00:35:13 -!- Quadrescence [~quad@unaffiliated/quadrescence] has quit [Quit: This computer has gone to sleep] 01:38:34 sea-theorem [~sea@fsf/member/sea4ever] has joined #sbcl 01:38:55 Hi all, I want to debug a running sbcl and make it eval a line of code 01:39:06 I'm not familiar with the source, so which function can I call to make it do that? 02:06:05 -!- davazp` [~user@31.200.179.50] has quit [Remote host closed the connection] 02:11:55 Nevermind, I managed to accomplish what I wanted to 02:22:01 Tribal [~Tribal@rcfreak0.com] has joined #sbcl 02:32:50 kanru` [~kanru@118-163-10-190.HINET-IP.hinet.net] has joined #sbcl 02:37:56 -!- christoph4 [~christoph@ppp-188-174-69-237.dynamic.mnet-online.de] has quit [Read error: Operation timed out] 02:40:30 -!- Tribal [~Tribal@rcfreak0.com] has quit [Quit: Leaving] 02:52:16 christoph4 [~christoph@ppp-93-104-178-147.dynamic.mnet-online.de] has joined #sbcl 03:09:15 Tribal [~Tribal@rcfreak0.com] has joined #sbcl 03:22:12 attila_lendvai [~attila_le@92.47.226.219] has joined #sbcl 03:22:12 -!- attila_lendvai [~attila_le@92.47.226.219] has quit [Changing host] 03:22:12 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 03:24:44 -!- Tribal [~Tribal@rcfreak0.com] has quit [Quit: Leaving] 03:28:46 Tribal [tribal@rcfreak0.com] has joined #sbcl 04:07:34 -!- Tribal [tribal@rcfreak0.com] has quit [Quit: Leaving] 04:10:49 Strigoides [~owen@114-134-0-126.lightwire.co.nz] has joined #sbcl 04:11:25 drmeister: it's only for the value cell. all other properties are global (e.g. CL doesn't have a facility for directly doing dynamically scoped functions) 04:11:43 slyrus [~chatzilla@adsl-99-183-240-185.dsl.pltn13.sbcglobal.net] has joined #sbcl 04:11:51 Tribal [tribal@rcfreak0.com] has joined #sbcl 04:12:11 jsnell_: Thanks - I should have known that. 04:12:28 one array per thread. but it's not the only possible design, you could equally well have an N-element array for each symbol (N=number of threads) 04:13:37 And can this all be handled within symbol-value and setf-symbol-value? 04:15:14 I'm asking because I've written a single thread Common Lisp compiler and I'm thinking about multi-threading. 04:16:03 you'd need to handle it for binding and unbinding as well, of course 04:16:17 Yes. 04:16:28 but really the hard parts of threading are going to be elsewhere 04:16:38 What are the hard parts? GC? 04:19:10 -!- sea-theorem [~sea@fsf/member/sea4ever] has left #sbcl 04:19:28 that can certainly be hard, depending on your ambitions. but I was mostly thinking of getting locking right for all of the (possibly hidden) global state 04:19:58 stuff like the internal data structures of the CLOS implementation 04:20:51 Yeah, I really don't know what the issues are yet. 04:21:52 I'm hosting the ECL source code on a new core written in C++ - I'm commenting out any multi-threading code and I was toying with leaving it in and trying to get it to work. 04:22:29 what in sbcl's clos requires that lock anyway? "efficient method lookup with PCL" implied generalizing to multi-threadedness would be easy, was it not? 04:23:08 There was some multithreading code in the generic function dispatch code if I remember correctly. 04:23:35 But I've mostly been commenting it out and ignoring it. 04:24:11 Mostly -> invariably 04:25:08 Then I read something about how dynamic variables make multithreading easy and I started wondering if I was doing the right thing. 04:25:31 Well, not make multithreading easy but they fit nicely with multithreading. 04:34:35 -!- slyrus [~chatzilla@adsl-99-183-240-185.dsl.pltn13.sbcglobal.net] has quit [Ping timeout: 256 seconds] 04:48:27 -!- foreignFunction [~niksaak@ip-4761.sunline.net.ua] has quit [Ping timeout: 252 seconds] 05:10:26 teggi [~teggi@113.172.61.193] has joined #sbcl 05:24:18 ehaliewicz [~user@50-0-51-11.dsl.static.sonic.net] has joined #sbcl 05:40:14 prxq [~mommer@mnhm-4d010804.pool.mediaWays.net] has joined #sbcl 05:51:17 -!