00:15:06 drl [~lat@110.139.229.172] has joined #sbcl 00:17:12 -!- rpg [~rpg@206.117.31.188] has quit [Ping timeout: 240 seconds] 00:37:46 -!- borkman [~user@S0106001111de1fc8.cg.shawcable.net] has quit [Ping timeout: 258 seconds] 00:52:19 -!- drl [~lat@110.139.229.172] has quit [Remote host closed the connection] 02:50:41 drl [~lat@110.139.229.172] has joined #sbcl 03:42:49 -!- drl [~lat@110.139.229.172] has quit [Read error: Connection reset by peer] 04:03:56 drl [~lat@110.139.229.172] has joined #sbcl 05:00:44 -!- cow-orker [~foobar@pogostick.net] has quit [Ping timeout: 258 seconds] 05:00:49 cow-orker [~foobar@pogostick.net] has joined #sbcl 05:01:38 -!- redline6561 is now known as redline6561_nop 05:12:20 -!- tsuru [~charlie@adsl-98-87-25-188.bna.bellsouth.net] has quit [Read error: Connection reset by peer] 05:13:25 tsuru [~charlie@adsl-74-179-28-14.bna.bellsouth.net] has joined #sbcl 05:19:36 -!- akovalen` is now known as akovalenko 05:22:47 -!- drl [~lat@110.139.229.172] has quit [Remote host closed the connection] 05:56:31 -!- Krystof [~user@81.174.155.115] has quit [Ping timeout: 248 seconds] 06:01:31 sdemarre [~serge@91.176.40.215] has joined #sbcl 07:21:45 Krystof [~user@81.174.155.115] has joined #sbcl 07:25:11 Blkt [~user@89-96-199-46.ip13.fastwebnet.it] has joined #sbcl 07:29:04 good morning everyone 08:00:09 pkhuong: 08:00:27 I've lost track of our division-by-multiply: anything new here? 08:23:42 EmmanuelOga [~emmanuel@host227.190-31-137.telecom.net.ar] has joined #sbcl 08:24:09 -!- EmmanuelOga [~emmanuel@host227.190-31-137.telecom.net.ar] has left #sbcl 09:43:10 -!- Blkt [~user@89-96-199-46.ip13.fastwebnet.it] has quit [Remote host closed the connection] 09:43:43 Blkt [~user@89-96-199-46.ip13.fastwebnet.it] has joined #sbcl 09:52:05 Neronus [christian@heraklit.ayous.org] has joined #sbcl 10:11:20 oiiii [~oiiii@82.71.241.25] has joined #sbcl 10:17:30 -!- akovalenko [~anton@95.73.55.37] has quit [Ping timeout: 260 seconds] 10:19:19 akovalenko [~anton@95.73.107.239] has joined #sbcl 12:29:32 -!- _8david` [~user@port-92-195-41-59.dynamic.qsc.de] has quit [Ping timeout: 256 seconds] 13:04:06 _8david` [~user@port-92-195-41-59.dynamic.qsc.de] has joined #sbcl 13:11:39 -!- tsuru [~charlie@adsl-74-179-28-14.bna.bellsouth.net] has quit [Remote host closed the connection] 13:22:05 -!- whoops [u549@gateway/web/irccloud.com/x-emvpxaolxvhgpoyp] has quit [Read error: Connection reset by peer] 13:28:15 nyef [~nyef@c-174-63-105-188.hsd1.ma.comcast.net] has joined #sbcl 13:28:23 G'morning all. 13:29:05 I was thinking about the "static array" thing some more, and I have some questions and possible solution approaches. 13:30:15 fe[nl]ix: Would it be "sufficient" for your solution if you could use AREF and (SETF AREF) on alien arrays, or are you looking for something "more complete"? 13:30:43 (Alien arrays are still arrays, what happens if we actually place them below CL:ARRAY in the type lattice?) 13:36:35 whoops [u549@gateway/web/irccloud.com/x-xjeulhfkgthmaijq] has joined #sbcl 13:39:10 nyef: I use cffi, which only uses SAPs 13:39:58 And do you know how hard it isn't to wrap a SAP in an alien array header? 13:41:20 nyef: are alien strings still strings, and alien functions functions? Treating arrays as a special case looks awfully inconsistent (maybe it's still the right thing to do, however) 13:43:35 alien numbers are numbers 13:43:54 akovalenko: Alien functions aren't quite functions, but we could probably make them funcallable. 13:44:24 Alien strings... have all that deport/naturalize junk required to deal with external formats. 13:45:20 Krystof: alien numbers are numbers, but the reverse is not true: (let ((x 0)) (addr x)) ;; there is no way to make it work 13:45:57 gienah [~mwright@ppp121-44-28-16.lns20.syd6.internode.on.net] has joined #sbcl 13:47:46 akovalenko: "No way"? Really? I'll grant that if the compiler doesn't know the type of X, but beyond that it should be feasable. 