00:01:09 psilord [~psilord@c-69-180-173-249.hsd1.mn.comcast.net] has joined #sbcl 00:10:34 crixxus [~Rob@135-23-80-105.cpe.pppoe.ca] has joined #sbcl 00:11:35 -!- crixus [~Rob@135-23-80-105.cpe.pppoe.ca] has quit [Ping timeout: 260 seconds] 00:15:04 -!- Quadrescence [~quad@unaffiliated/quadrescence] has quit [Quit: This computer has gone to sleep] 00:27:11 -!- crixxus [~Rob@135-23-80-105.cpe.pppoe.ca] has quit [Ping timeout: 272 seconds] 00:51:47 -!- segv- [~mb@95-91-241-76-dynip.superkabel.de] has quit [Remote host closed the connection] 00:57:25 -!- senj [~senj@unaffiliated/senj] has quit [Quit: Sleep Now] 01:07:51 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [Remote host closed the connection] 01:09:32 drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has joined #sbcl 01:21:14 crixus [~Rob@135-23-80-105.cpe.pppoe.ca] has joined #sbcl 01:24:48 hargettp [~hargettp@c-76-119-232-173.hsd1.ma.comcast.net] has joined #sbcl 01:27:15 -!- crixus [~Rob@135-23-80-105.cpe.pppoe.ca] has quit [Ping timeout: 260 seconds] 01:27:25 crixus [~Rob@135-23-80-105.cpe.pppoe.ca] has joined #sbcl 01:28:57 ebrasca [~chatzilla@host-198.58.elzappero.net] has joined #sbcl 01:41:25 -!- crixus [~Rob@135-23-80-105.cpe.pppoe.ca] has quit [Ping timeout: 248 seconds] 01:50:26 -!- hargettp [~hargettp@c-76-119-232-173.hsd1.ma.comcast.net] has quit [Quit: Linkinus - http://linkinus.com] 01:51:09 -!- Bike [~Glossina@wl-nat101.it.wsu.edu] has quit [Ping timeout: 272 seconds] 01:53:52 -!- ASau` is now known as ASau 01:57:27 -!- oleo [~oleo@xdsl-87-79-199-36.netcologne.de] has quit [Ping timeout: 252 seconds] 01:58:46 oleo [~oleo@xdsl-78-35-166-156.netcologne.de] has joined #sbcl 02:01:13 yacks [~py@103.6.159.103] has joined #sbcl 02:02:52 Bike [~Glossina@gannon-wless-gw.resnet.wsu.edu] has joined #sbcl 02:14:26 echo-area [~user@182.92.247.2] has joined #sbcl 02:14:33 -!- loke_ [~loke@203.127.16.194] has quit [Remote host closed the connection] 02:15:35 pranavrc [~pranavrc@122.164.235.134] has joined #sbcl 02:15:35 -!- pranavrc [~pranavrc@122.164.235.134] has quit [Changing host] 02:15:35 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 02:17:51 loke_ [~loke@203.127.16.194] has joined #sbcl 02:24:05 crixus [~Rob@135-23-80-105.cpe.pppoe.ca] has joined #sbcl 02:38:17 -!- crixus [~Rob@135-23-80-105.cpe.pppoe.ca] has quit [Ping timeout: 272 seconds] 02:38:40 -!- slyrus [~chatzilla@173-228-44-92.dsl.static.sonic.net] has quit [Ping timeout: 246 seconds] 02:39:03 Quadrescence [~quad@unaffiliated/quadrescence] has joined #sbcl 02:42:33 crixus [~Rob@135-23-80-105.cpe.pppoe.ca] has joined #sbcl 02:44:56 prxq_ [~mommer@x2f6c5ef.dyn.telefonica.de] has joined #sbcl 02:48:48 -!- prxq [~mommer@x2f6a59a.dyn.telefonica.de] has quit [Ping timeout: 272 seconds] 02:53:28 -!- Quadrescence [~quad@unaffiliated/quadrescence] has quit [Quit: Leaving] 02:55:17 davazp [~user@14.Red-79-152-116.dynamicIP.rima-tde.net] has joined #sbcl 02:56:37 -!- crixus [~Rob@135-23-80-105.cpe.pppoe.ca] has quit [Ping timeout: 248 seconds] 02:59:13 -!- ebrasca [~chatzilla@host-198.58.elzappero.net] has quit [Remote host closed the connection] 03:23:36 mogglebam [~marco@99-95-170-156.lightspeed.sndgca.sbcglobal.net] has joined #sbcl 03:24:04 Hi, I need to redefine a vop. 03:24:37 How do I do that and then recompile code based on the changed definition? 03:24:59 Specifically I'm working with %read-cycle-counter 03:26:22 I want to remove the cpuid calls and see how that affects the cycle count 03:30:10 crixus [~Rob@135-23-80-105.cpe.pppoe.ca] has joined #sbcl 03:37:59 -!