- ehaliewicz [~user@50-0-51-11.dsl.static.sonic.net] has left #sbcl 06:02:11 slyrus [~chatzilla@adsl-99-183-240-185.dsl.pltn13.sbcglobal.net] has joined #sbcl 06:31:50 -!- Bike [~Glossina@75-175-78-193.ptld.qwest.net] has quit [Ping timeout: 252 seconds] 06:33:41 Bike [~Glossina@75-175-78-193.ptld.qwest.net] has joined #sbcl 06:37:37 cmm [~cmm@bzq-79-180-138-123.red.bezeqint.net] has joined #sbcl 06:37:43 -!- cmm- [~cmm@bzq-79-183-105-210.red.bezeqint.net] has quit [Read error: Operation timed out] 06:54:34 scymtym [~user@89.31.118.161] has joined #sbcl 06:55:23 at 20:00 my time (19:00 UTC) is the IRC meeting for GSoC slot deduplication, which sounds scarier than it is 06:55:30 I will try to remember to be there 08:35:23 -!- Bike [~Glossina@75-175-78-193.ptld.qwest.net] has quit [Ping timeout: 260 seconds] 08:55:59 pkhuong: so, brief thoughts on the patch, which looks correct for our use of it 08:56:28 in the past, change-ref-leaf called derive-node-type, so implicitly the contract of callers of change-ref-leaf was to change the ref to something that was type-compatible 08:58:02 if we now do from-scratch recomputation of that node (in modarith), effectively the new contract is that the new value will cause that value's node to be of a different type, but will not change incompatibly any downstream types 09:04:27 so. My question would then be whether modarith's use of it is then safe (thinking about nested (logand #xfoo (+ x (+ y (+ z #xstupidconstant)))), and imagining the constant replacement and its type propagation 09:05:02 and my answer is that I am not at all sure 09:07:39 so. Will build and poke 09:23:57 Quadrescence [~quad@unaffiliated/quadrescence] has joined #sbcl 09:30:40 ; 73: L5: EBD6 JMP L2 09:30:40 ; 75: L6: EBD4 JMP L2 09:30:40 ; 77: L7: EBD2 JMP L2 09:30:40 ; 79: L8: EBD0 JMP L2 09:30:44 stupid compiler 09:34:04 stassats` [~stassats@wikipedia/stassats] has joined #sbcl 09:34:36 Krystof, do you know if the compiler fixes: JMP A ; A: JMP B ? 09:34:55 well clearly not from that example 09:35:01 unless it's a bug 09:35:20 it does not 09:36:03 pkhuong: so my extremely limited poking doesn't reveal a problem with the patch along those lines, but your thoughts would still be appreciated 09:36:13 I think I'll ping emarsden too to see if he has bandwidth for random testing 09:54:30 -!- kanru [~kanru@118-163-10-190.HINET-IP.hinet.net] has quit [Remote host closed the connection] 10:08:03 kanru [~kanru@118-163-10-190.HINET-IP.hinet.net] has joined #sbcl 10:31:36 davazp [~user@178.167.188.52.threembb.ie] has joined #sbcl 10:42:43 -!- kanru` [~kanru@118-163-10-190.HINET-IP.hinet.net] has quit [Ping timeout: 240 seconds] 12:26:05 -!- drmeister [~drmeister@pool-173-59-25-70.phlapa.fios.verizon.net] has quit [Remote host closed the connection] 12:27:10 segv- [~mb@95-91-241-81-dynip.superkabel.de] has joined #sbcl 12:35:02 drmeister [~drmeister@pool-173-59-25-70.phlapa.fios.verizon.net] has joined #sbcl 12:39:07 -!- stassats` [~stassats@wikipedia/stassats] has quit [Ping timeout: 240 seconds] 12:40:43 -!- drmeister [~drmeister@pool-173-59-25-70.phlapa.fios.verizon.net] has quit [Remote host closed the connection] 12:44:55 -!- davazp [~user@178.167.188.52.threembb.ie] has quit [Ping timeout: 264 seconds] 12:58:19 Hydan [~hydan@ip-89-103-110-5.net.upcbroadband.cz] has joined #sbcl 13:02:23 -!- Quadrescence [~quad@unaffiliated/quadrescence] has quit [Quit: This computer has gone to sleep] 13:03:14 Quadrescence [~quad@unaffiliated/quadrescence] has joined #sbcl 13:06:40 -!- Hydan [~hydan@ip-89-103-110-5.net.upcbroadband.cz] has quit [Remote host closed the connection] 13:08:38 Hydan [~hydan@ip-89-103-110-5.net.upcbroadband.cz] has joined #sbcl 13:15:37 LiamH [~none@pdp8.nrl.navy.mil] has joined #sbcl 13:19:53 -!