00:00:25 (6) (slime,warning/warning) "BUILD.lisp":1:6 (pt=7). Caught error during fontification while searching for forms that are suppressed by reader-conditionals. The error was: (error "Unbalanced parentheses"). 00:01:16 tcr is hunting fontification bugs of all kind the whole week 00:01:32 should I send a bug report? 00:01:46 it certainly wouldn't hurt 00:01:57 wasn't there just a bugfix to that like, yesterday? 00:02:32 ztzg [n=dash@dslb-084-057-007-159.pools.arcor-ip.net] has joined #lisp 00:02:41 yeah, make sure you are on head and also provide file on which it fails 00:03:32 Also, *features*.. 00:06:18 -!- nvoorhies [n=nvoorhie@adsl-75-36-204-63.dsl.pltn13.sbcglobal.net] has quit [Read error: 104 (Connection reset by peer)] 00:07:09 nvoorhies [n=nvoorhie@adsl-75-36-204-63.dsl.pltn13.sbcglobal.net] has joined #lisp 00:07:15 prip [n=_prip@host134-123-dynamic.32-79-r.retail.telecomitalia.it] has joined #lisp 00:08:17 -!- Zenton` [n=user@80.29.234.11] has quit [Read error: 110 (Connection timed out)] 00:09:09 -!- elias` [n=me@unaffiliated/elias/x-342423] has quit [Success] 00:14:11 -!- prip [n=_prip@host134-123-dynamic.32-79-r.retail.telecomitalia.it] has quit [Remote closed the connection] 00:14:32 prip [n=_prip@host134-123-dynamic.32-79-r.retail.telecomitalia.it] has joined #lisp 00:16:11 -!- iaindalton is now known as iain|away 00:16:42 -!- dstatyvka [i=ejabberd@pepelaz.jabber.od.ua] has left #lisp 00:17:01 -!- kejsaren [n=kejsaren@111.25.95.91.static.ras.siw.siwnet.net] has quit [Read error: 110 (Connection timed out)] 00:20:25 ltriant [n=ltriant@202.136.38.162] has joined #lisp 00:22:00 Vegan [n=sdfpme@119.128.75.221] has joined #lisp 00:23:50 foom: to my bug? I've cvs up -dP'ed today 00:24:28 -!- iain|away is now known as iaindalton 00:25:23 *Fare* needs a better model for multiple-build dependencies. 00:25:48 I now have a good "low-level model" for what it means to build things. 00:26:09 I need a good model for what it means to describe systems. 00:26:19 fusss [n=chatzill@ip70-179-113-121.dc.dc.cox.net] has joined #lisp 00:26:27 heya 00:26:54 let's focus first on building a single system, and get back the functionality I had with the working prototype of XCVB 00:27:15 any one know if there is an standalone package for cmucl's REMOTE and WIRE services now removed from sbcl? 00:27:40 -!- dysinger [n=tim@166.129.81.115] has quit [Connection timed out] 00:27:51 never heard of them before. :) 00:28:23 rstandy [n=rastandy@net-93-144-106-44.t2.dsl.vodafone.it] has joined #lisp 00:28:43 -!- fusss [n=chatzill@ip70-179-113-121.dc.dc.cox.net] has quit [Read error: 104 (Connection reset by peer)] 00:29:23 fusss [n=chatzill@ip70-179-113-121.dc.dc.cox.net] has joined #lisp 00:29:51 i think it could be brought back as something on top of cl-prevalence + usocket 00:31:31 i'm assuming cl-prevalence is the most complete serialization package out there, unless someone has better recommendations 00:33:54 there are allot of nice, clean, high-level lisp APIs to be had in stuff like gbbopen, stuff from xerox and all the things in the AI repo, not to mention mining cmucl and other "adult" lisps for inspiration. 00:34:32 -!- The-Kenny [n=moritz@p5087FB50.dip.t-dialin.net] has quit [Read error: 110 (Connection timed out)] 00:34:53 or rucksack, elephant... 00:36:17 those are in my working set 00:39:21 iaindalton, I found this somewhere in my inbox, dated 2009-12-14: . 00:39:51 -!- projections [n=projecti@88.235.101.2] has left #lisp 00:40:00 davazp [n=user@56.Red-79-153-148.dynamicIP.rima-tde.net] has joined #lisp 00:42:59 Riastradh: that's it. Must have been a friendly time-traveler. 00:44:30 -!- Tordek [n=tordek@host132.190-227-38.telecom.net.ar] has quit [Read error: 60 (Operation timed out)] 00:46:16 Tordek [n=tordek@host182.190-137-199.telecom.net.ar] has joined #lisp 00:48:25 -!- cmm [n=cmm@bzq-82-81-167-159.red.bezeqint.net] has quit [Remote closed the connection] 00:53:26 -!- proq [n=user@38.100.211.40] has quit [Read error: 104 (Connection reset by peer)] 00:53:45 nyef [n=nyef@vcwl1-61.daktel.net] has joined #lisp 00:53:49 Hello all. 00:54:01 Modius [n=Modius@adsl-69-150-57-177.dsl.austtx.swbell.net] has joined #lisp 00:55:16 hello nyef 00:58:50 *nyef* sighs. 00:59:04 Why is it that sometimes one of my terminal windows just goes away? 00:59:18 I enter a command, and it isn't logout or exit, and it vanishes. :-/ 00:59:41 must have some Gnome in there 00:59:51 Yeah, might be MGT. 01:00:22 Hrm. Not MGT. 01:00:27 nyef: run screen... at least it makes it bearable (or write a decent CL shell, it may not solve your problem, but it is quite cool) 01:00:30 It's XFCE's "Terminal" program, whatever it is. 01:00:44 that's disappointing. I'd have expected them to use xterm. 01:01:06 -!- chessguy [n=chessguy@67-130-43-2.dia.static.qwest.net] has quit ["Leaving"] 01:01:15 Yeah, well, I'd have expected them to stick with Xt and a custom widget set, but they didn't do that either. 01:01:16 erronerously pressing C-d or a binding that does the equivalent in XFCE? 01:01:36 shows how much attention I pay. 01:03:01 madnificent: Seems unlikely, given that it only happens when I hit enter to execute a new command. 01:03:01 -!- iaindalton is now known as iain|away 01:03:16 And C-d takes effect immediately. 01:03:50 -!- rstandy [n=rastandy@net-93-144-106-44.t2.dsl.vodafone.it] has quit [Read error: 110 (Connection timed out)] 01:04:34 man-oh-man-oh-man does it help to use a heap to store the open nodes and a has table to store the closed nodes when doing an A* search...total running time went from about 3 seconds to .011 01:05:20 of course 01:07:39 (not terribly surprising, of course: it reduces a bunch of O(n) operations to O(1) operations...and since they're repeated rougly k^n times in the worst case, that's BAD) 01:08:28 it really is true: lisp makes it very easy to write correct-but-horribly inefficient code:-) 01:08:50 -!- fisxoj [n=fisxoj@149.43.252.13] has quit [Read error: 113 (No route to host)] 01:08:52 Fortunately, Lisp also tends to provide profilers to point out where such code is. 01:10:43 yup. anyway, back to figuring out how to make nice little dungeons for @ to do battle with r and X 01:10:49 Ooh. 01:10:50 eadmund: whereas C makes it very hard to write any code at all ? 01:11:37 it makes it very easy to get lovely segmentation faults, though! 01:11:42 and, on mac, bus error 01:11:49 eadmund, what do you think of dto's rlx? 01:11:57 (Linux rarely says that, for some reason) 01:11:59 or of langband, etc 01:12:24 rsynnott: I think it does on some platforms, but yes, it tends to SIGSEGV instead. 01:13:12 rsynnott, because X86 emulates unaligned access by default, whereas 68K and PPC don't, so Mac has evolved the "let's disabled unaligned access" meme? 01:13:42 hmm, now I think of it, I'm not sure if it even does it on Intel macos 01:13:47 The real question is, how cheap can you make an unaligned access fault on an x86oid? 01:14:31 Fare: rlx looks cool, but it's GUI...i dig console graphics. i might end up stealing some ideas (and maybe code if the lciense allows) from langband 01:14:34 (Thus enabling the unaligned-access fault version of a read barrier.) 01:14:50 -!- Samy [n=sbahra@c-76-21-209-249.hsd1.dc.comcast.net] has quit [Read error: 110 (Connection timed out)] 01:14:59 nyef: faulting will always be expensive on an x86oid 01:15:21 The question remains, how expensive? 01:15:21 unless and until they add some feature with "shadow registers", etc. 01:15:56 at the very least, you have to save a lot of registers to memory -- which is SLOW. 01:16:09 even if you read them back from a cache 01:16:42 I mean, the stack spill isn't that bad if you can handle the trap quickly and entirely in ring 0, as there's that funky register renaming for the top several stack slots. 01:18:24 Oh well, something for the LispOS people to figure out, maybe. 01:18:48 <_3b> nyef: speaking of 'profilers to point out where such code is', know anything about walking stacks through omit-frame-pointer code? 01:18:59 Yes, why? 01:19:15 <_3b> that's my current guess as to what is breaking things for me 01:19:23 Reeealy? 01:20:03 Well, it's a bit system-dependent, but... ISTR that modern libgcc_a or whatever it's called has some barely-documented API for it. 01:20:22 <_3b> yeah, my simple test case (which hopefully represents my real problem as well) calls sleep, sleep sets ebx to 0xffffffff, sb-sprof tries to dereference $rbp 01:20:27 libunwind, maybe. 01:20:33 It essentially uses the same eh_frame DWARF info registered for exception handling to walk backwards. 01:22:08 If you've got the inclination, there is a lot of work available in this area (DWARF eh_frame info, platform unwind support and lisp runtime interoperation). 01:22:08 <_3b> (calls C sleep through ffi that is, not SBCL's SLEEP) 01:23:39 *_3b* has already wasted too much time just getting linux set up again enough to run sbcl to profile this stuff, not sure how much i can put into fixing sbcl too :/ 01:23:44 not sure how to provide a default personality to use the EBP on CL code. 01:26:02 -!- fusss [n=chatzill@ip70-179-113-121.dc.dc.cox.net] has quit [Read error: 110 (Connection timed out)] 01:30:04 http://gist.github.com/112527 Hey all, I'm working on this snippet of elisp code. I'm trying to figure out if there's a way to get whatever swank:eval-and-grab-output does to stick (such as (in-package :foo) forms) 01:31:33 -!- nyef [n=nyef@vcwl1-61.daktel.net] has quit [Read error: 104 (Connection reset by peer)] 01:31:48 nyef_ [n=nyef@vcwl1-61.daktel.net] has joined #lisp 01:31:57 Stupid bloody hotel internet. 01:32:06 <_3b> hmm, libunwind can't be called from signal handlers 01:32:28 <_3b> or at least some functions can't 01:34:52 Anyway, what I was going to ask, had I not been so rudely disconnected, was if you could use the other profiler to get what you needed done done? 01:35:01 <_3b> sigprof-handler gets called in each thread being profiled, right? 01:35:36 <_3b> hmm, not sure if the other profiler would work or not (would probably be less convenient at minimum though) 01:35:58 _3b: what do you need? 01:36:22 <_3b> trying to profile a running GL app 01:36:40 <_3b> turn on profiling, start it doing something specific i want measured, wait a bit, turn it off again 01:36:52 -!- sykopomp [n=sykopomp@unaffiliated/sykopomp] has quit [Remote closed the connection] 01:37:03 I mean, what sort of profiling info? 01:37:28 <_3b> just basic idea of what is using the most time 01:37:46 sykopomp [n=user@unaffiliated/sykopomp] has joined #lisp 01:39:18 I believe you could use sb-sprof and only store IP. 01:39:51 <_3b> turn off the stack walking part you mean? 01:40:26 Right. And you'd be able to get at the names of foreign functions after the fact (i.e. outside the handler) 01:45:32 -!- eadmund [n=ruhl+irc@66.7.168.156] has quit ["ERC Version 5.2 (IRC client for Emacs)"] 01:50:41 borsman [n=quassel@76.177.217.216] has joined #lisp 01:51:31 -!- gigamonkey` [n=user@adsl-99-24-218-217.dsl.pltn13.sbcglobal.net] has quit [Remote closed the connection] 01:54:12 -!- saikat [n=saikat@adsl-99-135-67-134.dsl.pltn13.sbcglobal.net] has quit [] 02:01:44 -!- schoppenhauer [n=schoppen@unaffiliated/schoppenhauer] has quit [Read error: 110 (Connection timed out)] 02:05:57 -!- dreish [n=dreish@minus.dreish.org] has quit [] 02:07:48 -!- ltriant [n=ltriant@202.136.38.162] has quit ["Get MacIrssi - http://www.sysctl.co.uk/projects/macirssi/"] 02:16:19 -!- mrsolo [n=mrsolo@nat/yahoo/x-fd5ac85bcd030ffd] has quit [Read error: 110 (Connection timed out)] 02:19:22 -!- DarkRavin [n=thedarkr@adsl-147-121-66.bhm.bellsouth.net] has quit [Remote closed the connection] 02:21:25 -!- S11001001 [n=sirian@pdpc/supporter/active/S11001001] has quit [Remote closed the connection] 02:21:29 S11001001 [n=sirian@74-137-151-39.dhcp.insightbb.com] has joined #lisp 02:24:33 -!- MrEdward [i=Ed@AAubervilliers-154-1-28-37.w90-3.abo.wanadoo.fr] has quit ["« La POO c'est bien beau, mais en C au moins on va droit au but. »"] 02:24:57 -!- Odin- [n=sbkhh@s121-302.gardur.hi.is] has quit [] 02:25:42 -!- SandGorgon [n=user@122.162.135.206] has quit ["Leaving"] 02:26:02 SandGorgon [n=user@122.162.135.206] has joined #lisp 02:27:00 -!- S11001001 [n=sirian@pdpc/supporter/active/S11001001] has quit [Remote closed the connection] 02:27:08 S11001001 [n=sirian@74-137-151-39.dhcp.insightbb.com] has joined #lisp 02:35:31 <_3b> nyef_, pkhuong: for a bandaid fix, does it sound reasonable to reject any FP that isn't (sb-di:control-stack-valid-p) instead of just those below 4096? 02:35:42 -!- mjf [n=mjf@r3a98.net.upc.cz] has quit [Read error: 110 (Connection timed out)] 02:36:05 Seems reasonable to me, but I'm really not familiar with the statistical profiler. 02:36:46 if it's good enough for your current use case, that sounds like a decent band aid. 02:37:02 -!- Fare [n=Fare@ita4fw1.itasoftware.com] has quit ["Leaving"] 02:38:11 <_3b> so far i seem to mostly be getting memory faults at fairly low addresses, so it would probably catch it 02:38:13 JAS415 [n=jon@ip24-250-13-137.ri.ri.cox.net] has joined #lisp 02:38:44 -!- ikki [n=ikki@201.155.75.146] has quit ["Leaving"] 02:38:56 <_3b> is JES in the comments jsnell? 02:39:10 yes. 02:39:45 DarkRavin [n=thedarkr@adsl-147-121-66.bhm.bellsouth.net] has joined #lisp 02:43:59 semyon421 [n=semyon@217.67.122.44] has joined #lisp 02:45:22 -!- nvoorhies [n=nvoorhie@adsl-75-36-204-63.dsl.pltn13.sbcglobal.net] has quit [Read error: 104 (Connection reset by peer)] 02:46:11 nvoorhies [n=nvoorhie@adsl-75-36-204-63.dsl.pltn13.sbcglobal.net] has joined #lisp 02:46:59 *nyef_* sighs. 02:47:09 -!- S11001001 [n=sirian@74-137-151-39.dhcp.insightbb.com] has quit [Remote closed the connection] 02:47:22 I knew that the GLX implementation in CLX was incomplete, but I didn't expect it to be -this- incomplete. 02:48:04 <_3b> probably half the stuff it does support doesn't even exist in gl3.1 :p 02:48:16 The utter lack of glXRenderLarge support is a killer. 02:48:52 *_3b* suspects the lack of direct rendering would be annoying too, but has as yet been too lazy to actually test that theory :) 02:49:39 Umm... It might be, but I suspect that for gentle use it should hold up fairly well. 02:50:02 Right now I'm just annoyed that the obvious way to load a texture isn't implemented. 02:50:24 <_3b> yeah, that does sounds a bit incomplete :) 02:50:45 FIXMEs and just plain wierd comments all over it, too. 02:50:50 And no documentation to speak of. 02:51:02 <_3b> i understand cl-opengl works with CLX, if you aren't completely allergic to loading cffi :) 02:51:35 <_3b> (not that it is much better documented, but at least it theoretically still has active devs you can ask about it) 02:51:43 *nyef_* actually is somewhat allergic to cffi. 02:52:08 <_3b> actually, i guess if you wanted, it might not be too hard to port it to sb-alien 02:52:12 <_3b> (or whatever it is called) 02:52:26 pixpop [n=chatzill@adsl-76-232-55-30.dsl.lsan03.sbcglobal.net] has joined #lisp 02:52:27 <_3b> at least the low level part 02:52:42 does anyone have mcclim running nicely on linux? 02:52:44 -!- semyon421 [n=semyon@217.67.122.44] has quit [brown.freenode.net irc.freenode.net] 02:52:44 -!- galdor [n=galdor@bur91-2-82-231-160-213.fbx.proxad.net] has quit [brown.freenode.net irc.freenode.net] 02:52:44 -!- boyscared [n=bm3719@c-68-32-124-6.hsd1.va.comcast.net] has quit [brown.freenode.net irc.freenode.net] 02:52:44 -!- Jarvellis [n=jarv@dsl-217-155-101-22.zen.co.uk] has quit [brown.freenode.net irc.freenode.net] 02:52:44 -!- Guest51399 [n=user@72.14.228.89] has quit [brown.freenode.net irc.freenode.net] 02:52:44 <_3b> the higher level stuff probably does too much cffi stuff to be convenient 02:52:52 semyon421 [n=semyon@217.67.122.44] has joined #lisp 02:52:52 galdor [n=galdor@bur91-2-82-231-160-213.fbx.proxad.net] has joined #lisp 02:52:52 boyscared [n=bm3719@c-68-32-124-6.hsd1.va.comcast.net] has joined #lisp 02:52:52 Guest51399 [n=user@72.14.228.89] has joined #lisp 02:53:07 JAS415: yes. use clbuild. 02:53:12 i have it running but it isn't what i'd qualify as particularly usable 02:53:15 ah ok 02:53:26 i was using asdf i think 02:53:54 different things... asdf is a systems definition facility. clbuild is an installer for asdf systems. 02:54:06 <_3b> hmm, or maybe it would be more annoying than i thought, forgot about the types 02:54:12 And asdf-install is just broken anyway. 02:54:20 _3b: "types"? 02:54:28 right, i did an asdf-install on it from sbcl and fixed dependencies in what it downloaded 02:54:37 but i'll try clbuild 02:55:08 <_3b> nyef_: i think it defines cffi types for the GL types, instead of using the low level types directly 02:55:15 JAS415: don't use asdf-install... it's borken by design :) 02:56:00 Wow, someone actually used those! 02:57:16 Umm... Hrm. 02:57:43 I'm underwhelmed by the use, however. 02:58:18 <_3b> which parts? 02:58:33 *_3b* doesn't actually remember the details very well anymore 02:58:34 ensure-double 02:59:27 On the one side, translate-to-foreign. Cool. On the other side, WTF is this even necessary? 02:59:39 <_3b> where is that? 02:59:48 http://common-lisp.net/cgi-bin/darcsweb/darcsweb.cgi?r=cl-opengl-cl-opengl;a=headblob;f=/gl/types.lisp 02:59:52 -!- JAS415 [n=jon@ip24-250-13-137.ri.ri.cox.net] has quit [Read error: 104 (Connection reset by peer)] 03:00:03 -!- borsman [n=quassel@76.177.217.216] has quit [Connection timed out] 03:00:08 <_3b> ah, types.lisp ... that should have been obvious :p 03:00:51 *_3b* doesn't really know how that part works 03:00:54 Really, nothing here has called for any real alien-type trickery on the level of a good GUID type. 03:01:23 <_3b> yeah, the only complicated bit would have been doing the enum types right, and we gave up on that 03:01:41 -!- semyon421 [n=semyon@217.67.122.44] has quit ["leaving"] 03:04:48 *hefner* doesn't think clbuild will make mcclim any more "usable" 03:05:26 -!- athos [n=philipp@92.250.250.68] has quit ["leaving"] 03:06:04 Maybe not, but clbuild will likely obtain a more usable version of mcclim than the last release, whenever it was. 03:06:13 -!- pixpop [n=chatzill@adsl-76-232-55-30.dsl.lsan03.sbcglobal.net] has left #lisp 03:06:44 fusss [n=chatzill@ip70-179-113-121.dc.dc.cox.net] has joined #lisp 03:09:47 wow. MOVLPS: "... cannot be used for register to register move" because the reg-reg case has been punned into MOVHLPS, which can only be used for register to register. 03:10:35 ... What CPU is that? Sounds like some of the ARM instruction encoding games. 03:11:12 SSE. 03:11:18 Ah. 03:12:34 It actually makes sense, in a HW sort of way: both leave the upper half of the destination unchanged when it's a register. 03:14:08 S11001001 [n=sirian@74-137-151-39.dhcp.insightbb.com] has joined #lisp 03:14:11 <_3b> cool, looks like that test does reject a wide variety of unusual looking FP values when profiling my real code 03:14:46 _3b: you could also check that the addresses go up. 03:15:37 <_3b> i think the stack walking code already does that, there is only a problem if the initial value is bad 03:16:03 Okay, I clearly need to re-think things before getting any further into OpenGL. 03:16:17 ick 03:16:43 And it's getting late enough that I should be signing off for the evening. 03:16:50 G'night all. 03:16:55 -!- nyef_ [n=nyef@vcwl1-61.daktel.net] has quit ["Sleep beckons."] 03:18:12 mcclim has what, three dependencies? clbuild seems like overkill :) 03:21:12 -!- S11001001 [n=sirian@pdpc/supporter/active/S11001001] has quit [Remote closed the connection] 03:21:15 S11001001 [n=sirian@74-137-151-39.dhcp.insightbb.com] has joined #lisp 03:24:12 anyone feel like the BNF in the Syntax sections of the hyperspec are unreadable? BNF just don't play nice with s-exps, an eye sore imo. 03:25:49 it has a huge number of non-terminals that can't be all kept in one's head, so why even bother with pure BNF? show examples with italized bnf rules thrown in where necessary 03:28:25 people really love formalisms 03:29:56 same explanation for the formal semantics definition being in R6RS 03:30:13 no one likes formalisms .. until they're gone :-) how many times have you taken a "language" and wondered where the hell is the 4 page grammar? 03:31:11 I like prose, because it gives me something to read and the writer can use colorful metaphors and other structures 03:31:14 *can* use 03:31:27 some guy announced a new FP a few weeks ago, and the manual contents looked like this: 1) Containers, 2) Socket, 3) Regex, 4) DB, etc. stupid library documentation passing for language manual. 