- christoph_debian [~christoph@ppp-46-244-238-187.dynamic.mnet-online.de] has quit [Read error: Operation timed out] 03:44:24 -!- crixus [~Rob@135-23-80-105.cpe.pppoe.ca] has quit [Ping timeout: 260 seconds] 03:52:08 christoph_debian [~christoph@ppp-46-244-227-234.dynamic.mnet-online.de] has joined #sbcl 04:05:31 -!- LiamH [~none@96.231.217.60] has quit [Quit: Leaving.] 04:08:37 -!- scymtym_ [~user@ip-5-147-115-29.unitymediagroup.de] has quit [Ping timeout: 248 seconds] 04:11:40 crixus [~Rob@135-23-80-105.cpe.pppoe.ca] has joined #sbcl 04:23:19 attila_lendvai [~attila_le@5.76.185.228] has joined #sbcl 04:23:19 -!- attila_lendvai [~attila_le@5.76.185.228] has quit [Changing host] 04:23:19 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 04:26:03 -!- crixus [~Rob@135-23-80-105.cpe.pppoe.ca] has quit [Ping timeout: 272 seconds] 04:38:26 ubii [~ubii@198.45.198.1] has joined #sbcl 04:38:26 -!- ubii [~ubii@198.45.198.1] has quit [Changing host] 04:38:26 ubii [~ubii@unaffiliated/ubii] has joined #sbcl 05:03:08 -!- yacks [~py@103.6.159.103] has quit [Read error: Connection reset by peer] 05:05:05 mogglebam: compile your new VOP definitions, compile the circular defun if any, compile any use of %r-c-c. 05:06:25 More practical (i.e. not vulnerable to out of order execution) way to estimate the overhead caused by cpuid: %r-c-c back to back, without anything between. 05:07:09 yacks [~py@103.6.159.103] has joined #sbcl 05:11:23 pkhuong: thankyou. I'm confused though -- when I'm in-package SB-VM, I get The function SB-VM::%READ-CYCLE-COUNTER is undefined. 05:11:37 When I try to call (%r-c-c) 05:12:23 There's no circular defun for %r-c-c, (defun %read-cycle-counter () (%read-cycle-counter)). 05:14:51 Ohhh 05:15:02 Thankyou. Wow, I clearly do not understand how VOPs get called. 05:16:06 I've been trying to read through this http://common-lisp.net/project/cmucl/doc/CMUCL-design.pdf 05:16:18 because it seems like the most complete documentation on SBCL's guts 05:16:23 but it is slow going 05:16:23 they're just instruction templates for the compiler, i thought? 05:18:26 Afaict yes, but I don't really understand the compiler either 05:18:34 -!- davazp [~user@14.Red-79-152-116.dynamicIP.rima-tde.net] has quit [Remote host closed the connection] 05:19:13 well i mean, it's only "called" when the compiler sees (%read-cycle-counter), which makes it put the vop, but that means there's no actual %read-cycle-counter functcion, which is what the defun is for. 05:21:09 *mogglebam* nods 05:27:15 ltbarcly [~textual@pool-108-42-99-156.snfcca.fios.verizon.net] has joined #sbcl 05:34:52 omg1314 [~omg1314@61.135.169.73] has joined #sbcl 05:54:43 -!- oleo [~oleo@xdsl-78-35-166-156.netcologne.de] has quit [Quit: Leaving] 06:06:07 -!- omg1314 [~omg1314@61.135.169.73] has quit [Ping timeout: 272 seconds] 06:18:30 -!- ltbarcly [~textual@pool-108-42-99-156.snfcca.fios.verizon.net] has quit [Quit: Computer has gone to sleep.] 06:54:05 -!- prxq_ is now known as prxq 07:09:07 -!- yacks [~py@103.6.159.103] has quit [Quit: Leaving] 07:12:02 angavrilov [~angavrilo@217.71.227.190] has joined #sbcl 07:14:21 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [Remote host closed the connection] 07:22:51 -!