00:03:55 Looks like we only need to pin progname and pty-name 00:06:27 I don't think that's needed either, the alien machinery should be pinning them automatically 00:10:50 I'm worried about signals in spawn... otoh, that's already an issue with non-gc signals. 00:16:22 homie [~homie@xdsl-84-44-209-251.netcologne.de] has joined #sbcl 00:17:38 doesn't without-gcing block all signals? 00:21:55 they're not masked at the unix level 00:23:32 ah, true. so the worry is syscalls being interrupted rather than signal handlers running during spawn? 00:24:57 right, and we'd now expose lots more failures to users 00:25:24 why don't you just move everything to the C runtime, switching to posix_spawn or something similar ? 00:25:31 -!- homie [~homie@xdsl-84-44-209-251.netcologne.de] has quit [Quit: none] 00:25:41 fe[nl]ix: that's already the case. 00:25:57 or, do you mean blocking signals in the C runtime? 00:26:55 yes 00:27:14 doesn't solve the deadlock. 00:31:37 homie [~homie@xdsl-84-44-209-251.netcologne.de] has joined #sbcl 00:32:20 -!- homie [~homie@xdsl-84-44-209-251.netcologne.de] has quit [Client Quit] 00:33:16 a way would be to call fork directly 00:33:57 easy on Linux, not sure about the rest 00:39:02 homie [~homie@xdsl-84-44-209-251.netcologne.de] has joined #sbcl 00:44:12 -!- homie [~homie@xdsl-84-44-209-251.netcologne.de] has quit [Quit: none] 00:46:07 homie [~homie@xdsl-84-44-209-251.netcologne.de] has joined #sbcl 00:54:27 -!- homie [~homie@xdsl-84-44-209-251.netcologne.de] has quit [Quit: none] 01:01:34 homie [~homie@xdsl-84-44-209-251.netcologne.de] has joined #sbcl 01:13:32 -!- wbooze [~wbooze@xdsl-84-44-209-251.netcologne.de] has quit [Quit: none] 01:21:06 wbooze [~wbooze@xdsl-84-44-209-251.netcologne.de] has joined #sbcl 01:24:00 -!- LiamH [~healy@pool-108-7-208-203.bstnma.fios.verizon.net] has quit [Quit: Leaving.] 01:53:56 LiamH [~healy@70-88-245-53-ma-nh-me-ne.hfc.comcastbusiness.net] has joined #sbcl 02:53:24 -!- kanru [~kanru@kanru-1-pt.tunnel.tserv15.lax1.ipv6.he.net] has quit [Ping timeout: 260 seconds] 02:58:39 kanru [~kanru@kanru-1-pt.tunnel.tserv15.lax1.ipv6.he.net] has joined #sbcl 03:12:42 -!- LiamH [~healy@70-88-245-53-ma-nh-me-ne.hfc.comcastbusiness.net] has quit [Quit: Leaving.] 04:46:30 Vivitron [~Vivitron@pool-98-110-213-33.bstnma.fios.verizon.net] has joined #sbcl 05:47:17 i want to subtract a &body from a &whole to get the lambda args, anyone can remind me the built-in function to do that? 05:54:02 sdemarre [~serge@91.176.154.9] has joined #sbcl 05:56:49 got answer, ldiff 06:52:32 angavrilov [~angavrilo@217.71.227.190] has joined #sbcl 07:06:02 esden [esden@repl.esden.net] has joined #sbcl 07:06:21 -!- esden [esden@repl.esden.net] has left #sbcl 07:12:36 -!- sdemarre [~serge@91.176.154.9] has quit [Ping timeout: 246 seconds] 07:34:09 edgar-rft [~GOD@HSI-KBW-078-043-123-191.hsi4.kabel-badenwuerttemberg.de] has joined #sbcl 08:22:00 nikodemus [~nikodemus@cs27127210.pp.htv.fi] has joined #sbcl 08:22:00 -!- ChanServ has set mode +o nikodemus 08:22:24 morning 08:43:24 gko [~user@114-34-168-13.HINET-IP.hinet.net] has joined #sbcl 09:26:09 morning 09:57:56 <|3b|> nikodemus (and/or Krystof): if lp#1050768 is specifically about the deftransform not handling unknown types, should there be another bug for whatever causes it to forget the array type in the test case? 10:03:15 the real bug is in the deftransform 10:03:31 the other issue is the overly conservative conservative type 10:03:38 but that's strictly not a bug 10:04:18 well, maybe a bug, but a much less important one 10:04:35 <|3b|> seems like it could be a performance regression, if it can be triggered in other situations 10:05:26 true 10:11:06 superjudge [~mjl@c83-250-198-227.bredband.comhem.se] has joined #sbcl 10:12:01 the reason node-conservative-type goes bonkers here is that the derived type is STRING, which a fairly complex type 10:28:15 <|3b|> is derived type of string different from declaring it to be a string? 