03:32:17 i think the succinct specs are for tool developers, not necessarily language users 03:32:48 the source is the language manual :) 03:34:02 which source? the part of it the that cold loads the part that loads itself and dumps a core which is warm loaded, or other parts of the language implementation? ;-) 03:34:59 well clearly whichever is used in the context of the code being parsed/evalled :P 03:36:34 gigamonkey` [n=user@adsl-99-35-218-254.dsl.pltn13.sbcglobal.net] has joined #lisp 03:37:51 i think I am undergoing through that past-restarts eureka when one thinks "I know! I know! I can use restarts to implement faux-continuations to handle user interaction on the web" .. must resist. 03:38:13 -!- drwhen [n=d@66-230-84-151-rb1.fai.dsl.dynamic.acsalaska.net] has quit [Read error: 110 (Connection timed out)] 03:38:37 drwhen [n=d@66-230-84-151-rb1.fai.dsl.dynamic.acsalaska.net] has joined #lisp 03:39:17 s/${GRAMMATICAL_ERRORS}/${CORRECTIONS}/g 03:40:14 -!- CrazyEddy [n=luteal@wrongplanet/CrazyEddy] has quit [Read error: 113 (No route to host)] 03:41:04 Tordek__ [n=tordek@host19.190-137-194.telecom.net.ar] has joined #lisp 03:41:46 -!- S11001001 [n=sirian@pdpc/supporter/active/S11001001] has quit [Remote closed the connection] 03:41:51 S11001001 [n=sirian@74-137-151-39.dhcp.insightbb.com] has joined #lisp 03:42:13 found it! 03:42:31 http://www.cs.cmu.edu/afs/cs/project/ai-repository/ai/lang/lisp/code/io/rpc/ 03:42:39 saikat [n=saikat@69.181.127.247] has joined #lisp 03:42:52 CrazyEddy [n=jestings@wrongplanet/CrazyEddy] has joined #lisp 03:42:58 fusss: if nothing else, people e have already written continuation libs for web stuff 03:43:11 *rsynnott* is a little sceptical of that approach 03:44:08 fusss: what do you want that for? 03:44:14 rsynnott: yeah, invoke-restart invites plenty of fascinating possibilities about clean-message passing and "reactive systems" .. then you find out you just need an epoll FFI :-) 03:45:33 rsynnott: long story, will share it later, but for now I have 15 minutes to get beer before the 12AM close-off 03:46:28 rsynnott: summary; something like cmucl's WIRE/REMOTE but for sbcl .. probably unneeded, given serialization + sockets, and xml-rpc goes even through firewalls, so it's probaby a moot exercise 03:46:43 but the Big Idea is distributed clos objects :-) 03:46:57 -!- S11001001 [n=sirian@74-137-151-39.dhcp.insightbb.com] has quit [Read error: 60 (Operation timed out)] 03:47:20 ah, I see 03:47:31 XML-rPC is terribly limited, though 03:47:34 remote GF invocation, instance migration and other mental wankery. was jelous of EJBs at first, the I really studied them further over a week, now I'm sympathetic ;-) 03:47:51 Facebook's thrift thing could be useful, as an intermediate step 03:48:14 base64'ed s-exp or some such .. thrifting? gotta jot that. brb later, hammered! ;-) 03:48:21 -!- fusss [n=chatzill@ip70-179-113-121.dc.dc.cox.net] has quit ["ChatZilla 0.9.84 [Firefox 3.0.10/2009042316]"] 03:50:14 -!- manuel_ [n=manuel@HSI-KBW-082-212-032-239.hsi.kabelbw.de] has quit [] 03:52:15 S11001001 [n=sirian@74-137-151-39.dhcp.insightbb.com] has joined #lisp 03:55:14 -!- Tordek [n=tordek@host182.190-137-199.telecom.net.ar] has quit [Read error: 110 (Connection timed out)] 03:55:18 jsoft [n=Administ@unaffiliated/jsoft] has joined #lisp 03:56:26 I hate formalisms, except as a scope-limited device to convince me that there really isn't much behind an idea after all. 04:01:20 -!- davazp [n=user@56.Red-79-153-148.dynamicIP.rima-tde.net] has quit [Remote closed the connection] 04:01:20 :) 04:07:01 -!- S11001001 [n=sirian@74-137-151-39.dhcp.insightbb.com] has quit [Read error: 60 (Operation timed out)] 04:10:48 envi^laptop [n=envi@115.94.144.228] has joined #lisp 04:12:25 chavo_ [n=user@c-66-41-11-10.hsd1.mn.comcast.net] has joined #lisp 04:18:25 WhiteFlame [i=foo@ip68-231-38-131.ph.ph.cox.net] has joined #lisp 04:20:11 xinming_ [n=hyy@125.109.74.165] has joined #lisp 04:20:35 -!- SandGorgon [n=user@122.162.135.206] has quit ["Leaving"] 04:20:52 -!- DarkRavin [n=thedarkr@adsl-147-121-66.bhm.bellsouth.net] has quit ["Ex-Chat"] 04:20:55 SandGorgon [n=user@122.162.135.206] has joined #lisp 04:21:18 -!- SandGorgon [n=user@122.162.135.206] has quit [Connection reset by peer] 04:24:23 -!- Phoodus [i=foo@ip68-231-38-131.ph.ph.cox.net] has quit [Read error: 110 (Connection timed out)] 04:25:53 SandGorgon [n=user@122.162.135.206] has joined #lisp 04:27:10 dysinger [n=tim@166.129.250.107] has joined #lisp 04:27:11 DarkRavin [n=thedarkr@adsl-147-121-66.bhm.bellsouth.net] has joined #lisp 04:29:15 S11001001 [n=sirian@74-137-151-39.dhcp.insightbb.com] has joined #lisp 04:29:39 jlf` [n=user@netblock-208-127-247-67.dslextreme.com] has joined #lisp 04:31:20 -!- S11001001 [n=sirian@74-137-151-39.dhcp.insightbb.com] has quit [Remote closed the connection] 04:31:45 S11001001 [n=sirian@74-137-151-39.dhcp.insightbb.com] has joined #lisp 04:33:05 -!- chavo_ [n=user@c-66-41-11-10.hsd1.mn.comcast.net] has quit [Remote closed the connection] 04:33:29 -!- S11001001 [n=sirian@pdpc/supporter/active/S11001001] has quit [Client Quit] 04:33:40 -!- Buganini [n=buganini@security-hole.info] has quit [Remote closed the connection] 04:36:01 -!- iain|away is now known as iaindalton 04:37:36 -!- xinming [n=hyy@125.109.254.12] has quit [Read error: 110 (Connection timed out)] 04:38:05 TekLok [n=TekLok@c-98-247-9-228.hsd1.wa.comcast.net] has joined #lisp 04:43:42 nego [n=nego@c-67-173-168-255.hsd1.il.comcast.net] has joined #lisp 04:44:18 -!- TekLok [n=TekLok@c-98-247-9-228.hsd1.wa.comcast.net] has quit ["Have you read your SICP today?"] 04:51:09 -!- nego [n=nego@c-67-173-168-255.hsd1.il.comcast.net] has quit ["Changing server"] 05:01:25 -!- jlf [n=user@unaffiliated/jlf] has quit [Read error: 104 (Connection reset by peer)] 05:04:46 jlf [n=user@unaffiliated/jlf] has joined #lisp 05:05:14 nego [n=nego@c-67-173-168-255.hsd1.il.comcast.net] has joined #lisp 05:06:13 I am new to lisp and I had a general question. What is the language best used for? I don't want to flame the channel, so if there are any links that answer the question that would be great. 05:06:31 <_3b> programming 05:06:46 -!- nvoorhies [n=nvoorhie@adsl-75-36-204-63.dsl.pltn13.sbcglobal.net] has quit [Read error: 104 (Connection reset by peer)] 05:09:16 *hefner* uses lisp for audiovisual experiments, data frobnification, and random puzzle solving 05:10:47 -!- Nshag [i=user@Mix-Orleans-106-3-178.w193-248.abo.wanadoo.fr] has quit ["Quitte"] 05:11:49 sexybaaron: there's a classic quote on the matter, google for "please don't assume lisp is only useful for" 05:11:52 fusss [n=chatzill@ip70-179-113-121.dc.dc.cox.net] has joined #lisp 05:13:09 -!- Vegan [n=sdfpme@119.128.75.221] has quit [Remote closed the connection] 05:15:03 fusss_ [n=chatzill@ip70-179-113-121.dc.dc.cox.net] has joined #lisp 05:15:04 -!- fusss [n=chatzill@ip70-179-113-121.dc.dc.cox.net] has quit [Read error: 104 (Connection reset by peer)] 05:15:06 -!- fusss_ is now known as fusss 05:16:58 if you say "a language" you're assuming they're all the same. Lisp is best compared to a "system", it's better a better system than Unix and Win32, imo :-) 05:20:10 Thanks, this is the kind of info I need to know. 05:20:13 -!- kidd2 [n=kidd@82.Red-79-150-113.dynamicIP.rima-tde.net] has quit [Remote closed the connection] 05:22:08 the best way to approach lisp is to get burned out delivery a fairly complex solution in any other tool of your "choice". once you hit that usual wall and acquire the necessary cynicism and hatred of commercial development, you will take to lisp like dyslexic kid to anagrams 05:22:23 s/delivery/delivering/ 05:23:56 after that, it's getting you off of lisp is like deporting a Haitian immigrant out of Brooklyn. you will always come back. 05:25:58 kidd2 [n=kidd@82.Red-79-150-113.dynamicIP.rima-tde.net] has joined #lisp 05:26:02 then you'll build some unmaintainable monstrosity using twenty libraries that you end up maintaining local forks of, with abysmal performance, and realizing you just can't win, you'll retire to a llama farm, clinging desperately to your last scraps of sanity. 05:26:21 -!- gigamonkey` [n=user@adsl-99-35-218-254.dsl.pltn13.sbcglobal.net] has quit [Read error: 110 (Connection timed out)] 05:28:21 deylen [n=deylen@94-194-104-130.zone8.bethere.co.uk] has joined #lisp 05:28:29 fuck that! sooner or later, you will realize all good software is just a thin macro layer over system calls. the self-hatred stops then. also the bleeding, if you were an emo who cut himself (better take that butter knife to Jersey and stab Kernighan, Thompson, Ritches and Susan Sarandon) 05:28:57 s/Ritches/Ritchey/ 05:29:33 this is getting meta, and vaguely biographical. 05:29:55 -!- Tordek__ is now known as Tordek 05:30:07 Quick question: Is there an easy way to generate an export symbols list for a (defpackage ) call? I want to allow access to most of the functions in the package. Am I going about this the wrong way? 05:31:04 do-external-symbols? but it includes even imported symbols. 05:31:34 fusss: that point is the point where you realize you need to write your own lisp, and ideally your own operating system 05:31:43 mrsolo [n=mrsolo@adsl-68-126-209-44.dsl.pltn13.pacbell.net] has joined #lisp 05:32:15 fusss, hmm that might work, do you know if lisp minds you exporting all the imported stuff? 05:32:37 fusss_ [n=chatzill@ip70-179-113-121.dc.dc.cox.net] has joined #lisp 05:32:54 <_3b> you can export anything you want... other programmers might object to you exporting things without actually thinking about it though :) 05:33:09 -!- fusss [n=chatzill@ip70-179-113-121.dc.dc.cox.net] has quit [Read error: 104 (Connection reset by peer)] 05:33:20 -!- fusss_ is now known as fusss 05:33:31 <_3b> but something like generating a list using code, then editing that by hand might be reasonable 05:34:04 _3b, hmm yeah that probably would work quite well. Ok thanks :) 05:34:06 -!- deylen [n=deylen@94-194-104-130.zone8.bethere.co.uk] has quit [] 05:34:12 <_3b> also, slime has hotkeys for toggling whether a symbol is exported, so you can go through your code and export them that way too 05:34:28 _3b: like, how? 05:34:58 <_3b> i think it is C-c x 05:35:31 i must type (do-external-symbols (sym (find-package FOO)) (print sym))) to figure out what a package FOO does, so often it's getting tiring. i really should read the docs first. 05:36:28 _3b: i would feel iffy about having such semantic side-effects carried out through editor keychords. best let me type the form and evaluate it manually. 05:47:03 Sun's EJB stuff is quite cool, imo. distributed reference passing; they pass around a "stub", a class "descriptor" that goes across the network cheaper than sending the actual instance. the standard call by value does depth-first copying of all the non-volatile slots of the instance and copies their instances if they're classes. then it serializes that and sends it across the network. 05:49:12 -!- jlf` [n=user@netblock-208-127-247-67.dslextreme.com] has quit [Read error: 110 (Connection timed out)] 05:49:16 the EJB reference passing stuff sends what, pretty much, amounts to class description, and instantiates that remotely. i still haven't found out what goes into that "stub", a piece of info that's cheaper to send that very high-level instance that might have huge dependency graphs (if the class members are themselves classes with other members, etc.) 05:50:08 and from what i read, distributed garbage collection boils down to just reference counting, with some decoration 05:51:28 this is the "Wilson survey" of distributed garbage collection, not pretty :-( www-sor.inria.fr/publi/SDGC_iwmm95.html 05:52:43 -!- jlf [n=user@unaffiliated/jlf] has quit [Read error: 110 (Connection timed out)] 05:54:09 *fusss* realizes none of that made sense to him on a second reread 05:54:46 ehu` [n=chatzill@82-170-33-173.ip.telfort.nl] has joined #lisp 05:55:42 -!- dysinger [n=tim@166.129.250.107] has quit [] 05:57:21 -!- mrsolo [n=mrsolo@adsl-68-126-209-44.dsl.pltn13.pacbell.net] has quit [Read error: 110 (Connection timed out)] 05:59:45 -!- anekos is now known as awayekos 06:00:05 nite 06:00:06 -!- fusss [n=chatzill@ip70-179-113-121.dc.dc.cox.net] has quit ["ChatZilla 0.9.84 [Firefox 3.0.10/2009042316]"] 06:03:24 cmm [n=cmm@bzq-82-81-167-159.red.bezeqint.net] has joined #lisp 06:04:25 -!- jmbr_ [n=jmbr@22.32.220.87.dynamic.jazztel.es] has quit [Read error: 110 (Connection timed out)] 06:04:42 bdowning_ [n=bdowning@mnementh.lavos.net] has joined #lisp 06:04:48 -!- bdowning [n=bdowning@mnementh.lavos.net] has quit [Read error: 60 (Operation timed out)] 06:05:40 -!- bdowning_ is now known as bdowning 06:09:20 -!- existentialmonk [n=carcdr@64-252-60-75.adsl.snet.net] has quit [Read error: 104 (Connection reset by peer)] 06:15:30 Good morning. 06:22:25 S11001001 [n=sirian@74-137-151-39.dhcp.insightbb.com] has joined #lisp 06:26:45 -!- iaindalton [n=iain@host-72-174-169-236.cdc-ut.client.bresnan.net] has left #lisp 06:33:29 Quadre` [n=quad@c-24-118-241-200.hsd1.mn.comcast.net] has joined #lisp 06:34:33 -!- Quadrescence [n=quad@unaffiliated/quadrescence] has quit [Nick collision from services.] 06:35:31 -!- Quadre` is now known as Quadrescence 06:36:22 manuel_ [n=manuel@HSI-KBW-082-212-032-239.hsi.kabelbw.de] has joined #lisp 06:52:21 -!- jsoft [n=Administ@unaffiliated/jsoft] has quit ["leaving"] 06:59:13 dv_ [n=dv@85-127-119-187.dynamic.xdsl-line.inode.at] has joined #lisp 07:04:22 Athas [n=athas@0x50a157d6.alb2nxx15.dynamic.dsl.tele.dk] has joined #lisp 07:05:28 jmbr [n=jmbr@92.32.220.87.dynamic.jazztel.es] has joined #lisp 07:10:13 Adamant [n=Adamant@c-76-29-188-22.hsd1.ga.comcast.net] has joined #lisp 07:15:22 sepult [n=sepult@xdsl-87-78-29-196.netcologne.de] has joined #lisp 07:20:59 ignas [n=ignas@ctv-79-132-160-221.vinita.lt] has joined #lisp 07:21:06 daniel_ [i=daniel@unaffiliated/daniel] has joined #lisp 07:22:31 mrsolo [n=mrsolo@adsl-68-126-209-44.dsl.pltn13.pacbell.net] has joined #lisp 07:23:39 legumbre_ [n=user@r190-135-29-63.dialup.adsl.anteldata.net.uy] has joined #lisp 07:24:37 -!- hjpark [n=user@116.40.135.22] has quit [Remote closed the connection] 07:26:34 nikodemus [n=nikodemu@cs181229041.pp.htv.fi] has joined #lisp 07:28:32 TekLok_ [n=TekLok@c-98-247-9-228.hsd1.wa.comcast.net] has joined #lisp 07:32:47 -!- legumbre [n=user@r190-135-38-164.dialup.adsl.anteldata.net.uy] has quit [Read error: 60 (Operation timed out)] 07:34:14 milanj [n=milan@79.101.169.94] has joined #lisp 07:36:50 -!- daniel [i=daniel@unaffiliated/daniel] has quit [Read error: 110 (Connection timed out)] 07:37:13 morning 07:38:11 Davse_Bamse [n=davse@4306ds4-soeb.0.fullrate.dk] has joined #lisp 07:38:25 quire pratice is more work than you think, my throat is still sore 07:39:14 anyhow, this is a Lisp channel, so I'll try to think Lispy thought's 07:39:30 jthing: Choir* 07:39:53 -!- ia__ [n=ia@89.169.189.230] has quit [Read error: 60 (Operation timed out)] 07:40:08 sorry, I think I sing better than I spell ) 07:41:10 It's Kor in Norski (Norwegian) 07:42:35 Any Lisp problems? 07:42:49 what do you mean? 07:42:59 are you here to fix my code? 07:43:11 in which case, I can totally tag out and -you- can be the one that spends all night on it :) 07:43:36 sykopomp: Any problems you wish to share? 07:45:32 gemelen [n=shelta@shpd-78-36-165-84.static.vologda.ru] has joined #lisp 07:45:40 As it were I do coding somewhat better than mind reading ;) 07:46:22 jthing: http://github.com/sykopomp/sheeple/blob/701900cfaea193e0520dbf01f9797376c8e089c8/TODO.org Pick something that looks interesting. Patches welcome. :P 07:47:28 inline caching is always oodles of fun for the whole family 07:47:28 sigh a Toad'o (TODO) 07:47:43 haha 07:47:57 can't keep sane without them. 07:48:04 not when I code, at least. 07:48:27 -!- manuel_ [n=manuel@HSI-KBW-082-212-032-239.hsi.kabelbw.de] has quit [] 07:50:40 Tankado [n=Woodruff@bzq-84-110-171-172.red.bezeqint.net] has joined #lisp 07:51:30 -!- bulibuta [n=bulibuta@unaffiliated/bulibuta] has quit [Read error: 110 (Connection timed out)] 07:52:17 Vegan [n=sdfpme@119.128.75.221] has joined #lisp 07:54:39 -!- beach [n=user@ABordeaux-158-1-27-178.w90-55.abo.wanadoo.fr] has quit [Read error: 104 (Connection reset by peer)] 07:55:55 ia__ [n=ia@89.169.189.230] has joined #lisp 07:56:26 -!- araujo [n=araujo@gentoo/developer/araujo] has quit ["Leaving"] 07:56:44 beach [n=user@ABordeaux-158-1-27-178.w90-55.abo.wanadoo.fr] has joined #lisp 07:57:44 sight, maybe 80% of yesterday's work reimplemented 07:57:50 sigh, even 07:58:00 ejs [n=eugen@185-163-124-91.pool.ukrtel.net] has joined #lisp 08:01:07 sounds like the typical Toad'o bit (TODO) 08:01:33 big pun to sykopomp 08:04:15 Anyhow I am happy, most of my code has actually worked.. 08:04:38 No toads in sight 08:11:56 nikodemus: as in: the end of reimplementation is in sight :)? 08:12:05 chris2 [n=chris@dslb-094-216-067-251.pools.arcor-ip.net] has joined #lisp 08:12:21 something like that 08:13:01 so are you angry or elated? 08:13:15 -!- Davse_Bamse [n=davse@4306ds4-soeb.0.fullrate.dk] has quit ["leaving"] 08:14:20 *jthing* finds the mood in the latter stage of a program to be crucial 08:15:39 Otherwise you get those #§! bugs 08:16:55 Take my advice and sit on it for a week. 08:18:15 -!- ejs [n=eugen@185-163-124-91.pool.ukrtel.net] has quit [Read error: 104 (Connection reset by peer)] 08:19:34 joachifm [n=joachim@bjo1-1x-dhcp388.studby.uio.no] has joined #lisp 08:21:36 -!- mrsolo [n=mrsolo@adsl-68-126-209-44.dsl.pltn13.pacbell.net] has quit [Read error: 110 (Connection timed out)] 08:21:58 existentialmonk [n=carcdr@64-252-60-75.adsl.snet.net] has joined #lisp 08:22:39 tcr [n=tcr@host145.natpool.mwn.de] has joined #lisp 08:28:13 -!- ignas [n=ignas@ctv-79-132-160-221.vinita.lt] has quit [Read error: 110 (Connection timed out)] 08:31:21 -!- slash_ [n=Unknown@whgeh0138.cip.uni-regensburg.de] has quit [Client Quit] 08:34:10 Jarvellis [n=jarv@dsl-217-155-101-22.zen.co.uk] has joined #lisp 08:34:59 Modius_ [n=Modius@99.179.100.6] has joined #lisp 08:40:14 kiuma_ [n=kiuma@93-36-1-24.ip57.fastwebnet.it] has joined #lisp 08:40:48 -!- yahooooo [n=yahooooo@c-76-104-183-185.hsd1.wa.comcast.net] has quit [Read error: 110 (Connection timed out)] 08:43:03 -!- kiuma_ is now known as kiuma 08:43:31 tombom [i=tombom@wikipedia/Tombomp] has joined #lisp 08:49:26 -!- Modius [n=Modius@adsl-69-150-57-177.dsl.austtx.swbell.net] has quit [Read error: 110 (Connection timed out)] 08:53:32 hkBst [n=hkBst@gentoo/developer/hkbst] has joined #lisp 08:55:51 -!- envi^laptop [n=envi@115.94.144.228] has quit ["Leaving"] 08:56:37 -!- kuwabara [n=kuwabara@cerbere.qosmos.eu] has quit [Read error: 110 (Connection timed out)] 08:59:30 Davidbrcz [n=david@nsc.ciup.fr] has joined #lisp 09:03:40 -!- chris2 [n=chris@dslb-094-216-067-251.pools.arcor-ip.net] has quit ["Leaving"] 09:04:13 yahooooo [n=yahooooo@c-76-104-183-185.hsd1.wa.comcast.net] has joined #lisp 09:10:41 delYsid [n=user@chello084115136207.3.graz.surfer.at] has joined #lisp 09:11:12 Hun [n=hun@pd956be5d.dip0.t-ipconnect.de] has joined #lisp 09:14:18 ...and now for the final missing part... 09:16:47 my bariton is failing? 09:16:51 lol 09:17:38 Just take that week.. 09:18:11 what are you working on, nikodemus ? .. the array-stuff? 09:18:21 mikezor_ [n=mikael@c-ede270d5.04-404-7570701.cust.bredbandsbolaget.se] has joined #lisp 09:26:42 yes, making things like (make-array (list 3) :element-type 'single-float :initial-contents (list x y z)) efficient and able to stack allocate 09:27:24 hooray, it works again. next, more test-cases 09:29:30 congrats nikodemus 09:32:27 Foofie [n=innocent@86.80-203-225.nextgentel.com] has joined #lisp 09:32:32 -!- Adamant [n=Adamant@unaffiliated/adamant] has quit [] 09:34:12 ruediger [n=the-rued@62-47-135-171.adsl.highway.telekom.at] has joined #lisp 09:34:16 -!- mikezor [n=mikael@c-ede270d5.04-404-7570701.cust.bredbandsbolaget.se] has quit [Read error: 110 (Connection timed out)] 09:38:50 rjack [n=rjack@adsl-ull-234-19.51-151.net24.it] has joined #lisp 09:39:00 nikodemus: What determines when stack allocation is possible? 09:40:17 -!- ruediger [n=the-rued@62-47-135-171.adsl.highway.telekom.at] has quit ["Leaving"] 09:41:33 -!- saikat [n=saikat@69.181.127.247] has quit [] 09:41:55 r1nu- [n=r1nu-@ppp-94-68-16-210.home.otenet.gr] has joined #lisp 09:41:59 morning! 09:42:11 hello r1nu- 09:43:05 is so good without univ today :D i slept a lot of hours :D 09:43:39 now you are ready to hack some Lisp 09:44:41 yes i have to finish a project ... i must finish it until 20 may. 09:45:19 r1nu-: A Lisp project? 09:45:44 beach yes... 09:46:00 Nice! What is it supposed to do? 09:46:13 ARI. 09:46:51 beach: manual dynamic-extent declarations initiate the dx analysis, which then reasons about "otherwise inaccessible" parts 09:47:19 nikodemus: OK. 09:47:29 but the final crunch is that various compiler transforms need to be there for things to stay (or move into) stack allocatable form 09:48:25 got the sack allocatable bit 09:48:28 stack 09:48:45 :initial-contents used to disable all MAKE-ARRAY transforms -- manually SETF'ing the elements was an order of magnitude faster than using (LIST X Y Z) as :INITIAL-CONTENTS 09:49:02 ahh 09:49:20 No more List crap 09:49:46 That was a real Wart 09:50:30 -!