00:03:14 -!- ASau [~user@p54AFF2B1.dip0.t-ipconnect.de] has quit [Ping timeout: 264 seconds] 00:04:59 ASau [~user@p54AFE921.dip0.t-ipconnect.de] has joined #sbcl 00:09:40 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [Remote host closed the connection] 01:05:11 drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has joined #sbcl 01:20:37 I blame unicode 01:41:28 -!- Bicyclidine [45a62f69@gateway/web/freenode/ip.69.166.47.105] has quit [Ping timeout: 245 seconds] 01:47:37 fridim__ [~fridim@bas2-montreal07-2925317577.dsl.bell.ca] has joined #sbcl 02:16:05 ggkitsas [~ggkitsas@150.140.215.17] has joined #sbcl 02:20:59 edgar-rft [~GOD@HSI-KBW-109-193-013-113.hsi7.kabel-badenwuerttemberg.de] has joined #sbcl 02:22:26 -!- scymtym_ [~user@ip-78-94-192-97.unitymediagroup.de] has quit [Ping timeout: 264 seconds] 02:45:14 -!- prxq [~mommer@x2f65496.dyn.telefonica.de] has quit [Ping timeout: 264 seconds] 02:57:38 prxq [~mommer@x2f6d8b9.dyn.telefonica.de] has joined #sbcl 03:31:44 -!- fridim__ [~fridim@bas2-montreal07-2925317577.dsl.bell.ca] has quit [Ping timeout: 252 seconds] 03:38:53 -!- christoph_deb [~christoph@ppp-88-217-53-197.dynamic.mnet-online.de] has quit [Ping timeout: 252 seconds] 03:51:42 christoph_deb [~christoph@ppp-88-217-53-34.dynamic.mnet-online.de] has joined #sbcl 04:35:01 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 04:53:29 -!- foom [jknight@nat/google/x-upxrahxdhizqpdym] has quit [Ping timeout: 244 seconds] 04:59:24 foom [jknight@nat/google/x-youbdbbrrmpocscc] has joined #sbcl 05:33:17 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Ping timeout: 252 seconds] 05:54:57 andreh [~andreh@189.27.17.129.dynamic.adsl.gvt.net.br] has joined #sbcl 06:04:25 sdemarre [~serge@118.70-64-87.adsl-dyn.isp.belgacom.be] has joined #sbcl 06:05:50 angavrilov [~angavrilo@217.71.227.190] has joined #sbcl 06:54:27 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 07:05:07 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 07:37:25 -!- oleo [~oleo@xdsl-84-44-179-183.netcologne.de] has quit [Ping timeout: 264 seconds] 07:50:01 oleo [~oleo@xdsl-87-79-196-157.netcologne.de] has joined #sbcl 07:55:37 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Ping timeout: 240 seconds] 08:08:23 -!- andreh [~andreh@189.27.17.129.dynamic.adsl.gvt.net.br] has quit [Ping timeout: 252 seconds] 08:24:53 -!- ggkitsas [~ggkitsas@150.140.215.17] has quit [Ping timeout: 252 seconds] 08:42:06 yacks [~py@103.6.159.103] has joined #sbcl 10:13:37 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Ping timeout: 240 seconds] 10:51:00 -!- kludge` [~comet@unaffiliated/espiral] has quit [Ping timeout: 265 seconds] 10:56:41 kludge` [~comet@unaffiliated/espiral] has joined #sbcl 11:30:17 xani [~user@178.183.136.121.dsl.dynamic.t-mobile.pl] has joined #sbcl 13:07:35 andreh [~andreh@189.27.17.129.dynamic.adsl.gvt.net.br] has joined #sbcl 13:09:38 fiveop [~fiveop@p5DDC79F5.dip0.t-ipconnect.de] has joined #sbcl 13:15:13 -!- andreh [~andreh@189.27.17.129.dynamic.adsl.gvt.net.br] has quit [Ping timeout: 264 seconds] 13:53:01 -!- sdemarre [~serge@118.70-64-87.adsl-dyn.isp.belgacom.be] has quit [Ping timeout: 264 seconds] 13:54:38 Aliens of type DOUBLE-FLOAT cannot be passed as arguments to CALL-OUT. 13:55:04 Where did I uncomment somthing or forgot to add something, to get that error compiling sbcl? 13:58:14 I'd expect src/compiler/target/c-call.lisp 13:59:23 specifically, do you have a :arg-tn alien type method defined for double-float? 14:02:34 no 14:02:53 grepping for the error message looks like you're right 14:06:51 no-one's told me whether I managed to build a binary that works anywhere else yet 14:06:57 sheesh, does anyone read sbcl-announce? 14:07:37 also, is no-one excited that I did an sbcl release on a plane? no? 14:08:28 :) 14:08:28 *|3b|* tries to work python into a snakes on a plane joke but fails 14:08:55 a missed opportunity! 14:23:26 looks like I need to read up on the ARM C ABI 14:35:04 sdemarre [~serge@118.70-64-87.adsl-dyn.isp.belgacom.be] has joined #sbcl 14:39:49 -!- sdemarre [~serge@118.70-64-87.adsl-dyn.isp.belgacom.be] has quit [Ping timeout: 264 seconds] 14:42:32 gcc's assembly outpout confuses me 14:43:23 caller: flds s0, .L10 ; load argument in float register 14:43:27 matija [~matija@188-230-156-48.dynamic.t-2.net] has joined #sbcl 14:43:33 -!- matija [~matija@188-230-156-48.dynamic.t-2.net] has left #sbcl 14:43:59 callee: fsts s0, (stackaddress) \n ldr r3, (stackaddress)\n fmrs s0, r3 14:44:20 essentially, store float register argument on stack, load it into general purpose register and move it from gpr to float register 14:44:23 ... 14:49:10 fiveop: -O2? 14:49:23 nyef [~nyef@pool-70-109-132-189.cncdnh.east.myfairpoint.net] has joined #sbcl 14:49:32 G'morning all. 14:49:32 nyef, memo from pkhuong: does unconditionally setting optimize to NIL in pack-tn hide the issue? I'd frankly be OK with disabling a probably useless "optimisation" as a temporary band-aid. 14:50:46 I might have made a little more progress on finding the blown VOP yesterday if my ethernet interface hadn't packed it in, even through a full shutdown. /-: 14:51:28 Something that would be nice to add to the trace files is a full listing of the contents of the constant vector for a component. 14:56:24 leuler [~user@p548F9BED.dip0.t-ipconnect.de] has joined #sbcl 14:56:29 fridim__ [~fridim@bas2-montreal07-2925317577.dsl.bell.ca] has joined #sbcl 14:57:19 pkhuong: I want to understand the ABI, my functions would get optimized away 14:57:39 O2 will remove noise that O1 doesn't. 14:58:13 float s(float a) { return a; } is a function I would hope any optimization would get rid of 14:58:15 just use different compilation units if you want to see both sides. 14:58:35 fiveop: I don't see how. The function must exist to be called. 14:59:01 if I split it into different units you're right, but the output is fine 14:59:04 If you only care about one side (e.g., callers), write a prorotype. 14:59:26 ... You're at the point of floating-point FFI? 15:00:02 that's a fair point, actually 15:00:22 what call to a foreign function failed to compile, fiveop? 15:01:18 If this is the ARM port, I'd say /anything/ with a floating-point parameter. 15:03:23 sin 15:03:31 I'm trying to get code/irrat to work 15:03:40 hi nyef 15:04:47 The next two questions that come to mind are "have you gotten to a cold REPL yet?" and "do we actually need code/irrat to get to a cold REPL?" 15:05:22 Sounds like good progress, at least. 15:05:48 I got everything compiling except for the float things. Then I tried to replace your !cold-init with the proper one, which segfaulted 15:06:21 Hrm. 15:06:25 debugging the problem with Krystof help showed that a few function calles in it tries to make a hash-table, which has float slots for thresholds etc 15:06:30 +'s 15:06:45 So, I was forced to keep working on float. 15:07:10 I have registers and instructions setup (though I have no idea, whether any of that works :)) 15:07:34 We shouldn't be getting a pure segfault there, though. 15:07:56 the segfault happens in #'< 15:08:02 somewhere in make-hash-table 15:08:31 You're missing my point: The runtime should be handling segfaults. 15:08:40 and do what? 15:09:05 Hrm. We're on cheneygc, so that might just be a guard page violation or something. 15:09:15 also we might be too early to do anything useful 15:09:35 At the very least, shouldn't we be getting an LDB prompt or similar lose() report? 15:09:47 but I thought all you were going to try to do was float passing between lisp routines 15:10:07 and some float comparisons, I guess 15:10:22 make-target-2.sh: line 39: 357 Segmentation fault (core dumped) ./src/runtime/sbcl --core output/cold-sbcl.core --lose-on-corruption --no-sysinit --no-userinit < make-target-2.lisp 15:10:34 (I'm not totally surprised to discover that signal handling doesn't work as I expect: show me a platform where it *did* work as expected first time) 15:12:05 But whatever this is caused by, hash tables need floats and cold init needs hash tables fairly quickly 15:12:18 so I thought, going ahead and adding float support was a good next step 15:12:21 Right, but they shouldn't need irrat yet. 15:13:06 I wanted to just add everything float related 15:13:15 Oh! 15:13:28 We need to set a sigaltstack, 15:14:43 The register and instruction code I wrote, is the one I am most unsure about, whether it will work later on. So if you ever have some spare time for it, please look at https://github.com/fiveop/sbcl-arm/commit/cbe9407ddf3bb86867a5e323e6c3db73e1793732 :) 15:14:52 If we're hitting a stack guard page or something like that, then the machine is "in lisp", which means that the stack pointer is within the guard page, and the kernel doesn't know to switch the stacks back for C operation. 15:17:15 Something seems very wrong about your storage bases. 15:17:19 That's the ASM trace of the segfault: http://paste.lisp.org/display/141287 15:17:31 Don't the single and double registers overlap? 15:17:39 yes 15:18:17 physically 15:18:30 So they need to share an SB. 15:18:43 Otherwise PACK can't detect conflicts. 15:19:05 I'll see if I can polish up what I have so far this weekend. 15:19:37 Okay, that's neat. Why is r9 zero there? 15:20:17 That's OLD-FP-ARG... 15:21:35 we are in return-single there 15:22:44 I thought this was the single-value case of return-multiple? 15:23:36 oh yes 15:25:35 So, OLD-FP-ARG is clearly zero on entry to the VOP, which leads to the question of what's clobbering it and why. 15:27:32 (I'd recommend taking a trace file and checking the code for the affected function.) 15:27:53 And, speaking of trace files and functions, is there an easy way to tell, from a trace file, which function is involved in a CALL-NAMED? 15:29:37 I've narrowed my symptoms down to a single IR1 (and IR2) block, and the parameters passed to a function that that block calls, but there are two CALL-NAMED operations involved, and I want to know if my symptom is visible on the first or the second instance, and if it shows on the second instance what the first function might be. 15:29:51 (Well, what the first function IS, really.) 15:30:17 And I don't have the machine, a core file, or fasl files handy, just the trace files. 15:42:13 -!- fridim__ [~fridim@bas2-montreal07-2925317577.dsl.bell.ca] has quit [Ping timeout: 264 seconds] 15:44:58 -!- bege [~bege@S0106001d7e5132b0.ed.shawcable.net] has quit [Ping timeout: 240 seconds] 15:52:09 Okay, the /second/ CALL-NAMED takes a value which isn't bound until after the call with the broken parameters. 15:52:40 bege [~bege@S0106001d7e5132b0.ed.shawcable.net] has joined #sbcl 15:53:13 Which means that my shortlist is MOVE-TO-SINGLE, MOVE-TO-WORD/FIXNUM, %SINGLE-FLOAT/SIGNED, //SINGLE-FLOAT, and MOVE-FROM-SINGLE. 15:53:38 Or something got corrupted within the NUMBER-DISPATCH (which seems rather unlikely). 15:54:47 are you talking about your sparc problem? 15:55:07 Yeah. 15:55:43 Hrm. Or something is clobbering NUMBER before entry to the block, which could happen given that there's a COERCE not fully accounted for. 15:57:14 ... Okay, now treating the COERCE as fully accounted for. 15:59:48 And I see it, but I don't know what it is that I'm seeing. 16:02:12 http://paste.lisp.org/display/141431 16:02:20 Is there a way to set a break point in a lisp function in gdb? 16:02:32 Yes, as long as you haven't taken a GC yet. 16:02:42 break points at memory locations do not seem to work 16:02:54 They don't? WFM on SPARC. 16:02:59 break *0xwhatever 16:03:04 exactly 16:03:29 Now, the function addresses reported in the map file aren't directly usable for this. 16:03:39 Cannot access memory at address 0x512f81a5 16:03:43 oh right :P 16:04:09 you might need to add 27 or subtract 1 or something 16:04:15 There's a displacement applied in any call operation. Apply that displacement to the reported address and you should get a word-aligned address. 16:04:22 oh, and you need to add the break point after the core is loaded 16:04:32 so set a breakpoint in call_into_lisp 16:04:38 then when you hit that, set the breakpoint you actually want 16:04:42 More likely to be 23, and yes, you'll want to break at call_into_lisp before you set your core breakpoints. 16:04:57 it's been a while for me :-) 16:05:30 23 decimal or hex? 16:05:58 Decimal, but really, double-check what the displacement should be beforehand. 16:06:14 ... There should be a copy of the displacement in call_into_lisp anyway. 16:06:40 nyef: I'll take a look at the VOPs you singled out. 16:07:50 now I'm not sure we handle nfp packing correctly. 16:08:29 And because only the "minor" platforms use NFP in the first place, it wouldn't have been caught. 16:08:40 %SINGLE-FLOAT/SIGNED looks good to me. 16:10:02 but PPC would fail? 16:10:44 Not necessarily? 16:11:05 We'd need to get a trace file from PPC to see what it does here. 16:11:57 and re NFP, we force alignment in VOPs, right? 16:12:48 Not sure. I know that we force alignment of NFP on allocation, with a rather misleading comment on the alignment process, but... 16:13:37 Oh, and the SC definition probably has something to do with the alignment, with :size 2 or something. 16:14:16 no, it's :size 0 16:14:29 ... Wha? 16:14:36 initial nfp siez is 0 16:14:41 number stack 16:14:50 No, the SC not the SB. 16:15:08 oh yeah, but that's only relative to the stack frame. 16:15:10 Ah, :ELEMENT-SIZE 2 :ALIGNMENT 2. 16:15:57 If you look at the definition of BYTES-NEEDED-FOR-NON-DESCRIPTOR-STACK-FRAME in call.lisp, you'll find that it aligns, and that the comment says "must be multiples of 8 bytes on the PMAX". 16:16:12 For ANY platform, hence "misleading comment". (-: 16:18:22 This morning I am grateful for: Finally getting the ball on the SPARC build failure out of my court for a while. 16:19:30 PPC probably escapes because it uses a DOUBLE-STACK temporary there. 16:19:37 fot it. 16:19:40 got it, too. 16:19:56 sparc/float.lisp:867 16:20:27 I see it now. 16:20:28 we store x on the stack, at the offset for our single-reg temp 16:20:57 and we should just have a function/macro to decode these things. 16:21:28 It may be Wednesday before I can do a test build on this, and later than that if my hardware still refuses to cooperate. 16:21:54 And I agree, we should have something that takes a stack TN directly for these operations. 16:22:42 I'll check the whole file. 16:23:29 And, speaking of, fiveop: You might want to fix BYTES-FOR-NON-DESCRIPTOR-STACK-FRAME to round up to double-word alignment, and fix the comment so that it says something about preserving alignment for double-float values. 16:24:46 I also added logic to align SB sizes... that'd make more sense than fixing stuff in b-f-n-d-s-f 16:26:18 see meta-vmdef. I had to disentangle initial size, minimal increment, and alignment to make smaller stack frames work. 16:28:43 sparc/float.lisp: because using the named constants generated in objdef.lisp would be too easy. 16:29:58 ah, never mind. 16:31:04 An awful lot of stuff in the various backends possibly pre-dates some of the more convenient constants. Or was written by people who either didn't know or didn't care, and hasn't been well-maintained since. 16:32:23 in this case, I was thrown off by the "data" variant-var ;) 16:33:48 You may recall that I went through the x86oid backends at one point to gather up stack frame offset calculations into one place. 16:34:23 -!- Krystof [~user@81.174.155.115] has quit [Ping timeout: 252 seconds] 16:34:32 Maybe we need a rule: "If you're using TN-OFFSET directly within a VOP definition, you're doing something wrong." 16:35:51 Krystof [~user@81.174.155.115] has joined #sbcl 16:35:51 -!- ChanServ has set mode +o Krystof 16:37:29 (I wonder if we'd find more bugs by perturbing regalloc) 16:41:46 not even sure this builds on < sparcv9. According to our VOPs, we can expect hardware double float div, but not necessarily negation or abs. 16:43:53 for whatever reason I now get a segfault the first time cold-init tries to call debug_print 16:44:56 somehow the address stored in the fixup in the print vop has nothing to do with the address of call_into_c 16:45:16 maybe I should do a full build again :) 16:49:16 I think sparc wants a function to move 32 bit integer values between GPR/control stack and FPR/number stack. 16:54:14 erh, GPR/number stack. 16:56:14 I think that my take-away from this is that all of the backends could do with some going-over looking for repeated idioms to macroize and functionize away. 16:58:05 I made it to make-single-float. Pretty sure I haven't seen any offset confusion yet, apart from %single-float/signed. 17:19:00 Building SBCL on x86-64 with fewer than 63 fixnum bits has been broken since May 2013. It's easy to fix and I'd like to do that despite the feature being apparantly unused. Or do you folks insist that support for this feature should be dropped rather sooner than later? 17:21:23 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 17:22:23 Actually, I'd prefer for the feature to be maintained indefinitely, as the variation helps to enforce separation between a few different concepts. 17:24:21 Are we now out of code freeze again? 17:24:56 Apparently yes. There have already been at least five commits since. 17:25:50 After a full rebuild I still get the segfault, when trying to call call_into_c from !cold-init (via the print vop). 17:28:07 The print vop is trying to jump to the address 0xe002e59e, which it reads from a fixup. The thing is that call_into_c is located at 0x20640 according to GDB. The fixup for debug_print also contains an address which is different from the one reported by gdb, though, I haven't check what call_into_c does with the value. 17:28:30 Could someone tell me where these fixup constants are actually calculated an emitted? 17:29:12 The only thing I can find in compilers/arm are two functions, emit-absolute-fixup and emit-relative-fixup, that are in a comment block. 17:29:22 so they are not even available 17:29:40 Have a look in SYS:SRC;COMPILER;ARM;PRINT.LISP.LATEST. 17:30:21 I don't have that file 17:30:28 ... Yes you do. 17:30:43 Sorry, you're right. SHOW.LISP.LATEST. 17:30:48 My bad. 17:31:01 I see that 17:31:07 inst word (make-fixup "call_into_c" :foreign)) 17:31:11 Yup. 17:31:39 Now, SYS:SRC;COMPILER;FIXUP.LISP.LATEST. 17:32:39 Then look at DUMP-FIXUPS in compiler/dump, and anything to do with fixups in compiler/generic/genesis. 17:33:13 where do you copy those file paths from? 17:33:33 What do you mean? 17:33:55 Not used to LPN syntax? 17:38:10 I don't know what LPN means 17:38:25 Logical PathName. 17:38:58 Does the file processed by load-cold-foreign-symbol-table survive the compile process, and if so what's its name? 17:39:09 don't answer 17:39:10 ... 17:39:38 Do you mean... "nm"? (-: 17:40:07 but I have a proper question 17:40:24 should the fixups for such foreign symbols contain the values indicated in that file? 17:40:37 if so, something goes wrong 17:40:48 Yes, unless you have enabled linkage-table. 17:41:02 If you have enabled linkage-table then there's more work to do. 17:41:35 There might be a declared "linkage table space", so you can check to see if the fixups are pointing to within that space. 17:42:11 If I have to enable that using a --with argument to config, then I don't. 17:43:17 Have a look in src/runtime/genesis/config.h to see if anything leaps out at you with respect to LINKAGE_TABLE. 17:44:54 nope 17:46:47 Okay, check the actual fixup bits in memory to see if they're correct. If they are, then the issue is with loading the fixup value at runtime. If they are not, then the issue is with patching the fixup value into the core during genesis. 17:47:45 When you say 'in memory' at runtime, you mean when I have the core loaded in gdb? 17:48:24 Yeah. 17:48:48 Or attack the core file with a hexdumper if you know how to navigate that way. 17:51:26 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Read error: Operation timed out] 17:53:28 This is what I was talking about: http://paste.lisp.org/display/141432 17:55:40 the examined value for debug_print is wrong 17:55:43 the line needs to read 17:56:07 Why is it adding 2 to PC? 17:56:44 PC should be word-aligned, as should the location of the fixup. 17:58:33 Recheck your definition of load-from-label and its dependencies. 17:58:53 (Or whatever it got replaced with, it's been a while since I looked at your tree.) 17:59:29 Its still your definition of load-from-label 18:02:14 Which I now try to understand. You added that on Janurary 8th 18:04:23 Change the use of LDB to MASK-FIELD. 18:08:30 both? 18:08:52 I'm away for a bit and will then continue. 18:09:08 The one to LOAD-CHUNK, mainly, though both would work. 18:12:39 The problem probably didn't occur before, because I commented out most of !cold-init to see which parts failed. 18:12:44 The offsets wer smaller. 18:12:44 :) 18:12:49 +e 18:12:52 now, I'm away 18:23:20 fridim__ [~fridim@bas2-montreal07-2925317577.dsl.bell.ca] has joined #sbcl 18:39:17 That fixed it, as it should :) 18:39:43 back to the problem in make-hash-table ;) 18:42:39 Uh-oh. 18:43:12 Fix up COMPUTE-LRA as well, please. 18:46:48 k 18:53:48 pnpuff [~InGaAsP@unaffiliated/pnpuff] has joined #sbcl 18:55:02 -!- pnpuff [~InGaAsP@unaffiliated/pnpuff] has left #sbcl 19:04:18 I have a different segfault this time (again) but in <. 19:04:41 This time the output from gdb actually makes sense, because the segfault happens when a str instruction is exectued. 19:04:49 we are in copy-more-args 19:05:00 in DO-REGS 19:40:05 Oh! Have you been doing "handle SIGSEGV pass nostop print"? 19:43:35 -!- oleo [~oleo@xdsl-87-79-196-157.netcologne.de] has quit [Ping timeout: 252 seconds] 19:44:40 no, but the segvs have been coming way before they should and also when run at the command-line 19:44:58 fiveop: that sounds good! 19:45:09 why is that good? 19:45:17 because it's different! 19:45:20 :) 19:45:33 -!- angavrilov [~angavrilo@217.71.227.190] has quit [Remote host closed the connection] 19:45:44 nyef: What are you talking about? 19:46:17 oleo [~oleo@xdsl-78-35-141-42.netcologne.de] has joined #sbcl 19:47:16 at a later point in the boot, once gc is enabled, if you're still getting segvs you'll need to tell gdb to pass them through to the underlying lisp 19:47:40 that's what he's saying? 19:48:47 Also applies to stack guard pages. 19:49:08 Is that another signal? 19:49:43 Still SIGSEGV, I believe. 19:50:25 And the guard pages would trigger much earlier, as they are used to control the amount of stack space that gets scrubbed during GC. 19:50:46 well currently I get a segsegv, when it tries to store a value from stack on copy-more-args in a call to < 19:51:14 What's the frame point there? 19:51:16 I tried to make a small sample in arm-boostrap.lisp with (defun var-arg-fun (a &rest more) ...) 19:51:19 but that didn't trigger it 19:51:39 let me restore that situation again 19:53:54 CFP: 0xb6bd4fe0; CSP: 0xb6bd5004; Wert der gelesen werden soll: CFP+32 (0xb6bd5000), wert 0 19:54:06 sorry 19:54:16 address of the value that should be read ;) 19:54:29 (my mind jumped to german and I didn't notice) 19:55:15 Have a look at your memory map in /proc/pid/maps: I bet you there's memory there, but it's mapped either read-only or no-access. 19:57:41 b6bd5000-b6bd6000 r-xp 00000000 00:00 0 [stack:1086] 19:57:44 there we go 19:57:49 b69d7000-b6bd5000 rwxp 00000000 00:00 0 19:57:51 That'd be a guard page. 19:58:20 Stack grows upwards, so you'll also see an rwxp page leading up to b6d5000. 19:58:57 Umm... So at this point we really DO need an altstack, or to disable the whole guard-page noise temporarily.. 19:59:16 What is an alt-stack? 19:59:26 man 2 sigaltstack 19:59:26 Sorry, I couldn't find anything for 2 sigaltstack. 20:00:15 man sigaltstack 20:00:15 http://developer.apple.com/documentation/Darwin/Reference/ManPages/man2/sigaltstack.2.html 20:00:20 Something like that. 20:01:15 found it 20:01:17 Crib the logic in x86-linux-os.c, arch_os_thread_init(), the bit that's conditional on LISP_FEATURE_C_STACK_IS_CONTROL_STACK, but don't crib the conditional, because it's always false on ARM. 20:01:48 LiamH [~none@96.231.217.60] has joined #sbcl 20:01:49 I'm probably going to have to go over all of that logic AGAIN before we're done, but for now it should be workable. 20:07:57 I don't understand what this helps us. So, single handlers will use a different stack. Does it allow the runtime to handle the segfault itself? 20:08:16 That's the idea, yes. 20:08:43 do we have a handler for that? 20:09:02 In normal operation, SIGSEGV is used for dealing with stack guard pages, various aspects of the GC, and so on. 20:09:07 ' 20:09:15 The handler should already be installed. 20:10:07 But the hardware stack pointer is pointing to the Lisp stack, which is only supposed to be used for boxed data, and which might be pointing to a read-only guard page, either of which situation will lead to badness. 20:10:19 Hence the use of the altstack. 20:16:37 that didn't change anything 20:16:54 (assuming running target-1 and host-2 after the change the runtime suffices) 20:17:22 oh wait 20:18:00 no it is as I said 20:20:11 and it might change something, for some reason host-2 failed and I didn't notice, since I send the output to a file 20:20:45 Hrm. 20:21:20 but wait 20:21:27 And between my Pi being nowhere near where I am and my network port acting up, I can't really mess with this directly myself. /-: 20:21:33 we shouldn't be anywhere near stack exhaustion 20:21:55 We're not near EXHAUSTION, we're triggering the guard page used as a high water mark when scrubbing the stack. 