00:16:56 -!- ltt_ [~ltt_@201.75.216.14] has quit [Quit: Textual IRC Client: www.textualapp.com] 00:30:27 -!- LiamH [~none@96.231.226.143] has quit [Quit: Leaving.] 00:34:48 drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has joined #sbcl 00:54:25 -!- edgar-rft [~GOD@HSI-KBW-109-193-013-113.hsi7.kabel-badenwuerttemberg.de] has quit [Quit: execution disconnected because no effect exists] 01:05:13 ehaliewicz [~user@50-0-51-28.dsl.static.sonic.net] has joined #sbcl 01:20:17 -!- ltbarcly [~textual@li94-204.members.linode.com] has quit [Ping timeout: 252 seconds] 01:22:23 ltt_ [~ltt_@201.75.216.14] has joined #sbcl 01:24:57 ltbarcly [~textual@li94-204.members.linode.com] has joined #sbcl 01:54:47 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 02:08:39 ... I know that I wrote this code. I know that it was quite some time ago (2012 at the LATEST), and since well before then I've tried to write code such that I could leave it alone for a while, forget ALL about it, come back to it, and be able to figure out what it does. 02:09:58 In the ARM port, in insts.lisp, in the COMPUTE-LRA instruction, there is a LABELS form. Three of the functions in this form have a parameter named, quite literally, MAGIC-VALUE, and there is no explanation given. 02:10:26 Whiskey Tango Foxtrot?!? 02:12:02 ... Ah. "magic value" is the technical term, from assem.lisp, EMIT-CHOOSER. 02:14:04 Clearly, I need to add this to my list of things to fix in SBCL. 02:27:38 -!- Bike [~Glossina@71-222-62-106.ptld.qwest.net] has quit [Ping timeout: 265 seconds] 02:29:05 Bike [~Glossina@71-222-119-201.ptld.qwest.net] has joined #sbcl 02:38:03 -!- Bike [~Glossina@71-222-119-201.ptld.qwest.net] has quit [Ping timeout: 272 seconds] 02:39:26 Bike [~Glossina@75-164-166-157.ptld.qwest.net] has joined #sbcl 02:45:39 -!- davazp [~user@70.Red-79-152-117.dynamicIP.rima-tde.net] has quit [Ping timeout: 272 seconds] 03:02:08 -!- ltt_ [~ltt_@201.75.216.14] has quit [Quit: Textual IRC Client: www.textualapp.com] 03:17:05 LiamH [~none@96.231.226.143] has joined #sbcl 03:29:40 -!- ltbarcly [~textual@li94-204.members.linode.com] has quit [Quit: Computer has gone to sleep.] 03:38:20 -!- christoph_debian [~christoph@ppp-46-244-229-3.dynamic.mnet-online.de] has quit [Ping timeout: 245 seconds] 03:52:52 christoph_debian [~christoph@ppp-46-244-239-69.dynamic.mnet-online.de] has joined #sbcl 03:55:00 wbooze [~oleo@xdsl-78-35-159-204.netcologne.de] has joined #sbcl 03:55:25 -!- oleo [~oleo@xdsl-78-35-153-18.netcologne.de] has quit [Ping timeout: 272 seconds] 03:56:21 -!- wbooze is now known as oleo 03:56:53 Vivitron [~Vivitron@c-50-172-44-193.hsd1.il.comcast.net] has joined #sbcl 04:00:23 -!- ehaliewicz [~user@50-0-51-28.dsl.static.sonic.net] has quit [Ping timeout: 272 seconds] 04:12:13 -!- nyef [~nyef@pool-70-109-134-244.cncdnh.east.myfairpoint.net] has quit [Quit: G'night all.] 04:31:31 slyrus [~chatzilla@64.129.1.15] has joined #sbcl 04:59:11 -!- angavrilov [~angavrilo@217.71.227.190] has quit [Ping timeout: 260 seconds] 04:59:27 -!- LiamH [~none@96.231.226.143] has quit [Quit: Leaving.] 05:33:20 -!- brucem [~bmitchene@waywardmonkeys.com] has quit [Quit: ZNC - http://znc.sourceforge.net] 05:34:49 brucem [~bmitchene@waywardmonkeys.com] has joined #sbcl 05:36:14 -!- |3b| [bbb@2600:3c00::f03c:91ff:fedf:5b65] has quit [Ping timeout: 240 seconds] 05:36:51 |3b|` [bbb@2600:3c00::f03c:91ff:fedf:5b65] has joined #sbcl 05:38:45 -!- |3b|` is now known as |3b| 06:00:19 ltbarcly [~textual@pool-108-42-99-156.snfcca.fios.verizon.net] has joined #sbcl 06:03:18 ltbarcly_ [~textual@li94-204.members.linode.com] has joined #sbcl 06:04:27 -!