- Fufie [n=innocent@86.80-203-225.nextgentel.com] has quit [Read error: 110 (Connection timed out)] 09:51:13 nikodemus: and the same with (VECTOR X Y Z) ? 09:51:45 seelenquell [n=seelenqu@pD9E45779.dip.t-dialin.net] has joined #lisp 09:55:41 stassats: in :INITIAL-CONTENTS? yes 09:56:39 and now it's ok? 09:56:52 nikodemus: perhaps you should get some sleep? 09:57:21 rstandy [n=rastandy@net-93-144-199-191.t2.dsl.vodafone.it] has joined #lisp 09:57:34 no offence, but you have been working all night. 09:59:17 anyhow nikodemus your work sounds truly usefull for a change.Kudos to you! 10:00:39 attila_lendvai_ [n=ati@adsl-89-132-1-79.monradsl.monornet.hu] has joined #lisp 10:06:57 eno__ [n=eno@adsl-70-137-136-250.dsl.snfc21.sbcglobal.net] has joined #lisp 10:08:57 dstatyvka [i=ejabberd@pepelaz.jabber.od.ua] has joined #lisp 10:12:30 jthing: it's 1300 here, and i've had a good night's sleep 10:12:47 stassats: yes 10:12:59 or rather, will be soon 10:13:37 One of the infamous Toad'o (TODO)'s 10:14:30 Anyone heard of 'do it right the first time'? 10:16:52 Odin- [n=sbkhh@s121-302.gardur.hi.is] has joined #lisp 10:17:03 -!- eno [n=eno@nslu2-linux/eno] has quit [Read error: 110 (Connection timed out)] 10:17:07 *jthing* feels TODO's keep me pissed off.. 10:18:47 oh, it's implemented all right. just not committed yet 10:20:05 athos [n=philipp@92.250.250.68] has joined #lisp 10:20:33 no offence nikodemus, I am sure your work is great, just winding 10:21:00 Zenton` [n=user@80.29.224.221] has joined #lisp 10:21:23 The-Kenny [n=moritz@p5087CE4E.dip.t-dialin.net] has joined #lisp 10:21:37 It is spelled "offense". 10:21:58 so it is 10:22:23 gottesmm [n=gottesmm@c121h014.wless.reed.edu] has joined #lisp 10:23:08 sorry, offense 10:25:33 dwave [n=ask@084202072214.customer.alfanett.no] has joined #lisp 10:27:21 -!- Odin- [n=sbkhh@s121-302.gardur.hi.is] has quit [] 10:28:00 you where using Ubuntu, jthing ? XChat has a built-in spell checker.. it underlines typos in real-time, as you're typing .. (it's very nice for me, because it enables me to say stupid things -- but still seem clever .. lol) 10:28:25 Inostadl: I use Opera 10:28:32 drafael [n=tapio@ip-118-90-129-14.xdsl.xnet.co.nz] has joined #lisp 10:28:49 ok 10:29:34 ERC has flyspell mode. 10:30:04 if i take offence i stop being passive agressive -- you'll know when it happens :) 10:30:10 beach. I use flyspell in emacs 10:30:42 jthing: ERC is the IRC client for Emacs. 10:31:24 yes, perhaps I should use that 10:32:49 I mean it is clear I can't spell for shit! 10:33:05 :) 10:38:51 -!- ehu` [n=chatzill@82-170-33-173.ip.telfort.nl] has quit [Read error: 110 (Connection timed out)] 10:40:13 *p_l* started dumping core of his failed commonqt run from yesterday... it's going to be big xD 10:40:27 -!- drafael [n=tapio@ip-118-90-129-14.xdsl.xnet.co.nz] has quit ["Leaving."] 10:40:34 manuel_ [n=manuel@213.144.1.98] has joined #lisp 10:40:47 core? nasty! 10:41:44 Spyderco [n=nash@194.45.110.65] has joined #lisp 10:42:12 well, theoretically that code still runs and is stopped in select(), but I have a feeling SBCL's GC want's to commit seppuku seeing the breakage at address 0 ;-) 10:43:13 drafael [n=tapio@ip-118-90-129-14.xdsl.xnet.co.nz] has joined #lisp 10:43:56 address 0? that sounds omnious 10:44:18 -!- Tankado [n=Woodruff@bzq-84-110-171-172.red.bezeqint.net] has quit [] 10:45:04 Well p_i I guess you have you answer.. 10:45:07 -!- manuel_ [n=manuel@213.144.1.98] has quit [Client Quit] 10:45:42 Then of cource it helps to know the answer. 10:45:52 lol 10:52:20 your 10:54:36 *jthing* realizes he really can't spell for shit. 10:56:09 And thus Opera (at least in it's current form) is not the ideal choice 10:59:33 -!- vy [n=user@213.139.194.186] has quit [Remote closed the connection] 11:01:01 projections [n=projecti@88.235.101.2] has joined #lisp 11:02:46 -!- abeaumont [n=abeaumon@84.76.48.250] has quit [Read error: 110 (Connection timed out)] 11:06:58 hsaliak [n=hsaliak@bb116-15-29-118.singnet.com.sg] has joined #lisp 11:07:29 ThomasI [n=thomas@unaffiliated/thomasi] has joined #lisp 11:08:52 -!- Foofie is now known as Fufie 11:16:47 blandest [n=blandest@softhouse.is.ew.ro] has joined #lisp 11:18:03 -!- dwave [n=ask@084202072214.customer.alfanett.no] has quit ["Be back later"] 11:20:09 -!- Vegan [n=sdfpme@119.128.75.221] has quit [Remote closed the connection] 11:21:19 Odin- [n=sbkhh@s121-302.gardur.hi.is] has joined #lisp 11:21:40 bgs100 [n=ian@unaffiliated/bgs100] has joined #lisp 11:22:15 -!- blandest [n=blandest@softhouse.is.ew.ro] has quit ["Leaving."] 11:22:45 -!- Odin- [n=sbkhh@s121-302.gardur.hi.is] has quit [Client Quit] 11:23:02 -!- daniel_ is now known as daniel 11:23:53 -!- tombom [i=tombom@wikipedia/Tombomp] has quit ["Peace and Protection 4.22.2"] 11:28:15 pkhuong: around? 11:28:54 -!- jmbr [n=jmbr@92.32.220.87.dynamic.jazztel.es] has quit [Read error: 110 (Connection timed out)] 11:31:55 _JFT_ [n=_JFT_@modemcable183.11-202-24.mc.videotron.ca] has joined #lisp 11:32:59 -!- milanj [n=milan@79.101.169.94] has quit ["Leaving"] 11:34:27 Yuuhi [i=benni@p5483FFCA.dip.t-dialin.net] has joined #lisp 11:34:54 mehrheit [n=user@lan-84-240-55-160.vln.skynet.lt] has joined #lisp 11:34:59 ttt-- [n=ubuntu@78-23-124-196.access.telenet.be] has joined #lisp 11:34:59 8.1G core.sbcl 11:35:09 nice coredump... 11:36:56 that's sort of a lot... 11:41:02 -!- The-Kenny [n=moritz@p5087CE4E.dip.t-dialin.net] has quit [Read error: 110 (Connection timed out)] 11:42:59 p_l: what do you have in there? application data? 11:43:32 nikodemus: SBCL + CommonQT + Currency converter example + QT data 11:43:38 I suppose GDB dumped *everything* 11:43:44 including mmapped but not used space 11:44:16 oh, you mean that's the vm size? 11:44:16 *p_l* bzips it 11:44:40 nikodemus: Well, I just got into GDB and had it dump a corefile 11:45:04 jmbr [n=jmbr@250.32.220.87.dynamic.jazztel.es] has joined #lisp 11:45:28 that should include the empty space as well, i believe 11:46:00 good thing it wasn't CCL, I wouldn't have enough diskspace for that 11:46:16 nicely bzipped down to 17M 11:47:19 what are you planning to do with that core? (i assume you realize you can't run it with --core) 11:48:07 nikodemus: It's a dump after something went badly. Just in case it might be helpful for someone better with FFI to find out what exactly happened 11:48:39 *nikodemus* impatiently waits for alpha to finish building 11:50:06 p_l: more power to them that can. i would suggest you also zip up the runtime that produced it 11:50:58 vy [n=user@213.139.194.186] has joined #lisp 11:51:01 yeah 11:51:25 http://paste.lisp.org/display/80300 it's from this one 11:51:31 (error, that is) 11:52:28 what platform? 11:53:19 x86_64 linux 11:53:44 hm, there are quite a few signal handling fixes in since 1.0.26.1 11:53:55 can you reproduce it on eg. 1.0.28? 11:55:00 -!- ASau [n=user@193.138.70.52] has quit [Remote closed the connection] 11:55:12 ASau [n=user@193.138.70.52] has joined #lisp 11:55:13 I'll see 11:59:08 -!- kleppari [n=spa@bitbucket.is] has quit ["leaving"] 12:03:27 So how big FFTs do people compute anyway? 12:03:54 on normal hw? 12:05:51 thehcdreamer [n=thehcdre@93.37.246.44] has joined #lisp 12:12:26 -!- hsaliak [n=hsaliak@bb116-15-29-118.singnet.com.sg] has quit [] 12:15:35 -!- dto [n=user@pool-98-118-1-212.bstnma.fios.verizon.net] has quit [Read error: 110 (Connection timed out)] 12:18:10 Pandala [n=Pandala@cbl33-1-82-228-229-180.fbx.proxad.net] has joined #lisp 12:18:45 kleppari [n=spa@bitbucket.is] has joined #lisp 12:18:52 Vegan [n=sdfpme@119.128.75.221] has joined #lisp 12:19:58 yay, i didn't break the build on alpha 12:21:22 hsaliak [n=hsaliak@bb116-15-29-118.singnet.com.sg] has joined #lisp 12:21:41 klausi [n=klausi@port-92-193-7-18.dynamic.qsc.de] has joined #lisp 12:22:00 p_l: Was that question for me? 12:22:16 nego_ [n=nego@c-67-173-168-255.hsd1.il.comcast.net] has joined #lisp 12:22:26 beach: yeah. I kind of figured some of that realtime radio filters might computer quite big FFTs :) 12:22:40 p_l: Answer: Yes, on PC/GNU/Linux. 12:23:00 elias` [n=me@resnet-pat-254.ucs.ed.ac.uk] has joined #lisp 12:23:19 one could use something like CUDA 12:23:32 p_l: What is a "realtime radio filter"? 12:23:34 *nikodemus* commits 12:24:17 beach: realtime digital analysis and filtering of radio signal, used AFAIK in telecommunications and radars 12:24:22 i think he means software radio 12:24:36 "software" 12:25:02 usually used in radios that would change their comm schemes 12:25:23 I see. 12:25:48 well, software radio uses the same kind of thing, I guess 12:26:59 beach: whence the question? 12:29:57 -!- Pandala [n=Pandala@cbl33-1-82-228-229-180.fbx.proxad.net] has quit ["This computer has gone to sleep"] 12:30:24 -!- rstandy [n=rastandy@net-93-144-199-191.t2.dsl.vodafone.it] has quit [Remote closed the connection] 12:30:50 Pandala [n=Pandala@cbl33-1-82-228-229-180.fbx.proxad.net] has joined #lisp 12:31:42 nikodemus: I am again thinking of implementing FFT, and the strategy used is determined by max size. For instance pre-computing e^2 pi k / N is a win, but only if if fits snuggly in the cache. Also, the "bit-reversal" algorithm could be precomputed for different N. 12:32:33 in hardware, i would compose the bit-reversal by catting a few 4-LUTs together 12:32:59 For the applications we use, i.e. sound, a typical size 1024. 12:33:25 Hun: I am thinking of using variable radix. I think sqrt(N) is optimal. 12:34:16 Hun: and as far as I can tell, this complicates the bit-reversal into somthing that no longer is. 12:34:39 -!- hsaliak [n=hsaliak@bb116-15-29-118.singnet.com.sg] has quit [] 12:34:57 -!- nego [n=nego@c-67-173-168-255.hsd1.il.comcast.net] has quit [Read error: 110 (Connection timed out)] 12:35:46 Also, it would be cute to implement it in Common Lisp. Then, SBCL maintainers could figure out ways that it would compile into very efficient code. :) 12:36:08 :) 12:37:20 Like, the "bit-reversal" would be sequences of things like (rotatef (aref v 1) (aref v 9) (aref v 8) (aref v 4)) 12:38:08 we do something like that in our vector cpu (research project) 12:38:28 it has a VSHUF opcode that shuffles whole vector registers according to its parameters :) 12:38:38 I see. 12:39:06 -!- _JFT_ [n=_JFT_@modemcable183.11-202-24.mc.videotron.ca] has quit [] 12:39:58 rstandy [n=rastandy@net-93-144-199-191.t2.dsl.vodafone.it] has joined #lisp 12:40:37 executes in a single cycle per instruction. it works by dividing the entry vector into variable-sized chunks and encoding a new permutation for them. you can reverse the bytes in a 32x32-bit vector by doing a reverse shuffle in 32, 16 and 8 bit chunks 12:44:00 gigamonkey` [n=user@adsl-99-35-218-254.dsl.pltn13.sbcglobal.net] has joined #lisp 12:44:02 jmbr_ [n=jmbr@136.33.220.87.dynamic.jazztel.es] has joined #lisp 12:44:02 Is there a good package for using python-libraries? Python-On-Lisp is alpha Nov 2006. 12:44:54 there is cl-python 12:45:01 but i don't know a thing about it 12:45:54 stassats: Thanks, I'll look it up. 12:47:22 meingbg: Clpython is a interpreter of python written in Common Lisp. It's not necessarily able to actually understand the library you have in mind. But extending clpython to make it so would of course be awesome. 12:47:23 meingbg: don't count on using python libraries that require FFI 12:47:34 I recall some lisp project compiling lisp to python bytecode. 12:47:55 isn't the compiler in CMU and SBCL called Python? ;) 12:48:11 photon [n=photon@unaffiliated/photon] has joined #lisp 12:49:20 grkz [n=qsvans@pdpc/supporter/active/grkz] has joined #lisp 12:51:20 HG` [n=wells@xdsles005.osnanet.de] has joined #lisp 12:51:49 tritchey [n=tritchey@c-68-58-88-241.hsd1.in.comcast.net] has joined #lisp 12:52:52 LostMonarch [n=roby@host215-20-dynamic.48-82-r.retail.telecomitalia.it] has joined #lisp 12:54:10 -!- Spyderco [n=nash@194.45.110.65] has quit [] 12:57:28 stassats: tcr: So Clpython reads the source code in the py library and makes it run on lisp? I was more looking for an FFI. 12:57:57 as i said, i know nothing about it 12:58:13 sreeram [n=sreeram@122.174.73.39] has joined #lisp 12:58:18 except that it exists 13:00:25 -!- jmbr [n=jmbr@250.32.220.87.dynamic.jazztel.es] has quit [Read error: 110 (Connection timed out)] 13:02:26 -!- gigamonkey` [n=user@adsl-99-35-218-254.dsl.pltn13.sbcglobal.net] has quit [Read error: 110 (Connection timed out)] 13:04:22 meingbg: pythononlisp at its current stage may be good enough. Have you tried it? 13:06:57 guys, i'm a noob in lisp, but i'm fond of this language ! 13:07:08 hooray! 13:07:53 Nshag [i=user@Mix-Orleans-106-3-195.w193-248.abo.wanadoo.fr] has joined #lisp 13:07:55 jmbr [n=jmbr@108.33.220.87.dynamic.jazztel.es] has joined #lisp 13:08:34 jmbr_: Haven't tried it, and can't find it on the web either? 13:10:45 -!- tritchey [n=tritchey@c-68-58-88-241.hsd1.in.comcast.net] has quit [] 13:15:05 hum 13:15:36 does anybody has an idea of a small project/program to write to become better in lisp ? 13:17:10 pkhuong: Could you perhaps commit the format destroyed-constant-args bit? 13:17:39 Davidbrcz: projecteuler.net 13:18:12 meingbg: i know this site 13:18:14 Davidbrcz, As an exercise for myself, I wrote a small program the ciphers a message using a key cipher. The user sets the key and message 13:18:23 i've done some exercices on it 13:18:28 (and deciphers) 13:18:34 but it is half boring 13:18:37 just maths 13:18:39 and maths 13:18:44 and again maths 13:18:51 -!- LostMonarch [n=roby@host215-20-dynamic.48-82-r.retail.telecomitalia.it] has quit ["raise RuntimeError"] 13:19:00 Davidbrcz: So what kind of programs would you be interested in? 13:19:29 beach: hum 13:19:33 dunno 13:19:53 every thing, but not just stupid maths calcul 13:19:54 Davidbrcz: Any among these: http://dept-info.labri.fr/~strandh/Lisp-projects/ 13:21:19 -!- thehcdreamer [n=thehcdre@93.37.246.44] has quit [] 13:22:27 ruediger [n=ruediger@62-47-135-171.adsl.highway.telekom.at] has joined #lisp 13:22:30 hi 13:22:38 hello ruediger 13:22:39 beach: hum 13:22:50 LiamH [n=nobody@pool-141-156-215-133.res.east.verizon.net] has joined #lisp 13:23:30 beach: this site presents some finished project 13:23:43 not exam 13:23:51 but i keep it into my bookmarks 13:24:18 Davidbrcz: Few projects ever become "finished". 13:24:48 -!- jmbr_ [n=jmbr@136.33.220.87.dynamic.jazztel.es] has quit [Read error: 110 (Connection timed out)] 13:25:39 beach: hum 13:25:40 Davidbrcz: You are hard to please! No math, no word processor, no text editor, no web browser, no email client, no accounting system. 13:26:10 i think it is a little big for a beginner 13:26:30 beach: no 13:26:44 i agree with writting a text editor 13:26:53 when i'll be good in lisp 13:27:05 not a the begin 13:27:22 Davidbrcz, do you know about hash-tables (or alists or plists) yet? 13:27:32 Davidbrcz: do you have any programming experience in any other language? 13:27:39 beach: C++ 13:27:47 i'm good 13:27:57 bgs100: yep 13:28:07 why ? 13:28:50 meingbg: let me find the svn repository... 13:28:56 jmbr: thanks 13:29:57 beach: An interesting list you have there. The assembly system and accounting application seems interesting. 13:30:17 Davidbrcz: If you are "good" at C++, writing an accounting system in Lisp should be very easy. 13:30:58 meingbg: Yeah, I like the accounting system myself. 13:31:32 meingbg: It is probably almost trivial to write, even with a GUI. 13:31:57 meingbg: svn co svn://common-lisp.net/project/python-on-lisp/svn python-on-lisp 13:32:04 froydnj: aroundp 13:32:23 beach: :o 13:32:27 i miss it 13:32:31 i'll try 13:32:36 thanks beach 13:32:40 jmbr: ok, thank you 13:32:51 Davidbrcz: No problem. 13:33:08 meingbg: it's a pity there's currently nothing better than python-on-lisp :-/ I'd like to write something along the lines of chicken scheme's pyffi but I currently lack the time 13:34:25 hum 13:34:29 jmbr: Perhaps you could continue improving Pyffi from it's current state? 13:34:37 solve a sudoku could be good no ? 13:34:55 Davidbrcz: you didn't want math? 13:34:57 Davidbrcz: Sure, if you solve it the way humans do. Otherwise it is too easy. 13:35:08 meingbg: there is nothing math-like about Sudoku. 13:35:37 meingbg: i would avoid maths when i'm learning a language 13:35:55 I guess you mean python-on-lisp (which is for common lisp). It's current state is so primitive it might be faster to start from scratch 13:36:01 -!- danlei [n=user@pD954F7DF.dip.t-dialin.net] has quit [Read error: 110 (Connection timed out)] 13:36:56 Davidbrcz: I understand, and actually I agree. 13:38:00 beach: Combinatorics is a branch of pure mathematics, but I'm unsure weather or not it can be applied to Sudoku's. 13:38:28 I guess so. 13:39:25 if you want to reason about soduku, like how many possibilities, etc, 13:39:44 beach: About that accounting system - I have had plans to construct an accounting system, but the idea would then be to make it more generalized than the average system. 13:39:44 http://en.wikipedia.org/wiki/Dancing_Links 13:40:00 -!- grkz [n=qsvans@pdpc/supporter/active/grkz] has quit [Read error: 110 (Connection timed out)] 13:40:29 meingbg: In what way is the average system not generalized? 13:41:24 hum, OOP is avaible in lisp 13:41:33 is it useful ? 13:41:37 extremely 13:41:38 ? 13:41:59 Davidbrcz: A lot of modern Lisp code would be OO. 13:42:20 ok, i didn't know that 13:42:23 beach: The average systems I've seen have one-dimensional account-spaces. 13:42:46 i needed to study the eval/apply part of sicp to understand how one can really use inheritance to build stuff :) 13:42:50 beach have you write OO code? 13:42:55 with lisp of course.. 13:42:59 meingbg: I'm afraid I don't know what that means. Could you give an example? 13:43:18 r1nu-: Some yes :) Like McCLIM, Gsharp, Climacs. 13:43:50 xmm have you noticed something better? (better than other OO languages..) 13:45:22 r1nu-: Me? I don't think you can compare programming paradigms that simply. It depends on the problem at hand. But Lisp is good for quite a few paradigms. 13:46:41 beach: Accounts are generally defined by it's number (accounts are in N-space, that is to to say), and a transaction is moving money from one account to another, or perhaps from/to multiple accounts. 13:47:15 meingbg: I am following so far. 13:50:25 nyef [n=nyef@vcwl1-61.daktel.net] has joined #lisp 13:50:31 G'morning all. 13:50:36 hello nyef 13:51:33 beach: If we generalize the account space from 1 dimension to n dimensions, we suddenly have more than an accounting system - we have a system wherewith we can describe things that happen. A view can then be used to extract the information needed by conventional accounting systems. There would be n dimensions, and every account would, instead of being identified by a natural number, have a name and a parent account. Every tree would be 13:51:33 called a dimension. The debit-credit zero difference requirement would then apply to each such dimension, even though a single transaction can operate on multiple dimensions. 13:51:39 -!- HG` [n=wells@xdsles005.osnanet.de] has quit [Client Quit] 13:52:18 *nyef* is building SBCL HEAD to see if nikodemus' latest commit affects the latin-9 bug. 13:52:48 chris2 [n=chris@dslb-094-216-067-251.pools.arcor-ip.net] has joined #lisp 13:53:36 meingbg: I am willing to believe you, but I guess I don't see what it would be useful for. 13:54:04 failure-p was set when creating ... from-host, compiler/generic/vm-ir2tran. 13:54:45 gigamonkey` [n=user@adsl-99-35-218-254.dsl.pltn13.sbcglobal.net] has joined #lisp 13:54:47 Cleaning up, checking for repeatability prior to beginning a bisection. 13:56:01 meingbg: Actually, a useful accounting system already has to have *some* tree structure, because there are different types of accounts like cost accounts, revenue accounts, supplier accounts, client accounts, tax accounts, etc, that must be treated differently when you do the annual report. 13:56:01 tcr: there's a comment wrt the hairiness of determining when constant arguments to format are clobbered. I might have been too aggressive and overlooked a couple edge cases. 13:56:30 beach: Then look into how accounting systems work today. There are things like "accounts in the range 2400-2499 are called so-and-so-accounts" etc. So the easy-to-implement-in-C system with account numbers is used to create an actual tree structure. 13:56:46 Anyone else seeing build failures of SBCL HEAD? 13:56:54 The-Kenny [n=moritz@p5087CE4E.dip.t-dialin.net] has joined #lisp 13:56:57 beach: Oh, you already said that. 13:57:28 nyef: let me try 13:57:50 beach: Yes, and some accounts are used in a way that could be implemented with a view instead. 