- mogglebam [~marco@99-95-170-156.lightspeed.sndgca.sbcglobal.net] has quit [Ping timeout: 265 seconds] 07:32:15 Hi, I'm facing this issue - https://github.com/cl-babel/babel/issues/12 while trying to load babel. One of the suggested solutions to the problem seems to be trying to clear out the old version of alexandria from .sbcl/site/, but my .sbcl folder has only one file (trusted-uids.lisp). How do I clear out alexandria and install a new version? 07:32:57 I'll ask on #lisp as well 07:53:29 jewel [~jewel@105-237-68-160.access.mtnbusiness.co.za] has joined #sbcl 07:59:38 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Ping timeout: 264 seconds] 08:24:01 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 09:02:59 drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has joined #sbcl 09:08:31 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [Ping timeout: 272 seconds] 09:19:33 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Ping timeout: 272 seconds] 09:23:05 attila_lendvai [~attila_le@5.76.185.228] has joined #sbcl 09:23:06 -!- attila_lendvai [~attila_le@5.76.185.228] has quit [Changing host] 09:23:06 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 09:31:04 ltbarcly [~textual@pool-108-42-99-156.snfcca.fios.verizon.net] has joined #sbcl 09:32:45 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Quit: Ping timeout: ] 09:36:17 -!- ltbarcly [~textual@pool-108-42-99-156.snfcca.fios.verizon.net] has quit [Ping timeout: 272 seconds] 10:07:08 -!- echo-area [~user@182.92.247.2] has quit [Remote host closed the connection] 10:31:01 -!- kludge` [~comet@unaffiliated/espiral] has quit [Ping timeout: 248 seconds] 10:37:06 kludge` [~comet@unaffiliated/espiral] has joined #sbcl 10:52:06 drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has joined #sbcl 10:56:45 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [Ping timeout: 252 seconds] 11:15:14 fiveop [~fiveop@p5DDC458B.dip0.t-ipconnect.de] has joined #sbcl 11:35:39 poglesbyg_ [~poglesbyg@pc-116-91.eduroam.uib.no] has joined #sbcl 11:41:19 scymtym: are you there? 11:42:37 oh, actually, I've just found the missing piece 11:43:12 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 11:46:23 the answer to "why not :defined" in ftype redefinitions is I think that a :defined definition isn't in fact a declaration at all; it's recording what we think the current ftype of the function object is 11:46:43 so it's not in any way condition-worthy to change its type 11:53:34 Hi, I'm facing this issue - https://github.com/cl-babel/babel/issues/12 while trying to load babel. One of the suggested solutions to the problem seems to be trying to clear out the old version of alexandria from .sbcl/site/, but my .sbcl folder has only one file (trusted-uids.lisp). How do I clear out alexandria and install a new version? 11:55:12 look at .cache/common-lisp/ 11:57:58 fiveop, tried removing alexandria from there as well, but I still get the error. update-all-dists seems to tell me that I have the latest version of quicklisp 11:59:50 Everything was fine until I updated ubuntu from 11.10 to 12.04 12:04:55 you could try find ~ -name '*.fasl' | xargs rm 12:05:10 (remove all fasl files under your home directory) 12:05:43 I seem to remember some thing might put fasls under /var/ somewhere too; you could try chasing those down 12:07:21 -!- poglesbyg_ [~poglesbyg@pc-116-91.eduroam.uib.no] has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz] 12:12:09 Krystof, removed all fasl files from ~ and /var/ but no luck :( 12:13:00 I wish I could give more information, but that's all I have 12:16:20 -!