- Strigoides [~owen@114-134-0-126.lightwire.co.nz] has quit [Quit: leaving] 13:32:20 foreignFunction [~niksaak@ip-4761.sunline.net.ua] has joined #sbcl 13:33:20 nyef [~nyef@c-50-157-244-41.hsd1.ma.comcast.net] has joined #sbcl 13:33:30 G'morning all. 13:47:17 hello 13:48:05 Anything interesting going on at this point? 13:48:45 nyef, it has been quiet 13:49:08 Fair enough. 13:49:40 from pinging emarsden, he says that he has (for the last two months or so) been getting substantial amounts of heap corruption from the random integer tester 13:49:52 Lovely. 13:49:56 I think we need to be running this ourselves more often 13:51:33 Given that it's not a regression for this month, is it still going to hold up code freeze? 13:52:32 what's a random integer tester? 13:53:55 (oh, i see) 14:00:06 nyef: no, it won't hold code freeze 14:00:35 it was in response to "we think we've fixed a bunch of modarith bugs; can you run the tester to see if it agrees?" 14:00:40 which is unfortunate 14:02:15 Mmm. So, time to set up the random tester and start bisecting? 14:02:36 maybe. Or thinking very hard... 14:04:29 Bisection helps with providing a more accurate starting point for the thinking. 14:04:52 davazp [~user@178.167.188.52.threembb.ie] has joined #sbcl 14:14:43 -!- aeth [~Michael@wesnoth/umc-dev/developer/aethaeryn] has quit [Quit: ...] 14:15:45 Mm... you're quite right. I think we have to clobber all the derived types of parent node when we change something. 14:17:00 Quadrescence: we do a bit of jmp-jmp tensioning at the IR1 level, but we sometimes miss opportunities. It'd work better at the IR2 level 14:17:46 wbooze [~wbooze@xdsl-78-35-153-83.netcologne.de] has joined #sbcl 14:27:40 Krystof: which we do... what also worries me is that we don't always do modular replacement for functions. 14:28:49 shouldn't we do something similar to what we do for lambda-var then? 14:29:03 anyway, that's for later. let's just put that fire out. 14:31:52 Krystof: I think the patch is strictly better than what we had before. 14:40:25 does the following sound right? we only *need* to perform replacement when using signed modarith to implement unsigned operations 14:40:36 so, when is it useful to do that? 14:50:58 drmeister [~drmeister@farnsworth.chem.temple.edu] has joined #sbcl 14:51:21 I think that the *need* statement is right 14:51:55 I think it's useful in the cases that we do do it -- when the result is tagged anyway 14:52:05 it saves the shifts to untag / tag 14:52:35 I think you're also right that the patch is strictly better than before 14:53:59 -!- drmeister [~drmeister@farnsworth.chem.temple.edu] has quit [Remote host closed the connection] 14:57:04 -!- wbooze [~wbooze@xdsl-78-35-153-83.netcologne.de] has quit [Ping timeout: 256 seconds] 14:57:50 actually, what worries me is why are we using m-s-f when we use fixnum modular arith for unsigned ops. 15:05:31 wbooze [~wbooze@xdsl-78-35-153-83.netcologne.de] has joined #sbcl 15:13:36 interesting. we insert a full call to two-arg-and instead of AND for fixnums, under (optimize space) ... 15:16:39 whoops 15:17:57 drmeister [~drmeister@farnsworth.chem.temple.edu] has joined #sbcl 15:28:53 so, I think the "right"t way to do this is to move the m-s-f/logand insertion logic to CUT-LVAR. 15:30:16 if we punted on some of the node (either it's not a ref/combination, or it's a combination without modular goodness), filter-lvar in cut-lvar 15:33:35 -!- teggi [~teggi@113.172.61.193] has quit [Remote host closed the connection] 15:38:22 hm. is rsync too non-portable to be used by sbcl build script? 15:38:35 in asdf-module.mk, the last 2 lines can be simply 15:38:37 rsync -vrl ./ "$(BUILD_ROOT)$(INSTALL_DIR)/" 15:38:41 -!