00:17:07 -!- TimKack` [~user@c-2ec2972d-74736162.cust.telenor.se] has quit [Ping timeout: 252 seconds] 00:24:01 TimKack [~tkack@c-2ec2972d-74736162.cust.telenor.se] has joined #sbcl 00:28:23 -!- TimKack [~tkack@c-2ec2972d-74736162.cust.telenor.se] has quit [Client Quit] 00:52:12 -!- Posterdati [~tapioca@host223-230-dynamic.16-87-r.retail.telecomitalia.it] has quit [Remote host closed the connection] 00:57:41 Posterdati [~tapioca@host223-230-dynamic.16-87-r.retail.telecomitalia.it] has joined #sbcl 00:59:38 -!- kwmiebach_ [~kwmiebach@164-177-155-66.static.cloud-ips.co.uk] has quit [Ping timeout: 245 seconds] 01:00:57 kwmiebach [~kwmiebach@164-177-155-66.static.cloud-ips.co.uk] has joined #sbcl 01:07:01 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 01:53:18 gko [~gko@114-34-168-13.HINET-IP.hinet.net] has joined #sbcl 03:16:20 slyrus_ [~chatzilla@99-28-161-110.lightspeed.miamfl.sbcglobal.net] has joined #sbcl 03:30:33 -!- huangjs [~huangjs@190.8.100.83] has quit [Read error: Operation timed out] 03:34:14 echo-area [~user@123.120.227.68] has joined #sbcl 03:38:02 huangjs [~huangjs@190.8.100.83] has joined #sbcl 04:09:47 plutoid [~pluto@58.39.164.31] has joined #sbcl 04:48:01 -!- rpg [~rpg@216.243.156.16.real-time.com] has quit [Quit: rpg] 04:49:50 rpg [~rpg@216.243.156.16.real-time.com] has joined #sbcl 04:56:24 -!- huangjs [~huangjs@190.8.100.83] has quit [Ping timeout: 252 seconds] 04:57:04 hefner [~hefner@static-71-166-162-194.washdc.east.verizon.net] has joined #sbcl 05:04:38 -!- kwmiebach [~kwmiebach@164-177-155-66.static.cloud-ips.co.uk] has quit [Ping timeout: 245 seconds] 05:06:21 kwmiebach [~kwmiebach@164-177-155-66.static.cloud-ips.co.uk] has joined #sbcl 05:28:29 -!- hefner [~hefner@static-71-166-162-194.washdc.east.verizon.net] has quit [Quit: hefner] 05:56:08 -!- slyrus_ [~chatzilla@99-28-161-110.lightspeed.miamfl.sbcglobal.net] has quit [Ping timeout: 240 seconds] 06:08:00 huangjs [~huangjs@190.8.100.83] has joined #sbcl 06:10:42 sdemarre [~serge@91.176.167.230] has joined #sbcl 06:31:59 edgar-rft [~user@HSI-KBW-078-043-123-191.hsi4.kabel-badenwuerttemberg.de] has joined #sbcl 06:47:22 angavrilov [~angavrilo@217.71.227.190] has joined #sbcl 06:51:10 nikodemus [~nikodemus@178-55-57-180.bb.dnainternet.fi] has joined #sbcl 06:51:11 -!- ChanServ has set mode +o nikodemus 08:08:50 -!- Phoodus [~foo@wsip-68-107-217-139.ph.ph.cox.net] has quit [Ping timeout: 260 seconds] 09:32:37 minion: paste #129411? 09:32:38 watch out, you'll make krystof angry 09:32:42 minion: paste 129411 09:32:44 you speak nonsense 09:33:04 argh. http://paste.lisp.org/display/129411 <- do these make sense? 09:52:54 -!- plutoid [~pluto@58.39.164.31] has quit [Quit: Leaving] 10:41:25 -!- homie` [~levgue@xdsl-78-35-161-187.netcologne.de] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 10:55:08 pkhuong: nyef had a lifetime visualizer somewhere 10:55:28 nikodemus: ah right. looked linux-only, though 11:15:04 nikodemus: impure tests are run in their own image, right? 11:22:46 yes 11:32:58 I randomly get failures in defstruct.impure 11:33:14 not repeatable, seems to only happen when running on 11 cores. 11:36:49 at least, I haven't seen it happen when pinned on a single core. 11:37:48 and I haven't yet managed to trigger it by running the test by itself. 11:38:00 actually, I just deleted my log, but i think it's package, not defstruct. 11:39:49 homie [~levgue@xdsl-78-35-128-28.netcologne.de] has joined #sbcl 12:11:38 I like the idea of a test failure which only happens on 11 cores 12:13:43 as opposed to some other prime number of cores ? 12:34:04 could someone run this to check that it fails in HEAD? http://paste.lisp.org/display/129412 12:35:08 Kryztof: fails 12:35:33 thank you 12:37:48 by the way, i modified the original test case and got a different error, (defun foo (x) (declare (optimize (speed 0) (space 0)) (type (integer -228645653448155482 -228645653447928749) x)) (floor 1.0 x)) then (foo -228645653448155482) signals invalid number of arguments: 1 12:38:37 stassats: further corroborates Kryztof's hypothesis. 