00:09:51 FareWell [~Fare@173-9-65-97-NewEngland.hfc.comcastbusiness.net] has joined #sbcl 00:25:43 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 00:30:21 KDr2 [~kdr2@125.33.181.192] has joined #sbcl 02:09:12 -!- specbot [~specbot@tiger.common-lisp.net] has quit [Remote host closed the connection] 02:09:16 specbot [~specbot@tiger.common-lisp.net] has joined #sbcl 02:10:36 echo-area [~user@182.92.247.2] has joined #sbcl 02:10:49 permanente [~profmakx@2a01:4f8:150:3181:2::20] has joined #sbcl 02:10:54 christop` [~user@oteiza.siccegge.de] has joined #sbcl 02:11:33 -!- christoph_debian [~user@oteiza.siccegge.de] has quit [Read error: No route to host] 02:12:12 -!- profmakx [~profmakx@unaffiliated/profmakx] has quit [Read error: Connection reset by peer] 02:20:34 -!- echo-area [~user@182.92.247.2] has quit [Remote host closed the connection] 02:24:28 echo-area [~user@182.92.247.2] has joined #sbcl 02:35:43 -!- FareWell [~Fare@173-9-65-97-NewEngland.hfc.comcastbusiness.net] has quit [Quit: Leaving] 03:33:47 stassats [~stassats@wikipedia/stassats] has joined #sbcl 04:04:06 -!- easye [~user@213.33.70.157] has quit [Remote host closed the connection] 05:13:50 slyrus [~chatzilla@99-28-161-110.lightspeed.miamfl.sbcglobal.net] has joined #sbcl 05:13:58 easye [~user@213.33.70.157] has joined #sbcl 05:25:27 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 244 seconds] 05:26:13 -!- Fare [fare@nat/google/x-kkgilackdhxopmxh] has quit [Ping timeout: 248 seconds] 05:30:59 nikodemus [~nikodemus@188-67-36-149.bb.dnainternet.fi] has joined #sbcl 05:31:00 -!- ChanServ has set mode +o nikodemus 05:56:47 o/ i'm done for this morning, in case someone is wondering :) 06:04:07 -!- nikodemus [~nikodemus@188-67-36-149.bb.dnainternet.fi] has quit [Quit: This computer has gone to sleep] 07:04:38 -!- Phoodus [~foo@ip72-223-116-248.ph.ph.cox.net] has quit [Ping timeout: 240 seconds] 07:07:19 -!- edgar-rft [~user@HSI-KBW-078-043-123-191.hsi4.kabel-badenwuerttemberg.de] has quit [Quit: Game over] 07:08:47 prxq [~mommer@mnhm-5f75e1e1.pool.mediaWays.net] has joined #sbcl 07:09:37 jdz [~jdz@193.206.22.97] has joined #sbcl 07:31:17 -!- Kryztof [~user@81.174.155.115] has quit [Ping timeout: 245 seconds] 09:22:07 attila_lendvai [~attila_le@178-164-243-202.pool.digikabel.hu] has joined #sbcl 09:22:07 -!- attila_lendvai [~attila_le@178-164-243-202.pool.digikabel.hu] has quit [Changing host] 09:22:07 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 09:23:32 -!- homie [~levgue@xdsl-78-35-131-83.netcologne.de] has quit [Ping timeout: 252 seconds] 09:28:10 -!- echo-area [~user@182.92.247.2] has quit [Ping timeout: 256 seconds] 09:33:54 homie [~levgue@xdsl-78-35-132-154.netcologne.de] has joined #sbcl 09:49:14 -!- KDr2 [~kdr2@125.33.181.192] has quit [Remote host closed the connection] 10:45:49 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 10:52:02 attila_lendvai [~attila_le@178-164-243-202.pool.digikabel.hu] has joined #sbcl 10:52:03 -!- attila_lendvai [~attila_le@178-164-243-202.pool.digikabel.hu] has quit [Changing host] 10:52:03 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 11:58:26 stassats [~stassats@pppoe.95-55-188-17.dynamic.avangarddsl.ru] has joined #sbcl 11:58:28 -!