00:09:54 -!- segv- [~mb@95-91-241-60-dynip.superkabel.de] has quit [Remote host closed the connection] 00:12:48 -!- echo-area [~user@114.254.101.235] has quit [Remote host closed the connection] 00:15:42 -!- robgssp [~user@cpe-24-93-28-218.rochester.res.rr.com] has quit [Ping timeout: 264 seconds] 00:25:19 -!- rpg [~rpg@216.243.156.16.real-time.com] has quit [Quit: rpg] 00:26:31 -!- Bike [~Glossina@71-214-85-151.ptld.qwest.net] has quit [Ping timeout: 256 seconds] 00:28:04 Bike [~Glossina@67-5-220-211.ptld.qwest.net] has joined #sbcl 00:31:55 Bike_ [~Glossina@67-5-228-54.ptld.qwest.net] has joined #sbcl 00:34:38 -!- Bike [~Glossina@67-5-220-211.ptld.qwest.net] has quit [Ping timeout: 256 seconds] 00:37:49 Bike [~Glossina@67-5-233-99.ptld.qwest.net] has joined #sbcl 00:38:28 -!- Bike_ [~Glossina@67-5-228-54.ptld.qwest.net] has quit [Ping timeout: 252 seconds] 00:47:28 LiamH [~none@pool-173-48-164-128.bstnma.fios.verizon.net] has joined #sbcl 01:00:35 Quadrescence [~quad@unaffiliated/quadrescence] has joined #sbcl 01:13:52 -!- Hydan [~hydan@ip-89-103-110-5.net.upcbroadband.cz] has quit [Ping timeout: 276 seconds] 01:27:56 davazp` [~user@213.37.200.68.dyn.user.ono.com] has joined #sbcl 01:32:22 -!- davazp` [~user@213.37.200.68.dyn.user.ono.com] has quit [Remote host closed the connection] 01:43:46 -!- fisxoj [~fisxoj@c-24-12-190-29.hsd1.il.comcast.net] has quit [Ping timeout: 256 seconds] 02:31:31 echo-area [~user@182.92.247.2] has joined #sbcl 02:33:43 Bike_ [~Glossina@67-5-206-225.ptld.qwest.net] has joined #sbcl 02:34:18 -!- Bike [~Glossina@67-5-233-99.ptld.qwest.net] has quit [Disconnected by services] 02:34:24 -!- Bike_ is now known as Bike 02:37:52 -!- christoph4 [~christoph@ppp-188-174-89-217.dynamic.mnet-online.de] has quit [Read error: Operation timed out] 02:52:21 christoph4 [~christoph@ppp-188-174-67-28.dynamic.mnet-online.de] has joined #sbcl 03:05:01 chu [~user@unaffiliated/chu] has joined #sbcl 03:14:05 good news, the new cut-to-width fixes another one of emarsden's test cases. 03:14:25 I'll think about it some more and hopefully commit tomorrow. 03:47:19 ehaliewicz [~user@50-0-51-11.dsl.static.sonic.net] has joined #sbcl 03:48:45 -!- LiamH [~none@pool-173-48-164-128.bstnma.fios.verizon.net] has quit [Quit: Leaving.] 03:58:37 attila_lendvai [~attila_le@92.47.185.182] has joined #sbcl 03:58:37 -!- attila_lendvai [~attila_le@92.47.185.182] has quit [Changing host] 03:58:37 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 04:05:42 foreignFunction [~niksaak@ip-4761.sunline.net.ua] has joined #sbcl 04:23:50 robgssp [~user@cpe-24-93-28-218.rochester.res.rr.com] has joined #sbcl 04:31:20 Krystof [~user@81.174.155.115] has joined #sbcl 04:31:20 -!- ChanServ has set mode +o Krystof 06:13:58 -!- ASau [~user@p5797F121.dip0.t-ipconnect.de] has quit [Ping timeout: 252 seconds] 06:19:04 -!- chu [~user@unaffiliated/chu] has left #sbcl 06:20:59 benkard [~benkard@mnch-4d04dc04.pool.mediaWays.net] has joined #sbcl 06:28:06 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 06:47:47 -!- ehaliewicz [~user@50-0-51-11.dsl.static.sonic.net] has quit [Ping timeout: 256 seconds] 06:48:16 teggi [~teggi@123.20.106.53] has joined #sbcl 07:03:17 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 07:28:14 ASau [~user@p5797EA1A.dip0.t-ipconnect.de] has joined #sbcl 07:45:20 -!- benkard [~benkard@mnch-4d04dc04.pool.mediaWays.net] has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz] 07:58:49 rudi_ [~rudi@1x-193-157-241-188.uio.no] has joined #sbcl 08:00:21 -!- rudi_ is now known as rudi 08:02:41 benkard [~benkard@mnch-4d04dc04.pool.mediaWays.net] has joined #sbcl 08:16:57 -!- benkard [~benkard@mnch-4d04dc04.pool.mediaWays.net] has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz] 08:19:19 slyrus [~chatzilla@adsl-99-183-240-185.dsl.pltn13.sbcglobal.net] has joined #sbcl 08:42:38 Hydan [~hydan@ip-89-103-110-5.net.upcbroadband.cz] has joined #sbcl 08:45:52 chu [~user@unaffiliated/chu] has joined #sbcl 09:06:46 pranavrc [~pranavrc@122.164.173.103] has joined #sbcl 09:06:46 -!