- DeadTrickster [~dead@62.122.188.214] has quit [Remote host closed the connection] 12:17:02 poglesbyg_ [~poglesbyg@pc-116-91.eduroam.uib.no] has joined #sbcl 12:17:10 I tried running (alexandria:format-symbol t '#:~a-code-point-counter "UTF16") as suggested at https://github.com/cl-babel/babel/issues/9#issuecomment-4430276, but that returns an error as well - #:~A-CODE-POINT-COUNTER fell through ETYPECASE expression. Wanted one of (STRING SIMPLE-STRING). 12:18:49 dunno. Not an sbcl problem, though 12:18:52 good luck 12:19:11 alright thanks 12:27:24 -!- poglesbyg_ [~poglesbyg@pc-116-91.eduroam.uib.no] has quit [Quit: Textual IRC Client: www.textualapp.com] 12:40:44 drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has joined #sbcl 12:45:08 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [Ping timeout: 252 seconds] 13:00:24 pranavrc: debian packaged sbcl's used to have common-lisp-controller which used to cause a lot of headache... not sure if that is still the case 13:02:02 attila_lendvai, oh. Let me try installing from source. I have 1.0.55 debian right now 13:02:08 thanks for the heads up 13:20:04 yacks [~py@103.6.159.103] has joined #sbcl 13:41:23 LiamH [~none@96.231.217.60] has joined #sbcl 14:03:27 segv- [~mb@95-91-241-32-dynip.superkabel.de] has joined #sbcl 14:05:50 -!- fiveop [~fiveop@p5DDC458B.dip0.t-ipconnect.de] has quit [] 14:29:28 drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has joined #sbcl 14:33:53 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [Ping timeout: 245 seconds] 14:48:00 oleo [~oleo@xdsl-78-35-130-3.netcologne.de] has joined #sbcl 15:03:10 slyrus [~chatzilla@107.201.5.56] has joined #sbcl 15:06:53 Krystof: thanks for the explanation earlier 15:12:16 -!- LiamH [~none@96.231.217.60] has quit [Quit: Leaving.] 15:17:29 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Ping timeout: 242 seconds] 15:20:34 nyef [~nyef@pool-108-7-220-91.bstnma.fios.verizon.net] has joined #sbcl 15:20:44 G'morning all. 15:24:03 crixus [~Rob@135-23-80-105.cpe.pppoe.ca] has joined #sbcl 15:24:31 hi nyef 15:26:32 attila_lendvai [~attila_le@5.76.185.228] has joined #sbcl 15:26:32 -!- attila_lendvai [~attila_le@5.76.185.228] has quit [Changing host] 15:26:32 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 15:31:25 So, I've run into a little bit of a register pressure situation that I, as yet, haven't figured out how to handle. 15:32:16 I've got wired and component-live TNs filling up all of my unboxed registers coming up on CALL-OUT, and there's a MOVE-SAP VOP that needs a LOAD-TN slot. 15:33:08 My best guess is that this is for the FUNCTION argument to CALL-OUT, which is (for register pressure reasons) stored on SAP-STACK anyway. 15:35:10 As near as I can tell, the VOP is going to do something like "SAP-MOVE t109[NS0]>t108[NL2] => t109[NS0] ... My bad, that first t109 should have been a t31. 15:37:25 -!- slyrus [~chatzilla@107.201.5.56] has quit [Ping timeout: 246 seconds] 15:37:59 The TN identity is different, but it's LOCATION= even before the register has been assigned. 15:38:44 -!- crixus [~Rob@135-23-80-105.cpe.pppoe.ca] has quit [Ping timeout: 265 seconds] 15:46:37 slyrus [~chatzilla@107.201.5.56] has joined #sbcl 15:58:08 nyef: load-if? 15:58:36 The TN seems to get assigned, even if it never gets used. 15:58:51 load-if is already in play. 15:59:21 crixus [~Rob@135-23-80-105.cpe.pppoe.ca] has joined #sbcl 15:59:58 milanj [~milanj@82.117.199.26] has joined #sbcl 16:01:42 (AFAICT, load-if merely suppresses code generation, it doesn't affect lifetimes or anything like that.) 16:01:55 oh yeah, too late 16:04:21 -!