- stassats [~stassats@pppoe.95-55-188-17.dynamic.avangarddsl.ru] has quit [Changing host] 11:58:28 stassats [~stassats@wikipedia/stassats] has joined #sbcl 12:10:55 gko [~gko@114-34-168-13.HINET-IP.hinet.net] has joined #sbcl 13:12:54 LiamH [~none@pdp8.nrl.navy.mil] has joined #sbcl 13:21:08 Fare [fare@nat/google/x-xjlprvgdmezanpjs] has joined #sbcl 13:26:08 tcr1 [~tcr@host110-2.natpool.mwn.de] has joined #sbcl 13:29:51 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 13:31:38 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 246 seconds] 14:04:29 sheep_ [~peter@pjstirling.plus.com] has joined #sbcl 14:18:48 tyson1 [~Ian@74.13.218.50] has joined #sbcl 14:21:24 attila_lendvai [~attila_le@178-164-243-202.pool.digikabel.hu] has joined #sbcl 14:21:24 -!- attila_lendvai [~attila_le@178-164-243-202.pool.digikabel.hu] has quit [Changing host] 14:21:24 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 14:27:42 Okay, full backup of my last build made, time to do the smoke-test thing and try to build 1.0.28 (my build-host version). 14:28:31 ... which, of course, turns out to be older than the --xc-host= build option. 14:29:52 unhandled SIMPLE-ERROR in thread # {10029710F3}>: 14:29:52 FAILURE-P was set when creating "obj/from-xc/src/compiler/target-disassem.lisp-obj". 14:30:02 I can't compile 1.0.57.9 :-( 14:30:04 head is broken 14:30:40 leuler [~user@p54905287.dip.t-dialin.net] has joined #sbcl 14:31:15 I'll start a build and see if it's obvious, but I don't relaly have time for sbcling today 14:31:16 attila_lendvai1 [~attila_le@178-164-243-68.pool.digikabel.hu] has joined #sbcl 14:31:16 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Disconnected by services] 14:36:04 if you're bissecting-- 1.0.57.1 compiles fine. 14:36:41 edgar-rft [~user@HSI-KBW-078-043-123-191.hsi4.kabel-badenwuerttemberg.de] has joined #sbcl 14:37:58 target-disassem.lisp:1011. The last SETF's missing value form seems to be loc. 14:54:49 -!- tyson1 [~Ian@74.13.218.50] has left #sbcl 15:06:44 milanj [~milanj_@79-101-181-135.dynamic.isp.telekom.rs] has joined #sbcl 15:26:44 ASau [~user@93-80-192-112.broadband.corbina.ru] has joined #sbcl 15:28:51 -!- jdz [~jdz@193.206.22.97] has quit [Ping timeout: 252 seconds] 15:42:09 and another sb-concurrency test failure :( 15:42:53 -!- homie [~levgue@xdsl-78-35-132-154.netcologne.de] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 15:46:07 ... Fun and games, huh? Maybe I should have grabbed that dual-cpu G5 yesterday after all, just to have a multi-core PPC test machine to work with. 15:46:31 no doubt that'd help root out some nice bugs 15:47:10 homie [~levgue@xdsl-78-35-132-154.netcologne.de] has joined #sbcl 15:48:32 Yeah, but I don't have space right now to set it up, nor time to spend on what amounts to another project machine. 15:48:54 And there'd still be a good chance it'd take a thermal failure or a blown cap. 15:49:02 -!- leuler [~user@p54905287.dip.t-dialin.net] has quit [Quit: ERC Version 5.1.2 $Revision: 1.796.2.6 $ (IRC client for Emacs)] 15:50:14 *attila_lendvai1* arghs seeing Nikodemus' repeated changes to debug.lisp leading to endless rebasing and merging... 16:26:32 nikodemus [~nikodemus@188-67-36-149.bb.dnainternet.fi] has joined #sbcl 16:26:32 -!- ChanServ has set mode +o nikodemus 16:27:57 morning nikodemus! 16:31:21 nikodemus, thanks a lot. See above in the logs for bug in 1.0.57.7 that breaks the build and suggested fix 16:31:54 target-disassem.lisp:1011. The last SETF's missing value form seems to be loc. 16:32:27 fwiw, that builds for me. not sure if it's _right_, but it builds :) 16:32:42 sdemarre [~serge@91.176.153.88] has joined #sbcl 16:33:17 -!