00:08:55 wbooze [~wbooze@xdsl-78-35-131-161.netcologne.de] has joined #sbcl 00:29:26 Quadrescence: ping 00:29:33 pong 00:29:46 PkhuONG 00:30:19 can you try applying pvk.ca/misc/aligned_thread_struct.diff on top of HEAD? 00:30:43 sure 00:37:23 pkhuong, http://paste.lisp.org/+2SW7 00:37:38 well, fuck. 00:37:46 we'll have to find some way to make the TLS trick work. 00:38:54 -!- wbooze [~wbooze@xdsl-78-35-131-161.netcologne.de] has quit [Ping timeout: 252 seconds] 00:40:23 pkhuong, is there at least an understanding on why it's erroring? 00:40:54 nope. 00:40:58 dang 00:41:05 in fact, slyrus: have you tried your pathc on 10.8? 00:47:03 wbooze [~wbooze@xdsl-84-44-211-180.netcologne.de] has joined #sbcl 01:03:27 rbarraud [~rbarraud@222-155-139-54.jetstream.xtra.co.nz] has joined #sbcl 01:04:02 -!- rbarraud [~rbarraud@222-155-139-54.jetstream.xtra.co.nz] has quit [Read error: Connection reset by peer] 01:09:49 -!- wbooze [~wbooze@xdsl-84-44-211-180.netcologne.de] has quit [Ping timeout: 272 seconds] 01:14:05 pkhuong: 10.8? yeah, that's what I've been using all day 01:14:46 slyrus: you pinged? 01:15:17 hey jsnell, I was hoping we were going to have a quick fix for the 10.8 problem and I wanted to see if you wouldn't mind that going in while we're frozen, even though it's not a "regression". 01:15:23 slyrus: well, the x86 version with %fs hack seems mainline-worthy to me. 01:15:55 and, jsnell, of course, if you want to weigh in on the mach thread games we're playing, that's welcome too :) 01:16:42 esp if you can think of a way to read an x64 thread's TLS with ptrace-style access. 01:17:47 fine by me, if you're fairly confident it doesn't make things worse on older versions of OS X :-) 01:17:59 hah! :) 01:18:19 jsnell: if anything, it might help a bit there as well 01:19:06 neat 01:20:35 -!- chturne [~chturne@host86-148-233-236.range86-148.btcentralplus.com] has quit [Ping timeout: 272 seconds] 01:21:18 wbooze [~wbooze@xdsl-84-44-211-180.netcologne.de] has joined #sbcl 01:22:15 -!- wbooze [~wbooze@xdsl-84-44-211-180.netcologne.de] has quit [Client Quit] 01:23:15 wbooze [~wbooze@xdsl-84-44-211-180.netcologne.de] has joined #sbcl 01:26:34 mm... It seems like CCL forces addresses in the lower 32 bit. 01:33:04 looks like we can expect random failures too. 01:33:19 http://svn.clozure.com/publicsvn/openmcl/trunk/source/lisp-kernel/thread_manager.c <- allocate_tcr for !win32, 64 bit darwin. 01:35:33 -!- hargettp [~hargettp@pool-71-174-139-98.bstnma.east.verizon.net] has quit [Quit: Leaving...] 01:39:22 ok. I'll allocate space for 1G pointers and double indirect. 01:40:14 Fare [~Fare@c-68-81-138-209.hsd1.pa.comcast.net] has joined #sbcl 02:01:38 moment of truth... 02:02:13 ok, so far so good. 02:02:50 I'm sure there's some reason why this is a bad idea, but until somebody points it out to me, this seems kind of cute. 02:03:19 pkhuong: rwlock for all_threads_lock, just grab a read lock in catch_exception_raise and everybody's happy. 02:10:39 -!