00:01:55 chu [~user@unaffiliated/chu] has joined #sbcl 00:03:32 Bike [~Glossina@71-222-56-127.ptld.qwest.net] has joined #sbcl 00:08:23 -!- ehaliewicz [~user@50-0-51-11.dsl.static.sonic.net] has quit [Ping timeout: 256 seconds] 00:35:19 -!- cmm [~cmm@bzq-109-67-119-85.red.bezeqint.net] has quit [Read error: Operation timed out] 00:36:06 cmm [~cmm@bzq-79-177-123-220.red.bezeqint.net] has joined #sbcl 00:43:44 -!- echo-area [~user@114.254.102.122] has quit [Remote host closed the connection] 00:45:03 Fare [~fare@150.Red-83-60-131.dynamicIP.rima-tde.net] has joined #sbcl 00:51:40 -!- chu [~user@unaffiliated/chu] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 00:54:45 -!- maxm [~user@75.99.81.170] has quit [Read error: Connection reset by peer] 00:54:53 maxm [~user@unaffiliated/maxm] has joined #sbcl 01:02:23 -!- foreignFunction [~niksaak@ip-4761.sunline.net.ua] has quit [Quit: Leaving.] 01:20:38 chu [~user@unaffiliated/chu] has joined #sbcl 01:33:18 LiamH [~none@pool-173-48-164-128.bstnma.fios.verizon.net] has joined #sbcl 02:16:56 -!- Fare [~fare@150.Red-83-60-131.dynamicIP.rima-tde.net] has quit [Remote host closed the connection] 02:17:34 -!- wbooze [~wbooze@xdsl-78-35-172-184.netcologne.de] has quit [Ping timeout: 252 seconds] 02:38:17 -!- christoph4 [~christoph@ppp-188-174-170-91.dynamic.mnet-online.de] has quit [Read error: Operation timed out] 02:52:29 christoph4 [~christoph@ppp-188-174-89-217.dynamic.mnet-online.de] has joined #sbcl 03:16:26 foreignFunction [~niksaak@ip-4761.sunline.net.ua] has joined #sbcl 03:29:42 -!- foreignFunction [~niksaak@ip-4761.sunline.net.ua] has quit [Quit: Leaving.] 03:30:43 -!- LiamH [~none@pool-173-48-164-128.bstnma.fios.verizon.net] has quit [Quit: Leaving.] 03:45:42 ehaliewicz [~user@50-0-51-11.dsl.static.sonic.net] has joined #sbcl 03:46:28 yacks [~py@180.151.36.168] has joined #sbcl 03:58:55 -!- robgssp [~user@cpe-24-93-28-218.rochester.res.rr.com] has quit [Remote host closed the connection] 04:00:00 robgssp [~user@cpe-24-93-28-218.rochester.res.rr.com] has joined #sbcl 04:10:45 echo-area [~user@182.92.247.2] has joined #sbcl 04:42:55 -!- asedeno [asedeno@nat/google/x-ecodotthpsavzxph] has quit [Ping timeout: 260 seconds] 04:44:32 asedeno [~asedeno@216.239.33.112] has joined #sbcl 04:54:02 -!- drmeister [~drmeister@pool-71-185-168-200.phlapa.fios.verizon.net] has quit [Remote host closed the connection] 05:06:42 -!- asedeno [~asedeno@216.239.33.112] has quit [Ping timeout: 260 seconds] 05:07:58 prxq [~mommer@mnhm-590c0575.pool.mediaWays.net] has joined #sbcl 05:08:18 asedeno [asedeno@nat/google/x-ggovbyshjkqtlvkv] has joined #sbcl 05:15:37 teggi [~teggi@113.173.4.86] has joined #sbcl 05:23:04 attila_lendvai [~attila_le@92.47.184.117] has joined #sbcl 05:23:04 -!- attila_lendvai [~attila_le@92.47.184.117] has quit [Changing host] 05:23:04 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 05:34:35 *Krystof* cries 05:35:30 time from release to first complaint about release: 7 hours 05:36:19 Krystof: bright side: you have users. 05:36:30 on it. 05:36:43 Krystof: You may recall that that is a luxury that I don't really get to enjoy. :) 05:36:58 easy bug too. the code even looks wrong :\ 05:39:13 the rc's looking really good now, eh? ;) 05:39:18 *sigh* 05:43:40 brucem: I take that point. Who knew that fedora used sbcl to distribute some obscure verification tool from SRI? 05:45:12 PVS isn't that obscure. Now I'm really scared of introducing IR1 changes. 05:46:32 it's not in debian, so it doesn't exist :-) 05:47:48 -!- Blkt [~user@dynamic-adsl-62-10-11-72.clienti.tiscali.it] has quit [Ping timeout: 256 seconds] 05:51:20 Probably has a funky license. It's SRI after all. 05:56:45 astonishingly according to the download page it's GPL 05:57:03 (I can't actually get the reduced test case to error in the way described in the mail) 05:57:56 it's really nifty. 05:58:15 we derive a NIL return type for nconc, and proceed to call, then fall through to the arg count error trap 05:58:26 The final form must be wrapped in a defun. 