20:22:00 oh 20:22:58 -!- segv- [~mb@95-91-242-78-dynip.superkabel.de] has quit [Ping timeout: 244 seconds] 20:23:18 I'm not sure how to turn it off (or even if we CAN turn it off). 20:23:48 what do you mean, when scrubbing the stack? 20:24:05 but wait 20:24:22 the other cheneygc ports don't have altstacks 20:24:46 what makes this port different? 20:24:53 The other cheneygc ports have enough registers for both Lisp and C stack pointers. 20:25:00 ah 20:25:17 ok, fair point 20:25:24 *Krystof* goes back to writing 20:26:08 nyef: Am I supposed to run it with "handle SIGSEGV pass nostop print" after the change? 20:26:09 fiveop: During GC, we scavenge the data on the control stack up to the "high water mark" of the current stack pointer. After that we "scrub" the stack (overwrite it with zeros) to the high water mark of the soft guard page. 20:26:42 but the segfault is trigged in < 20:26:49 Yes, or pass stop print, but to issue cont commands when it stops (after taking a look at the situation, in case we run into a non-guard-page related fault). 20:27:13 Yes, it's triggered in <, because the soft guard page needs to be moved to set a new high-water mark. 20:27:25 ah 20:27:29 now I see 20:27:49 maybe that worked before, I never used cont before and now it jumps to a different address 20:27:59 We should probably start writing documentation on this at some point. 20:28:13 I don't believe that SIGSEGV defaults to pass on gdb, but I could be wrong. 20:29:26 no it does not 20:29:42 oh it does 20:33:47 So what is the following: 20:34:11 -!- yacks [~py@103.6.159.103] has quit [Ping timeout: 252 seconds] 20:35:55 When I get the sigsegv in <, I cont. I get another sigsegv in 0xa378, a function unknown to gdb between _init and calloc: it tries to push something to stack guard page. 20:36:24 I cont, which lands me with another sigsegv in 0xa378 again 20:36:32 cont again, program terminate 20:36:32 s 20:37:53 Hrm. Something still isn't right. 20:38:16 Do you have the libc debug symbols installed? 20:39:39 I'd say yes: http://paste.lisp.org/display/141434 20:41:22 you could live dangerously, and break in the sigsegv_handler 20:41:27 I have a back trace 20:42:05 #0 0x0000a378 in ?? () 20:42:05 #1 0x00013fc8 in lower_thread_control_stack_guard_page (th=0xb6dd7000) at interrupt.c:1513 20:43:07 Hrm. 20:43:53 Get back to that first SIGSEGV, the one in #'< check the stack and frame pointers, stepi, and check them and the program counter again, make sure that the altstack is actually coming into play. 20:44:38 -!- slyrus [~chatzilla@107.201.5.56] has quit [Ping timeout: 240 seconds] 20:46:16 their values change, whether they are correct I do not know 20:46:42 Okay, that's a start. 20:49:20 http://paste.lisp.org/display/141435 20:49:41 here, you also have /proc/pid/maps before and after 20:50:11 Hrm. That's not right... 20:51:55 Probably doesn't help that we operate with a mixed stack discipline, either. C stacks go one way, the Lisp control stack goes the other way. 20:56:30 That... is odd. 20:56:58 Is it setting the new guard page as though it's an expand-down stack? 20:57:38 You can't ask me 20:57:44 Or is that a low-water mark guard page? 20:57:59 What are the program counter values before and after? 20:58:57 pc 0x4f41ca54 0x4f41ca54 20:59:02 pc 0xa378 0xa378 20:59:08 but what's more interesting is probably 20:59:22 #1 0x00013fc8 in lower_thread_control_stack_guard_page (th=0xb6dd7000) at interrupt.c:1513 20:59:25 #2 0x000140d0 in lower_control_stack_guard_page () at interrupt.c:1536 20:59:28 #3 0x000141bc in handle_guard_page_triggered (context=0xb6bd4d10, addr=0xb6bd5000 "") at interrupt.c:1556 20:59:31 after 21:00:29 Joy. The altstack isn't getting applied. 21:00:47 Or it just plain Doesn't Work, which has been known to happen with some systems. 21:03:01 or it doesn't work under gdb 21:03:10 that's been known too 21:03:16 That's another possibility, yes. 21:03:48 I have been known to write my own debugging tools for this kind of damage. /-: 21:05:04 I still get a segfault without gdb, though ofc I don't know where 21:09:31 Okay, how about this: In thread.c, initial_thread_trampoline(), there are a bunch of calls to protect_*_stack{,_hard}_guard_page(). Disable them for now. 21:10:09 I'm not sure if that will help or not, TBH. 21:11:00 where are the implementations of these functions (it is compiling) 21:11:30 Probably in validate.c. 21:11:55 oh right, its a macro 21:15:56 Why do lots of source files contain 0x0c characters (new page/form feed)? 21:17:07 For a couple of reasons, one of which is easier navigation in emacs. 21:17:52 Probably easiest to think of it as some stylistic choice from back in the day, and neither remove nor add them. (-: 21:22:11 it matters when you print them out 21:22:23 as you might well have to when the time comes to merge this port ) 21:24:21 CORRUPTION WARNING in SBCL pid 2370: 21:24:21 Memory fault at b6e18014 (pc=0x4f41ca54, sp=(nil)) 21:24:21 The integrity of this image is possibly compromised. 21:24:24 that's new :) 21:25:09 This is the result of removing all the guard page calls from initial_thread_trampoline. 21:28:45 nyef: I currently don't think that I'll get a lot further with this problem unless I first understand fully, how our memory - and in particular stack - managament works 21:29:16 Yeah, you might not. There's a lot going on, and I twisted certain bits quite a bit when setting up the ARM port as well. 21:30:29 I will commit the changes, that we know where improvements or fixes (compute-lra and load-from-label) 21:31:44 I would prefer continueing with the float things. There I have a better idea what I need to do, even though I have to read a lot :) 21:32:42 Which brings me to another question: Do you have a document stating how single and double float values are passed according to the C ABI. I only found a document that said something general about coprocessor registers. 21:33:26 I used gcc output to figure out the convention, though an official documentation is always better :) 21:34:21 There are at least two conventions, unfortunately. 21:35:57 The document that might be interesting would be the AAPCS, though I don't know if that one covers any specifics for the floating-point. 21:36:00 depending on whether the chip does floats in hardware or software? 21:36:49 -!- xani [~user@178.183.136.121.dsl.dynamic.t-mobile.pl] has quit [Ping timeout: 264 seconds] 21:36:59 Almost. There's a variant where there's a hardware FPU but the floats are passed in the integer registers and on the stack. 21:38:12 On my Pi, gcc passed them in float registers 21:38:30 the fun thing is 21:38:32 So you have a hard-float ABI. 21:38:48 What about when you have more parameters than registers? 21:39:02 In void(float,float,double) and void(float,double,float) the two floats will be passed in the same registers (s0,s1) 21:39:12 it goes to the stack 21:39:22 And mixed with integer parameters? 21:39:35 It will use registers 21:39:41 as far as possible 21:39:43 and then the stack 21:39:57 I actually didn't test what happens when I have float and non-float stack registers 21:40:16 But the only thing that makes sense is to put them on the stack in order 21:40:20 (to me at least :)) 21:40:47 Another thing I didn't test, is whether doubles have to be double word align on the stack. 21:41:32 Now, see, this is one of the reasons why I didn't do this part of c-call. (-: 21:42:07 you mean define-alien-type-method for double and single? 21:43:02 If I understand it correctly, arg-state is a arch dependend structure, in which you (as the arch implementor) can keep track what you have done so far with the arguments for a single alien call. 21:43:06 I hope you get them in order :) 21:44:21 Oh, don't think that I won't leave it unimplemented if I think I can get away with it. (-: 21:44:32 what do you mean? 21:45:06 Don't we need to support passing float args to alien functions? 21:45:23 Sure, but we won't need to support stack floats for quite a while. 21:47:06 probably not 21:48:54 So I'd just arrange for the register values to work, and signal an error if it goes as far as stack values. 21:50:09 The different arg passing conventions would need to be a build option. 21:50:17 Yup. Already planned. 21:51:08 Also don't forget that there are at least three or for variations on the VFP unit, plus an MMX FPU on some systems, plus soft-float-only systems, and that's just off the top of my head. 21:53:04 It would be nice, if you could upload the work you have done so far in vm.lisp and insts.lisp. 21:53:59 I would continue with adding register passing to c-call and see how I progress in getting things to compile (not necessarily run :)) 21:55:44 I haven't done anything in insts.lisp, actually. 21:56:25 Oh, there's also the FPA unit, which was another FPU option. 21:56:50 I would start with the hardware I have at hand, and then go from there. 21:57:12 As would I, but I'd make sure to distinguish which bits need to be altered for other hardware. 22:01:00 And right-right now, I don't have any hardware set up. 22:03:11 my gcc has 17 options for -mfpu most (if not all) of which seem to be for ARM :) 22:04:17 *nyef* sighs. 22:04:38 Once again, I find myself not knowing where the sdcard with the debian chroot for my android tablet has gotten to. 22:05:12 Of course, said chroot image might be corrupt, as well. That's been known to happen. /-: 22:05:44 The only other piece of hardware I have here is a beaglebone, which is still taped into its ESD bag. 22:08:13 What does single-float-hidden-bit represent? It seems to correspond to the most significant fraction bit. 22:08:22 But why hidden? 22:09:09 the least significant of the exponent actually (*-float-trapping-nan-bit is the most significant fraction bit) 22:09:27 Hang on, that bit I /did/ manage to add to my local tree. 22:10:08 I added it as well, since it was the same for every arch I looked at :) 22:10:15 I just don't understand what those two mean. 22:10:36 The rest is relatively self explanatory 22:10:43 http://paste.lisp.org/display/141439 22:11:19 IEEE float presumes that the mantissa of a normalized float has a leading 1 bit that isn't physically present. 22:16:18 And the most significant bit of the mantissa distinguishes between quiet and signalling NaNs (see http://de.wikipedia.org/wiki/IEEE_754#Keine_Zahl_.28NaN.29 or http://en.wikipedia.org/wiki/NaN). 22:16:31 nyef: I have the same constants except for the rounding modes 22:16:51 I have nearest (0), positive (1), negative (2), and zero (3) 22:16:56 Hrm. 22:17:02 I may or may not have got those right. 22:17:19 page 886 of the 2005 ARM ARM 22:17:20 One of the points, however, is the comment. 22:18:50 but the difference here would be a weird difference in interpretation 22:19:47 good night 22:20:12 Sleep well. 22:20:16 leuler: thank you, its also mentioned in the ARM ARM 22:20:27 and thanks everyone for the help :) 22:20:33 -!- fiveop [~fiveop@p5DDC79F5.dip0.t-ipconnect.de] has quit [] 23:13:44 FareWell [fare@nat/google/x-ywmfvhwcaqdthdzv] has joined #sbcl 23:14:15 Ouch. Is there a way to invalidate the subclasses of a class you're redefining, to avoid a circularity error while redefining them? 23:16:11 sdemarre [~serge@118.70-64-87.adsl-dyn.isp.belgacom.be] has joined #sbcl 23:16:37 -!- ASau [~user@p54AFE921.dip0.t-ipconnect.de] has quit [Remote host closed the connection] 23:33:09 *FareWell* reaches for the hammer, aka unintern 23:39:50 -!- FareWell [fare@nat/google/x-ywmfvhwcaqdthdzv] has left #sbcl 23:46:39 ASau [~user@p54AFE921.dip0.t-ipconnect.de] has joined #sbcl 23:51:17 -!- sdemarre [~serge@118.70-64-87.adsl-dyn.isp.belgacom.be] has quit [Read error: Operation timed out] 00:02:01 -!- ASau [~user@p54AFE921.dip0.t-ipconnect.de] has quit [Ping timeout: 240 seconds] 00:03:02 ASau` [~user@p5083DE93.dip0.t-ipconnect.de] has joined #sbcl 00:03:12 -!- leuler [~user@p548F9BED.dip0.t-ipconnect.de] has quit [Quit: ERC Version 5.1.2 $Revision: 1.796.2.6 $ (IRC client for Emacs)] 01:08:44 slyrus [~chatzilla@107.201.5.56] has joined #sbcl 01:31:12 -!- ASau` is now known as ASau 02:35:21 attila_lendvai [~attila_le@5.251.216.208] has joined #sbcl 02:35:21 -!- attila_lendvai [~attila_le@5.251.216.208] has quit [Changing host] 02:35:21 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 02:45:49 -!- prxq [~mommer@x2f6d8b9.dyn.telefonica.de] has quit [Ping timeout: 264 seconds] 02:47:01 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Ping timeout: 264 seconds] 02:58:11 prxq [~mommer@x2f64950.dyn.telefonica.de] has joined #sbcl 03:03:59 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 03:19:24 izirku [~izirku@pool-173-71-18-32.dllstx.fios.verizon.net] has joined #sbcl 03:38:49 -!- christoph_deb [~christoph@ppp-88-217-53-34.dynamic.mnet-online.de] has quit [Ping timeout: 240 seconds] 03:48:51 Umm... Is anyone else having trouble building SBCL HEAD on x86-64/linux? 03:50:19 *nyef* tries building the latest release as a start to a bisection search. 03:53:01 christoph_deb [~christoph@ppp-88-217-58-110.dynamic.mnet-online.de] has joined #sbcl 04:08:01 commit b4b05aac3ab3e74fb26eb6ebb5e0b44f2841eaff breaks the build for me on x86-64/linux. 04:08:53 Fails to build the new definition of (DEFSTRUCT ARG). 04:33:28 yauz [4fca40aabe@osgiliath.yauz.de] has joined #sbcl 04:40:34 -!- ASau [~user@p5083DE93.dip0.t-ipconnect.de] has quit [Remote host closed the connection] 04:42:33 -!- izirku [~izirku@pool-173-71-18-32.dllstx.fios.verizon.net] has quit [] 04:42:57 nyef: builds for me 04:44:15 I get a warning about (SB-PCL::SLOT-ACCESSOR :GLOBAL POSITION SB-PCL::READER) being undefined and its name being reserved and thus unportable. 04:47:03 what's the build host? 04:48:24 ASau` [~user@p5083DE93.dip0.t-ipconnect.de] has joined #sbcl 04:48:39 1.0.23. 04:50:19 sounds old. 04:54:24 -!- ASau` [~user@p5083DE93.dip0.t-ipconnect.de] has quit [Remote host closed the connection] 04:56:50 Doesn't mean it shouldn't build. 04:58:11 ASau` [~user@p5083DE93.dip0.t-ipconnect.de] has joined #sbcl 05:01:07 -!- ASau` [~user@p5083DE93.dip0.t-ipconnect.de] has quit [Remote host closed the connection] 05:04:30 ASau` [~user@p5083DE93.dip0.t-ipconnect.de] has joined #sbcl 05:15:34 unless a bug was fixed along the way 05:21:48 Mmm. It's after midnight, I'll worry about it in the morning. 05:24:00 In fact, I'm going to sign off now. Have a good rest-of-your-day. 05:24:03 -!- nyef [~nyef@pool-70-109-132-189.cncdnh.east.myfairpoint.net] has quit [Quit: G'night all.] 05:45:38 angavrilov [~angavrilo@217.71.227.190] has joined #sbcl 05:55:44 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Ping timeout: 253 seconds] 06:29:07 -!- LiamH [~none@96.231.217.60] has quit [Quit: Leaving.] 06:41:18 sdemarre [~serge@112.84-64-87.adsl-dyn.isp.belgacom.be] has joined #sbcl 07:01:30 ggkitsas [~ggkitsas@150.140.215.17] has joined #sbcl 07:44:02 -!- ams [ams@Psilocybe.Update.UU.SE] has quit [Ping timeout: 264 seconds] 08:19:25 -!- oleo [~oleo@xdsl-78-35-141-42.netcologne.de] has quit [Ping timeout: 264 seconds] 08:23:37 -!- ggkitsas [~ggkitsas@150.140.215.17] has quit [Ping timeout: 264 seconds] 08:32:10 oleo [~oleo@xdsl-87-78-76-233.netcologne.de] has joined #sbcl 09:05:06 ams [ams@Psilocybe.Update.UU.SE] has joined #sbcl 10:00:37 -!- kanru [~kanru@118-163-10-190.HINET-IP.hinet.net] has quit [Ping timeout: 265 seconds] 10:26:36 tmh_ [~e@unaffiliated/tmh/x-0013998] has joined #sbcl 10:31:17 -!- kludge` [~comet@unaffiliated/espiral] has quit [Ping timeout: 252 seconds] 10:36:37 kludge` [~comet@unaffiliated/espiral] has joined #sbcl 11:11:41 Poenikatu [~kvirc@pdpc/supporter/bronze/poenikatu] has joined #sbcl 11:12:35 I've downloaded sbcl 1.1.16 and unpacked the tarball. The INSTALL file says issue the command "sh make.sh" to build SBCL. Can't: the file aint there 11:13:02 Could somebody provide said file? 11:13:30 Would it be the same as the make.sh file from sbcl 1.1.15? 11:15:23 <|3b|> looks like it is there 11:15:47 |3b|: In the tarball? 11:15:50 <|3b|> yeah 11:16:04 <|3b|> -rwxr-xr-x csr21/1000 4309 2014-02-28 09:53 sbcl-1.1.16/make.sh 11:18:02 |3b|: Piss. Just noticed that I'd downloaded the binary pkg. Sorry folks. :-( 11:23:55 fiveop [~fiveop@p5DDC79F5.dip0.t-ipconnect.de] has joined #sbcl 11:24:18 White_Flame [~quassel@wsip-68-107-217-139.ph.ph.cox.net] has joined #sbcl 11:25:03 When I instantiate a 3-element struct, the constructor allocates 48 bytes, but only writes to 5 slots, 2 headers and 3 values. (x86-64, v1.1.9, checked the changelogs) 11:25:39 Is it wasting 8 bytes of the tlab, or is that expected? 11:29:25 everything in sbcl is double-word aligned 11:29:37 ah, duh, makes sense 11:30:14 so even when shuffled off to older generations, it'll still effectively take 48 bytes each 11:32:10 yes 11:32:18 think of this as an opportunity for a fourth slot for free 11:34:27 well, I would ideally prefer just the 24 bytes taken for the 3 slots... 11:34:40 a la cons, but nothing's that cheap anymore 11:35:27 both array and struct plop into 48-byte range. I guess dual-cons is 32 bytes, but that's more expensive to set up 11:48:51 -!- Poenikatu [~kvirc@pdpc/supporter/bronze/poenikatu] has left #sbcl 11:49:45 nyef [~nyef@pool-70-109-132-189.cncdnh.east.myfairpoint.net] has joined #sbcl 11:49:55 G'morning all. 11:51:52 good morning 11:55:24 -!- tmh_ [~e@unaffiliated/tmh/x-0013998] has quit [Remote host closed the connection] 11:59:10 A few weeks ago I learned about the -i flag to git add. It's awesome :) 12:02:48 nyef: where does the input in float-input-denormal-trap-bit come from in your changes? 12:03:21 I found a float-denormal-trap-bit used for x86. Are their semantics different? 12:03:27 I don't remember, tbh. 12:08:36 Either it comes from the ARM ARM, from whatever port I cribbed the original version of the definitions from, or some glitch in my wetware. 12:10:48 Why do we sometimes use functions/macros to encapsulate some instructions and then we have things like COMPUTE-LRA, multiple instructions in one define-instruction? 12:11:13 I know that the latter gives different disassemble output. But do we want that? 12:11:36 It's at least partly tradition. 12:12:17 And there's also scheduler interaction to worry about on some backends, possibly including ARM at some point (although the scheduler really is bleeding useless). 12:13:29 Not having a 1:1 correspondence between our instructions and the architecture instructions feels like a design flaw. 12:14:22 Which level of "architecture instructions"? Some architectures define instructions at the assembler level that get broken up into multiple instructions in the final code. 12:15:29 like push? 12:16:33 Specific examples aren't coming to me off the top of my head, I'm afraid. 12:16:40 Preferably, I would be able to compare trace-file output with gdb disassembly directly 12:16:48 fiveop: if you've just learnt about git add -i, wait till you discover magit 12:17:06 Krystof: I heard of that, when git was new. 12:17:25 Never tried it though 12:17:28 Krystof: HEAD doesn't build from 1.0.28. 12:18:18 1.1.16 does, though. 12:23:00 I pushed yesterdays changes to compute-lra and load-from-label as well as a fix to the vfp instructions https://github.com/fiveop/sbcl-arm 12:23:43 Cool. 12:24:05 stassats [~stassats@wikipedia/stassats] has joined #sbcl 12:24:17 I could push my other float related changes, but I'd rather wait for your VM file :) 12:26:48 Okay, I might be able to do something about that today. 12:27:16 Otherwise, it might be Wednesday or later. 12:27:17 where can i complain about people using internal features of sbcl? 12:28:00 *stassats* looks at slime as the frequent offender 12:28:06 I'd start with the people doing it. 12:28:53 breaking the internal features more often would be a discouragement 13:32:40 pranavrc [~pranavrc@122.164.196.176] has joined #sbcl 13:32:40 -!- pranavrc [~pranavrc@122.164.196.176] has quit [Changing host] 13:32:40 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 13:34:25 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 240 seconds] 13:41:08 leuler [~user@p548F9B6B.dip0.t-ipconnect.de] has joined #sbcl 13:41:32 tmh_ [~e@unaffiliated/tmh/x-0013998] has joined #sbcl 13:44:49 -!- tmh_ [~e@unaffiliated/tmh/x-0013998] has quit [Remote host closed the connection] 13:51:45 -!- sdemarre [~serge@112.84-64-87.adsl-dyn.isp.belgacom.be] has quit [Remote host closed the connection] 13:52:02 LiamH [~none@96.231.217.60] has joined #sbcl 14:01:09 nyef: ok, looking 14:01:20 Krystof: It's the first commit after the release. 14:01:34 Something about using WITH-SLOTS on a slot called POSITION. 14:02:45 Now, the right answer might be "SBCL that old is broken, upgrade your build host", but I'm not conversant enough with that part of the spec to be able to make such a determination myself. 14:05:16 it's dubious 14:05:27 with-slots on a structure isn't portable anyway 14:06:30 Okay, so me rewriting the entire thing to use accessors instead would be appropriate? 14:11:50 I'd take that 14:12:09 Okay, I'll put that together next, then. 14:12:30 thanks 14:18:46 On another topic, has anyone been taking care of the bug database recently? 14:21:14 I suspect that no-one's gone through and marked things as "Fix Released" from 1.1.16 14:21:32 Stas is quite diligent at that generally, though 14:22:08 I'm more thinking the 86 "New bugs" which may not have been confirmed/triaged and assigned an importance. 14:30:25 And done. 14:32:42 attila_lendvai [~attila_le@5.251.216.208] has joined #sbcl 14:32:42 -!- attila_lendvai [~attila_le@5.251.216.208] has quit [Changing host] 14:32:42 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 14:51:03 pnpuff [~l@unaffiliated/pnpuff] has joined #sbcl 14:51:31 -!- pnpuff [~l@unaffiliated/pnpuff] has left #sbcl 14:59:02 krishnaTejaY__ [~krishna@14.139.82.6] has joined #sbcl 15:00:03 stassats [~stassats@wikipedia/stassats] has joined #sbcl 15:09:11 -!- ASau` is now known as ASau 15:15:34 -!- stassats [~stassats@wikipedia/stassats] has quit [Remote host closed the connection] 15:16:13 stassats [~stassats@wikipedia/stassats] has joined #sbcl 15:48:13 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 264 seconds] 15:53:10 -!- fikusz [~fikusz@catv-89-132-137-62.catv.broadband.hu] has quit [Quit: Leaving] 15:58:47 fikusz [~fikusz@catv-89-132-137-62.catv.broadband.hu] has joined #sbcl 16:09:35 ggkitsas [~ggkitsas@150.140.215.17] has joined #sbcl 16:13:18 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 16:13:28 attila_lendvai [~attila_le@5.251.216.208] has joined #sbcl 16:13:28 -!- attila_lendvai [~attila_le@5.251.216.208] has quit [Changing host] 16:13:28 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 16:19:46 stassats [~stassats@wikipedia/stassats] has joined #sbcl 16:20:37 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Ping timeout: 264 seconds] 16:21:48 -!- stassats [~stassats@wikipedia/stassats] has quit [Client Quit] 16:29:31 stassats [~stassats@wikipedia/stassats] has joined #sbcl 16:44:34 >From feb10e8d6ca8de45f9c7623da9f33d18f769bfa6 16:44:36 oh, come on, it's 2014 16:44:41 how can I still be suffering from this? 16:53:03 where's feb10e8d6ca8de45f9c7623da9f33d18f769bfa6? 16:55:17 different repo 16:55:23 I was complaining about the ">From" 17:41:25 -!- edgar-rft [~GOD@HSI-KBW-109-193-013-113.hsi7.kabel-badenwuerttemberg.de] has quit [Quit: activity vanished because computer stinks] 17:49:14 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Quit: Ping timeout: ] 17:59:02 -!- antoszka [~antoszka@unaffiliated/antoszka] has quit [Ping timeout: 264 seconds] 18:11:51 -!- krishnaTejaY__ [~krishna@14.139.82.6] has quit [Read error: Connection reset by peer] 18:30:01 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 240 seconds] 18:55:18 antoszka [~antoszka@unaffiliated/antoszka] has joined #sbcl 19:19:58 anyone have a freebsd build that works on freebsd9? 19:20:56 stassats [~stassats@wikipedia/stassats] has joined #sbcl 19:22:27 slyrus: I have one somewhere. the trick was to install the 8.x compat package. 19:23:01 oh, I installed the 6x compat package... 19:24:02 or you could do a Krystof and cross compile from clisp ;) 19:24:35 when do we get to make https://github.com/sbcl/sbcl the canonical repository? 19:42:25 -!- ggkitsas [~ggkitsas@150.140.215.17] has quit [Ping timeout: 240 seconds] 19:48:34 why would we? 20:29:23 -!- angavrilov [~angavrilo@217.71.227.190] has quit [Remote host closed the connection] 20:36:13 -!- oleo [~oleo@xdsl-87-78-76-233.netcologne.de] has quit [Ping timeout: 264 seconds] 20:41:50 oleo [~oleo@xdsl-84-44-177-5.netcologne.de] has joined #sbcl 20:57:45 SB-CONCURRENCY-TEST::FRLOCK.1 fails on fbsd 9 20:58:21 Krystof: i'm not an sf.net fan 20:58:28 who is? 21:00:32 Krystof, apparently 21:00:53 i would think that Krystof is not a fan of changes 21:00:58 heh 21:01:00 indeed 21:01:27 I am not quite the most conservative person I know 21:01:28 but it's close 21:02:02 though sourceforge changes things on a whim once a year or two 21:02:08 slyrus: frlock is plain broken on non-TSO platforms :\ 21:02:25 "TSO" platforms? 21:02:48 total store order? 21:03:10 yeah. 21:03:49 And this is/isn't something that could be fixed up by sprinkling it with memory-barrier fairy dust? 21:04:28 yeah, it's just one barrier that's in the wrong position, iirc. 21:05:12 But I haven't even had time to log on the compile farm's many-way PPC to replicate the failure. 