13:57:53 meingbg: Well, I implemented an accounting system for my company, and I used inheritance rather than relying on a particular numbering system. 13:58:20 gigamonk` [n=user@adsl-99-17-207-207.dsl.pltn13.sbcglobal.net] has joined #lisp 13:59:01 beach: And then what? Did you make some kind of export utility to convert to standardized account numbers? 13:59:24 meingbg: No, that wasn't required. 14:00:26 meingbg: How standardized are those account numbers internationally? 14:00:34 nyef: yep 14:00:44 FAILURE-P was set when creating "obj/from-host/src/compiler/generic/vm-ir2tran.lisp-obj". 14:01:14 nikodemus: Are you busy right now? 14:01:21 beach: In sweden we have no formal requirement, but most companies use either EU-BAS-97 or the older national BAS-96 14:01:36 stassats: I'm on x86-64 linux, you? 14:01:43 me too 14:01:53 let me check on 32 14:02:16 beach: I think EU-bas-97 is some kind of european thing, that is, not international. 14:02:26 Okay, I'm doing a bisection. 14:02:30 meingbg: Looks plausible. 14:02:47 nickw [n=nick@93-97-203-158.zone5.bethere.co.uk] has joined #lisp 14:03:14 oh damn, it's already six o'clock and still not found that slime/cmucl bug 14:03:37 meingbg: When I looked into this, I couldn't find a single standard, so it would have been foolish to use numbers to determine account types. Instead, I have classes that represent account types, and the user can assign any old number (or none) to any account. 14:03:57 beach: Well, yes. What exactly do you mean by using inheritance in your accounting system? 14:04:48 meingbg: Like I have a class called cost-account, and all cost accounts like telephone, travel, office supplies etc inherit from it. 14:05:29 beach: So how is the "travel" account for example, identified? 14:05:56 meingbg: I just use Lisp identifiers. 14:06:04 i.e. symbols 14:06:16 beach: OK, sounds good. 14:06:45 -!- ace4016 [i=ace4016@cpe-76-168-248-118.socal.res.rr.com] has quit ["night"] 14:06:50 levy [n=ati@apn-89-223-204-129.vodafone.hu] has joined #lisp 14:07:02 levy_ [n=levy@apn-89-223-204-129.vodafone.hu] has joined #lisp 14:07:19 beach: So then you kind of did what I had in mind. 14:07:37 meingbg: OK, I see. 14:07:38 nyef: depends :) 14:07:50 oh, what, build failures? 14:07:53 -!- levy_ [n=levy@apn-89-223-204-129.vodafone.hu] has left #lisp 14:07:56 ouou 14:08:07 meingbg: I just never looked into average systems, so I didn't know what kind of limitations they would have. 14:08:22 tritchey [n=tritchey@c-68-58-88-241.hsd1.in.comcast.net] has joined #lisp 14:08:23 Yeah, and it's in your latest range of commits. I'm doing a bisection now. 14:08:44 what's the warning? 14:09:18 Bunch of style-warnings about unused variables in generic/vm-ir2tran that I saw. 14:09:27 beach: That's probably the best way to go about things anyways. Make your own system while you have your mind unpolluted, next look into standard/other systems and make your version 2 system the best of two worlds.dddd 14:09:30 hm 14:09:32 rdd [n=user@c83-250-152-109.bredband.comhem.se] has joined #lisp 14:09:37 is (sheet-transformation) supposed to work in mcclim? 14:09:45 meingbg: That is often how I work, yes. 14:09:46 you're on linux/x86-64? 14:09:47 Once I get the bisection done I'll get a build log for you. 14:09:47 unused variables and unknown ignored variables 14:09:50 Yes. 14:10:10 *nikodemus* builds on x86-64 14:10:37 and also error "; SB!VM::DATA-VECTOR-SET-WITH-OFFSET/SIMPLE-BASE-STRING is not the name of a defined VOP." 14:10:56 aha 14:10:56 -!- Pandala [n=Pandala@cbl33-1-82-228-229-180.fbx.proxad.net] has quit ["Leaving"] 14:11:01 Better. I didn't see that, but I also wasn't looking. 14:11:04 Pandala [n=Pandala@cbl33-1-82-228-229-180.fbx.proxad.net] has joined #lisp 14:11:10 *nikodemus* sees it too 14:11:16 i think i know what's wrong 14:11:45 Yeah, bisect points to 1.0.28.51. 14:11:52 -!- gigamonkey` [n=user@adsl-99-35-218-254.dsl.pltn13.sbcglobal.net] has quit [Read error: 110 (Connection timed out)] 14:12:06 HG` [n=wells@xdslfb238.osnanet.de] has joined #lisp 14:12:13 ikki [n=ikki@189.228.133.213] has joined #lisp 14:12:51 beach: If you also implement "dimensions" the way I described, you would be able to store information in transactions about for example who was involved and how the transaction was done in real world, and make views such as "Things sold cash" or "Services bought from this company" 14:12:52 -!- jmbr [n=jmbr@108.33.220.87.dynamic.jazztel.es] has quit [Read error: 54 (Connection reset by peer)] 14:13:02 jmbr [n=jmbr@108.33.220.87.dynamic.jazztel.es] has joined #lisp 14:13:32 beach: Or add multi-currency support with settings, no coding. 14:13:54 beach: (To answer your question about usefulness) 14:14:05 meingbg: I see. I do store information in transactions currently, because the accounts are never updated. 14:15:02 beach: Which gives you the ability of calculating any account's value on-demand. 14:15:15 Yes. 14:15:15 beach: as well as any account's historical value. 14:15:30 Again, yes. No information is ever destroyed. 14:15:45 nikodemus: initialize-vector always inlines (?). I think that something that blits with an inline loop by default and inlines the loop on short arguments would be preferable. 14:15:47 meingbg: It also allows me to modify an old transaction at the end of the year :) 14:16:18 -!- HG` [n=wells@xdslfb238.osnanet.de] has quit [Client Quit] 14:16:25 beach: :) 14:16:32 is there an mcclim guru somewhere around? 14:16:41 *beach* hides 14:17:12 levy: I wouldn't call myself a guru, but there are some other people around as well, so please state your problem. 14:18:06 I am working on a game which simulates bodies moving around using gravity (solving a differential equation with Runge Kutta 4) 14:18:25 and I would like to have infinite zooming and scrolling 14:18:36 beach: At the end of the year, a common accounting task is to summarize the year in different ways. Usually a lot of accounts are used for this, but in your system it would simply be a view, right? 14:19:03 I found sheet-transformation, but I could not make it work 14:19:05 meingbg: It could be. I just didn't go that far. 14:19:48 initializing the :sheet-transformation of a clim-stream-pane subclass instance to (make-scaling-transformation 2 2) causes the result of my custom drawing to continuously jump up and down a bit when (redisplay-frame-panes) is called from the background thread 14:20:01 beach: I see. It would probably just confuse the accountants anyway... 14:20:09 eh, I probably do something wrong, but I don't see what 14:20:57 meingbg: that would be me! 14:21:00 pkhuong: initalize-vector is only used when the elements are either constants of arguments :INITIAL-CONTENTS (LIST ...) 14:21:05 ok, i have a fix 14:21:28 beach: Oh, I see :) 14:21:42 pkhuong: so a loop would be pretty pointless 14:21:53 :initial-element always gets a loop, on the other hand 14:22:23 my display function just draws a couple of circles, nothing fancy there and if I remove the :sheet-transformation initarg then it draws as expected 14:22:44 s/constants of/constants or/ 14:22:45 levy: The sheet transformation determines how coordinates are transformed from the coordinate system of the sheet to that of its parent. 14:23:01 levy: You might want to set the medium transformation instead. 14:23:21 nikodemus: Are these transforms also supposed to clear up the latin-9 bug? 14:23:30 nyef: no idea 14:23:33 beach, ah I'l give it a try 14:23:40 they're supposed to make things faster :) 14:23:40 Fair enough. 14:23:44 levy: You typically don't touch the sheet transformation yourself. It is used by things like layout panes, etc. 14:23:56 nikodemus: I've generated really humongous initial-content lists. 14:24:32 beach, I told you that I'm probably doing something wrong :) 14:24:54 levy: Yep. Looks like you were right. 14:25:30 It occurred to me that it would be possible to make the underlying cause of the latin-9 bug subtler by having the cross-compiler transform calls to make-array with an element-type specified to use the upgraded element-type, which would mean that make-array sees an already-upgraded type at runtime, and thus wouldn't upgrade to T. 14:25:52 beach, thx! 14:26:03 pkhuong: Where's that comment? 14:26:37 levy: Let me know if it works. 14:26:44 tcr: fndb.lisp, just above the defknown for FORMAT 14:27:12 pkhuong: how would you make the initialization a loop? 14:27:20 I'd rather have the code pregenerate a specialised vector identical to the destination as a constant + blitting loop for medium/long vectors. 14:27:52 oh, you mean *constant* lists 14:27:59 yeah, that could be done 14:28:00 beach, works like a charm... 14:28:13 levy: Excellent! 14:28:32 Careful with doing that in the cross-compiler, though. 14:28:47 The code for dumping specialized vectors is a little funky-looking. 14:29:11 why isn't there an (or multiple) RFC(s) or something of the likes that expresses how lisps should handle things like networking (and other things that aren't part of the spec)? 14:29:29 only 4 SSE2 instruction variants left! 14:29:38 davazp [n=user@56.Red-79-153-148.dynamicIP.rima-tde.net] has joined #lisp 14:29:41 pkhuong: extremely cool 14:29:47 are you adding them to x86 as well? 14:29:51 pkhuong: ? are you implementing support for SSE2 (sbcl?) ? 14:29:59 not yet :) 14:30:13 danlei [n=user@pD9E2ED2F.dip.t-dialin.net] has joined #lisp 14:30:42 I guess I could add all the definitions to x86 and leave the problem of alignment for another set of commits. 14:31:04 borsman [n=quassel@76.177.217.216] has joined #lisp 14:32:24 -!- The-Kenny [n=moritz@p5087CE4E.dip.t-dialin.net] has quit [Read error: 110 (Connection timed out)] 14:33:29 i have vector data alignment done 14:33:37 -!- vy [n=user@213.139.194.186] has quit [Remote closed the connection] 14:33:41 so... 14:33:42 stack? 14:33:49 stack too, iirc 14:34:14 need to make sure it doesn't clash with the latest array stuff, but should be all there 14:34:17 so I can ask for :alignment 4 for sse-stack? 14:34:32 oh, no 14:34:52 sorry. just that vector data is 16 byte aligned on both stack and heap 14:35:09 Pretty good already. 14:35:35 -!- sepult [n=sepult@xdsl-87-78-29-196.netcologne.de] has quit [Read error: 110 (Connection timed out)] 14:35:42 -!- gigamonk` [n=user@adsl-99-17-207-207.dsl.pltn13.sbcglobal.net] has quit [Read error: 110 (Connection timed out)] 14:36:00 I'm also fairly certain that xmm15 can be used safely on x86-64, now that we have code to move from fp-zero SCs (XOR). 14:36:04 sepult [n=sepult@xdsl-87-78-131-159.netcologne.de] has joined #lisp 14:36:19 -!- rstandy [n=rastandy@net-93-144-199-191.t2.dsl.vodafone.it] has quit [Read error: 104 (Connection reset by peer)] 14:36:55 a defoptimizer declines when it returns NIL, right? 14:37:41 beach, am I right when I think that the gtkcairo backend is faster than CLX? 14:38:15 levy: I don't know. I haven't used the gtkairo backend very much. 14:38:15 beach, I am having some strange performance problems when drawing let's say 100 points individually 14:38:43 -!- Pandala [n=Pandala@cbl33-1-82-228-229-180.fbx.proxad.net] has quit ["Leaving"] 14:38:49 levy: Oh, that's possible, 'cause each point would generate an X protocol request. 14:39:17 beach, is there a way to work around that? 14:39:43 beach, I would like to draw the trajectory of bodies 14:40:16 er a polyline would be probably better 14:40:32 tcr: iirc it depends on the optimizer: some are not allowed to decline 14:40:47 but derive-type optimizers iirc returns nil for "dunno", yes 14:40:52 levy: for individual points, draw-points or draw-points* might actually generate a single request. 14:41:06 If there isn't a CLIM function for drawing multiple points at once, would it make sense for the backend to batch a series of consecutive point draw operations (and line draw operations) in order to send them all at once? 14:41:28 nyef: there are two such functions. 14:41:42 Fair enough. 14:41:48 oh, I see 14:41:55 -!- sepult [n=sepult@xdsl-87-78-131-159.netcologne.de] has quit ["KVIrc 3.4.0 Virgo http://www.kvirc.net/"] 14:43:41 lelelele [n=leif@h59ec2263.c46-04-13.dyn.perspektivbredband.net] has joined #lisp 14:43:49 jsoft [n=Administ@unaffiliated/jsoft] has joined #lisp 14:43:59 -!- drafael [n=tapio@ip-118-90-129-14.xdsl.xnet.co.nz] has quit ["Leaving."] 14:49:05 Does sbcl optimize (cond (nil a) (t b)) => b and similar? 14:49:11 -!- davazp [n=user@56.Red-79-153-148.dynamicIP.rima-tde.net] has quit [Remote closed the connection] 14:49:17 yes. 14:49:27 DISASSEMBLE is your friend 14:50:27 ok, thanks. 14:51:01 what do you think of the book, Artificial Intelligence: A Modern Approach? 14:53:00 lacking the knowledge of ASM itself, disassemble doesn't tell me all too much 14:53:08 dreish [n=dreish@minus.dreish.org] has joined #lisp 14:53:53 time to acquire that knowledge, it's simple 14:53:59 madnificent: learning is good for you :) 14:54:13 Isn't there a screenshot showing slime repl with graphical output? 14:54:29 but you should also become able to recognize some things (like full calls) pretty quickly 14:54:40 madnificent: when there are some labels and branches, it didn't get optimized away 14:54:42 leo2007: It is a good book, the standard introductionary book to AI too me thinks :) 14:55:13 pjb: http://collison.ie/blog/2008/06/pictures-in-the-slime-repl 14:55:19 nikodemus: I must choose what I learn. Time is limited. I simply wanted to express that disassemble isn't helpful for everyone 14:55:33 Hun: ah 14:56:05 also when there are some fdefinitions and calls, the function calls didn't get inlined 14:56:50 -!- jsoft [n=Administ@unaffiliated/jsoft] has quit [Remote closed the connection] 14:57:10 adeht: thanks, but it seems I don't get data from collision.ei :-( 14:57:24 transpose on `ei' 14:57:36 also, collison 14:58:11 Oops. But I copy-and-pasted correctly the URI. 15:00:25 lelelele: in terms of lisp programming, is it a good one? 15:00:39 beach, what do you suggest if I want to draw a polyline with a 100 points but each segment with different color fading from black to white? 15:00:46 lelelele: its web pages says it has 90% market share 15:00:47 beach, performancewise 15:01:02 leo2007: it's not concentrated on lisp, afaik 15:01:32 leo2007: for that there is Paradigms of AI Programming 15:01:39 there are code examples in several languages, one of them lisp, on the web site 15:02:00 iirc 15:02:02 nikodemus pasted "examples for madnificient" at http://paste.lisp.org/display/80335 15:03:32 stassats and adeht, nice. 15:03:36 levy: That, to me, sounds like something to consider using opengl for. 15:04:00 (Or did you mean the lines would be a solid color, just a different color for each line?) 15:04:11 jsoft [n=Administ@unaffiliated/jsoft] has joined #lisp 15:04:15 do you know of other universities other than MIT and berkeley that use lisp in their research and lectures? 15:04:39 I try to insert a BREAK to the DEFTRANSFORM of funcall, but it doesn't seem to be triggered; what am I missing? 15:04:42 leo2007: IIRC there isn't a single line of lisp in that book. 15:05:02 each line is just a few pixels, together they make a spline like thing fading into the background 15:05:04 -!- klausi [n=klausi@port-92-193-7-18.dynamic.qsc.de] has quit ["Leaving"] 15:05:20 tcr: What sort of "doesn't seem to be triggered"? 15:05:22 it just shows the path the body went through in the past 15:05:32 leo2007: Oh sorry. I see people already replied. Yes, it is not focused on any language in particular. But a good book on AI :) 15:06:31 nyef: I don't get into the debugger when I C-c C-c 15:06:53 ah, the source-transform gets run first 15:07:29 I wonder when the deftransform gets a chance to run in light of that source-transform 15:07:36 Ooh. Slime repl with graphical output. I've wanted that at times. 15:09:02 nyef: check out the slime-devel thread, I submitted an almost-contrib, though I've not updated it to work with latest slime 15:09:24 Yeah, reading the thread now. 15:09:51 What I ended up doing is working in the emacs *scratch* buffer and calling swank to invoke a lisp function that returns a PNM body. 15:10:24 http://www.lisphacker.com/temp/shot1.png 15:11:02 ok, build fix committed 15:11:33 tcr: if the source transform picks it, the deftransform never runs 15:12:57 nikodemus: the question is when does the source-transform not run? 15:14:45 eh, this is a pretty funny game like thing 15:16:31 rstandy [n=rastandy@net-93-144-199-191.t2.dsl.vodafone.it] has joined #lisp 15:18:42 envi^home [n=envi@220.121.234.156] has joined #lisp 15:18:51 in case anybody wants to give it a try: darcs get http://common-lisp.net/project/cl-dwim/darcs/comet 15:19:36 -!- adeht [n=death@bzq-84-110-251-220.red.bezeqint.net] has quit ["ERC Version 5.3 (IRC client for Emacs)"] 15:19:59 you can drag bodies, do infinite scroll and zoom, add new bodies and delete existing ones, speed up and down simulation (Runge Kutta 4th order and the newton laws rulez) 15:20:08 tcr: IIRC, they only run on the user's code. If a deftransform expands into FUNCALL, you'll have to rely on the deftransform. Source transforms have the advantage of requiring less fiddling by the system -> higher performance. 15:20:25 tcr: given sufficiently hairy code where the funcall is not apparent in the source 15:20:33 pkhuong: i hope you're wrong 15:20:47 adeht [n=death@bzq-84-110-251-220.red.bezeqint.net] has joined #lisp 15:20:59 -!- Modius_ [n=Modius@99.179.100.6] has quit [Read error: 54 (Connection reset by peer)] 15:21:03 source transforms should definitely run on transformed code as well 15:21:18 -!- Athas [n=athas@0x50a157d6.alb2nxx15.dynamic.dsl.tele.dk] has quit [Read error: 104 (Connection reset by peer)] 15:21:22 Athas [n=athas@0x50a157d6.alb2nxx15.dynamic.dsl.tele.dk] has joined #lisp 15:21:22 Modius_ [n=Modius@99.179.100.6] has joined #lisp 15:21:24 *nikodemus* checks 15:21:35 compiler macros don't, do they? 15:22:26 they too should 15:22:54 yes, transform-call passes the result into the normal compiler codepath via transform-inline-lambda 15:23:08 i mean ir1-convert-inline-lambda 15:23:21 screenshot at: http://imagebin.org/49263 15:23:37 and notinline inhibits all three -- source-transforms, compiler macros, and deftransforms 15:23:45 -!- jsoft [n=Administ@unaffiliated/jsoft] has quit ["leaving"] 15:24:56 -!- Quadrescence [n=quad@unaffiliated/quadrescence] has quit [Read error: 110 (Connection timed out)] 15:24:59 nikodemus: ah, like (apply #'funcall #'list '(1 2 3))! 15:26:54 hm unfortunately not 15:27:44 hum 15:27:46 yeah. i'm pretty sure i've seen it fire, but it takes a bit of trying :) 15:27:56 guys, how i could do something like that 15:28:02 -!- Athas [n=athas@0x50a157d6.alb2nxx15.dynamic.dsl.tele.dk] has quit [Connection reset by peer] 15:28:02 -!- Modius_ [n=Modius@99.179.100.6] has quit [Read error: 104 (Connection reset by peer)] 15:28:06 Athas [n=athas@0x50a157d6.alb2nxx15.dynamic.dsl.tele.dk] has joined #lisp 15:28:14 something like what? 15:28:16 Davidbrcz: Like what? 15:28:17 (defun foo (f P) (f P)) 15:28:25 Modius_ [n=Modius@99.179.100.6] has joined #lisp 15:28:25 and use it like 15:28:36 (funcall f P) 15:28:42 ok 15:28:59 At the same time, given the existence of funcall, why bother? 15:30:42 nyef: because i'm creating a function do to som of f(i) with i between K and P 15:30:50 and f is undefined 15:32:11 -!- nyef [n=nyef@vcwl1-61.daktel.net] has quit [Remote closed the connection] 15:32:29 -!- ThomasI [n=thomas@unaffiliated/thomasi] has quit ["Bye Bye!"] 15:32:45 -!- billstclair [n=billstcl@unaffiliated/billstclair] has quit [Read error: 113 (No route to host)] 15:35:20 levy: nice! 15:35:53 dto [n=user@pool-98-118-1-212.bstnma.fios.verizon.net] has joined #lisp 15:37:14 beach, thx, started yesterday evening... :-) 15:37:32 -!- Modius_ [n=Modius@99.179.100.6] has quit [Client Quit] 15:37:55 shmho [n=user@58.142.15.103] has joined #lisp 15:38:24 jsoft [n=Administ@unaffiliated/jsoft] has joined #lisp 15:38:41 Modius [n=Modius@99.179.100.6] has joined #lisp 15:39:42 Modius_ [n=Modius@99.179.100.6] has joined #lisp 15:40:27 -!- Modius_ [n=Modius@99.179.100.6] has quit [Client Quit] 15:40:45 -!- Modius [n=Modius@99.179.100.6] has quit [Client Quit] 15:41:02 Modius [n=Modius@99.179.100.6] has joined #lisp 15:41:06 alec [n=alec@pool-96-237-2-171.bstnma.east.verizon.net] has joined #lisp 15:41:39 -!- Vegan [n=sdfpme@119.128.75.221] has quit [Remote closed the connection] 15:44:10 reaver__ [n=reaver@h15a2.n2.ips.mtn.co.ug] has joined #lisp 15:44:57 levy: I find whipping up a GUI using McCLIM is a breeze. 15:45:08 drafael [n=tapio@ip-118-90-129-14.xdsl.xnet.co.nz] has joined #lisp 15:45:31 is there something like a 10 minute introduction to clim somewhere? 