10:37:24 yes and no 10:37:48 here is should not matter (i actually have it fixed in my tree already, just no tests yet) 10:38:19 but in general we need to be a lot more careful about derived types 10:38:33 because they may be based on assumptions that can change 10:38:44 clhs array-in-bounds-p 10:38:45 http://www.lispworks.com/reference/HyperSpec/Body/f_ar_in_.htm 10:39:14 does anyone except me think that (array-in-bounds-p "foo" 0 0) should return nil instead of signaling an error? 10:40:20 <|3b|> well, number of subscripts is supposed to match rank of array 10:40:47 where does it say that? 10:40:57 note also, exceptional situations: none 10:41:13 <|3b|> "subscripts---a list of integers of length equal to the rank of the array. " in arguments and values 10:41:14 oh, in arguments, i see 10:41:16 hm 10:42:04 clozure at least agrees with us there, so good 10:43:37 nikodemus: hey how're you doing? you're on fire lately! :) 10:44:44 is that a good thing? 10:46:31 of course 10:47:49 nah. just had to fix my earlier booboos, which then snowballed into a couple of unrelated fixes 10:48:10 and the stack allocation thing was part of a wish 10:48:47 of course, if you compare to last 4 months before that... yeah, then it's pretty good :) 10:49:28 but by that metric I'm on fire too! 10:49:38 *Krystof* feels the heat 10:49:39 :) 10:49:53 relative metrics are the best 10:50:25 <|3b|> nikodemus: looks like AREF doesn't specify an error for extra indices either, just the same constraint on subscripts 10:50:37 but lichtblau is truly on fire :) 10:50:49 just wondering, why we can't help nik work on sbcl full-time? 10:51:05 because i don't want to, anymore 10:51:09 also, because enough money to live comfortably is a lot of money 10:51:10 i see 10:51:13 it's more fun as a hobby 10:51:40 i see. what's the new thing you're working on? i bet it's a lot of fun 10:51:46 and, yeah, being beholden to users and their totally unreasonable demands is terrible! 10:52:00 if someone gave me a 5-year stipend to do "whatever i want on sbcl", that would be a different thing... 10:52:25 huangjs: http://www.zenrobotics.com/ 10:52:38 we're hiring, by the way! clever people needed :) 10:53:06 cool 10:53:40 on the negative side, it's clojure, java, and c++. on the positive side i get to stick lasers on robots 10:54:10 what do you think of clojure? comparing CL? 10:54:27 it's a lisp, fair enough 10:54:43 it's lisp1, which bothers me there even more than in scheme, but not entirely sure why 10:54:59 it has some nice features 10:55:25 using [] for bindings bothers me much less than i thought -- it's a non-issue, and even defensible 10:55:43 it's horribly inefficient from my perspective 10:56:01 hwo about meta programming? as neat as CL? 10:56:38 i'm used to writing lisp code that performs within a reasonable fraction of C/C++. clojure needs to work hard to be within a reasonable fraction of java 10:56:50 haven't done much of that, really 10:56:54 i see. 10:57:04 there are also lots of features i haven't had the chance to explore 10:57:16 no multiple values 10:57:32 &key is there, sort of, but it's not nearly as good 10:57:40 no multiple values... 10:57:59 hate the lazy sequences -- not in principle, but in implementation details 10:58:09 it's too easy to realize on by accident 10:58:30 it's too easy to accidentally make something lazy because you refactor things and remove the only thing that realized it 10:58:45 well, not implmentation details, but API details 10:59:29 one thing stopped me reading is conj I think, which is specified by performance reason... 10:59:41 don't like the way sequences are overly generic. it has nice parts, but you really need to take care if you care what the real type of your sequence is -- it's easy to end up holding something diffent 11:00:40 support for print-readably / print-escape type stuff is incomplete 11:01:01 (easy to serialize something in a manner that loses information) 11:01:34 no small objects -- but that's a java thing, really. a native implementation probably would have them 11:02:26 heap consing like mad, enough so that GC pauses are an issue even with fancy java gcs. (well, maybe not if you use a realtime java) 11:02:35 how do you like the concurrency part? 11:02:46 