- pranavrc [~pranavrc@122.164.173.103] has quit [Changing host] 09:06:46 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 09:30:31 -!- scymtym_ [~user@ip-5-147-122-209.unitymediagroup.de] has quit [Ping timeout: 260 seconds] 09:56:05 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Quit: Ping timeout: ] 09:57:05 -!- Quadrescence [~quad@unaffiliated/quadrescence] has quit [Quit: This computer has gone to sleep] 10:05:41 -!- echo-area [~user@182.92.247.2] has quit [Remote host closed the connection] 10:37:16 -!- maxm [~user@unaffiliated/maxm] has quit [Remote host closed the connection] 10:38:46 maxm [~user@unaffiliated/maxm] has joined #sbcl 10:53:33 slyrus_ [~chatzilla@adsl-99-183-240-185.dsl.pltn13.sbcglobal.net] has joined #sbcl 10:54:18 -!- slyrus [~chatzilla@adsl-99-183-240-185.dsl.pltn13.sbcglobal.net] has quit [Ping timeout: 245 seconds] 10:54:28 -!- slyrus_ is now known as slyrus 10:58:49 segv- [~mb@95-91-241-60-dynip.superkabel.de] has joined #sbcl 11:45:17 -!- segv- [~mb@95-91-241-60-dynip.superkabel.de] has quit [Remote host closed the connection] 11:53:48 segv- [~mb@95-91-241-60-dynip.superkabel.de] has joined #sbcl 12:09:18 -!- chu [~user@unaffiliated/chu] has quit [Ping timeout: 245 seconds] 12:18:12 scymtym_ [~user@ubi-1-235-197.dhcp.uni-bielefeld.de] has joined #sbcl 12:20:09 stassats [~stassats@wikipedia/stassats] has joined #sbcl 12:21:47 can't institute my (mod x) changes, because during cold-init it apparently can't call numeric-type-low during !typecheckfuns 12:21:58 and that seems to only happen on x86-64, not x86 12:22:23 internal error #31 (Object is of the wrong type.) SC: 27, Offset: 14 $1= 0x1002ed8593: instance pointer 12:25:26 *stassats* is not sure how can that be and cold init is too cold to see anything 12:27:52 -!- drmeister [~drmeister@pool-71-185-168-200.phlapa.fios.verizon.net] has quit [Remote host closed the connection] 12:27:55 have you got transformations for (typep x '(mod y)) ? 12:27:58 or maybe it's usage of type-specifier 12:28:12 nope, only for type check 12:28:14 s 12:28:43 ok, I think you need them (so that !typecheckfuns-cold-init can generate type check functions without doing type specifier stuff) 12:29:30 i added a test for the vop into hairy-type-check-template-name 12:30:07 i can avoid calling specifier-type by just doing `(mod ,(1+ hi)) 12:30:24 typespec-specifier 12:30:30 or specifier-typespec 12:30:39 or however it's called 12:30:42 hm 12:31:09 I think I mean that !typecheckfuns-cold-init looks like (progn (setf (gethash '(mod 3) *table*) (lambda (x) (typep x '(mod 3)))) ...) 12:31:42 if the typep call isn't transformed by the compiler into something simpler, there will be calls involving type accessors way too early 12:31:53 why does it work on x86 then? 12:31:54 but maybe I'm misunderstanding your symptoms 12:32:10 i wish i understood them 12:32:27 can I see your patch? 12:32:31 i'll paste a diff when i'm out of idea 12:32:31 s 12:32:46 ok 12:34:06 and in my on live sbcl testing, (typep x '(mod 3))) isn't really transformed into check-... flavor vops 12:34:37 maybe cold-init does something different 12:35:01 it doesn't have to be -- it just has to be not a full call to anything 12:35:51 benkard [~benkard@dhcp-138-246-84-64.dynamic.eduroam.mwn.de] has joined #sbcl 12:36:30 ok, `(mod ,(1+ hi)) instead of type-specifier doesn't work either, i'm going to paste the thing 12:37:30 and why would it matter during cold-init? does cold init perform compilation? 12:37:42 or is it just something compiled in the xc being wrong? 