15:46:17 beach, especially with the REPL at hand 15:46:17 madnificent: that would be as strange as a 10 minute introduction to Common Lisp. I could give you the buzzwords in 10 minutes perhaps. Is that what you want? 15:46:32 I want to install cl-ppcre via asdf-install. I choose a personal installation since I don't run sbcl as root. But for some reason it tries to write something to /usr/local/lib/sbcl/site/cl-unicode-0.1.1/build/util.fasl. Which of course fails 15:46:39 what's wrong? how can I fix this? 15:46:57 beach: buzzwords plus a tiny example would be it 15:46:58 I also work with C++ and Java (programming for food) too and it is such a pain that there is no REPL 15:47:00 levy: Yeah. For Gsharp, I created my own font. So I needed a font viewer. It took only a few hours to write and it is 171 lines of Lisp/CLIM code. 15:47:30 beach, that LOC is always the case wit lisp :) 15:47:37 levy: letting lispers code java is like torture :( 15:47:38 beach: a 10 minute screencast that shows you whipping up a gui would be nice, though 15:47:43 Lack of REPL in java is torture... 15:47:49 (incf nikodemus) 15:47:51 levy: With a different GUI library? 15:47:58 madnificent, there is worse than Java, say C++ 15:48:03 madnificent: If you start reading the getting-started in the McCLIM documentation and stop after 10 minutes, then you should be fine. 15:48:15 beach: :D thanks! 15:48:33 beach, I meant in general, I always end up with short and concise code 15:48:40 nikodemus: the closest thing we have to that is gilberth's slides at LSM. 15:48:49 levy: Sure, yes. 15:48:55 but i agree. getting into clim is not that hard, really. granted, my apps have been pretty trivial, but they've done what they were supposed to 15:49:02 p_l: everything in java is torture 15:49:16 beach, clim has very good concepts I think 15:49:36 levy: can i ask you about cl-def? you use it, right? 15:49:46 nikodemus, sure 15:49:49 levy: Yes, I agree. That's why I wanted it for Gsharp. And that's why we have McCLIM. :) 15:49:55 basically, what is it good for? 15:50:33 each definer has a couple of options 15:50:41 -!- attila_lendvai_ is now known as attila_lendvai 15:50:51 *attila_lendvai* scrolls back 15:51:15 for example the function definer can export the name, declaim it inline, optimize, etc. 15:51:51 I hate keeping package definition export list up to date 15:52:32 ok. anything else? 15:53:05 ehu` [n=chatzill@82-170-33-173.ip.telfort.nl] has joined #lisp 15:53:16 same for class, exporting accessors, slot names, etc. 15:53:17 deffoo-bar-where-to-put-dashes 15:54:07 also, definitions themselves can have arguments. (def (foo :some-args-for-the-definition-itself 42) name ...) 15:54:13 wow macroexpansion happens late in sbcl 15:54:28 HG` [n=wells@85.8.71.44] has joined #lisp 15:55:10 tcr: what do you mean? 15:55:31 which, as levy said, includes things like centrally controlled optimization levels for function definitions. i can turn them off globally with a switch when hunting a nasty bug... 15:55:32 levy: Slime can do it for you. C-c x adds the symbol at point to your defpackage form 15:56:02 when i read code that uses it, i see a lot of (def function ...) (def method ..) (def macro ..), which is a rather trivial layer -- not a problem to read 15:56:18 tcr, why not keeping it in place? 15:56:28 Buganini [n=buganini@security-hole.info] has joined #lisp 15:57:10 nikodemus: I'd have guessed that there's first a macroexpansion run which expands everything. But process-toplevel-form only expands the toplevel form itself, but not recursively. 15:57:11 tcr: the problem with C-c x is a bit philosophical: i'd like to see what is exported at the same point where it is defined. 15:57:12 but then i run across (def definer foo ...) and later (def foo ...), and i *really* miss (defmacro deffoo ...) having been used instead, because macroexpansions from (def definer foo ...) have been a tad opaque to me 15:57:49 nikodemus, well in those cases it's a nop and it does not make more sense than just having consistency, you are absolutely right 15:58:06 except the exporting/optimization flags... 15:58:24 jimi_hendrix [n=user@unaffiliated/jimihendrix/x-735601] has joined #lisp 15:58:26 attila_lendvai: Sure I want that, too, but I think exported symbols should be fontified specially in the CADR of a deffoo form 15:58:31 whats a good socket lib? 15:58:52 jimi_hendrix: iolib 15:58:55 anyways, thanks for the explanation 15:59:07 jimi_hendrix: many projects use usocket, I think. If you don't need win32 support, iolib (it has IPv6!) 16:00:15 tcr, and what about optimized or inlined functions, and all the combinations, etc 16:00:22 nikodemus: where did you end up looking at the macroexpansion of (def definer foo ...)? just curious, because i've never did except when working on cl-def... 16:00:30 ruediger_ [n=ruediger@62-47-133-248.adsl.highway.telekom.at] has joined #lisp 16:00:32 LostMonarch [n=roby@host215-20-dynamic.48-82-r.retail.telecomitalia.it] has joined #lisp 16:00:50 i was looking at cl-serializer, iirc 16:00:56 jimi_hendrix: if you need something lightweight and portable, then use usocket. if you need something serious, less portable then iolib 16:01:45 levy: What about it? 16:02:32 tcr: (def (function ioe) foo ...) defines a function that has optimize declarations (unless centrally disabled), declaims it inline, and exports its name. 16:02:41 tcr, how do you color a function which is exported and inlined? what about inlined, internal, optimized? 16:03:14 wow that's ugly 16:03:15 btw, my tool for checking exported symbols is basically fuzzy completion... :) 16:03:38 -!- HG` [n=wells@85.8.71.44] has quit [Client Quit] 16:03:53 benny99 [n=benny@p5486E4E6.dip.t-dialin.net] has joined #lisp 16:04:20 tcr, well that's a matter of taste and being used to something 16:04:29 "ugly" is subjective. although i wouldn't say "beautiful" to qualify it either. i'd use "practical"... 16:05:20 jsoft_ [n=Administ@unaffiliated/jsoft] has joined #lisp 16:05:46 when are we going to forget keeping code in text files? 16:06:31 The-Kenny [n=moritz@p5087CE4E.dip.t-dialin.net] has joined #lisp 16:06:39 yeah, that's the ultimate "ugly"... everything else that fools around with characters is just background noise... :) 16:07:47 cl-def also leaves room for things like (def (constant :test 'string=) +foo+ "bar") 16:08:55 -!- antoszka [n=antoszka@unaffiliated/antoszka] has quit [Read error: 113 (No route to host)] 16:09:54 jmbr_ [n=jmbr@7.33.220.87.dynamic.jazztel.es] has joined #lisp 16:11:06 schoppenhauer [n=schoppen@unaffiliated/schoppenhauer] has joined #lisp 16:14:29 -!- lelelele [n=leif@h59ec2263.c46-04-13.dyn.perspektivbredband.net] has quit [Remote closed the connection] 16:15:38 mcspiff [n=user@drmons0501w-142177079167.pppoe-dynamic.ns.aliant.net] has joined #lisp 16:16:26 hsaliak [n=hsaliak@bb116-15-29-118.singnet.com.sg] has joined #lisp 16:16:28 -!- ruediger [n=ruediger@62-47-135-171.adsl.highway.telekom.at] has quit [Read error: 110 (Connection timed out)] 16:19:38 beach, thanks for the help once again! 16:19:42 have to go now 16:20:00 -!- levy [n=ati@apn-89-223-204-129.vodafone.hu] has quit ["..."] 16:22:43 -!- jmbr_ [n=jmbr@7.33.220.87.dynamic.jazztel.es] has quit [Read error: 104 (Connection reset by peer)] 16:23:00 nvoorhies [n=nvoorhie@adsl-75-36-204-63.dsl.pltn13.sbcglobal.net] has joined #lisp 16:23:00 jmbr_ [n=jmbr@7.33.220.87.dynamic.jazztel.es] has joined #lisp 16:23:05 -!- Athas [n=athas@0x50a157d6.alb2nxx15.dynamic.dsl.tele.dk] has quit [Remote closed the connection] 16:24:56 -!- jmbr [n=jmbr@108.33.220.87.dynamic.jazztel.es] has quit [Read error: 110 (Connection timed out)] 16:25:10 antoszka [n=antoszka@unaffiliated/antoszka] has joined #lisp 16:30:11 kami- [n=user@unaffiliated/kami-] has joined #lisp 16:30:23 hello 16:30:33 -!- existentialmonk [n=carcdr@64-252-60-75.adsl.snet.net] has quit [Remote closed the connection] 16:33:38 attila_lendvai: are there recommendations for the use of :direct-store in cl-perec? 16:35:30 kami-: hi! use it to control in which sql table the slot should go. for example if you have a nameable-mixin, then it may be preferable to store the slot's column in the subclasses' table so that you can install individual unique constraints 16:36:44 also it has performance implications, because if there's a slot in a base class' table, then the two tables need to be joined to query an instance 16:36:45 attila_lendvai: and keep the super-class's tables smaller (in terms of number of rows) 16:36:51 -!- bgs100 [n=ian@unaffiliated/bgs100] has quit [Read error: 110 (Connection timed out)] 16:37:01 kami-: or completely nonexistent 16:37:08 tell me guys, I'm trying to use sb-bsd-sockets and sb-thread in some code which use asdf, and I can't figure how to do properly 16:37:14 I've an asd files 16:37:21 -!- hsaliak [n=hsaliak@bb116-15-29-118.singnet.com.sg] has quit [] 16:37:29 I defined inside a system 16:37:40 attila_lendvai: so, to minimise the number of joins, I would select push-down? 16:37:54 then I've a file pkg.lisp which defines my package, , which :use :sb-thread 16:38:02 but I can't add :sb-bsd-sockets to :use 16:38:16 why is sb-bsd-sockets handled differently from sb-thread ? 16:39:36 kami-: yes, but keep the other implications in mind also 16:40:21 kami-: querying :push-down'd slots for example needs to query all the subclasses' table, etc.. the usual sql stuff 16:42:12 attila_lendvai: yes. I'll keep that in mind. 16:42:27 saikat [n=saikat@69.181.127.247] has joined #lisp 16:46:57 -!- rdd [n=user@c83-250-152-109.bredband.comhem.se] has quit [Remote closed the connection] 16:47:10 hsaliak [n=hsaliak@bb116-15-29-118.singnet.com.sg] has joined #lisp 16:47:13 -!- tritchey [n=tritchey@c-68-58-88-241.hsd1.in.comcast.net] has quit [] 16:47:15 attila_lendvai: I have another problem regarding the UI. My users don't like the way the standard commands in the inspector/viewer expand the inspector. They don't 'understand' what to do. 16:47:26 -!- hsaliak [n=hsaliak@bb116-15-29-118.singnet.com.sg] has quit [Client Quit] 16:48:12 kami-: "dommands"? you mean the links rendered for instance references that open the details view inline when clicked? 16:48:47 attila_lendvai: yes. detail view, selection/search for slot values which are entities etc. 16:49:31 attila_lendvai: I though maybe they a kind of drill down would better suit them 16:49:36 s/though/thought 16:50:13 kami-: what is a drill down if not the current behavior? 16:50:45 -!- awayekos is now known as anekos 16:51:01 kami-: also, don't let yourself guided merely by your user's *first* responses... 16:51:24 attila_lendvai: I thought of something like replacing the viewer with the instance-selector and keep a kind of bread crumb navigation at the top of the component? 16:52:00 people usually complain for anything unusual when they meet it first, but it does not mean they'll keep on complaining after a getting-used-to period 16:53:27 ...and to be honest i'm a bit lost here, because i've spent too much of my time with the old codebase recently... :| 16:53:49 galdor: :use doesn't load the relevant files, only uses a package that must already exist. I am guessing sb-bsd-sockets is a package that lives in a different file that you need to load, or that you need to indicate as a dependency in the ASDF file. 16:54:52 attila_lendvai: your app in production, does it have the same layout as the wui demo? 16:54:55 beach: mhh; i'm using several files, so I'm searching for the cleanest solution 16:54:56 in your defsystem, so :depends-on (:sb-bsd-sockets) 16:55:15 sb-thread is part of the core if sbcl is built with threads, so you don't need to load it 16:55:20 nikodemus: ok, just done that 16:55:22 -!- nikodemus [n=nikodemu@cs181229041.pp.htv.fi] has quit ["This computer has gone to sleep"] 16:55:36 so now I need to require and use-package :sb-bsd-sockets anywhere I want to use it ? 16:57:20 galdor: as nikodemus just told you, to load the relevant system, :depends-on in the system definition will do. But :use is a different animal altogether because it has to do with what symbols you import to a package. You should be very careful using it, and consider explicitly prefixing symbols instead. 16:59:24 tritchey [n=tritchey@c-68-58-88-241.hsd1.in.comcast.net] has joined #lisp 16:59:33 -!- chris2 [n=chris@dslb-094-216-067-251.pools.arcor-ip.net] has quit [Operation timed out] 17:00:02 beach: appart from name conflicts risks, is there a reason not to use it ? 17:00:20 I mean typing things like sb-bsd-sockets:sockopt-reuse-address is really horrible 17:00:25 chris2 [n=chris@dslb-094-216-067-251.pools.arcor-ip.net] has joined #lisp 17:01:15 galdor: Your system might not compile in the future when new exported symbols are added to the package you :use. But if you are willing to take that risk, go ahead. 17:01:39 is there a way to explicitely import a list of symbols ? 17:01:57 like use Foo::Bar (sym1, sym2) in perl for example 17:02:05 clhs import 17:02:05 http://www.lispworks.com/reference/HyperSpec/Body/f_import.htm 17:02:56 ah cool, thank you! 17:03:01 there is also the :import-from keyword to defpackage. 17:03:07 -!- r1nu- [n=r1nu-@ppp-94-68-16-210.home.otenet.gr] has quit [] 17:04:55 -!- sreeram [n=sreeram@122.174.73.39] has quit [] 17:08:34 Argh. SBCL hackers: can you think of a way to check the sanity of SB-DISASSEM's pattern database early during the build process? 17:09:10 -!- schoppenhauer [n=schoppen@unaffiliated/schoppenhauer] has quit [Read error: 113 (No route to host)] 17:10:04 -!- mcspiff [n=user@drmons0501w-142177079167.pppoe-dynamic.ns.aliant.net] has quit [Remote closed the connection] 17:10:37 HG` [n=wells@xdslhh195.osnanet.de] has joined #lisp 17:12:38 rdd [n=user@c83-250-152-109.bredband.comhem.se] has joined #lisp 17:17:16 -!- DrForr [n=drforr@eldwist.darkuncle.net] has quit [Remote closed the connection] 17:18:09 ken-p [n=ken-p@84.92.70.37] has joined #lisp 17:18:34 -!- HG` [n=wells@xdslhh195.osnanet.de] has quit [Client Quit] 17:20:19 -!- reaver__ [n=reaver@h15a2.n2.ips.mtn.co.ug] has quit [Read error: 110 (Connection timed out)] 17:22:48 X-Scale2 [i=email@89-180-180-54.net.novis.pt] has joined #lisp 17:25:34 nikodemus [n=nikodemu@cs181229041.pp.htv.fi] has joined #lisp 17:27:03 -!- X-Scale [i=email@89.180.229.68] has quit [Nick collision from services.] 17:27:22 Quadrescence [n=quad@unaffiliated/quadrescence] has joined #lisp 17:27:32 -!- X-Scale2 is now known as X-Scale 17:29:47 ejs [n=eugen@223-168-124-91.pool.ukrtel.net] has joined #lisp 17:32:26 -!- nikodemus [n=nikodemu@cs181229041.pp.htv.fi] has quit ["This computer has gone to sleep"] 17:33:00 -!- ejs [n=eugen@223-168-124-91.pool.ukrtel.net] has quit [Client Quit] 17:33:39 -!- borsman [n=quassel@76.177.217.216] has quit [Read error: 104 (Connection reset by peer)] 17:35:31 -!- tritchey [n=tritchey@c-68-58-88-241.hsd1.in.comcast.net] has quit [] 17:36:56 -!- projections [n=projecti@88.235.101.2] has left #lisp 17:37:32 -!- jsoft [n=Administ@unaffiliated/jsoft] has quit [Connection reset by peer] 17:38:17 -!- jsoft_ [n=Administ@unaffiliated/jsoft] has quit [Read error: 104 (Connection reset by peer)] 17:38:34 -!- tcr [n=tcr@host145.natpool.mwn.de] has quit ["Leaving."] 17:39:45 -!- chris2 [n=chris@dslb-094-216-067-251.pools.arcor-ip.net] has quit [Read error: 131 (Connection reset by peer)] 17:40:21 jophish [n=jophish@dial-80-47-3-226.access.uk.tiscali.com] has joined #lisp 17:40:29 chris2 [n=chris@dslb-094-216-067-251.pools.arcor-ip.net] has joined #lisp 17:41:10 -!- LostMonarch [n=roby@host215-20-dynamic.48-82-r.retail.telecomitalia.it] has quit ["raise RuntimeError"] 17:43:45 -!- chris2 [n=chris@dslb-094-216-067-251.pools.arcor-ip.net] has quit [SendQ exceeded] 17:44:29 chris2 [n=chris@dslb-094-216-067-251.pools.arcor-ip.net] has joined #lisp 17:48:36 -!- joachifm [n=joachim@bjo1-1x-dhcp388.studby.uio.no] has quit [Client Quit] 17:49:53 -!- shmho [n=user@58.142.15.103] has quit [Remote closed the connection] 17:50:31 -!- benny99 [n=benny@p5486E4E6.dip.t-dialin.net] has quit ["Leaving"] 17:51:06 -!- Buganini [n=buganini@security-hole.info] has quit [Read error: 113 (No route to host)] 17:54:26 araujo [n=araujo@gentoo/developer/araujo] has joined #lisp 17:57:18 -!- Davidbrcz [n=david@nsc.ciup.fr] has quit ["Ex-Chat"] 17:59:29 is it possible to portably get info about alignment of certain types? 18:00:07 no 18:00:26 Athas [n=athas@0x50a157d6.alb2nxx15.dynamic.dsl.tele.dk] has joined #lisp 18:03:09 I'm using SBCL with Slime. Is it possible to create a thread with its own REPL? 18:04:37 jsoft [n=Administ@unaffiliated/jsoft] has joined #lisp 18:07:00 ruediger_: open another connection? 18:07:16 -!- nvoorhies [n=nvoorhie@adsl-75-36-204-63.dsl.pltn13.sbcglobal.net] has quit [Read error: 104 (Connection reset by peer)] 18:07:27 attila_lendvai: you mean I should create a new swank server in the thread? 18:07:34 ruediger_: If it isn't, perhaps you would be interested in the CLIM listener. 18:07:48 nvoorhies [n=nvoorhie@adsl-75-36-204-63.dsl.pltn13.sbcglobal.net] has joined #lisp 18:08:11 ruediger_: i mean M-x slime-connect again to the same port, then the usual slime tools to switch between connections -- if it works 18:09:53 -!- nvoorhies [n=nvoorhie@adsl-75-36-204-63.dsl.pltn13.sbcglobal.net] has quit [Read error: 104 (Connection reset by peer)] 18:09:55 nvoorhies_ [n=nvoorhie@adsl-75-36-204-63.dsl.pltn13.sbcglobal.net] has joined #lisp 18:12:53 -!- mehrheit [n=user@lan-84-240-55-160.vln.skynet.lt] has quit [Read error: 113 (No route to host)] 18:13:35 pkhuong: So how big FFTs do you typically see? 18:14:01 Fast Fourier Transform? 18:14:14 yes jthing! 18:14:23 *_3b* has seen lots of 8x8 ones, but I'm assuming that isn't your target range :) 18:14:37 schoppenhauer [n=schoppen@unaffiliated/schoppenhauer] has joined #lisp 18:15:11 _3b: Well, for analyzing sound, we can easily do 1024 or more 18:16:04 I'm essentially interested in knowing whether there is a domain that uses sizes larger than the size of the cache. 18:16:16 [and whether we care about those domains] 18:17:56 beah, the Intel core duo has a 4096 kb cache and is more optimal for Lisp if that is what you mean... 18:18:34 beach: larger than L2? Definitely. I'm told the mersenne prime search does FFTs on the order of 33M elements ;) 18:18:37 jthing: No, as I said, I am interested in knowing how big FFTs people do. 18:19:01 pkhuong: OK, so they have a different problem :) 18:19:24 -!- ikki [n=ikki@189.228.133.213] has quit ["Leaving"] 18:19:40 pkhuong: so it does 18:20:51 *beach* now thinks he will just target applications related to sound. 18:21:16 postamar [n=postamar@x-132-204-253-5.xtpr.umontreal.ca] has joined #lisp 18:21:25 -!- lemoinem [n=swoog@modemcable110.189-201-24.mc.videotron.ca] has quit [Read error: 113 (No route to host)] 18:22:08 right, I'm told large FFTs are common in other non-signal-processing tasks (or for non-realtime convolutions). 18:23:42 the AFG library perhaps? 18:24:09 beach: are you working on http://common-lisp.net/project/cl-wav-synth/ ? 18:25:24 lemoinem [n=swoog@modemcable110.189-201-24.mc.videotron.ca] has joined #lisp 18:26:09 chris2_ [n=chris@dslb-094-216-067-251.pools.arcor-ip.net] has joined #lisp 18:26:51 -!- legumbre_ is now known as legumbre 18:28:54 michaelw: Nope. But it's a nice project. 18:30:19 dwave [n=ask@212.251.243.4] has joined #lisp 18:34:41 pkhuong: I don't know whether you read the scrollbacks, but it seems to me that if your FFTs are significantly smaller than the cache, you should precompute "bit reversal" algorithms, and e^-2pik/N coefficients. If not, getting a cache miss when trying to access them, might be more expensive than re-computing them. 18:35:56 michaelw: My ambition is to have the fastest FFT around, that it be written in Lisp, and that it be distributed with major implementations of Common Lisp. :) 18:36:56 -!- rjack [n=rjack@adsl-ull-234-19.51-151.net24.it] has quit ["leaving"] 18:37:02 Good luck, but my bet would be on FORTRAN 18:39:00 jthing: false. Try FFTW. works with generated code, where CL should excel 18:39:24 ahh 18:40:30 FFTW is OCaml generating C, afaik 18:40:42 beach: I'd probably start by reading the papers linked at 18:41:37 -!- chris2 [n=chris@dslb-094-216-067-251.pools.arcor-ip.net] has quit [Connection timed out] 18:42:17 beach: make something faster than FFTW with dynamic generation of code and name it FFTE ;-) 18:43:14 SBCL is good, but not that good.. Sigh 18:44:00 tsuru [n=user@c-69-245-36-64.hsd1.tn.comcast.net] has joined #lisp 18:44:43 michaelw: I read those a long time ago. Also, you should know that I have the world record of bit-reversal algorithms, beating that of my co-author by a factor 5. 18:45:20 p_l: The name, I am not sure, but you guessed the rest. 18:45:26 jthing: yes, SBCL is good. Just now, look what it says to me: WARNING: Ignoring unknown optimization quality SANITY in: (OPTIMIZE (SPEED 3) (SPACE 2) (DEBUG 0) (SANITY 0)) 18:45:27 For one thing it lacks a peephole optimizer. 