21:05:28 it fails unreliably on PPC 21:06:18 sourceforge has had substantially less downtime than github in the last 3 years 21:59:01 -!- stassats [~stassats@wikipedia/stassats] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 22:20:37 Bicyclidine [45a62f6e@gateway/web/freenode/ip.69.166.47.110] has joined #sbcl 22:21:21 -!- fiveop [~fiveop@p5DDC79F5.dip0.t-ipconnect.de] has quit [] 22:22:36 krzysz00 [~krzys@76.78.130.20] has joined #sbcl 22:32:02 -!- leuler [~user@p548F9B6B.dip0.t-ipconnect.de] has quit [Quit: ERC Version 5.1.2 $Revision: 1.796.2.6 $ (IRC client for Emacs)] 23:02:21 -!- ASau [~user@p5083DE93.dip0.t-ipconnect.de] has quit [Quit: away.] 23:29:24 scymtym_ [~user@ip-78-94-192-97.unitymediagroup.de] has joined #sbcl 23:32:16 -!- oleo [~oleo@xdsl-84-44-177-5.netcologne.de] has quit [Read error: Connection reset by peer] 23:32:19 -!- Bicyclidine [45a62f6e@gateway/web/freenode/ip.69.166.47.110] has quit [Quit: Page closed] 23:34:48 oleo [~oleo@xdsl-84-44-177-5.netcologne.de] has joined #sbcl 23:34:54 Bicyclidine [45a62f6e@gateway/web/freenode/ip.69.166.47.110] has joined #sbcl 23:40:15 davazp [~user@167.Red-88-15-120.dynamicIP.rima-tde.net] has joined #sbcl 23:41:18 -!- slyrus [~chatzilla@107.201.5.56] has quit [Ping timeout: 240 seconds] 23:58:33 slyrus [~chatzilla@107.201.5.56] has joined #sbcl 00:12:37 -!- davazp [~user@167.Red-88-15-120.dynamicIP.rima-tde.net] has quit [Read error: Connection reset by peer] 01:16:16 Okay, time for an IR1-hacking question. If I have two separate LVARs, and I want to have the compiler generate code to overwrite one with the contents of the other, how can I do this? 01:18:17 I am quite willing to burn an IR1 block to do this, and in fact am PLANNING to do so. 01:22:01 My current best guess is to introduce a BIND node to map the source LVAR to a LAMBDA-VAR and then use a REF to overwrite the destination LVAR. 02:46:25 -!- prxq [~mommer@x2f64950.dyn.telefonica.de] has quit [Ping timeout: 264 seconds] 02:54:23 -!- Bicyclidine [45a62f6e@gateway/web/freenode/ip.69.166.47.110] has quit [Ping timeout: 245 seconds] 02:57:05 kanru [~kanru@118-163-10-190.HINET-IP.hinet.net] has joined #sbcl 02:58:37 prxq [~mommer@x2f6df9c.dyn.telefonica.de] has joined #sbcl 03:16:58 -!- scymtym_ [~user@ip-78-94-192-97.unitymediagroup.de] has quit [Ping timeout: 240 seconds] 03:24:30 -!- fridim__ [~fridim@bas2-montreal07-2925317577.dsl.bell.ca] has quit [Read error: Operation timed out] 03:38:59 -!- krzysz00 [~krzys@76.78.130.20] has quit [Ping timeout: 244 seconds] 03:39:13 -!- christoph_deb [~christoph@ppp-88-217-58-110.dynamic.mnet-online.de] has quit [Ping timeout: 240 seconds] 03:47:51 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 03:50:04 krzysz00 [~krzys@76.78.130.177] has joined #sbcl 03:53:19 christoph_deb [~christoph@ppp-88-217-38-138.dynamic.mnet-online.de] has joined #sbcl 04:05:48 -!- nyef [~nyef@pool-70-109-132-189.cncdnh.east.myfairpoint.net] has quit [Quit: G'night all.] 04:09:53 -!- krzysz00 [~krzys@76.78.130.177] has quit [Ping timeout: 252 seconds] 04:18:45 -!- LiamH [~none@96.231.217.60] has quit [Quit: Leaving.] 04:25:53 krzysz00 [~krzys@76.78.130.20] has joined #sbcl 05:31:49 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 05:35:32 yacks [~py@103.6.159.103] has joined #sbcl 05:43:04 LiamH [~none@96.231.217.60] has joined #sbcl 05:44:43 -!- LiamH [~none@96.231.217.60] has quit [Client Quit] 05:45:42 pranavrc_ [~pranavrc@122.164.46.96] has joined #sbcl 05:46:49 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Ping timeout: 240 seconds] 05:48:27 you can't overwrite an lvar: lvars are write once 05:49:09 so yeah, splicing in a bind node that dominates all the relevant nodes is probably best 05:50:23 but i'm really not sure how well ir1 deals with such heavy mutation. never converting away the lambda var would be even better, if possible 06:54:51 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 06:54:58 attila_lendvai [~attila_le@5.251.216.208] has joined #sbcl 06:54:58 -!- attila_lendvai [~attila_le@5.251.216.208] has quit [Changing host] 06:54:58 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 07:03:35 -!- krzysz00 [~krzys@76.78.130.20] has quit [Ping timeout: 244 seconds] 07:06:28 angavrilov [~angavrilo@217.71.227.190] has joined #sbcl 07:14:38 -!- pranavrc_ [~pranavrc@122.164.46.96] has quit [Ping timeout: 240 seconds] 07:16:36 pranavrc [~pranavrc@122.174.14.173] has joined #sbcl 07:16:36 -!- pranavrc [~pranavrc@122.174.14.173] has quit [Changing host] 07:16:36 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 07:24:13 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Ping timeout: 264 seconds] 07:29:55 pranavrc [~pranavrc@122.164.207.71] has joined #sbcl 07:29:55 -!- pranavrc [~pranavrc@122.164.207.71] has quit [Changing host] 07:29:55 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 07:40:00 pranavrc_ [~pranavrc@122.164.110.248] has joined #sbcl 07:41:18 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Ping timeout: 240 seconds] 07:50:59 -!- yacks [~py@103.6.159.103] has quit [Ping timeout: 240 seconds] 08:13:13 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Ping timeout: 240 seconds] 08:32:03 yacks [~py@122.179.97.23] has joined #sbcl 08:52:33 attila_lendvai [~attila_le@5.251.216.208] has joined #sbcl 08:52:33 -!- attila_lendvai [~attila_le@5.251.216.208] has quit [Changing host] 08:52:33 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 09:07:22 ltbarcly [~textual@80.179.186.2.forward.012.net.il] has joined #sbcl 09:20:17 -!- oleo [~oleo@xdsl-84-44-177-5.netcologne.de] has quit [Read error: Operation timed out] 09:33:13 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Ping timeout: 264 seconds] 09:33:57 jdz [~jdz@212.36.34.246] has joined #sbcl 09:34:00 attila_lendvai [~attila_le@5.251.216.208] has joined #sbcl 09:34:00 -!- attila_lendvai [~attila_le@5.251.216.208] has quit [Changing host] 09:34:00 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 09:36:37 oleo [~oleo@xdsl-78-35-135-247.netcologne.de] has joined #sbcl 10:09:25 -!- ltbarcly [~textual@80.179.186.2.forward.012.net.il] has quit [Quit: Computer has gone to sleep.] 10:25:11 -!- jdz [~jdz@212.36.34.246] has quit [Quit: Leaving] 10:32:25 -!- kludge` [~comet@unaffiliated/espiral] has quit [Ping timeout: 240 seconds] 10:36:48 kludge` [~comet@unaffiliated/espiral] has joined #sbcl 11:13:41 -!- pranavrc_ [~pranavrc@122.164.110.248] has quit [Remote host closed the connection] 11:18:33 nyef [~nyef@pool-70-109-132-189.cncdnh.east.myfairpoint.net] has joined #sbcl 11:18:46 G'morning all. 11:19:05 morning 11:23:19 davazp [~user@167.Red-88-15-120.dynamicIP.rima-tde.net] has joined #sbcl 11:38:41 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 11:53:04 I does anyone have any idea why sbcl doesn't enter the slime debugger when I press C-c C-c (only some 20-30 seconds later)? it's a fairly normal code with no voodoo in it. there are some print-object methods, and it may be a slime issue... 11:54:03 some heavy consing is involved, because sometimes I see gc heap exhausted logs in the inferior lisp buffer 12:00:41 krzysz00 [~krzys@76.78.130.20] has joined #sbcl 12:01:19 Do you have many threads going, or just a small handful? 12:04:20 *nyef* really has no idea. 12:17:09 I think it's something big, e.g. an array, being printed for the slime backtrace, but it's merely a wild guess 12:17:33 *attila_lendvai* is trying with sbcl and slime head, but it doesn't work, "Symbol "BASIC-DEFINITION" not found in the SB-INT package." 12:25:02 That'd be a SLIME bug: SB-INT is a private package of SBCL. If anything that uses it gets broken by an SBCL change, we officially don't care. 12:25:57 I know, I rolled back to the latest sbcl release until slime gets updated 12:26:35 I've also noticed that I had a patch of my own messing with nested error handling, so I try master 12:27:52 and voila, it works. now, how on earth did I break this? 12:31:28 ... Bloody MAYBE-DELETE-EXIT, screwing up perfectly good IR1 structure. /-: 12:31:49 *attila_lendvai* becomes skeptical and rebases his change to be super sure (https://github.com/attila-lendvai/sbcl/commit/90fd9e9385857737c75c500f416aaab424a4ef71) 13:04:44 ltbarcly [~textual@93.94.40.14] has joined #sbcl 13:06:59 -!- yacks [~py@122.179.97.23] has quit [Quit: Leaving] 13:09:37 -!- ltbarcly [~textual@93.94.40.14] has quit [Read error: Connection reset by peer] 13:10:24 ltbarcly [~textual@93.94.40.14] has joined #sbcl 13:10:29 jrm [~jrm@freebsd/hacker/jrm] has joined #sbcl 13:21:56 -!- ltbarcly [~textual@93.94.40.14] has quit [Quit: Computer has gone to sleep.] 13:33:56 yacks [~py@103.6.159.103] has joined #sbcl 13:45:38 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [] 13:53:26 LiamH [~none@96.231.217.60] has joined #sbcl 13:53:38 -!- asedeno [~asedeno@66.102.14.26] has quit [Ping timeout: 264 seconds] 13:53:51 asedeno [~asedeno@66.102.14.25] has joined #sbcl 14:17:15 attila_lendvai: I've had that for years, but only on os x. 14:42:22 jdz [~jdz@212.36.34.246] has joined #sbcl 14:42:58 the "quid-pro-quo" system fails to compile due to its use of sb-int:arg-list (http://cl-test-grid.appspot.com/blob?key=cc9ry6mb3y); i can file a bug against that system, right? 14:43:25 scymtym: Right, any use of SB-INT is explicitly unsupported by SBCL. If it breaks, it's not our problem. 14:43:36 nyef: thank you 15:02:13 michael_lee [~michael_l@113.135.104.168] has joined #sbcl 15:05:33 segv- [~mb@95-91-241-251-dynip.superkabel.de] has joined #sbcl 15:21:13 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [Remote host closed the connection] 15:34:56 drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has joined #sbcl 15:51:34 Okay, this nonlinear LVAR thing is really starting to bug me. I know where the IR1 structure is getting trashed, but I'm drawing a massive blank on how to keep it from getting trashed. 15:53:51 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [Remote host closed the connection] 15:54:15 I think it would be nice if the encapsulate reworking led to a supported encapsulation API 15:55:18 ideally one which could support advice, and wouldn't need changing if we ever implemented e.g. rewriting of some nops so that we could modify functions without changing the identity of function objects 15:56:16 You mean... provide an explicit space to overwrite using :ENCAPSULATE NIL, so that it's actually thread-safe? 15:56:25 ltbarcly [~textual@93.94.40.14] has joined #sbcl 16:00:21 -!- ltbarcly [~textual@93.94.40.14] has quit [Read error: Connection reset by peer] 16:00:39 ltbarcly [~textual@93.94.40.14] has joined #sbcl 16:01:43 -!- ltbarcly [~textual@93.94.40.14] has quit [Client Quit] 16:03:46 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 16:22:18 -!- yacks [~py@103.6.159.103] has quit [Quit: Leaving] 16:30:11 fridim_ [~fridim@173.231.115.58] has joined #sbcl 16:39:45 yes 16:40:37 brown` [user@nat/google/x-ywfhelhlqkrajcky] has joined #sbcl 16:56:48 drmeister [~drmeister@155.247.96.196] has joined #sbcl 16:56:53 root_empire [~michael_l@117.32.237.91] has joined #sbcl 17:00:04 -!- michael_lee [~michael_l@113.135.104.168] has quit [Ping timeout: 265 seconds] 17:09:14 -!- drmeister [~drmeister@155.247.96.196] has quit [Remote host closed the connection] 17:11:34 drmeister [~drmeister@155.247.96.196] has joined #sbcl 17:37:49 -!- angavrilov [~angavrilo@217.71.227.190] has quit [Remote host closed the connection] 17:40:33 -!- drmeister [~drmeister@155.247.96.196] has quit [Remote host closed the connection] 17:44:04 drmeister [~drmeister@155.247.96.196] has joined #sbcl 17:54:02 -!- krzysz00 [~krzys@76.78.130.20] has quit [Ping timeout: 252 seconds] 17:56:27 wbooze [~wbooze@xdsl-78-35-135-247.netcologne.de] has joined #sbcl 18:19:24 fiveop [~fiveop@p5DDC79F5.dip0.t-ipconnect.de] has joined #sbcl 18:21:28 xani [~user@178.183.144.254.dsl.dynamic.t-mobile.pl] has joined #sbcl 18:22:48 Why are there irix related conditionals in the code? 18:23:36 Because irix was a supported platform at some point? 18:23:48 So it is not supported anymore? 18:23:57 Do you have an irix box? 18:24:03 I certainly don't. 18:24:20 Next you'll be asking about the HPPA code. 18:24:29 According to wikipedia the whole OS is no longer supported since last December. 18:24:36 No 18:25:11 Would someone mind, if I removed the irix related code? 18:25:22 Is it in the way? 18:25:49 It makes things dirty. 18:26:32 In runtime.c we have #ifdef irix #include .. #include .. #endif. Both headers have already been included unconditionally by that point. 18:28:46 -!- davazp [~user@167.Red-88-15-120.dynamicIP.rima-tde.net] has quit [Remote host closed the connection] 18:33:21 -!- jdz [~jdz@212.36.34.246] has quit [Ping timeout: 265 seconds] 18:36:03 jdz [~jdz@212.36.34.246] has joined #sbcl 18:59:12 -!- jdz [~jdz@212.36.34.246] has quit [Ping timeout: 240 seconds] 19:02:36 -!- drmeister [~drmeister@155.247.96.196] has quit [Read error: Connection reset by peer] 19:03:08 drmeister [~drmeister@155.247.96.196] has joined #sbcl 19:05:01 jdz [~jdz@212.36.34.246] has joined #sbcl 19:18:19 izirku [~IceChat9@sys-its-g56vnw1-yev.cc.unt.edu] has joined #sbcl 19:19:54 -!- izirku [~IceChat9@sys-its-g56vnw1-yev.cc.unt.edu] has quit [Client Quit] 19:20:24 izirku [~IceChat9@sys-its-g56vnw1-yev.cc.unt.edu] has joined #sbcl 19:27:58 -!- jdz [~jdz@212.36.34.246] has quit [Ping timeout: 240 seconds] 19:31:01 jdz [~jdz@212.36.34.246] has joined #sbcl 19:34:13 Where is *WILD-TYPE* defined, and what does it actually mean in practical terms? 19:37:24 to what the object it points to resolves ? 19:37:38 like file, directory .... ? 19:39:16 Ah, found it. There's a !COLD-INIT-FORMS MACROLET FROB in code/late-type which does the damage. 19:40:01 it's for things like (array *) 19:40:11 which is different from (array t) 19:40:38 I'm more dealing with an LVAR associated with an EXIT here. (-: 19:41:54 I think I have the CAST node that I'm trying to create build properly now, and possibly the start of the block that I need, so next I need to get the block linked into the component (and control flow) correctly. 19:43:41 Hacking IR1 at this level makes my head hurt. 20:09:43 *nyef* sighs. 20:10:06 "ctran-strats-block"? REALLY? Clearly, I don't quite know what I'm doing here. /-: 20:11:00 irrat compiles 20:11:08 Congratulations! 20:12:13 drink! 20:12:56 I added alien-type-methods for floats, but throw an error if we get past needing 7 double or 15 single registers. 20:12:59 So no stack args yet 20:13:53 "failed AVER: SUCC". Lovely. How did I break this? 20:14:25 ltbarcly [~textual@80.179.186.2.forward.012.net.il] has joined #sbcl 20:16:33 ... By using the wrong block somewhere, for starters. 20:16:48 Consequently, I didn't implement moving arguments from the stack 20:18:18 -!- ltbarcly [~textual@80.179.186.2.forward.012.net.il] has quit [Client Quit] 20:18:41 krzysz00 [~krzys@76.78.130.20] has joined #sbcl 20:21:38 -!- slyrus [~chatzilla@107.201.5.56] has quit [Ping timeout: 240 seconds] 20:24:18 fiveop: so where are you currently stuck? Still on the stack guard page stuff? 20:24:43 slyrus [~chatzilla@107.201.5.56] has joined #sbcl 20:27:18 scymtym_ [~user@ip-78-94-192-97.unitymediagroup.de] has joined #sbcl 20:35:45 I am stuck there. But I currently don't try to overcome that. If I try myself at it again, I'll do it on free weekend day. 20:36:21 I'm hoping to have some time to look into the guard page damage next weekend or sooner. 20:37:43 I went through the change logs and made notes in build-order for all files that have conditionals to remove float related code. 20:38:00 Next, I'll try to get rid of those 20:41:51 I... have a working patch for lp#309099. 20:42:34 It's bloody BEAUTIFUL to see the cleanup being spliced into the correct location, and the test case return a correct value instead of garbage. 20:45:02 Also fixes lp#518099 and lp#533930. 20:53:36 -!- drmeister [~drmeister@155.247.96.196] has quit [Ping timeout: 264 seconds] 20:56:18 -!- krzysz00 [~krzys@76.78.130.20] has quit [Ping timeout: 240 seconds] 20:59:28 http://paste.lisp.org/display/141469 if anyone cares to see the results. (-: 21:04:12 wow 6 year old bug squashed! nice! 21:04:22 older than 6 years old 21:04:29 6 years is when we started using launchpad 21:04:37 you'll probably say that it's time to move to github issues or something 21:05:05 lp is much better that whatever sf.net had 21:05:05 https://bugs.launchpad.net/bugs/is much better that whatever sf.net had 21:05:18 Heh. 21:05:25 haha 21:06:24 I can certainly understand why nobody else fixed this one. It took most of a day just to figure out the actual mechanism by which it was broken, let alone devise a fix. 21:17:24 -!- fiveop [~fiveop@p5DDC79F5.dip0.t-ipconnect.de] has quit [] 21:29:01 ltbarcly [~textual@80.179.186.2.forward.012.net.il] has joined #sbcl 21:32:30 -!- ltbarcly [~textual@80.179.186.2.forward.012.net.il] has quit [Client Quit] 21:34:56 drmeister [~drmeister@155.247.96.196] has joined #sbcl 21:40:01 -!- xani [~user@178.183.144.254.dsl.dynamic.t-mobile.pl] has quit [Ping timeout: 240 seconds] 21:45:16 Okay, feedback time. Does the following commit (and its parent) look reasonable to those of you familiar with IR1 hacking? http://repo.or.cz/w/sbcl/nyef.git/commitdiff/bbbb745fc78901d6fe6b3f21fd5099a26570279a 21:46:34 pkhuong: If you're around, this means you. (-: 21:46:54 ha 21:48:02 There's no way I'm pushing this to HEAD without a review. 21:48:04 xani [~user@178.183.147.143.dsl.dynamic.t-mobile.pl] has joined #sbcl 21:48:32 -!- drmeister [~drmeister@155.247.96.196] has quit [Read error: Connection reset by peer] 21:48:58 drmeister [~drmeister@155.247.96.196] has joined #sbcl 21:50:23 It may not be the third of the high-importance bugs that I was hoping for, but at least it's a fifth of them! (-: 21:51:13 it is pretty awesome 21:51:20 maybe you can be called "apd mark II" 22:04:06 One problem that I have is that this is the first major piece of IR1 hacking I've done, and while it seems to work I have no idea if I'm doing everything correctly or not. 22:05:13 -!- wbooze [~wbooze@xdsl-78-35-135-247.netcologne.de] has quit [Ping timeout: 240 seconds] 22:05:37 -!- oleo [~oleo@xdsl-78-35-135-247.netcologne.de] has quit [Ping timeout: 240 seconds] 22:06:00 oleo [~oleo@xdsl-78-35-173-51.netcologne.de] has joined #sbcl 22:19:20 -!- drmeister [~drmeister@155.247.96.196] has quit [Read error: Connection reset by peer] 22:19:53 drmeister [~drmeister@155.247.96.196] has joined #sbcl 22:35:03 -!- xani [~user@178.183.147.143.dsl.dynamic.t-mobile.pl] has quit [Quit: rcirc on GNU Emacs 24.3.1] 22:57:48 -!- drmeister [~drmeister@155.247.96.196] has quit [Remote host closed the connection] 22:58:42 Krystof: Do you have an angle for adding support for newline styles in external formats? 23:06:43 -!- jdz [~jdz@212.36.34.246] has quit [Read error: Connection timed out] 23:10:51 drmeister [~drmeister@166.170.22.11] has joined #sbcl 23:13:51 fridim__ [~fridim@173.231.115.58] has joined #sbcl 23:14:16 jdz [~jdz@212.36.34.246] has joined #sbcl 23:19:58 -!- jdz [~jdz@212.36.34.246] has quit [Ping timeout: 265 seconds] 23:21:34 jdz [~jdz@212.36.34.246] has joined #sbcl 23:29:38 -!- jdz [~jdz@212.36.34.246] has quit [Ping timeout: 252 seconds] 23:29:59 jdz [~jdz@212.36.34.246] has joined #sbcl 23:34:14 -!- fikusz [~fikusz@catv-89-132-137-62.catv.broadband.hu] has quit [Read error: Operation timed out] 23:37:38 -!- drmeister [~drmeister@166.170.22.11] has quit [Read error: Connection reset by peer] 23:46:58 nyef: i once started by making external-formats composite objects containing a character-coding and a newline-coding 23:50:05 That's one of the possibilities I'd consider, yes. 23:50:07 fikusz [~fikusz@catv-89-132-137-62.catv.broadband.hu] has joined #sbcl 23:51:02 The reason I'm asking is that I'd like to see the number of "high-importance" bugs in the tracker knocked down a bit, and external-format line-endings accounts for two out of twelve unfixed bugs. 23:51:20 (out of fifteen high-importance bugs.) 23:52:53 in addition, having support for newline-styles would be really useful in practice 23:54:16 Mmm. Who knows when they're going to run into one of those VMS-style text files, which have a 16-bit little-endian line length count followed by as many ASCII characters, followed by an optional alignment byte for each line, after all. (-: 23:55:19 important use-case indeed :) 23:56:26 I can't believe that lp#1248181 is all that important, though. Who still uses solaris? 23:57:05 How about the IBM "blocked" style files, which have a series of "blocks" containing a 16bit big endian block length, then within that, a sequence of records, consisting of a 16bit big-endian record-length and data (in EBCDIC of course)? 23:57:19 Surely that needs to be supported! 23:57:37 -!- jdz [~jdz@212.36.34.246] has quit [Ping timeout: 240 seconds] 23:58:02 Right. But CRLF-style line endings are clearly bogus, nobody uses those. 23:58:31 We'll need to support CR-style line-endings though, somebody might have a 68k mac lying around. 23:58:58 jdz [~jdz@212.36.34.246] has joined #sbcl 00:01:02 I'm not entirely convinced that lp#659173 is valid, as it seems to require using an ftype designator in a manner not supported by the standard. 00:01:25 lp 659173 00:01:25 https://bugs.launchpad.net/bugs/659173 00:02:20 yeah, that's lying to the compiler, isn't it 00:02:33 maybe could warn still? 00:03:01 What grounds are there for a warning, though? 00:04:13 basically the same reason you get a warning from doing the same thing with defun? 00:04:39 foom: I don't suppose you're conversant enough with IR1 to be able to comment on my recent branch? 00:05:48 nyef: probably not, but I'll take a look anyways. :) 00:06:02 Do you have the link, or shall I re-post it? 00:06:08 http://repo.or.cz/w/sbcl/nyef.git/commitdiff/bbbb745fc78901d6fe6b3f21fd5099a26570279a ? 00:06:26 That'd be the one. 00:06:43 *foom* -> dinner. 00:08:00 -!- fridim__ [~fridim@173.231.115.58] has quit [Ping timeout: 264 seconds] 00:15:21 krzysz00 [~krzys@76.78.130.20] has joined #sbcl 00:21:12 -!- xymox [lechuck@unaffiliated/contempt] has quit [Ping timeout: 264 seconds] 00:23:05 fridim__ [~fridim@65.93.79.33] has joined #sbcl 00:26:15 xymox [lechuck@unaffiliated/contempt] has joined #sbcl 00:30:58 -!- xymox [lechuck@unaffiliated/contempt] has quit [Read error: Connection reset by peer] 00:33:12 -!- jdz [~jdz@212.36.34.246] has quit [Ping timeout: 264 seconds] 00:33:39 jdz [~jdz@212.36.34.246] has joined #sbcl 00:38:11 xymox [~xymox@unaffiliated/contempt] has joined #sbcl 00:42:09 -!- xymox [~xymox@unaffiliated/contempt] has quit [Remote host closed the connection] 00:48:49 xymox [lechuck@unaffiliated/contempt] has joined #sbcl 00:53:13 -!- xymox [lechuck@unaffiliated/contempt] has quit [Ping timeout: 240 seconds] 00:57:08 xymox [lechuck@unaffiliated/contempt] has joined #sbcl 01:06:12 -!- Bike [~Glossina@gannon-wless-gw.resnet.wsu.edu] has quit [Quit: restartin'] 01:07:42 Bike [~Glossina@gannon-wless-gw.resnet.wsu.edu] has joined #sbcl 01:09:12 -!- xymox [lechuck@unaffiliated/contempt] has quit [Ping timeout: 265 seconds] 01:12:33 xymox [lechuck@unaffiliated/contempt] has joined #sbcl 01:19:51 -!- xymox [lechuck@unaffiliated/contempt] has quit [Ping timeout: 265 seconds] 01:20:18 drmeister [~drmeister@166.170.22.11] has joined #sbcl 01:20:51 xymox [lechuck@unaffiliated/contempt] has joined #sbcl 01:22:20 -!- drmeister [~drmeister@166.170.22.11] has quit [Read error: Connection reset by peer] 01:28:24 -!- xymox [lechuck@unaffiliated/contempt] has quit [Ping timeout: 264 seconds] 01:28:43 -!- root_empire [~michael_l@117.32.237.91] has quit [Read error: Connection reset by peer] 01:29:38 xymox [lechuck@unaffiliated/contempt] has joined #sbcl 01:38:20 drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has joined #sbcl 01:42:40 michael_lee [~michael_l@117.32.237.91] has joined #sbcl 02:04:28 Ooh. Can we mark lp#484143 as a duplicate of lp#309096 ? 02:05:00 -!- jdz [~jdz@212.36.34.246] has quit [Ping timeout: 264 seconds] 02:09:43 jdz [~jdz@212.36.34.246] has joined #sbcl 02:14:58 -!- jdz [~jdz@212.36.34.246] has quit [Ping timeout: 240 seconds] 02:15:12 ASau [~user@46.115.151.45] has joined #sbcl 02:15:43 jdz [~jdz@212.36.34.246] has joined #sbcl 02:26:18 -!- slyrus [~chatzilla@107.201.5.56] has quit [Ping timeout: 240 seconds] 02:27:25 slyrus [~chatzilla@107.201.5.56] has joined #sbcl 02:28:09 lp484143 02:28:16 lp 484143 02:28:17 https://bugs.launchpad.net/bugs/484143 02:35:43 I've long thought that 484143 was given far too high an importance, and 309096 simply confirms my suspicion, as it's the same issue at heart. 02:38:02 -!- ASau [~user@46.115.151.45] has quit [Remote host closed the connection] 02:38:38 ASau [~user@46.115.151.45] has joined #sbcl 02:40:53 attila_lendvai [~attila_le@5.251.216.208] has joined #sbcl 02:40:53 -!- attila_lendvai [~attila_le@5.251.216.208] has quit [Changing host] 02:40:53 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 02:43:13 prxq_ [~mommer@x2f6cbcf.dyn.telefonica.de] has joined #sbcl 02:46:00 -!- prxq [~mommer@x2f6df9c.dyn.telefonica.de] has quit [Ping timeout: 240 seconds] 02:49:18 -!- slyrus [~chatzilla@107.201.5.56] has quit [Ping timeout: 240 seconds] 02:52:04 yacks [~py@103.6.159.103] has joined #sbcl 02:53:00 -!