- crixus [~Rob@135-23-80-105.cpe.pppoe.ca] has quit [Ping timeout: 265 seconds] 16:06:55 Is this SAP-MOVE being emitted by EMIT-COERCE-VOP in represent? 16:06:56 drmeister [~drmeister@166.170.21.65] has joined #sbcl 16:09:29 no clue. Have you looked at the ir2tran? 16:10:00 I don't see any direct references, and it IS a MOVE-VOP. 16:10:17 you wouldn't see a direct reference. 16:10:20 only 'move 16:10:35 Not even seeing that, AFAICT. 16:10:36 (whichever ir2-convert method can insert call-out vops) 16:10:56 The %ALIEN-FUNCALL IR2-CONVERT optimizer in aliencomp. 16:11:38 Nope, just a direct reference to (LVAR-TN CALL BLOCK FUNCTION). 16:11:39 -!- drmeister [~drmeister@166.170.21.65] has quit [Read error: Connection reset by peer] 16:11:52 In the VOP* form for CALL-OUT. 16:12:21 rpg [~rpg@23-25-144-217-static.hfc.comcastbusiness.net] has joined #sbcl 16:14:53 One of the problems is that we don't know if the register TNs are live or not until we're in the middle of PACK. 16:16:13 Hrm. What if CALL-OUT can take the TN in either SC? 16:17:30 perhaps. 16:18:42 forcing most args to the stack in make-call-out-tn might also work :\ 16:18:57 unspill them in call-out. 16:18:58 That's exactly the route that I didn't want to go. 16:19:22 Although unspilling within the VOP would reduce the complexity... 16:19:47 No, they still end up processed by MOVE-ARG. That's no good. 16:20:12 yeah they get move-arged, but to the stack 16:20:37 Which means that the stack frame allocation will need adjusting in call-out as well. 16:22:23 but.. why do you have so few registers available? 16:24:12 There are only sixteen registers total, including the program counter, we need an interior pointer for the link register, we need the current stack and frame pointers, we need null in a register, and we need reg_CODE. 16:24:29 That leaves ten registers for descriptor-regs and non-descriptor-regs. 16:24:32 crixus [~Rob@69.77.176.98] has joined #sbcl 16:26:07 Function calling requires at least three arg-passing registers (a limitation of the SBCL stack model on systems with split stacks), plus at least two other boxed registers for things like the function to call and such. 16:26:27 That leaves five non-descriptor-regs, which is just *barely* enough most of the time. 16:27:30 Ugh. It's still trying to pack a load-tn. But possibly for a different VOP? 16:29:56 I might need to do something horrible to this IR2-CONVERT optimizer. 16:33:20 -!- gko [gko@2400:8900::f03c:91ff:fe70:e605] has quit [Ping timeout: 245 seconds] 16:35:58 And I just got a build, having done something horrible to %ALIEN-FUNCALL IR2-CONVERT, so that's progress. 16:35:59 gko [gko@2400:8900::f03c:91ff:fe70:e605] has joined #sbcl 16:42:13 MORE HORRIBLE THINGS 16:43:24 Okay, what I do is reverse ARGS and ARG-TNS so that it moves the stack arguments into place first, so that the wired register args aren't live while that's happening. 16:49:55 drmeister [~drmeister@155.247.96.196] has joined #sbcl 16:50:53 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Ping timeout: 240 seconds] 16:57:15 -!- jewel [~jewel@105-237-68-160.access.mtnbusiness.co.za] has quit [Ping timeout: 260 seconds] 17:01:57 And I think I got it. The compiler is generating fairly stupid code for the stack args overall, but it's at least generating code. 17:04:37 -!