18:45:53 beach: cheers. Though I don't know where bit reversal would be useful in FFT. 18:45:55 pjb: never seen that one 18:47:30 -!- Schmidt [i=lsc@c193-150-248-188.bredband.comhem.se] has quit [Remote closed the connection] 18:47:38 *_3b* votes for just dumping the FFTs on a GPU :) 18:47:40 michaelw: If you reorganize the elements initially, you won't have to use temporary space. Bit-reversal is the name of some algorithms that exchange elements of indexes between some index i and the bit-reversal of i. 18:48:03 bombshelter13_ [n=bombshel@209-161-234-188.dsl.look.ca] has joined #lisp 18:48:27 michaelw: Turns out this only works for radix 2, so we need to do some research for other values of the radix, not to mention variable-size radix. 18:49:09 jsoft_ [n=Administ@unaffiliated/jsoft] has joined #lisp 18:49:15 -!- DarkRavin [n=thedarkr@adsl-147-121-66.bhm.bellsouth.net] has quit ["Ex-Chat"] 18:50:09 jthing: perhaps I shouldn't throw the stone, but sometimes you talk too much, in particular about this of which you have no idea. 18:50:48 s/this/things/ 18:50:50 I studied physics beach 18:51:04 jthing: notice I typed sanity instead of safety ;-) 18:51:06 oh, so you *must* be smart then. 18:51:32 jthing: but the effect is that sbcl refuses to compile a program without sanity. 18:51:50 whatever sorry if I insulted you 18:52:08 jthing: I deal with people who studied physics every day (or close to it), and my main problem with them is that they think that they are smarter than everyone else. 18:53:18 beach: of course, they explain chemistry, and therefore biology. But we CS guys are even better: we explain physics! :-) 18:53:29 jthing: In fact, now that I think about it, that might be your problem as well, i.e., that you cannot possibly think that there are smart people out there who did *not* study physics. 18:53:54 "I studied physics beach" - apropos of nothing 18:54:23 beach: In fairness, people who think they are smarter than everybody else are by no means confined to physics :) 18:54:46 rsynnott: There are smart people in physics, like you and Xof, but there are quite a number of assholes as well. 18:55:08 I am soory you feel so strongly about this beach 18:55:13 me? I'm not in physics 18:55:17 rsynnott: That might be true. 18:55:18 or smart ;) 18:56:16 jthing: don't feel sorry about it. It's just a natural reaction of one with more than 20 years of experience trying to deal with it. 18:56:48 In my mind this is not a superiority contest. On that you can take my word! beach 18:56:50 rsynnott: Sorry, misunderstood. 18:57:12 DarkRavin [n=thedarkr@adsl-147-121-66.bhm.bellsouth.net] has joined #lisp 18:57:33 jthing: Then please talk only when you know what you are talking about. 18:57:49 will do 19:00:24 jthing: Thanks! In addition, you are a Scandinavian, and they are *very* attached to doing the right thing. Norwegians less so than the Swedes perhaps. I don't know. 19:02:16 me neither :) 19:02:50 -!- dfox [n=dfox@r11jn246.net.upc.cz] has quit [Read error: 104 (Connection reset by peer)] 19:02:57 |stern| [n=seelenqu@pD9E476AB.dip.t-dialin.net] has joined #lisp 19:03:36 oh, cool, wolfram alpha launched 19:04:13 beach: unrolling the recursion + scheduling should get that implicitly. 19:04:35 rsynnott: what's that? 19:04:57 pjb: I wouldn't put it that way, but what I have seen recently is that CS is taking over other domains because they run out of methods and options (what? the eigenvalues are *not* independent?, well, then, uh, I don't know what to do). At that point, it takes a hacker (i.e., a CS guy) to make things progress. Unfortunatly for other domains, this mostly happens in CS. 19:05:16 pkhuong: Which part? 19:05:51 kmcorbett [n=Keith@c-76-119-215-57.hsd1.ma.comcast.net] has joined #lisp 19:05:59 jthing: http://www.wolframalpha.com/ - mathematica on the web, sort of 19:06:09 thanks 19:07:05 -!- chris2_ [n=chris@dslb-094-216-067-251.pools.arcor-ip.net] has quit ["Leaving"] 19:07:13 slash_ [n=Unknown@p4FF0B027.dip.t-dialin.net] has joined #lisp 19:07:19 -!- jsoft [n=Administ@unaffiliated/jsoft] has quit [Read error: 110 (Connection timed out)] 19:07:32 chris2 [n=chris@dslb-094-216-067-251.pools.arcor-ip.net] has joined #lisp 19:08:06 rsynnott: yeah but it's not as cool as I hoped 19:08:28 it doesn't understand most of my queries 19:08:35 -!- dwave [n=ask@212.251.243.4] has quit ["Be back later"] 19:08:50 and some are plain wrong. Like "conductivity of Si at 0K" 19:10:36 beach: precomputing e^2pik/n at just the right granularity 19:11:35 rudi [n=rudi@84-119-52-95.dynamic.xdsl-line.inode.at] has joined #lisp 19:11:40 ruediger_: can we reach 0K? 19:11:53 we can be pretty close, I guess 19:12:12 but real absolute zero would probably require freezing time ;p 19:12:14 -!- antoszka [n=antoszka@unaffiliated/antoszka] has quit ["+++ killed by SIGSEGV +++"] 19:12:18 madnificent: no. But "conductivity of Si at 0.1K" is still plain wrong 19:12:26 or even 10K 19:12:28 p_l: close is meaningless (as it is relative) 19:12:36 ruediger_: ok, those are fair questions 19:12:52 it gives always the conductivity of 300K afair 19:13:02 ruediger_: true, thanks for pointing that out 19:13:18 -!- envi^home [n=envi@220.121.234.156] has quit [Read error: 54 (Connection reset by peer)] 19:13:22 silly knowledge machine 19:13:27 -!- jophish [n=jophish@dial-80-47-3-226.access.uk.tiscali.com] has quit [Connection timed out] 19:13:51 actually I have a split feeling about wolfram alpha. In some way it is really really cool 19:14:06 but sometimes it's so disappointing 19:14:24 e.g. you type in op amp and you get a nice mini simulation 19:14:34 but on the other hand it trades conductivity as a constant 19:15:13 (I mean not understanding every question or not having every data available at the moment is understandable) 19:15:21 pkhuong: I think I know *precisely* what you mean, and this is looking more and more like PhD topic, but you may have already suggested it of course. :) 19:15:30 at least the syntax is nicer than Mathematica's :D 19:16:45 I'm sure it will improve 19:16:51 TERRIBLY slow at the moment, though 19:17:14 -!- eno__ is now known as eno 19:17:33 -!- mgr [n=mgr@psychonaut.psychlotron.de] has quit [Read error: 113 (No route to host)] 19:18:51 pkhuong: That is *so* neat! At the top level you use a combination of precomputed values and some epsilon, and at lower level all the values are precomputed. But where do you draw the limit? Very impressive! 19:19:01 rsynnott: somebody probably typed in "DOES P=NP?" and it tries to proof that now ;) http://abstrusegoose.com/strips/how_to_solve_it.PNG 19:19:09 -!- seelenquell [n=seelenqu@pD9E45779.dip.t-dialin.net] has quit [Read error: 101 (Network is unreachable)] 19:19:18 I believe that's what FFTW does. 19:19:27 rlol 19:19:35 someone should input the answer from Concrete Maths for such case 19:19:40 hear that beach 19:20:37 pkhuong: Well, I guess that's not surprising, given that it's their domain of research, and it took me a few minutes to think of it. 19:21:29 desu [i=desu@unaffiliated/desu] has joined #lisp 19:22:43 I'm hoping to find a way to model the problem and take into account the size of the register file, SIMD and the desire to minimise the number of (concurrent) memory access streams. It would then become a simple matter of shoving that in CPLEX and waiting a couple hours ;) 19:23:20 "Wolfram|Alpha isn't sure what to do with your input." on my first question... :-/ 19:23:30 pkhuong: You want CPLEX? My wife works in that team. 19:23:39 pkhuong: I guess you'd also need to model how you call stuff in it 19:24:19 legumbre_ [n=user@r190-135-29-126.dialup.adsl.anteldata.net.uy] has joined #lisp 19:24:19 beach: I'm in a logistics and transport research center; we have a couple licenses for now (: 19:24:34 p_l: you don't call stuff in kernels. 19:24:43 pkhuong: I am offering you access to the dev team. 19:25:31 pkhuong: but I suppose there are differences between jmp and call (or GO and PUSHGO) that might useful :) 19:26:20 beach: Hopefully that won't be needed. I like my boxes black! 19:27:27 pjb: get used to it :) 19:27:57 mgr [n=mgr@psychonaut.psychlotron.de] has joined #lisp 19:27:58 -!- jfactor [n=jfactor@student167-17.hampshire.edu] has quit [Read error: 110 (Connection timed out)] 19:28:05 pkhuong: I can appreciate that. I just thought I would like you to know that my wife is in the CPLEX team, and that *might* help, though probably less now that Ilog has been bought by IBM. 19:28:12 It's not my day, sbcl refuses to compile wit (sanity 0), wolfram doesn't get my questions... 19:28:12 -!- legumbre [n=user@r190-135-29-63.dialup.adsl.anteldata.net.uy] has quit [Read error: 60 (Operation timed out)] 19:29:47 pjb: you mean (safety 0) ? 19:30:10 Yes :-) 19:30:18 -!- postamar [n=postamar@x-132-204-253-5.xtpr.umontreal.ca] has left #lisp 19:30:23 I had a lapsus. 19:30:33 Wolfram|Alpha ne comprend pas encore le français. 19:30:55 stassats` [n=stassats@wikipedia/stassats] has joined #lisp 19:31:18 antoszka [n=antoszka@unaffiliated/antoszka] has joined #lisp 19:31:32 That sucks. 19:31:43 -!- stassats [n=stassats@wikipedia/stassats] has quit [Read error: 104 (Connection reset by peer)] 19:31:48 ruediger_: that comic was really good 19:31:58 beach: don't try common lisp or lisp then :-( 19:33:04 ich spechens neithther froggy not kraut ;) 19:33:24 nor 19:34:13 (french or germain) 19:35:10 vy [n=user@88.229.232.85] has joined #lisp 19:35:14 pjb: My memory is getting worse and worse, so you need to keep the context clear when talking to me. Sorry about that. 19:35:20 jthing: spreche not spechens 19:35:38 -!- rudi [n=rudi@84-119-52-95.dynamic.xdsl-line.inode.at] has quit ["Client exciting"] 19:35:49 ruediger_: sorry 19:36:00 beach: Wolfram|Alpha ne comprend pas encore le français. That sucks. beach: don't try common lisp or lisp then :-( 19:36:59 beach: looking forward to the summer vacation, are you? 19:37:12 pjb: I can't say I care about Wolfram, other than their great site for various math problems. 19:37:36 beach: do you say that because they don't use Lisp? 19:37:40 some nice statistics and graphs, too 19:37:59 actually said that Macsyma took such a turn for the worse 19:38:04 madnificent: You bet! But partially for reasons you can't imagine! 19:38:17 *madnificent* wondersr 19:38:33 -!- antoszka [n=antoszka@unaffiliated/antoszka] has quit ["+++ killed by SIGSEGV +++"] 19:38:57 if the universe would be perfect we'd all use Symbolics Alpha introduced 10 years ago and really working ;) 19:39:26 -!- kleppari [n=spa@bitbucket.is] has quit ["Lost terminal"] 19:39:47 pjb: I think we have an innocent but simple misunderstanding here. 19:40:31 ruediger_: Sounds right! :) 19:40:46 ruediger_: as in Symbolics for the now effectively non-existent alpha chip, or a symbolics version of Wolfram Alpha? 19:41:06 Wolfram Alpha 19:41:20 but why not both? ;) 19:41:30 -!- legumbre_ is now known as legumbre 19:41:36 well, Itanium ;) 19:41:55 (Intel's Itanium partners were very enthusiastic to kill their own chips as soon as possible, sadly) 19:41:56 Yeah if if it worked we would all use it (in the future) and thus back date it ;) 19:42:30 kleppari [n=spa@bitbucket.is] has joined #lisp 19:43:21 rsynnott: actually I think the Itanium is a good idea 19:43:42 just intel was too careful and thus allowed amd to introduce this crappy x86/64 stuff 19:43:45 sorry for the sarcasm.. 19:44:30 Killed were PA-RISC and Alpha, afaik... and Compaq was in intel's pocket since the company was created afaik 19:44:31 ruediger_: The real problem was the 800 series 19:44:45 In principle, certainly, but they don't seem to have done a terribly good job developing the itanium 19:44:57 (and they are clearly prioritising the x86-64) 19:45:23 rsynnott: they had to pull people out of Itanium to not completely lose in x86-64 market 19:45:31 yeah 19:45:38 any sbcl binaries that run on tiger ? 19:45:40 or else amd would've overrun them 19:46:05 lisp [n=djdouma@rrcs-24-43-85-58.west.biz.rr.com] has joined #lisp 19:46:19 I think Intel's mistake was not pushing the ia64 into the desktop market. They could have easily bundled it with a x86 cpu 19:46:26 -!- lisp [n=djdouma@rrcs-24-43-85-58.west.biz.rr.com] has left #lisp 19:46:33 and I guess a part of this is Intel being uppity and trying to decide what market wants (they didn't intend to give any consumer 64bit system for next few years) 19:46:47 so now we are stuck with the x86/64 shit 19:46:49 p_l: yeah 19:47:01 ruediger_: From the 900 series onward they hvae been doing ok (better SPECS than AMD.) 19:47:06 have 19:47:19 djdouma [n=djdouma@rrcs-24-43-85-58.west.biz.rr.com] has joined #lisp 19:47:44 ruediger_: Microsoft was terribly unhelpful there 19:47:52 *p_l* recalls some info on how SPEC is not a good way to compare CPUs anymore due to chipmakers having special optimizing modes for it 19:48:07 sad0ur [n=sad0ur@r2bk213.net.upc.cz] has joined #lisp 19:48:21 dfox [n=dfox@r11jn246.net.upc.cz] has joined #lisp 19:48:32 rsynnott: Intel not introducing anything like DEC's recompiler also lowered IA64's chances. 19:48:37 even on servers, Windows Itanium support is relatively limited 19:48:41 clynbech [n=user@cpe.atm2-0-1071129.arcnxx16.customer.tele.dk] has joined #lisp 19:48:50 -!- bombshelter13_ [n=bombshel@209-161-234-188.dsl.look.ca] has quit [Read error: 104 (Connection reset by peer)] 19:49:00 rsynnott: that's why the should have bundled the IA64 with a x86 19:49:22 ruediger_: and ensure no-one uses ia64? ;P 19:49:25 rotfl 19:49:32 no :p 19:49:44 the hw x86 emulator was a flop 19:49:47 apps requiring more than 4gb ram would use ia64 19:50:13 it may take off yet 19:50:20 I doubt it 19:50:23 Seriosly the Core Duo chip works well with Lisp 19:50:25 ruediger_: I suppose such scheme would have lower acceptance than NT/alpha 19:50:29 if they can ever manage to make it competitive against Power6/7 19:50:41 (basically a UNIX market anyway) 19:50:47 rsynnott: they'd have to change the ISA seriously to do that, I think 19:50:53 gigamonk` [n=user@adsl-99-17-207-207.dsl.pltn13.sbcglobal.net] has joined #lisp 19:51:07 (the end of Sun may help them with this, if Oracle kills Rock) 19:51:09 there's a problem with integer performance basically by design 19:51:25 It's the parelell execution of the smithsfield aritecture which is fucked 19:51:42 too long pipelines 19:52:12 true. But afaik IA64 for some reason has fucked up integer performance compared to floating-point 19:52:21 -!- clynbech [n=user@cpe.atm2-0-1071129.arcnxx16.customer.tele.dk] has quit [Client Quit] 19:52:46 with x86_64 it is up to the software developers to make their apps scale. with the ia64 it was up to the hardware 19:52:46 and most software doesn't use so much floating-point 19:52:50 -!- mattrepl [n=mattrepl@ip68-100-82-124.dc.dc.cox.net] has quit [] 19:53:02 -was +is 19:53:42 interesting really that at a time when it is easier to move between architectures than ever before, there are onl about four serious ones left 19:54:11 ruediger_: can you elaborate on that? (sw devs make their apps scale) 19:54:37 klausi [n=klausi@port-92-193-7-18.dynamic.qsc.de] has joined #lisp 19:55:16 madnificent: moore's law gives us a steadily increasing number of transistors 19:55:16 rsynnott: four? 19:55:34 and on IA64 it is possible to use them in a way that is transparent to the software 19:55:40 like adding a new integer unit 19:55:49 x86, POWER, SPARC and ARM 19:56:01 gotcha 19:56:04 and thus doing the calculations in parallel without the software developer having to change one thin 19:56:25 ruediger_: thank you :) 19:56:27 ruediger_: except writing an AGI to actually optimize it properly ;-) 19:56:30 nyef [n=nyef@vcwl1-61.daktel.net] has joined #lisp 19:56:31 rsynnott: ARM is in another segment, and SPARC is just about dead. 19:56:54 Hello all. 19:56:54 SPARC isn't dead yet, not in server market, and MIPS is getting back in HPC 19:56:57 but on x86_64 we can't scale that way. So the increasing density of transistors is used to add just more cores 19:57:11 ruediger_: We're talking about the same IA64 with VLIW and little/no speculation/reordering, right? 19:57:14 and now it is up to the developer to use the the cores 19:57:26 ruediger_: ILP can't be exploited too long 19:57:31 pkhuong: I'm more talking about the general concept 19:57:44 I think it actually dates back to the HP-RISC. 19:57:47 ruediger_: They all add more cores 19:57:49 the whole prefixing thing 19:58:02 -!- djdouma [n=djdouma@rrcs-24-43-85-58.west.biz.rr.com] has quit [] 19:58:08 pkhuong: well, in any sectors (to an extent ARM, POWER and even x86 compete in embedded market) 19:58:28 including SPACK (16 last I checked) 19:58:29 The same concept of trusting the *compiler* to find ILP? If anything, x86 is much better at transparently using additional units. 19:58:34 I mean writing a compiler for x86 is hard because the instruction set doesn't really represent what's going on in the hardware 19:58:38 POWER, x86 and SPARC compete on servers, and x86 has no competition on desktop/workstation 19:58:57 ruediger_: no, writing a compiler for IA64 is hard. 19:59:02 ruediger_: I'm a little surprised that Intel hasn't allowed developers to access the underlying hardware 19:59:05 ARM and MIPS and PPC compete in embedded 19:59:15 and thus tricks working on one model will break performance on a different model 19:59:20 rsynnott: AMD did, no one used it 19:59:26 rsynnott: AMD did once. Nobody used it. 19:59:27 pkhuong: the compiler has to do an amount of scheduling itself, no/ 19:59:29 ah 19:59:36 p_l: ^5! 19:59:40 ruediger_: that's true 19:59:44 (on ia64, i mean) 19:59:48 AMD did? What chipsets was that for? 19:59:56 nyef: K6 afaik 20:00:01 and possibly K5 20:00:18 and on IA64 the whole "intelligence" is in the compiler 20:00:18 (oh, that's a bit confused; I meant for x86 with the underlying hardware access, but IA64 for the scheduling by compiler) 20:00:18 So... nothing I'm actually likely to have handy. 20:00:32 pkhuong: yeah it's probably harder because you have to do more stuff 20:00:35 K6 was based on NextGen stuff and there was possibility to disable hw rewriter in decoder 20:00:58 but I guess it's just harder because most of the research in that area is missing, yet 20:01:04 p_l: bit of an all-or-nothing deal, then? 20:01:14 also, Transmeta 256bit VLIW designs had documented internal CPU 20:01:30 rsynnott: it could be probably switched at least on per app basis 20:01:37 rsynnott: it's little known part 20:01:46 no one cared to use it 20:02:06 yep, I never even heard of it before 20:02:10 probably because transmeta wasn't that popular 20:02:19 hadn't realised developers were allowed use transmeta directly either 20:02:38 btw, I have SPECint results for first gen Itanium @800MHz compared to 1GHz 21264 and 2GHz Pentium 4... 20:02:53 ruedige: they don't mostly 20:02:59 and SPECfp 20:03:17 p_l: did the itanium loose? 20:03:53 ruediger_: If your target market uses mostly integer operations, choosing Itanium is a fucked-up idea ;P 20:04:05 :) 20:04:15 of course, most stuff IS integer-centric 20:04:17 true 20:04:23 it had around 68% of Pentium 4 performance 20:04:28 and that's of the old Pentium 4 20:04:29 p_l: but actually it could scale towards integer operations. By adding more integer units it should execute them faster 20:04:35 (if they don't depend on each other) 20:04:49 More like: if you're doing anything but FP arithmetic in perfectly nested loops, don't go with Itanium. 20:05:23 ruediger_: yes, but then stuff like amount of transistors (and size problems related to it) as well as power usage start to get in 20:05:34 ruediger_: not more so than x86-64! The ISA restricts the number and types of work packets that may be combined. 20:05:39 it's easier to extend 21264 with another I-box than IA-64 :P 20:05:53 amount of transistors/cm^2 is increasing exponentially :) 20:06:16 ruediger_: and effective lightspeed is getting slower, resulting in cpus having problems 20:06:21 sigh 20:06:37 at least that was explanation I had seen :) 20:06:51 p_l: yeah but that's why intel seems to put a lot of effort into photonics 20:06:58 (and I probably used wrong terms...) 20:07:10 You are all stuck with your old X86 based processors 20:07:19 ruediger_: that's for interconnect. 20:07:23 ruediger_: You mean, they're trying to set up a photonic boom? 20:07:25 *p_l* whacks jthing with his Alpha 20:07:33 nyef: boom? 20:07:43 *_3b* has a nice nvidia processor to go with the intel 20:07:47 jthing: what processor do you use? 20:08:09 _3b: any idea when larrabee will be available on shelves? 20:08:13 (actually I'm sitting next to a PowerPC :-) and a HP-RISC machine *g*) 20:08:24 (A play on "sonic boom", but also used for its other meaning of a period of strong economic growth.) 