- slyrus [~chatzilla@adsl-99-190-99-176.dsl.pltn13.sbcglobal.net] has quit [Ping timeout: 250 seconds] 02:11:36 I don't understand the system enough to convince myself there's no potential deadlock (: 02:14:23 slyrus [~chatzilla@adsl-99-190-99-176.dsl.pltn13.sbcglobal.net] has joined #sbcl 02:17:26 well, the threads/GC suggest there's actual deadlock, not potential deadlock :( 02:18:18 threads/GC tests, that is 02:19:42 I'll post something that's closer to what CCL does as soon as it builds 02:33:18 slyrus: can you try pvk.ca/misc/ccl-like-exception-ports.diff on top of what you have on github? 02:44:27 building 02:50:45 must be a good sign if it hasn't died yet? 02:55:58 heh 02:56:01 slow build :) 02:56:12 need SSD! 02:57:07 but, yes, so far so good :) 02:58:13 it's a hack and random changes to malloc will bring thread creation to a halt. 02:58:14 echo-area [~user@182.92.247.2] has joined #sbcl 02:58:41 ack: arithmetic error FLOATING-POINT-OVERFLOW signalled 02:58:48 in ::: Running :GC-DEADLOCK 03:00:53 who knows? I sure don't. 03:02:49 this time it completed 03:02:54 ship it! 03:04:41 oops. ldb this time. 03:04:47 ooh, how? 03:04:56 -!- echo-area [~user@182.92.247.2] has quit [Quit: bbl] 03:05:03 Fault @ 0x1002df5530, page 1470 not marked as write-protected 03:05:47 echo-area [~user@182.92.247.2] has joined #sbcl 03:05:50 ok, that's bad. 03:06:31 in tests? 03:07:29 yes, threads.impure.lisp 03:08:07 the good news is that it's not faulting in catch_exception_raise (: 03:16:45 I don't understand how these changes could cause that symptom. 03:22:59 slyrus: especially given that you didn't observe that in the earlier race-ful commits? 03:23:53 I hear ya 03:28:01 how's x86? 03:44:14 building 03:44:42 great. unreproduceable failure in tests here as well. 03:57:02 x86 works 04:20:26 cool, I might be on to something (: 04:21:23 oh? 04:21:38 do you want me to commit my first batch of changes and then you can commit on top of that? 04:22:12 including the race-ful stuff? 04:22:38 I'm really not hot on the idea of having the all_threads traversal in mainline 04:34:54 slyrus: change the test in find_receive_port to "if ((mach_port_t)curr != (unsigned long)curr) goto skip;" seems to work here. 05:18:28 -!- slyrus [~chatzilla@adsl-99-190-99-176.dsl.pltn13.sbcglobal.net] has quit [Ping timeout: 246 seconds] 05:45:43 -!- psilord1 [~psilord@c-69-180-173-249.hsd1.mn.comcast.net] has quit [Quit: Leaving.] 06:15:02 What's the status update on OS 10.8 threading? 06:15:52 we have something that seems to work for x86, and I believe pvk.ca/misc/ccl-like-exception-ports2.diff, on top of https://github.com/slyrus/sbcl/commits/mach-thread-cleanup-2, will also work on x86-64. 06:16:57 so both of those are necessary for x86-64? 06:17:36 I might give them a whirl and try to put them through their paces 06:17:54 well, the ad hoc diff makes it not-obviously-thread-unsafe. 06:18:29 that'd be nice, thanks. I'll read the any result tomorrow morning. 06:31:01 sdemarre [~serge@91.176.57.77] has joined #sbcl 06:33:23 pkhuong, your patch is being rejected on HEAD 07:00:04 -!- kanru [~kanru@66.80.220.195] has quit [Ping timeout: 272 seconds] 08:11:36 TimKack [~user@213.208.236.186] has joined #sbcl 09:11:47 -!- sdemarre [~serge@91.176.57.77] has quit [Ping timeout: 272 seconds] 10:07:28 hargettp [~hargettp@pool-71-174-139-98.bstnma.east.verizon.net] has joined #sbcl 10:24:12 sdemarre [~serge@91.176.57.77] has joined #sbcl 10:33:38 scymtym [~user@2001:638:504:2093:226:b9ff:fe7d:3e1f] has joined #sbcl 10:46:12 -!