00:25:20 -!- kanru [~kanru@kanru-1-pt.tunnel.tserv15.lax1.ipv6.he.net] has quit [Ping timeout: 260 seconds] 00:26:07 kanru [~kanru@kanru-1-pt.tunnel.tserv15.lax1.ipv6.he.net] has joined #sbcl 01:13:37 -!- kanru [~kanru@kanru-1-pt.tunnel.tserv15.lax1.ipv6.he.net] has quit [Quit: ZNC - http://znc.sourceforge.net] 01:16:37 kanru [~kanru@kanru-1-pt.tunnel.tserv15.lax1.ipv6.he.net] has joined #sbcl 02:48:42 Fare [fare@nat/google/x-hrccwkumgncxqcia] has joined #sbcl 04:15:55 -!- Fare [fare@nat/google/x-hrccwkumgncxqcia] has quit [Read error: Operation timed out] 05:07:38 -!- asedeno_work [asedeno@nat/google/x-ldgpjdvcstmxpxmf] has quit [Ping timeout: 246 seconds] 05:20:09 asedeno_work [asedeno@nat/google/x-zornhrjqnagcyzel] has joined #sbcl 06:00:05 -!- wbooze [~wbooze@xdsl-78-35-156-46.netcologne.de] has quit [Ping timeout: 256 seconds] 06:07:13 sdemarre [~serge@91.176.154.9] has joined #sbcl 06:55:31 angavrilov [~angavrilo@217.71.227.190] has joined #sbcl 07:14:58 -!- sdemarre [~serge@91.176.154.9] has quit [Ping timeout: 246 seconds] 07:27:41 edgar-rft [~GOD@HSI-KBW-078-043-123-191.hsi4.kabel-badenwuerttemberg.de] has joined #sbcl 08:05:19 attila_lendvai [~attila_le@37.99.86.139] has joined #sbcl 08:05:19 -!- attila_lendvai [~attila_le@37.99.86.139] has quit [Changing host] 08:05:19 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 08:17:57 nikodemus [~nikodemus@cs27127210.pp.htv.fi] has joined #sbcl 08:17:58 -!- ChanServ has set mode +o nikodemus 11:25:18 danlentz [~danlentz@c-76-117-5-156.hsd1.nj.comcast.net] has joined #sbcl 11:25:54 nikodemus: so, fancy robotics for you these days? sounds exciting. 11:26:25 (What programming languages are you guys using for which tasks?) 11:44:19 lichtblau: octave and r for data analysis. C++ for some sensor stuff. clojure for most of the system, but with plenty of stuff written directly in java or in C++ also being called -- because clojure is way too slow sometimes 11:45:11 different people use different things for ad-hoc tasks 11:45:32 i grab CL the first thing most of the time, some go for python, so for whatever 11:45:39 s/so/some 11:46:29 (i've been mostly writing C++ for the past few weeks, actually) 11:48:39 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 11:51:50 -!- nikodemus [~nikodemus@cs27127210.pp.htv.fi] has quit [Quit: This computer has gone to sleep] 11:57:04 aha, sounds to me like you're in a prototyping phase where you're often trying out lots of new approaches (?) 12:03:39 cmm [~cmm@bzq-79-183-222-211.red.bezeqint.net] has joined #sbcl 12:05:35 -!- cmm- [~cmm@bzq-79-183-213-148.red.bezeqint.net] has quit [Ping timeout: 256 seconds] 12:18:40 -!- |3b| [foobar@cpe-72-179-19-4.austin.res.rr.com] has quit [Remote host closed the connection] 12:19:22 |3b| [foobar@cpe-72-179-19-4.austin.res.rr.com] has joined #sbcl 13:05:50 nikodemus` [~nikodemus@cs27127210.pp.htv.fi] has joined #sbcl 13:05:50 -!- ChanServ has set mode +o nikodemus` 13:11:10 gko [~user@114-34-168-13.HINET-IP.hinet.net] has joined #sbcl 13:13:54 ok, aside from one final walker bogosity (which isn't exactly new, just changed it's shape) i think i've amended my sins now 13:14:06 its, even 13:15:52 lichtblau: well, yes and no. two first sales are already public, so we're productizing really, but there's also a lot of prototyping going on at the same time... 13:17:21 (http://www.zenrobotics.com/press-releases/maes-containers/ and http://www.zenrobotics.com/press-releases/zenrobotics-pr_2012-09-13/) 13:21:00 nice. So just how cold is the climate in Helsinki anyway? :-) 13:42:58 summers are short but not too snowy :) 13:44:49 (actually you can think of weather in helsinki as a continental version of london -- summers are warmer and winters are colder) 13:52:17 So let me elaborate here on #sbcl on my twitter replies to your blog... My thinking is that a truly lightweight, efficient threading/processing concept (think erlang) implies userland threads. 