- scymtym [~user@89.31.118.161] has quit [Ping timeout: 248 seconds] 15:39:02 is rsync (commonly) available on mingw platforms? 15:39:22 I don't know anything about mingw 15:40:06 I need a local patch to do that in any case, because the tar/chown doesn't work with my build system. 15:40:11 But if it needs to stay local, that's fine. 15:41:34 (doesn't work because tar *really* *really* wants to set the permissions and owner of any dir it touches, but some dirs get pre-created and owned by another user.) 15:41:57 (which I'll admit is a bit of an unusual case) 15:56:19 so, the answer to why do we work hard to preserve the sign bit when compiling unsigned ops as fixnum arithmetic: modular shifting. 15:59:35 ah yes 16:01:34 and, nope. we only do that for left shifts 16:07:13 -!- davazp [~user@178.167.188.52.threembb.ie] has quit [Ping timeout: 240 seconds] 16:35:23 -!- slyrus [~chatzilla@adsl-99-183-240-185.dsl.pltn13.sbcglobal.net] has quit [Ping timeout: 256 seconds] 17:07:11 -!- wbooze [~wbooze@xdsl-78-35-153-83.netcologne.de] has quit [Quit: none] 17:09:36 foreignFunction1 [~niksaak@ip-4761.sunline.net.ua] has joined #sbcl 17:10:21 -!- foreignFunction [~niksaak@ip-4761.sunline.net.ua] has quit [Read error: Connection reset by peer] 17:13:30 Bike [~Glossina@75-175-78-193.ptld.qwest.net] has joined #sbcl 17:30:41 -!- Vivitron [~Vivitron@pool-98-110-213-33.bstnma.fios.verizon.net] has quit [Remote host closed the connection] 17:42:01 akovalenko [~user@77.51.76.204] has joined #sbcl 17:49:41 wbooze [~wbooze@xdsl-78-35-153-83.netcologne.de] has joined #sbcl 17:50:32 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 17:51:00 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 17:51:19 -!- akovalenko [~user@77.51.76.204] has quit [Ping timeout: 256 seconds] 17:52:17 stassats` [~stassats@wikipedia/stassats] has joined #sbcl 17:55:52 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Client Quit] 17:57:13 pkhuong: is the patch you posted yesterday destined for commit, or did you figure out that it's broken in some way? 18:06:33 I'd quite like to freeze this evening if possible 18:06:50 I have a busy weekend, mostly absent from connectivity 18:09:04 well, the if/if commit can always be reverted instead 18:09:32 *stassats`* is in a voluntary freeze, too many commits to sort through already 18:10:01 I think I'm just waiting for pkhuong's go/no-go on a couple of things 18:11:29 i'd like some opinion on 18:11:31 lp 1182204 18:11:31 https://bugs.launchpad.net/bugs/1182204 18:12:22 what's better, store specifier-type of types, or throw away class wrapper unconditionally when the types are not :primitive 18:14:58 (c) have the type checker function for deftypes be a funcallable object which lazily compile, and have a hook on deftype which invalidates the compiled version? 18:15:10 (d) fall back on "the consequences are undefined" in 3.2.2.3 18:15:27 it's already d 18:16:30 i think that c is not necessary, since it's understandable that after changing deftype recompilation is needed, just that in this case recompilation doesn't do anything 18:16:51 since we check against the type specifier, which remains the same, even though the expansion is different 18:24:13 Oh, right 18:26:55 davazp [~user@31.200.128.17] has joined #sbcl 18:32:48 Bike_ [~Glossina@75-175-75-143.ptld.qwest.net] has joined #sbcl 18:35:14 -!- Bike [~Glossina@75-175-78-193.ptld.qwest.net] has quit [Ping timeout: 256 seconds] 18:35:29 -!- Bike_ is now known as Bike 18:36:55 hm, so what is this "LAMBDA (SB-PCL::CACHE SB-PCL::EMF SB-PCL::MISS-FN)" when loading contribs? 