- nikodemus [~nikodemus@188-67-36-149.bb.dnainternet.fi] has quit [Quit: This computer has gone to sleep] 16:42:08 Fade [~fade@outrider.deepsky.com] has joined #sbcl 16:42:46 -!- sheep_ [~peter@pjstirling.plus.com] has quit [Quit: Ex-Chat] 16:43:22 I have a 32bit debian squeeze system that's showing an invalid exit status from packages.impure.lisp in the test suite on 1.0.57 16:43:34 test failed, expected 104 return code, got 1 16:48:24 ASau` [~user@95-28-124-123.broadband.corbina.ru] has joined #sbcl 16:50:12 nikodemus [~nikodemus@188-67-36-149.bb.dnainternet.fi] has joined #sbcl 16:50:12 -!- ChanServ has set mode +o nikodemus 16:51:38 -!- gko [~gko@114-34-168-13.HINET-IP.hinet.net] has quit [] 16:52:37 -!- ASau [~user@93-80-192-112.broadband.corbina.ru] has quit [Ping timeout: 248 seconds] 16:54:27 Fade: Anything more specific than that? 16:54:55 that's just what was reported at the end of the test run 16:55:11 the rest of it passed by pretty fast. is there any other info I could dig out? 16:55:51 You can run just that one test file, to see if it's repeatable, and to try and get data that won't scroll by too quickly. 16:56:05 Pass the filename as a parameter to run-tests.sh. 16:59:29 http://paste.lisp.org/+2RYX 17:04:55 -!- pchrist [~spirit@gentoo/developer/pchrist] has quit [Quit: leaving] 17:09:13 pchrist [~spirit@gentoo/developer/pchrist] has joined #sbcl 17:20:12 -!- nikodemus [~nikodemus@188-67-36-149.bb.dnainternet.fi] has quit [Quit: This computer has gone to sleep] 17:20:52 Ugh. Can't even build 1.0.28 on this box without it catching a SIGBUS. 17:23:12 Guess the next thing to do is find out if the build failure from yesterday is repeatable or not. 17:29:29 -!- slyrus [~chatzilla@99-28-161-110.lightspeed.miamfl.sbcglobal.net] has quit [Ping timeout: 252 seconds] 17:37:20 -!- sdemarre [~serge@91.176.153.88] has quit [Ping timeout: 246 seconds] 17:41:30 -!- tcr1 [~tcr@host110-2.natpool.mwn.de] has quit [Quit: Leaving.] 17:44:28 when I load one .so, it complains that it can't find the symbols that are defined in the next .so 17:44:31 that's annoying. 17:45:19 Fare: load the so before defining alien functions. 17:45:21 I can hush sb-kernel:undefined-alien-style-warning wholesale, but oh well. 17:45:46 pkhuong, I'm defining the alien functions when building the image, loading the .so at runtime. 17:45:53 ah. 17:46:20 handler-bind'ed it away. 17:47:20 how's france to you? 17:47:51 cold and rainy 17:48:40 yea, we've had undefined-alien-style-warning totally hushed for like 6 years. 17:49:00 cause otherwise you can't load a saved image without it spewing a crapload of shit at you. 18:01:40 I wonder why I'm only catching that now. 18:01:54 maybe it's the "upgrading from ancient 1.0.41" effect? 18:04:47 ... 1.0.41 is ancient? 18:05:01 *nyef* glances at the 1.0.28 he's using as a build host. 18:16:38 ITA had a patch in sbcl for a while just disabling it 18:16:52 we removed it because there is better warning control now. 18:17:39 ("now" is certainly a long time ago, but the local patch survived well past its introduction) 18:20:10 foom: oh, ok. 18:21:16 -!- Vivitron [~Vivitron@pool-98-110-213-33.bstnma.fios.verizon.net] has quit [Ping timeout: 276 seconds] 18:21:49 better warning control = you're actually using that xcvb-driver thing? oh, yup, you are. Sweet. 