13:52:25 But the reverse isn't true: Just because you have userland threads doesn't mean they're fast, because the whole language/implementation/ecosystem needs to be aligned with that design goal. 13:52:33 And I don't see that happen in CL or SBCL; those userland threads will be as heavyweight as the native ones. So what's the specific benefit? 13:53:44 Doesn't mean I'm opposed to userland threads, and maybe I'll come around to (what I percieve as) Anton's viewpoint that they're useful for one implemention detail, and that that's sufficient reason to do them. 13:53:59 But which users really want them for what? 13:57:34 lichtblau: i think wanting to use them for blocking IO stuff is reasonable -- but i'm not sufficiently well versed in the kind of servers that would do that to say if it's a really good idea 13:58:29 at the same time, if there is zero differentiation between userland and native threads in lisp land, then for debugging purposes i could see the point in being able to suspend threads, etc 13:58:52 but i'm not convinced myself that either is a particularly good idea 14:00:04 the blog was a response to a wish: "tell me what userland threads in sbcl would entail in your mind" 14:01:12 but at the same time, we /could/ make lighterweight threads as well -- it's not like every thread needs a huge stack 14:01:29 not maybe featherweight threads, but... much smaller ones 14:04:14 by the way, did you see this? https://github.com/copyleft/cl-html5-parser 14:06:35 leuler [~user@p548FCD9B.dip.t-dialin.net] has joined #sbcl 14:26:19 -!- edgar-rft [~GOD@HSI-KBW-078-043-123-191.hsi4.kabel-badenwuerttemberg.de] has quit [Quit: sudden death] 14:27:55 -!- danlentz [~danlentz@c-76-117-5-156.hsd1.nj.comcast.net] has quit [Quit: Colloquy for iPhone - http://colloquy.mobi] 14:28:16 danlentz [~danlentz@c-76-117-5-156.hsd1.nj.comcast.net] has joined #sbcl 14:29:49 -!- nikodemus` [~nikodemus@cs27127210.pp.htv.fi] has quit [Quit: This computer has gone to sleep] 14:38:02 -!- danlentz [~danlentz@c-76-117-5-156.hsd1.nj.comcast.net] has quit [Quit: Colloquy for iPhone - http://colloquy.mobi] 15:01:32 slyrus_ [~chatzilla@adsl-99-183-243-129.dsl.pltn13.sbcglobal.net] has joined #sbcl 15:33:36 asciilifeform [~asciilife@pool-71-191-43-115.washdc.fios.verizon.net] has joined #sbcl 15:44:49 nikodemus [~nikodemus@cs27127210.pp.htv.fi] has joined #sbcl 15:44:50 -!- ChanServ has set mode +o nikodemus 15:46:18 -!- gko [~user@114-34-168-13.HINET-IP.hinet.net] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 16:03:45 LiamH [~healy@pool-108-7-208-203.bstnma.fios.verizon.net] has joined #sbcl 16:23:52 nikodemus: "POSIX.1-2008 removes the specifications of makecontext() and swapcontext(), citing portability issues, and recommending that applications be rewritten to use POSIX threads instead." 16:24:43 quoth the manpage 16:46:39 wbooze [~wbooze@xdsl-87-79-199-28.netcologne.de] has joined #sbcl 17:03:32 haha 17:07:57 -!- LiamH [~healy@pool-108-7-208-203.bstnma.fios.verizon.net] has quit [Quit: Leaving.] 17:21:42 LiamH [~healy@pool-108-7-208-203.bstnma.fios.verizon.net] has joined #sbcl 18:30:06 -!- LiamH [~healy@pool-108-7-208-203.bstnma.fios.verizon.net] has quit [Quit: Leaving.] 18:45:18 LiamH [~healy@pool-108-7-208-203.bstnma.fios.verizon.net] has joined #sbcl 18:55:47 sdemarre [~serge@91.176.154.9] has joined #sbcl 19:44:34 -!- nikodemus [~nikodemus@cs27127210.pp.htv.fi] has quit [Quit: Leaving] 19:46:20 -!- LiamH [~healy@pool-108-7-208-203.bstnma.fios.verizon.net] has quit [Quit: Leaving.] 19:53:57 I'm reminded of gnu pth, which now uses pthreads + big mutex everywhere. 19:55:17 -!- leuler [~user@p548FCD9B.dip.t-dialin.net] has quit [*.net *.split] 19:55:18 -!- luis [~luis@nhop.r42.eu] has quit [*.net *.split] 19:55:18 -!- fe[nl]ix [~quassel@pdpc/supporter/professional/fenlix] has quit [*.net *.split] 19:55:18 -!- daimrod [~daimrod@sbrk.org] has quit [*.net *.split] 19:55:19 -!- cmm [~cmm@bzq-79-183-222-211.red.bezeqint.net] has quit [*.net *.split] 19:55:19 -!