06:03:41 ok. the fixed type deriver is both shorter and more precise. I'll commit tomorrow morning. 06:06:40 thanks 06:07:13 one advantage of a long release cycle is a subsequent short release cycle 06:10:07 does LANG,LC_ALL=C lead to *default-external-format*=:ascii? 06:11:01 I think it does. 06:11:05 yes 06:11:33 pkhuong, Krystof: thanks 06:12:07 (well, strictly it depends on how your system is configured. Try LANG=C locale -k LC_CTYPE | grep charmap 06:12:09 ) 06:12:25 and if a system does not load with that locale/external format, the system's asd-file is at fault? 06:12:44 (still working on cl-test-grid regression testing) 06:13:21 scymtym_: I think the default assumption is now utf-8. 06:14:05 re nconc error, it would be a really good idea to make terminate ir2 blocks that don't fall through (and don't end with branches/returns) with some clearer error trap. 06:14:37 pkhuong: it thought that as well but running cl-test-grid with LANG,LC_ALL=C leads to new load failures which seem to be caused by *default-external-format* being :ascii 06:15:23 right, so you need to set an utf-8 locale. 06:15:55 it will try that as a workaround 06:16:17 ideally, systems with non-ascii source files would say so in their asd files 06:16:37 I think some of them have non-ascii asd files (: 06:16:51 :) 06:31:54 -!- ehaliewicz [~user@50-0-51-11.dsl.static.sonic.net] has quit [Ping timeout: 264 seconds] 06:35:30 ehaliewicz [~user@50-0-51-11.dsl.static.sonic.net] has joined #sbcl 06:50:02 pkhuong: you're right, a few systems failed while reading the asd files 07:32:48 benkard [~benkard@mnch-4d04dc04.pool.mediaWays.net] has joined #sbcl 07:49:06 -!- benkard [~benkard@mnch-4d04dc04.pool.mediaWays.net] has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz] 07:55:53 -!- scymtym_ [~user@ip-5-147-122-209.unitymediagroup.de] has quit [Ping timeout: 256 seconds] 07:57:26 leoc`` [~leoc.git@p5DDBA34A.dip0.t-ipconnect.de] has joined #sbcl 07:58:25 -!- leoc` [~leoc.git@p5DDB92A3.dip0.t-ipconnect.de] has quit [Ping timeout: 246 seconds] 08:11:19 benkard [~benkard@mnch-4d04dc04.pool.mediaWays.net] has joined #sbcl 08:13:02 -!- benkard [~benkard@mnch-4d04dc04.pool.mediaWays.net] has quit [Client Quit] 08:15:32 benkard [~benkard@mnch-4d04dc04.pool.mediaWays.net] has joined #sbcl 08:40:26 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 08:57:36 -!- Hydan [~hydan@ip-89-103-110-5.net.upcbroadband.cz] has quit [Remote host closed the connection] 09:11:56 Sagane [~Sagane@177.100-226-89.dsl.completel.net] has joined #sbcl 09:13:28 hlavaty [~user@friedrichstrasse.knowledgetools.de] has joined #sbcl 09:18:02 stassats [~stassats@wikipedia/stassats] has joined #sbcl 09:18:56 huh, sleep on windows can't sleep shorter than 0.01 second 09:20:19 on threaded sbcl, it uses some kind of timers 09:20:42 on non-threaded it uses millisleep 09:21:45 huh, that doesn't work well too 09:23:03 looks like it's inherent, can't sleep shorter than that 09:24:24 stassats: why do you need a shorter sleep than that? 09:25:09 i want to run a 1000 of them to check for consing 09:30:06 It's interesting how motivating the latest modular arithmetic bug is to work on all sorts of other issues 09:30:15 ohh 09:30:26 figures it was some special requirement. :-) 09:30:46 loke_: another thing is "linux does it" 09:33:13 leuler [~user@p548FC843.dip0.t-ipconnect.de] has joined #sbcl 09:35:48 stassats: Well, yeah I'm sure. :-) That said, I'm sure there is a way to do it on windows? 09:36:03 googling says "pretty much no" 09:36:21 not below 10ms 09:36:41 10ms, it's an eternity for the cpu 09:37:03 why hello srctran.lisp -- long time no see 09:38:50 and i forgot how to write double floats 09:39:11 "Syntax of a Float" to the rescue 09:39:42 *stassats* got used to 1d0 and couldn't figure how to do 1.5d0 09:42:07 oh, and just as i'm committing, Krystof fixed it 09:42:23 sorry :) 09:42:37 why hello ldb> -- long time no see 09:42:42 though, i still have some more tests to add 09:43:38 and the place where you put TRUNCATE won't solve consing problems 09:43:38 can we trace internal functions (defined using labels)? 