- ltbarcly [~textual@pool-108-42-99-156.snfcca.fios.verizon.net] has quit [Ping timeout: 245 seconds] 06:13:02 -!- slyrus [~chatzilla@64.129.1.15] has quit [Read error: Connection reset by peer] 06:13:37 -!- echo-area [~user@123.112.228.105] has quit [Read error: Connection reset by peer] 06:14:19 echo-area [~user@123.112.228.105] has joined #sbcl 06:14:46 slyrus [~chatzilla@64.129.1.15] has joined #sbcl 06:27:43 -!- ubii [~ubii@198.45.198.1] has quit [Quit: Leaving] 06:27:51 angavrilov [~angavrilo@217.71.227.190] has joined #sbcl 07:06:18 -!- crixus [~Rob@135-23-80-105.cpe.pppoe.ca] has quit [Quit: Leaving] 07:30:14 -!- Vivitron [~Vivitron@c-50-172-44-193.hsd1.il.comcast.net] has quit [Ping timeout: 264 seconds] 08:04:07 ehaliewicz [~user@50-0-51-28.dsl.static.sonic.net] has joined #sbcl 09:00:44 stassats [~stassats@wikipedia/stassats] has joined #sbcl 09:05:41 <|3b|> stassats: make-array :element-type stuff broke drei 09:07:01 <|3b|> it does :element-type '(or null foo) where foo is a class defined in same file 09:08:44 thanks 09:08:54 do people really expect (or null foo) to get specialized? 09:09:04 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [Remote host closed the connection] 09:09:24 *|3b|* wouldn't 09:10:30 and even as a declaration of the result of aref, it's not going to be useful 09:11:40 <|3b|> well, it could affect things like typep or typecase 09:12:14 and it's not :element-type changes, it's upgraded-array-element-type chanes 09:13:43 (make-array 10 :element-type 'unknown) is alright, since it probably checks that it's unknown before calling upgraded-array-element-type, but (make-array 10 :element-type '(or null unknown)) 09:14:05 and i changed upgraded-array-element-type to error not just with 'unknown, but with (or null unknown) too 09:15:36 <|3b|> could they error at runtime instead of compile time? 09:16:32 that's what it does, the transform for make-array just for unknown-type-p, not contains-unknown-type-p 09:16:36 *|3b|* gets no error from (make-array 10 :element-type 'unknown) 09:16:37 checks 09:16:50 <|3b|> compiling or running 09:16:59 <|3b|> '(or null unknown) errors at compile time 09:17:49 <|3b|> but returns an array when evaluated directly 09:19:08 *|3b|* is testing on "1.1.14.41-a189a69" x8664 linux 09:19:56 i wonder how many instances of unknown-type-p being used is susceptible to similar problems 09:22:15 at least the following is wrong too: (sb-kernel:type= (sb-kernel:specifier-type '(simple-array (or null unknown1))) (sb-kernel:specifier-type '(simple-array (or null unknown2)))) => T 09:22:47 forward referenced types can cause misoptimizations 09:25:55 <|3b|> yeah, i noticed (declare (type unknown x)) (if (typep x 'unknown) ...) didn't optimize as well for defclass in same file 09:26:35 there's a ticket for that 09:27:17 or two 09:27:20 lp 310120 09:27:20 https://bugs.launchpad.net/bugs/310120 09:27:40 prxq [~mommer@x2f6bfa8.dyn.telefonica.de] has joined #sbcl 09:28:08 <|3b|> cool, i don't have to make one then :) 09:30:06 lp 1082967 09:30:06 https://bugs.launchpad.net/bugs/1082967 09:30:12 that probably makes it a duplicated 09:30:13 -!- slyrus [~chatzilla@64.129.1.15] has quit [Ping timeout: 246 seconds] 09:47:54 undefined types are tricky 09:50:37 (upgraded-array-element-type '(cons unknown)) will still not signal errors 09:51:07 the rationale was for (upgraded-array-element-type '(and fixnum unknown)) to signal an error, since it returned FIXNUM, which is not exactly right 09:51:23 returning T for (cons unknown) is harmless 09:51:54 even though it can mean that unknown becomes NIL, then (upgraded-array-element-type '(cons nil nil)) => NIL 09:52:08 did i mention unknown types being tricky? 