- ASau [~user@46.115.151.45] has quit [Quit: reboot] 02:57:16 ASau [~user@46.115.151.45] has joined #sbcl 03:07:13 slyrus [~chatzilla@107.201.5.56] has joined #sbcl 03:13:59 edgar-rft [~GOD@HSI-KBW-109-193-013-113.hsi7.kabel-badenwuerttemberg.de] has joined #sbcl 03:22:18 -!- jdz [~jdz@212.36.34.246] has quit [Ping timeout: 240 seconds] 03:24:13 jdz [~jdz@212.36.34.246] has joined #sbcl 03:38:05 -!- fridim__ [~fridim@65.93.79.33] has quit [Ping timeout: 265 seconds] 03:38:25 -!- christoph_deb [~christoph@ppp-88-217-38-138.dynamic.mnet-online.de] has quit [Ping timeout: 240 seconds] 03:46:10 -!- jdz [~jdz@212.36.34.246] has quit [Ping timeout: 244 seconds] 03:46:39 jdz [~jdz@212.36.34.246] has joined #sbcl 03:52:39 christoph_deb [~christoph@ppp-88-217-43-129.dynamic.mnet-online.de] has joined #sbcl 04:07:13 -!- krzysz00 [~krzys@76.78.130.20] has quit [Read error: Connection reset by peer] 04:15:38 -!- LiamH [~none@96.231.217.60] has quit [Quit: Leaving.] 04:17:36 -!- jdz [~jdz@212.36.34.246] has quit [Ping timeout: 264 seconds] 04:23:18 -!- slyrus [~chatzilla@107.201.5.56] has quit [Ping timeout: 240 seconds] 04:24:47 jdz [~jdz@212.36.34.246] has joined #sbcl 04:31:07 -!- jdz [~jdz@212.36.34.246] has quit [Ping timeout: 252 seconds] 04:31:54 -!- nyef [~nyef@pool-70-109-132-189.cncdnh.east.myfairpoint.net] has quit [Quit: G'night all.] 04:33:39 jdz [~jdz@212.36.34.246] has joined #sbcl 04:42:08 -!- scymtym_ [~user@ip-78-94-192-97.unitymediagroup.de] has quit [Ping timeout: 253 seconds] 04:45:26 slyrus [~chatzilla@107.201.5.56] has joined #sbcl 05:01:12 -!- jdz [~jdz@212.36.34.246] has quit [Ping timeout: 240 seconds] 05:05:32 ltbarcly [~textual@80.179.186.2.forward.012.net.il] has joined #sbcl 05:07:01 jdz [~jdz@212.36.34.246] has joined #sbcl 05:12:41 -!- ltbarcly [~textual@80.179.186.2.forward.012.net.il] has quit [Quit: Computer has gone to sleep.] 05:25:25 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 05:40:50 I'll take a look at the ir1 magic tomorrow pm. 05:42:00 -!- jdz [~jdz@212.36.34.246] has quit [Ping timeout: 240 seconds] 05:43:31 jdz [~jdz@212.36.34.246] has joined #sbcl 05:56:43 -!- oleo [~oleo@xdsl-78-35-173-51.netcologne.de] has quit [Quit: Leaving] 06:01:00 -!- ASau [~user@46.115.151.45] has quit [Quit: off.] 06:08:18 -!- jdz [~jdz@212.36.34.246] has quit [Ping timeout: 240 seconds] 06:10:26 jdz [~jdz@212.36.34.246] has joined #sbcl 06:31:32 -!- prxq_ is now known as prxq 06:32:11 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Remote host closed the connection] 06:32:28 pranavrc [~pranavrc@122.164.110.248] has joined #sbcl 06:32:38 -!- pranavrc [~pranavrc@122.164.110.248] has quit [Changing host] 06:32:38 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 06:45:57 pranavrc_ [~pranavrc@122.174.8.253] has joined #sbcl 06:48:02 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Ping timeout: 244 seconds] 06:50:19 -!- pranavrc_ [~pranavrc@122.174.8.253] has quit [Ping timeout: 252 seconds] 06:51:47 sdemarre [~serge@112.84-64-87.adsl-dyn.isp.belgacom.be] has joined #sbcl 06:52:59 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 07:02:00 -!- jdz [~jdz@212.36.34.246] has quit [Ping timeout: 264 seconds] 07:06:54 jdz [~jdz@212.36.34.246] has joined #sbcl 07:07:48 angavrilov [~angavrilo@217.71.227.190] has joined #sbcl 07:21:08 -!- jdz [~jdz@212.36.34.246] has quit [Ping timeout: 252 seconds] 07:21:44 jdz [~jdz@212.36.34.246] has joined #sbcl 07:33:18 -!- jdz [~jdz@212.36.34.246] has quit [Ping timeout: 240 seconds] 07:34:40 -!- yacks [~py@103.6.159.103] has quit [Quit: Leaving] 07:38:54 jdz [~jdz@212.36.34.246] has joined #sbcl 07:50:02 -!- sdemarre [~serge@112.84-64-87.adsl-dyn.isp.belgacom.be] has quit [Ping timeout: 244 seconds] 07:59:09 pranavrc_ [~pranavrc@122.164.96.141] has joined #sbcl 07:59:45 yacks [~py@122.179.78.185] has joined #sbcl 08:00:00 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Ping timeout: 240 seconds] 08:00:14 -!- yacks [~py@122.179.78.185] has quit [Max SendQ exceeded] 08:01:01 yacks [~py@122.179.78.185] has joined #sbcl 08:04:24 -!- pranavrc_ [~pranavrc@122.164.96.141] has quit [Ping timeout: 264 seconds] 08:06:24 pranavrc [~pranavrc@122.164.171.238] has joined #sbcl 08:06:24 -!- pranavrc [~pranavrc@122.164.171.238] has quit [Changing host] 08:06:24 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 08:35:59 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 08:51:17 -!- jdz [~jdz@212.36.34.246] has quit [Ping timeout: 265 seconds] 08:51:34 jdz [~jdz@212.36.34.246] has joined #sbcl 09:02:28 pranavrc_ [~pranavrc@122.164.134.200] has joined #sbcl 09:03:38 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Ping timeout: 240 seconds] 09:07:24 -!- pranavrc_ [~pranavrc@122.164.134.200] has quit [Ping timeout: 264 seconds] 09:07:55 pranavrc [~pranavrc@122.164.122.159] has joined #sbcl 09:07:55 -!- pranavrc [~pranavrc@122.164.122.159] has quit [Changing host] 09:07:55 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 09:14:38 pranavrc_ [~pranavrc@122.164.111.134] has joined #sbcl 09:15:58 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Ping timeout: 240 seconds] 09:17:02 ehaliewicz [~user@50-0-50-115.dsl.dynamic.sonic.net] has joined #sbcl 09:20:46 -!- jdz [~jdz@212.36.34.246] has quit [Ping timeout: 265 seconds] 09:21:09 jdz [~jdz@212.36.34.246] has joined #sbcl 09:24:01 pranavrc [~pranavrc@122.164.233.93] has joined #sbcl 09:24:01 -!- pranavrc [~pranavrc@122.164.233.93] has quit [Changing host] 09:24:01 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 09:26:01 -!- pranavrc_ [~pranavrc@122.164.111.134] has quit [Ping timeout: 240 seconds] 09:36:05 -!- jdz [~jdz@212.36.34.246] has quit [Quit: Leaving] 09:42:50 i'm running an old version of sbcl, but i just noticed this runs without a division by zero error http://paste.lisp.org/display/141475 10:07:10 the CLHS says "might signal" 10:07:26 so you have no guarantees about a division by zero error 10:07:57 "This means that the situation has undefined consequences; however, if an error is signaled, it is of the specified type. For example, ``open might signal an error of type file-error.''" 10:26:14 pranavrc_ [~pranavrc@122.174.31.51] has joined #sbcl 10:27:29 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Ping timeout: 240 seconds] 10:30:51 -!- kludge` [~comet@unaffiliated/espiral] has quit [Ping timeout: 265 seconds] 10:36:27 attila_lendvai [~attila_le@5.251.216.208] has joined #sbcl 10:36:27 -!- attila_lendvai [~attila_le@5.251.216.208] has quit [Changing host] 10:36:27 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 10:36:34 kludge` [~comet@unaffiliated/espiral] has joined #sbcl 10:40:24 -!- yacks [~py@122.179.78.185] has quit [Ping timeout: 264 seconds] 10:41:15 -!- ehaliewicz [~user@50-0-50-115.dsl.dynamic.sonic.net] has quit [Remote host closed the connection] 11:01:29 yacks [~py@122.179.78.185] has joined #sbcl 11:12:32 is Jan Moringen here? 12:02:18 -!- yacks [~py@122.179.78.185] has quit [Quit: Leaving] 12:12:49 nyef [~nyef@pool-70-109-132-189.cncdnh.east.myfairpoint.net] has joined #sbcl 12:13:01 G'morning all. 12:32:02 attila_lendvai: just arrived 12:57:17 -!- segv- [~mb@95-91-241-251-dynip.superkabel.de] has quit [Ping timeout: 253 seconds] 13:09:06 eudoxia [~eudoxia@r179-24-17-177.dialup.adsl.anteldata.net.uy] has joined #sbcl 13:16:18 yacks [~py@103.6.159.103] has joined #sbcl 13:44:29 -!- Bike [~Glossina@gannon-wless-gw.resnet.wsu.edu] has quit [Ping timeout: 240 seconds] 13:45:38 -!- pranavrc_ [~pranavrc@122.174.31.51] has quit [] 13:56:48 fridim__ [~fridim@bas2-montreal07-2925317685.dsl.bell.ca] has joined #sbcl 14:04:35 davazp [~user@167.Red-88-15-120.dynamicIP.rima-tde.net] has joined #sbcl 14:17:44 scymtym: wrote a mail since then... but looking at the responses it looks like a bigger can of worms than I thought 14:18:06 *attila_lendvai* goes offline 14:18:11 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 14:23:48 LiamH [~none@129-2-129-146.wireless.umd.edu] has joined #sbcl 14:26:32 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [Remote host closed the connection] 14:27:10 drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has joined #sbcl 14:31:13 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [Ping timeout: 240 seconds] 14:38:24 segv- [~mb@95-91-241-251-dynip.superkabel.de] has joined #sbcl 14:43:34 -!- LiamH [~none@129-2-129-146.wireless.umd.edu] has quit [Quit: Leaving.] 14:46:17 drmeister [~drmeister@166.170.22.11] has joined #sbcl 14:49:49 wbooze [~wbooze@xdsl-87-79-199-91.netcologne.de] has joined #sbcl 14:50:49 oleo [~oleo@xdsl-87-79-199-91.netcologne.de] has joined #sbcl 14:51:36 jdz [~jdz@212.36.34.246] has joined #sbcl 14:53:01 ASau [~user@176.5.184.127] has joined #sbcl 15:08:35 -!- eudoxia [~eudoxia@r179-24-17-177.dialup.adsl.anteldata.net.uy] has quit [Quit: Leaving] 15:11:00 -!- drmeister [~drmeister@166.170.22.11] has quit [Ping timeout: 264 seconds] 15:17:19 drmeister [~drmeister@155.247.96.196] has joined #sbcl 15:49:32 -!- wbooze [~wbooze@xdsl-87-79-199-91.netcologne.de] has quit [Quit: none] 16:12:21 -!- yacks [~py@103.6.159.103] has quit [Quit: Leaving] 16:14:18 -!- slyrus [~chatzilla@107.201.5.56] has quit [Ping timeout: 240 seconds] 16:41:16 Bike [~Glossina@gannon-wless-gw.resnet.wsu.edu] has joined #sbcl 16:52:33 slyrus [~chatzilla@107.201.5.56] has joined #sbcl 16:59:07 I think I just managed to trap 1.1.5.23 in an infinite loop by way of trying to compile what looked to me like a reasonable LOOP form. 17:00:18 ... Nevermind, not a reasonable LOOP form. 17:02:42 pranavrc [~pranavrc@122.164.161.85] has joined #sbcl 17:02:42 -!- pranavrc [~pranavrc@122.164.161.85] has quit [Changing host] 17:02:42 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 17:04:38 -!- slyrus [~chatzilla@107.201.5.56] has quit [Ping timeout: 240 seconds] 17:06:00 On the other hand, I might want to report a bug about what happens when someone passes the compiler an infinitely-long function-call form. 17:09:27 slyrus [~chatzilla@107.201.5.56] has joined #sbcl 17:18:42 attila_lendvai [~attila_le@5.251.216.208] has joined #sbcl 17:18:42 -!- attila_lendvai [~attila_le@5.251.216.208] has quit [Changing host] 17:18:42 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 17:21:26 fiveop [~fiveop@p5DDC79F5.dip0.t-ipconnect.de] has joined #sbcl 17:27:23 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 17:32:18 -!- Bike [~Glossina@gannon-wless-gw.resnet.wsu.edu] has quit [Ping timeout: 240 seconds] 17:58:20 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Quit: Ping timeout: ] 18:03:37 -!- michael_lee [~michael_l@117.32.237.91] has quit [Read error: Connection reset by peer] 18:17:28 -!- fiveop [~fiveop@p5DDC79F5.dip0.t-ipconnect.de] has quit [] 18:28:25 -!- edgar-rft [~GOD@HSI-KBW-109-193-013-113.hsi7.kabel-badenwuerttemberg.de] has quit [Quit: memory access abandoned because no continuation expected] 18:31:18 -!- slyrus [~chatzilla@107.201.5.56] has quit [Ping timeout: 240 seconds] 18:47:55 Bike [~Glossina@gannon-wless-gw.resnet.wsu.edu] has joined #sbcl 18:54:02 -!- fikusz [~fikusz@catv-89-132-137-62.catv.broadband.hu] has quit [Quit: Leaving] 18:55:50 -!- angavrilov [~angavrilo@217.71.227.190] has quit [Remote host closed the connection] 18:56:25 fikusz [~fikusz@catv-89-132-137-62.catv.broadband.hu] has joined #sbcl 18:56:37 -!- milosn [~milosn@94.12.79.143] has quit [Read error: No route to host] 18:56:57 milosn [~milosn@94.12.79.143] has joined #sbcl 19:15:48 -!- drmeister [~drmeister@155.247.96.196] has quit [Ping timeout: 264 seconds] 19:43:06 drmeister [~drmeister@155.247.96.196] has joined #sbcl 19:55:31 slyrus [~chatzilla@173-228-44-92.dsl.static.sonic.net] has joined #sbcl 20:03:46 scymtym_ [~user@ip-78-94-192-97.unitymediagroup.de] has joined #sbcl 20:32:50 nyef: I think it makes sense. 20:34:18 nyef: is there a reason to believe we'll never have to deal with multiple predecessors? 20:35:11 and we're replacing an EXIT node which always terminates its block 20:35:29 so that's one problem we don't have to deal with. 20:36:00 -!- drmeister [~drmeister@155.247.96.196] has quit [Ping timeout: 265 seconds] 20:37:47 Multiple predecessors for which node/block? 20:38:20 Hrm. For the EXIT? 20:41:56 Or are you thinking about the explicit CAR on the result of BLOCK-SUCC? 20:43:41 does the exit node necessarily start its block? 20:43:59 shouldn't we link the exit's block to our new one and then flush the exit? 20:44:07 The exit node necessarily ENDS its block, but does not necessarily start it. 20:44:35 I see, yeah 20:45:16 And if you look at the sequence of operations, we adjust the block links well before we delete the EXIT node. 20:46:21 The one change that I'd make to this mess at this point is to use DESTRUCTURING-BIND to get my hands on ENTRY-BLOCK instead of simply using CAR. 20:47:17 I know that the only case where BLOCK-SUCC is not a singleton is if the last node in the block is an IF, and in this case it has to be an EXIT instead, but I'd like the extra runtime validation. 20:47:38 I'd avert singleton-ness for clarity 20:47:40 *aver 20:48:46 Mmm... I think I'd rather have a SINGLETON-VALUE function in that instance, or just use DS-B and add a comment to the effect that an EXIT node only ever has one successor or something like that. 20:50:00 Anyway, the patch looks reasonable, and I haven't done anything obviously wrong with the IR1 structures? 20:50:20 Nothing obvious to me. 20:50:23 drmeister [~drmeister@155.247.96.196] has joined #sbcl 20:50:29 Cool, thanks. 20:50:47 I'll possibly tweak the ENTRY-BLOCK stuff and then commit, well in time for the next release. (-: 20:52:37 As I mentioned yesterday, this is my first major IR1 hack, so I really wanted extra eyes and feedback on it before I committed it, even if it DID survive a build and appear to fix the issue. 20:54:16 cheryllium [~chatzilla@128.237.213.55] has joined #sbcl 20:55:40 sdemarre [~serge@112.84-64-87.adsl-dyn.isp.belgacom.be] has joined #sbcl 20:58:32 Hello all, I'm a student interested in working with sbcl for the google summer of code 20:58:57 I was wondering if I could discuss project ideas with anyone here? 20:59:21 Yes, this would be one good place to discuss project ideas. 20:59:48 I'm not a mentor, but I hope that I might at least serve as a good sounding board. 21:00:26 And the first thing to ask is if you've seen http://www.sbcl.org/gsoc2013/ideas/ yet? 21:00:29 yup! 21:00:49 I was looking through that page, and there's just so many ideas, I can't decide which to focus on 21:00:58 I'm also a little intimidated by it 21:01:38 the reason I was excited to hear that sbcl would be part of gsoc is that I use sbcl very often. I love coding in lisp, so I'd love to contribute as well 21:01:56 the thing is, I don't have any experience writing compilers 21:02:36 I learn very quickly, but I am wondering exactly how much knowledge of compilers is required for most of the projects on the page? 21:02:37 Not all of these ideas are compiler ideas, some of them are runtime things, debugging tools, and so on. 21:03:14 I've coded in C, but not assembly. what level of experience do you think would be required for success with these projects? 21:03:28 It would vary on a per-project basis. 21:03:50 Most of the projects on that page have a "prerequisites" section and a "student learning" section. 21:05:10 yes... I'm just afraid there is a level of background knowledge required that might not be mentioned in the prerequisites... 21:05:31 you're right though 21:05:41 it is per-project 21:06:52 for instance, here is one project I'm interested in http://www.sbcl.org/gsoc2014/ideas/#sec-1.2 21:07:30 The GMP thing? 21:07:33 yes 21:07:50 integrating GMP with sbcl 21:09:05 one question is, why would we want to integrate GMP? would there be a significant speed-up in big number calculations, for instance, or more space efficiency? 21:10:31 Ah, there's a new list. Hadn't realized that. 21:13:11 SB-GMP is pretty much done. 21:14:42 xani [~user@178.183.147.143.dsl.dynamic.t-mobile.pl] has joined #sbcl 21:14:53 which projects under section 1 would be of most use to the community? 21:14:57 for a student to work on 21:16:09 If you want to work with section 1, I'd think that 1.6 or 1.7 would be fairly useful. 21:16:33 ah nice, 1.6 is the other one I was interested in 21:17:04 Okay, what on earth is 6.21 "scheduling pass"? It's massively under-described, and doesn't the assembler already HAVE a scheduling option? 21:18:56 unfortunately, the prereq "Basic understanding of shared-memory concurrency." 21:19:08 I don't have that 21:19:29 on the other hand, it does say "basic" understanding, so it's possible for me to learn it given the correct resources and a litle time 21:22:38 are there any gsoc mentors in here? 21:24:00 Another possible angle is to look through the launchpad bug tracker to see if there are any interesting bugs/feature requests there. 21:24:30 are you in #lisp? I had an idea of working on a browser-based repl for sbcl, and some people there are telling me I can do that for gsoc 21:24:54 pnpuff [~l@unaffiliated/pnpuff] has joined #sbcl 21:24:58 I am in #lisp, but I'm mostly ignoring that conversation. I find the idea of a sandboxed REPL to be somewhat boring. 21:25:14 But apparently that might just be me. (-: 21:25:23 you could speak up with that opinion 21:25:56 -!- pnpuff [~l@unaffiliated/pnpuff] has left #sbcl 21:26:13 I could. Or I could get back to actual paying work, or continue trying to figure out a specific bug in the compiler, or... 21:26:40 just saying 21:35:49 Hrm. Some of these projects, especially some of the section six projects, are on my list of "possible weekend hacks". 21:36:29 And some of them are on my list of "massive nest of related projects that will take a lot of work, but would be worth doing anyway". 21:37:09 are there any that would be approachable by a college freshman who hasn't done this kind of thing before? 21:37:56 Honestly? I have no idea. 21:38:27 Krystof might know, I think he's in touch with the whole "college" thing. 21:39:26 well, sort of 21:39:32 I am mostly aware of how little my students know 21:39:41 sometimes I am aware how little I know too, though 21:40:47 there is no need to belittle me. we all start from somewhere. 21:40:59 -!- xani [~user@178.183.147.143.dsl.dynamic.t-mobile.pl] has quit [Ping timeout: 240 seconds] 21:41:27 these projects may be your "possible weekend hacks", nyef, so they are listed here for us larval engineers as a learning experienc. 21:43:30 xani [~user@178.183.146.1.dsl.dynamic.t-mobile.pl] has joined #sbcl 21:43:53 I'd have no objection to a student doing something that nyef could do in a weekend 21:44:17 That's fair. 6.1 "Free-er form displaced arrays" has been something I've thought about off-and-on for a while. 6.15 "Core relocation" has a branch from several years ago that almost worked, and the reason that it didn't has since been fixed, so that shouldn't be too hard. 6.18 "Allocation pools" involves more than a little GC hacking, but I've done some of the preliminary thinking for it already... On the other hand, some of the 21:44:17 other projects scare me. 21:44:25 I do think that you would be best served finding out quickly, now, whether you are actually "compatible" with SBCL 21:44:35 have you read Paul Khuong's blog post from last year? 21:44:37 Krystof, what do you mean? 21:44:43 no, I have not. I'll look it up now 21:44:43 http://www.pvk.ca/Blog/2013/04/13/starting-to-hack-on-sbcl/ 21:44:46 thank you 21:45:19 ah nice, this is an excellent starting point. 21:46:48 Krystof: Would you happen to know what "6.21 Scheduling pass" is supposed to be? 21:47:15 no 21:48:12 -!- sdemarre [~serge@112.84-64-87.adsl-dyn.isp.belgacom.be] has quit [Ping timeout: 264 seconds] 21:48:26 Damn. That's one of four that look like they might be interesting and are insufficiently described for me to be able to figure out what they might be. 21:49:02 (The others being 6.2 Autodxification of higher-order functions, 6.6 MPFR for long floats, and 6.23 Parametric recursive types." 21:49:20 ... And I seem to be having trouble with balancing parens against quotes this afternoon. 21:49:34 what do you guys think of the random number generator idea? 21:49:51 1.3 21:51:26 Hrm. That one looks like it could be neat, as it would have fairly minimal impact on the main system, with most of the code ending up in a contrib, and possibly establishing a common interface so that other implementations could also use the same kind of selectability. 21:52:35 I'm also thinking that 2.1 seems within my skill level 21:52:48 I'd expect that anyone who needs an actual cryptographically strong PRNG would wind up implementing their own for portability reasons, but having one or more already available would be nice. 21:54:27 2.1 could be interesting as well. Hacker's Delight might have some useful information in that direction, or the PPC Compiler Writers Guide... 21:54:48 hmm, I guess I'm looking mainly at 1.3 and 2.1. The integers thing actually sounds very interestng to me. I'll download the source now and play around a bit. 21:54:53 one last question though 21:55:18 is it a problem if I am using windows? should I attempt to acquire a linux computer for this, or is it possible to do some development on windows 8? 21:56:47 If you're going to be using windows, another project idea for you is figuring out how to add line-ending support to our external-format system. 21:57:09 If windows turns out to be too much of a pain, I'd recommend installing virtualbox and using a Linux VM. 21:57:50 what about porting more recent versions of sbcl to windows? 21:58:36 for instance, it appears we are up to 1.1.16, and the highest windows download is 1.1.12 21:58:54 I know nothing about the kind of work it would take to do this, of course. 21:59:11 That shouldn't be a porting effort, that should be a compiling effort. 21:59:32 If it doesn't work, then you have a bug-hunt on your hands. 22:00:28 For some of these download versions it's not that a newer SBCL isn't expected to build out-of-the-box, it's that nobody has built and packaged a download version yet. 22:01:08 *nyef* glares at the Linux/MIPS entries in the download table. 22:01:27 would it be a worthwhile effort to build and package a download version for windows, for the newer sbcl? 22:02:08 probably not as a gsoc project, but just something nice to do 22:02:11 as a contribution 22:02:13 on my own time. 22:03:13 Well, your first test, and something that you probably need to do anyway, is to build SBCL HEAD anyway, even if you don't do the whole distribution packaging thing. 22:03:20 If you can 22:04:06 If you can't build SBCL HEAD then you're going to have problems doing any development on SBCL itself. 22:05:09 Krystof: What's our current policy on builds-for-download, btw? 22:06:54 ad-hoc 22:08:55 So, if I were to set up a build-farm that did x86/Linux, PPC/Linux, MIPSle/Linux, x86/FreeBSD, x86-64/FreeBSD, and PPC/Darwin (yeah, right) builds, we might keep an awful lot of the platform table more up-to-date? 22:10:15 that sounds like a hilarious amount of fun! 22:10:45 if those could do two or three builds a month that would definitely be helpful 22:12:38 -!- Bike [~Glossina@gannon-wless-gw.resnet.wsu.edu] has quit [Ping timeout: 240 seconds] 22:14:35 Looks like four boxes, three if PPC/Darwin is a no-go or there's a usable option for running both PPC builds on the same hardware. 22:25:19 I still don't understand how we have those two MIPS builds in the table. I wasn't ever able to build 1.0.28 MIPSle without patching the source. 22:26:59 I could blame signal handling, or maybe compiler-dependent bugs in our runtime, or any of a number of things really :) 22:28:12 For MIPSle? My position is that the MIPS port was broken from day one. 22:28:22 Reliably, repeatably, broken. 22:28:34 I built it 22:28:42 When? 22:28:48 ages and ages ago 22:30:21 Must have been before June 16, 2005 or after June 9, 2012, from what I've been able to figure out. 22:32:41 Oh, and the "porting" page on sbcl.org might be just a touch out-of-date. (-: 22:33:16 I suck at release management and project leadership 22:34:11 can't even have an up-to-date website 22:34:48 maybe I should crowdfund one 22:34:57 Heh. 22:35:38 My personal website disappeared a while back, and I didn't notice for a while... and I still haven't set it back up. 22:48:44 I have hosts available for openbsd builds on x86, x86-64 and powerpc fwiw 22:49:15 I used to even do nightly builds but they seem to have stopped at some point 22:49:25 Do we really want to keep pretending to support PPC/Darwin? 22:49:37 however I am terrible at building and uploading releases 22:50:26 -!- cheryllium [~chatzilla@128.237.213.55] has quit [Read error: Connection reset by peer] 22:50:50 cheryllium [~chatzilla@128.237.213.55] has joined #sbcl 22:53:31 BE RISC provides a nice target for finding bugs :P 22:54:32 Bike [~Glossina@gannon-wless-gw.resnet.wsu.edu] has joined #sbcl 22:57:37 -!- oleo [~oleo@xdsl-87-79-199-91.netcologne.de] has quit [Ping timeout: 240 seconds] 22:58:34 oleo [~oleo@xdsl-84-44-179-167.netcologne.de] has joined #sbcl 22:59:29 -!- drmeister [~drmeister@155.247.96.196] has quit [Remote host closed the connection] 23:00:04 drmeister [~drmeister@155.247.96.196] has joined #sbcl 23:04:22 god, building this on windows does not look fun. 23:04:32 -!- drmeister [~drmeister@155.247.96.196] has quit [Ping timeout: 244 seconds] 23:04:33 So set up a Linux VM instead then. 23:04:39 I just ordered a new laptop. 23:05:33 I realized at some point that it would be nice to have a linux machine (exclusively linux). 