13:48:24 nyef: the type of X is T, the compiler knows that much for sure. Now what? 13:49:26 akovalenko: Now ADDR should signal an error that there is no usable alien type corresponding to the lisp type T. Or just construct such a type, and throw an type error when passing it off somewhere else. 13:50:24 Now, taking the ADDR counts as closing over the variable, and if the variable isn't marked as d-x we'll need specialized value-cells, but that's still feasable. 13:52:49 -!- redline6561_nop is now known as redline6561 13:54:35 Oh, and integrating WITH-ALIEN and LET should help with some alien-boxing issues. 13:56:29 Something about intermediate values passed between alien-funcall forms being boxed unnecessarily. 13:56:52 nyef: integrating WITH-ALIEN with unboxed "lispy" values is too tight a coupling between compiler implementation and FFI. Btw, are there unboxed double-machine-word values in SBCL currently? 13:57:13 double-float on 32-bit platforms? 13:57:31 How do you mean "too tight a coupling"? 13:57:44 well, I had in mind (unsigned-byte 64) 13:58:07 Umm... Someone had a patch for... "sb-regpair", maybe? 14:00:51 nyef: what if SBCL had no support for unboxed integer variables at all? With current approach to aliens, it would have no bearing on whether we have (unsigned 8)..(unsigned 64) alien types. With any let+addr integration, every FFI type would have to be supported as unboxed type by the compiler, or so it seems 14:02:05 *akovalenko* is reminded of CFFI-based WITH-FOREIGN-VALUE and ADDR stuff that I've posted 2 weeks ago; http://paste.lisp.org/+2OGO 14:03:35 But we already have FFI types supported as unboxed types due to naturalize/deport. 14:05:25 And your worst-case scenario is having the compiler say, for some types, "unable to convert to unboxed/alien representation". 14:07:44 yep, if you add a can of non-portability worms (i.e. something that SBCL was able to convert on x86, but refuses to convert on mips, although alien type per se is valid across these archs) 14:09:46 Yes and no. Your real portability problems come with changing n-word-bits. 14:10:12 So you'll be able to cover long-long on 64-bit architectures, but not on 32-bit architectures. 14:14:57 We tend to be pretty careful about keeping things in sync between backends when doing game-changing things to the data representations. 14:15:58 Take the d-x closure / "ancestor frame" stuff. I went so far as to add the VOPs required to the HPPA backend, even though it probably doesn't even still compile. 14:18:27 -!- gienah [~mwright@ppp121-44-28-16.lns20.syd6.internode.on.net] has quit [Quit: leaving] 14:26:37 homie [~levgue@xdsl-84-44-177-189.netcologne.de] has joined #sbcl 14:40:09 ... float.pure.lisp / RANGE-REDUCTION is expected to fail? 14:42:22 tsuru [~charlie@adsl-74-179-28-14.bna.bellsouth.net] has joined #sbcl 14:58:45 I was going to try and get sbcl running on openbsd/hppa but my hardware no longer powers on 15:00:34 antgreen [user@nat/redhat/x-lzrdtiysxzndbacj] has joined #sbcl 15:05:08 Probably just as well, once you got it working, I'd be trying to convince you to make gencgc work as preparation for threading. :-P 15:06:20 maybe by the time I could manage that then openbsd would have a pthreads implementation that sbcl could use :) 15:07:28 I thought obsd didn't /want/ threading for some reason? 15:10:32 no, just lack of resources and interest 15:11:12 even now, there's really only one guy working on the new rthreads kernel bits 15:18:37 -!- sbryant- is now known as sbryant 15:34:21 Krystof: it's the same as an old intel paper, actually. 