20:08:27 ruediger_: A 940 from Intel 20:08:33 <_3b> pkhuong: no idea, i don't follow the hardware closely unless i can afford to buy something :) 20:08:39 Jabberwockey [n=jens@port-7665.pppoe.wtnet.de] has joined #lisp 20:08:57 jthing: what's that? an Atom? 20:09:15 isn't that the weird old Intel RISC chip? 20:09:16 *p_l* needs to get a newer Alpha. Plus a SPARC. And some FPGAs to learn VHDL so he can implement his MMIX variant 20:09:18 ruediger_: No A Dell 20:09:34 rsynnott: Atom is actually the only pure x86 cpu on market 20:09:34 nyef: oh ok :) 20:09:43 (oh, that's an i960) 20:09:52 <_3b> 940 as in core i7? 20:10:01 no 20:10:06 right 20:10:06 i960 is pretty old iirc 20:10:11 ruediger_: yep 20:10:14 VERY old 20:10:15 p_l: what about the geode? 20:10:21 early 90s stuff 20:10:38 and via has some x86 cpu's too 20:10:40 madnificent: Geode isn't used in desktops/laptops, though 20:10:52 ruediger_: that would be the 800 series 20:10:54 but that's the sad fact about intel. Every couple of years they bring out a nice new architecture. But they fail to really implement it into the market 20:11:04 madnificent: And I'm not sure about Via not using a microcoded RISC 20:11:13 I'll take amd64 over IA64 any day. 20:11:17 ruediger_: totally sucks 20:12:05 as is the atom... but counting netbooks and the likes in does yield some via's and (I think) one geode (if not now, then in the near future) 20:12:32 geode is probably microcoded risc as well, though, no? 20:12:48 p_l: I wonder about the via cpu being a hidden RISC cpu though :) 20:13:03 amd plans on making a low power cpu in it's fusion line 20:13:21 and apparently Fusion got speeded up 20:14:06 the Intel 900 series had that goal in mind p_I 20:14:40 900-series as in what exactly? I'm getting lost 20:15:39 http://www.2cpu.com/articles/106_1.html 20:15:49 ignas [n=ignas@ctv-79-132-160-221.vinita.lt] has joined #lisp 20:16:45 -!- gigamonk` [n=user@adsl-99-17-207-207.dsl.pltn13.sbcglobal.net] has quit [Remote closed the connection] 20:16:51 gigamonk` [n=user@adsl-99-17-207-207.dsl.pltn13.sbcglobal.net] has joined #lisp 20:18:30 well, I'm more of AMD fan (the fact that I currently sit on intel system is more of university choice...). 20:19:04 -!- Jabberwockey [n=jens@port-7665.pppoe.wtnet.de] has quit ["Verlassend"] 20:19:07 lol 20:19:11 *rsynnott* has never quite understood brand loyalty in processors :) 20:19:38 -!- dfox [n=dfox@r11jn246.net.upc.cz] has quit [Remote closed the connection] 20:19:50 rsynnott: it wasn't really brand loyalty. It's just that after Pentium 4 it took i7 to reestablish any belief in competent design at intel ;-) 20:20:05 dfox [n=dfox@r11jn246.net.upc.cz] has joined #lisp 20:20:29 p4 was a bit of a disaster all right :) 20:20:37 that, and finding various weird things as to how the "competition" worked on cpu market 20:20:44 Intel Core Duo has pretty good design 20:21:04 jthing: ironically, more a modernised p3 than anything else 20:21:31 (the p3 was deliberately killed in the early days of the p4, incidentally, as dangerous internal competition) 20:21:33 Indeed Intel has some truly shitty process ors (like the 820) 20:21:37 jthing: The last time I was buying a computer where I did choose the CPU, Core2 was a) overpriced b) slower in 64bit mode than my chosen cpu 20:21:43 processors 20:22:04 -!- SandGorgon [n=user@122.162.135.206] has quit [Read error: 60 (Operation timed out)] 20:22:15 and I knew from beginning that 99% of code I'm going to run is 64bit :) 20:22:15 antoszka [n=antoszka@unaffiliated/antoszka] has joined #lisp 20:22:37 (of course, since I'm an apple user I'm basically at the mercy of whatever Apple decides to use...) 20:22:54 P_I: the point is that Core Duo runs Lisp wery well 20:23:00 rsynnott: *sniff* powerpc *sniff* 20:23:30 ruediger_: well, IBM was unhelpful there 20:23:52 does x86-64 has tagged arithmetic instructions? 20:24:07 (I liked the way that they huffily announced a low-power laptop-suitable G5 a few days after Apple announced its Intel transition) 20:24:12 rsynnott: I don't know what happened in the background 20:24:19 Don't think so 20:24:28 rsynnott: 20:24:51 hmm, was it ARM or Sparc with tagged arithmetic instructions? 20:25:02 sparc, probably. 20:25:35 I don't remember seeing any when I did a set of ARM instruction definitions for an assembler. 20:25:46 sparc definately 20:26:09 disumu [n=disumu@p54BCB65A.dip.t-dialin.net] has joined #lisp 20:26:13 -!- jsoft_ [n=Administ@unaffiliated/jsoft] has quit ["zzz"] 20:26:49 nyef: could it be a start of sbcl port to arm? ;-) 20:27:04 yeah that'd be great 20:27:07 sparc has support for tags in the 2 low order bits. 20:27:26 I HATE ARM 20:27:38 sorry 20:27:42 hrhr 20:27:45 p_l: Yes, it was. Been thinking a bit about picking that up again. 20:27:55 minion: arm port log? 20:27:55 arm port log: http://www.lisphacker.com/projects/sbcl-arm/port-log.txt 20:28:06 tritchey [n=tritchey@c-68-58-88-241.hsd1.in.comcast.net] has joined #lisp 20:28:08 good to see I wasn't the only one thinking of it :) 20:28:18 jthing: dangerous attitude 20:28:37 MIPS64 would also be nice 20:28:39 (given that everyone in the industrialised world uses at least one ARM every day :) ) 20:29:12 (ah, there's mips port already) 20:29:23 MIPS32 only, though, isn't it? 20:29:30 Still, should make things a bit easier. 20:29:34 yeah 20:29:36 with sbcl support for arm an open calculator based on maxima would be possible :) 20:29:41 Just, please, don't do what was done for Alpha. 20:29:41 Won't you need to do a lot of trickery to reduce memory usage? 20:29:47 rsynnott: Probably due to my time at Opera 20:29:57 it would be nice to see SBCL running on SiCortex ^_^ 20:30:13 ah, yes, I suppose they'd have been ARM-centric :) 20:30:24 nop 20:30:27 -!- schoppenhauer [n=schoppen@unaffiliated/schoppenhauer] has quit [Remote closed the connection] 20:30:46 But they do support ARM 20:31:14 well, it's their big platform, in practice, isn't it? 20:31:22 no 20:31:42 there must surely be more people using mobile opera than desktop opera 20:31:57 yes 20:32:51 I thought opera finances itself by licensing opera for arm 20:33:23 sorry ro dissapoint you 20:33:27 the desktop version seems very popular in Russia and eastern europe, but almost unused elsewhere 20:33:49 -!- gigamonk` [n=user@adsl-99-17-207-207.dsl.pltn13.sbcglobal.net] has quit [Read error: 110 (Connection timed out)] 20:33:52 (which is odd, really; it's quite a nice browser) 20:33:57 they are in fact financenced by 'google' 20:34:00 jthing: so what's the financial model of opera software? 20:34:09 by google? 20:34:15 yes 20:34:16 I used it on windows til Chrome came out 20:34:20 lol so google financed mozilla, opera and chrome 20:34:29 ruediger_: probably for setting default search to google, and so on 20:34:30 -financed +finances 20:34:37 rsynnott: hmm, ok 20:34:44 but is that enough to keep the company running? 20:34:54 I'd be surprised 20:34:56 ruediger_: only Opera that I am aware 20:35:12 Mozilla certainly receives considerable money from Google 20:35:16 ruediger_: The pay per click bit 20:35:45 mozilla gets the majority from Google 20:36:04 56.8 million in 2006 20:36:26 (and I heard that Nokia is focusing more on mozilla since Apple released the iPhone and they don't want to support Webkit anymore :)) 20:37:09 It keeps running :) 20:37:29 I wonder how too sometimes. 20:38:08 ruediger_: I doubt it, Nokia's version of WebKit is very well prepared and integrated nicely into symbian 20:38:32 Google also went with webkit for its mobile platform 20:38:43 rsynnott: and for their own browser 20:38:43 mozilla on phones has never been a great success 20:38:48 yep 20:38:49 so it did 20:39:07 but I heard that Nokia started supporting Mozilla especially Fennek 20:39:25 because they don't want to depend on Webkit due to it being controlled by Apple 20:39:33 but this could be a rumor of course 20:39:55 Opera has always done better on mobile phones 20:40:22 well, Apple's 'control' over webkit is lower these days 20:40:23 That's their other share of the profit 20:40:35 they're still the big contributor, but it is quite legitimately open 20:41:19 wasn't WebKit LGPL? 20:41:36 (Microsoft really created quite a few monsters when they killed MacOS IE and deprioritised win32 IE :) ) 20:41:37 yeah 20:41:43 p_l: khtml was, certainly 20:41:44 -!- alec [n=alec@pool-96-237-2-171.bstnma.east.verizon.net] has quit ["leaving"] 20:41:58 there's something weird about webkit license; significant parts are a BSD-like thing 20:42:12 rsynnott: rotfl 20:42:47 clynbech [n=user@cpe.atm2-0-1071129.arcnxx16.customer.tele.dk] has joined #lisp 20:42:47 rsynnott: no, it seems the genuine deel 20:43:03 wwthe funny thing about webkit is that nokia's webkit-based MiniMap browser was better at iPhone 1.0 "apps" than Safari on iPhone :P 20:43:28 I guess the biggest advantage of Firefox on the Desktop is actually that they provide such a flexible and dynamic platform. Which is probably the biggest disadvantage on Embedded systems 20:44:08 firefox has recently had much slower javascript than opera or webkit, too, which is quite a big deal for mobile 20:44:11 s 20:44:14 ruediger_: it's also pretty big disadvantage given how bloated it got 20:44:19 on desktops, that is 20:44:32 Actually Opera under linux is buggy, and that is the major problem 20:44:41 p_l: yeah but that's why they have so many addons 20:44:41 Mmm... I'm really not liking firefox 3 compared to firefox 2. 20:44:49 and due to the addons I stick with Firefox 20:45:05 nyef: ? ff3 is so much better than ff2 imho 20:45:17 *p_l* uses JIT JS and various tricks with optimizing to get enough speed out of FF... cause no other browser has the functionality I want 20:45:29 interestingly, the latest webkit does native javascript compilation 20:45:42 before ff3 I used to browse some stuff with Safari because it didn't had so much memory leaks. But since ff3 Safari3 is far worse :) 20:45:42 (using LLVM, apparently, though only in Intel) 20:45:52 ruediger_: Adding bookmarks, which is an operation I do with some frequency, takes far more keystrokes under ff3. 20:45:53 rsynnott: using llvm? cool 20:46:08 nyef: with del.icio.us addin it's just C-d :) 20:46:13 rsynnott: That would be the google thingiie 20:46:20 ruediger_: safari 4 is quite good 20:46:38 but with tracemonkey mozilla has a pretty fast js implementation, too 20:46:41 jthing: the one I was reading about was the webkit/apple javascript engine, not the google v8 thing 20:46:46 And ff3 adds the site icon to the bookmark list immediately, ff2 waited until the first visit, a behavior which I found useful. 20:46:51 -!- kpreid [n=kpreid@cpe-67-249-58-190.twcny.res.rr.com] has quit [] 20:46:53 Apple uses quite a lot of LLVM stuff 20:46:56 -!- vy [n=user@88.229.232.85] has left #lisp 20:47:00 most browsers are crap 20:47:07 truth 20:47:17 kpreid [n=kpreid@cpe-67-249-58-190.twcny.res.rr.com] has joined #lisp 20:47:17 p_l: apple usus llvm? 20:47:19 -!- delYsid [n=user@chello084115136207.3.graz.surfer.at] has quit [Read error: 104 (Connection reset by peer)] 20:47:26 i have to run them inside virtual machines because of security issues 20:47:27 guaqua: yeah 20:47:33 guaqua: a lot 20:47:34 firefox especially, such a pile of crap 20:47:39 url? 20:47:42 Crome is tbe onkly engine I know with 20:47:51 guaqua: check info on Apple's OpenGL info 20:47:53 or where it is used? 20:47:53 guaqua: they support llvm heavily. They use it e.g. to compile shaders on gpus that won't support them natively 20:47:53 there are strong rumours that MacOS 10.6 will use a gcc with an LLVM backend, at least for C/ObjC 20:47:54 <_3b> apple uses LLVM for GL drivers too from what i understand 20:48:01 like gallium will do for linux 20:48:16 at the moment, besides the javascript thing, they use it for OpenGL on machines with limited graphics cards 20:48:22 and they are developing clang in order to replace GCC in XCode 20:48:31 they use it for safari js engine? 20:48:41 Crome is the onky engine I know with Dynamic java support. 20:48:44 guaqua: not current safari 20:48:47 only 20:48:51 Tracemonkey also includes some JIT compilation 20:48:58 jthing: dynamic java support? 20:49:01 though I think it's in the recent nightly webkit releases 20:49:03 p_l: yes 20:49:15 christ I can't spell for shit 20:49:20 p_l: they are faster than v4 (the chrome engine) in some benchmarks 20:49:32 micro benchmarks 20:49:39 3.1b3 with tracemonkey set to enabled (you need to get to about:config to enable full features) works quite nice 20:49:43 ruediger_: yes. Same idea as java 20:49:56 jthing: sorry I can't follow you 20:50:26 of course, fast javascript, at least on the desktop, is pretty useless 20:50:27 ruediger_: dymaic recompilaton of javaScript 20:50:32 oh ok 20:50:45 because webdevs have to support IE, which has slow javascript even in IE8, anyway 20:50:46 rsynnott: I beg to differ 20:50:53 ok you are right :D 20:51:01 yeah IE is a big problem 20:51:16 IE is SLOW 20:51:22 agreed 20:51:50 and I guess that MS doesn't has the interest to boost JS in IE because it'll only help google *conspiracy mode on* 20:51:54 jthing: I beg to differ - IE8b2 was faster than FF3 on my windows xp pc :P 20:51:56 (same problem likely to apply to HTML5; Safari and Opera are supporting latest draft now, but IE probably won't until at least IE9) 20:52:17 microsoft is probably doing us all a favor by holding up "progress" 20:52:37 hefner: lol. Probably 20:52:38 ruediger_: or possibly just showing more of their recent inability to do anything 20:52:55 also apparently IE team at MS is working together with mozilla to make sure new HTML/JS APIs will be identical in both FF and IE 20:52:55 (it's really quite amazing how little they now get done for the number of people they employ) 20:53:00 recent? 20:53:14 rsynnott: yeah I guess they have a pretty crappy management 20:53:30 well, somewhere along the rode between XP and the imaginary longhorn 20:53:38 I mean they are one of the biggest software companies and they neither produce innovation nor their normal software in time 20:54:03 rofl 20:54:04 ruediger_: I blame managers on requiring "backward compatibility at bug-level" all the time 20:54:25 they're managing that in Win7 through virtualising XP 20:54:42 which on the face of it seems reasonable, but will actually require them to maintain XP basically forever 20:54:51 rsynnott: I guess programmers forced it down management throats 20:54:53 -!- sykopomp [n=user@unaffiliated/sykopomp] has quit [Remote closed the connection] 20:54:56 a ms insider told me e.g. that they usually have a dev team to produce a new software version. After it's done the dev team moves to the next piece of software and a new dev team takes over the other piece of software 20:55:30 so they have different people working to maintain the software than to develop it 20:55:31 though it looks like the OEMs will force them to do that anyway; HP has got permission to offer downgrade to XP on all machines until at least the end of 2010, and I'm sure there'll be further extensions 20:55:52 really? that sounds ridiculous 20:56:04 rsynnott: that XP deal will probably kill MS 20:56:19 they don't have that much choice 20:56:27 -!- antoszka [n=antoszka@unaffiliated/antoszka] has quit ["+++ killed by SIGSEGV +++"] 20:56:50 sykopomp [n=user@unaffiliated/sykopomp] has joined #lisp 20:57:20 a company forced to make all its internal software work on Vista/Win7 and replace all their hardware to handle the new OS, and retrain their employees, might get dangerous ideas about Apple or Linux 20:57:20 I don't understand the virtual XP thing. Windows is windows. If windows is suddenly not windows, you might as well buy a mac. 20:57:24 IMHO they should use that XP sandbox thingy, streamline it and apply it to all legacy code 20:57:35 ruediger_: That is completely untrue thogh.. 20:57:53 jthing: it is? 20:58:05 somebody told me that they develop visual c++ that way 20:58:11 hefner: it's more about software written outside MS than inside. Mostly stuff written by various bespoke companies 20:58:16 Seriously they are not that stupid 20:58:51 -c++ +studio 20:58:56 ruediger_: And it is 20:59:11 <_3b> hefner: the difference is MS can give you a free XP license to run your old apps in, but apple can't :/ 20:59:31 still, Windows 7 looks delicious. Though I wouldn't be surprised if my install would break quite a lot of apps ;) 20:59:33 ruediger_: What about C++ studio? 20:59:59 jthing: that they have one team for maintaining and one team for developing the new version 21:00:08 sbahra [n=sbahra@c-76-21-209-249.hsd1.dc.comcast.net] has joined #lisp 21:00:20 odd, though. Win7 is really nothing new, and the next windows mobile doesn't look like anything special 21:00:30 ruediger_: no they don'y 21:00:32 sorry, I just can't grasp this idea that running a virtualized Windows XP is a reasonable solution to backward compatibility, given that they haven't actually dropped support for Win32 so as to have a good excuse for it. 21:00:33 rsynnott: Windows 7 is Vista distilled :P 21:00:43 their only product that people get excited about is the xbox, and that's still a commercial failure 21:00:48 jthing: are you sure? The source I got this from was really trust worthy 21:01:12 hefner: it seems overkill, all right 21:01:18 <_3b> presumably the idea is that now they can drop win32 if they want 21:01:21 hefner: that virtual XP thing is more for broken apps that break on systems that don't behave *exactly* like the one they were written on, down to build number 21:01:32 and if they don't get it perfect, it will just confuse people 21:01:33 I mean, this isn't like Mac OS X and Classic, this is like if Mac OS 8 contained a virtual Mac OS 6 box. 21:01:49 I guess they bought some tech journalists and bloggers to make windows7 propaganda. 21:01:57 and yes, MS plans on dropping winapi 21:02:04 i mean it's the same as vista 21:02:14 Glad I use ubuthu.. 21:02:22 (the current Virtual PC is probably not up to the job; I find using WinXP in it more awkward than using WinXP on an old version of VMWare fusion) 21:02:26 antoszka [n=antoszka@unaffiliated/antoszka] has joined #lisp 21:02:35 but actually vista was probably the best windows. They finally introduced stuff that is common on unix since the 90s :D 21:02:39 ruediger_: As I said, it's Vista minus various problems of Vista 21:02:51 good night 21:02:54 hefner: and their XP mode is just a virtualiser, IIUC. Might as well run parallels or VMware ;) 21:02:57 and then they go and bring out things like the Zune, which never made any sort of sense 21:02:59 -!- kami- [n=user@unaffiliated/kami-] has quit ["ERC Version 5.3 (IRC client for Emacs)"] 21:02:59 ruediger_: true 21:03:01 p_l: but the public response is completely different 21:03:07 weird; the used to be quite competent 21:03:21 (also, the stupid table thing) 21:03:36 that XP mode is more than virtualisation - it contains AFAIK a shitload of stuff for integration of XP-mode apps with rest of the systenm 21:03:36 what? 21:03:39 *system 21:03:57 I guess that ms sees big rivals coming like google and apple 21:04:04 are we talking about m$ and windoze on #lisp? wtf?! 21:04:17 ruediger_: they have failed dramatically to respond to them, though 21:04:21 yeah 21:04:32 that's what happens if you are a fat monopoly 21:04:47 ms completely missed the internet 21:04:53 p_l: well, screw'em. surely these people can afford vmware and a winxp disk. 21:04:57 a FAT monopoly 21:05:24 probably, the XP mode is in fact part of the work for the future sandbox thing 21:05:43 -!- attila_lendvai [n=ati@adsl-89-132-1-79.monradsl.monornet.hu] has quit ["..."] 21:05:46 after all, Blackcomb was supposed to drop native apps completely 21:06:03 they lost the browser market, they lost the internet, they lost high class pcs, they lost mp3s ... 21:06:11 and they lost against winxp :D 21:06:16 ... We really are going to have to do quite a bit of work to get SBCL up-to-date with the modern world, aren't we? 21:06:18 (Vista flopped, the stupid Zune thing was ill-conceived from the start, and their response to iphone seems to be "hah! Windows Mobile is perfect! Everybody loves it!" 21:06:32 and the less said about Windows Live the better 21:06:45 nyef: yeah, ARM support :-) 21:06:49 rotfl 21:07:00 Windows Mobile? Right. They only say that because people stopped WinCEing at the name. 21:07:01 p_l: I suspect that Blackcomb is more or less a Cairo/Taligent product 21:07:05 (totally imaginary) 21:08:01 but ms is in fighting mode I guess. at least they don't want to loose on the netbooks 21:08:01 or Longhorn, for that matter :) 21:08:04 hmmm... when did they lost high class pcs? 21:08:06 if that happens they are gone 21:08:07 I agree Vista was a flop 21:08:19 I think Vista flop was more a PR thing 21:08:36 ruediger_: they've never been able to force the manufacturers to use Vista, there 21:08:42 But do you seriously think you can overcome it! 21:08:48 and I see no reason why it'll be any different with Win7 21:08:58 p_l: I have to google now. But isn't Apple selling more systems in the class of >1200$ at the moment 21:09:16 rsynnott: yeah they lost against winxp 21:09:26 ruediger_: certainly, Apple has more than 50% of the laptop market in that price category 21:09:31 probably not on desktops 21:09:38 ruediger_: my point 21:09:51 but that's just as bad 21:09:53 but the super-expensive desktop market has always been tiny, though the price bounds for it have moved down 21:10:05 rsynnott: t 21:10:45 and they're busily damaging that market with their own xbox; most of what market exists there is games-related 21:11:06 yes absolutely 21:11:15 but didn't they win against sony 21:11:48 nop 21:11:54 I heard that the ps3 is a big commercial failure 21:12:08 so it is 21:12:10 ruediger_: As far as I have seen it, outside of DTP people and similar, Microsoft isn't losing any market in PCs. 21:12:38 It's easy not to lose market share if the market is shrinking. 