00:29:30 -!- rbarraud [~rbarraud@118-92-1-3.dsl.dyn.ihug.co.nz] has quit [Ping timeout: 265 seconds] 00:32:41 rbarraud [~rbarraud@118-92-1-3.dsl.dyn.ihug.co.nz] has joined #sbcl 01:04:46 _3b`` [foobar@cpe-72-179-19-4.austin.res.rr.com] has joined #sbcl 01:48:59 -!- nyef [~nyef@pool-64-222-178-106.man.east.myfairpoint.net] has quit [Quit: G'night all.] 02:18:43 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 240 seconds] 02:20:07 -!- attila_lendvai [~attila_le@catv-89-133-169-124.catv.broadband.hu] has quit [Ping timeout: 265 seconds] 03:01:07 -!- Kaer [b@c-2fcce253.97-16-64736c12.cust.bredbandsbolaget.se] has quit [Read error: Connection reset by peer] 03:15:27 -!- hargettp [~anonymous@96.237.121.128] has quit [Quit: hargettp] 03:28:45 -!- Blkt [~user@dynamic-adsl-94-37-236-50.clienti.tiscali.it] has quit [Ping timeout: 265 seconds] 07:11:09 tcr [~tcr@cpc5-bour5-2-0-cust340.15-1.cable.virginmedia.com] has joined #sbcl 07:22:11 -!- tcr [~tcr@cpc5-bour5-2-0-cust340.15-1.cable.virginmedia.com] has quit [Read error: Connection reset by peer] 07:22:27 tcr [~tcr@cpc5-bour5-2-0-cust340.15-1.cable.virginmedia.com] has joined #sbcl 08:04:27 -!- rbarraud [~rbarraud@118-92-1-3.dsl.dyn.ihug.co.nz] has quit [Read error: Connection reset by peer] 08:07:42 rbarraud [~rbarraud@118-92-22-174.dsl.dyn.ihug.co.nz] has joined #sbcl 09:25:38 nikodemus [~nikodemus@cs181199216.pp.htv.fi] has joined #sbcl 09:25:38 -!- ChanServ has set mode +o nikodemus 09:47:16 afternoon 09:50:53 Hi nikodemus. What's on today's agenda? 09:56:13 stassats [~stassats@wikipedia/stassats] has joined #sbcl 10:56:33 Blkt [~user@dynamic-adsl-94-37-236-50.clienti.tiscali.it] has joined #sbcl 10:59:14 good day everyone 11:00:28 today: housecleaning, mostly. i don't think i'll have time for much sbcl work today 11:15:40 rbarraud_ [~rbarraud@118-92-8-84.dsl.dyn.ihug.co.nz] has joined #sbcl 11:15:54 -!- rbarraud [~rbarraud@118-92-22-174.dsl.dyn.ihug.co.nz] has quit [Ping timeout: 252 seconds] 11:21:35 What are the costs of with-pinned-objects? 11:22:16 push a word on stack, pop it back out 11:22:28 well, word per object 11:25:00 which of course means that the tail-position of w-p-o isn't a real tail-position 11:25:32 thanks 11:33:53 -!- nikodemus [~nikodemus@cs181199216.pp.htv.fi] has quit [Quit: This computer has gone to sleep] 11:51:57 hargettp [~anonymous@96.237.121.128] has joined #sbcl 11:58:18 -!- hargettp [~anonymous@96.237.121.128] has quit [Quit: hargettp] 12:04:02 hargettp [~anonymous@96.237.121.128] has joined #sbcl 12:15:38 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 264 seconds] 13:03:17 -!- Blkt [~user@dynamic-adsl-94-37-236-50.clienti.tiscali.it] has quit [Ping timeout: 276 seconds] 13:05:58 Why do EQ and EQL hash tables die on float NaNs? Neither operation cares by itself. 13:07:53 This is especially strange with EQ 13:08:25 And means that you cannot have nan constants, because the compiler dies on them inside ctype-of 13:27:37 angavrilov_: for hash tables, at least, it depends on the trap mask. 13:29:16 I interpret EQ as comparing the bits of the boxed pointer, and EQL as comparing the bits of the boxed number. I don't see where the trap mask comes into it 13:29:57 Anyway, you cannot change the trap mask for the compiler itself 13:31:46 and there are known issues with floating point computations and the compiler. 13:34:35 actually, no, hash tables wfm on linux/x86-64. 13:36:58 Dies here: (setf (gethash (sb-kernel:make-single-float #x7fceadde) (make-hash-table :test #'eq)) 3) 13:39:26 it's the compilation that dies, not the operation itself. 