15:37:49 your reading on the subject is more complete than mine :) 15:39:29 maybe the review part is what div by mul really needs. 15:39:34 so that we stop reinventing the wheel 15:40:43 ... We keep reinventing the wheel, adding more sides each time. 15:55:29 I had a patch for lossage? 15:56:54 pkhuong: I'd tend to agree 15:58:09 nyef: do I still have a lowtag left for SSE? (: 16:08:00 Krystof: the new, and interesting thing in that post is the saturating increment step. 16:19:21 -!- oiiii [~oiiii@82.71.241.25] has quit [Remote host closed the connection] 16:23:37 -!- Blkt [~user@89-96-199-46.ip13.fastwebnet.it] has quit [Remote host closed the connection] 16:52:41 pkhuong: No, no lowtags left. Feel free to steal a couple back, though. 16:53:57 It seems that there are a couple of widetags left, though. 16:54:51 Going back to our discussion earlier, does it make sense to have (setf (fdefinition ...) ) ? 16:55:16 it doesn't work? 16:55:41 With an arbitrary alien function pointer? I wouldn't have thought so. 16:56:06 ah. I don't know that I like this at all (: 16:58:36 -!- redline6561 [~redline65@li69-162.members.linode.com] has quit [Read error: Connection reset by peer] 17:04:47 nyef: (setf (fdefinition ) ...) in order to redirect linkage table entry to a lisp callback? Then it's madness :) 17:06:01 ... "this is Sparta"? 17:07:14 Sparta won't impress anyone who has seen Soviet Russia :) 17:07:20 redline6561 [~redline65@li69-162.members.linode.com] has joined #sbcl 17:07:35 It's a movie reference. 17:07:49 *akovalenko* knows :) 17:08:11 Fair enough, then. 17:08:38 *akovalenko* just spotted a fixnum_value assumption in win32/threads runtime. Seemingly (and hopefully) a single one.. 17:14:30 It's not such a big deal on the 32-bit ports. 17:14:56 Largely because of the constraints on n-fixnum-tag-bits on such platforms. 17:17:00 "win32" is kind of misleading here (I normally test amd64 version first :) 17:17:47 Ahh. Fair enough, then. 17:59:56 antgreen` [user@nat/redhat/x-vbtkkxobxkoddltk] has joined #sbcl 18:01:27 -!- antgreen [user@nat/redhat/x-lzrdtiysxzndbacj] has quit [Ping timeout: 255 seconds] 18:40:15 rpg [~rpg@12.130.119.109] has joined #sbcl 19:17:41 antgreen [user@nat/redhat/x-agsttibeoequhxml] has joined #sbcl 19:23:36 -!- antgreen` [user@nat/redhat/x-vbtkkxobxkoddltk] has quit [Ping timeout: 240 seconds] 19:23:47 -!- rpg [~rpg@12.130.119.109] has quit [Quit: rpg] 19:59:20 -!- homie [~levgue@xdsl-84-44-177-189.netcologne.de] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 19:59:25 -!- antgreen [user@nat/redhat/x-agsttibeoequhxml] has quit [Remote host closed the connection] 20:22:20 -!- pchrist [~spirit@gentoo/developer/pchrist] has quit [Ping timeout: 252 seconds] 20:24:07 pchrist [~spirit@gentoo/developer/pchrist] has joined #sbcl 21:07:18 prxq [~mommer@mnhm-4d013b3c.pool.mediaWays.net] has joined #sbcl 21:19:39 danlarkin [~dan@danlarkin.org] has joined #sbcl 21:23:14 -!- nyef [~nyef@c-174-63-105-188.hsd1.ma.comcast.net] has quit [Quit: G'night all.] 21:37:51 -!- sdemarre [~serge@91.176.40.215] has quit [Ping timeout: 248 seconds] 22:07:03 Fare [~Fare@74.125.59.116] has joined #sbcl 22:33:58 -!- prxq [~mommer@mnhm-4d013b3c.pool.mediaWays.net] has quit [Quit: Leaving] 22:48:12 gienah [~mwright@ppp121-44-28-16.lns20.syd6.internode.on.net] has joined #sbcl 22:53:33 rpg [~rpg@216.243.156.16.real-time.com] has joined #sbcl 22:59:09 -!- tsuru [~charlie@adsl-74-179-28-14.bna.bellsouth.net] has quit [Remote host closed the connection] 23:04:59 -!- pchrist [~spirit@gentoo/developer/pchrist] has quit [Ping timeout: 276 seconds] 23:09:09 -!- rpg [~rpg@216.243.156.16.real-time.com] has quit [Quit: rpg] 23:13:55 -!- Fare [~Fare@74.125.59.116] has quit [Quit: Leaving] 23:15:03 Fare [~Fare@74.125.59.116] has joined #sbcl 23:48:44 pchrist [~spirit@gentoo/developer/pchrist] has joined #sbcl