- hargettp [~hargettp@pool-71-174-139-98.bstnma.east.verizon.net] has quit [Quit: Leaving...] 10:49:21 hargettp [~hargettp@pool-71-174-139-98.bstnma.east.verizon.net] has joined #sbcl 11:03:31 -!- echo-area [~user@182.92.247.2] has quit [Remote host closed the connection] 11:08:57 -!- Fare [~Fare@c-68-81-138-209.hsd1.pa.comcast.net] has quit [Ping timeout: 272 seconds] 11:21:04 -!- hargettp [~hargettp@pool-71-174-139-98.bstnma.east.verizon.net] has quit [Quit: Leaving...] 11:28:24 hargettp [~hargettp@pool-71-174-139-98.bstnma.east.verizon.net] has joined #sbcl 11:33:28 -!- hargettp [~hargettp@pool-71-174-139-98.bstnma.east.verizon.net] has quit [Quit: Leaving...] 12:21:03 LiamH [~none@pdp8.nrl.navy.mil] has joined #sbcl 12:36:01 Quadrescence: it's on top of slyrus's stuff. 12:36:27 -!- tsuru` [~charlie@adsl-74-179-29-153.bna.bellsouth.net] has quit [Read error: Operation timed out] 12:37:24 tsuru` [~charlie@adsl-74-179-28-75.bna.bellsouth.net] has joined #sbcl 13:07:09 chturne [~chturne@host86-148-233-236.range86-148.btcentralplus.com] has joined #sbcl 13:14:22 -!- wbooze [~wbooze@xdsl-84-44-211-180.netcologne.de] has quit [Read error: Connection reset by peer] 13:24:39 -!- minion [~minion@tiger.common-lisp.net] has quit [Remote host closed the connection] 13:24:42 minion [~minion@tiger.common-lisp.net] has joined #sbcl 13:32:59 slyrus [~chatzilla@adsl-99-190-99-176.dsl.pltn13.sbcglobal.net] has joined #sbcl 13:41:05 -!- huangjs [~huangjs@190.8.100.83] has quit [Read error: Connection reset by peer] 13:41:31 huangjs [~huangjs@190.8.100.83] has joined #sbcl 13:52:41 aceluck [~aceluck@183.78.10.115] has joined #sbcl 13:53:47 Hi. I'm having trouble with sbcl crashing in threaded code on Mountain Lion. Is there any way I can help resolve this? 13:54:10 <|3b|> if that is 10.8, they are working on thread stuff 13:54:10 By f.ex. giving remote login to my mountain lion machine? 13:54:25 Yes, it's 10.8. Good to hear 13:54:46 well, we're pretty sure we have a fix 13:54:56 two sec, I'm pushing to github 13:55:04 If I can help testing it in any way... 13:55:12 <|3b|> cool, i don't have to dig through logs for patches to link to :) 13:56:44 aceluck: can you install an older, unthreaded, sbcl (e.g. 1.0.54 I think) and build https://github.com/pkhuong/sbcl/tree/fix-10.8? 13:57:28 If I build without threading option, it works 13:57:44 except I need threading for running hunchentoot and other things 13:57:56 ok, you already have a working SBCL, good. 13:58:09 I still have a machine with Lion, so I'm working on that for now 13:58:27 but I also have a laptop with 10.8, and was wondering if I could help somehow 13:58:38 yes, this is what the instructions are about. 13:59:40 oh sorry, misread that. Sure , I can do that. Uhm, could it be built with clozure? 13:59:43 it's easier to run the tests after building on the same machine, and unthreaded SBCL should work fine on 10.8. You need a CL to build SBCL, and I suggested you install an unthreaded binary on the 10.8 machine to be able to build that tree 14:00:11 if have a working clozure (ccl64) running on the 10.8 machine 14:00:34 I haven't tried in ages. Probably, but I wouldn't try to debug two issues at the same time. 14:01:04 sure 14:01:10 let me give it a shot 14:03:11 -!