09:56:37 i see now a better reasoning, disregarding the type becoming NIL (since that wouldn't really happen to legitimate programs) 09:57:08 if unknown becomes '(and fixnum (signed-byte 8)), then FIXNUM specialization is bogus, since it now should be (signed-byte 8) 10:00:03 for upgraded-array-element-type, it's easy, since it can just error, but (make-array 10 :element-type '(and fixnum unknown)) at runtime will specialize to T, but hopefully the type will be known when the runtime comes 10:00:23 i'm thinking of making make-array error at runtime, but that may break existing code 10:01:09 |3b|: and your thing is fixed 10:30:50 -!- kludge` [~comet@unaffiliated/espiral] has quit [Ping timeout: 245 seconds] 10:33:56 kludge` [~comet@unaffiliated/espiral] has joined #sbcl 10:57:39 drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has joined #sbcl 11:02:16 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [Ping timeout: 246 seconds] 11:07:30 -!- ehaliewicz [~user@50-0-51-28.dsl.static.sonic.net] has quit [Remote host closed the connection] 12:08:00 ehaliewicz [~user@50-0-51-28.dsl.static.sonic.net] has joined #sbcl 12:35:53 -!- ehaliewicz [~user@50-0-51-28.dsl.static.sonic.net] has quit [Remote host closed the connection] 12:46:46 drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has joined #sbcl 12:51:23 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [Ping timeout: 260 seconds] 13:11:25 -!- ltbarcly_ [~textual@li94-204.members.linode.com] has quit [Ping timeout: 265 seconds] 13:21:32 ltbarcly [~textual@pool-108-42-99-156.snfcca.fios.verizon.net] has joined #sbcl 13:31:23 -!- scymtym [~user@2001:638:504:2093:baca:3aff:fe83:e736] has quit [*.net *.split] 13:31:42 scymtym [~user@2001:638:504:2093:baca:3aff:fe83:e736] has joined #sbcl 13:39:00 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 260 seconds] 13:46:33 leuler [~user@p548F8C67.dip0.t-ipconnect.de] has joined #sbcl 13:48:10 nyef [~nyef@pool-70-109-134-244.cncdnh.east.myfairpoint.net] has joined #sbcl 13:48:21 G'morning all. 13:54:02 ltt_ [~ltt_@201.75.216.14] has joined #sbcl 13:55:52 -!- ltt_ [~ltt_@201.75.216.14] has quit [Client Quit] 13:55:58 Good afternoon nyef. 14:06:56 ... Can we usefully embed cold-sbcl.map into cold-sbcl.core, in such a way that the GC can treat the addresses as additional weak pointers? 14:07:35 (Key word here is "usefully". It does us no good if we can't get at the up-to-date map when we need it.) 14:09:53 The upshot would be that we'd still have a usable map after a GC, or after core relocation, and could then push the first GC even earlier without affecting debuggability. 14:10:39 -!- luis [~luis@kerno.org] has quit [Ping timeout: 252 seconds] 14:12:58 luis [~luis@kerno.org] has joined #sbcl 14:20:15 ltt_ [~ltt_@201.75.216.14] has joined #sbcl 14:35:31 drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has joined #sbcl 14:39:52 LiamH [~none@96.231.226.143] has joined #sbcl 14:40:47 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [Ping timeout: 272 seconds] 14:45:18 -!- xymox [lechuck@s0.barwen.ch] has quit [Changing host] 14:45:18 xymox [lechuck@unaffiliated/contempt] has joined #sbcl 14:54:53 -!- ltt_ [~ltt_@201.75.216.14] has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz] 15:15:15 On the one hand, I'm glad that gdb is actually usable for working with SBCL on ARM. On the other, I'm not a fan of the UI. 15:19:22 stassats [~stassats@wikipedia/stassats] has joined #sbcl 15:45:33 ltt_ [~ltt_@201.75.216.14] has joined #sbcl 16:01:53 drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has joined #sbcl 16:13:37 -!