12:39:31 and when -228645653448151381 is passed, it's (FOO #) in the backtrace 12:44:13 fix coming. 12:44:49 here's a quick hint, by the way: the source of the error is obvious if you set sb-c::*check-consistency* to t before compiling the function 12:45:48 how do you feel about eliminating open FP ranges in the compiler as well? 12:46:01 that's more-or-less what I've done 12:46:07 k. 12:46:16 you can take a look at what I've done and then do it better :-) 12:46:21 pushed 12:49:44 the major question in my mind about open fp ranges is whether there is use lower-bound-exclusive-zero and upper-bound-exclusive-one 12:51:21 the minor question which I don't have time to attack this afternoon is how the interval code got so messy that it had multiple interval type conversion paths. The first one I fixed turned out not to be called by a higher-level one which also handled open intervals 12:54:07 um, "turned out to be called" 12:54:10 anyway, it's a mess 12:54:17 Kryztof: I was thinking of doing that in make-numeric-type. 12:54:22 kind of like we do with integers 12:55:07 you mean by canonizing (foo) to ? 12:55:14 yeah. 12:55:28 where delta = machine epsilon. 12:55:39 only for #-sb-xc-host, obviously. 12:56:37 I think that's fine, but I think I completely misunderstood your question 12:57:05 I think that what I've done has mostly eliminated open FP ranges in the compiler. I think you're proposing to remove them from, I dunno, the type system :) 12:57:11 yes :) 12:57:35 stassats` [~stassats@wikipedia/stassats] has joined #sbcl 12:57:37 I don't have any objection, and in fact it moves us incrementally closer to a portable CL implementation of IEEE754, so that we can finally shake off the dependency on the host floating point implementation being sane 12:58:08 anyway. I think I need to mark 39 variously terrible courseworks this afternoon :-( 12:58:54 our derivation stuff is probably still all sorts of wrong wrt rounding mode, though. 12:59:03 (or flush to zero) oh well 13:00:10 "incrementally" :-) 13:03:10 we should really check for nans 13:04:02 does compiling sbcl includes many fp operations? 13:04:15 stassats: no. 13:04:46 so, why does it matter that the host implementation has proper ieee754? 13:05:07 it messes up the results of type derivation. 13:05:31 hefner [~hefner@static-71-166-162-194.washdc.east.verizon.net] has joined #sbcl 13:15:11 take a grep for cross-float-infinity-kludge, for make-unportable-float, and for various other existing cross-compiling float contortions 13:15:50 consider that lispworks has historically had short-float == IEEE single and single-, double-, long- be IEEE double (I don't know if that's still the case) 13:15:57 clisp doesn't support denormals 13:43:57 *Kryztof* cries. Did I say "variously terrible"? 13:44:45 are they all terrible? 13:50:44 yes 13:53:48 a sum total of zero submissions did the colour space manipulation correctly :-( 13:56:06 despite an explicit statement in the brief that it was unaceptable, and explicit feedback on this issue from the last submission, several of them have submitted docx files 13:58:09 attila_lendvai [~attila_le@188-143-57-222.pool.digikabel.hu] has joined #sbcl 13:58:25 -!- attila_lendvai [~attila_le@188-143-57-222.pool.digikabel.hu] has quit [Changing host] 13:58:25 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 14:03:59 We all know students don't actually listen to us when we open our mouths. 14:05:53 Kryztof: you know, coerce-for-bound is a broken interface... and everywhere it's used, it actually means coerce-interval. 14:06:10 -!