18:37:00 didn't happen before 18:37:40 I think that's also one of pkhuong's last-minute issues 18:45:37 heap allocating floats is strange, sometimes it decides to do so with literal floats at compile time, sometimes at run-time 18:46:03 yes 18:46:13 even if i just pass it around 18:46:38 (bar 1d0) allocates at ct, (sleep 1d0) at rt 18:47:35 it used to be we never allocated at runtime, then with inline constants, I made it always allocate at runtime, and then nikodemus added some heuristic to avoid allocating arguments to full calls 18:48:24 so, resort to (load-time-value 1d0) now? 18:50:14 -!- Krystof is now known as sbcl_Krystof 18:50:31 (sorry for the stupid nick. Googly IRC meeting in 10 minutes) 18:51:06 -!- wbooze [~wbooze@xdsl-78-35-153-83.netcologne.de] has quit [Ping timeout: 256 seconds] 18:52:44 there. 18:55:05 done? Thank you all for a fantastic month! :-) 19:01:07 -!- Bike [~Glossina@75-175-75-143.ptld.qwest.net] has quit [Quit: Reconnecting] 19:01:11 Bike_ [~Glossina@75-175-75-143.ptld.qwest.net] has joined #sbcl 19:07:13 -!- Bike_ is now known as Bike 19:17:43 -!- drmeister [~drmeister@farnsworth.chem.temple.edu] has quit [Remote host closed the connection] 19:26:25 -!- sbcl_Krystof is now known as Krystof 19:30:07 good, no more pcl saying "hi" 19:36:20 leuler [~user@p548FDB51.dip0.t-ipconnect.de] has joined #sbcl 19:36:48 GSoC "deduplication" meeting over: no change to our current status 19:36:57 offical results on Monday 19:37:22 uh oh 19:37:30 debugger invoked on a SB-INT:BUG in thread #: full call to SB-KERNEL:DATA-VECTOR-REF-WITH-OFFSET 19:38:06 pkhuong: that's one of your recentest commits 19:38:13 fuck it. revert it. Next to no one complains about the bug it fixes. 19:39:14 <|3b|> hmm, "::: Running (:SEMAPHORE-NOTIFICATION :WAIT-ON-SEMAPHORE)" for at least 5 min now i think 19:39:19 revert it and add a test case 19:39:23 stassats`: do you have a test case? 19:39:29 not yet 19:40:48 Krystof: The unicode normalization tests fail (naturally) under (not :sb-unicode). The tests should IMO either be skipped or marked as failed under that condition. I'd like to commit something to that effect. Which do you prefer? 19:41:20 drmeister [~drmeister@farnsworth.chem.temple.edu] has joined #sbcl 19:42:10 oh whoops I forgot to do that 19:42:29 the existing tests can be skipped, I think 19:42:43 they explicitly test normalization of Unicode data, so it doesn't make sense to run them under non-unicode 19:42:57 thank you for catching that 19:43:44 OK, I'll prepare a patch. I assume it's OK to commit that during the freeze. 19:44:56 yes! 19:50:30 -!- nyef [~nyef@c-50-157-244-41.hsd1.ma.comcast.net] has quit [Quit: G'night all.] 19:52:58 <|3b|> and running body of that test in a loop errored with "The assertion (= N (+ K (LENGTH SAFE))) failed with N = 54, 19:53:00 <|3b|> (+ K (LENGTH SAFE)) = 53." 19:53:05 <|3b|> after 47 tries 19:53:42 |3b|: what platform? 19:53:54 <|3b|> x8664 linux 19:54:21 extra details, then; I've never seen that one fail, and that is my platform too 19:54:49 -!- leuler [~user@p548FDB51.dip0.t-ipconnect.de] has quit [Quit: ERC Version 5.1.2 $Revision: 1.796.2.6 $ (IRC client for Emacs)] 19:56:16 *|3b|* hasn't seen it before either, and i've run tests on builds from last few days a few times 19:56:34 <|3b|> i think that was second run through tests on pull from a few minutes ago 19:56:59 <|3b|> built with --fancy and dynamic-space-size 8gb 19:57:23 hm 19:57:23 *|3b|* supposes i should remove my small local changes and see if i can reproduce it still 19:58:08 <|3b|> running in a loop again, possibly locked up after 23 tries 19:58:35 <|3b|> and ignores C-c 20:00:21 <|3b|> ah, interrupted that time, waiting for 'thread result lock' mutex 20:00:53 the test is racy... 20:01:06 <|3b|> lp 1038034 20:01:06 https://bugs.launchpad.net/bugs/1038034 20:01:15 *|3b|* thought that last bit sounded familiar 20:01:23 <|3b|> so i guess not a regression 20:01:40 pkhuong: http://paste.lisp.org/display/137298 20:03:09 -!