**** BEGIN LOGGING AT Sun Mar 01 02:59:59 2015 Mar 01 11:47:45 Hi, I want to install ipsec-tools 0.8.2 from meta-oe (it already has recipe). However, it does not install any .conf files related to racoon or setkey. Can someone help? Mar 01 12:35:38 anyone active during this nice weekend? Mar 01 12:35:44 hi Mar 01 12:37:09 define active ;) Mar 01 12:37:37 ah, great. Mar 01 12:37:43 so anyone using sysvinit in here? Mar 01 12:37:51 I am struggling with implementing my ntpd use case for busybox Mar 01 12:38:00 I would like to use this start-stop-daemon facility Mar 01 12:38:16 what I would like to achieve is to run the daemon in the background to sync up from a remote peer; all cool, done Mar 01 12:38:28 but now I have another board and I would like to sync that up from the one which goes to the internet Mar 01 12:38:49 therefore, the "main" board would have a client daemon as well as an ntpd server running Mar 01 12:39:05 it is trivial to run these, but I do not quite get how to handle the start, stop and restart for the instances separately. Mar 01 12:39:13 since the busybox binary run would be the same Mar 01 12:39:30 so when I say "start-stop-daemon -K -n ntpd" -> it would kill both instances. Mar 01 12:39:32 you don't need a client and a server Mar 01 12:39:38 ntpd is both Mar 01 12:39:44 of course I could parse the ps output on my own, but I am wondering if there is a friendlier way. Mar 01 12:39:51 abelloni: nope Mar 01 12:39:52 however, I'm not sure that is the case for the busybox applet Mar 01 12:39:57 you start the server with a specific option Mar 01 12:40:05 it is not. Mar 01 12:40:29 abelloni: -l Run as server on port 123 Mar 01 12:41:57 so basically I am lost at how to get their own life cycle. Mar 01 12:42:04 with regards to start, stop and restart Mar 01 12:42:11 I could probably wrap them into a script or even binary. Mar 01 12:42:17 that is another hackish way. Mar 01 12:42:45 but ultimately, I think start-stop-daemon ought to provide an identifier Mar 01 12:42:52 I thought -n was for that as you specify -a for the daemon already. Mar 01 12:42:55 but apparently not! Mar 01 12:43:24 such an option could help users using things like ntpd from busybox where the same executable is used for both client and server purposes, you see? Mar 01 12:44:44 I could also launch and get pid, store it, etc, but it is kind of tedious, all. Mar 01 12:45:08 from what I see in the code, using -l it will be both client and server Mar 01 12:45:53 oh, really? Mar 01 12:46:11 that would be good if it was just documentation failure. Mar 01 12:46:41 I wouldn't agree that this is documentation failure :) Mar 01 12:46:53 why? Mar 01 12:46:58 see, I got confused Mar 01 12:47:02 and wasted a lot of time Mar 01 12:47:18 if it clearly writes that it is also a client when you run it as server, then I could have spared a lot of time. Mar 01 12:47:27 but I am still unsure that is the case, to be honest. Mar 01 12:50:34 I don't see why it wouldn't be the case Mar 01 12:51:16 it just listens on the port Mar 01 12:57:18 because being a server is radically different from being a client Mar 01 12:57:46 there may be exceptions, but that usually holds true 99%. Mar 01 12:57:51 at least in my experience anyhow. Mar 01 12:57:59 but we would need to know what are the exceptions. Mar 01 12:58:15 it was not clear to me after all! Mar 01 13:00:55 hmm, I cannot quite get the code just yet. Mar 01 13:01:13 like I said, that's what on would expect from an ntp daemon anyway Mar 01 13:01:16 one Mar 01 13:01:26 like I said, that is not what I expected. Mar 01 13:01:31 feel free to sent a documentation patch Mar 01 13:01:41 already requested the change ;-) Mar 01 13:01:57 well, I just asked. Mar 01 13:02:02 because I cannot get my head around to the code. Mar 01 13:02:06 it is a bit complex... Mar 01 13:04:07 http://git.busybox.net/busybox/tree/networking/ntpd.c#n2180 Mar 01 13:04:14 that sets up the fd Mar 01 13:04:28 then http://git.busybox.net/busybox/tree/networking/ntpd.c#n2355 Mar 01 13:04:36 processes incoming packets Mar 01 13:05:24 recv_and_process_client_pkt(/*G.listen_fd*/); :O Mar 01 13:06:48 that commented file descriptor is "funky" Mar 01 13:07:36 that is the ugliness added by busybox Mar 01 13:07:46 http://git.busybox.net/busybox/tree/networking/ntpd.c#n1976 Mar 01 13:08:09 adding globals everywhere Mar 01 13:09:57 I think busybox's codebase is one of the worst ones I have ever seen Mar 01 13:10:04 not because of this, but for goto abuses, etc. Mar 01 13:12:52 ok, thanks. Mar 01 13:32:58 hmm, I think I will try to improve my one year old patch on the mailing list to finally to get in :) Mar 01 13:34:12 hopefully with more luck this time ;-) Mar 01 13:44:39 JaMa: are you around Mar 01 13:45:24 I am looking for a good example of LSB header to finally get my busybox ntpd patch in :) Mar 01 13:46:37 because even now, pretty much no busybox daemon has such a header Mar 01 13:46:44 and that was the reason for my feature to be blocked ;-) Mar 01 13:46:57 but since no one accepted it like this, just me, I will update it, but I need a prime example... Mar 01 13:48:48 lpapp: e.g. a5f631eefdfaeca22d8d06b63d1e27fb934a8bc5 in meta-oe adds LSB header Mar 01 13:49:31 search "LSB header" in git log to easily find many more examples Mar 01 13:52:20 JaMa: ah, ok, I was using git log -S lsb Mar 01 13:52:26 but that gave me lots of irrelevant results, thanks. Mar 01 14:03:17 JaMa: git show a5f631eefdfaeca22d8d06b63d1e27fb934a8bc5 Mar 01 14:03:17 fatal: bad object a5f631eefdfaeca22d8d06b63d1e27fb934a8bc5 Mar 01 14:03:23 which repository are you referring to? I am in poky. Mar 01 14:05:29 JaMa: bad object in openembedded-core, too. Mar 01 14:05:48 ah, sorry, you wrote meta-oe .... how silly I am! Mar 01 14:06:03 apologies :) Mar 01 14:19:53 hey, Jama! ./meta-oe/recipes-connectivity/gateone/gateone/gateone-init:4:# Required-Start: networking -> isn't the hash missing before networking? Mar 02 02:49:00 afaik, yocto has two cross compiler versions..how to select one over another? **** ENDING LOGGING AT Mon Mar 02 02:59:58 2015