21:12:39 (despite it having the best hardware. If you could upgrade the ram it could probably be quite nice for scientific computing or even as a workstation) 21:13:04 I might be mistaken, but as far as I had seen, Macs are a niche. 21:13:08 p_l: hmm. Depends 21:13:14 unless you count macs sold as PCs 21:13:17 (Or, rather, it's easy to lose money while not losing market share if the market is shrinking.) 21:13:24 sure most companies run their default system which is a PC with win2k or xp and IE6 21:13:34 (especially large companies) 21:13:45 Heh. Yeah, IE6 still lives. 21:13:59 but if you look at the universities there are a lot of people using macbooks and similar 21:14:06 I have an IE6 VM -just- so I can enter time and expenses for my current job. 21:14:18 ruediger_: the worst problem is those large companies have software, quite often written by themselves, even, that requires exactly the same config as the dev's machine 21:14:20 and ps3 is still a failure 21:14:34 nyef: into an activex plugin? those were the days ... ;) 21:14:50 ruediger_: The common quote when we see a macbook here is "ok, who get's to pwn this machine?" :P 21:15:02 :D 21:15:09 but look at non technical studies 21:15:19 apple is a big trend 21:15:20 it's cool 21:15:29 Sure, go ahead and pwn my box. 21:15:30 ruediger_: Humanities are target of jokes even without macs, it would be too cruel :P 21:15:52 pkhuong: It started when a mac owner started touting how secure it s 21:15:53 *it is 21:16:11 a sys admin told me once that they switched to Microsoft in the 90s because everybody came from the university with their illegal copy of ms office 21:16:25 heh 21:16:37 ineiros_ [n=ineiros@rogue.hut.fi] has joined #lisp 21:16:46 and companies will adopt if they just get students with apple experience 21:17:20 borsman [n=quassel@76.177.217.216] has joined #lisp 21:17:27 stassats`: do you know if your patch to slime-fuzzy will be accepted? 21:17:36 and getting the technical staff to change is probably the easiest thing. But getting the secretaries to change :) 21:17:57 -!- clynbech [n=user@cpe.atm2-0-1071129.arcnxx16.customer.tele.dk] has quit [Connection timed out] 21:18:05 pkhuong: your box? 21:18:19 The secretaries will start changing once people start using openoffice more. 21:18:23 ace4016 [i=ace4016@cpe-76-168-248-118.socal.res.rr.com] has joined #lisp 21:18:24 ruediger_: install a solitaire and minesweeper clone and you have secretaries ready to change. The technical people *I* know are much harder to swithc 21:18:27 *switch 21:18:35 :D 21:18:54 depends on the age 21:18:56 my main problem with macs is their owner's attitude, most of the time 21:19:01 yeah 21:19:03 -!- ineiros [n=ineiros@kosh.hut.fi] has quit ["leaving"] 21:19:16 ruediger_: I decided a long time ago that my time was worth too much to waste time fiddling with linux on a laptop. Win XP + coLinux, and now OS X. 21:19:24 -!- ineiros_ is now known as ineiros 21:20:03 heh. For me, Macs were always outside price range and move to intel killed any interest from me 21:20:15 OSX was my major system for 4 years 21:20:24 but I switched back to linux 21:20:31 (yes, I was drooling over PowerMac G5 with PCI-Express) 21:20:46 Yeah, a G5 laptop would have been interesting if it were still a contender, but the move to intel kindof killed the point. 21:20:50 The OS sucks, frankly, but that's why I have a desktop. 21:20:57 nyef: absolutely 21:21:38 I was so happy that I bought my iBook when I heard that they wanted to switch to intel 21:21:41 gigamonk` [n=user@adsl-99-17-207-207.dsl.pltn13.sbcglobal.net] has joined #lisp 21:22:10 at the time I was still hoping that they'd either release a g5 laptop or sth. based on cell (but with a more powerful ppc) 21:22:25 -!- DarkRavin [n=thedarkr@adsl-147-121-66.bhm.bellsouth.net] has quit ["Ex-Chat"] 21:22:44 but in the end I think linux is the better system 21:22:46 If I was to spend >20K USD on a desktop machine that isn't going to run games, I'd get SiCortex PDS. If a gaming rig - something designed with Windows in mind (i7?). For laptop: Something with Linux, maybe even on ARM 21:22:56 -!- sad0ur [n=sad0ur@r2bk213.net.upc.cz] has quit ["leaving"] 21:23:01 G5-based laptops: because battery life is overrated. 21:23:03 heh, when I was in university, at least in the last couple of years, it was always the CS and maths students who had apple laptops, more than the humanities students 21:23:10 pkhuong: :-D 21:23:33 pkhuong: but there are power saving g5s available now. 21:23:39 ruediger_: there was NEARLY a g5 laptop :) 21:23:51 yeah 21:23:59 IBM released a suitably low energy chip days after the switchover announcement 21:24:05 yeah 21:24:06 http://www.youtube.com/watch?v=JFbceSyo-Xg 21:24:08 bad timing 21:24:26 (and they'd been talking to that PPC manufacturer that they later bought about a laptop chip) 21:24:33 sorry, I am pissed off 21:24:35 intel - because we don't want you to notice we fudged up battery benchmarks, we paid BIOS writers to cripple competition ;-) (joke... I hope) 21:24:42 ruediger_: probably a deliberate announcement on IBM's part, to save face 21:24:44 rsynnott: freescale? 21:24:57 nope, funny name, they bought it a few months back 21:25:06 PA Semi 21:25:07 rsynnott: apple was a small customer for ibm, comparatively 21:25:29 rsynnott: yeah rumor was that Apple was pissed due to Microsoft getting the newest powerpcs for their xbox and IBM not doing a thing for intel. But there were so many rumors at the time 21:25:38 (it was perceived that the failure of G5 was largely IBM's fault) 21:28:21 apple also looks set to go up against exchange, shortly 21:28:34 (will be very interesting to see what microsoft do if they do) 21:29:35 mrsolo [n=mrsolo@adsl-68-126-209-44.dsl.pltn13.pacbell.net] has joined #lisp 21:29:41 rsynnott: against exchange? 21:30:38 MacOS 10.6 will be both an exchange client and a client for an apple clone of exchange 21:30:41 nego [n=nego@c-67-173-168-255.hsd1.il.comcast.net] has joined #lisp 21:30:54 MacOs 10.6 server will provide the server component 21:31:18 pricing will presumably be quite competitive; apple server pricing has always been far lower than MS 21:32:16 *p_l* notices that he has never seen an apple server in the wild, except for the two on K-PAX page 21:32:52 no, for the moment they're not attractive except as fileserver/domain controllers for primarily apple networks 21:33:07 (no client licensing required, so far cheaper than Win2003) 21:33:25 rsynnott: ah, their "unlimited clients" license? 21:33:40 if they were to release a decent windows client for their exchange clone, though, it could do well; exchange is stupidly expensive 21:33:57 yep; comes with the machine 21:34:03 I wonder if MS could be persuaded to drop per-client licensing 21:34:18 very unlikely, unless there's serious competition 21:34:37 Aren't there free exchange server replacements with full outlook interop available already? 21:34:41 (actually, they haven't done it even where there IS serious competition; SQL server being a good example) 21:34:59 Though I guess what I would very like is to see more competent windows admins... 21:35:00 nyef: generally not supported by a large company, though 21:35:16 I suppose I can see that. 21:35:20 nyef: yeah. 21:35:38 every now and than you read sth. about that in a new linux distro review 21:35:40 nyef: also none of them are drop-in compatible - it's jumping through flaming hoops for Outlook interop 21:35:54 nothing except exchange supports exchange protocol 21:36:09 Hmm... Oh well. 21:36:43 the closest thing to outlook interop I had seen so far was something that allowed Outlook to use IMAP4 as storage 21:37:10 the "competition" uses usually their own software 21:37:19 SQL Server is actually quite good at what it does, but its pricing is simply too high 21:37:59 but aren't it's competitors (the one that they see themselves as competition) similarly priced? 21:38:11 -!- gigamonk` [n=user@adsl-99-17-207-207.dsl.pltn13.sbcglobal.net] has quit [Read error: 110 (Connection timed out)] 21:38:44 well, that's the problem, really; it's choosing to compete only against oracle and db2 21:38:48 nyef: so, do you have any idea what the magic incantation is to change the name of the instruction when both args are registers? 21:39:04 rsynnott: there's also sybase 21:39:35 *p_l* wouldn't buy Oracle stuff... except for Rdb, but they didn't develop it themselves 21:39:36 it could potentially compete against mysql/postgres, which are used to run very big webservices, but its pricing model doesn't make sense for that 21:39:59 pkhuong: For sb-disassem? I'd have thought that using templates that don't have overlapping encoding spaces would suffice. 21:40:22 plenty of companies pay a relative fortune for MySQL support and enterprise tools, but it's still less than they'd be paying for SQL Server 21:40:36 (and crucially, they were able to get started with mysql/postgres for free) 21:40:41 I guess mysql isn't really a competition - those using it would be probably content with BDB/TC I guess (or did they repair subqueries?) 21:40:56 gigamonk` [n=user@adsl-99-17-207-207.dsl.pltn13.sbcglobal.net] has joined #lisp 21:41:06 I can't see many mysql users being too happy with BDB 21:41:13 it is still a relational database 21:41:18 I wonder if there are big installations of those exchange compatible servers 21:41:23 (but no, subqueries on MySQL are still best avoided) 21:41:48 rsynnott: :( So still, "If you can, avoid MySQL" still stands in my rulebook 21:42:25 oh, absolutely 21:42:51 unfortunately, if I get the job I'm applying for (not likely), I'll have to work with it 21:43:19 however, there are very big functioning mysql applications 21:45:07 such as? 21:46:57 Google AdWords, Facebook spring to mind 21:47:27 -!- nego_ [n=nego@c-67-173-168-255.hsd1.il.comcast.net] has quit [Read error: 110 (Connection timed out)] 21:47:51 Doesn't AdWords work on BigTable? 21:48:02 <_3b> what would be a good macro character for embedding markup into code? (reader macro would just ignore the markup, so marked up code would still be compilable) 21:50:05 p_l: the interface where people buy ads works on mysql, apparently 21:50:22 there was some interesting documentation on their attempt to move to oracle a while back 21:50:28 _3b: I like ; for that purpose. :-P 21:50:45 oh, no, not markup in code! 21:51:07 <_3b> nyef: i'd like to be able to do something like (this is old code _(this part changed)_ more old code) 21:54:08 drakej [n=Steven@216-67-20-72-rb2.nwc.dsl.dynamic.acsalaska.net] has joined #lisp 21:54:09 <_3b> haven't decided if i need more than just highlighting some regions or not though 21:54:18 -!- tritchey [n=tritchey@c-68-58-88-241.hsd1.in.comcast.net] has quit [] 21:55:14 <_3b> rsynnott: not intended for real code, just code on a web page 21:56:10 -!- X-Scale [i=email@89-180-180-54.net.novis.pt] has left #lisp 21:56:26 ah, ok :) 21:57:06 <_3b> but i'd rather something that could be a reader macro instead of some existing markup, since then i can keep using slime while i edit it :) 21:58:36 <_3b> (at least when the multi-mode.el stuff isn't flaking out :) 22:06:38 -!- hkBst [n=hkBst@gentoo/developer/hkbst] has quit [Read error: 104 (Connection reset by peer)] 22:07:37 mcspiff [n=user@drmons0501w-142177079167.pppoe-dynamic.ns.aliant.net] has joined #lisp 22:08:15 nyef: I'd have to pull out one of fields making up reg/mem in a new instruction format for only 2 instructions that frankly piss me off. 22:08:17 pina [n=finale@unaffiliated/pina] has joined #lisp 22:08:31 -!- mcspiff [n=user@drmons0501w-142177079167.pppoe-dynamic.ns.aliant.net] has left #lisp 22:08:56 mcspiff [n=user@drmons0501w-142177079167.pppoe-dynamic.ns.aliant.net] has joined #lisp 22:11:28 -!- smithzv [n=smithzv@c-67-173-240-139.hsd1.co.comcast.net] has quit [Read error: 110 (Connection timed out)] 22:13:08 I'd have a small question 22:13:18 I write a function which returns a lambda form 22:13:29 I can use it as a function with funcall 22:13:40 but I'd want to use it directly 22:13:49 -!- Yuuhi [i=benni@p5483FFCA.dip.t-dialin.net] has quit ["ERC Version 5.3 (IRC client for Emacs)"] 22:13:53 so I use setq to assign the lambda form to a symbol 22:14:00 but I still can't use it as a function 22:14:04 is there a way to do this ? 22:14:27 setf fdefinition 22:14:27 clhs symbol-function 22:14:27 http://www.lispworks.com/reference/HyperSpec/Body/f_symb_1.htm 22:14:28 mejja [n=user@c-f6b6e555.023-82-73746f38.cust.bredbandsbolaget.se] has joined #lisp 22:18:02 mhh 22:18:04 thank you 22:18:09 gonna try to figure out how to use it 22:19:06 -!- ruediger_ [n=ruediger@62-47-133-248.adsl.highway.telekom.at] has quit ["Leaving"] 22:20:16 -!- gigamonk` [n=user@adsl-99-17-207-207.dsl.pltn13.sbcglobal.net] has quit [Read error: 60 (Operation timed out)] 22:21:09 it works perfectly, thank you guys! 22:22:51 ziarkaen [n=ziarkaen@87.115.92.73] has joined #lisp 22:23:32 -!- ehu` [n=chatzill@82-170-33-173.ip.telfort.nl] has quit [Read error: 110 (Connection timed out)] 22:23:51 ruediger [n=the-rued@62-47-133-248.adsl.highway.telekom.at] has joined #lisp 22:24:16 maybe you really want funcall 22:24:27 mhh I don't think so 22:24:40 unless your intention is for the generated function to stay around forever 22:24:47 I have a log-message function which takes among other args a log level 22:25:04 I'd like a simple function for each log level, ltrace, linfo, etc. 22:25:34 -!- nickw [n=nick@93-97-203-158.zone5.bethere.co.uk] has quit [Read error: 54 (Connection reset by peer)] 22:25:56 galdor pasted "symbol-function example" at http://paste.lisp.org/display/80351 22:26:08 i find it conveniant 22:26:21 note that I'm a lisp beginner, if it's an error, just tell me :) 22:26:57 it's an error 22:27:29 non-'idiomatic', at least 22:28:09 write a macro DEFINE-LOG-LEVEL or something which expands into the right code (possibly DEFUNs) 22:28:21 you might also like to step one level out and delete the `log-func' function and just call some implementation of curry directly 22:28:40 assuming you have some utilities package imported 22:28:45 puchacz [n=puchacz@87-194-5-99.bethere.co.uk] has joined #lisp 22:28:48 oh just didn't see I could use curry 22:29:08 alexandria, arnesi, and metatilities-base all include curry 22:31:26 -!- benny [n=benny@i577A13B4.versanet.de] has quit [Read error: 110 (Connection timed out)] 22:33:19 since I'm rewriting this part, I'd add another question, why have I to respect some declaration order for functions 22:33:29 I mean if func a use func b, b has to be declared before a 22:33:35 no 22:34:12 perhaps I did something wrong, but if I define b after a, sbcl tells me that b is undefined 22:34:21 -!- mrsolo [n=mrsolo@adsl-68-126-209-44.dsl.pltn13.pacbell.net] has quit [Read error: 110 (Connection timed out)] 22:34:33 -!- Nshag [i=user@Mix-Orleans-106-3-195.w193-248.abo.wanadoo.fr] has quit [Remote closed the connection] 22:34:39 that's because you're evalling one at a time, if you do both in a compilation unit the message won't show up 22:35:08 regardless you can ignore the message if you like 22:35:23 tcr [n=tcr@host145.natpool.mwn.de] has joined #lisp 22:35:24 I'm evaluating using asdf 22:35:36 I mean asdf:oop 'adsf:load-op :mypackage 22:35:48 so they are in separate files? 22:35:48 I don't use slime (don't hit me) 22:35:55 a and b ? no 22:35:58 -!- atarashi [n=code@y225183.dynamic.ppp.asahi-net.or.jp] has quit ["Leaving"] 22:47:21 galdor: using slime isn't actually mandatory, but it is probably a good idea, and will save you annoyance in the lnog run 22:47:47 -!- puchacz [n=puchacz@87-194-5-99.bethere.co.uk] has quit ["Konversation terminated!"] 22:51:06 -!- gottesmm [n=gottesmm@c121h014.wless.reed.edu] has quit [Read error: 60 (Operation timed out)] 22:52:52 mjf [n=mjf@r3a98.net.upc.cz] has joined #lisp 22:53:45 rsynnott: I'm really used to vim 22:54:08 my fingers are kind of mapped to vim commands 22:54:17 so was I :) 22:54:56 -!- drakej [n=Steven@216-67-20-72-rb2.nwc.dsl.dynamic.acsalaska.net] has left #lisp 22:55:18 jfactor [n=jfactor@student167-189.hampshire.edu] has joined #lisp 22:58:21 so were quite a few among us 22:59:13 ikki [n=ikki@189.228.133.213] has joined #lisp 23:02:40 galdor pasted "logging" at http://paste.lisp.org/display/80352 23:02:53 does it seems more idiomatic ? 23:04:40 Zenton`` [n=user@80.29.224.179] has joined #lisp 23:06:36 galdor: No, why use macros? 23:08:03 because michaelw adviced me to 23:08:14 it seems fine, make the code shorter 23:08:50 *makes, sorry 23:09:06 -!- ziarkaen [n=ziarkaen@87.115.92.73] has quit [Read error: 110 (Connection timed out)] 23:09:45 Not really. (setf (fdefinition 'ltrace) (make-log-function 'trace)) 23:10:32 Yes, but it's nicer to use defun, since you benefit from implementation defined behavior, and may add documentation string. 23:10:54 -!- antoszka [n=antoszka@unaffiliated/antoszka] has quit ["+++ killed by SIGSEGV +++"] 23:11:27 r1nu- [n=r1nu-@ppp-94-68-16-210.home.otenet.gr] has joined #lisp 23:12:49 drunkenmonkey [i=55c89673@gateway/web/ajax/mibbit.com/x-9962ad0438392cc4] has joined #lisp 23:12:58 jophish [n=jophish@dial-80-47-3-226.access.uk.tiscali.com] has joined #lisp 23:12:59 pjb: Where do you get all that time from to argue in c.l.l? 23:13:16 Lazy saturday. 23:13:31 ()()(((Fu)(()(ck())())()(Li)()()()sp)()(dumb)()((asses))()( 23:13:42 and C++ compilations in the week :-) 23:14:24 pjb: but if i use defun and not defmacro, I can't set the the name of the specialized log function right ? 23:14:29 drunkenmonkey: Very clever. 23:14:39 no it wasn't 23:14:45 you must be stupid 23:15:03 galdor: yes, that's why I said it's nicer to use defun in you macro, than make-log-function which would use lambda. 23:15:38 nikodemus [n=nikodemu@cs181229041.pp.htv.fi] has joined #lisp 23:15:46 -!- Zenton` [n=user@80.29.224.221] has quit [Read error: 104 (Connection reset by peer)] 23:15:48 pjb: ah ok, thought you wanted me not to use defmacro :) 23:15:58 galdor: My point was that you not necessarily need macros. Newcomers often try to use macros for when higher-order function would do, too. 23:16:16 well lispians are better than haskexuals, at least there is no ass involved 23:16:32 but still an abomination 23:16:40 I more and more like lisp, but I use soo many time to find more elegant ways to do things 23:16:43 galdor: I'd probably go the macro route myself, but only for the implementation-dependent extensions regarding DEFUN vs (SETF FDEFINITION) of my implementation of choice 23:18:36 -!- sbahra [n=sbahra@c-76-21-209-249.hsd1.dc.comcast.net] has quit ["Leaving"] 23:19:16 shesh 23:19:21 you people are all dead inside 23:19:40 antoszka [n=antoszka@unaffiliated/antoszka] has joined #lisp 23:19:50 enclosed in parenthesises 23:20:06 damn this trolling sucked 23:20:10 -!- drunkenmonkey [i=55c89673@gateway/web/ajax/mibbit.com/x-9962ad0438392cc4] has left #lisp 23:20:27 Victory! 23:22:57 -!- mcspiff [n=user@drmons0501w-142177079167.pppoe-dynamic.ns.aliant.net] has quit [Read error: 104 (Connection reset by peer)] 23:26:31 pjb: what happened? 23:26:50 The troll left, despised. 23:27:01 ah, drunkenmonkey? had him on ignore ^^; 23:28:15 Ignoring him worked perfectly well. 23:28:20 yes its sunday today :D 23:33:15 dwave [n=ask@212.251.243.4] has joined #lisp 23:40:16 sbahra [n=sbahra@c-76-21-209-249.hsd1.dc.comcast.net] has joined #lisp 23:44:08 -!- tdy [n=tdy@unaffiliated/tdy] has quit ["quitting..."] 23:44:58 -!- mejja [n=user@c-f6b6e555.023-82-73746f38.cust.bredbandsbolaget.se] has quit [Remote closed the connection] 23:46:06 -!- Athas [n=athas@0x50a157d6.alb2nxx15.dynamic.dsl.tele.dk] has quit [Remote closed the connection] 23:46:48 any relation to gigamonkey there? 23:47:28 He'd probably like you to think so. 23:48:20 -!- tcr [n=tcr@host145.natpool.mwn.de] has quit ["Leaving."]