- hefner [~hefner@static-71-166-162-194.washdc.east.verizon.net] has quit [Quit: hefner] 14:07:38 hefner [~hefner@static-71-166-162-194.washdc.east.verizon.net] has joined #sbcl 14:07:42 http://paste.lisp.org/display/129414 <- how about something like that? 14:08:38 slyrus_ [~chatzilla@99-28-161-110.lightspeed.miamfl.sbcglobal.net] has joined #sbcl 14:14:40 no objections. And yes, coerce-for-bound is broken 14:15:23 the most sad thing about the assignment is that I bet I get pushback on my grading from the department. ("You can't fail so many of them; it looks bad") 14:20:22 CL-USER> (sb-c::specifier-type '(float * (1.0))) 14:20:22 # 14:20:36 people might find this representation strange (: 14:23:30 oh well. Can't fix the interval thing without rethinking it all. 14:24:10 I can't wait for the next complaint we get from someone about the error message from typechecking of the argument to RANDOM, or the loop step-by variable 14:26:05 pkhuong: arguably the unparser for float types should check whether the printed representation is substantially shorter if it uses an exclusive bound 14:26:13 Kryztof: right (: 14:26:25 What's wrong with the "interval thing"? 14:26:56 -!- echo-area [~user@123.120.227.68] has quit [Remote host closed the connection] 14:28:11 Quadrescence: it works with half-open (mathematical) intervals, and tries to convert them to concrete types like integers or floats. 14:28:52 What does [0, 228645653448150900) mean, when converted to a float interval? 14:30:03 ok... so we might be able to sanely convert and widen the interval a bit. 14:30:03 I suppose the conservative approach is to take the least float greater than or equal to the upper bound 14:40:23 Quadrescence: that's not always going to work. 14:40:39 Why? 14:40:51 Sometimes, it means [0, 228645653448150900-1) 14:41:18 How does that not work with an interval which is a superset? 14:41:47 So you turn open ends into closed ones? Yes, that works... it's also a bit disappointing. 14:42:46 Well there's that, or you can represent the intervals exactly, and never let them be floats. Maybe rational approximations. 14:43:24 prxq [~mommer@mnhm-4d01024c.pool.mediaWays.net] has joined #sbcl 14:44:01 Quadrescence: unfortunately, at some point, we need floats in the real world. 14:44:11 pkhuong, For types? Why? 14:44:41 because the values are floats. 14:44:50 So, why do the endpoints have to be floats? 14:45:04 to be be able to sanely test types. 14:46:25 Hm, not sure how non-float endpoints impede the testing of types. We know the range represents a float, that's the base of the compound type. Can we not reliably perform cl:> and cl:= with rationals and floats? 14:46:49 a very simple way to ensure atrocious performance. 14:47:52 we'd also miss out on some range propagation if we don't round at each step, but that's another issue. 14:50:59 yeah, we could take the max value of (prev-float (float bound)) and (float (1- bound))... 14:53:10 but, most likely still broken when rounding modes are involved. 14:56:43 ah damn. Can't tell if a value is meant to be a rational or an integer. 14:58:51 yes 14:58:54 that is the main problem 14:59:05 things would be so much simpler with no language support for ratio 15:02:49 can't fix that before floor-rem-bound-aux 15:26:49 attila_lendvai1 [~attila_le@84-236-57-207.pool.digikabel.hu] has joined #sbcl 15:26:49 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Disconnected by services] 15:31:52 Kryztof: ok... so http://paste.lisp.org/+2RUU/1 also fixes the floor issue. 15:32:23 Does it seem right to you? 15:33:09 no, it's wrong (: 15:35:31 excellent, I shall not read it 15:35:55 whatever was wrong with it, write a test case quickly before you forget 15:36:14 -!- nikodemus [~nikodemus@178-55-57-180.bb.dnainternet.fi] has quit [Quit: This computer has gone to sleep] 15:37:29 plutoid [~pluto@58.39.164.31] has joined #sbcl 15:40:22 it's a trivial (nth-value 1 (floor 1.5 1)) /= 0 15:53:34 attila_lendvai [~attila_le@84-236-57-207.pool.digikabel.hu] has joined #sbcl 15:53:34 -!- attila_lendvai [~attila_le@84-236-57-207.pool.digikabel.hu] has quit [Changing host] 15:53:34 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 15:53:43 -!- attila_lendvai1 [~attila_le@84-236-57-207.pool.digikabel.hu] has quit [Quit: Leaving.] 16:01:36 -!- sdemarre [~serge@91.176.167.230] has quit [Ping timeout: 272 seconds] 16:02:36 -!- froydnj [nfroyd@people.mozilla.com] has quit [Quit: Lost terminal] 16:04:38 nikodemus [~nikodemus@178-55-57-180.bb.dnainternet.fi] has joined #sbcl 16:04:38 -!