23:06:49 once I download it, is it possible to run the test suite right away? 23:08:16 You need to do a build before the test suite will work. 23:09:03 to build from the github source, wolud I follow the instructions here? http://sbcl.sourceforge.net/getting.html 23:10:41 That might work. I usually keep an older binary from sourceforge around and have a script that invokes make.sh with appropriate parameters to select that build host. 23:22:18 -!- fikusz [~fikusz@catv-89-132-137-62.catv.broadband.hu] has quit [Ping timeout: 240 seconds] 23:33:46 nyef: for 2.1, there's this blog post I've been meaning to write for ~3 years to generalise and unify all div-by-mul strength reductions I've seen. 23:34:58 re scheduling pass, it's kind of brain dead (iirc) and doesn't work at all on x86oids. 23:35:18 I believe I mostly put it up because some one had expressed interest in that kind of work. 23:35:51 -!- xani [~user@178.183.146.1.dsl.dynamic.t-mobile.pl] has quit [Remote host closed the connection] 23:38:59 nyef: autodx is just making sure that if a function is only passed to standard HOF (e.g., map or sort), they're declared DX. 23:39:51 There are many ways to achieve that, some simple (with source transforms, even) others not so much. the problem with the simple approaches is that they hinder other optimisations when the map function actually gets inlined. 23:40:01 (otoh, not sure that inlining as much as we do is a good idea) 23:48:13 some sort of LAP to generate type adapters instead of boxing everything when calling functions would be awesome, also ;) 23:48:18 (now, free booze) 23:50:08 drmeister [~drmeister@166.170.22.11] has joined #sbcl 23:56:53 Okay, yeah, the scheduler is entirely brain-dead, and isn't configured for x86oids. I was looking into it recently in the context of our problems with SPARC (operating under an assumption of bogus lifetime information, which turned out not to be the issue), so I know a couple of things about it. 23:58:11 What might be interesting is to represent assembly code as another IR that we can then do optimizations on. 00:06:16 -!- segv- [~mb@95-91-241-251-dynip.superkabel.de] has quit [Ping timeout: 265 seconds] 00:10:21 so I've git cloned the repo on a linux vm, and run sh make.sh in the directory 00:10:30 I think I am running into an error though 00:11:10 Because I get "make-host-1.sh: 34: make-host-1.sh: sbcl: not found" 00:11:18 I am supposed to have an sbcl binary installed first? 00:12:38 Yes. Also look at the --xc-host= parameter to make.sh. 00:13:18 You might try the -h parameter to make.sh to start with. 00:13:29 ah I see, I need to compile it with another lisp implementation? 00:13:35 (I just installed sbcl, and it is compiling now.) 00:14:24 the page says that it uses another common lisp to compile SBCL, how does that work? is the compiler for this written in lisp? how does it "use" another lisp? 00:15:13 Yes, the compiler is written in Lisp. 00:16:19 oh no, another error. I think Im missing another dependency. it says "zlib.h: No such file or directory" 00:16:24 The build process starts by building Python (the compiler) as a cross-compiler, then uses that cross-compiler to build the entire system. 00:16:52 hm, I didn't know that Python was still used to compile lisps today. 00:16:54 Install the zlib development package, or build --without-sb-core-compression 00:16:57 Python the compiler 00:17:16 "Python" is the name of the compiler in CMUCL and SBCL. 00:18:54 -!- drmeister [~drmeister@166.170.22.11] has quit [Read error: Connection reset by peer] 00:24:44 drmeister [~drmeister@166.170.22.11] has joined #sbcl 00:26:16 yes I know 00:26:29 man this is compiling for a while now... 00:26:29 Sorry, I couldn't find anything for this is compiling for a while now.... 00:26:38 LOL 00:26:47 accidentally called specbot. 00:27:20 Takes about an hour for me on a PPC, twenty minutes to half an hour on my ThinkPad x61... And 5-10 minutes on a fairly modern machine. 00:29:02 nice, it was like 10 minutes for me. I'm running the tests now. this is pretty exciting. 00:29:35 The tests probably take longer. (-: 00:29:48 Should be 4-5 expected failures IIRC. 00:31:07 -!- drmeister [~drmeister@166.170.22.11] has quit [Read error: Connection reset by peer] 00:34:50 -!- cheryllium [~chatzilla@128.237.213.55] has quit [Read error: Connection reset by peer] 00:35:57 cheryllium [~chatzilla@128.237.213.55] has joined #sbcl 00:36:03 sorry my computer crashed... :/ 00:36:29 I dunno why though (it's windows). probably having music playing, 400 tabs open on firefox, irc and a vm running killed my ram... 00:38:52 Or some rather CPU-intensive test cases in the test suite kicked your CPU into thermal overload. 02:11:49 ccl-logbot [~ccl-logbo@setf.clozure.com] has joined #sbcl 02:11:49 02:11:49 -!- names: ccl-logbot fridim_ DalekBaldwin foom oleo psilord drmeister @Krystof xymox scymtym_ |3b| fikusz ubii jdz davazp` nyef Bike christoph_deb prxq Posterdati ivan`` luis nicdev scymtym fe[nl]ix Blkt_ kludge` jsnell daimrod pkhuong mood faheem minion hzp joshe ferada Munksgaard brucem jaimef flip214 pchrist specbot yauz antoszka kanru asedeno izirku milosn redline6561 jrm bege ams White_Flame reb 02:12:01 -!- kludge` [~comet@unaffiliated/espiral] has quit [Ping timeout: 240 seconds] 02:13:15 kludge` [~comet@unaffiliated/espiral] has joined #sbcl 02:20:45 DalekBal` [~user@ip68-231-209-84.oc.oc.cox.net] has joined #sbcl 02:20:49 -!- DalekBal` [~user@ip68-231-209-84.oc.oc.cox.net] has left #sbcl 02:39:34 ccl-logbot [~ccl-logbo@setf.clozure.com] has joined #sbcl 02:39:34 02:39:34 -!- names: ccl-logbot foom drmeister davazp` reb White_Flame ams bege 02:39:34 -dickson.freenode.net:#sbcl- [freenode-info] help freenode weed out clonebots -- please register your IRC nick and auto-identify: http://freenode.net/faq.shtml#nicksetup 02:56:20 prxq [~mommer@x2f68f92.dyn.telefonica.de] has joined #sbcl 03:02:34 nyef [~nyef@pool-70-109-132-189.cncdnh.east.myfairpoint.net] has joined #sbcl 03:02:34 ubii [~ubii@198.45.198.1] has joined #sbcl 03:02:34 attila_lendvai [~attila_le@5.251.216.208] has joined #sbcl 03:05:50 Blkt [~Blkt@2a01:4f8:150:80a1::aaaa] has joined #sbcl 03:06:06 -!- nyef [~nyef@pool-70-109-132-189.cncdnh.east.myfairpoint.net] has quit [*.net *.split] 03:06:07 -!- Blkt [~Blkt@2a01:4f8:150:80a1::aaaa] has quit [*.net *.split] 03:06:07 -!- attila_lendvai [~attila_le@5.251.216.208] has quit [*.net *.split] 03:06:07 -!- ubii [~ubii@198.45.198.1] has quit [*.net *.split] 03:07:07 kludge` [~comet@arthur.silvertree.org] has joined #sbcl 03:11:58 Blkt [~Blkt@2a01:4f8:150:80a1::aaaa] has joined #sbcl 03:11:58 nyef [~nyef@pool-70-109-132-189.cncdnh.east.myfairpoint.net] has joined #sbcl 03:11:58 ubii [~ubii@198.45.198.1] has joined #sbcl 03:11:58 attila_lendvai [~attila_le@5.251.216.208] has joined #sbcl 03:16:47 -!- davazp` [~user@177.Red-83-63-237.staticIP.rima-tde.net] has quit [Remote host closed the connection] 03:30:19 Bike [~Glossina@gannon-wless-gw.resnet.wsu.edu] has joined #sbcl 03:30:40 fikusz [~fikusz@catv-89-132-137-62.catv.broadband.hu] has joined #sbcl 03:31:36 jrm [~jrm@freebsd/hacker/jrm] has joined #sbcl 03:50:07 -!- nyef [~nyef@pool-70-109-132-189.cncdnh.east.myfairpoint.net] has quit [Quit: G'night all.] 03:52:38 christoph_deb [~christoph@ppp-46-244-238-183.dynamic.mnet-online.de] has joined #sbcl 04:08:08 yacks [~py@103.6.159.103] has joined #sbcl 04:13:32 pranavrc [~pranavrc@122.164.212.234] has joined #sbcl 04:35:30 jiacobucci [~jiacobucc@pool-173-73-8-83.washdc.fios.verizon.net] has joined #sbcl 04:43:41 -!- jrm [~jrm@freebsd/hacker/jrm] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 05:26:20 -!- ubii [~ubii@198.45.198.1] has quit [Read error: Connection reset by peer] 05:26:44 ubii [~ubii@198.45.198.1] has joined #sbcl 05:49:11 ASau [~user@46.115.48.210] has joined #sbcl 06:05:41 -!- ASau [~user@46.115.48.210] has quit [Quit: off] 07:46:46 -!- pranavrc [~pranavrc@122.164.212.234] has quit [Remote host closed the connection] 07:53:29 angavrilov [~angavrilo@217.71.227.190] has joined #sbcl 07:53:29 slyrus [~chatzilla@107.201.5.56] has joined #sbcl 07:53:29 edgar-rft [~GOD@HSI-KBW-109-193-013-113.hsi7.kabel-badenwuerttemberg.de] has joined #sbcl 07:53:29 milosn [~milosn@94.12.79.143] has joined #sbcl 07:53:29 lacedaemon [~quassel@pdpc/supporter/professional/fenlix] has joined #sbcl 07:53:29 psilord [~psilord@c-69-180-173-249.hsd1.mn.comcast.net] has joined #sbcl 07:53:29 jdz [~jdz@212.36.34.246] has joined #sbcl 07:53:29 Krystof [~user@81.174.155.115] has joined #sbcl 07:53:29 xymox [lechuck@unaffiliated/contempt] has joined #sbcl 07:53:29 Posterdati [~kvirc@host208-231-dynamic.2-79-r.retail.telecomitalia.it] has joined #sbcl 07:53:29 ivan`` [~ivan@unaffiliated/ivan/x-000001] has joined #sbcl 07:53:29 luis [~luis@kerno.org] has joined #sbcl 07:53:29 nicdev [~user@kilimanjaro.rafpepa.com] has joined #sbcl 07:53:29 scymtym [~user@2001:638:504:2093:baca:3aff:fe83:e736] has joined #sbcl 07:53:29 redline6561 [~redline65@li69-162.members.linode.com] has joined #sbcl 07:53:29 izirku [~IceChat9@sys-its-g56vnw1-yev.cc.unt.edu] has joined #sbcl 07:53:29 asedeno [~asedeno@66.102.14.25] has joined #sbcl 07:53:29 kanru [~kanru@118-163-10-190.HINET-IP.hinet.net] has joined #sbcl 07:53:29 antoszka [~antoszka@unaffiliated/antoszka] has joined #sbcl 07:53:29 yauz [4fca40aabe@osgiliath.yauz.de] has joined #sbcl 07:53:29 specbot [~specbot@common-lisp.net] has joined #sbcl 07:53:29 pchrist [spirit@gentoo/developer/pchrist] has joined #sbcl 07:53:29 flip214 [~marek@unaffiliated/flip214] has joined #sbcl 07:53:29 jaimef [jaimef@166.84.6.60] has joined #sbcl 07:53:29 brucem [~bmitchene@waywardmonkeys.com] has joined #sbcl 07:53:29 Munksgaard [munksgaard@hub.pronoia.dk] has joined #sbcl 07:53:29 ferada [~ferada@37.221.196.86] has joined #sbcl 07:53:29 joshe [~joshe@onyx.elsasser.org] has joined #sbcl 07:53:29 hzp [~user@188-67-206-253.bb.dnainternet.fi] has joined #sbcl 07:53:29 minion [~minion@common-lisp.net] has joined #sbcl 07:53:29 faheem [~faheem@bulldog.duhs.duke.edu] has joined #sbcl 07:53:29 mood [~mood@146.185.164.46] has joined #sbcl 07:53:29 pkhuong [~pkhuong@23.233.218.119] has joined #sbcl 07:53:29 daimrod [daimrod@sbrk.org] has joined #sbcl 07:53:29 jsnell [~jsnell@ash.snellman.net] has joined #sbcl 07:53:29 -!- dickson.freenode.net has set mode +o Krystof 07:56:47 xani [~user@178.183.148.36.dsl.dynamic.t-mobile.pl] has joined #sbcl 07:56:47 -!- attila_lendvai is now known as Guest51106 08:07:15 -!- yacks [~py@103.6.159.103] has quit [Quit: Leaving] 08:17:46 pranavrc [~pranavrc@122.164.212.234] has joined #sbcl 08:17:47 -!- pranavrc [~pranavrc@122.164.212.234] has quit [Changing host] 08:17:47 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 08:19:00 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Read error: Connection reset by peer] 08:19:31 pranavrc [~pranavrc@122.164.212.234] has joined #sbcl 08:19:31 -!- pranavrc [~pranavrc@122.164.212.234] has quit [Changing host] 08:19:31 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 08:23:59 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Ping timeout: 244 seconds] 08:40:28 yacks [~py@122.179.102.149] has joined #sbcl 08:41:54 pranavrc [~pranavrc@122.164.212.234] has joined #sbcl 08:41:54 -!- pranavrc [~pranavrc@122.164.212.234] has quit [Changing host] 08:41:55 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 08:51:07 scymtym: it's the newcomer's equivalent to forgetting to increment version.lisp-expr :) 08:52:13 -!- prxq [~mommer@x2f68f92.dyn.telefonica.de] has quit [Quit: Leaving] 08:56:55 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [] 08:57:29 -!- Guest51106 is now known as attila_lendvai 08:58:07 -!- attila_lendvai is now known as Guest76115 08:58:07 -!- Guest76115 [~attila_le@5.251.216.208] has quit [Changing host] 08:58:07 Guest76115 [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 09:00:35 pranavrc [~pranavrc@122.164.212.234] has joined #sbcl 09:00:35 -!- pranavrc [~pranavrc@122.164.212.234] has quit [Changing host] 09:00:35 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 09:16:42 ltbarcly [~textual@pool-108-42-99-156.snfcca.fios.verizon.net] has joined #sbcl 10:32:24 -!- kludge` [~comet@arthur.silvertree.org] has quit [Ping timeout: 240 seconds] 10:37:05 kludge` [~comet@unaffiliated/espiral] has joined #sbcl 11:25:24 stassats [~stassats@wikipedia/stassats] has joined #sbcl 11:25:53 found a way to shave off 992KB for defstruct accessors: not declaring their type 11:26:18 the source transforms already the necessary THE and TRULY-THE 11:29:33 segv- [~mb@95-91-211-99-dynip.superkabel.de] has joined #sbcl 11:32:22 -!- xani [~user@178.183.148.36.dsl.dynamic.t-mobile.pl] has quit [Remote host closed the connection] 11:40:57 andreh [~andreh@189.27.4.212.dynamic.adsl.gvt.net.br] has joined #sbcl 11:42:39 -!- andreh [~andreh@189.27.4.212.dynamic.adsl.gvt.net.br] has quit [Client Quit] 11:43:04 davazp [~user@177.Red-83-63-237.staticIP.rima-tde.net] has joined #sbcl 11:45:17 madgoblin [~madgoblin@78.24.216.166] has joined #sbcl 11:52:21 -!- Guest76115 [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 12:09:37 -!- Bike [~Glossina@gannon-wless-gw.resnet.wsu.edu] has quit [Ping timeout: 265 seconds] 12:09:51 Bike [~Glossina@gannon-wless-gw.resnet.wsu.edu] has joined #sbcl 12:11:01 eudoxia [~eudoxia@r186-52-11-120.dialup.adsl.anteldata.net.uy] has joined #sbcl 12:20:45 |3b| [bbb@2600:3c00::f03c:91ff:fedf:5b65] has joined #sbcl 12:28:50 -!- stassats [~stassats@wikipedia/stassats] has quit [Read error: Operation timed out] 13:17:58 -!- ltbarcly [~textual@pool-108-42-99-156.snfcca.fios.verizon.net] has quit [Quit: Computer has gone to sleep.] 13:32:05 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [Remote host closed the connection] 13:32:40 drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has joined #sbcl 13:36:43 -!- yacks [~py@122.179.102.149] has quit [Quit: Leaving] 13:37:05 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [Ping timeout: 244 seconds] 13:40:57 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [] 13:45:38 -!- lacedaemon is now known as fe[nl]ix 13:47:49 drmeister [~drmeister@166.170.23.251] has joined #sbcl 13:49:56 oleo [~oleo@xdsl-78-35-182-157.netcologne.de] has joined #sbcl 13:54:26 -!- drmeister [~drmeister@166.170.23.251] has quit [Read error: Connection reset by peer] 14:06:06 yacks [~py@103.6.159.103] has joined #sbcl 14:09:59 stassats [~stassats@wikipedia/stassats] has joined #sbcl 14:20:43 -!- madgoblin [~madgoblin@78.24.216.166] has quit [Remote host closed the connection] 14:28:42 uh oh, a barrage of test failures on ppc 14:32:25 drmeister [~drmeister@155.247.96.196] has joined #sbcl 14:36:13 prxq [~mommer@x2f68f92.dyn.telefonica.de] has joined #sbcl 15:03:18 -!- drmeister [~drmeister@155.247.96.196] has quit [Remote host closed the connection] 15:06:36 xani [~user@178.183.148.36.dsl.dynamic.t-mobile.pl] has joined #sbcl 15:16:28 found a too optimistic defstruct constructor parsing => (defstruct (c (:constructor make-c (&aux (a b c . whatever))))) => no error 15:17:21 -!- eudoxia [~eudoxia@r186-52-11-120.dialup.adsl.anteldata.net.uy] has quit [Quit: Leaving] 15:17:38 -!- yacks [~py@103.6.159.103] has quit [Ping timeout: 265 seconds] 15:17:43 I'm reminded of cliini's comment about &aux in lisp500 15:17:59 bug: "&aux not actually implemented". resolution: "cunning plan to flush out the heretics" 15:22:13 (&aux 123) is accepted also 15:23:10 (defstruct (c (:constructor make-c (&aux nil))) nil) => fine, (defstruct (c (:constructor make-c (&aux (nil 10)))) nil) => NIL names a defined constant, and cannot be used as a local 15:24:22 yacks [~py@103.6.159.103] has joined #sbcl 15:26:14 drmeister [~drmeister@155.247.96.196] has joined #sbcl 15:27:58 fridim__ [~fridim@65.93.77.146] has joined #sbcl 15:53:12 stassats` [~stassats@wikipedia/stassats] has joined #sbcl 15:54:51 -!- kludge` [~comet@unaffiliated/espiral] has quit [Disconnected by services] 15:55:09 kludge` [~comet@unaffiliated/espiral] has joined #sbcl 15:58:46 after fixing all the problems caused by it, i have 896KB of savings 16:00:36 so, 2MB of additional space doesn't sound as bad, though those 896KB could have been saved with the previous approach too, so it's still 3MB technically 16:02:06 -!- minion [~minion@common-lisp.net] has quit [Disconnected by services] 16:02:08 minion [~minion@common-lisp.net] has joined #sbcl 16:02:32 -!- specbot [~specbot@common-lisp.net] has quit [Ping timeout: 264 seconds] 16:04:38 specbot [~specbot@common-lisp.net] has joined #sbcl 16:08:21 jiacobucci1 [~jiacobucc@pool-173-73-8-83.washdc.fios.verizon.net] has joined #sbcl 16:08:32 -!- jiacobucci [~jiacobucc@pool-173-73-8-83.washdc.fios.verizon.net] has quit [Excess Flood] 16:09:07 -!- stassats [~stassats@wikipedia/stassats] has quit [Remote host closed the connection] 16:09:20 -!- drmeister [~drmeister@155.247.96.196] has quit [Remote host closed the connection] 16:09:40 drmeister [~drmeister@155.247.96.196] has joined #sbcl 16:10:59 -!- slyrus [~chatzilla@107.201.5.56] has quit [Ping timeout: 252 seconds] 16:18:56 -!- davazp [~user@177.Red-83-63-237.staticIP.rima-tde.net] has quit [Remote host closed the connection] 16:20:28 -!- yacks [~py@103.6.159.103] has quit [Ping timeout: 265 seconds] 16:24:20 davazp [~user@177.Red-83-63-237.staticIP.rima-tde.net] has joined #sbcl 16:27:38 -!- drmeister [~drmeister@155.247.96.196] has quit [Ping timeout: 240 seconds] 16:39:17 fiveop [~fiveop@p5DDC79F5.dip0.t-ipconnect.de] has joined #sbcl 16:55:02 senj [~senj@unaffiliated/senj] has joined #sbcl 16:57:11 slyrus [~chatzilla@107.201.5.56] has joined #sbcl 17:33:44 ccl-logbot [~ccl-logbo@setf.clozure.com] has joined #sbcl 17:33:44 17:33:44 -!- names: ccl-logbot sdemarre slyrus crixus eudoxia_ DGASAU` Bike scymtym oleo segv- round-robin kanru zhengyangl kludge` edgar-rft tmh_ xymox christoph_debian prxq_ psilord ferada asedeno flip214 joshe Posterdati fikusz @Krystof pkhuong milosn foom luis minion specbot irsol jaimef pchrist daimrod Blkt |3b| ams White_Flame reb` scymtym_ antoszka Munksgaard fe[nl]ix jsnell_ brucem ivan`` nicdev redline6561 yauz hzp faheem mood 17:33:44 -leguin.freenode.net:#sbcl- [freenode-info] if you're at a conference and other people are having trouble connecting, please mention it to staff: http://freenode.net/faq.shtml#gettinghelp 17:40:03 attila_lendvai [~attila_le@147.30.12.25] has joined #sbcl 17:40:03 -!- attila_lendvai [~attila_le@147.30.12.25] has quit [Changing host] 17:40:03 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 17:52:23 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 18:14:24 Hydan [~0x6463@ip-89-103-110-16.net.upcbroadband.cz] has joined #sbcl 18:14:38 f2312 [~roland@209-195-77-29.cpe.distributel.net] has joined #sbcl 18:33:13 -!- round-robin [~bubo@91.224.149.58] has quit [Ping timeout: 240 seconds] 18:46:59 -!- f2312 [~roland@209-195-77-29.cpe.distributel.net] has quit [Quit: leaving] 18:47:59 -!- prxq_ is now known as prxq 2014-03-22T14:14:08Z ccl-logbot joined #sbcl 2014-03-22T14:14:08Z 2014-03-22T14:14:08Z names: ccl-logbot xymox flip214_ krzysz00 ivan`` luis faheem nicdev yauz hzp LiamH scymtym fiveop drmeister pranavrc kludge` jaimef Hydan angavrilov edgar-rft yacks christoph_debian oleo prxq crixxxus slyrus Guest50092 DGASAU` tmh_ ferada asedeno flip214 joshe Posterdati fikusz @Krystof pkhuong milosn foom minion specbot irsol pchrist daimrod Blkt |3b| ams White_Flame scymtym_ Munksgaard antoszka fe[nl]ix jsnell_ brucem mood redline6561 2014-03-22T14:15:52Z flip214 quit (Write error: Broken pipe) 2014-03-22T14:16:15Z asedeno quit (Remote host closed the connection) 2014-03-22T14:18:53Z asedeno joined #sbcl 2014-03-22T14:32:07Z krzysz00 quit (Quit: leaving) 2014-03-22T14:41:30Z LiamH quit (Quit: Leaving.) 2014-03-22T14:50:36Z Hydan quit (Ping timeout: 245 seconds) 2014-03-22T14:55:21Z attila_lendvai joined #sbcl 2014-03-22T14:55:21Z attila_lendvai quit (Changing host) 2014-03-22T14:55:21Z attila_lendvai joined #sbcl 2014-03-22T15:07:51Z ccl-logbot joined #sbcl 2014-03-22T15:07:51Z 2014-03-22T15:07:51Z names: ccl-logbot attila_lendvai asedeno xymox flip214_ ivan`` luis faheem yauz hzp nicdev scymtym fiveop drmeister pranavrc kludge` jaimef angavrilov edgar-rft yacks christoph_debian oleo prxq crixxxus slyrus Guest50092 DGASAU` tmh_ ferada joshe Posterdati fikusz @Krystof pkhuong milosn foom minion specbot irsol pchrist daimrod Blkt |3b| ams White_Flame scymtym_ antoszka Munksgaard fe[nl]ix jsnell_ brucem redline6561 mood 2014-03-22T15:11:48Z leuler joined #sbcl 2014-03-22T15:15:14Z Bike joined #sbcl 2014-03-22T15:30:58Z oleo quit (Ping timeout: 265 seconds) 2014-03-22T15:31:23Z oleo joined #sbcl 2014-03-22T15:39:42Z Bike quit (Quit: NINE HOUUUUUR) 2014-03-22T15:46:50Z fiveop quit 2014-03-22T16:57:30Z Bike joined #sbcl 2014-03-22T17:10:57Z Bike quit (Read error: Connection reset by peer) 2014-03-22T17:25:02Z slyrus quit (Ping timeout: 246 seconds) 2014-03-22T17:27:36Z Bike joined #sbcl 2014-03-22T17:29:59Z Bike quit (Read error: Connection reset by peer) 2014-03-22T17:32:56Z Bike joined #sbcl 2014-03-22T17:37:27Z Bike quit (Ping timeout: 252 seconds) 2014-03-22T17:38:43Z Bike joined #sbcl 2014-03-22T18:04:11Z pranavrc quit (Quit: Ping timeout: ∞) 2014-03-22T18:07:27Z Hydan joined #sbcl 2014-03-22T18:13:41Z Bike quit (Ping timeout: 246 seconds) 2014-03-22T18:14:14Z Bike joined #sbcl 2014-03-22T18:23:25Z Bike quit (Read error: Connection reset by peer) 2014-03-22T18:23:49Z Bike joined #sbcl 2014-03-22T18:28:58Z Bike quit (Ping timeout: 240 seconds) 2014-03-22T18:34:28Z Bike joined #sbcl 2014-03-22T18:47:52Z attila_lendvai quit (Quit: Leaving.) 2014-03-22T18:52:40Z Bike quit (Read error: Connection reset by peer) 2014-03-22T18:52:59Z eudoxia joined #sbcl 2014-03-22T18:56:59Z Bike joined #sbcl 2014-03-22T18:57:06Z eudoxia quit (Client Quit) 2014-03-22T19:00:14Z yacks quit (Ping timeout: 246 seconds) 2014-03-22T19:05:05Z Bike quit (Read error: Connection reset by peer) 2014-03-22T19:05:53Z slyrus joined #sbcl 2014-03-22T19:11:09Z Bike joined #sbcl 2014-03-22T19:36:01Z slyrus quit (Ping timeout: 240 seconds) 2014-03-22T19:44:32Z Bike quit (Remote host closed the connection) 2014-03-22T19:44:51Z Bike joined #sbcl 2014-03-22T19:50:01Z Bike quit (Ping timeout: 240 seconds) 2014-03-22T20:58:43Z oleo: morning 2014-03-22T21:37:37Z Bike joined #sbcl 2014-03-22T21:42:49Z Bike quit (Ping timeout: 240 seconds) 2014-03-22T21:43:15Z Bike joined #sbcl 2014-03-22T22:29:35Z edgar-rft quit (Quit: mental process destroyed because connection corrupted) 2014-03-22T22:30:35Z jaimef quit (Ping timeout: 246 seconds) 2014-03-22T22:40:03Z Bike quit (Read error: Connection reset by peer) 2014-03-22T22:40:25Z Bike joined #sbcl 2014-03-22T22:40:26Z jaimef joined #sbcl 2014-03-22T22:45:02Z Hydan quit (Remote host closed the connection) 2014-03-22T23:03:31Z LiamH joined #sbcl 2014-03-22T23:04:27Z Bike quit (Ping timeout: 268 seconds) 2014-03-22T23:05:59Z Bike joined #sbcl 2014-03-22T23:28:22Z leuler quit (Quit: ERC Version 5.1.2 $Revision: 1.796.2.6 $ (IRC client for Emacs)) 2014-03-22T23:34:18Z Bike quit (Ping timeout: 240 seconds) 2014-03-22T23:58:22Z LiamH quit (Quit: Leaving.) 2014-03-23T00:43:08Z LiamH joined #sbcl 2014-03-23T00:50:25Z nyef joined #sbcl 2014-03-23T00:50:33Z nyef: Hello all. 2014-03-23T00:50:48Z nyef: I have the SPARC build limping along again. 2014-03-23T01:07:53Z pkhuong: good evening nyef. Good news (: 2014-03-23T01:18:56Z tmh_ quit (Ping timeout: 245 seconds) 2014-03-23T01:26:04Z slyrus joined #sbcl 2014-03-23T01:38:28Z nyef: Test suite results aren't great, but they're about right for a still-somewhat-working-but-neglected backend. 2014-03-23T02:39:09Z prxq_ joined #sbcl 2014-03-23T02:42:18Z prxq quit (Ping timeout: 240 seconds) 2014-03-23T02:44:52Z kanru joined #sbcl 2014-03-23T02:51:34Z LiamH quit (Quit: Leaving.) 2014-03-23T02:53:25Z nyef: Okay, preliminary commits put together. Tomorrow I'll make sure that they still build on x86-64/linux and sparc/linux, then push. 2014-03-23T02:53:33Z nyef: Oh, and update NEWS. 2014-03-23T02:59:38Z nyef quit (Quit: G'night all.) 2014-03-23T03:08:48Z Bike joined #sbcl 2014-03-23T03:13:16Z tmh_ joined #sbcl 2014-03-23T03:13:59Z pkhuong: awesome, thanks... old ports are teh suck. 2014-03-23T03:16:52Z pranavrc joined #sbcl 2014-03-23T03:33:38Z psilord joined #sbcl 2014-03-23T03:39:22Z christoph_debian quit (Ping timeout: 265 seconds) 2014-03-23T03:53:12Z christoph_debian joined #sbcl 2014-03-23T03:59:53Z drmeister quit (Quit: Leaving...) 2014-03-23T04:18:18Z oleo quit (Ping timeout: 240 seconds) 2014-03-23T04:18:56Z oleo joined #sbcl 2014-03-23T04:57:13Z psilord quit (Quit: Leaving.) 2014-03-23T04:58:48Z yacks joined #sbcl 2014-03-23T05:57:16Z scymtym quit (Ping timeout: 245 seconds) 2014-03-23T07:20:54Z attila_lendvai joined #sbcl 2014-03-23T07:28:48Z _8hzp joined #sbcl 2014-03-23T07:32:50Z hzp quit (Ping timeout: 265 seconds) 2014-03-23T07:38:37Z edgar-rft joined #sbcl 2014-03-23T07:53:12Z Bike quit (Remote host closed the connection) 2014-03-23T07:53:49Z Bike joined #sbcl 2014-03-23T07:58:54Z prxq_ is now known as prxq 2014-03-23T08:19:15Z Krystof: thanks, nyef! 2014-03-23T08:19:31Z Krystof: expect freeze post those commits 2014-03-23T08:27:42Z attila_lendvai quit (Quit: Leaving.) 