Conversation with #inferno at Mon Sep 19 14:32:55 2011 on powerman-asdf@irc.freenode.net (irc) (14:32:55) niven.freenode.net: (notice) [freenode-info] channel flooding and no channel staff around to help? Please check with freenode support: http://freenode.net/faq.shtml#gettinghelp (14:40:26) bvalek2 [50bb91f7@gateway/web/freenode/ip.80.187.145.247] entered the room. (16:45:32) bvalek2 left the room (quit: Quit: Page closed). (17:06:19) yiyus: rog: does your infauth code work with current inferno-os? (17:06:35) yiyus: is it different to what is in inferno-npe? (17:07:49) rog: good question. it should do, as it doesn't use anything funky (17:08:00) rog: what's inferno-npe? (17:08:39) yiyus: noah evans repository: '$p' || reboot exit with errors: $status (17:08:50) yiyus: http://code.google.com/p/inferno-npe/ (17:08:57) yiyus: (wrong paste) (17:10:51) yiyus: i just want to cpu from inferno, i don't know if there has been any improvement since you wrote that code (17:42:25) vsrinivas: yiyus: does work with npe. (17:42:29) vsrinivas: we have it in tree. (17:42:58) vsrinivas: rog: inferno-npe was npe's tree; but lots of folks have added 'interesting' bits. (17:46:15) yiyus: vsrinivas: thanks, i will give inferno-npe a try (17:46:21) vsrinivas: dunno if there's a single list of changes, but off the top of my head -- added a Rendezvous module to the inferno libs; builds on solaris 5.11; mjl-'s execatidle() change to the GC; ipv6 support; in front of the inferno tree allocator there is a per-kproc cache of small allocations; slightly better dragonflybsd support code; my lazy TClunk code (optional, mount -J) (17:46:27) yiyus: do you know why it did not make it into inferno-os? (17:46:47) vsrinivas: yiyus: some of the changes were too experimental; no idea wrt infauth thought. (17:46:49) vsrinivas: *though. (17:47:43) vsrinivas: a tiny bit in -npe was there because I was using it professionally -- added nmealib; wasn't allowed to publish the limbo module for it though. (17:48:06) vsrinivas: oh, someone did a devsrv-equivalent (devserve) (17:48:44) yiyus: well, i will give a try to inferno-npe as a drawterm replacement (17:48:51) vsrinivas: :) go for it! (17:49:22) vsrinivas: i tried to replace the only sem_* code in the linux support stuff with futexes, but didn't find any measurable performance difference. (17:49:28) vsrinivas: so the linux support code is the same. (17:50:56) vsrinivas: oh, yea... added /dev/vmstat; list of counters easy to readout; (17:51:40) vsrinivas: vmstat_entry(char *name, int *var, lock *lk): -- make a counter at @var visible under @name; take lock before readout. (17:51:55) vsrinivas: -npe hasn't been synced with -os in a while, sadly. (17:52:05) vsrinivas: pre-pthread-support-code iirc (17:53:55) yiyus: i want to try 9cpu, and once i see it working i don't mind doing some work to make it work with inferno-os (17:54:10) yiyus: maybe acme-sac would be faster, but i could not find updated linux binaries (18:00:51) rog: yiyus: it should just work (18:00:59) rog: yiyus: have you tried following the instructions? (18:01:44) yiyus: not yet, i was not sure if starting with inferno-os+infauth, inferno-npe, or acme-sac, that's why i asked (18:01:50) yiyus: i will try in a few hours (18:11:58) vsrinivas: is the spki/sdsi work still happening? (18:14:14) vsrinivas: yiyus: just tried it as drawterm; works well. (18:14:43) vsrinivas: out-of the box inferno-npe : run auth/factotum, auth/feedkey ; make sure #^ is mounted at /chan; (18:14:48) vsrinivas: 9cpu -r -h (18:14:49) vsrinivas: tada! (19:11:16) powerman-asdf: vsrinivas: what does mjl- change to execatidle()/GC? any chance it's related to http://code.google.com/p/inferno-os/issues/detail?id=274 ? (19:21:32) bvalek2 [50bb9f0b@gateway/web/freenode/ip.80.187.159.11] entered the room. (19:45:12) vsrinivas: powerman-asdf: yea. it was (19:45:30) vsrinivas: was meant to solve that bug (19:50:56) powerman-asdf: cool. no way to get that chance without doing manual diff -os and -npe, I suppose? (19:57:34) bvalek2 left the room (quit: Ping timeout: 252 seconds). (19:58:33) vsrinivas: both are in hg, so you can just cherry pick the patch (20:07:51) powerman-asdf: probably I don't know hg good enough for this :( (20:27:16) rog is now known as rog-away (21:08:09) Fish [~Fish@9fans.fr] entered the room. (00:16:37) Fish left the room (quit: Quit: WeeChat 0.3.5). (00:45:23) rog-away left the room (quit: Ping timeout: 260 seconds). (08:48:23) Fish [~Fish@9fans.fr] entered the room. (09:11:00) Fish left the room (quit: Quit: WeeChat 0.3.5). (09:31:48) rog-away [~rog@host-92-30-159-91.as13285.net] entered the room. (09:48:26) acmeuser [~acmeuser@118.172.69.48.adsl.dynamic.totbb.net] entered the room. (09:48:26) acmeuser left the room. (10:15:05) bvalek2 [50bb9fec@gateway/web/freenode/ip.80.187.159.236] entered the room. (10:37:12) bvalek2 left the room (quit: Ping timeout: 252 seconds). (10:40:02) bvalek2 [50bb9fec@gateway/web/freenode/ip.80.187.159.236] entered the room. (11:08:44) bvalek2 left the room (quit: Ping timeout: 252 seconds). (12:16:51) rog-away left the room (quit: Quit: Konversation terminated!). (12:17:00) rog-away [~rog@host-92-30-159-91.as13285.net] entered the room. (12:30:05) rog-away is now known as wrtp (12:35:42) bvalek2 [50bb9fec@gateway/web/freenode/ip.80.187.159.236] entered the room. (14:50:47) wrtp_test_away [~wrtp_test@ec2-50-17-137-101.compute-1.amazonaws.com] entered the room. (14:51:10) wrtp_test_away left the room (quit: Client Quit). (14:57:47) wrtp_test_away [~wrtp_test@ec2-107-20-122-136.compute-1.amazonaws.com] entered the room. (15:11:02) wrtp_test_away left the room (quit: Quit: Coyote finally caught me). (15:34:32) wrtp_test_away [~wrtp_test@ec2-107-20-118-153.compute-1.amazonaws.com] entered the room. (15:49:51) wrtp_tes` [~wrtp_test@host-92-30-159-91.as13285.net] entered the room. (15:51:18) wrtp_tes` left the room (quit: Client Quit). (15:58:34) powerman-asdf: someone asked me about recommended books about software architecture design, and only really good book I was able to recall was Uncle Bob's (and even that I was reading before I go deep in Inferno world, so I'm not sure about my current attitude to it) (15:59:45) powerman-asdf: I mostly dislike Fowler; GoF doesn't do much more than just name patters. (15:59:56) powerman-asdf: so, can you recommend any other books?