- huangjs [~huangjs@190.8.100.83] has quit [Read error: Connection reset by peer] 14:11:48 antgreen [user@nat/redhat/x-upmgbrqgodnmpnga] has joined #sbcl 14:17:31 hargettp [~hargettp@65.219.237.244] has joined #sbcl 14:25:05 pkhuong: I tried building, it failed self tests for sb-rt and sb-concurrency 14:25:40 (or failed building those modules, not sure) 14:27:27 sb-rt?! 14:28:09 for sb-rt it says: Error while trying to load definition for system sb-rt from pathname sb-rt.asd: The value -933 is not of type UNSIGNED-BYTE 14:29:18 For sb-concurrency it fails in test: SB-CONCURRENCY-TEST::QUEUE.T.1 14:29:36 fatal error in SBL pid ***: mach_port_allocate_name failed with return code 5 14:30:02 aceluck: I don't think you built the right tree. 14:30:41 is there a git command I can run to verify ? 14:30:47 git branch 14:31:06 it just says master 14:31:10 or just "git checkout fix-10.8" 14:31:14 I did clone from the url you posted 14:31:38 yeah, there's a fair number of branches there (: 14:31:41 okay, I did the checkout 14:31:44 I'll build again 14:32:03 it's on the fix-10.8 branch now 14:32:08 I built the master before 14:32:12 pkhuong: Nice job on having a fix so quickly. :) 14:33:02 slyrus's the one with a 10.8 machine. I just mess around on a 10.6 MBA. 14:33:21 wbooze [~wbooze@xdsl-78-35-186-73.netcologne.de] has joined #sbcl 14:33:24 Will kudos to all :) 14:33:58 I will happily try doing some builds tonight; my Mac is at home. 14:37:30 -!- easye [~user@213.33.70.157] has quit [Remote host closed the connection] 14:37:46 easye [~user@213.33.70.157] has joined #sbcl 14:43:50 psilord [~psilord@23-25-144-217-static.hfc.comcastbusiness.net] has joined #sbcl 15:29:54 _travis_ [~nonya@c-98-244-123-185.hsd1.va.comcast.net] has joined #sbcl 15:30:11 fyi pkhuong 's branch works fine on my 10.8 15:32:02 Sweet :) 15:34:50 -!- aceluck [~aceluck@183.78.10.115] has quit [Remote host closed the connection] 15:36:31 -!- hargettp [~hargettp@65.219.237.244] has quit [Read error: Connection reset by peer] 15:36:36 hargettp_ [~hargettp@65.219.237.244] has joined #sbcl 15:42:11 -!- hargettp_ [~hargettp@65.219.237.244] has quit [Read error: Connection reset by peer] 15:42:16 hargettp [~hargettp@65.219.237.244] has joined #sbcl 15:57:36 -!- TimKack [~user@213.208.236.186] has quit [Remote host closed the connection] 16:28:02 dtw [~dtw@pdpc/supporter/active/dtw] has joined #sbcl 16:38:59 -!- hargettp [~hargettp@65.219.237.244] has quit [Quit: Colloquy for iPhone - http://colloquy.mobi] 17:23:13 -!- wbooze [~wbooze@xdsl-78-35-186-73.netcologne.de] has quit [Ping timeout: 246 seconds] 17:32:24 michaelw_ [~michaelw@tor.foldr.org] has joined #sbcl 18:03:19 stassats [~stassats@wikipedia/stassats] has joined #sbcl 18:21:40 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 248 seconds] 18:33:41 slyrus_ [~chatzilla@173-228-44-92.dsl.static.sonic.net] has joined #sbcl 18:40:46 kanru [~kanru@199.195.142.182] has joined #sbcl 19:21:02 redscare [~user@ool-4b7ff65e.static.optonline.net] has joined #sbcl 19:40:38 -!