- ltbarcly [~textual@pool-108-42-99-156.snfcca.fios.verizon.net] has quit [Quit: Computer has gone to sleep.] 16:22:55 -!- stassats [~stassats@wikipedia/stassats] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 16:27:26 -!- oleo [~oleo@xdsl-78-35-159-204.netcologne.de] has quit [Ping timeout: 246 seconds] 16:41:53 oleo [~oleo@xdsl-87-79-195-127.netcologne.de] has joined #sbcl 16:46:22 echo-are` [~user@114.254.107.157] has joined #sbcl 16:48:43 -!- echo-area [~user@123.112.228.105] has quit [Ping timeout: 272 seconds] 16:50:25 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [Remote host closed the connection] 16:55:19 ferada [~ferada@37.221.196.86] has joined #sbcl 16:59:12 hi, when I have rather lengthy assembly routine which should be funcallable, but probably not inlined every time like with define-vop, what is the best way? vop + jump to assembly routine? 17:01:16 -!- fikusz [~fikusz@catv-89-132-137-62.catv.broadband.hu] has quit [Quit: Leaving] 17:01:53 fikusz [~fikusz@catv-89-132-137-62.catv.broadband.hu] has joined #sbcl 17:09:24 "Best" is subjective, but VOP+JUMP is a plausible option. 17:09:51 Or, actually, have a look at the assembly-routines already, most of them automatically have a VOP that calls them. 17:11:02 We actually go to a little bit of trouble to disable producing a VOP for some of the routines. 17:16:53 Or, if you want to be ABLE to inline the routine, produce the VOP, don't mark it :FAST-SAFE, and produce a "stub" for it (like (defun car (x) (car x))), with appropriate declarations so that the compiler inlines it (without a :fast-safe VOP option, you'll probably need (SAFETY 0), for example), and done. The compiler usually won't inline it, it'll be funcallable, and it can be inlined if necessary. 17:19:51 Well, I think I'll try your last suggestion; otherwise, if there is no VOP, you can't actually call an assembly-routine as a regular function, or how would that work? 17:21:40 (defun foo (...) (%foo ...)) (define-vop ... (:translate %foo)) works as well 17:22:16 helps with debugging etc. 17:23:02 Have a look at the UNWIND assembly-routine. It is exposed as the function %CONTINUE-UNWIND through a :TRANSLATE option. 17:23:47 Will do. Thanks for all the suggestions. 17:24:24 ltbarcly [~textual@pool-108-42-99-156.snfcca.fios.verizon.net] has joined #sbcl 17:38:08 -!- ltt_ [~ltt_@201.75.216.14] has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz] 17:42:22 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Quit: Ping timeout: ] 17:42:49 drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has joined #sbcl 17:50:20 ltt_ [~ltt_@201.75.216.14] has joined #sbcl 18:13:47 -!- ltt_ [~ltt_@201.75.216.14] has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz] 18:16:44 ltt_ [~ltt_@201.75.216.14] has joined #sbcl 18:25:51 arm-port-2 branch updated. 18:33:01 slyrus [~chatzilla@64.129.1.15] has joined #sbcl 18:40:45 -!- ltt_ [~ltt_@201.75.216.14] has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz] 18:40:50 -!- segv- [~mb@cpeB-151.mvcable.net] has quit [Ping timeout: 245 seconds] 18:45:30 Okay, why can't the compiler coerce or load a POSITIVE-FIXNUM into an UNSIGNED-REG? 18:45:42 it can't? 18:46:13 right. so, I guess the idea is that it'll go fixnum -> signed -> unsigned 18:46:16 Well, that's what the error message says when I try to build primordial-extensions on ARM. 18:46:26 you're missing a move-vop then. 18:46:32 Probably. 