12:37:55 hm. Reasonable questions 12:39:37 http://paste.lisp.org/display/137456 12:40:19 "(not sb-xc-host)" is just a left-over of trying everything 12:42:26 can you get any more information on what $1 and $2 are? 12:42:40 if i knew how 12:43:07 does print $1 work? 12:43:20 yes 12:44:30 http://paste.lisp.org/display/137456#1 12:45:02 so that's the hash table 12:45:34 what's $2 ? :-) 12:45:56 fdefn for call-with-recursive-system-lock 12:46:03 huh 12:46:14 that is more broken than I expected 12:48:04 and just tested, x86 compiled fine, passed the tests, and the vop is actually being applied 12:48:56 to user code, and to the internal code 12:49:47 maybe my power-of-two mask is wrong, i spent a lot of time on it 12:50:15 (logior (- (* (1+ hi) (ash 1 n-fixnum-tag-bits))) fixnum-tag-mask) 12:51:37 fixnum-tag-mask is #b00000....001 12:51:54 do you really mean to or those bits in? 12:53:01 basically, i need to call TEST on #b1111111000001 to see if either any of the high bits or the tag bits are set 12:53:29 and i use negative numbers for sign extension 12:55:40 maybe i should just take #xFFFFFF and punch wholes in it, because cooking up the correct two's complement is troublesome for me 12:55:52 and what does the constantize do? 12:56:39 puts into the unboxed constant section if it's larger than 32 bits 12:56:46 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 12:57:10 maybe it doesn't do a good job at sign-extension 12:57:19 and that's the differrence between x86 and x86-64 12:58:20 let me see if i can avoid calling this vop with >32 bits limits 13:01:28 that didn't help 13:01:40 ok, another difference between x86 and x86 is n-fixnum-tag-bits 13:03:20 (while I'm reading, isn't (* x (ash 1 y)) equal to (ash x y)? 13:03:31 i think i'm shifting too much 13:04:35 or maybe not 13:04:38 *stassats* sighs 13:05:53 no, the mask is correct, in fact, when i disable power of two optimizations, it still fails 13:06:10 so, the problem is in cmp then 13:08:08 and you're right about the ash equivalence 13:08:29 (i hope it doesn't fail because the code is not pretty) 13:10:35 huh, disabling the CMP part gives internal error #2 (Object is not of type LIST.) 13:10:38 SC: 27, Offset: 4 0x45900000010: even fixnum: 1195074650116 13:10:38 13:11:07 might that not be falling through to some other test? 13:11:07 and just in SB!KERNEL::!COLD-INIT this time 13:11:44 it should just generate (< hi x low) thing in this case 13:13:12 is ja signed or unsigned? 13:16:05 unsigned 13:16:51 i disabled that branch by not invoking the vop when it's not a power of two 13:17:12 let me try some other error code 13:21:47 *stassats* is convinced that all the rebuilding during this thing increased the outside temperature by at least one degree 13:22:08 what do you know, using object-not-fixnum-error works 13:22:59 now, trying r11 for the temporary 13:25:23 does it make sense to use result for the temporary? 13:25:53 hm, I suppose there's a chance that result is the same as value 13:26:16 moving to temp-reg-tn doesn't work either 13:29:27 nope 13:29:37 now it fails in random-layout-clos-hash 13:30:10 well, my special error code, object-not-mod-error worked alright before 13:31:32 http://xkcd.com/349/ 13:33:41 disassemble totally should be hyperlink for labels 13:34:06 hyperlinked 13:37:57 can i disassemble from ldb? 13:38:39 sort of 13:38:51 if you know where in memory you're looking, you can attach gdb and disassemble from there 13:38:59 maybe a more accurate answer would have been "no" 13:39:10 right, how con i know where the function in the frame lives? 13:39:27 there should be output/cold-init.map 13:39:38 that provides addresses for named functions 13:39:55 (there should totally be a built-in ldb disassembler) 13:47:02 drmeister [~drmeister@farnsworth.chem.temple.edu] has joined #sbcl 13:48:34 -!- benkard [~benkard@dhcp-138-246-84-64.dynamic.eduroam.mwn.de] has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz] 13:49:18 psilord [~pkeller@23-25-144-217-static.hfc.comcastbusiness.net] has joined #sbcl 13:51:13 nothing enlightening so far 13:51:40 i don't actually see the sign of my vop at that place 13:56:24 could it be a boxing/unboxing issue? 13:57:11 or rather, tagging untagging 13:57:20 as in, an already untagged fixnum is passed to the vop 13:58:00 I don't think so 14:00:19 basically, something like cmp rsi 4 jumps to the break, and kaboom 14:02:00 doesn't look at all like what the vop compiles to 14:02:10 -!- christoph4 [~christoph@ppp-188-174-67-28.dynamic.mnet-online.de] has quit [Ping timeout: 276 seconds] 14:06:06 stassats: wouldn't (rational 0 10) match your hairy-type-check-template? 