09:44:00 Sagane_ [~Sagane@177.100-226-89.dsl.completel.net] has joined #sbcl 09:44:03 stassats: I admit to not thinking very hard about the consing 09:44:17 I wanted to get it to the position of being able to add test cases 09:44:23 well, it could be all very simple if not for bloody consing 09:46:45 -!- Sagane [~Sagane@177.100-226-89.dsl.completel.net] has quit [Ping timeout: 248 seconds] 09:47:36 http://stackoverflow.com/questions/7827062/is-there-a-windows-equivalent-of-nanosleep 09:47:45 stassats: that one at lease gives some suggests 09:47:48 suggestions 09:48:05 yes, i've seen that 09:52:38 ok, moved truncate instead of / 09:52:43 no consing again 09:56:32 marked the tests on non-consiness of sleep as not failing on all other platforms, but i have actually no idea whether they do 09:57:01 maybe they compute floats in some other fashion, or something else conses instead 09:59:37 and what do you know, they do fail on x86 10:00:37 Krystof: is it correct that defining a method on make-instance of a class whose metaclass is standard-class is implementation-defined ? 10:01:24 *stassats* doesn't get it, how could i test it before then? 10:08:02 just (values (truncate 1000000000 (denominator s))) conses on x86 10:08:26 Quadresce_ [~quad@unaffiliated/quadrescence] has joined #sbcl 10:08:26 or even just (truncate 1000000000 <5>) 10:09:18 Krystof: http://paste.lisp.org/+2Y1U 10:09:26 and i don't get how 10:10:06 1000000000 is a bignum on x86. 10:10:14 rudi_ [~rudi@1x-193-157-195-224.uio.no] has joined #sbcl 10:10:48 samskull1 [~samskulls@S0106001111de1fc8.cg.shawcable.net] has joined #sbcl 10:11:04 -!- rudi_ is now known as rudi 10:11:29 -!- ehaliewicz [~user@50-0-51-11.dsl.static.sonic.net] has quit [Remote host closed the connection] 10:11:56 aiee hold on I have a bug report next to me relating to an els talk 10:12:13 oh, damn, i forget 1-bit fixnum tags are only on x84 10:12:17 somehow a function's type has been derived as (or (double-float 0) (rational 0) (single-float 0.0)) 10:12:29 (and the (double-float 0) causes problems later because 0 is not a double-float) 10:13:23 -!- Quadrescence [~quad@unaffiliated/quadrescence] has quit [Read error: Connection reset by peer] 10:13:23 -!- samskulls [~samskulls@S0106001111de1fc8.cg.shawcable.net] has quit [Remote host closed the connection] 10:13:24 -!- samebchase [~samuel@codesurfers.net] has quit [Remote host closed the connection] 10:13:41 so, truncate twice? 10:14:04 i mean, how more ugly can split-seconds-for-sleep get 10:14:46 I believe truncate with a bignum input always conses, even if the result(s) is/are fixnums. 10:15:03 -!- Quadresce_ [~quad@unaffiliated/quadrescence] has quit [Quit: This computer has gone to sleep] 10:15:08 davazp [~user@195.Red-88-5-174.dynamicIP.rima-tde.net] has joined #sbcl 10:16:30 samebchase [~samuel@codesurfers.net] has joined #sbcl 10:17:11 ok, (* rem 10 (truncate 100000000 (denominator seconds))) doesn't cons... 10:17:31 and truncate should get it together too 10:17:54 it's because of (bignum-truncate number (make-small-bignum divisor)) 10:18:24 so, until someone write bignum-truncate-by-fixnum 10:25:25 -!- zulu_inuoe [~quassel@184.89.111.53] has quit [Ping timeout: 252 seconds] 10:29:58 -!- chu [~user@unaffiliated/chu] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 10:30:45 great, when split-seconds-for-sleep is compiled during the build, it conses, when i compile it live, it doesn't 10:31:59 or am i modifying different trees 10:32:17 of course, that was sbcl, not sbcl-x86 10:32:58 -!- benkard [~benkard@mnch-4d04dc04.pool.mediaWays.net] has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz] 10:33:57 benkard [~benkard@mnch-4d04dc04.pool.mediaWays.net] has joined #sbcl 10:42:33 argh argh argh SO MANY BUGS 10:44:41 finally, ::: Success (SLEEP :NON-CONSING) 10:46:41 ok, i hope it's the last of SLEEP commits 10:48:26 interesting way to start a test-case: (in-package :common-lisp) 10:50:07 -!