2014-03-23T09:26:01Z angavrilov quit (Ping timeout: 245 seconds) 2014-03-23T09:28:31Z pranavrc quit (Remote host closed the connection) 2014-03-23T10:45:37Z angavrilov joined #sbcl 2014-03-23T10:54:53Z oleo quit (Ping timeout: 265 seconds) 2014-03-23T10:54:57Z |3b| quit (Remote host closed the connection) 2014-03-23T10:55:00Z |3b| joined #sbcl 2014-03-23T10:55:49Z oleo joined #sbcl 2014-03-23T11:27:25Z luis` joined #sbcl 2014-03-23T11:29:50Z oleo quit (*.net *.split) 2014-03-23T11:29:50Z |3b| quit (*.net *.split) 2014-03-23T11:29:52Z flip214_ quit (*.net *.split) 2014-03-23T11:29:53Z luis quit (*.net *.split) 2014-03-23T11:33:12Z flip214 joined #sbcl 2014-03-23T11:33:12Z flip214 quit (Changing host) 2014-03-23T11:33:12Z flip214 joined #sbcl 2014-03-23T11:38:44Z oleo joined #sbcl 2014-03-23T11:45:36Z |3b|` joined #sbcl 2014-03-23T12:19:45Z |3b|` is now known as |3b| 2014-03-23T12:28:16Z fiveop joined #sbcl 2014-03-23T12:28:41Z LiamH joined #sbcl 2014-03-23T13:22:15Z |3b| quit (Remote host closed the connection) 2014-03-23T13:22:17Z |3b| joined #sbcl 2014-03-23T13:22:19Z |3b|` joined #sbcl 2014-03-23T13:22:19Z |3b| quit (Remote host closed the connection) 2014-03-23T13:23:39Z |3b|` quit (Remote host closed the connection) 2014-03-23T13:24:16Z |3b|` joined #sbcl 2014-03-23T13:27:54Z |3b|` is now known as |3b| 2014-03-23T13:34:39Z pranavrc joined #sbcl 2014-03-23T13:34:39Z pranavrc quit (Changing host) 2014-03-23T13:34:39Z pranavrc joined #sbcl 2014-03-23T13:50:11Z Kevin_Xi joined #sbcl 2014-03-23T13:50:19Z Kevin_Xi left #sbcl 2014-03-23T13:57:20Z leuler joined #sbcl 2014-03-23T14:00:11Z xymox quit (Ping timeout: 245 seconds) 2014-03-23T14:24:35Z |3b|` joined #sbcl 2014-03-23T14:28:55Z |3b| quit (Ping timeout: 265 seconds) 2014-03-23T15:14:21Z attila_lendvai joined #sbcl 2014-03-23T15:14:27Z attila_lendvai quit (Changing host) 2014-03-23T15:14:28Z attila_lendvai joined #sbcl 2014-03-23T16:31:45Z |3b|` is now known as |3b| 2014-03-23T17:06:42Z oleo quit (Ping timeout: 268 seconds) 2014-03-23T17:28:15Z pranavrc quit (Remote host closed the connection) 2014-03-23T17:34:15Z oleo joined #sbcl 2014-03-23T17:35:02Z fiveop quit 2014-03-23T17:59:18Z scymtym joined #sbcl 2014-03-23T18:06:16Z yacks quit (Quit: Leaving) 2014-03-23T19:37:38Z angavrilov quit (Remote host closed the connection) 2014-03-23T20:03:58Z crixxxus quit (Ping timeout: 240 seconds) 2014-03-23T20:07:50Z crixus joined #sbcl 2014-03-23T21:03:28Z Bike_ joined #sbcl 2014-03-23T21:03:39Z Bike_ is now known as Bicyclidine 2014-03-23T21:14:58Z danlentz joined #sbcl 2014-03-23T21:26:08Z attila_lendvai quit (Quit: Leaving.) 2014-03-23T21:30:59Z crixus quit (Ping timeout: 246 seconds) 2014-03-23T21:55:14Z crixus joined #sbcl 2014-03-23T21:56:20Z edgar-rft quit (Quit: lifeform experiment terminated because everything exploded) 2014-03-23T22:20:25Z slyrus quit (Ping timeout: 240 seconds) 2014-03-23T22:53:26Z leuler quit (Quit: ERC Version 5.1.2 $Revision: 1.796.2.6 $ (IRC client for Emacs)) 2014-03-23T23:17:44Z crixus quit (Ping timeout: 246 seconds) 2014-03-23T23:20:25Z jrm joined #sbcl 2014-03-23T23:30:19Z crixus joined #sbcl 2014-03-23T23:52:10Z crixxus joined #sbcl 2014-03-23T23:55:32Z crixus quit (Ping timeout: 246 seconds) 2014-03-24T00:06:16Z slyrus joined #sbcl 2014-03-24T00:18:38Z Bicyclidine quit (Quit: but now i must go) 2014-03-24T01:02:16Z crixxxus joined #sbcl 2014-03-24T01:05:18Z crixxus quit (Ping timeout: 240 seconds) 2014-03-24T01:08:39Z crixxus joined #sbcl 2014-03-24T01:12:25Z crixxxus quit (Ping timeout: 240 seconds) 2014-03-24T01:26:49Z kanru quit (Ping timeout: 268 seconds) 2014-03-24T01:38:57Z nyef joined #sbcl 2014-03-24T01:39:06Z nyef: Hello all. 2014-03-24T01:39:52Z nyef: I'm not likely to be here long tonight. 2014-03-24T01:39:56Z nyef: pkhuong: Ping? 2014-03-24T01:41:55Z pkhuong: nyef: pong 2014-03-24T01:42:23Z nyef: Two things. First is that I pushed the first of two commits required to fix the SPARC build. 2014-03-24T01:43:27Z nyef: Second is that I'm looking over your little stack VM post, and I have a question: Does the VM support having different net stack effects over parallel branches (such as the forth "IF DROP THEN' sequence)? 2014-03-24T01:43:49Z pkhuong: yes. The stack effect is dynamic, per instruction. 2014-03-24T01:44:28Z pkhuong: (I'm still waiting for my mail client to download the commit emails. I suspect t-mo is throttling me) 2014-03-24T01:45:22Z nyef: So... *code-base* (well, RSI) gets updated every time a PUSH or POP occurs? 2014-03-24T01:46:14Z pkhuong: no. push jumps to the copy of the next opcode that corresponds to the adjusted stack pointer 2014-03-24T01:46:18Z pkhuong: same for pop. 2014-03-24T01:46:43Z nyef: But with multiple code paths with differing numbers of PUSH or POP operations, wouldn't there be a problem? 2014-03-24T01:46:48Z pkhuong: PUSH and POP are specialised for each value of the stack pointer themselves, so this is all static. 2014-03-24T01:47:43Z pkhuong: PUSH-for-tos@1 jumps to [next opcode]-for-tos-@0, regardless of the next opcode 2014-03-24T01:48:32Z pkhuong: the VM opcode stream only encodes opcodes, not the TOS pointer. The latter is implicit, and is determined by which copy of the primop is executed at runtime. 2014-03-24T01:49:25Z nyef: Okay, I need to read through the code a bit more thoroughly. 2014-03-24T01:50:13Z pkhuong: and jumping to the right copy just entails adding a constant offset to the index of the opcode. 2014-03-24T01:52:06Z nyef: Okay, I think I start to understand this now. 2014-03-24T02:03:39Z nyef: Well, it's certainly clever, though I'm not sure I'm really comfortable with such a stack design for my own VMs. 2014-03-24T02:04:38Z pkhuong: the fixed size rotating stack, or the implementation? 2014-03-24T02:05:14Z nyef: The fixed size rotating stack. 2014-03-24T02:06:06Z pkhuong: I'm not convinced either, but willing to try and see where it leads me 2014-03-24T02:06:10Z nyef: On the other hand, it's an essentially arbitrary size... 2014-03-24T02:07:59Z nyef: I'm also massively unconvinced about direct-threaded code. 2014-03-24T02:08:38Z krzysz00 joined #sbcl 2014-03-24T02:09:04Z nyef: (Having a background in indirect-threaded techniques, some experience with a token-threaded system, and looking at the possibilities of a subroutine-threaded implementation.) 2014-03-24T02:11:06Z pkhuong: subroutine threading is clever, but I'm not sure that it's a good idea on current machines, at least with small forth-style opcodes 2014-03-24T02:12:18Z nyef: Given that you can inline the primitive operations? 2014-03-24T02:12:38Z pkhuong: right, if you're willing to write half of a native code compiler, things are different 2014-03-24T02:13:26Z nyef: Yeah, but how much effort IS that in Forth, really? 2014-03-24T02:13:46Z pkhuong: fwiw, I don't think the trick of encoding the data stack pointer in IP is married to direct threading; it was just easier to implement. 2014-03-24T02:14:33Z nyef: Yeah, I can see that. I was actually trying to figure a token-threaded version. 2014-03-24T02:30:16Z kanru joined #sbcl 2014-03-24T02:42:58Z prxq quit (Ping timeout: 240 seconds) 2014-03-24T02:55:44Z prxq joined #sbcl 2014-03-24T03:27:15Z nyef quit (Quit: G'night all.) 2014-03-24T03:38:18Z christoph_debian quit (Ping timeout: 240 seconds) 2014-03-24T03:39:36Z yacks joined #sbcl 2014-03-24T03:47:51Z LiamH quit (Quit: Leaving.) 2014-03-24T03:51:50Z christoph_debian joined #sbcl 2014-03-24T05:07:42Z pranavrc joined #sbcl 2014-03-24T05:07:42Z pranavrc quit (Changing host) 2014-03-24T05:07:42Z pranavrc joined #sbcl 2014-03-24T05:16:20Z crixxxus joined #sbcl 2014-03-24T05:17:27Z scymtym quit (Ping timeout: 268 seconds) 2014-03-24T05:20:21Z crixxus quit (Ping timeout: 252 seconds) 2014-03-24T05:55:37Z oleo quit (Read error: No route to host) 2014-03-24T05:56:02Z oleo joined #sbcl 2014-03-24T06:01:58Z crixxxus quit (Ping timeout: 240 seconds) 2014-03-24T06:04:22Z krzysz00 quit (Ping timeout: 252 seconds) 2014-03-24T06:05:28Z oleo quit (Quit: Leaving) 2014-03-24T06:56:48Z angavrilov joined #sbcl 2014-03-24T06:57:58Z DGASAU` quit (Ping timeout: 240 seconds) 2014-03-24T07:16:52Z danlentz_ joined #sbcl 2014-03-24T07:18:04Z _8hzp` joined #sbcl 2014-03-24T07:19:23Z jrm2 joined #sbcl 2014-03-24T07:19:33Z danlentz quit (Ping timeout: 268 seconds) 2014-03-24T07:19:33Z danlentz_ is now known as danlentz 2014-03-24T07:19:39Z _8hzp quit (Ping timeout: 240 seconds) 2014-03-24T07:20:10Z jrm quit (Ping timeout: 268 seconds) 2014-03-24T07:22:01Z yacks quit (Quit: Leaving) 2014-03-24T07:45:51Z yacks joined #sbcl 2014-03-24T08:07:44Z DGASAU joined #sbcl 2014-03-24T09:30:01Z kludge` quit (Ping timeout: 240 seconds) 2014-03-24T09:34:04Z kludge` joined #sbcl 2014-03-24T10:41:20Z jdz joined #sbcl 2014-03-24T11:01:06Z krzysz00 joined #sbcl 2014-03-24T11:37:32Z hlavaty joined #sbcl 2014-03-24T12:06:43Z c4h joined #sbcl 2014-03-24T12:57:38Z yacks quit (Quit: Leaving) 2014-03-24T12:59:15Z pranavrc quit 2014-03-24T13:00:20Z eudoxia joined #sbcl 2014-03-24T13:07:33Z eudoxia quit (Quit: leaving) 2014-03-24T13:07:40Z eudoxia joined #sbcl 2014-03-24T13:13:13Z segv- joined #sbcl 2014-03-24T13:30:48Z eudoxia quit (Quit: leaving) 2014-03-24T13:31:10Z eudoxia joined #sbcl 2014-03-24T13:32:53Z eudoxia quit (Client Quit) 2014-03-24T13:33:03Z eudoxia joined #sbcl 2014-03-24T13:43:14Z krzysz00 quit (Ping timeout: 252 seconds) 2014-03-24T13:49:01Z yacks joined #sbcl 2014-03-24T14:04:10Z jrm2 is now known as jrm 2014-03-24T14:04:11Z jrm quit (Changing host) 2014-03-24T14:04:11Z jrm joined #sbcl 2014-03-24T14:12:56Z krzysz00 joined #sbcl 2014-03-24T14:57:08Z krzysz00 quit (Ping timeout: 246 seconds) 2014-03-24T15:21:49Z crixus joined #sbcl 2014-03-24T15:23:16Z pranavrc joined #sbcl 2014-03-24T15:23:16Z pranavrc quit (Changing host) 2014-03-24T15:23:16Z pranavrc joined #sbcl 2014-03-24T15:28:17Z crixus quit (Ping timeout: 252 seconds) 2014-03-24T15:36:17Z c4h quit (Quit: quit) 2014-03-24T15:50:12Z crixus joined #sbcl 2014-03-24T15:53:57Z pjb` joined #sbcl 2014-03-24T15:56:42Z pjb` is now known as pjb 2014-03-24T16:03:05Z attila_lendvai joined #sbcl 2014-03-24T16:03:05Z attila_lendvai quit (Changing host) 2014-03-24T16:03:05Z attila_lendvai joined #sbcl 2014-03-24T16:13:07Z crixus quit (Ping timeout: 264 seconds) 2014-03-24T16:21:24Z crixus joined #sbcl 2014-03-24T16:29:37Z eudoxia quit (Ping timeout: 268 seconds) 2014-03-24T16:31:01Z eudoxia joined #sbcl 2014-03-24T16:43:43Z slyrus quit (Ping timeout: 264 seconds) 2014-03-24T16:50:34Z oleo joined #sbcl 2014-03-24T17:27:44Z crixus quit (Read error: Connection reset by peer) 2014-03-24T17:28:18Z ASau joined #sbcl 2014-03-24T17:51:37Z kanru quit (Ping timeout: 240 seconds) 2014-03-24T17:55:35Z eudoxia quit (Quit: leaving) 2014-03-24T18:08:41Z pranavrc quit (Quit: Ping timeout: ∞) 2014-03-24T18:41:09Z yacks quit (Quit: Leaving) 2014-03-24T18:41:38Z cmack joined #sbcl 2014-03-24T19:17:03Z slyrus joined #sbcl 2014-03-24T19:49:59Z angavrilov quit (Remote host closed the connection) 2014-03-24T19:53:48Z fiveop joined #sbcl 2014-03-24T21:05:29Z Bike_ joined #sbcl 2014-03-24T21:12:19Z Bike_ is now known as Bicyclidine 2014-03-24T21:27:22Z scymtym joined #sbcl 2014-03-24T22:02:32Z fiveop quit 2014-03-24T22:03:13Z slyrus quit (Ping timeout: 268 seconds) 2014-03-24T22:15:25Z ccl-logbot joined #sbcl 2014-03-24T22:15:25Z 2014-03-24T22:15:25Z names: ccl-logbot scymtym Bicyclidine cmack ASau oleo attila_lendvai pjb segv- hlavaty jdz kludge` DGASAU _8hzp` danlentz jrm christoph_debian prxq |3b| flip214 luis` Bike tmh_ jaimef mood redline6561 brucem jsnell_ fe[nl]ix antoszka Munksgaard scymtym_ White_Flame ams Blkt daimrod pchrist irsol specbot minion foom milosn pkhuong @Krystof fikusz Posterdati joshe ferada Guest50092 yauz nicdev faheem ivan`` asedeno 2014-03-24T22:28:07Z eudoxia joined #sbcl 2014-03-24T22:33:25Z ASau` joined #sbcl 2014-03-24T22:33:40Z ASau quit (Remote host closed the connection) 2014-03-24T22:34:53Z ASau` is now known as ASau 2014-03-24T22:35:05Z _8hzp` quit (Remote host closed the connection) 2014-03-24T22:54:31Z attila_lendvai quit (Ping timeout: 252 seconds) 2014-03-25T00:38:08Z jrm quit (Quit: ERC Version 5.3 (IRC client for Emacs)) 2014-03-25T00:59:14Z Bicyclidine quit (Quit: leaving) 2014-03-25T01:16:18Z pjb quit (Ping timeout: 240 seconds) 2014-03-25T01:47:32Z yacks joined #sbcl 2014-03-25T01:50:36Z psilord joined #sbcl 2014-03-25T01:53:21Z eudoxia quit (Quit: leaving) 2014-03-25T01:56:41Z segv- quit (Ping timeout: 265 seconds) 2014-03-25T02:11:00Z kanru joined #sbcl 2014-03-25T02:25:34Z slyrus joined #sbcl 2014-03-25T02:43:48Z prxq quit (Ping timeout: 268 seconds) 2014-03-25T02:56:10Z prxq joined #sbcl 2014-03-25T02:57:31Z oleo quit (Ping timeout: 264 seconds) 2014-03-25T02:58:27Z oleo joined #sbcl 2014-03-25T03:39:05Z christoph_debian quit (Ping timeout: 246 seconds) 2014-03-25T03:52:40Z christoph_debian joined #sbcl 2014-03-25T03:56:37Z segv- joined #sbcl 2014-03-25T04:28:49Z michael_lee joined #sbcl 2014-03-25T04:33:41Z scymtym quit (Ping timeout: 252 seconds) 2014-03-25T04:52:23Z attila_lendvai joined #sbcl 2014-03-25T05:15:35Z attila_lendvai quit (Quit: Leaving.) 2014-03-25T05:23:50Z pranavrc joined #sbcl 2014-03-25T05:58:55Z pranavrc quit 2014-03-25T06:00:27Z pranavrc joined #sbcl 2014-03-25T06:03:45Z oleo quit (Quit: Leaving) 2014-03-25T06:40:23Z angavrilov joined #sbcl 2014-03-25T07:38:54Z yacks quit (Ping timeout: 265 seconds) 2014-03-25T07:42:17Z attila_lendvai joined #sbcl 2014-03-25T07:47:16Z ASau quit (Ping timeout: 252 seconds) 2014-03-25T07:59:11Z segv- quit (Remote host closed the connection) 2014-03-25T08:03:17Z yacks joined #sbcl 2014-03-25T09:32:39Z kludge` quit (Ping timeout: 268 seconds) 2014-03-25T09:36:26Z kludge` joined #sbcl 2014-03-25T10:48:58Z hlavaty quit (Remote host closed the connection) 2014-03-25T11:20:13Z pjb joined #sbcl 2014-03-25T11:37:00Z eudoxia joined #sbcl 2014-03-25T11:47:02Z segv- joined #sbcl 2014-03-25T11:50:59Z ams quit (Changing host) 2014-03-25T11:50:59Z ams joined #sbcl 2014-03-25T12:18:59Z attila_lendvai quit (Quit: Leaving.) 2014-03-25T12:26:34Z attila_lendvai joined #sbcl 2014-03-25T12:32:57Z attila_lendvai1 joined #sbcl 2014-03-25T12:32:57Z attila_lendvai quit (Disconnected by services) 2014-03-25T12:32:57Z attila_lendvai1 quit (Changing host) 2014-03-25T12:32:57Z attila_lendvai1 joined #sbcl 2014-03-25T12:39:20Z yacks quit (Ping timeout: 252 seconds) 2014-03-25T12:51:15Z DGASAU` joined #sbcl 2014-03-25T12:52:06Z jdz quit (Quit: Leaving) 2014-03-25T12:55:29Z DGASAU quit (Ping timeout: 265 seconds) 2014-03-25T12:58:23Z pranavrc quit 2014-03-25T13:00:20Z psilord quit (Quit: Leaving.) 2014-03-25T13:05:46Z hlavaty joined #sbcl 2014-03-25T13:25:12Z DGASAU` is now known as DGASAU 2014-03-25T13:25:53Z tmh_ quit (Changing host) 2014-03-25T13:25:53Z tmh_ joined #sbcl 2014-03-25T13:44:26Z yacks joined #sbcl 2014-03-25T14:01:03Z jdz joined #sbcl 2014-03-25T14:02:38Z edgar-rft joined #sbcl 2014-03-25T14:14:11Z psilord joined #sbcl 2014-03-25T14:21:52Z pranavrc joined #sbcl 2014-03-25T14:31:43Z drmeister joined #sbcl 2014-03-25T14:56:49Z oleo joined #sbcl 2014-03-25T15:41:45Z oleo quit (Ping timeout: 265 seconds) 2014-03-25T15:42:04Z oleo joined #sbcl 2014-03-25T15:49:42Z eudoxia quit (Quit: leaving) 2014-03-25T17:22:17Z jdz quit (Quit: Leaving) 2014-03-25T17:28:34Z drmeister quit (Remote host closed the connection) 2014-03-25T17:30:07Z drmeister joined #sbcl 2014-03-25T17:33:03Z drmeiste_ joined #sbcl 2014-03-25T17:33:52Z drmeister quit (Read error: No route to host) 2014-03-25T18:00:29Z drmeiste_ quit (Ping timeout: 265 seconds) 2014-03-25T18:04:05Z drmeister joined #sbcl 2014-03-25T18:20:55Z yacks quit (Ping timeout: 264 seconds) 2014-03-25T18:22:07Z michael_lee quit (Quit: Ex-Chat) 2014-03-25T18:22:18Z hlavaty quit (Ping timeout: 240 seconds) 2014-03-25T18:29:02Z pranavrc quit (Remote host closed the connection) 2014-03-25T18:34:47Z angavrilov quit (Remote host closed the connection) 2014-03-25T19:33:49Z slyrus quit (Ping timeout: 252 seconds) 2014-03-25T19:50:38Z attila_lendvai1 quit (Ping timeout: 240 seconds) 2014-03-25T19:55:08Z drmeister quit (Remote host closed the connection) 2014-03-25T20:06:36Z cmack quit (Remote host closed the connection) 2014-03-25T20:07:43Z irsol quit (Ping timeout: 264 seconds) 2014-03-25T20:09:01Z kludge` quit (Changing host) 2014-03-25T20:09:01Z kludge` joined #sbcl 2014-03-25T20:09:38Z irsol joined #sbcl 2014-03-25T20:16:18Z irsol quit (Ping timeout: 240 seconds) 2014-03-25T20:17:58Z irsol joined #sbcl 2014-03-25T20:18:15Z drmeister joined #sbcl 2014-03-25T20:18:45Z drmeister quit (Remote host closed the connection) 2014-03-25T20:19:12Z ASau joined #sbcl 2014-03-25T20:19:20Z drmeister joined #sbcl 2014-03-25T20:24:59Z irsol quit (Ping timeout: 246 seconds) 2014-03-25T20:27:13Z irsol joined #sbcl 2014-03-25T20:37:14Z irsol quit (Ping timeout: 246 seconds) 2014-03-25T20:39:18Z drmeister quit (Remote host closed the connection) 2014-03-25T20:39:44Z irsol joined #sbcl 2014-03-25T20:50:22Z irsol quit (Ping timeout: 268 seconds) 2014-03-25T20:52:22Z irsol joined #sbcl 2014-03-25T21:00:17Z irsol quit (Ping timeout: 265 seconds) 2014-03-25T21:04:18Z irsol joined #sbcl 2014-03-25T21:09:24Z drmeister joined #sbcl 2014-03-25T21:12:31Z irsol quit (Ping timeout: 264 seconds) 2014-03-25T21:14:25Z irsol joined #sbcl 2014-03-25T21:22:16Z drmeister quit (Remote host closed the connection) 2014-03-25T21:24:08Z irsol quit (Ping timeout: 246 seconds) 2014-03-25T21:26:13Z irsol joined #sbcl 2014-03-25T21:35:34Z irsol quit (Ping timeout: 265 seconds) 2014-03-25T21:36:46Z irsol joined #sbcl 2014-03-25T21:36:57Z scymtym joined #sbcl 2014-03-25T21:40:11Z drmeister joined #sbcl 2014-03-25T21:41:35Z nicdev` joined #sbcl 2014-03-25T21:44:23Z ams quit (Ping timeout: 252 seconds) 2014-03-25T21:45:37Z nicdev quit (Read error: Connection reset by peer) 2014-03-25T21:45:38Z Munksgaard quit (Ping timeout: 245 seconds) 2014-03-25T21:45:56Z nicdev` is now known as nicdev 2014-03-25T21:47:41Z fe[nl]ix quit (Ping timeout: 245 seconds) 2014-03-25T21:47:55Z Munksgaard joined #sbcl 2014-03-25T21:48:10Z fe[nl]ix joined #sbcl 2014-03-25T21:48:29Z irsol quit (Read error: Connection reset by peer) 2014-03-25T21:50:19Z irsol joined #sbcl 2014-03-25T21:50:40Z irsol quit (Read error: Connection reset by peer) 2014-03-25T21:51:37Z ltbarcly_ joined #sbcl 2014-03-25T21:52:57Z drmeister quit (Remote host closed the connection) 2014-03-25T22:01:16Z psilord quit (Quit: Leaving.) 2014-03-25T22:01:35Z irsol joined #sbcl 2014-03-25T22:02:27Z ltbarcly_ quit (Read error: Connection reset by peer) 2014-03-25T22:03:11Z ltbarcly_ joined #sbcl 2014-03-25T22:04:46Z Bike_ joined #sbcl 2014-03-25T22:15:24Z drmeister joined #sbcl 2014-03-25T22:23:10Z ltbarcly_ quit (Quit: Computer has gone to sleep.) 2014-03-25T22:29:07Z ltbarcly_ joined #sbcl 2014-03-25T22:35:11Z drmeister quit (Read error: Connection reset by peer) 2014-03-25T22:45:28Z ltbarcly_ quit (Quit: Computer has gone to sleep.) 2014-03-25T22:48:23Z Bike_ quit (Remote host closed the connection) 2014-03-25T22:48:37Z Bike_ joined #sbcl 2014-03-25T23:06:06Z irsol quit (Changing host) 2014-03-25T23:06:06Z irsol joined #sbcl 2014-03-25T23:11:11Z slyrus joined #sbcl 2014-03-25T23:15:44Z fourier joined #sbcl 2014-03-25T23:15:47Z segv- quit (Ping timeout: 246 seconds) 2014-03-25T23:42:00Z psilord joined #sbcl 2014-03-25T23:43:58Z fourier quit (Ping timeout: 240 seconds) 2014-03-25T23:55:31Z drmeister joined #sbcl 2014-03-26T00:05:41Z Bike_ quit (Quit: leaving) 2014-03-26T00:30:20Z slyrus quit (Ping timeout: 246 seconds) 2014-03-26T00:45:25Z drmeister quit (Remote host closed the connection) 2014-03-26T01:25:53Z drmeister joined #sbcl 2014-03-26T02:35:58Z prxq_ joined #sbcl 2014-03-26T02:39:29Z prxq quit (Ping timeout: 246 seconds) 2014-03-26T03:12:51Z yacks joined #sbcl 2014-03-26T03:36:48Z christoph_debian quit (Read error: Operation timed out) 2014-03-26T03:37:16Z edgar-rft quit (Quit: conversation interrupted because no future happened) 2014-03-26T03:44:56Z oleo quit (Ping timeout: 246 seconds) 2014-03-26T03:45:23Z oleo joined #sbcl 2014-03-26T03:52:23Z christoph_debian joined #sbcl 2014-03-26T04:14:20Z slyrus joined #sbcl 2014-03-26T04:32:52Z scymtym quit (Ping timeout: 268 seconds) 2014-03-26T04:49:42Z attila_lendvai joined #sbcl 2014-03-26T05:24:18Z tmh_ quit (Ping timeout: 240 seconds) 2014-03-26T05:24:57Z pranavrc joined #sbcl 2014-03-26T05:24:57Z pranavrc quit (Changing host) 2014-03-26T05:24:57Z pranavrc joined #sbcl 2014-03-26T05:31:30Z tmh_ joined #sbcl 2014-03-26T05:46:19Z michael_lee joined #sbcl 2014-03-26T05:58:38Z angavrilov joined #sbcl 2014-03-26T06:03:41Z oleo quit (Quit: Leaving) 2014-03-26T06:06:53Z c4h joined #sbcl 2014-03-26T07:22:59Z ASau quit (Ping timeout: 246 seconds) 2014-03-26T07:28:17Z yacks quit (Ping timeout: 240 seconds) 2014-03-26T07:55:23Z yacks joined #sbcl 2014-03-26T08:19:57Z c4h quit (Quit: quit) 2014-03-26T08:32:01Z c4h joined #sbcl 2014-03-26T08:33:01Z prxq_ is now known as prxq 2014-03-26T08:38:12Z DGASAU` joined #sbcl 2014-03-26T08:38:55Z DGASAU quit (Ping timeout: 268 seconds) 2014-03-26T08:43:54Z DGASAU` is now known as DGASAU 2014-03-26T08:44:37Z michael_lee quit (Ping timeout: 240 seconds) 2014-03-26T08:57:20Z ams joined #sbcl 2014-03-26T08:57:27Z ams quit (Changing host) 2014-03-26T08:57:28Z ams joined #sbcl 2014-03-26T08:58:47Z michael_lee joined #sbcl 2014-03-26T09:31:43Z kludge` quit (Ping timeout: 264 seconds) 2014-03-26T09:35:55Z kludge` joined #sbcl 2014-03-26T09:44:35Z DGASAU quit (Quit: reboot) 2014-03-26T09:54:15Z DGASAU joined #sbcl 2014-03-26T10:01:40Z DGASAU quit (Remote host closed the connection) 2014-03-26T10:05:39Z attila_lendvai quit (Quit: Leaving.) 2014-03-26T10:24:37Z DGASAU joined #sbcl 2014-03-26T10:27:13Z c4h quit (Quit: quit) 2014-03-26T10:29:09Z leoc joined #sbcl 2014-03-26T10:43:31Z jdz joined #sbcl 2014-03-26T11:04:52Z attila_lendvai joined #sbcl 2014-03-26T11:47:08Z hlavaty joined #sbcl 2014-03-26T12:36:57Z yacks quit (Ping timeout: 268 seconds) 2014-03-26T12:43:59Z attila_lendvai quit (Quit: Leaving.) 2014-03-26T13:04:42Z pranavrc quit (Ping timeout: 268 seconds) 2014-03-26T13:17:06Z eudoxia joined #sbcl 2014-03-26T13:18:18Z yacks joined #sbcl 2014-03-26T13:25:34Z segv- joined #sbcl 2014-03-26T13:46:28Z drmeister quit (Read error: Connection reset by peer) 2014-03-26T14:04:27Z xymox joined #sbcl 2014-03-26T14:11:14Z xymox quit (Read error: Connection reset by peer) 2014-03-26T14:11:52Z xymox joined #sbcl 2014-03-26T14:43:00Z oleo joined #sbcl 2014-03-26T14:58:41Z Krystof_ joined #sbcl 2014-03-26T14:59:37Z Krystof quit (Ping timeout: 240 seconds) 2014-03-26T15:06:30Z scymtym joined #sbcl 2014-03-26T15:13:59Z ASau joined #sbcl 2014-03-26T15:27:37Z pranavrc joined #sbcl 2014-03-26T15:27:37Z pranavrc quit (Changing host) 2014-03-26T15:27:37Z pranavrc joined #sbcl 2014-03-26T15:31:11Z jdz quit (Quit: Leaving) 2014-03-26T15:49:53Z jdz joined #sbcl 2014-03-26T16:16:51Z ASau quit (Ping timeout: 245 seconds) 2014-03-26T16:19:37Z oleo quit (Ping timeout: 240 seconds) 2014-03-26T16:32:52Z oleo joined #sbcl 2014-03-26T16:35:50Z Krystof_: http://google-opensource.blogspot.de/2014/03/steel-bank-common-lisp-wrap-up-post-new.html 2014-03-26T16:36:17Z Krystof_: (I didn't write the last sentence! I can feel the excitement from here) 2014-03-26T17:30:12Z ASau joined #sbcl 2014-03-26T17:30:46Z jaimef quit (Excess Flood) 2014-03-26T17:32:08Z michael_lee quit (Quit: Ex-Chat) 2014-03-26T17:34:16Z slyrus quit (Ping timeout: 265 seconds) 2014-03-26T17:36:08Z jaimef joined #sbcl 2014-03-26T17:52:12Z pranavrc quit (Quit: Ping timeout: ∞) 2014-03-26T18:38:23Z eudoxia quit (Quit: leaving) 2014-03-26T18:42:36Z slyrus joined #sbcl 2014-03-26T19:56:22Z yacks quit (Quit: Leaving) 2014-03-26T19:58:06Z slyrus quit (Ping timeout: 245 seconds) 2014-03-26T20:10:53Z Krystof_ quit (Read error: Operation timed out) 2014-03-26T20:32:06Z attila_lendvai joined #sbcl 2014-03-26T20:40:48Z Krystof_ joined #sbcl 2014-03-26T20:56:51Z Krystof_ is now known as Krystof 2014-03-26T20:56:57Z ChanServ has set mode +o Krystof 2014-03-26T21:04:46Z edgar-rft joined #sbcl 2014-03-26T22:04:50Z drmeister joined #sbcl 2014-03-26T22:14:02Z easy-iPad joined #sbcl 2014-03-26T22:25:09Z jrm joined #sbcl 2014-03-26T22:31:56Z easy-iPad quit (Quit: Outta here?) 2014-03-26T22:45:33Z ltbarcly_ joined #sbcl 2014-03-26T22:53:36Z ltbarcly_ quit (Quit: Computer has gone to sleep.) 2014-03-26T22:55:38Z slyrus joined #sbcl 2014-03-26T23:27:01Z drmeister quit (Remote host closed the connection) 2014-03-26T23:29:22Z eudoxia joined #sbcl 2014-03-26T23:32:19Z segv- quit (Ping timeout: 264 seconds) 2014-03-26T23:35:44Z attila_lendvai quit (Quit: Leaving.) 2014-03-26T23:44:02Z drmeister joined #sbcl 2014-03-27T00:00:18Z drmeister quit (Read error: Connection reset by peer) 2014-03-27T00:08:31Z slyrus quit (Ping timeout: 245 seconds) 2014-03-27T00:10:49Z drmeister joined #sbcl 2014-03-27T00:45:37Z eudoxia quit (Ping timeout: 240 seconds) 2014-03-27T00:57:20Z ltbarcly_ joined #sbcl 2014-03-27T01:12:54Z drmeister quit (*.net *.split) 2014-03-27T01:12:54Z ASau quit (*.net *.split) 2014-03-27T01:12:54Z oleo quit (*.net *.split) 2014-03-27T01:12:54Z ams quit (*.net *.split) 2014-03-27T01:12:55Z psilord quit (*.net *.split) 2014-03-27T01:12:55Z |3b| quit (*.net *.split) 2014-03-27T01:15:37Z edgar-rft quit (Ping timeout: 240 seconds) 2014-03-27T01:49:20Z oleo joined #sbcl 2014-03-27T01:49:20Z drmeister joined #sbcl 2014-03-27T01:49:20Z ASau joined #sbcl 2014-03-27T01:49:20Z ams joined #sbcl 2014-03-27T01:49:20Z psilord joined #sbcl 2014-03-27T01:49:20Z |3b| joined #sbcl 2014-03-27T01:50:52Z oleo is now known as Guest55547 2014-03-27T02:23:06Z michael_lee joined #sbcl 2014-03-27T02:36:37Z prxq_ joined #sbcl 2014-03-27T02:40:07Z prxq quit (Ping timeout: 264 seconds) 2014-03-27T02:52:06Z ltbarcl__ joined #sbcl 2014-03-27T02:53:23Z ltbarcly_ quit (Ping timeout: 252 seconds) 2014-03-27T03:16:26Z ltbarcl__ quit (Quit: Computer has gone to sleep.) 