14:06:34 it would... 14:06:46 well, let's try for the integer class 14:07:41 I don't know whether that matters. 14:08:16 it should, eventually 14:08:21 maybe not for this thing 14:10:05 at least it miscompiles it here (on a live patched sbcl), so, good catch either way 14:10:18 christoph4 [~christoph@ppp-188-174-6-30.dynamic.mnet-online.de] has joined #sbcl 14:11:56 and the cold init still dutifully fails 14:16:35 leuler [~user@p548FC0B4.dip0.t-ipconnect.de] has joined #sbcl 14:17:25 the code actually looks like an unoptimized MOD check 14:20:14 wbooze [~wbooze@xdsl-78-35-182-60.netcologne.de] has joined #sbcl 14:26:24 hlavaty`` [~user@friedrichstrasse.knowledgetools.de] has joined #sbcl 14:28:00 -!- hlavaty` [~user@friedrichstrasse.knowledgetools.de] has quit [Ping timeout: 245 seconds] 14:36:32 sb!xc:fixnump 14:36:41 in hairy-type-check-template-name 14:37:09 why? 14:37:35 i'm compiling with the same sbcl, how would that matter? 14:37:52 are you compiling x86 with x86-64 host? 14:37:59 nope, x86-64 with x86-64 14:38:12 shame. It's still wrong (you still need sb!xc:fixnump) 14:38:12 x86 compiled on x86-64 works perfectly 14:38:49 sb!xc:fixnump is undefined 14:39:33 or rather, the symbol is not found in the sb-xc package 14:39:35 guh 14:39:37 sorry 14:39:55 maybe fixnump is the right thing 14:40:00 and (typep x 'fixnum) is the wrong thing 14:40:16 (fixnump x) ? 14:40:48 stassats: yes, sorry 14:40:50 the error appears to be somewhere in gethash3, and there's a lot of mod checks, but none looks suspect 14:41:23 *stassats* will go for a jog before losing his mind completely on this 14:45:32 -!- rudi [~rudi@1x-193-157-241-188.uio.no] has quit [Quit: Client exciting.] 15:06:22 -!- slyrus [~chatzilla@adsl-99-183-240-185.dsl.pltn13.sbcglobal.net] has quit [Ping timeout: 252 seconds] 15:17:42 benkard [~benkard@ppp-188-174-95-19.dynamic.mnet-online.de] has joined #sbcl 15:27:57 ha, restricting the vop only for 0 to 1 works 15:30:02 *stassats* is doing a binary search for a good range 15:40:38 attila_lendvai [~attila_le@92.47.185.182] has joined #sbcl 15:40:38 -!- attila_lendvai [~attila_le@92.47.185.182] has quit [Changing host] 15:40:38 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 15:52:42 -!- pkhuong [~pkhuong@24.37.239.86] has quit [Remote host closed the connection] 15:59:53 fails when translating 0 to 4 16:07:41 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Read error: Connection reset by peer] 16:08:12 attila_lendvai [~attila_le@92.47.185.182] has joined #sbcl 16:08:12 -!- attila_lendvai [~attila_le@92.47.185.182] has quit [Changing host] 16:08:12 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 16:13:59 -!- Hydan [~hydan@ip-89-103-110-5.net.upcbroadband.cz] has quit [Remote host closed the connection] 16:14:18 Hydan [~hydan@ip-89-103-110-5.net.upcbroadband.cz] has joined #sbcl 16:38:40 arrdem [~arrdem@dhcp-53-132.ece.utexas.edu] has joined #sbcl 16:49:10 -!- wbooze [~wbooze@xdsl-78-35-182-60.netcologne.de] has quit [Quit: none] 16:59:02 -!- benkard [~benkard@ppp-188-174-95-19.dynamic.mnet-online.de] has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz] 17:15:16 found which part is that, psxhash, it has (integer 0 4) 17:21:56 -!- foreignFunction [~niksaak@ip-4761.sunline.net.ua] has quit [Quit: Leaving.] 17:26:51 prxq [~mommer@mnhm-590c2d40.pool.mediaWays.net] has joined #sbcl 17:28:04 wbooze [~wbooze@xdsl-78-35-182-60.netcologne.de] has joined #sbcl 17:39:25 pkhuong [~pkhuong@modemcable086.239-37-24.mc.videotron.ca] has joined #sbcl 17:39:35 Krystof: something like http://paste.lisp.org/display/137463 ? 17:52:32 benkard [~benkard@ppp-188-174-95-19.dynamic.mnet-online.de] has joined #sbcl 17:56:27 jsnell_ [~jsnell@ash.snellman.net] has joined #sbcl 17:59:32 -!- reb [user@nat/google/x-ugorupxmidxyggsr] has quit [Ping timeout: 264 seconds] 17:59:33 -!- jsnell [~jsnell@ash.snellman.net] has quit [Ping timeout: 264 seconds] 17:59:33 -!- Krystof [~user@81.174.155.115] has quit [Ping timeout: 264 seconds] 18:04:27 I was about to commit my only currently finished contribution (fix for lp#1085729) when during testing it surprisingly a test in dynamic-extent.impure.lisp failed under x86. 