18:47:48 ok. So IR1 proved the move safe, but IR2 can't implement it... because we don't chain moves between TNs, in IR2. 18:48:25 I'm going to guess that I need to add the WORD-MOVE VOP. 18:49:26 ah. yeah, that move would be hard to implement without any word-move vop (: 18:51:46 I should, in fact, probably just go through and add in all of the bits for move.lisp at this time. 19:13:39 -!- Bike [~Glossina@75-164-166-157.ptld.qwest.net] has quit [Ping timeout: 272 seconds] 19:14:53 Bike [~Glossina@67-5-213-235.ptld.qwest.net] has joined #sbcl 19:16:49 How should I interpret # ? It's clearly a constant, but which bit is the value, and what's the other side? 19:17:45 And why is the compiler trying to pass it to VOP MOVE, since it apparently isn't supposed to go there? 19:18:03 That's a nil, in constant slot 8 19:18:18 or whatever 536870908 is. 19:18:21 But it's listed as a POSITIVE-FIXNUM? 19:18:38 wait 32 bit platform? 19:18:41 With a last digit of 8, it can't be NIL. 19:18:42 Yeah. 19:19:06 Okay, what's the !1 bit in the TN representation, then? 19:19:47 that's the tn ID 19:19:51 first TN you printed, I guess. 19:19:55 Ah, okay. 19:20:15 depends on the ir2 phase. We have bogus MOVEs for a while. 19:20:42 my kingdom for more types ;) 19:22:51 Is this likely to be because it's a CONSTANT rather than an IMMEDIATE? 19:23:24 oh, yes. That sounds right. That's m-p-f? 19:24:02 $ARCH/vm.lisp: immediate-constant-sc. 19:24:50 I... might have to back-and-fill here, immediate-constant-sc only likes (unsigned-byte 8) values. 19:25:05 not mpf, it's even, but yeah, still a fixnum. 19:28:35 I don't know what I was thinking when I set up LOAD-IMMEDIATE, I can make it do a lot more than what I originally limited it to. 19:34:54 -!- ltbarcly [~textual@pool-108-42-99-156.snfcca.fios.verizon.net] has quit [Quit: Computer has gone to sleep.] 19:43:58 drmeiste_ [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has joined #sbcl 19:46:57 sdemarre [~serge@91.176.245.254] has joined #sbcl 19:49:14 emartenson__ [~loke@203.127.16.194] has joined #sbcl 19:50:31 Subfusc_ [~Subfusc@tjenen.de] has joined #sbcl 19:52:08 -!- Subfusc [~Subfusc@tjenen.de] has quit [Ping timeout: 248 seconds] 19:53:04 brucem_ [~bmitchene@waywardmonkeys.com] has joined #sbcl 19:53:29 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [*.net *.split] 19:53:30 -!- |3b| [bbb@2600:3c00::f03c:91ff:fedf:5b65] has quit [*.net *.split] 19:53:33 -!- reb [user@nat/google/x-allzaxamjmqmdtkj] has quit [*.net *.split] 19:53:33 -!- foom [~jknight@2620:15c:6:14:be30:5bff:fedf:6db6] has quit [*.net *.split] 19:53:33 -!- loke_ [~loke@203.127.16.194] has quit [*.net *.split] 19:53:35 -!- gko [gko@2400:8900::f03c:91ff:fe70:e605] has quit [*.net *.split] 19:53:42 angavrilov_ [~angavrilo@217.71.227.190] has joined #sbcl 19:56:01 ltt_ [~ltt_@201.75.216.14] has joined #sbcl 19:57:26 ppc srawi 19:57:27 Shift Right Algebraic Word Immediate: http://www.nersc.gov/vendor_docs/ibm/asm/srawi.htm#idx817 19:57:37 tsuru [~charlie@adsl-74-179-198-164.bna.bellsouth.net] has joined #sbcl 19:57:46 -!- brucem [~bmitchene@waywardmonkeys.com] has quit [Ping timeout: 260 seconds] 19:57:47 -!- daimrod [daimrod@sbrk.org] has quit [Ping timeout: 260 seconds] 19:57:50 -!- angavrilov [~angavrilo@217.71.227.190] has quit [Read error: Connection reset by peer] 19:57:53 -!- sdemarre [~serge@91.176.245.254] has quit [Read error: Connection reset by peer] 20:02:47 -!- Subfusc_ is now known as Subfusc 20:02:52 -!- kludge` [~comet@unaffiliated/espiral] has quit [Disconnected by services] 20:03:19 kludge` [~comet@unaffiliated/espiral] has joined #sbcl 20:04:06 nyef_ [~nyef@pool-70-109-134-244.cncdnh.east.myfairpoint.net] has joined #sbcl 20:04:19 Tribal [tribal@rcfreak0.com] has joined #sbcl 20:06:32 loke_ [~loke@203.127.16.194] has joined #sbcl 20:08:22 -!