i like the java part of it :) 11:03:10 i like jvm, except for not having argument lists in backtraces, and no way to get live backtraces 11:03:53 clojure has made me want a fast cl on top of jvm 11:04:16 because jvm /is/ pretty cool 11:04:50 and it's nice to have a crapload of libraries from a reputable source, with known licences 11:04:55 i currently also use node, python and trying a little bit Go, Go is really nice in concurrency. you mentioned user-space threads, how complex will the impl be? 11:05:14 huangjs: i'm not working on them 11:05:20 i know 11:05:37 but you're the one who prbably knows most how complex it could be :) 11:06:31 so i don't know really. i'd guess in the same ballpark as porting sbcl's native threads to a new unixoid OS that needs /some/ special help -- or a bit easier 11:07:04 1-3 guru weeks, 1-3 months if you need to become a guru first 11:07:13 haha 11:12:52 -!- homie [~homie@xdsl-84-44-209-251.netcologne.de] has quit [Ping timeout: 248 seconds] 11:13:54 -!- wbooze [~wbooze@xdsl-84-44-209-251.netcologne.de] has quit [Ping timeout: 264 seconds] 11:23:58 -!- huangjs [~huangjs@69.84.244.131] has quit [Quit: Ex-Chat] 11:50:56 homie [~homie@xdsl-78-35-152-14.netcologne.de] has joined #sbcl 12:25:48 -!- Quadrescence [~quad@unaffiliated/quadrescence] has quit [Ping timeout: 240 seconds] 12:27:01 -!- nikodemus [~nikodemus@cs27127210.pp.htv.fi] has quit [Quit: Leaving] 12:38:06 Quadrescence [~quad@unaffiliated/quadrescence] has joined #sbcl 12:50:39 -!- superjudge [~mjl@c83-250-198-227.bredband.comhem.se] has quit [Quit: Lost terminal] 12:52:42 wbooze [~wbooze@xdsl-78-35-152-14.netcologne.de] has joined #sbcl 12:55:43 -!- wbooze [~wbooze@xdsl-78-35-152-14.netcologne.de] has quit [Client Quit] 12:58:54 wbooze [~wbooze@xdsl-78-35-152-14.netcologne.de] has joined #sbcl 13:02:55 -!- homie [~homie@xdsl-78-35-152-14.netcologne.de] has quit [Quit: none] 13:04:33 homie [~homie@xdsl-78-35-152-14.netcologne.de] has joined #sbcl 13:21:42 leuler [~user@p548FB643.dip.t-dialin.net] has joined #sbcl 13:59:57 superjudge [~mjl@c83-250-198-227.bredband.comhem.se] has joined #sbcl 14:16:29 nikodemus` [~nikodemus@cs27127210.pp.htv.fi] has joined #sbcl 14:16:30 -!- ChanServ has set mode +o nikodemus` 14:20:17 -!- homie [~homie@xdsl-78-35-152-14.netcologne.de] has quit [Quit: none] 14:22:23 homie [~homie@xdsl-78-35-152-14.netcologne.de] has joined #sbcl 14:24:01 -!- nikodemus` [~nikodemus@cs27127210.pp.htv.fi] has quit [Quit: This computer has gone to sleep] 15:00:03 -!- leuler [~user@p548FB643.dip.t-dialin.net] has quit [Quit: ERC Version 5.1.2 $Revision: 1.796.2.6 $ (IRC client for Emacs)] 15:16:20 -!- Quadrescence [~quad@unaffiliated/quadrescence] has quit [Ping timeout: 260 seconds] 15:33:24 LiamH [~healy@pool-108-7-208-203.bstnma.fios.verizon.net] has joined #sbcl 16:08:22 Quadrescence [~quad@unaffiliated/quadrescence] has joined #sbcl 16:23:08 -!- Vivitron [~Vivitron@pool-98-110-213-33.bstnma.fios.verizon.net] has quit [Remote host closed the connection] 16:37:25 -!- edgar-rft [~GOD@HSI-KBW-078-043-123-191.hsi4.kabel-badenwuerttemberg.de] has quit [Quit: perpetual darkness] 16:46:20 -!- gko [~user@114-34-168-13.HINET-IP.hinet.net] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 17:53:40 -!- LiamH [~healy@pool-108-7-208-203.bstnma.fios.verizon.net] has quit [Quit: Leaving.] 18:27:54 -!- H4ns [hans@netzhansa.com] has left #sbcl 19:29:16 LiamH [~healy@pool-108-7-208-203.bstnma.fios.verizon.net] has joined #sbcl 19:56:28 -!- LiamH [~healy@pool-108-7-208-203.bstnma.fios.verizon.net] has quit [Quit: Leaving.] 20:09:40 sdemarre [~serge@91.176.154.9] has joined #sbcl 20:34:00 -!- slyrus_ [~chatzilla@adsl-99-183-243-129.dsl.pltn13.sbcglobal.net] has quit [Ping timeout: 246 seconds] 20:34:43 slyrus_ [~chatzilla@adsl-99-183-243-129.dsl.pltn13.sbcglobal.net] has joined #sbcl 20:56:41 -!- angavrilov [~angavrilo@217.71.227.190] has quit [Read error: Operation timed out] 21:13:54 -!- sdemarre [~serge@91.176.154.9] has quit [Ping timeout: 246 seconds] 23:23:46 LiamH [~healy@pool-108-7-208-203.bstnma.fios.verizon.net] has joined #sbcl