- ChanServ has set mode +o nikodemus 16:18:37 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Ping timeout: 265 seconds] 16:57:51 attila_lendvai [~attila_le@188-143-63-193.pool.digikabel.hu] has joined #sbcl 16:57:51 -!- attila_lendvai [~attila_le@188-143-63-193.pool.digikabel.hu] has quit [Changing host] 16:57:51 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 17:03:05 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Ping timeout: 265 seconds] 17:04:01 attila_lendvai [~attila_le@188-143-63-193.pool.digikabel.hu] has joined #sbcl 17:04:01 -!- attila_lendvai [~attila_le@188-143-63-193.pool.digikabel.hu] has quit [Changing host] 17:04:02 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 17:07:36 attila_lendvai1 [~attila_le@188-143-56-86.pool.digikabel.hu] has joined #sbcl 17:07:36 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Disconnected by services] 17:09:37 -!- plutoid [~pluto@58.39.164.31] has quit [Quit: Leaving] 17:26:17 -!- attila_lendvai1 [~attila_le@188-143-56-86.pool.digikabel.hu] has quit [Ping timeout: 245 seconds] 17:27:21 attila_lendvai [~attila_le@84-236-66-115.pool.digikabel.hu] has joined #sbcl 17:27:21 -!- attila_lendvai [~attila_le@84-236-66-115.pool.digikabel.hu] has quit [Changing host] 17:27:21 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 17:37:13 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Ping timeout: 272 seconds] 17:41:25 attila_lendvai [~attila_le@188-143-65-31.pool.digikabel.hu] has joined #sbcl 17:41:25 -!- attila_lendvai [~attila_le@188-143-65-31.pool.digikabel.hu] has quit [Changing host] 17:41:25 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 17:41:32 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Client Quit] 17:42:07 Phoodus [~foo@wsip-68-107-217-139.ph.ph.cox.net] has joined #sbcl 18:00:54 -!- rpg [~rpg@216.243.156.16.real-time.com] has quit [Quit: rpg] 18:09:43 tresta [~narog@s213-103-192-37.cust.tele2.se] has joined #sbcl 18:57:19 -!- nikodemus [~nikodemus@178-55-57-180.bb.dnainternet.fi] has quit [Quit: This computer has gone to sleep] 20:10:28 -!- minion [~minion@pppoe.178-66-22-152.dynamic.avangarddsl.ru] has quit [Disconnected by services] 20:10:30 -!- specbot [~specbot@pppoe.178-66-22-152.dynamic.avangarddsl.ru] has quit [Disconnected by services] 20:10:32 specbot [~specbot@pppoe.178-66-55-243.dynamic.avangarddsl.ru] has joined #sbcl 20:11:15 minion [~minion@pppoe.178-66-55-243.dynamic.avangarddsl.ru] has joined #sbcl 20:14:03 -!- stassats` [~stassats@wikipedia/stassats] has quit [Ping timeout: 244 seconds] 20:14:17 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 252 seconds] 20:19:22 -!- Phoodus [~foo@wsip-68-107-217-139.ph.ph.cox.net] has quit [Ping timeout: 272 seconds] 20:25:27 stassats [~stassats@wikipedia/stassats] has joined #sbcl 20:50:39 -!- angavrilov [~angavrilo@217.71.227.190] has quit [Ping timeout: 260 seconds] 21:03:12 sdemarre [~serge@91.176.167.230] has joined #sbcl 21:03:47 pkhuong: Going back a bit, since I see that you're active, I seem to recall doing a couple of versions of the VOP lifetime visualizer, one as an embryonic emacs mode with the intent to make editing the lifetimes easier. 21:06:18 I'll take a look. 21:06:56 http://www.lisphacker.com/temp/voplife.el 21:07:41 Note that voplife-test-function is supposed to grovel the parameters out of a VOP definition automatically, but I'm not sure how well it works. 21:07:57 probably won't work very well with a macro anyway. 21:08:31 Probably not, but there's another entry point, with usage examples at the bottom. 21:12:40 Jini [~pidgin@195.112.102.205] has joined #sbcl 21:29:43 -!- sdemarre [~serge@91.176.167.230] has quit [Quit: Leaving.] 21:30:20 -!- prxq [~mommer@mnhm-4d01024c.pool.mediaWays.net] has quit [Remote host closed the connection] 22:09:34 TimKack [~tkack@c-2ec2309d-74736162.cust.telenor.se] has joined #sbcl 23:26:31 blackwol` [~blackwolf@ool-4575fc51.dyn.optonline.net] has joined #sbcl 23:28:25 -!- blackwolf [~blackwolf@ool-4575fc51.dyn.optonline.net] has quit [Ping timeout: 260 seconds] 23:45:42 -!- Jini [~pidgin@195.112.102.205] has quit [Quit: Leaving.]