu/bot/ubuntulog] has joined #upstart | ||
=== Topic for #upstart: Upstart 0.3.8 | http://upstart.ubuntu.com/ | http://upstart.ubuntu.com/wiki/UpstartOnGentoo | ||
=== Topic (#upstart): set by Keybuk at Fri Mar 16 18:29:42 2007 | ||
=== juergbi [n=juerg@80-219-17-127.dclient.hispeed.ch] has joined #upstart | ||
=== mbiebl [n=michael@e180065090.adsl.alicedsl.de] has joined #upstart | ||
=== j_ack [n=rudi@p508d9de1.dip0.t-ipconnect.de] has joined #upstart | ||
=== j_ack_ [n=rudi@p508DBD73.dip0.t-ipconnect.de] has joined #upstart | ||
mbiebl | Keybuk: hi | 09:04 |
Keybuk | hey | 09:44 |
mbiebl | I noticed when I symlink an upstart job file into /etc/event.d, upstart does not recognize when the file is changed. | 09:55 |
Keybuk | right | 09:55 |
mbiebl | Is that a limitation of inotify? | 09:55 |
Keybuk | or a feature | 09:55 |
Keybuk | more a filesystem limitation | 09:56 |
Keybuk | since when changing a file, one would have to search the entire disk for symlinks to it, in order to update their directory entries | 09:56 |
mbiebl | why that? | 09:57 |
Keybuk | because the alternative would be every time you make a symlink, storing a back link somewhere in the linked-to file's inode | 09:57 |
Keybuk | and that simply doesn't work with relative symlinks ;) | 09:58 |
Keybuk | since they can be on filesystems that can be mounted anywhere | 09:58 |
Keybuk | thus refer to different paths | 09:58 |
Keybuk | (not to mention that it wouldn't work with cross-filesystem symlinks) | 09:58 |
mbiebl | Hm, maybe I don't get it. If /etc/event.d/foo is a symlink to /lib/upstart/foo, why would you have to store a back link? | 09:59 |
Md | the solution would be to use inotify on each symlink in the directory. may or may not be worth the effort (see the debian-devel thread about udev and the same problem) | 09:59 |
Keybuk | because upstart watches /etc/event.d for changes | 10:01 |
Keybuk | and changing /lib/upstart/foo doesn't change that directory | 10:01 |
Keybuk | it changes the /lib/upstart directory | 10:01 |
Keybuk | and there's nothing there telling the filesystem that there's a symlink in the /etc/event.d directory and you might want to touch that too | 10:01 |
Keybuk | personally I think it's a feature, since it means you can have a set of files that are explicitly *not* watched | 10:01 |
mbiebl | I guess, you're right | 10:02 |
=== Md_ [i=md@freenode/staff/md] has joined #upstart | ||
=== Md_ is now known as Md | ||
=== Keybuk [n=scott@82.108.80.247] has joined #upstart |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!