- redscare [~user@ool-4b7ff65e.static.optonline.net] has left #sbcl 19:45:17 wbooze [~wbooze@xdsl-78-35-186-73.netcologne.de] has joined #sbcl 20:32:42 -!- antgreen [user@nat/redhat/x-upmgbrqgodnmpnga] has quit [Remote host closed the connection] 20:57:45 -!- sdemarre [~serge@91.176.57.77] has quit [Quit: Leaving.] 20:57:50 -!- dtw [~dtw@pdpc/supporter/active/dtw] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 21:01:18 -!- _travis_ [~nonya@c-98-244-123-185.hsd1.va.comcast.net] has quit [Ping timeout: 264 seconds] 21:18:19 pkhuong, I think I mispatched (I pulled slyrus's stuff and applied your patch, which inadvertently applied to HEAD, and not his stuff) 21:18:50 Quadrescence: no, the patch was over slyrus's stuff. I put it up on github, if you want to try. It should be fine 21:20:06 Right, yeah, going to try now 21:26:24 pkhuong, I think reddit brought your blog down. 21:27:13 which is very sad, since there's zero dynamic content on it. 21:32:42 _travis_ [~nonya@c-24-127-49-108.hsd1.va.comcast.net] has joined #sbcl 21:33:36 pkhuong, I still get a warning during build for sb-concurrency, but the tests seem to pass 21:37:10 -!- LiamH [~none@pdp8.nrl.navy.mil] has quit [Quit: Leaving.] 21:37:45 pkhuong, while the tests went though okay, in the end, I got exit status 1 21:39:42 http://paste.lisp.org/+2SWK 21:41:15 can you try and re-run the tests with logging to find what the error is 21:41:17 ? 21:41:31 it'll probably be nondeterministic as well. 21:41:45 how? 21:42:04 I run in screen. 21:44:53 -!- slyrus_ [~chatzilla@173-228-44-92.dsl.static.sonic.net] has quit [Ping timeout: 250 seconds] 21:53:18 pkhuong, I ran again... apparent success 21:54:17 try again? These things aren't going to be deterministic. 21:54:56 although, hard failures should get a bug report dialog, I think. 21:57:36 I don't know if this is relevant. I did sh run-tests.sh > sbcl-tests, and this was the last thing that was shown to my terminal window. 21:59:28 I note that DESTRUCTURING-BIND is implemented using PARSE-DEFMACRO, and PARSE-DEFMACRO is implemented in terms of DESTRUCTURING-BIND. Can throw me a bone as to how that's working? Is one of the predefined somewhere? 21:59:39 *Can someone 22:00:16 parse-defmacro is a compiled function. 22:00:51 So, a working destructuring-bind macro only needed to exist in the *compiler* for it to be compiled into the right function. 22:02:37 -!- psilord [~psilord@23-25-144-217-static.hfc.comcastbusiness.net] has quit [Quit: Leaving.] 22:03:09 foom, do in a loose sense, destructuring-bind is being forward-declared in parse-defmacro? 22:03:13 *so 22:04:04 oops, forgot to paste: http://paste.lisp.org/display/130772#1 22:04:14 "i don't know if this ^^^ is relevant" 22:05:42 (well, parse-defmacro-lambda-list to be more precise) 22:06:52 Quadrescence: there might be some error spew around that. 22:15:54 -!- cmm [~cmm@bzq-109-64-234-229.red.bezeqint.net] has quit [Read error: Operation timed out] 22:19:34 cmm [~cmm@bzq-109-67-200-205.red.bezeqint.net] has joined #sbcl 22:41:19 -!- _travis_ [~nonya@c-24-127-49-108.hsd1.va.comcast.net] has quit [Ping timeout: 240 seconds]