- davazp [~user@195.Red-88-5-174.dynamicIP.rima-tde.net] has quit [Remote host closed the connection] 10:50:55 -!- benkard [~benkard@mnch-4d04dc04.pool.mediaWays.net] has quit [Ping timeout: 245 seconds] 10:53:44 benkard [~benkard@mnch-4d04dc04.pool.mediaWays.net] has joined #sbcl 10:59:50 managed to simplify that pvs thing 10:59:57 and it's really weird 11:07:51 looks like it's due to my nconc type derivation 11:11:28 and a similar thing was reported, i could have caught this thing too 11:14:22 and people really do wait until the releases to test things 11:33:52 chu [~user@unaffiliated/chu] has joined #sbcl 11:36:37 -!- benkard [~benkard@mnch-4d04dc04.pool.mediaWays.net] has quit [Ping timeout: 248 seconds] 11:41:06 benkard [~benkard@mnch-4d04dc04.pool.mediaWays.net] has joined #sbcl 11:46:36 -!- benkard [~benkard@mnch-4d04dc04.pool.mediaWays.net] has quit [Quit: Textual IRC Client: www.textualapp.com] 11:47:42 benkard [~benkard@mnch-4d04dc04.pool.mediaWays.net] has joined #sbcl 11:50:06 -!- echo-area [~user@182.92.247.2] has quit [Remote host closed the connection] 11:56:46 -!- benkard [~benkard@mnch-4d04dc04.pool.mediaWays.net] has quit [Ping timeout: 246 seconds] 12:05:27 benkard [~benkard@mnch-4d04dc04.pool.mediaWays.net] has joined #sbcl 12:06:07 drmeister [~drmeister@pool-71-185-168-200.phlapa.fios.verizon.net] has joined #sbcl 12:11:49 stassats: is non-consing sleep expected to work on x86 now? 12:12:59 LiamH [~none@pool-173-48-164-128.bstnma.fios.verizon.net] has joined #sbcl 12:15:29 yes 12:15:52 i think i forgot to pass --arch=x86 the previous time i was testing it 12:16:08 this time, i made sure i didn't 12:16:41 hm, looks like this nconc type derivation is still broken 12:16:42 damn 12:18:54 stassats: it seems to have failed on x86 here: https://ci.cor-lab.org/job/sbcl-master/label=ubuntu_quantal_32bit/lastCompletedBuild/testReport/, but i will do another build to make sure the most recent version is used 12:20:03 stassats: Did you read pkhuong's take on the PVS/nconc thing this morning? 12:20:13 no 12:21:48 segv- [~mb@95-91-241-60-dynip.superkabel.de] has joined #sbcl 12:22:27 well, i fixed the PVS problem, but introduce some another, and i don't have a test-case 12:23:07 time for think-hard-debugging 12:34:31 GOTCHA 12:34:36 (sb-kernel:specifier-type (sb-kernel:%simple-fun-type (funcall (defun foo () (flet ((bar (y) (declare (type float y)) (expt 2 y))) #'bar))))) 12:48:02 davazp [~user@195.Red-88-5-174.dynamicIP.rima-tde.net] has joined #sbcl 12:59:43 -!- LiamH [~none@pool-173-48-164-128.bstnma.fios.verizon.net] has quit [Quit: Leaving.] 13:02:49 echo-area [~user@114.254.101.235] has joined #sbcl 13:03:05 -!- davazp [~user@195.Red-88-5-174.dynamicIP.rima-tde.net] has quit [Read error: Connection reset by peer] 13:03:20 davazp [~user@195.Red-88-5-174.dynamicIP.rima-tde.net] has joined #sbcl 13:03:40 Krystof: Is it just that fixup-interval-expt needs to call bound-func in some more cases? 13:03:57 yes, or coerce-numeric-bound 13:04:08 but the difficult bit was finding an observable test case 13:04:11 stassats` [~stassats@wikipedia/stassats] has joined #sbcl 13:04:19 -!- stassats [~stassats@wikipedia/stassats] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 13:04:48 Aha 13:05:07 (or at least an observable test case that didn't involve recompiling cl-nlp and all its dependencies) 13:07:00 If its so difficult to observe, does that mean that some other parts of the compiler are too forgiving with respect to bad type specifiers? 13:11:43 yes, I think the numeric type constructor could be more defensive 13:12:32 it used not to be because we were less disciplined about splitting e.g. REAL types into (OR RATIONAL SINGLE-FLOAT DOUBLE-FLOAT) 13:29:00 nyef [~nyef@c-50-157-244-41.hsd1.ma.comcast.net] has joined #sbcl 13:29:08 G'morning all. 13:37:54 ha. Better recipe for reproduction: 13:37:57 (progn (defmethod foo ((x list) &optional (y 0.0)) (declare (type float y)) (expt 2 y)) (defmethod foo ((x list) &optional (y 0.0)) (declare (type float y)) (expt 2 y))) 13:40:09 -!