13:40:13 Yes I see, now that I've put the value in a var 13:40:18 <_3b``> lp 486812 13:40:23 I wonder what exactly dies though 13:40:41 bound derivation for numeric types. 13:41:10 <_3b``> hmm, bots not cooperating again? https://bugs.launchpad.net/sbcl/+bug/486812 13:41:16 <_3b``> angavrilov_: ^ that? 13:42:04 Yes 13:42:19 Disabling foldable won't help if you make a defconstant 13:42:52 I guess I'll have to use a defparameter for now :) 13:46:21 I've written VOPs to save float constants in the constant vector before... 13:46:38 *constant segment. 14:17:13 stassats [~stassats@wikipedia/stassats] has joined #sbcl 14:41:50 nikodemus [~nikodemus@cs181199216.pp.htv.fi] has joined #sbcl 14:41:50 -!- ChanServ has set mode +o nikodemus 15:13:10 Bronsa [~bronsa@host60-184-dynamic.21-79-r.retail.telecomitalia.it] has joined #sbcl 15:38:59 jweiss [~weissj@cpe-069-134-025-078.nc.res.rr.com] has joined #sbcl 16:08:22 -!- hargettp [~anonymous@96.237.121.128] has quit [Quit: hargettp] 16:10:42 -!- Bronsa [~bronsa@host60-184-dynamic.21-79-r.retail.telecomitalia.it] has quit [Quit: Lost terminal] 16:26:48 -!- chandler [~n@opendarwin/developer/chandler] has quit [Ping timeout: 240 seconds] 16:27:20 chandler [~n@new.unmutual.info] has joined #sbcl 16:27:45 -!- chandler is now known as Guest24533 16:35:17 hargettp [~anonymous@96.237.121.128] has joined #sbcl 16:36:17 -!- Guest24533 [~n@new.unmutual.info] has quit [Changing host] 16:36:17 Guest24533 [~n@opendarwin/developer/chandler] has joined #sbcl 16:36:34 -!- Guest24533 is now known as chandler 16:49:08 -!- cmm [~cmm@bzq-79-182-202-208.red.bezeqint.net] has quit [Ping timeout: 250 seconds] 16:49:55 cmm [~cmm@bzq-79-182-202-208.red.bezeqint.net] has joined #sbcl 16:51:00 Blkt [~user@dynamic-adsl-94-37-236-50.clienti.tiscali.it] has joined #sbcl 17:05:49 nyef [~nyef@pool-64-222-178-106.man.east.myfairpoint.net] has joined #sbcl 17:05:57 G'afternoon all. 17:29:33 -!- Blkt [~user@dynamic-adsl-94-37-236-50.clienti.tiscali.it] has quit [Quit: Error: do not makunbound t please!] 17:30:01 Blkt [~user@dynamic-adsl-94-37-236-50.clienti.tiscali.it] has joined #sbcl 17:41:58 hm. i think i know how to make the dreaded "could not optimize away %SAP-ALIEN: forced to do runtime allocation of alien-value structure" much less of a performance problem 17:42:23 (1) intern alien types (2) cache the naturalize & deport functions in alien types 17:44:19 for example, right now (sb-bsd-sockets:get-host-by-name "localhost") causes us to compute the naturalize-lambda for (unsigned 8) four times! 17:47:19 ... Nice. 17:48:13 try that call with (TIME ...) and note the number of forms interpreted :) 17:48:33 Not going there. 17:48:43 heh 17:48:57 I've got a possible angle on reducing the size of DX closures, but it depends on some changes that I haven't gotten working right just yet. 17:49:25 nice 17:49:51 are you looking at (let ((f (lambda ...))) (declare (dx f)) ...) as well? 17:50:03 That's on my list. 17:50:11 I was rather surprised to find that it didn't work. 17:50:15 *stassats* notices that with DRAKMA all the TIME 17:50:39 name resolution takes more time than the actual request 17:50:48 when on localhost, that is 17:50:55 super. i had a gander at it a while ago, but didn't get far. i *think* i figured out what needs to happen, but there was some extra information that i needed in a place where it wasn't available or something 17:51:36 Right now, though, I'd be happy if I could get a clean bootstrap with UWP cleanups being marked as DX. 17:51:40 hah 17:52:03 Since that's actually the one case that my current closure changes actually /should/ support. 