- edgar-rft [~GOD@HSI-KBW-149-172-63-75.hsi13.kabel-badenwuerttemberg.de] has quit [Quit: sigh...] 20:06:58 stassats`: thanks. looking at it. 20:07:13 it breaks: cl-ppcre 20:08:04 yeah... we have hand-rolled VOPs with stronger constraints. 20:08:26 revert, I'll add test cases, and come back with a complete patch next month. 20:15:45 I don't know how we derived that the MOD subform could evaluate to -1. 20:18:21 (length vector) could be 0 20:18:55 I think that doesn't actually mean that mod can be -1, but I'd bet that the derive-type doesn't know that 20:19:26 if I splice in a straight (mod x -1), it finds (integer 0 0) 20:22:21 what about (mod x )? 20:23:37 wait, where *is* there any information about MOD? 20:23:49 I see only a transform which fires only if the second arg is a constant 20:28:41 it inlines code that calls truncate, iirc 20:31:51 I see that effect in disassembly 20:32:27 ah, inline 20:32:54 maybe next month I will put that in sb-int:info :function :info where I can see it 20:33:04 so that I don't have to remember how to spell :inline-expansion-designator 20:35:00 wbooze [~wbooze@xdsl-78-35-131-21.netcologne.de] has joined #sbcl 20:38:12 ASau` [~user@p5797E072.dip0.t-ipconnect.de] has joined #sbcl 20:41:55 -!- ASau [~user@p4FF965AB.dip0.t-ipconnect.de] has quit [Ping timeout: 264 seconds] 20:43:48 I think I know why data-vector-ref doesn't fail as much as it should: if our logic fires, it's most probably in unsafe code, and, in unsafe code, the INDEX type check is optimised away (but guarantees that the VOPs apply) 20:51:12 (the bare -1 in mod gets transformed into logand, assuming that deftransforms fire before inlining) 20:52:12 -!- ASau` is now known as ASau 20:55:28 -!- drmeister [~drmeister@farnsworth.chem.temple.edu] has quit [Read error: Connection reset by peer] 20:56:07 drmeister [~drmeister@farnsworth.chem.temple.edu] has joined #sbcl 21:01:28 slyrus [~chatzilla@173-228-44-92.dsl.static.sonic.net] has joined #sbcl 21:06:52 -!- prxq [~mommer@mnhm-4d010804.pool.mediaWays.net] has quit [Remote host closed the connection] 21:17:55 -!- stassats` [~stassats@wikipedia/stassats] has quit [Ping timeout: 264 seconds] 21:45:53 -!- segv- [~mb@95-91-241-81-dynip.superkabel.de] has quit [Remote host closed the connection] 21:54:55 -!- drmeister [~drmeister@farnsworth.chem.temple.edu] has quit [Remote host closed the connection] 21:57:15 -!- slyrus [~chatzilla@173-228-44-92.dsl.static.sonic.net] has quit [Ping timeout: 256 seconds] 22:01:43 -!- LiamH [~none@pdp8.nrl.navy.mil] has quit [Quit: Leaving.] 22:55:24 ehaliewicz [~user@50-0-51-11.dsl.static.sonic.net] has joined #sbcl 22:58:44 -!- davazp [~user@31.200.128.17] has quit [Ping timeout: 256 seconds] 23:30:26 drmeister [~drmeister@pool-71-185-168-200.phlapa.fios.verizon.net] has joined #sbcl 23:33:32 *|3b|* builds a windows sbcl, gets failed tests on sb-concurrency, and apparent lockup in (:condition-variable :notify-multiple) 23:35:53 -!- Quadrescence [~quad@unaffiliated/quadrescence] has quit [Quit: This computer has gone to sleep] 23:36:12 <|3b|> last thing it printed was (condition-broadcast queue) followed by 10 pairs of 'woke up' and 'condition-wait' 23:51:17 <|3b|> on 2nd try running tests, i get lots of expected failures, "Unexpected success: debug.impure.lisp / BACKTRACE-INTERRUPTED-CONDITION-WAIT 23:51:41 <|3b|> and invalid exit status from threads.impure.lisp 23:51:41 Krystof: for cut-to-width, something like https://github.com/pkhuong/sbcl/commit/9b780a98704e23768e7a46a9736d9ca5a12587a9 seems more likely to plug all the holes 23:54:53 -!- foreignFunction1 [~niksaak@ip-4761.sunline.net.ua] has quit [Quit: Leaving.] 23:59:11 <|3b|> and running just threads.impure.lisp i got another lockup, and a segfault (or whatever exception c0000005 is)