- rudi [~rudi@1x-193-157-195-224.uio.no] has quit [Quit: Client exciting.] 13:53:46 ok, the problem is with MEMBER types 13:56:48 how do i decide that '(member 1 2 3 (a b c)) is suitable for LIST? 13:59:34 do you mean "the value is not definitely disjoint from LIST"? 14:00:00 if so, check that the type intersection is not *empty-type* 14:00:33 yes 14:02:44 that seems to work well 14:04:01 maybe wait for pkhuong to wake up? He spent some time thinking about it last night 14:04:14 well, i already broke it 14:05:18 heh, true 14:09:50 -!- chu [~user@unaffiliated/chu] has left #sbcl 14:10:21 types are hard 14:10:55 yes! 14:13:25 the only consolation that append type derivation helped to catch one bug in sbcl already 14:14:01 (the misplaced ` around values in a transform one) 14:14:29 (i think) 14:18:40 *stassats`* waits for haswell to hit the stores, so that i can decrease sbcl compilation and testing times 14:24:08 hlavaty` [~user@friedrichstrasse.knowledgetools.de] has joined #sbcl 14:26:28 -!- hlavaty [~user@friedrichstrasse.knowledgetools.de] has quit [Ping timeout: 276 seconds] 14:32:19 -!- davazp [~user@195.Red-88-5-174.dynamicIP.rima-tde.net] has quit [Ping timeout: 276 seconds] 14:33:33 -!- scymtym [~user@2001:638:504:2093:226:b9ff:fe7d:3e1f] has quit [Remote host closed the connection] 14:34:01 scymtym [~user@2001:638:504:2093:226:b9ff:fe7d:3e1f] has joined #sbcl 14:37:06 -!- Sagane_ [~Sagane@177.100-226-89.dsl.completel.net] has quit [Quit: Leaving] 14:46:37 wbooze [~wbooze@xdsl-87-79-248-11.netcologne.de] has joined #sbcl 14:51:08 stassats`: should be in stores already (not that I'd expect any real improvement over an ivy bridge) 14:51:15 davazp [~user@195.Red-88-5-174.dynamicIP.rima-tde.net] has joined #sbcl 14:51:22 i have sandy bridge 14:51:36 and i5 at that 14:52:32 meh I have i5 too and can totally live with it 14:52:58 it even executes !cold-init on arm under qemu really quickly 14:54:09 i remember "totally living" with p4 era celerons 14:54:21 (admittedly !cold-init mostly calls primitive %halt) 14:55:07 but even a 20% increases should be nice 14:55:17 especially when rebuilding 10 times 14:59:18 -!- benkard [~benkard@mnch-4d04dc04.pool.mediaWays.net] has quit [Ping timeout: 252 seconds] 15:00:53 20% should be plausible. unfortunately the only interesting desktop haswell part wasn't launched, and will not even be sold retail :-( 15:02:04 benkard [~benkard@mnch-4d04dc04.pool.mediaWays.net] has joined #sbcl 15:02:27 now i have two gc-impure tests failing on x86, BUG-936304 and BUG-981106 15:03:25 now, where was I before that redefining-methods-with-expt bug hit? Oh yes, srctran.lisp 15:03:31 no 15:03:41 i saw it before 15:04:04 so, nconc can't cause that 15:13:30 Compile-time error: 15:13:30 Lock on package SB-INT violated when binding ARG-LIST as a local symbol-macro 15:13:31 while in package COMMON-LISP-USER. 15:13:35 huh. Nothing is ever simple 15:14:53 (running (sb-debug:arg 0) in the debugger in a traced :break t function) 15:19:03 foreignFunction [~niksaak@ip-4761.sunline.net.ua] has joined #sbcl 15:20:30 -!- Bike [~Glossina@71-222-56-127.ptld.qwest.net] has quit [Ping timeout: 264 seconds] 15:20:52 LiamH [~none@pool-173-48-164-128.bstnma.fios.verizon.net] has joined #sbcl 15:21:57 Bike [~Glossina@67-5-248-17.ptld.qwest.net] has joined #sbcl 15:38:47 I think emarsden's latest bug is (a) old (b) a result of not looking through CASTs 15:39:31 yes. I have a more principled cut to width that'll work. 15:39:34 (I think) 15:40:44 to generate reports like https://ci.cor-lab.org/job/sbcl-master/label=ubuntu_quantal_32bit/lastCompletedBuild/testReport/, i would like to slightly extend result collection and add junit-style reporting in test-util.lisp and run-tests.lisp 15:40:51 would such a change be acceptable? 