- yacks [~py@103.6.159.103] has quit [Quit: Leaving] 17:13:16 fiveop [~fiveop@p5DDC458B.dip0.t-ipconnect.de] has joined #sbcl 17:16:16 Hello fiveop. 17:19:42 hi 17:20:53 rpg_ [~rpg@198-74-7-110.fttp.usinternet.com] has joined #sbcl 17:22:25 fiveop: I got the register pressure during alien-funcall down to the point where it should be able to handle any number of arguments without choking. 17:22:38 I'm putting the commits together now. 17:22:58 did you see my new commits? 17:23:12 I cloned your repository last night. 17:23:17 I made sort of a log as well 17:23:23 Nice. 17:24:56 -!- rpg [~rpg@23-25-144-217-static.hfc.comcastbusiness.net] has quit [Ping timeout: 252 seconds] 17:25:30 https://gist.github.com/anonymous/dbd653610f26d09a9964 17:26:08 Nice. 17:26:26 If you could commit in the next 5 hours I can take your changes with me on a train ride tomorrow 17:27:01 -could 17:27:39 Okay, so it seems to me that the big remaining bits are static-funs, full-call assembly routines, the various arithmetic assembly routines, assorted template VOPs, and floating-point. 17:28:43 You're already kicking ass with the template VOPs, and the static-funs and full-call assembly routine support are both variations on the basic function calling mechanism 17:28:48 nyef: clever, re re-ordering 17:35:06 -!- rpg_ [~rpg@198-74-7-110.fttp.usinternet.com] has quit [Ping timeout: 252 seconds] 17:35:58 nyef: threads 17:36:09 Out of scope for the initial version. 17:36:23 The target is a functioning cheneygc system to start with. 17:36:34 ok 17:37:18 crixxus [~Rob@69.77.176.98] has joined #sbcl 17:37:37 So I'll continue to go through build-order and look for missing vops 17:38:08 Once we have working cheneygc we can start doing things like adding callbacks, moving to gencgc (required for threads), enabling dynamic-extent, messing about with predicated code (the cmove stuff that x86oids use is just the start of what we might be able to do with ARM). 17:38:59 Another approach you might take, rather than just slogging through build-order, is to compare the arm backend files to other platforms to see what VOPs they have that ARM doesn't. 17:39:11 Two days ago I tried to use PRINT and LIST in arm-bootstrap which both yield a seg fault. How do I go about debugging that? 17:39:32 nyef: that's correct. but then I get more than I need. 17:39:44 Hrm. A segfault? That's odd... 17:40:11 Well, PRINT probably wouldn't work anyway, because the streams wouldn't be set up yet. 17:40:23 (cons 1 2) worked though 17:40:31 -!- crixus [~Rob@69.77.176.98] has quit [Ping timeout: 252 seconds] 17:41:18 I have only a rough idea, what should and should not work. And it comes from looking at the file names in build-order 17:41:58 Have a look in ir2tran for code that's disabled on ARM. And then look at arm/alloc and compare it to the other backends. 17:43:08 why? 17:43:28 LIST and LIST* should be there. 17:44:01 I found it 17:44:37 ir2tran was also where I got in trouble at the end the last time 17:44:45 %pop-values is not implemented 17:45:01 That's another VOP, of course. 17:45:08 that has to do with stack pointers 17:45:22 Yes, as does %NIP-VALUES. 17:45:38 arithmetic VOPs are rather self explanatory; stack related stuff isn't if you don't know the details 17:45:45 Fair point. 17:46:01 I might not get to those today, but I'll see what I can do. 17:46:28 by the end of the year we might have something working :) 17:46:45 You mis-spelled "month", btw. d-: 17:47:45 :) 17:48:19 Seriously, my main worry at this point is the floating-point stuff. 