- Labrit [tribal@rcfreak0.com] has quit [Read error: Operation timed out] 20:08:29 -!- emartenson__ [~loke@203.127.16.194] has quit [Ping timeout: 265 seconds] 20:08:30 -!- nyef [~nyef@pool-70-109-134-244.cncdnh.east.myfairpoint.net] has quit [Ping timeout: 265 seconds] 20:08:32 -!- Krystof [~user@81.174.155.115] has quit [Ping timeout: 265 seconds] 20:13:29 Krystof_ [~user@81.174.155.115] has joined #sbcl 20:13:44 -!- nyef_ is now known as nyef 20:15:31 -!- Krystof_ is now known as Krystof 20:15:38 -!- ChanServ has set mode +o Krystof 20:17:10 When we're doing MOVE-FROM-UNSIGNED, are we sizing the bignum based on the MSB of the register that the value is in? 20:18:25 Labrit [tribal@rcfreak0.com] has joined #sbcl 20:18:48 -!- Tribal [tribal@rcfreak0.com] has quit [*.net *.split] 20:23:07 ltt__ [~ltt_@201.75.216.14] has joined #sbcl 20:33:59 LiamH1 [~none@96.231.228.133] has joined #sbcl 20:36:47 -!- ivan`` [~ivan@unaffiliated/ivan/x-000001] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 20:36:48 -!- nicdev [~user@kilimanjaro.rafpepa.com] has quit [Write error: Broken pipe] 20:36:52 -!- jaimef [jaimef@dns.mauthesis.com] has quit [*.net *.split] 20:36:55 -!- specbot [~specbot@common-lisp.net] has quit [*.net *.split] 20:37:05 -!- ltt_ [~ltt_@201.75.216.14] has quit [Ping timeout: 264 seconds] 20:37:12 -!- nyef [~nyef@pool-70-109-134-244.cncdnh.east.myfairpoint.net] has quit [Ping timeout: 264 seconds] 20:37:15 -!- oleo [~oleo@xdsl-87-79-195-127.netcologne.de] has quit [Ping timeout: 264 seconds] 20:37:18 -!- flip214 [~marek@unaffiliated/flip214] has quit [Ping timeout: 264 seconds] 20:37:23 flip214_ [~marek@86.59.100.100] has joined #sbcl 20:37:26 oleo [~oleo@xdsl-87-79-195-127.netcologne.de] has joined #sbcl 20:37:32 nicdev` [~user@kilimanjaro.rafpepa.com] has joined #sbcl 20:37:34 -!- Krystof [~user@81.174.155.115] has quit [Ping timeout: 259 seconds] 20:37:35 Vivitron [~Vivitron@c-50-172-44-193.hsd1.il.comcast.net] has joined #sbcl 20:37:50 -!- LiamH [~none@96.231.226.143] has quit [Ping timeout: 252 seconds] 20:37:54 ivan``_ [~ivan@unaffiliated/ivan/x-000001] has joined #sbcl 20:39:10 jaimef [jaimef@dns.mauthesis.com] has joined #sbcl 20:39:10 specbot [~specbot@common-lisp.net] has joined #sbcl 20:40:09 daimrod [daimrod@sbrk.org] has joined #sbcl 20:41:20 nyef [~nyef@pool-70-109-134-244.cncdnh.east.myfairpoint.net] has joined #sbcl 20:42:18 Krystof [~user@81.174.155.115] has joined #sbcl 20:42:18 -!- ChanServ has set mode +o Krystof 20:53:30 LiamH [~none@96.231.220.239] has joined #sbcl 20:54:43 -!- LiamH1 [~none@96.231.228.133] has quit [Ping timeout: 272 seconds] 20:56:09 on 64 bit? IIRC, yes. 20:56:37 ehaliewicz [~user@50-0-51-28.dsl.static.sonic.net] has joined #sbcl 21:00:39 edgar-rft [~GOD@HSI-KBW-109-193-013-113.hsi7.kabel-badenwuerttemberg.de] has joined #sbcl 21:17:13 -!- heddwch [~yoshi@76.8.3.189] has quit [Ping timeout: 246 seconds] 21:19:55 heddwch [~yoshi@76.8.3.189] has joined #sbcl 21:33:47 segv- [~mb@cpeB-151.mvcable.net] has joined #sbcl 21:35:27 LiamH1 [~none@pool-173-73-123-21.washdc.east.verizon.net] has joined #sbcl 21:36:37 -!- LiamH [~none@96.231.220.239] has quit [Ping timeout: 248 seconds] 21:37:37 Ugh. Even with the MOVE stuff sorted (or, I should say, apparently sorted) the compiler STILL can't handle DATA-VECTOR-REF/SIMPLE-BASE-STRING. /-: 21:38:32 This time something more fundamental, at least. Less magic to contend with. 21:44:37 -!