15:41:16 like Status Code: 404? 15:41:33 that report doesn't appear especially useful 15:42:14 stassats`: sorry, let me check 15:44:18 stassats`: should work now (same URL) 15:44:37 it does 15:45:22 nice 15:45:35 would be nice if it showed which git revision that is 15:47:01 build number and general info is here https://ci.cor-lab.org/view/rosetta-trunk/job/sbcl-master/label=ubuntu_quantal_64bit/lastCompletedBuild/ 15:47:10 you have only 936304 failing, i also have 981106 gc-impure test failing 15:47:44 it would be nice if it not only ran sbcl tests, but like cl-test-grid 15:47:45 i only saw 936304 failing on x86, so far (on laptop and build server) 15:47:58 that's the next step :) 15:48:09 i saw too, but now it's two failures 15:48:10 I can imagine gc failures being intermittent / sensitive to addresses/build environment 15:49:00 i have (disable :sb-doc) 15:49:04 maybe that's why! 15:49:34 stassats`: i can try without sb-doc 15:49:46 i'm trying with 15:50:01 would you be OK with email notifications for build status changes? 15:50:02 and yes, building and testing with all combination of features is a must 15:50:32 well, "must" 15:50:58 just the default is a great start. All possible combinations might end up quite time consuming, but definitely nice to have too 15:51:19 Krystof: considering scymtym considers building the whole quicklisp and running all the tests there 15:52:06 yeah 15:52:10 maybe --fancy and unchanged default features as the potentially most commonly used configurations? 15:52:12 Krystof: doubly-weak ? 15:52:18 weak on both key and value 15:52:32 ah, okay. 15:52:38 if value is reaped the value in the table is set to nil; if key is reaped the whole entry vanishes 15:52:38 scymtym: some things like sb-unicode or sb-thread produce unexpected failures 15:52:48 and are not affected by --fancy 15:52:51 or unexpected successes :) 15:53:15 Krystof: yeah, is the talk about why they're bad going to be online in any form? 15:53:28 i think i can schedule about four configurations in parallel 15:53:30 don't know. Not "bad" but tricky to use 15:53:42 it took him three? four? tries to get it right 15:54:13 -!- benkard [~benkard@mnch-4d04dc04.pool.mediaWays.net] has quit [Ping timeout: 276 seconds] 15:54:44 pkhuong: ok, I'll let you have first go at latest cut-to-width iteration :) 15:54:49 how generous of me 15:55:46 Krystof: ahh, we have them and MPS has some special support for it due to us, hence my curiosity. 15:56:53 poppingtonic [~user@212.49.88.98] has joined #sbcl 15:57:56 benkard [~benkard@mnch-4d04dc04.pool.mediaWays.net] has joined #sbcl 15:59:01 how well do CPUs cope with all the non-instructions in error traps? 15:59:32 or do they see RET and stop? 16:03:42 -!- davazp [~user@195.Red-88-5-174.dynamicIP.rima-tde.net] has quit [Read error: Connection reset by peer] 16:03:46 Hrm... I know that for some CPUs (at least PPC and ARM, possibly others), we occasionally use "conditional" traps, where the trap instruction itself has sufficient context that the runtime can figure out what to do, but we could get even more clever and store the error codes out-of-line for normal traps with a bit of work. 16:03:48 depends. some want a full cache line between the last instruction and the first non-instruction. 16:05:16 current x86oids tend to be less aggressive, and are OK with the setup, I expect. 16:05:36 (Hint: Traps have known addresses relative to the start of the component, and while the vestigial trace-table stuff is broken, it could be resurrected and would be *PERFECT* for this kind of thing.) 16:20:30 -!- Bike [~Glossina@67-5-248-17.ptld.qwest.net] has quit [Ping timeout: 264 seconds] 16:22:08 Bike [~Glossina@174-25-32-249.ptld.qwest.net] has joined #sbcl 16:22:38 sb-bsd-sockets tests open a listen socket with a fixed port number :( 16:22:55 so much for testing configurations in parallel 16:26:27 that can be fixed 16:26:30 -!