- flip214 [~marek@unaffiliated/flip214] has quit [*.net *.split] 19:55:19 -!- minion [~minion@tiger.common-lisp.net] has quit [*.net *.split] 19:56:01 fe[nl]ix [~quassel@pdpc/supporter/professional/fenlix] has joined #sbcl 19:56:01 daimrod [~daimrod@sbrk.org] has joined #sbcl 19:56:15 luis` [~luis@nhop.r42.eu] has joined #sbcl 19:56:16 cmm [~cmm@bzq-79-183-222-211.red.bezeqint.net] has joined #sbcl 19:56:16 flip214 [~marek@unaffiliated/flip214] has joined #sbcl 19:56:16 minion [~minion@tiger.common-lisp.net] has joined #sbcl 20:00:39 -!- fe[nl]ix [~quassel@pdpc/supporter/professional/fenlix] has quit [*.net *.split] 20:00:40 -!- daimrod [~daimrod@sbrk.org] has quit [*.net *.split] 20:01:04 fe[nl]ix [~quassel@pdpc/supporter/professional/fenlix] has joined #sbcl 20:01:04 daimrod [~daimrod@sbrk.org] has joined #sbcl 20:19:14 -!- sdemarre [~serge@91.176.154.9] has quit [Ping timeout: 246 seconds] 20:21:47 -!- angavrilov [~angavrilo@217.71.227.190] has quit [Ping timeout: 244 seconds] 20:44:56 edgar-rft [~GOD@HSI-KBW-078-043-123-191.hsi4.kabel-badenwuerttemberg.de] has joined #sbcl 21:46:32 -!- Posterdati [~quassel@host45-237-dynamic.6-87-r.retail.telecomitalia.it] has quit [Read error: Connection reset by peer] 21:49:23 Posterdati [~quassel@host45-237-dynamic.6-87-r.retail.telecomitalia.it] has joined #sbcl 22:23:04 -!- edgar-rft [~GOD@HSI-KBW-078-043-123-191.hsi4.kabel-badenwuerttemberg.de] has quit [Quit: this computer sucks] 22:35:43 -!- wbooze [~wbooze@xdsl-87-79-199-28.netcologne.de] has quit [Ping timeout: 246 seconds] 23:04:34 lichtblau: any insight into http://discontinuity.info/~pkhuong/gc-deadlock-gdb.txt or http://discontinuity.info/~pkhuong/gc-deadlock-backtrace.txt? 23:05:05 wbooze [~wbooze@xdsl-84-44-209-251.netcologne.de] has joined #sbcl 23:05:16 AFAICT, we have thread 3 and 2 in a deadlock 23:05:31 3 is in w/o gcing, and waits for a ptmalloc lock 23:05:37 while 2 is stopped in the middle of calloc 23:10:27 uhm. On Linux? Is that a new or old problem? 23:10:49 linux 23:11:17 I don't know if it's new. I don't think many people exercise concurrent run-program the way I currently am. 23:14:11 and which one is 0x7ffff1c80060 here? 23:17:46 isn't the thread struct allocated at the same time as the stack? 23:18:00 seems to be thread 3 23:33:15 this seems to be one of those things which look like obvious "yes, that would hang" bugs. But at 1:30 AM it's hard to say without adding "but ask me again in the morning". Random observations: 23:34:08 - if we didn't mess around with signals on top of foreign code stack frames, this particular problem shouldn't happen (yay safepoints!) 23:34:33 all right, thanks. 23:35:30 - but more fundamentally, even safepoint builds would have a problem, namely if Lisp code (not C code) used locks in without-gcing liberally. So basically the rule should be "then don't do that" 23:35:51 can we spawn w/ gcing? 23:36:44 - of course, on glibc you could add various __malloc_hook things that would wrap ordinary malloc with a C function that blocks gc_sigset around the malloc call. But if you have the same problem on other C libraries, that sort of workaround is going to be annoying. 23:49:15 Let's assume for the moment that everyone hates safepoints and their performance and we can't enable them by default on posix because of that (not my hope, but let's think it through). 23:49:34 In that case, we could/should still do the following: Take the CSP safepoints (i.e., not the global safepoints in Lisp code, just the per-thread bits in call_into_c) and use those even on non-safepoint builds. 23:51:15 well; afk. 23:52:00 all right. We'll see what the others think. 23:52:46 maybe there's something subtle going on, but isn't that without-gcing just a legacy cmucl kludge for pinning the arguments? 23:53:26 maybe we have to pin a variable number of values(?) 23:53:44 LiamH [~healy@pool-108-7-208-203.bstnma.fios.verizon.net] has joined #sbcl 23:57:04 don't think so, the argv / env lists are marshalled manually