- segv- [~mb@cpeB-151.mvcable.net] has quit [Ping timeout: 248 seconds] 22:04:10 Oops. I seem to have screwed up with negative fixnums in LOAD-IMMEDIATE. 22:10:57 thepreacher [~thepreach@87.112.72.196] has joined #sbcl 22:14:58 -!- slyrus [~chatzilla@64.129.1.15] has quit [Ping timeout: 260 seconds] 22:29:42 VOPs have to state every temporary in advance, right? 22:30:58 Usually, yes. 22:32:13 There are times when you can play fast and loose with the rules, but those times are few and far between, and recognizing them is a matter of experience. 22:37:37 -!- thepreacher [~thepreach@87.112.72.196] has quit [Quit: Leaving] 22:41:31 Ah well, one thing I don't have. This may not be the best idea anyway: I was trying to unpack a set of arguments from ":more t" to registers, but I'm getting an error when I use the VOP. 22:48:03 ltbarcly [~textual@pool-108-42-99-156.snfcca.fios.verizon.net] has joined #sbcl 22:48:58 -!- leuler [~user@p548F8C67.dip0.t-ipconnect.de] has quit [Quit: ERC Version 5.1.2 $Revision: 1.796.2.6 $ (IRC client for Emacs)] 22:50:14 why do you think you want to do that? 22:50:46 -!- edgar-rft [~GOD@HSI-KBW-109-193-013-113.hsi7.kabel-badenwuerttemberg.de] has quit [Quit: happening discontinued because all hope lost] 22:51:09 ltbarcly_ [~textual@li94-204.members.linode.com] has joined #sbcl 22:53:02 -!- ltbarcly [~textual@pool-108-42-99-156.snfcca.fios.verizon.net] has quit [Ping timeout: 264 seconds] 22:54:53 So that I only touch the actually necessary registers on a syscall. It looks like the alien funcall machinery might be more appropriate in the end, but it'll take some time until I understand that part enough. 22:55:55 It's just syscall as a VOP at the moment. 23:00:53 Ahh. 23:01:07 You have an IR2-CONVERT optimizer defined for this thing? 23:01:57 Umm... And maybe an LTN-ANNOTATE optimizer? 23:03:19 What you want to do is do something similar to what ALIEN-FUNCALL does in laying out the register parameters, so that the compiler will have them all in place prior to passing them to your VOP, and then your VOP just :IGNOREs the args since they're already where they need to be. 23:03:49 I'd generate a couple VOPs for however many arguments your syscalls can accept. 23:04:14 Or that, yeah. 23:04:50 It's not like you don't have a reasonable upper bound for number of arguments. 23:05:17 more reasonable on some OSes than others ;) 23:05:19 ... You're generating syscalls from SBCL? 23:05:24 That's the easiest way, yes. 23:05:43 This sounds like a neat project. 23:05:52 Yes, just an idea, I'll see how far that goes. 23:06:00 Less libc, more lisp (or so ...) 23:06:06 There's a CMUCL branch that replaced a lot of libc calls with direct syscells. 23:06:27 No impact on performance, but cute hack. Check in martin cracauer's CVS trees. 23:06:29 (Okay, it's also a project that I've looked into a time or to myself, though I don't recall getting as far as making VOPs for it, more just doing the feasability study.) 23:06:37 This https://github.com/xach/cmucl-direct-syscalls ? 23:06:59 kthx 23:07:55 sounds plausible. 23:10:27 -!- LiamH1 is now known as LiamH 23:16:27 -!- Bike [~Glossina@67-5-213-235.ptld.qwest.net] has quit [Ping timeout: 276 seconds] 23:18:05 Bike [~Glossina@71-222-33-141.ptld.qwest.net] has joined #sbcl 23:46:45 -!- Bike [~Glossina@71-222-33-141.ptld.qwest.net] has quit [Ping timeout: 248 seconds] 23:48:48 Bike [~Glossina@75-164-162-106.ptld.qwest.net] has joined #sbcl 23:49:55 -!- ltt__ [~ltt_@201.75.216.14] has quit [Quit: Textual IRC Client: www.textualapp.com] 23:56:54 segv- [~mb@cpeB-151.mvcable.net] has joined #sbcl