17:52:17 disassembler stuff: should i look at merging some of your patches next week, or is it better if you handle them when you have the time? 17:52:36 I should probably do it, because I know of a couple bugs in them. 17:52:44 And the sequencing of the changes is wrong. 17:52:48 ok 17:53:01 when should i next poke you about it? 17:53:14 next weekend? 17:53:27 Friday, I think? 17:53:35 ok 17:53:46 Might have attacked it by then, too. 17:54:27 I've also got a couple PPC tweaks, but they're more of a hack, and less important. 18:39:32 nikodemus pasted "this looks like a win" at http://paste.lisp.org/display/115638 18:40:28 nikodemus: Just a little bit of a win, yes. 18:40:30 Ship it! 18:44:21 arrays, structures and unions are still as bad as ever, though 18:44:46 because i didn't want to cache things in hash-tables with list-keys 18:44:55 So, is that just pointer types / integers? 18:45:20 And, hey, does alien type redefinition still fail silently? 18:45:32 did it? 18:45:56 Yeah, it'd ignore any redefinitions. 18:46:10 anywyas, integer types are stored in two vectors indexes by width upto 256 bytes, pointer types are cached in types they point to, singleton types in global variables] 18:46:20 Possibly just for either structs or pointers to (struct nil ...) types. 18:46:21 256 bits, eveb 18:46:38 nyef: i think i did something about that at some point, but could be a different issue 18:50:05 Mmm. And it's been a long while since last I did anything in that direction. 18:58:35 looking at parse-alien-record-type, redefinitions should take effect normally 18:58:55 Okay, fair enough. 18:59:22 Hrm... I'm 20 minutes into my build, why hasn't it failed yet? 19:05:57 Ah, right. Enabling SHOW slows things down. 19:12:57 Blkt` [~user@dynamic-adsl-94-37-226-26.clienti.tiscali.it] has joined #sbcl 19:14:30 -!- Blkt [~user@dynamic-adsl-94-37-236-50.clienti.tiscali.it] has quit [Ping timeout: 240 seconds] 19:47:30 -!- hargettp [~anonymous@96.237.121.128] has quit [Quit: hargettp] 20:05:25 ... Joy. I've got a closed-over LAMBDA-VAR which has a TN without a corresponding SAVE-TN. 20:08:18 -!- Blkt` is now known as Blkt 20:32:50 -!- Blkt [~user@dynamic-adsl-94-37-226-26.clienti.tiscali.it] has quit [Read error: Connection reset by peer] 20:33:28 Blkt [~user@dynamic-adsl-94-37-226-26.clienti.tiscali.it] has joined #sbcl 20:48:46 -!- nikodemus [~nikodemus@cs181199216.pp.htv.fi] has quit [Quit: Leaving] 20:49:44 -!- minion [~minion@common-lisp.net] has quit [Ping timeout: 264 seconds] 20:49:44 -!- lisppaste2 [~lisppaste@common-lisp.net] has quit [Ping timeout: 264 seconds] 20:51:48 -!- jweiss [~weissj@cpe-069-134-025-078.nc.res.rr.com] has quit [Ping timeout: 245 seconds] 20:55:20 jweiss [~weissj@cpe-069-134-025-078.nc.res.rr.com] has joined #sbcl 21:00:53 bleh, no bot. 21:01:02 I'll fix it soon. 21:01:10 Just have to start this build. 21:04:18 lisppaste2 [~lisppaste@common-lisp.net] has joined #sbcl 21:08:05 minion [~minion@common-lisp.net] has joined #sbcl 21:11:38 specbot [~specbot@common-lisp.net] has joined #sbcl 21:21:01 drdo [~user@98.192.108.93.rev.vodafone.pt] has joined #sbcl 21:49:17 -!- cmm [~cmm@bzq-79-182-202-208.red.bezeqint.net] has quit [Ping timeout: 272 seconds] 21:49:37 cmm [~cmm@bzq-79-182-202-208.red.bezeqint.net] has joined #sbcl 22:14:48 -!- tcr [~tcr@cpc5-bour5-2-0-cust340.15-1.cable.virginmedia.com] has quit [Quit: Leaving.] 22:19:28 -!- mbohun [~mbohun@ppp115-156.static.internode.on.net] has quit [Quit: Leaving] 23:23:54 -!- mega1 [~quassel@catv4E5CABA2.pool.t-online.hu] has quit [Ping timeout: 250 seconds] 23:55:23 -!- jweiss [~weissj@cpe-069-134-025-078.nc.res.rr.com] has quit [Ping timeout: 276 seconds]