17:50:02 have you figured which version you want to support? 17:50:24 I would like to support my RasperryPi 17:50:33 Same here. 17:50:39 That... might be vfp1. 17:51:11 Which has the earliest version of floating point operations of any ARM as far as I know. 17:51:27 Sounds about right. 17:52:07 I'll look at it later. Off to learn chinese :) 17:52:11 Have fun. 17:52:38 ah right, it's deprecated 17:52:46 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Quit: Ping timeout: ] 17:53:03 -!- milanj [~milanj@82.117.199.26] has quit [Ping timeout: 245 seconds] 17:53:43 After we get the Pi going, I'll consider the beaglebone... Or my GN10.1 again. 17:54:05 deprecated, and incompatible with eabi. joy. 17:55:38 I've mostly stopped paying attention to these little details until they actually prevent me from making any progress (like the syscall scheme changing at one point). 17:55:59 senj [~senj@unaffiliated/senj] has joined #sbcl 17:57:25 I gave it an hour and ran away in disgust. I have no time for sbcl anyway (: 17:57:51 I'm almost at that point myself, unfortunately. /-: 17:58:14 I really should be doing actual paying work right now. 18:03:49 ltbarcly [~textual@pool-108-42-99-156.snfcca.fios.verizon.net] has joined #sbcl 18:18:27 fiveop: I just pushed my changes to my public repo, including the %POP-VALUES bit. Nowhere near as tough as %NIP-VALUES is going to be. 18:26:49 Okay, apparently we do have VFP2, but we might need to support both soft and hard float ABIs. 18:37:27 pnpuff [~harmonic@unaffiliated/pnpuff] has joined #sbcl 18:59:19 rpg [~rpg@23-25-144-217-static.hfc.comcastbusiness.net] has joined #sbcl 19:37:28 pkhuong: aren't you finishing your phd right now? 19:38:19 fiveop: no, it's finished. 19:38:31 and now 19:39:39 work at appnexus starting in february. 20:20:07 jewel [~jewel@105-237-68-160.access.mtnbusiness.co.za] has joined #sbcl 20:29:35 Vivitron [~Vivitron@c-50-172-44-193.hsd1.il.comcast.net] has joined #sbcl 20:40:15 -!- senj [~senj@unaffiliated/senj] has quit [Quit: Sleep Now] 20:48:53 -!- jewel [~jewel@105-237-68-160.access.mtnbusiness.co.za] has quit [Ping timeout: 272 seconds] 20:51:51 jewel [~jewel@105-237-68-160.access.mtnbusiness.co.za] has joined #sbcl 20:52:36 host-1+host-2 now takes two minutes on my machine 20:56:24 We're getting there. 20:56:48 Are you into src/compiler yet? (-: 20:58:26 Oh, and PRINT might start working once you're past fd-streams and can run stream-cold-init-or-reset. 20:58:31 Might not, though. 20:59:18 -!- jewel [~jewel@105-237-68-160.access.mtnbusiness.co.za] has quit [Ping timeout: 245 seconds] 20:59:58 Just being able to do a WRITE-LINE to *standard-output* would be a nice milestone, btw. 21:08:02 nyef: almost 21:09:25 From now on, I'll put FIXME-ARM comments in code, at points where I am not at all sure if I do the right thing, for you to review. 21:10:15 preferably you should review everything, but those will be the points where I am particularly unsure :) 21:10:42 I don't have time or patience right now to review everything, I'm afraid. 21:11:01 In fact, right-right now I'm on the clock. /-: 21:11:22 I am talking about whenever you pull 21:11:33 My immediate plans for the ARM port are possibly nip-values and starting to rough in floating-point stuff, FWIW. 21:12:00 I figure that you have half a chance at making the static-fun and full-call assembly-routine stuff work yourself. 