18:05:05 The failure is dependent on the core layout, for example compiling the same sources in a directory with a shorter name makes the test pass. 18:05:36 ASau` [~user@p5797EA1A.dip0.t-ipconnect.de] has joined #sbcl 18:06:09 Fortunately, I could find a reproducer that works for me under x86 and x86-64 with current head. See http://paste.lisp.org/display/137464. 18:06:28 If someone could confirm that this paste leads to a memory fault I would be grateful. 18:07:19 it's failing all the time 18:07:19 now 18:07:19 i mean, it's not your patch 18:07:25 yes 18:08:01 -!- arrdem [~arrdem@dhcp-53-132.ece.utexas.edu] has quit [Ping timeout: 252 seconds] 18:08:02 -!- prxq [~mommer@mnhm-590c2d40.pool.mediaWays.net] has quit [Ping timeout: 252 seconds] 18:08:02 -!- ASau [~user@p5797EA1A.dip0.t-ipconnect.de] has quit [Ping timeout: 252 seconds] 18:08:02 -!- jdz [~jdz@85.254.212.34] has quit [Ping timeout: 252 seconds] 18:08:16 jdz [~jdz@85.254.212.34] has joined #sbcl 18:09:36 prxq [~mommer@mnhm-590c2d40.pool.mediaWays.net] has joined #sbcl 18:09:42 drmeiste_ [~drmeister@farnsworth.chem.temple.edu] has joined #sbcl 18:10:54 -!- asedeno [asedeno@nat/google/x-ggovbyshjkqtlvkv] has quit [*.net *.split] 18:10:54 -!- automaciej [~automacie@quinoa.blizinski.pl] has quit [*.net *.split] 18:10:55 -!- angavrilov [~angavrilo@217.71.227.190] has quit [*.net *.split] 18:10:55 -!- easye [~user@213.33.70.157] has quit [*.net *.split] 18:11:33 -!- drmeister [~drmeister@farnsworth.chem.temple.edu] has quit [Ping timeout: 259 seconds] 18:12:16 angavrilov [~angavrilo@217.71.227.190] has joined #sbcl 18:13:01 leuler: I get some layout failure when printing 18:13:29 asedeno [asedeno@nat/google/x-paluhrgdisxdlmmq] has joined #sbcl 18:13:33 loading this doesn't do anything 18:14:01 easye` [~user@213.33.70.157] has joined #sbcl 18:14:04 arrdem [~arrdem@146.6.53.132] has joined #sbcl 18:14:25 automaciej [~automacie@2001:6f8:1c00:3ef::2] has joined #sbcl 18:14:27 because of the ignored values 18:14:36 The value 70368666908624 is not of type SB-KERNEL:LAYOUT. with (print (list (nested-bad 42) (make-nested-good :bar *bar*))) 18:15:15 I have # for (instance-layout (nested-bar 42)) 18:15:53 that's on x86-64 18:16:17 yeah. wait, why are we returning a DX value? 18:17:05 jdz_ [~jdz@85.254.212.34] has joined #sbcl 18:17:34 with debug 1, The value # is not of type SB-KERNEL:LAYOUT. 18:19:39 -!- brucem [~bmitchene@waywardmonkeys.com] has quit [Ping timeout: 256 seconds] 18:21:36 I have annotated the paste with my result. 18:21:48 -!- teggi [~teggi@123.20.106.53] has quit [Remote host closed the connection] 18:22:14 yeah. we're returning a stack value. 18:22:24 which, no surprise, results in badness when we try to print it. 18:22:34 My SBCL is compiled without --fancy in case it makes a difference. 18:22:38 -!- jdz [~jdz@85.254.212.34] has quit [*.net *.split] 18:23:03 So is the test from dynamic-extent.impure.lisp bogus? 18:23:19 I'm recompiling an x86 of 1.1.7 without fancy 18:23:43 It's :CONSERVATIVE-NESTED-DX, by the way. 18:24:10 that, or we're overly agressive on DX contagion. 18:27:54 arrdem_ [~arrdem@dhcp-53-132.ece.utexas.edu] has joined #sbcl 18:28:38 automaci1j [~automacie@cl-1008.ham-02.de.sixxs.net] has joined #sbcl 18:30:32 -!- pipping [~pipping@exherbo/developer/pipping] has quit [Ping timeout: 256 seconds] 18:31:05 pipping [~pipping@tchaikovsky.exherbo.org] has joined #sbcl 18:31:47 gdb is not very helpful with our calling conventions 18:32:26 and disassemble is not as helpful too 18:32:31 yeah. 1.1.7/x86 no fancy. -bad is returning a stack pointer. What's the test for? 18:32:49 ;; NESTED-BAD should not stack-allocate :BAR due to the SETF. 18:33:57 wtf. the first equalp test in :conservative-nested-dx is bogus 18:34:22 so nested-bar doens't get DXed because of (setf *bar*), but nested itself does. 18:34:22 and it's perfectly fine. 18:34:31 -!- arrdem_ [~arrdem@dhcp-53-132.ece.utexas.edu] has quit [Ping timeout: 271 seconds] 18:34:39 ldb prints symbols in the backtrace, can it print addresses instead? 