- poppingtonic [~user@212.49.88.98] has quit [Ping timeout: 264 seconds] 16:26:48 *stassats`* got confused with the order of optimization/bug-fix/enhancement in NEWS again 16:31:10 scymtym: oh yeah, I keep hitting that as well. 16:31:46 pkhuong: i was thinking about reading the port number from an environment variable 16:32:11 that way, the build server could take care of avoiding clashes 16:34:32 poppingtonic [~user@212.49.88.98] has joined #sbcl 16:38:38 i see "fixup vector 0xc1acad0 has a bad widetag: 38" from debug.impure.lisp on x86 (in https://ci.cor-lab.org/view/rosetta-trunk/job/sbcl-master/featureset=2,label=ubuntu_quantal_32bit/24/consoleFull) 16:38:51 is this actually plausible or due to bad emulation / bad ram? 16:41:04 ... Is that #x38 or #x26? 16:41:28 (Really shouldn't be #x38, as that's a fixnum.) 16:42:05 i just copied the message 16:42:35 context is 16:42:37 ::: Running :GF-DISPATCH-BACKTRACE 16:42:38 fatal error encountered in SBCL pid 18151(tid 3084678848): 16:42:38 fixup vector 0xc1acad0 has a bad widetag: 38 16:42:41 16:43:10 bad ram isn't plausible if the error is consistent 16:44:18 Hrm... Not liking that output from room.test.sh... 16:44:40 just to rule out that possibility: it should be possible to build and test multiple SBCL trees in parallel on a single machine, right? 16:45:30 yes. 16:45:31 well, there at least used to be a hilarious bug on os x 16:46:00 where multiple sbcl processes would somehow interfere with each other, eventually causing unreproducible errors 16:47:00 Where on earth does that message come from? It's not formatted as a gc invariant message... 16:47:02 I'm not convinced we should be using the past tense. X86 tests make my macbook panic. 16:47:20 but since your logs say ubuntu, I guess that doesn't apply 16:47:50 nyef: gencgc.c:1921 16:48:30 gencgc_apply_code_fixups? 16:49:19 the machine was probably heavily loaded at that point 16:49:37 scymtym: but that would imply heap corruption 16:51:05 Quadrescence [~quad@unaffiliated/quadrescence] has joined #sbcl 16:53:50 stassats`: i have no idea what went wrong 16:54:25 i can serialize the builds and check whether that makes the problem go away 17:00:53 what if we standardized the place where arg-count checking happens, then for known calls with known number of arguments we could jump ahead of that, and then don't need invalid-arg-count-error error trap, since the place can be always the same? 17:01:41 I'd rather have multiple XEPs, like the preliminary CMUCL hack. 17:03:34 -!- poppingtonic [~user@212.49.88.98] has quit [Ping timeout: 256 seconds] 17:06:56 many times XEP inserts a safepoint at the end which is followed by a safepoint in no-arg-parsing-ep 17:07:13 should arg parsing even have safepoints? 17:14:31 is it an issue? 17:14:55 define issue? 17:15:53 does it affect anyone, how? 17:16:41 it increases code size and includes unnecessary instructions 17:18:19 consequently, it affects people who care about such things 17:23:44 -!- LiamH [~none@pool-173-48-164-128.bstnma.fios.verizon.net] has quit [Quit: Leaving.] 17:34:23 attila_lendvai [~attila_le@92.47.184.117] has joined #sbcl 17:34:23 -!- attila_lendvai [~attila_le@92.47.184.117] has quit [Changing host] 17:34:23 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 17:43:00 LiamH [~none@pool-173-48-164-128.bstnma.fios.verizon.net] has joined #sbcl 17:51:40 poppingtonic [~user@212.49.88.98] has joined #sbcl 17:57:45 -!- LiamH [~none@pool-173-48-164-128.bstnma.fios.verizon.net] has quit [Quit: Leaving.] 18:09:47 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 18:12:21 -!- pchrist_ is now known as pchrist 18:12:53 -!- wbooze [~wbooze@xdsl-87-79-248-11.netcologne.de] has quit [Ping timeout: 248 seconds] 18:21:10 fisxoj [~fisxoj@c-24-12-190-29.hsd1.il.comcast.net] has joined #sbcl 18:59:44 -!- fisxoj [~fisxoj@c-24-12-190-29.hsd1.il.comcast.net] has quit [Ping timeout: 256 seconds] 19:00:03 -!