18:27:09 Fare: Any luck with the SIGCHLD thing, btw? 18:28:23 haven't tried it yet 18:28:29 hope to try it today or tomorrow 18:32:32 Phoodus [~foo@wsip-68-107-217-139.ph.ph.cox.net] has joined #sbcl 18:33:29 -!- homie [~levgue@xdsl-78-35-132-154.netcologne.de] has quit [Read error: Connection reset by peer] 18:34:08 homie [~levgue@xdsl-78-35-132-154.netcologne.de] has joined #sbcl 18:44:09 froydnj [~nfroyd@v-1025.fw1.scl3.mozilla.net] has joined #sbcl 18:50:26 nikodemus [~nikodemus@188-67-36-149.bb.dnainternet.fi] has joined #sbcl 18:50:26 -!- ChanServ has set mode +o nikodemus 19:26:58 sdemarre [~serge@91.176.153.88] has joined #sbcl 19:29:17 -!- Quadrescence [~quad@unaffiliated/quadrescence] has quit [Quit: Leaving] 19:40:35 stassats [~stassats@wikipedia/stassats] has joined #sbcl 19:53:48 Kryztof [~user@81.174.155.115] has joined #sbcl 19:53:49 -!- ChanServ has set mode +o Kryztof 20:04:25 Maybe I can get the SIMD thing in a decent shape for a contrib. 20:07:46 (v:/+ (v:= x vector)) is 16 times as fast as cl:count for a double vector, and slightly faster than a hand-written map nil loop (which generate surprisingly good code) 20:10:30 do we not inline cl:count? 20:11:10 we do nothing with it, it seems. 20:11:27 except for bit vectors, which is admittedly very nice 20:17:39 if we could get the right performance by just transforming to map nil, that'd be nice (even nicer if we just used that for the actual implementation, too) 20:19:24 lots of sequence functions would benefit from extra transforms 20:19:47 many of them could do fine with widetag dispatch -- no need to bloat all call sites 20:20:03 but YAY for the SIMD contrib :) 20:20:37 nikodemus: it's mostly C code though (: 20:21:15 re bloating all call sites, I really think we should move most of those to out of line functions or static funs 20:21:35 sure, the SSE-packs would be cooler -- but i'll take what i can get :) 20:21:36 logcount on x8oids is incredibly silly, for instance. 20:21:51 oh yeah. and I should definitely merge in the SORT patch 20:22:34 it's pretty good and almost always better than the current stuff. 20:22:35 i've started writing a generalize widetag-dispatch framework a couple of times, but never got it really nice 20:23:10 slyrus [~chatzilla@173-228-44-92.dsl.static.sonic.net] has joined #sbcl 20:23:43 oh, wow. we have a type-error in error 20:24:37 *froydnj* wonders if a "yo dawg" would be too lowbrow for that... 20:24:52 never 20:25:17 infinite-error-protector is to blame -- it's (unless (infinite-error-protect) ...) meaning ERROR looks like it can return NIL 20:27:12 Kryztof: how hard do you think it would be to have a vector of CLOS hashes? 20:27:31 is this more "yo dawg"? 20:27:46 (what do you mean? :-) 20:28:30 no, it's just that (in theory, at least), we'd have very solid hash functions for CLOS dispatch if we each wrapper had a random hash for each position in the arglist. 20:28:41 or at least for the first few positions. 20:28:54 so wrappers used to have 8 hash values each 20:29:00 I removed them ages ago 20:29:07 you could probably just revert that :) 20:29:11 heh :) 20:29:41 what was the reason for removing them? 20:29:57 they didn't add anything noticeable 20:29:57 Quadrescence [~quad@unaffiliated/quadrescence] has joined #sbcl 20:30:36 8fee0ba99cd1b1038072bd3fc8f5d5338d80d2de 20:33:25 in this case I think the use of the multiple hashes was to choose dynamically which hash was best for each dispatch table 20:33:36 but I did the stats and 99% of them were using the 0th hash 20:34:14 I guess the difference would only show up in rare patterns like generic arithmetic 20:34:56 otherwise, I expect there are different sets of class for each argument. 