18:34:46 *stassats* can't figure out where the error comes from 18:34:49 arrdem_ [~arrdem@dhcp-53-132.ece.utexas.edu] has joined #sbcl 18:35:31 gdb should do fine with our calling convention 18:35:50 ASau`` [~user@p5797EA1A.dip0.t-ipconnect.de] has joined #sbcl 18:35:50 -!- specbot [~specbot@common-lisp.net] has quit [Disconnected by services] 18:35:51 -!- minion [~minion@common-lisp.net] has quit [Disconnected by services] 18:35:54 specbot [~specbot@common-lisp.net] has joined #sbcl 18:35:55 minion [~minion@common-lisp.net] has joined #sbcl 18:36:29 the backtrace it shows is nowhere near what ldb shows 18:36:36 Hydan` [~hydan@ip-89-103-110-5.net.upcbroadband.cz] has joined #sbcl 18:37:31 -!- Hydan [~hydan@ip-89-103-110-5.net.upcbroadband.cz] has quit [Write error: Broken pipe] 18:37:40 -!- ASau` [~user@p5797EA1A.dip0.t-ipconnect.de] has quit [Ping timeout: 245 seconds] 18:37:49 brucem [~bmitchene@waywardmonkeys.com] has joined #sbcl 18:38:04 -!- |3b| [foobar@cpe-72-177-66-41.austin.res.rr.com] has quit [Ping timeout: 245 seconds] 18:38:19 -!- arrdem [~arrdem@146.6.53.132] has quit [Write error: Broken pipe] 18:38:19 -!- fe[nl]ix [~quassel@pdpc/supporter/professional/fenlix] has quit [Excess Flood] 18:38:20 -!- automaciej [~automacie@2001:6f8:1c00:3ef::2] has quit [Remote host closed the connection] 18:38:37 samebcha1e [~samuel@codesurfers.net] has joined #sbcl 18:38:49 -!- scymtym_ [~user@ubi-1-235-197.dhcp.uni-bielefeld.de] has quit [Ping timeout: 246 seconds] 18:39:33 leuler: so, I think nested-bad should end with (true x), like nested-good, and that the first assert in :conservative-nested-dx can just be a call to nested-bad. 18:39:38 fe[nl]ix [~quassel@pdpc/supporter/professional/fenlix] has joined #sbcl 18:40:36 -!- benkard [~benkard@ppp-188-174-95-19.dynamic.mnet-online.de] has quit [Ping timeout: 245 seconds] 18:40:37 -!- automaci1j [~automacie@cl-1008.ham-02.de.sixxs.net] has quit [Ping timeout: 245 seconds] 18:40:37 automaci1j [~automacie@cl-1008.ham-02.de.sixxs.net] has joined #sbcl 18:40:37 -!- Ralt [Ralt@2a01:7e00::f03c:91ff:feae:6c69] has quit [Ping timeout: 245 seconds] 18:41:19 Ralt [Ralt@2a01:7e00::f03c:91ff:feae:6c69] has joined #sbcl 18:42:41 ivan``_ [~ivan@unaffiliated/ivan/x-000001] has joined #sbcl 18:43:34 pkhuong: So; what to do? Shall I open a launchpad bug? I don't understand DX good enough to be of any help here. Or do you simply want to change the test without further efforts by whoever? 18:43:49 asedeno_ [asedeno@nat/google/x-ayodyzbmrwexcbpp] has joined #sbcl 18:44:23 -!- wbooze [~wbooze@xdsl-78-35-182-60.netcologne.de] has quit [Ping timeout: 278 seconds] 18:45:07 send an email on the ML with our findings so far? 18:45:34 I'm not too confident in my understanding of DX contagion either, but I'm pretty sure the first equalp test is bogus. 18:46:17 benkard [~benkard@ppp-188-174-95-19.dynamic.mnet-online.de] has joined #sbcl 18:48:07 -!- arrdem_ is now known as arrdem 18:48:52 -!- asedeno [asedeno@nat/google/x-paluhrgdisxdlmmq] has quit [*.net *.split] 18:48:53 -!- samebchase [~samuel@codesurfers.net] has quit [*.net *.split] 18:48:54 -!- ivan`` [~ivan@unaffiliated/ivan/x-000001] has quit [*.net *.split] 18:53:03 I have to leave now. I will send an email and probably open a launchpad bug tomorrow. 18:57:44 -!- leuler [~user@p548FC0B4.dip0.t-ipconnect.de] has quit [Quit: ERC Version 5.1.2 $Revision: 1.796.2.6 $ (IRC client for Emacs)] 18:59:31 -!- ASau`` is now known as ASau 19:02:41 automaciej [~automacie@quinoa.blizinski.pl] has joined #sbcl 19:03:41 asedeno [asedeno@nat/google/x-nwhzjbwiqlitwxzf] has joined #sbcl 19:03:42 cmm- [~cmm@bzq-79-177-123-220.red.bezeqint.net] has joined #sbcl 19:05:39 -!- automaci1j [~automacie@cl-1008.ham-02.de.sixxs.net] has quit [Ping timeout: 245 seconds] 19:05:40 -!- asedeno_ [asedeno@nat/google/x-ayodyzbmrwexcbpp] has quit [Ping timeout: 245 seconds] 19:05:40 -!- cmm [~cmm@bzq-79-177-123-220.red.bezeqint.net] has quit [Ping timeout: 245 seconds] 19:07:19 ha, looks like even just using emit-constant breaks cold-init 19:09:41 why would that happen during cold-init? 19:10:08 beats me 19:10:33 -!