21:16:44 will this come up, during my way through build-order? 21:17:07 (I added '&& mplayer bing.mp3' to my build command now ...) 21:18:28 It probably will come up, and failing that it'll come up once you get to cold-init, at which point you'll need to kill the arm-bootstrap file itself and actually see how far the world spins when you give it a shove. (-: 21:28:34 :) 21:34:32 Once we're at that point, the debugging fun REALLY begins. 21:38:57 I get a value of (2^32-1) for Y (declared (:constant fixnum)) in FAST-LOGAND-C/FIXNUM=>FIXNUM. I didn't expect that 21:59:45 How can I disable forms in src/code/... only for the host-2 step? 22:00:29 Something to do with sbcl-xc-host. 22:02:07 I'll try #+sb-xc-host 22:11:18 rpg_ [~rpg@198-74-7-110.fttp.usinternet.com] has joined #sbcl 22:11:29 -!- rpg_ [~rpg@198-74-7-110.fttp.usinternet.com] has quit [Client Quit] 22:11:52 -!- rpg [~rpg@23-25-144-217-static.hfc.comcastbusiness.net] has quit [Read error: Connection reset by peer] 22:12:00 rpg [~rpg@23.25.144.217] has joined #sbcl 22:12:17 rpg_ [~rpg@198-74-7-110.fttp.usinternet.com] has joined #sbcl 22:12:20 -!- rpg_ [~rpg@198-74-7-110.fttp.usinternet.com] has quit [Client Quit] 22:16:14 -!- rpg [~rpg@23.25.144.217] has quit [Ping timeout: 240 seconds] 22:46:00 -!- pnpuff [~harmonic@unaffiliated/pnpuff] has quit [] 23:29:48 Do we want to consider supporting FPA ARM systems, or just stick with the VFP systems? 23:30:04 And what about systems with no FPU at all? 23:30:26 Does SBCL support any targets without an FPU? 23:30:28 -s 23:31:03 AFAIK, it does not, as of yet. 23:31:19 But that doesn't mean that it couldn't. 23:33:45 we would have to implement floating point operations in software 23:33:58 I think that would be at the very bottom of my wishlist 23:37:05 -!- segv- [~mb@95-91-241-32-dynip.superkabel.de] has quit [Ping timeout: 272 seconds] 23:37:08 arm/move.lisp can not be x compiled, because of an error in arm/move ;) 23:37:58 Fun. Anything I can help with? 23:39:52 -!- angavrilov [~angavrilo@217.71.227.190] has quit [Remote host closed the connection] 23:41:52 no, I'm going to bed 23:42:02 Sleep well. 23:44:43 one last thing 23:44:45 (inst add ,reg null-tn (static-symbol-offset ,symbol)) 23:44:58 can static-symbol-offset be larger than 255? :) 23:45:13 (that's from load-symbol in macros) 23:45:59 well, I think that's my problem. will fix it tomorrow 23:46:00 -!- fiveop [~fiveop@p5DDC458B.dip0.t-ipconnect.de] has quit [] 23:46:33 Yeah, that's probably it. 23:47:10 And, of course, the usual solution is to build up the displacement in parts either in LIP or in the target. 23:47:58 ... 1023 is the more likely limit, however, as the low two bits of static-symbol-offset are expected to be zero. 23:50:24 Umm... symbols are, what, six words long? That'd be a practical maximum of (floor 1024 6) => 170 static symbols. Seems somehow unlikely. 23:50:52 Even if it is an eight-bit displacement, 42 static symbols STILL seems a touch unlikely. 23:51:28 Okay, 55 symbols on x86-64/linux. I sit corrected. 23:53:33 scymtym_ [~user@ip-5-147-115-29.unitymediagroup.de] has joined #sbcl 23:54:08 segv- [~mb@95-91-241-32-dynip.superkabel.de] has joined #sbcl 23:55:51 ASau` [~user@p54AFFE57.dip0.t-ipconnect.de] has joined #sbcl 23:57:39 Yeah, low two bits clear, so up to 1020, thus 169 symbols or so. Entirely not likely to be the culprit. 23:59:37 -!- ASau [~user@p54AFF1B3.dip0.t-ipconnect.de] has quit [Ping timeout: 272 seconds] 23:59:44 -!- ASau` is now known as ASau