- teggi [~teggi@113.173.4.86] has quit [Remote host closed the connection] 19:06:13 -!- benkard [~benkard@mnch-4d04dc04.pool.mediaWays.net] has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz] 19:11:07 -!- foreignFunction [~niksaak@ip-4761.sunline.net.ua] has quit [Ping timeout: 246 seconds] 19:14:42 foreignFunction [~niksaak@ip-4761.sunline.net.ua] has joined #sbcl 19:24:35 -!- leuler [~user@p548FC843.dip0.t-ipconnect.de] has quit [Quit: ERC Version 5.1.2 $Revision: 1.796.2.6 $ (IRC client for Emacs)] 19:27:56 poppingt` [~user@212.49.88.98] has joined #sbcl 19:28:04 -!- poppingtonic [~user@212.49.88.98] has quit [Ping timeout: 256 seconds] 19:29:05 -!- poppingt` is now known as poppintonic 19:29:05 -!- poppintonic is now known as poppingtonic 19:29:38 -!- poppingtonic [~user@212.49.88.98] has left #sbcl 19:31:45 -!- danlentz [~danlentz@c-68-37-70-235.hsd1.nj.comcast.net] has quit [Remote host closed the connection] 19:35:38 danlentz [~danlentz@c-68-37-70-235.hsd1.nj.comcast.net] has joined #sbcl 19:54:02 Vivitron [~Vivitron@pool-98-110-213-33.bstnma.fios.verizon.net] has joined #sbcl 19:56:58 -!- Bike [~Glossina@174-25-32-249.ptld.qwest.net] has quit [Ping timeout: 246 seconds] 19:59:38 Bike_ [~Glossina@174-25-32-249.ptld.qwest.net] has joined #sbcl 20:01:14 rpg [~rpg@216.243.156.16.real-time.com] has joined #sbcl 20:05:06 Hydan [~hydan@ip-89-103-110-5.net.upcbroadband.cz] has joined #sbcl 20:06:22 -!- Hydan [~hydan@ip-89-103-110-5.net.upcbroadband.cz] has quit [Remote host closed the connection] 20:06:57 Hydan [~hydan@ip-89-103-110-5.net.upcbroadband.cz] has joined #sbcl 20:18:08 -!- Bike_ is now known as Bike 20:19:18 -!- Hydan [~hydan@ip-89-103-110-5.net.upcbroadband.cz] has quit [Remote host closed the connection] 20:21:11 Hydan [~hydan@ip-89-103-110-5.net.upcbroadband.cz] has joined #sbcl 20:26:27 Sagane [~Sagane@177.100-226-89.dsl.completel.net] has joined #sbcl 20:31:21 milanj [~milanj@cable-94-189-145-213.dynamic.sbb.rs] has joined #sbcl 20:38:56 ASau` [~user@p5797F121.dip0.t-ipconnect.de] has joined #sbcl 20:42:49 -!- ASau [~user@p5797EE7C.dip0.t-ipconnect.de] has quit [Ping timeout: 276 seconds] 20:43:19 scymtym_ [~user@ip-5-147-122-209.unitymediagroup.de] has joined #sbcl 20:43:39 -!- foreignFunction [~niksaak@ip-4761.sunline.net.ua] has quit [Quit: Leaving.] 20:46:30 -!- ASau` is now known as ASau 20:57:46 sdemarre [~serge@109.134.163.182] has joined #sbcl 21:07:19 -!- nyef [~nyef@c-50-157-244-41.hsd1.ma.comcast.net] has quit [Quit: G'night all.] 21:28:29 -!- sdemarre [~serge@109.134.163.182] has quit [Ping timeout: 256 seconds] 21:32:58 fisxoj [~fisxoj@c-24-12-190-29.hsd1.il.comcast.net] has joined #sbcl 21:40:57 davazp [~user@213.37.200.68.dyn.user.ono.com] has joined #sbcl 21:50:34 davazp` [~user@213.37.200.68.dyn.user.ono.com] has joined #sbcl 21:50:52 -!- davazp [~user@213.37.200.68.dyn.user.ono.com] has quit [Ping timeout: 256 seconds] 21:56:32 -!- davazp` [~user@213.37.200.68.dyn.user.ono.com] has quit [Ping timeout: 256 seconds] 22:00:58 -!- prxq [~mommer@mnhm-590c0575.pool.mediaWays.net] has quit [Remote host closed the connection] 22:12:56 -!- leoc`` [~leoc.git@p5DDBA34A.dip0.t-ipconnect.de] has quit [Ping timeout: 260 seconds] 22:31:46 Bike_ [~Glossina@71-214-84-152.ptld.qwest.net] has joined #sbcl 22:34:51 -!- Bike [~Glossina@174-25-32-249.ptld.qwest.net] has quit [Ping timeout: 260 seconds] 22:36:31 -!- Bike_ [~Glossina@71-214-84-152.ptld.qwest.net] has quit [Ping timeout: 264 seconds] 22:36:45 -!- Sagane [~Sagane@177.100-226-89.dsl.completel.net] has quit [Quit: Leaving] 22:37:03 Bike [~Glossina@71-214-85-151.ptld.qwest.net] has joined #sbcl 22:44:56 -!- milanj [~milanj@cable-94-189-145-213.dynamic.sbb.rs] has quit [Quit: Leaving] 23:08:00 -!- stassats` [~stassats@wikipedia/stassats] has quit [Ping timeout: 245 seconds] 23:14:43 -!- Krystof [~user@81.174.155.115] has quit [Ping timeout: 246 seconds] 23:45:18 -!- Quadrescence [~quad@unaffiliated/quadrescence] has quit [Quit: This computer has gone to sleep]