- segv- [~mb@95-91-241-60-dynip.superkabel.de] has quit [Ping timeout: 245 seconds] 19:10:39 maybe it just happens, but cold-init is just first when it can run 19:14:47 and only on x86-64 19:18:23 -!- samskull1 [~samskulls@S0106001111de1fc8.cg.shawcable.net] has quit [Ping timeout: 246 seconds] 19:19:06 -!- bege [~bege@S0106001d7e5132b0.ed.shawcable.net] has quit [Ping timeout: 246 seconds] 19:19:20 bege [~bege@S0106001d7e5132b0.ed.shawcable.net] has joined #sbcl 19:24:16 turns out *constants* isn't bound at that time 19:24:33 Quadrescence [~quad@unaffiliated/quadrescence] has joined #sbcl 19:24:49 wait, maybe i'm using the wrong symbol 19:25:09 doubtful. the first thing find-constant does is check for boundp ness 19:25:22 samskulls [~samskulls@S0106001111de1fc8.cg.shawcable.net] has joined #sbcl 19:27:36 it checks, but doesn't produce any errors 19:27:59 uh? 19:28:16 sdemarre [~serge@109.134.163.182] has joined #sbcl 19:28:56 segv- [~mb@95-91-241-60-dynip.superkabel.de] has joined #sbcl 19:34:37 -!- samebcha1e is now known as samebchase 19:37:34 it is unbound 19:38:01 when running in live sbcl, it's bound inside the vop 19:39:31 i guess object-not-mod-error it is 19:39:49 because i can't get object-not-type-error to work no matter what i try 19:50:19 scymtym_ [~user@ip-5-147-122-209.unitymediagroup.de] has joined #sbcl 19:51:38 does (logior (- (ash (1+ hi) n-fixnum-tag-bits)) fixnum-tag-mask) look right for testing (mod power-of-two) and fixnums at the same time? 19:51:44 i am not sure about anything anymore 19:52:11 (it doesn't cause the cold-init problem, just want to make sure it's correct) 19:52:37 HI in this case is (1- power-of-two) 19:52:39 hi is a power of two, or one less? 19:55:38 (lognot (fixnumize hi)) is more obviously right (: 19:55:49 but the logior looks ok. 19:58:19 indeed 19:58:22 thanks 20:01:53 I'm not sure that calling emit-constant after ltn is kosher. 20:02:16 well, cold-init is not sure either 20:02:31 i'm going the object-not-mod-error route anyway 20:07:26 what does :save-p :compute-only mean? 20:07:39 i know it's about lifetime, but what does that mean? 20:08:58 compute the save location for TNs, but don't force them there. Lets debugging code do its job, iirc. 20:10:31 I think an ir2-convert optimiser would be the easiest way to ensure you have a handle on a boxed constant. 20:11:18 some other time, i'm sick of this thing already 20:16:51 here's what i ended up with http://paste.lisp.org/display/137456#3 20:17:32 (typep x '(mod x)) needs to be optimized, and (<= 0 x N) as well 20:18:38 and implemented for other CPUs 20:18:49 -!- Ralt [Ralt@2a01:7e00::f03c:91ff:feae:6c69] has quit [*.net *.split] 20:18:50 -!- easye` [~user@213.33.70.157] has quit [*.net *.split] 20:18:50 -!- joshe [~joshe@opal.elsasser.org] has quit [*.net *.split] 20:18:51 -!- kanru [~kanru@118-163-10-190.HINET-IP.hinet.net] has quit [*.net *.split] 20:19:12 it doesn't really give big performance gains, but it's measurable and the code is smaller 20:21:38 I'd go with emit-move-arg-template and emit-constant. 20:22:28 i'd rather not go anywhere, i've wasted too much time on this 20:22:38 if you have the second argument in the immediate sc, you'll be able to do the loading yourself, only in the error path 20:23:28 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 20:24:12 maybe after i forget about my mental trauma, then i can try again 20:26:16 -!- drmeiste_ [~drmeister@farnsworth.chem.temple.edu] has quit [Remote host closed the connection] 20:26:24 easye` [~user@213.33.70.157] has joined #sbcl 20:28:43 Ralt [Ralt@2a01:7e00::f03c:91ff:feae:6c69] has joined #sbcl 20:28:43 joshe [~joshe@opal.elsasser.org] has joined #sbcl 20:29:34 and the dedicated error takes slightly less space 20:31:28 and of course, i should use fixnumize on the error code 20:31:29 -!- sdemarre [~serge@109.134.163.182] has quit [Ping timeout: 248 seconds] 20:35:05 drmeister [~drmeister@wirelessNAT188.wireless.temple.edu] has joined #sbcl 20:40:47 -!