2014-03-27T03:23:29Z yacks joined #sbcl 2014-03-27T03:38:55Z christoph_debian quit (Ping timeout: 264 seconds) 2014-03-27T03:52:09Z christoph_debian joined #sbcl 2014-03-27T04:07:56Z pkhuong quit (Ping timeout: 246 seconds) 2014-03-27T04:08:02Z pkhuong joined #sbcl 2014-03-27T04:46:01Z luis` quit (Ping timeout: 240 seconds) 2014-03-27T04:46:27Z luis` joined #sbcl 2014-03-27T05:07:57Z oleo joined #sbcl 2014-03-27T05:10:17Z Guest55547 quit (Ping timeout: 265 seconds) 2014-03-27T05:15:31Z edgar-rft joined #sbcl 2014-03-27T05:20:33Z michael_lee quit (Remote host closed the connection) 2014-03-27T05:25:29Z pranavrc joined #sbcl 2014-03-27T05:34:38Z scymtym quit (Ping timeout: 240 seconds) 2014-03-27T05:42:26Z attila_lendvai joined #sbcl 2014-03-27T05:42:26Z attila_lendvai quit (Changing host) 2014-03-27T05:42:26Z attila_lendvai joined #sbcl 2014-03-27T06:07:10Z oleo quit (Quit: Leaving) 2014-03-27T06:20:22Z ASau quit (Ping timeout: 265 seconds) 2014-03-27T06:25:37Z jrm quit (Ping timeout: 240 seconds) 2014-03-27T06:57:06Z jrm joined #sbcl 2014-03-27T07:02:44Z prxq_ is now known as prxq 2014-03-27T07:42:22Z slyrus joined #sbcl 2014-03-27T07:50:17Z yacks quit (Remote host closed the connection) 2014-03-27T08:01:07Z joshe quit (Ping timeout: 264 seconds) 2014-03-27T09:30:59Z kludge` quit (Ping timeout: 246 seconds) 2014-03-27T09:36:30Z kludge` joined #sbcl 2014-03-27T11:09:20Z ebrasca joined #sbcl 2014-03-27T11:56:41Z eudoxia joined #sbcl 2014-03-27T12:32:56Z pranavrc quit 2014-03-27T12:41:06Z drmeister quit (Remote host closed the connection) 2014-03-27T13:27:29Z pranavrc joined #sbcl 2014-03-27T13:27:29Z pranavrc quit (Changing host) 2014-03-27T13:27:29Z pranavrc joined #sbcl 2014-03-27T13:33:01Z segv- joined #sbcl 2014-03-27T13:49:06Z scymtym_: 206b3fa8 fails on linux-x86 for all feature-sets (but not on MacOS or linux-x86_64) (see https://ci.cor-lab.org/job/sbcl-master/featureset=1,label=ubuntu_quantal_32bit/382/consoleFull) 2014-03-27T13:49:14Z scymtym_: 0bd6b78d is currently running 2014-03-27T14:06:42Z eudoxia quit (Quit: leaving) 2014-03-27T14:08:37Z drmeister joined #sbcl 2014-03-27T14:11:13Z yacks joined #sbcl 2014-03-27T14:14:12Z michael_lee joined #sbcl 2014-03-27T14:27:03Z tmh_ quit (Changing host) 2014-03-27T14:27:03Z tmh_ joined #sbcl 2014-03-27T14:37:49Z oleo joined #sbcl 2014-03-27T14:59:37Z attila_lendvai quit (Quit: Leaving.) 2014-03-27T15:21:25Z eudoxia joined #sbcl 2014-03-27T15:41:50Z drmeister quit (Ping timeout: 255 seconds) 2014-03-27T16:08:30Z scymtym_: 8fd27491 seems to have fixed the failures on linux-x86 2014-03-27T16:20:25Z Krystof: thanks 2014-03-27T16:52:57Z jdz quit (Quit: Leaving) 2014-03-27T16:55:32Z attila_lendvai joined #sbcl 2014-03-27T16:55:32Z attila_lendvai quit (Changing host) 2014-03-27T16:55:32Z attila_lendvai joined #sbcl 2014-03-27T16:55:49Z ASau joined #sbcl 2014-03-27T16:56:27Z attila_lendvai quit (Client Quit) 2014-03-27T16:58:42Z attila_lendvai joined #sbcl 2014-03-27T16:58:42Z attila_lendvai quit (Changing host) 2014-03-27T16:58:42Z attila_lendvai joined #sbcl 2014-03-27T17:21:14Z yacks quit (Quit: Leaving) 2014-03-27T17:30:51Z nyef joined #sbcl 2014-03-27T17:31:15Z nyef: Hello all. 2014-03-27T17:35:40Z ebrasca: nyef: hi 2014-03-27T17:36:05Z oleo quit (Ping timeout: 265 seconds) 2014-03-27T17:37:00Z oleo joined #sbcl 2014-03-27T17:37:11Z redline6561: hey nyef. what's next now that sparc is fixed up? 2014-03-27T17:37:36Z ebrasca: can i make in unix shell : function args 2014-03-27T17:37:46Z ferada quit (Ping timeout: 245 seconds) 2014-03-27T17:37:54Z ebrasca: and it execute me script with this args 2014-03-27T17:38:13Z nyef: I get back to the paying work for a bit, then either more SPARC fixups or taking another look at ARM, or figuring out what's going on with gray streams, or... 2014-03-27T17:40:37Z ebrasca: how i pase arguments to a lisp script in a unix sheell? 2014-03-27T17:45:56Z nyef: Presumably, the same way you'd pass the arguments to any other script. Now, for actually RECEIVING them, have a look for something called *POSIX-ARGV* if you're using SBCL. 2014-03-27T17:48:31Z ltbarcly joined #sbcl 2014-03-27T17:50:36Z slyrus quit (Ping timeout: 245 seconds) 2014-03-27T17:51:20Z ebrasca: nyef:looking 2014-03-27T18:18:54Z pranavrc quit (Remote host closed the connection) 2014-03-27T18:37:06Z ebrasca: nyef:can gif me one example? 2014-03-27T18:37:48Z nyef: Have a look at tests/run-tests.lisp in the SBCL source tree. 2014-03-27T18:40:02Z root_empire joined #sbcl 2014-03-27T18:42:43Z Krystof: working ARM would make me happy 2014-03-27T18:42:55Z Krystof: it somehow feels like only the hard 90% of the work remains 2014-03-27T18:43:16Z michael_lee quit (Ping timeout: 265 seconds) 2014-03-27T18:44:16Z eudoxia quit (Quit: leaving) 2014-03-27T18:46:20Z root_empire quit (Read error: Connection reset by peer) 2014-03-27T18:51:12Z nyef: Mmm. Great swathes of the system at least compiling, but huge amounts of it not working, subtle gotchas all over the place, and two or three absolute showstoppers that require tearing out and rewriting chunks of the system... 2014-03-27T18:52:47Z ebrasca: nyef: thx for help me 2014-03-27T18:54:16Z ltbarcly quit (Quit: Computer has gone to sleep.) 2014-03-27T18:55:33Z foom: Anyone know why sbcl on linux/x86-64 doesn't use %gs for TLS? 2014-03-27T18:56:47Z nyef: I'm going to go with "it was easier not to". 2014-03-27T18:57:15Z nyef: More seriously, IIRC, at the time (and I don't know about now), there weren't any user-mode APIs for setting the GS segment base in Linux. 2014-03-27T18:57:53Z foom: There is now, arch_prctl(ARCH_SET_GS, addr); I think it's been there for a long time, but not sure. 2014-03-27T18:59:05Z nyef: Similar issues for Win64, IIRC. 2014-03-27T18:59:42Z nyef: Besides, it's a LOT simpler to just use a GPR. 2014-03-27T19:00:03Z foom: bonus free register, though! 2014-03-27T19:01:13Z fe[nl]ix: yummy 2014-03-27T19:02:54Z foom: And then you can use a GPR for nil, and not feel so bad about the waste. 2014-03-27T19:05:14Z slyrus joined #sbcl 2014-03-27T19:16:39Z foom: well, said syscall has existed at least since 2.6.12-rc2 (first commit in kernel git) 2014-03-27T19:27:49Z drmeister joined #sbcl 2014-03-27T19:28:52Z ltbarcly joined #sbcl 2014-03-27T19:33:22Z ltbarcly_ joined #sbcl 2014-03-27T19:35:38Z ltbarcly quit (Ping timeout: 240 seconds) 2014-03-27T19:52:01Z ferada joined #sbcl 2014-03-27T20:03:09Z michael_lee joined #sbcl 2014-03-27T20:03:18Z drmeister quit (Ping timeout: 255 seconds) 2014-03-27T20:19:59Z fiveop joined #sbcl 2014-03-27T20:20:33Z fiveop: nyef: hey :) 2014-03-27T20:20:37Z nyef: Hello. 2014-03-27T20:20:59Z fiveop: How are things? In particular your general schedule :) 2014-03-27T20:21:41Z nyef: Busyish, but I'm co-located with my ARM system again. 2014-03-27T20:25:13Z fiveop: I haven't worked on the arm port in the last few weeks. 2014-03-27T20:26:07Z fiveop: I was going through all the #-arm locations in src/code for float related things and tried to add the relevant vops, constants, etc. 2014-03-27T20:26:15Z nyef: Mmm. And my SBCL-hacking priority has been SPARC recently. 2014-03-27T20:31:01Z fiveop: If you ever find a little time for the arm port, polishing and pushing your float VM code would be what I'd find most helpful :) 2014-03-27T20:32:21Z nyef: I was actually thinking to investigate the signal handling and stack issues. 2014-03-27T20:33:17Z fiveop: that, I'd find second most helpful :P 2014-03-27T20:38:31Z ebrasca quit (Remote host closed the connection) 2014-03-27T20:41:48Z ebrasca joined #sbcl 2014-03-27T20:48:56Z Bike_ joined #sbcl 2014-03-27T20:57:41Z angavrilov quit (Remote host closed the connection) 2014-03-27T20:58:16Z ebrasca quit (Remote host closed the connection) 2014-03-27T21:02:41Z ebrasca joined #sbcl 2014-03-27T21:07:37Z fiveop quit 2014-03-27T21:22:32Z Bike_ is now known as Bicyclidine 2014-03-27T21:34:55Z attila_lendvai quit (Quit: Leaving.) 2014-03-27T21:35:03Z attila_lendvai joined #sbcl 2014-03-27T21:35:27Z michael_lee quit (Ping timeout: 255 seconds) 2014-03-27T21:49:44Z nyef: The docstring for the special variable - seems to be misleading. Consider the result of (EVAL '-) at the REPL in light of the docstring. 2014-03-27T21:53:28Z attila_lendvai quit (Quit: Leaving.) 2014-03-27T21:59:57Z ltbarcly_ quit (Ping timeout: 240 seconds) 2014-03-27T22:00:53Z scymtym joined #sbcl 2014-03-27T22:04:10Z ltbarcly joined #sbcl 2014-03-27T22:18:51Z ltbarcly quit (Ping timeout: 265 seconds) 2014-03-27T22:21:27Z ltbarcly joined #sbcl 2014-03-27T22:44:49Z ltbarcly quit (Quit: Computer has gone to sleep.) 2014-03-27T22:47:17Z ltbarcly joined #sbcl 2014-03-27T22:47:57Z segv- quit (Ping timeout: 240 seconds) 2014-03-27T22:51:38Z ltbarcly quit (Ping timeout: 240 seconds) 2014-03-27T23:12:29Z Bicyclidine quit (Quit: leaving) 2014-03-27T23:21:38Z slyrus quit (Ping timeout: 240 seconds) 2014-03-28T00:05:56Z ltbarcly joined #sbcl 2014-03-28T00:11:36Z ltbarcly quit (Quit: Textual IRC Client: www.textualapp.com) 2014-03-28T01:15:49Z ebrasca quit (Remote host closed the connection) 2014-03-28T01:37:53Z andreh joined #sbcl 2014-03-28T01:45:16Z drmeister joined #sbcl 2014-03-28T01:46:07Z nyef quit (Quit: G'night all.) 2014-03-28T02:05:19Z andreh quit (Ping timeout: 264 seconds) 2014-03-28T02:20:25Z drmeister quit (Read error: Connection reset by peer) 2014-03-28T02:38:03Z yacks joined #sbcl 2014-03-28T02:40:17Z prxq quit (Ping timeout: 240 seconds) 2014-03-28T02:49:37Z scymtym quit (Ping timeout: 240 seconds) 2014-03-28T02:53:20Z prxq joined #sbcl 2014-03-28T03:11:28Z slyrus joined #sbcl 2014-03-28T03:38:49Z christoph_debian quit (Ping timeout: 265 seconds) 2014-03-28T03:52:41Z christoph_debian joined #sbcl 2014-03-28T04:02:39Z michael_lee joined #sbcl 2014-03-28T04:18:02Z pranavrc joined #sbcl 2014-03-28T05:13:54Z attila_lendvai joined #sbcl 2014-03-28T05:14:35Z pranavrc_ joined #sbcl 2014-03-28T05:15:45Z pranavrc quit (Ping timeout: 265 seconds) 2014-03-28T05:19:16Z slyrus quit (Remote host closed the connection) 2014-03-28T05:55:19Z slyrus joined #sbcl 2014-03-28T05:58:58Z oleo quit (Quit: Leaving) 2014-03-28T06:22:24Z slyrus quit (Ping timeout: 255 seconds) 2014-03-28T06:26:15Z slyrus joined #sbcl 2014-03-28T06:28:47Z angavrilov joined #sbcl 2014-03-28T06:31:43Z slyrus quit (Ping timeout: 264 seconds) 2014-03-28T06:44:27Z ASau quit (Ping timeout: 255 seconds) 2014-03-28T07:07:08Z Kevin_Xi joined #sbcl 2014-03-28T07:07:34Z Kevin_Xi left #sbcl 2014-03-28T07:12:41Z slyrus joined #sbcl 2014-03-28T09:30:38Z kludge` quit (Ping timeout: 252 seconds) 2014-03-28T09:36:34Z kludge` joined #sbcl 2014-03-28T11:19:22Z segv- joined #sbcl 2014-03-28T11:32:09Z andreh_ joined #sbcl 2014-03-28T12:08:41Z attila_lendvai quit (Quit: Leaving.) 2014-03-28T12:11:23Z attila_lendvai joined #sbcl 2014-03-28T12:11:23Z attila_lendvai quit (Changing host) 2014-03-28T12:11:23Z attila_lendvai joined #sbcl 2014-03-28T12:39:44Z pranavrc_ quit 2014-03-28T12:41:39Z eudoxia joined #sbcl 2014-03-28T12:49:09Z drmeister joined #sbcl 2014-03-28T13:07:17Z drmeister quit (Read error: Connection reset by peer) 2014-03-28T13:31:39Z nyef joined #sbcl 2014-03-28T13:31:47Z nyef: G'morning all. 2014-03-28T13:35:46Z pranavrc joined #sbcl 2014-03-28T13:36:38Z ams: morning 2014-03-28T13:37:25Z fe[nl]ix: hi nyef 2014-03-28T13:39:33Z Krystof: morning 2014-03-28T13:39:40Z oleo joined #sbcl 2014-03-28T13:52:37Z ebrasca joined #sbcl 2014-03-28T13:56:56Z drmeister joined #sbcl 2014-03-28T14:10:12Z ebrasca quit (Remote host closed the connection) 2014-03-28T14:29:55Z pkhuong: good morning 2014-03-28T14:30:09Z fe[nl]ix: hi pkhuong 2014-03-28T14:54:51Z drmeister quit (Read error: Connection reset by peer) 2014-03-28T14:55:06Z drmeister joined #sbcl 2014-03-28T15:27:46Z drmeiste_ joined #sbcl 2014-03-28T15:30:40Z drmeister quit (Ping timeout: 268 seconds) 2014-03-28T15:33:00Z loke joined #sbcl 2014-03-28T15:33:02Z drmeister joined #sbcl 2014-03-28T15:33:15Z drmeiste_ quit (Ping timeout: 252 seconds) 2014-03-28T15:43:00Z drmeister quit (Ping timeout: 268 seconds) 2014-03-28T16:08:18Z eudoxia quit (Ping timeout: 255 seconds) 2014-03-28T16:09:50Z drmeister joined #sbcl 2014-03-28T16:11:49Z eudoxia joined #sbcl 2014-03-28T16:22:53Z drmeister quit (Read error: No route to host) 2014-03-28T16:23:14Z drmeister joined #sbcl 2014-03-28T16:36:51Z cantstanya joined #sbcl 2014-03-28T16:37:02Z cantstanya left #sbcl 2014-03-28T16:56:11Z ASau joined #sbcl 2014-03-28T17:01:19Z hlavaty quit (Remote host closed the connection) 2014-03-28T17:08:37Z attila_lendvai quit (Quit: Leaving.) 2014-03-28T17:09:11Z attila_lendvai joined #sbcl 2014-03-28T17:40:36Z pranavrc quit (Quit: Ping timeout: ∞) 2014-03-28T17:40:37Z yacks quit (Ping timeout: 240 seconds) 2014-03-28T18:24:36Z oleo quit (Ping timeout: 240 seconds) 2014-03-28T18:29:27Z eudoxia quit (Quit: leaving) 2014-03-28T18:38:05Z oleo joined #sbcl 2014-03-28T18:54:16Z drmeister quit (Ping timeout: 240 seconds) 2014-03-28T19:03:31Z slyrus quit (Ping timeout: 264 seconds) 2014-03-28T19:06:22Z slyrus joined #sbcl 2014-03-28T19:31:39Z Bike_ joined #sbcl 2014-03-28T19:32:02Z Bike_ is now known as Bicyclidine 2014-03-28T19:35:51Z scymtym joined #sbcl 2014-03-28T19:50:17Z slyrus quit (Ping timeout: 268 seconds) 2014-03-28T20:19:56Z attila_lendvai quit (Quit: Leaving.) 2014-03-28T20:21:08Z segv- quit (Ping timeout: 240 seconds) 2014-03-28T20:30:01Z slyrus joined #sbcl 2014-03-28T20:35:22Z angavrilov quit (Remote host closed the connection) 2014-03-28T20:38:41Z asedeno quit (Quit: leaving) 2014-03-28T20:39:52Z michael_lee quit (Remote host closed the connection) 2014-03-28T20:41:19Z leuler joined #sbcl 2014-03-28T21:13:47Z andreh__ joined #sbcl 2014-03-28T21:14:56Z Bicyclidine quit (Ping timeout: 265 seconds) 2014-03-28T21:15:21Z andreh_ quit (Ping timeout: 252 seconds) 2014-03-28T21:22:22Z Bicyclidine joined #sbcl 2014-03-28T21:39:51Z drmeister joined #sbcl 2014-03-28T21:51:15Z leuler quit (Quit: ERC Version 5.1.2 $Revision: 1.796.2.6 $ (IRC client for Emacs)) 2014-03-28T21:54:00Z andreh__ quit (Ping timeout: 255 seconds) 2014-03-28T21:55:09Z drmeister quit (Remote host closed the connection) 2014-03-28T22:04:15Z Bicyclidine quit (Ping timeout: 255 seconds) 2014-03-28T22:16:18Z drmeister joined #sbcl 2014-03-28T22:24:35Z Bicyclidine joined #sbcl 2014-03-28T22:39:15Z drmeister quit (Ping timeout: 268 seconds) 2014-03-28T22:55:38Z Bicyclidine quit (Ping timeout: 240 seconds) 2014-03-28T23:17:48Z drmeister joined #sbcl 2014-03-29T00:55:26Z edgar-rft quit (Quit: existence stopped by memory explosion) 2014-03-29T01:13:34Z drmeister quit (Read error: Connection reset by peer) 2014-03-29T01:14:18Z drmeister joined #sbcl 2014-03-29T01:33:42Z scymtym quit (Read error: Operation timed out) 2014-03-29T02:07:36Z danlentz quit (Quit: danlentz) 2014-03-29T02:08:52Z drmeister quit (Remote host closed the connection) 2014-03-29T02:37:44Z prxq_ joined #sbcl 2014-03-29T02:41:08Z prxq quit (Ping timeout: 240 seconds) 2014-03-29T02:42:43Z nyef quit (Quit: G'night all.) 2014-03-29T03:12:38Z drmeister joined #sbcl 2014-03-29T03:39:08Z christoph_debian quit (Ping timeout: 255 seconds) 2014-03-29T03:53:27Z christoph_debian joined #sbcl 2014-03-29T04:35:12Z pranavrc joined #sbcl 2014-03-29T04:37:30Z yacks joined #sbcl 2014-03-29T05:09:11Z joshe joined #sbcl 2014-03-29T05:36:16Z drmeister quit (Remote host closed the connection) 2014-03-29T06:10:00Z angavrilov joined #sbcl 2014-03-29T06:56:57Z ccl-logbot joined #sbcl 2014-03-29T06:56:57Z 2014-03-29T06:56:57Z names: ccl-logbot angavrilov joshe yacks pranavrc christoph_debian prxq_ slyrus oleo ASau loke kludge` ferada jrm luis` pkhuong mood redline6561 brucem jsnell_ antoszka scymtym_ White_Flame Blkt daimrod pchrist specbot minion foom milosn fikusz Posterdati Guest50092 yauz faheem ivan`` Bike flip214 kanru pjb nicdev Munksgaard fe[nl]ix irsol tmh_ DGASAU leoc xymox jaimef @Krystof ams psilord |3b| 2014-03-29T07:25:10Z drmeister joined #sbcl 2014-03-29T07:29:52Z drmeister quit (Ping timeout: 252 seconds) 2014-03-29T07:31:38Z yacks quit (Ping timeout: 240 seconds) 2014-03-29T07:36:34Z yacks joined #sbcl 2014-03-29T08:06:28Z ltbarcly joined #sbcl 2014-03-29T08:10:56Z ltbarcly_ joined #sbcl 2014-03-29T08:12:08Z ltbarcly quit (Ping timeout: 268 seconds) 2014-03-29T08:17:28Z enupten joined #sbcl 2014-03-29T08:21:29Z ltbarcly_ quit (Quit: Computer has gone to sleep.) 2014-03-29T08:23:10Z ltbarcly joined #sbcl 2014-03-29T08:36:36Z stassats joined #sbcl 2014-03-29T08:37:22Z pranavrc quit (Quit: Ping timeout: ∞) 2014-03-29T08:50:18Z stassats: something bad happens with the number stack, floating point registers and returning from a thread on PPC 2014-03-29T09:13:56Z drmeister joined #sbcl 2014-03-29T09:18:21Z drmeister quit (Ping timeout: 255 seconds) 2014-03-29T09:26:16Z stassats quit (Ping timeout: 240 seconds) 2014-03-29T09:31:36Z kludge` quit (Ping timeout: 240 seconds) 2014-03-29T09:36:38Z kludge` joined #sbcl 2014-03-29T09:59:50Z edgar-rft joined #sbcl 2014-03-29T10:00:23Z enupten quit (Ping timeout: 265 seconds) 2014-03-29T11:02:38Z drmeister joined #sbcl 2014-03-29T11:07:08Z drmeister quit (Ping timeout: 240 seconds) 2014-03-29T12:10:15Z ltbarcly quit (Ping timeout: 255 seconds) 2014-03-29T12:13:24Z ltbarcly joined #sbcl 2014-03-29T12:45:59Z stassats joined #sbcl 2014-03-29T12:51:25Z drmeister joined #sbcl 2014-03-29T12:55:36Z drmeister quit (Ping timeout: 240 seconds) 2014-03-29T13:52:56Z drmeister joined #sbcl 2014-03-29T14:25:50Z scymtym joined #sbcl 2014-03-29T14:33:48Z LiamH joined #sbcl 2014-03-29T14:47:25Z attila_lendvai joined #sbcl 2014-03-29T14:47:25Z attila_lendvai quit (Changing host) 2014-03-29T14:47:25Z attila_lendvai joined #sbcl 2014-03-29T15:16:37Z pkhuong: stassats: "Depending on the ABI, the stack pointer shall be 16-bytes (Sys V PPC ABI) or 2014-03-29T15:16:40Z pkhuong: 8-bytes (PPC EABI) aligned" 2014-03-29T15:16:58Z ltbarcly quit (Quit: Computer has gone to sleep.) 2014-03-29T15:17:32Z stassats: i need three things: a non local exit catcher, fixnum->float conversion and something saved on the number stack 2014-03-29T15:18:07Z stassats: http://paste.lisp.org/display/141838 2014-03-29T15:18:48Z stassats: gdb says it fails somewhere in call_into_lisp during register restoration 2014-03-29T15:19:15Z stassats: i presume it's the thread calling into lisp and returning 2014-03-29T15:20:32Z pkhuong: what happens if you force a 64 bit value (or 2 words) on the number stack? 2014-03-29T15:20:32Z stassats: and safety 0 is necessary too 2014-03-29T15:20:50Z stassats: a double-float? 2014-03-29T15:21:03Z pkhuong: for instance. 2014-03-29T15:22:00Z stassats: changing (* x 2) to (coerce x 'double-float) doesn't segfault 2014-03-29T15:24:07Z stassats: (let ((number (* x 2)) (number2 (* x 3))) (foo 0) (and (plusp number) (plusp number2))) doesn't crash either 2014-03-29T15:25:25Z stassats: it is a bit strange since the stack is supposed to be unwound when call_into_lisp finishes 2014-03-29T15:28:55Z stassats: though, (plusp number2) is probably optimized out, because (let ((number (* x 2)) (number2 (* x 3))) (foo 0) (and (plusp number) (plusp number2))) does crash 2014-03-29T15:29:01Z stassats: wrong paste 2014-03-29T15:29:05Z stassats: (+ number number2) 2014-03-29T15:32:14Z pkhuong: we're not clever enough to optimise (plusp number) away 2014-03-29T15:32:25Z pkhuong: *(plusp number2) 2014-03-29T15:32:43Z stassats: (and (plusp number) (plusp number2)) doesn't crash, (+ number number2) does crash 2014-03-29T15:33:12Z stassats: but it's not clear why the stack higher up is affected, is the number stack not adjusted properly after returning? 2014-03-29T15:34:00Z stassats: i probably need to unroll C_FULL_EPILOG macro to get gdb show something better 2014-03-29T15:34:27Z pkhuong: oh yeah, that's really annoying. Sometimes I just stare at asm instead. 2014-03-29T15:39:51Z stassats: fails at lwz 0,4(1) 2014-03-29T15:40:39Z stassats: https://github.com/sbcl/sbcl/blob/master/src/runtime/ppc-assem.S#L278 2014-03-29T15:41:27Z stassats: in the prolog it's mfcr 0 stw 0,8(1), but lwz 1,0(1) ; \ lwz 0,4(1) in the epilog 2014-03-29T15:41:31Z stassats: is 4 vs 8 right? 2014-03-29T15:41:47Z stassats: eh, it's the wrong way 2014-03-29T15:42:44Z stassats: still, it doesn't seem to be matching 2014-03-29T15:45:55Z stassats: but it couldn't have been broken all this time, could it 2014-03-29T15:47:56Z pkhuong: We've all seen some very old and stupid bugs. 2014-03-29T15:48:14Z stassats: ok, so, apparently the register number 1 has some wrong address 2014-03-29T15:48:54Z stassats: that's reg_NSP 2014-03-29T15:50:47Z stassats: i'm now confused about the relationship between NFP and NSP 2014-03-29T15:53:03Z pkhuong: I can guarantee you I'm just as confused about anything number-stack related. 2014-03-29T15:57:40Z hzp joined #sbcl 2014-03-29T16:06:18Z stassats: %single-float/signed VOP for some reason is concerned with NFP 2014-03-29T16:10:58Z stassats: it stores two parts of some magic as two words on the number stack and then reads it back as a double 2014-03-29T16:11:33Z stassats: is that really the best way to do integer to float conversion? 2014-03-29T16:16:19Z pkhuong: I expect that's the only way to move values from GPRs to FPRs. 2014-03-29T16:16:21Z stassats: looks like the answer is "yes" 2014-03-29T16:17:10Z stassats: since there's no 32-bit instruction for int->float 2014-03-29T16:22:18Z stassats: how does other code knows that %single-float/signed uses two NFP seats? 2014-03-29T16:23:24Z stassats: it probably doesn't matter, since it discards the values once it's done 2014-03-29T16:26:33Z stassats: wait, gdb says Cannot access memory at address 0x43300000 2014-03-29T16:26:45Z stassats: the half of the float magic is #x4330 2014-03-29T16:33:19Z attila_lendvai quit (Quit: Leaving.) 2014-03-29T16:33:28Z attila_lendvai joined #sbcl 2014-03-29T16:33:29Z attila_lendvai quit (Changing host) 2014-03-29T16:33:29Z attila_lendvai joined #sbcl 2014-03-29T16:37:59Z stassats: so, looks like NSP is pointing at the wrong thing or the right thing is overwritten 2014-03-29T16:40:09Z bege joined #sbcl 2014-03-29T16:48:28Z leuler joined #sbcl 2014-03-29T16:50:33Z LiamH quit (Quit: Leaving.) 2014-03-29T16:51:07Z LiamH joined #sbcl 2014-03-29T17:11:41Z attila_lendvai quit (Read error: Connection reset by peer) 2014-03-29T17:11:46Z attila_lendvai1 joined #sbcl 2014-03-29T17:11:46Z attila_lendvai1 is now known as attila_lendvai 2014-03-29T17:11:46Z attila_lendvai quit (Changing host) 2014-03-29T17:11:46Z attila_lendvai joined #sbcl 2014-03-29T17:15:09Z stassats: it doesn't crash on 1.1.13.51, time for bisecting 2014-03-29T17:25:49Z attila_lendvai quit (Read error: Connection reset by peer) 2014-03-29T17:26:01Z attila_lendvai joined #sbcl 2014-03-29T17:26:01Z attila_lendvai quit (Changing host) 2014-03-29T17:26:01Z attila_lendvai joined #sbcl 2014-03-29T17:33:21Z stassats: it's starting to look like another sb-regalloc problem 2014-03-29T17:38:17Z nyef joined #sbcl 2014-03-29T17:38:52Z nyef: Hello all. 2014-03-29T17:39:05Z stassats: nyef: you're just in time for ppc regalloc problems! 2014-03-29T17:39:14Z nyef: Yes, I saw the thread on sbcl-devel. 2014-03-29T17:39:20Z nyef: Fun and games, huh? 2014-03-29T17:39:48Z stassats: you bet 2014-03-29T17:39:59Z nyef: Unfortunately, and I'm sure you've heard me say this not so long ago about another platform, I don't have any of my PPC systems nearby, and won't for about another week. 2014-03-29T17:40:49Z stassats: i can provide traces 2014-03-29T17:41:08Z pkhuong: nyef: and GCC's compile farm has PPC systems 2014-03-29T17:41:14Z stassats: here's the code: http://paste.lisp.org/display/141838 2014-03-29T17:42:00Z attila_lendvai quit (Quit: Leaving.) 2014-03-29T17:42:12Z stassats: so, it fails while exiting from call_into_lisp: https://github.com/sbcl/sbcl/blob/master/src/runtime/ppc-assem.S#L278 2014-03-29T17:42:27Z attila_lendvai joined #sbcl 2014-03-29T17:42:49Z stassats: REG_1 is 0x43300000 at that moment, and that comes from https://github.com/sbcl/sbcl/blob/master/src/compiler/ppc/float.lisp#L468 2014-03-29T17:44:06Z nyef: And that's reg_nsp? 2014-03-29T17:44:08Z Krystof: I newly have compile-farm access 2014-03-29T17:44:29Z nyef: That's reg_nsp! 2014-03-29T17:44:53Z Krystof: (in that way of things, it's spring and therefore I am reconstructing a usable ~/.emacs from scratch) 2014-03-29T17:44:57Z stassats: good and bad traces: http://paste.lisp.org/display/141838#1 2014-03-29T17:44:59Z nyef: ... Okay, that strongly indicates a corrupted NFP on the stack. 2014-03-29T17:45:22Z stassats goes to a bigger screen to ponder on them 2014-03-29T17:45:40Z attila_lendvai quit (Client Quit) 2014-03-29T17:45:53Z attila_lendvai joined #sbcl 2014-03-29T17:45:53Z attila_lendvai quit (Changing host) 2014-03-29T17:45:53Z attila_lendvai joined #sbcl 2014-03-29T17:46:06Z nyef: I'll make a prediction: One of them is using NFP as a temporary, the other isn't. 2014-03-29T17:46:20Z stassats: the VOP is the same, both are using NFP 2014-03-29T17:46:39Z stassats: at least i think it's the same 2014-03-29T17:47:08Z nyef: Umm... Also include trace for TEST please. 