20:35:05 yes, could be 20:37:46 ... and sb-bsd-sockets just died with SIGBUS, same as last time. Looks like I've possibly got a repeatable build failure. (-: 20:41:57 Hrm. Is there anything special about SB-MD5.OTHER.INCREMENTAL 70? 20:53:58 (Yes, there is: It's the last full line of output before running the SB-MD5.MD5SUM-FILE.0.) 20:57:16 -!- sdemarre [~serge@91.176.153.88] has quit [Ping timeout: 276 seconds] 20:58:31 huangjs [~huangjs@190.8.100.83] has joined #sbcl 21:01:10 -!- nikodemus [~nikodemus@188-67-36-149.bb.dnainternet.fi] has quit [Ping timeout: 276 seconds] 21:05:51 nikodemus [~nikodemus@37-219-147-87.nat.bb.dnainternet.fi] has joined #sbcl 21:05:52 -!- ChanServ has set mode +o nikodemus 21:16:17 Got it! If the system hits an internal-error trap, it catches SIGBUS. 21:17:06 Now I just need to figure out WHY. 21:17:57 is that on MIPS? 21:18:02 Yup! 21:22:11 I've got a debian chroot on a 15 gig partition on a 500 gig disk (mostly unpartitioned space) attached to an E3000 running DD-WRT. 21:22:24 Takes about three and a half hours to fail an SBCL build. (-: 21:25:19 does qemu emulation not work ? 21:26:01 ... Never trust an emulator, always check against real hardware. 21:26:29 Besides, the E3000 is slow enough, running qemu on it would suck horribly. d-: 21:37:29 nikodemus: speaking of clojure-like features, how about support for value-less hash-tables a.k.a. sets? 21:40:05 theoretically, that could be done without sacrificing portability, say, allocate values vectors only when the first value is put, if value is NIL, don't create the value vector 21:40:14 and when there's no value vector, always return NIL 21:40:21 or T instead of NIL 21:40:23 slyrus: the difference in space usage would be tiny, actully. 21:41:18 really? 2x I would have thought... unless we already do what stassats suggests 21:41:39 we don't, last time i checked 21:42:04 and not 2x, the ht already has index vector, value vector and key vector 21:42:12 which are roughly of the same size 21:42:28 slyrus: for GC purposes, we store key-value pairs in a separate vector, and work with indices in that vector in the hash table itself. 21:43:34 not 2x, meaning 1.5 21:44:24 less than that with collisions, and the fact that the kv vector uses nearly exactly as much memory as needed for the data 21:44:56 while writes to the hash table are scattered, so that it uses the full ceiling-power-of-two words. 21:55:35 -!- LiamH [~none@pdp8.nrl.navy.mil] has quit [Quit: Leaving.] 22:15:01 -!- homie [~levgue@xdsl-78-35-132-154.netcologne.de] has quit [Ping timeout: 244 seconds] 22:18:34 -!- christop` is now known as christoph_debian 23:26:38 -!- slyrus [~chatzilla@173-228-44-92.dsl.static.sonic.net] has quit [Ping timeout: 240 seconds] 23:27:01 -!- nikodemus [~nikodemus@37-219-147-87.nat.bb.dnainternet.fi] has quit [Quit: This computer has gone to sleep] 23:36:02 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 252 seconds] 23:45:42 -!- prxq [~mommer@mnhm-5f75e1e1.pool.mediaWays.net] has quit [Remote host closed the connection]