- drmeister [~drmeister@wirelessNAT188.wireless.temple.edu] has quit [Ping timeout: 256 seconds] 20:42:16 drmeister [~drmeister@166.137.105.207] has joined #sbcl 20:51:18 -!- Quadrescence [~quad@unaffiliated/quadrescence] has quit [Quit: This computer has gone to sleep] 20:59:56 scymtym: no more sb-bsd-sockets listening on a predefined port 21:00:06 scymtym / scymtym_ 21:02:14 -!- drmeister [~drmeister@166.137.105.207] has quit [Remote host closed the connection] 21:06:14 stassats: thanks, i can remove the ad-hoc from the builder server than 21:07:04 s/than/then/ 21:10:27 i will probably still keep test execution serialized for now since there were so many issues with parallel execution (and remain even if everything is serialized) 21:10:47 s/and remain/and some remain/ 21:12:06 one of the sporadic ones is threads.impure.lisp.lurking-threads on x86_64 21:12:09 something should be done with coalescing error breaks, i have screens of on heavy inlined code which uses car (car (car 21:12:27 each one has its own error code 21:12:38 drmeister [~drmeister@pool-71-185-168-200.phlapa.fios.verizon.net] has joined #sbcl 21:13:57 stassats: I think there's a mapping from error code to source location at high enough debug 21:14:02 so, unfortunately, we can't. 21:14:26 wanted to paste all the breaks, got "paste too large" 21:16:29 it unfolds list comparison for (equal list (list a b c d e f)) to not cons a list, so, can't get rid of them either 21:16:37 but i can reduce safety 21:20:02 (safety 0) shaves 3KB of code 21:20:11 from 18KB 21:20:19 (in a single function) 21:24:28 on our build server, on a single x86 virtual machine, the "fixup vector 0xcc3a990 has a bad widetag: 38" error occurred six times in a row for default features but not a single time for either of the three feature sets --fancy, -sb-unicode and -sb-thread 21:24:57 are the single dashes without? 21:25:06 yes 21:25:36 basically 21:25:38 case ${featureset} in 21:25:38 1) 21:25:38 ;; 21:25:41 2) 21:25:45 OPTIONS=--fancy 21:25:48 ;; 21:25:49 3) 21:25:52 cat <<"EOF" > customize-target-features.lisp 21:25:56 (lambda (features) 21:26:00 (setf features (delete :sb-unicode features)) 21:26:02 features) 21:26:05 EOF 21:26:08 ;; 21:26:11 4) 21:26:14 cat <<"EOF" > customize-target-features.lisp 21:26:17 (lambda (features) 21:26:21 (setf features (delete :sb-thread features)) 21:26:21 features) 21:26:26 EOF 21:26:30 ;; 21:26:31 esac 21:26:34 seriously, paste here? 21:26:38 sorry, that was a bit long 21:27:07 sorry 21:29:59 i should some day investigate better in-line caching, without putting too much inline 21:30:22 just the load-time-value vector + a function 21:30:34 that will solve the multitude of error codes blowing up code size 22:00:03 -!- prxq [~mommer@mnhm-590c2d40.pool.mediaWays.net] has quit [Remote host closed the connection] 22:01:03 -!- benkard [~benkard@ppp-188-174-95-19.dynamic.mnet-online.de] has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz] 22:15:58 -!- drmeister [~drmeister@pool-71-185-168-200.phlapa.fios.verizon.net] has quit [Read error: Connection reset by peer] 22:16:33 drmeister [~drmeister@pool-71-185-168-200.phlapa.fios.verizon.net] has joined #sbcl 22:28:53 -!- Hydan` [~hydan@ip-89-103-110-5.net.upcbroadband.cz] has quit [Ping timeout: 252 seconds] 22:53:36 what is %simple-fun-next used for? 23:01:44 when GCing code components, I think? Or for debug, definitely. 23:01:48 /introspection 23:13:42 Quadrescence [~quad@unaffiliated/quadrescence] has joined #sbcl 23:16:43 -!- segv- [~mb@95-91-241-60-dynip.superkabel.de] has quit [Remote host closed the connection] 23:22:13 -!- psilord [~pkeller@23-25-144-217-static.hfc.comcastbusiness.net] has quit [Quit: Leaving.] 23:22:49 psilord [~psilord@c-69-180-173-249.hsd1.mn.comcast.net] has joined #sbcl 23:35:06 -!- easye` [~user@213.33.70.157] has quit [Remote host closed the connection] 23:35:43 easye [~user@213.33.70.157] has joined #sbcl 23:51:59 pkhuong: why did you preference check-simd-pack vop with #+x86-64 inside x86-64/type-vops.lisp? 23:52:13 a leftover?