2014-03-29T17:47:24Z stassats: right, it haven't been changed in about a decade 2014-03-29T17:48:19Z nyef: At some level, I knew there was something iffy going on with NFP/NSP, possibly on all precise targets. /-: 2014-03-29T17:50:23Z stassats: http://paste.lisp.org/display/141838#3 test trace 2014-03-29T17:50:29Z nyef: Thank you. 2014-03-29T17:52:01Z stassats: there may have been other changes, the good trace is from 1.1.13.5something 2014-03-29T17:52:21Z stassats: i'll get closer versions eventually 2014-03-29T17:53:11Z nyef: That's fine. I need to spend about half an hour or so getting my dinner started, then I can concentrate on this properly. 2014-03-29T17:55:59Z stassats: and it also doesn't depend on threads, i made a version of the test with signals and it failed on a thread-less sbcl too 2014-03-29T18:07:00Z stassats: closer versions: http://paste.lisp.org/display/141839 2014-03-29T18:07:07Z attila_lendvai quit (Ping timeout: 252 seconds) 2014-03-29T18:08:57Z stassats: first thing i see bad: STW $NL0,0($NFP) vs STW $NL0,8($NFP) 2014-03-29T18:10:49Z attila_lendvai joined #sbcl 2014-03-29T18:10:49Z attila_lendvai quit (Changing host) 2014-03-29T18:10:49Z attila_lendvai joined #sbcl 2014-03-29T18:11:11Z stassats: that's in move-to-word/fixnum 2014-03-29T18:12:48Z stassats: can't exactly pin point the culprit revision since it doesn't build on ppc in between 2014-03-29T18:14:12Z nyef: Why is that bad? 2014-03-29T18:14:38Z stassats: it should have been bad: ... good: ... 2014-03-29T18:15:16Z nyef: Ah. 2014-03-29T18:15:32Z nyef: NFP stack chains aren't stored on the number stack. 2014-03-29T18:16:05Z stassats: the offsets to NFP seems to be all different 2014-03-29T18:17:00Z nyef: Which is what you'd expect with PACK changes... 2014-03-29T18:18:09Z nyef: On the other hand... Does anything in the "good" version hit 0(NFP)? 2014-03-29T18:18:50Z stassats: yes, 0x4330 is stored there, vs 32 on the bad version 2014-03-29T18:19:02Z stassats: (0x43300000 is the fault address) 2014-03-29T18:32:24Z nyef: Umm... Okay, I found something that worries me. Look at the STWU instructions in XEP-ALLOCATE-FRAME. 2014-03-29T18:37:22Z stassats: what's worrying about it? 2014-03-29T18:38:51Z nyef: Actually, I'm not sure right now... Still getting everything set up for my dinner, and looking at code whenever I need to look something up in the recipe. 2014-03-29T18:39:09Z stassats: don't cook a bus error 2014-03-29T18:39:52Z nyef: "When life gives you a lemon, make lemonade. When life gives you a hemlock, don't make hemlockade. Always know the difference between a lemon and a hemlock." 2014-03-29T18:42:31Z stassats: lisp code appears to be happy with the stack, only the call_into_lisp doesn't like it 2014-03-29T19:04:40Z Krystof: right, if it's an ABI thing lisp code can just merrily roll along 2014-03-29T19:04:51Z Krystof: like on mips, all lisp code is one big function for the C abi 2014-03-29T19:05:15Z nyef: Almost any non-x86oid port is like that, actually. 2014-03-29T19:05:22Z ltbarcly joined #sbcl 2014-03-29T19:07:38Z stassats: so, it is supposed to find the location of the link register at 0(NSP) 2014-03-29T19:09:59Z stassats: or rather, it saves LR at 4(1) and it restores REG_1 from 0(1), but it's bogus 2014-03-29T19:11:54Z stassats: and not a single comment 2014-03-29T19:12:10Z stassats: "it's so obvious, let's not explain it" 2014-03-29T19:15:36Z stassats: the epilogue of call_into_lisp is rarely exercised, so it may have gone broken undetected for quite some time 2014-03-29T19:16:04Z stassats: only when exiting threads and signals handlers, c callbacks 2014-03-29T19:16:44Z nyef: And, quite often, threads wouldn't return "normally" anyway. 2014-03-29T19:20:29Z oleo quit (Ping timeout: 265 seconds) 2014-03-29T19:21:56Z oleo joined #sbcl 2014-03-29T19:22:33Z Krystof: stassats: meh, this is much easier than the bug when the ppc platform would faill to call into lisp when the initial function address's bit 13 was set 2014-03-29T19:26:24Z nyef: Okay, my dinner is finally in the oven. 2014-03-29T19:26:34Z stassats: on ppc, c stack grows downward, but the lisp stack grows upward, could there be a problem? 2014-03-29T19:26:45Z ltbarcly quit (Ping timeout: 255 seconds) 2014-03-29T19:26:55Z nyef: Shouldn't be, except possibly in the debugger. 2014-03-29T19:27:26Z nyef: We have the same issue on ARM, actually, although there it's complicated by having to share a stack pointer(!). 2014-03-29T19:27:44Z stassats: so, why in STWU $NSP,-24($NSP) is -24 negative? 2014-03-29T19:28:01Z nyef: Just about to look at that. 2014-03-29T19:30:09Z nyef: It's negative because the C stack grows downwards? 2014-03-29T19:30:42Z stassats: and NSP is the C stack? 2014-03-29T19:30:54Z nyef: Yes. Should be the C stack pointer, too. 2014-03-29T19:31:05Z nyef: And NFP should be hte C frame pointer, IIRC. 2014-03-29T19:31:28Z stassats: when i stack allocate a cons, its address grows 2014-03-29T19:32:04Z nyef: Our control stack is CFP and CSP. 2014-03-29T19:32:21Z nyef: And stack-allocated-boxed-objects go on the control stack. 2014-03-29T19:32:46Z nyef: We don't stack-allocate unboxed arrays, IIRC, and we CAN'T stack-allocate mixed objects such as structures with unboxed slots. 2014-03-29T19:32:57Z stassats: so, why does it touch the C stack? 2014-03-29T19:33:11Z ltbarcly joined #sbcl 2014-03-29T19:33:30Z nyef: Because the C stack is the number stack. 2014-03-29T19:33:38Z stassats: blimey 2014-03-29T19:34:47Z stassats: so, does %single-float/signed compute stack offsets right? 2014-03-29T19:35:16Z nyef: 'swhat I'm going to look into as soon as I've figured out if number-stack-displacement does what I think it does. 2014-03-29T19:37:36Z nyef: Looks reasonable at first glance. 2014-03-29T19:38:01Z nyef: Is NFP mapped to the C frame pointer, btw? 2014-03-29T19:42:41Z stassats: is STWU $NSP,-24($NSP) enough to store things at 36($NFP)? 2014-03-29T19:43:21Z nyef: No, it's only enough to cover through 16($NFP). 2014-03-29T19:43:30Z nyef: Err... Not even that. 2014-03-29T19:43:52Z stassats: it's STWU $NSP,-24($NSP) then ADDI $NFP, $NSP, 8 2014-03-29T19:44:09Z stassats: so, it seems like $NSP-16 2014-03-29T19:44:20Z nyef: Yeah, for a net frame size of 16 octets. 2014-03-29T19:45:37Z stassats: in the offending code, it doesn't touch any lower NFP offset, goes straight to 32 and 36 2014-03-29T19:45:43Z stassats: and 0 2014-03-29T19:46:40Z stassats: no, it does indirectly, through $A0 2014-03-29T19:51:08Z scymtym quit (Ping timeout: 255 seconds) 2014-03-29T19:51:27Z stassats: the bad code uses 32 and 36 for %single-float/signed, and upto 20($A0) where MR $A0, $NFP 2014-03-29T19:51:53Z stassats: but the coode uses 0 and 4 for %single-float/signed, and still up to 20($A0) 2014-03-29T19:52:00Z stassats: so, those 32 and 36 seem suspect 2014-03-29T19:54:38Z stassats: why does %single-float/signed multiply by n-word-bytes twice? 2014-03-29T19:54:58Z stassats: https://github.com/sbcl/sbcl/blob/master/src/compiler/ppc/float.lisp#L464 2014-03-29T19:57:09Z nyef: That's it, right there. 2014-03-29T19:57:11Z stassats: when the tn-offset is 0, high and low are 0 and 4, but if it's 2, they become 32 and 36, it seens that it should be (let* ((stack-offset (tn-offset temp))) 2014-03-29T19:57:13Z nyef: Good catch. 2014-03-29T19:58:06Z nyef: Either that, or temp-offset-low should be (+ stack-offset n-word-bytes), depending on if you want the strength reduction or not. 2014-03-29T19:58:51Z enupten joined #sbcl 2014-03-29T20:06:01Z stassats: testing, i wonder whether that can be counted as a regression? 2014-03-29T20:06:47Z Krystof: if it's a change just to compiler/ppc, I'll take it 2014-03-29T20:06:57Z Krystof: (and NEWS) 2014-03-29T20:06:59Z stassats: it is 2014-03-29T20:07:39Z nyef: It's not exactly a regression, it's a latent bug that was tickled by the new PACK changes, just like the similar bug on SPARC. 2014-03-29T20:10:57Z ltbarcly quit (Quit: Computer has gone to sleep.) 2014-03-29T20:12:05Z stassats: nothing seems to be failing anymore 2014-03-29T20:13:05Z nyef: Ship it, then desk-check the MIPS and Alpha versions of that VOP. d-: 2014-03-29T20:15:46Z nyef: Okay, the MIPS and Alpha versions don't have any obvious issues at first glance. 2014-03-29T20:16:03Z ltbarcly joined #sbcl 2014-03-29T20:20:48Z stassats: mips has a proper int-to-float instruction 2014-03-29T20:20:55Z stassats: and alpha does something saner too 2014-03-29T20:35:57Z ltbarcly quit (Quit: Computer has gone to sleep.) 2014-03-29T20:36:53Z xymox quit (Ping timeout: 252 seconds) 2014-03-29T20:40:13Z xymox joined #sbcl 2014-03-29T20:41:24Z nyef: Alpha does the SPARC thing, only without screwing up the temp TNs. 2014-03-29T20:42:31Z stassats quit (Ping timeout: 264 seconds) 2014-03-29T20:51:38Z ltbarcly joined #sbcl 2014-03-29T21:08:58Z angavrilov quit (Remote host closed the connection) 2014-03-29T21:13:45Z ASau quit (Ping timeout: 252 seconds) 2014-03-29T21:46:08Z xymox quit (Ping timeout: 268 seconds) 2014-03-29T21:49:07Z xymox joined #sbcl 2014-03-29T21:52:52Z ASau joined #sbcl 2014-03-29T22:20:05Z ltbarcly_ joined #sbcl 2014-03-29T22:22:16Z ltbarcly quit (Ping timeout: 240 seconds) 2014-03-29T22:45:56Z leuler quit (Quit: ERC Version 5.1.2 $Revision: 1.796.2.6 $ (IRC client for Emacs)) 2014-03-29T22:47:44Z attila_lendvai quit (Quit: Leaving.) 2014-03-29T23:10:20Z yacks quit (Ping timeout: 265 seconds) 2014-03-29T23:26:58Z LiamH quit (Quit: Leaving.) 2014-03-29T23:30:01Z Quadrescence joined #sbcl 2014-03-30T00:09:08Z Quadrescence quit (Quit: Leaving) 2014-03-30T00:21:52Z Quadrescence joined #sbcl 2014-03-30T00:32:55Z LiamH joined #sbcl 2014-03-30T00:39:47Z nyef: Hey, it looks like one of the GCC compile farm machines is an hppa box. And it's a multi-CPU system, at that. Should take about 90-120 minutes for a full build on that, assuming no bit-rot. 2014-03-30T00:56:37Z nyef: (Yes, I'm fully aware that the "no bit-rot" assumption is laughable for the HPPA case... And the ALPHA case, for that matter. 2014-03-30T00:56:39Z nyef: ) 2014-03-30T01:15:32Z scymtym joined #sbcl 2014-03-30T01:57:55Z edgar-rft quit (Quit: happening finished by invalid something) 2014-03-30T02:12:16Z LiamH quit (Quit: Leaving.) 2014-03-30T02:14:42Z ltbarcly_ quit (Quit: Computer has gone to sleep.) 2014-03-30T02:17:51Z ltbarcly joined #sbcl 2014-03-30T02:24:19Z ltbarcly quit (Quit: Computer has gone to sleep.) 2014-03-30T02:38:23Z prxq__ joined #sbcl 2014-03-30T02:39:09Z christoph_debian quit (Ping timeout: 265 seconds) 2014-03-30T02:41:38Z prxq_ quit (Ping timeout: 240 seconds) 2014-03-30T02:43:43Z ltbarcly joined #sbcl 2014-03-30T02:53:14Z christoph_debian joined #sbcl 2014-03-30T03:31:10Z yacks joined #sbcl 2014-03-30T03:40:08Z nyef quit (Quit: G'night all.) 2014-03-30T04:08:20Z pranavrc joined #sbcl 2014-03-30T04:36:26Z michael_lee joined #sbcl 2014-03-30T04:38:56Z krzysz00 joined #sbcl 2014-03-30T05:26:56Z angavrilov joined #sbcl 2014-03-30T05:40:47Z attila_lendvai joined #sbcl 2014-03-30T06:04:29Z krzysz00 quit (Ping timeout: 255 seconds) 2014-03-30T06:05:22Z ltbarcly quit (Quit: Computer has gone to sleep.) 2014-03-30T06:23:13Z enupten quit (Ping timeout: 265 seconds) 2014-03-30T06:30:36Z yacks quit (Ping timeout: 240 seconds) 2014-03-30T06:37:52Z pranavrc quit (Remote host closed the connection) 2014-03-30T06:53:26Z yacks joined #sbcl 2014-03-30T07:03:35Z scymtym quit (Ping timeout: 268 seconds) 2014-03-30T07:06:11Z enupten joined #sbcl 2014-03-30T07:18:36Z stassats joined #sbcl 2014-03-30T07:37:51Z oleo quit (Ping timeout: 265 seconds) 2014-03-30T07:38:12Z oleo joined #sbcl 2014-03-30T07:50:53Z enupten quit (Quit: Leaving) 2014-03-30T08:06:39Z attila_lendvai quit (Quit: Leaving.) 2014-03-30T08:09:39Z fikusz quit (Quit: Leaving) 2014-03-30T08:12:39Z fikusz joined #sbcl 2014-03-30T09:17:24Z jaimef quit (Ping timeout: 268 seconds) 2014-03-30T09:31:42Z jaimef joined #sbcl 2014-03-30T09:34:46Z kludge` quit (Ping timeout: 252 seconds) 2014-03-30T09:39:49Z kludge` joined #sbcl 2014-03-30T10:58:56Z xymox quit (Ping timeout: 265 seconds) 2014-03-30T11:01:25Z krzysz00 joined #sbcl 2014-03-30T11:02:41Z xymox joined #sbcl 2014-03-30T12:15:21Z stassats: sb-dynamic-cored sbcl strangely fails somewhere in PCL bootstrap 2014-03-30T12:18:32Z LiamH joined #sbcl 2014-03-30T12:28:08Z stassats: and it doesn't make any sense, as usual 2014-03-30T13:05:13Z ASau` joined #sbcl 2014-03-30T13:08:08Z ASau quit (Ping timeout: 240 seconds) 2014-03-30T13:38:52Z tmh_ quit (Ping timeout: 268 seconds) 2014-03-30T13:41:44Z stassats: i'm probably uncovering some other ppc bug, since i don't see foreign stuff anywhere close 2014-03-30T13:44:59Z tmh_ joined #sbcl 2014-03-30T14:38:31Z nyef joined #sbcl 2014-03-30T14:38:40Z nyef: G'morning all. 2014-03-30T14:41:08Z Krystof: nyef: there's more to hppa brokenness than just bitrot 2014-03-30T14:41:15Z Krystof: as far as I know, it only ever worked on PARISC 1 2014-03-30T14:41:26Z Krystof: I got icache problems when trying to make it work on PARISC 2 2014-03-30T14:45:42Z nyef: Yeah, I figured it'd be more than a little project. 2014-03-30T14:49:06Z stassats: ppc has more surprises, at least it's in warm init, but it just seems to fail to return from a function 2014-03-30T14:49:24Z nyef: Oh, joy. 2014-03-30T14:49:33Z nyef: I'm a week away from having access to my own PPC systems. 2014-03-30T14:49:45Z stassats: that's with sb-dynamic-core, but it really shouldn't affect bootstrapping PCL 2014-03-30T15:06:04Z scymtym joined #sbcl 2014-03-30T15:09:10Z stassats: is ldb ever not useless on ppc? 2014-03-30T15:13:40Z nyef: LDB is useful for a couple of things, but it's mostly useless. 2014-03-30T15:14:21Z nyef: Looking at interrupt contexts, objects in memory, and getting a backtrace is about the limit of what you can do with LDB, and even these don't always work. 2014-03-30T15:14:41Z stassats: i'm yet to see a useful backtrace on PPC 2014-03-30T15:15:09Z nyef: Non-x86oid backtraces don't include alien frames, for starters. 2014-03-30T15:15:41Z nyef: And then, depending on how you end up in LDB, there can be issues that truncate the backtrace or elide frames. 2014-03-30T15:16:18Z stassats: right now typing back gets me another memory fault 2014-03-30T15:17:04Z edgar-rft joined #sbcl 2014-03-30T15:17:14Z michael_lee quit (Quit: Ex-Chat) 2014-03-30T15:17:20Z nyef: And memory faults in SBCL are horribly nasty and broken anyway. /-: 2014-03-30T15:17:47Z nyef: Bet you can't even get an interrupt context out of it. 2014-03-30T15:18:49Z nyef: (Why not? Because the memory fault path in the runtime doesn't handle that case the way that every other path does for some reason.) 2014-03-30T15:21:13Z nyef: And PPC is one of the better-maintained non-x86oid backends. Fun, huh? 2014-03-30T15:22:24Z nyef: On the upside, gdb usually works okay on non-x86oid backends because they maintain the C call stack fairly well. 2014-03-30T15:23:04Z stassats: it looks like i can't call CLASS-WRAPPER on a class 2014-03-30T15:23:46Z stassats: or i just can't print it 2014-03-30T15:26:57Z pranavrc joined #sbcl 2014-03-30T15:30:04Z stassats: time to gdb, it says something interesting 2014-03-30T15:30:47Z stassats: it seems to be segfaulting in _dl_fixup, that's closer to sb-dynamic-core, but doesn't explain anything 2014-03-30T15:31:33Z stassats: hm, maybe it's searching for undefined_tramp and failing 2014-03-30T15:31:48Z stassats: but that seems to be too late to fail there, it would have failed earlier 2014-03-30T15:32:12Z nyef: Does it work without sb-dynamic-core? 2014-03-30T15:32:50Z stassats: it does 2014-03-30T15:33:21Z nyef: And this is a recent breakage? 2014-03-30T15:33:41Z stassats: i haven't tried enabling sb-dynamic-core on ppc before 2014-03-30T15:33:46Z nyef: Ah. 2014-03-30T15:34:15Z stassats: what it does is basically accesses runtime symbols through dlsym, not trough the table generated by nm at compile-time 2014-03-30T15:34:37Z stassats: that allows to recompile the runtime without having to recompile the whole thing 2014-03-30T15:35:28Z nyef: Umm... You can recompile the runtime, genesis-2, target-2-load, and done. 2014-03-30T15:35:53Z stassats: well, you know what i mean 2014-03-30T15:35:56Z nyef: Not as fast, I'll admit, but it's a decent shortcut. 2014-03-30T15:36:47Z stassats: when there was no sb-dynamic-core, i modified target-2 to perform even less work and it took 8 seconds 2014-03-30T15:36:59Z stassats: when i was debugging something in the runtime 2014-03-30T15:38:52Z stassats: maybe it's funcallable_instance_tramp 2014-03-30T15:39:51Z nyef: Oh, hell... 2014-03-30T15:40:37Z nyef: Does sb-dynamic-core even work if the trampoline addresses are tagged? 2014-03-30T15:40:51Z stassats: tagged? 2014-03-30T15:41:21Z nyef: Like on SPARC. 2014-03-30T15:41:42Z stassats: well, it apparently works enough to get to the warm init and compile things up to braid.lisp and fails on loading braid.fasl 2014-03-30T15:42:07Z nyef: Okay, PPC doesn't use tagged tramps, but SPARC does and ARM does. 2014-03-30T15:48:12Z nyef: (Maybe we could/should move the trampolines into read-only space instead of leaving them in the runtime?) 2014-03-30T15:49:07Z nyef: Funcallable_instance_tramp on ppc is tagged. 2014-03-30T15:50:01Z stassats: maybe trampolines are supposed to be accessible from C, that's why they are in x-assem.S and not static routines? 2014-03-30T15:50:07Z stassats: s/static/assembly/ 2014-03-30T15:50:15Z nyef: That said, this isn't a regression, so shouldn't hold up the release. 2014-03-30T15:50:38Z stassats: it shouldn't, i'm readying this for after the freeze 2014-03-30T15:53:38Z nyef: Only direct reference to a trampoline from the runtime is in verify_space() in gencgc.c, for funcallable_instance_tramp. 2014-03-30T15:54:42Z stassats: i think there's something in signal handling 2014-03-30T15:56:33Z Krystof: oh yes, I have to do a release 2014-03-30T15:56:41Z Krystof: I knew that there was something else I was procrastinating 2014-03-30T15:56:50Z nyef: stassats: I did a git grep for the three main tramps. 2014-03-30T16:02:35Z stassats: so, what are the implications of funcallable_instance_tramp being tagged? 2014-03-30T16:03:26Z nyef: It's not called directly by its address, it's treated as a SIMPLE-FUN and invoked by way of a displacement against its address. 2014-03-30T16:04:09Z stassats: that's probably the problem 2014-03-30T16:05:17Z nyef: Should be easy to move to an assembly routine. 2014-03-30T16:05:31Z nyef: Hrm. 2014-03-30T16:05:39Z nyef: Actually, no, that won't work. 2014-03-30T16:05:53Z nyef: The fixup address would be wrong in that case. 2014-03-30T16:06:19Z sdemarre joined #sbcl 2014-03-30T16:07:41Z nyef: I don't have an answer for you, but it's certainly an interesting problem. 2014-03-30T16:08:36Z stassats: maybe the indirect address should just be tagged too? 2014-03-30T16:08:50Z stassats: let me try that 2014-03-30T16:09:55Z stassats: is it just +1? 2014-03-30T16:10:07Z nyef: No, it's something like -23. 2014-03-30T16:10:15Z nyef: From the code itself. 2014-03-30T16:10:26Z nyef: There's the whole simple-fun header thing to consider as well. 2014-03-30T16:11:21Z stassats: that may be more problematic 2014-03-30T16:11:26Z nyef: Yup. 2014-03-30T16:11:45Z nyef: I'm looking for an angle now, but you may have to table this project for a little while. 2014-03-30T16:12:14Z stassats: why is this tagging necessary? 2014-03-30T16:13:12Z stassats: and i know what i can do, disable indirection for funcallable_instance_tramp, that'll make clear that that's problem indeed 2014-03-30T16:13:17Z nyef: I'm not sure in this case, but we need a general solution because SPARC and ARM use tagged undefined_tramp and closure_tramp because it's simpler/faster there. 2014-03-30T16:14:42Z nyef: Hrm. I may have an angle. 2014-03-30T16:18:11Z nyef: Yeah, I have an angle. And I can test that angle on my SPARC, but probably not today. 2014-03-30T16:18:55Z stassats: and it is indeed funcallable_instance_tramp tagging 2014-03-30T16:21:24Z stassats: i can cook up the necessary bits in arch_write_linkage_table_jmp 2014-03-30T16:21:38Z nyef: Okay, I can move funcallable_instance_tramp to an assembly routine without having to alter genesis. 2014-03-30T16:22:49Z nyef: And at one point I had a branch that moved undefined_tramp and closure_tramp to assembly routines as well, but that took a little doing to set up and I'm not quite sure where it got to. 2014-03-30T16:24:44Z stassats: well, you do that, i'll do the table cook up way and we'll see what's better 2014-03-30T16:28:06Z stassats: cooking the table looks really hacky 2014-03-30T16:30:37Z nyef: What I have in mind is also hacky, but it's hacky at build-time not at runtime. 2014-03-30T16:34:47Z stassats: funcallable_instance_tramp not having going through an indirection is a performance improvement too 2014-03-30T16:35:06Z stassats: undefined_tramp isn't really expected to be fast 2014-03-30T16:35:07Z nyef: Yup. Same would apply to closure_tramp. 2014-03-30T16:39:16Z stassats: ok, i don't really like the replication of tagging and fun_headers in the indirection table 2014-03-30T16:40:01Z stassats: it's easy to do, but not really nice 2014-03-30T16:40:48Z nyef: Let me see what I can come up with in terms of assembly-routines. I may be able to nail this one hard enough that it'll work on SPARC as well. 2014-03-30T16:41:26Z nyef: Might take a couple days, though, I'm supposed to be getting some work done. /-: 2014-03-30T16:43:57Z stassats: what's the main difficulty in moving tramps into asm routines? 2014-03-30T16:46:26Z nyef: The two difficulties are that you need to be able to construct a tagged pointer as the entry point (not possible, or at least not easy, with the current assemfile, but easily hacked in), and that undefined_tramp and closure_tramp are both required early in genesis, so you have to load the assem-obj files after creating the static symbols but before creating any fdefns, or something like that. I worked it out once. 2014-03-30T16:47:13Z nyef: The possible third difficulty is getting the overall object structure right for the trampoline SIMPLE-FUNs with respect to the overall code-object. 2014-03-30T16:47:18Z stassats: construct it untagged but change the address afterwards? 2014-03-30T16:49:21Z nyef: Yeah, I have a couple of angles for tweaking the address associated with an assembly-routine. 2014-03-30T16:51:49Z stassats: some addresses in *assembly-routines* do not seem to be even on x86-64 2014-03-30T16:52:11Z nyef: x86oids don't have fixed-width instructions. 2014-03-30T16:52:32Z stassats: but why isn't it aligned? 2014-03-30T16:52:58Z nyef: Because it's simply concatenated. I believe it's possible to align as well, and it will have to be for the trampolines. 2014-03-30T16:54:26Z nyef: But unless there's a bit of work done, the closest you'll come is to split the word containing the tagged address up into octets, emit some of them in one assembly-routine, then the rest in the assembly-routine with the appropriate name for what you're doing. 2014-03-30T16:55:44Z stassats: i wonder how much asm routines for things like generic-negate actually matter on modern CPUs 2014-03-30T16:58:35Z leuler joined #sbcl 2014-03-30T17:16:07Z slyrus quit (Ping timeout: 264 seconds) 2014-03-30T17:16:42Z sdemarre quit (Ping timeout: 255 seconds) 2014-03-30T17:33:16Z sdemarre joined #sbcl 2014-03-30T17:40:36Z stassats quit (Ping timeout: 268 seconds) 2014-03-30T17:40:36Z yacks quit (Ping timeout: 268 seconds) 2014-03-30T18:01:46Z ASau` is now known as ASau 2014-03-30T18:24:34Z pranavrc quit (Quit: Ping timeout: ∞) 2014-03-30T18:54:37Z Krystof: master unfrozen. Go wild. 2014-03-30T19:39:22Z angavrilov quit (Remote host closed the connection) 2014-03-30T20:12:08Z sdemarre quit (Ping timeout: 240 seconds) 2014-03-30T20:24:18Z leuler quit (Quit: ERC Version 5.1.2 $Revision: 1.796.2.6 $ (IRC client for Emacs)) 2014-03-30T20:36:42Z oleo quit (Ping timeout: 265 seconds) 2014-03-30T20:40:48Z prxq__ is now known as prxq 2014-03-30T21:06:20Z oleo joined #sbcl 2014-03-30T21:17:29Z jrm quit (Quit: ERC Version 5.3 (IRC client for Emacs)) 2014-03-30T21:32:30Z ltbarcly joined #sbcl 2014-03-30T21:47:18Z nyef sighs. 2014-03-30T21:47:52Z nyef: Our gray streams implementation is substantially crippled by the compiler fndb. 2014-03-30T21:48:23Z nyef: I've found some three cases thus far where the spec (or other requirements) say one thing, but the fndb says something else. 2014-03-30T21:49:01Z nyef: Two of which are highlighted by merely running gray-streams.impure.lisp. 2014-03-30T21:49:27Z ltbarcly quit (Quit: Computer has gone to sleep.) 2014-03-30T21:53:42Z fe[nl]ix: nyef: I keep getting an annoying warning about the fndb when I specialize on cl:close 2014-03-30T21:53:55Z nyef: Yup! That's one of the three that I'm aware of. 2014-03-30T21:54:41Z nyef: Now, it might be that "all" we need to do is to kill out the fndb entries for these functions, or at least make them more general, but I'm not going to say for certain until I've looked into it in more detail. 2014-03-30T21:55:27Z nyef: I've been bit by a gray stream function returning an integer instead of a boolean (legal, as far as I can tell, as the functions are supposed to be generalized booleans) and then the call site typechecking it. 2014-03-30T21:59:05Z fe[nl]ix: what are the other two ? 2014-03-30T22:01:39Z nyef: open-stream-p and stream-listen. 2014-03-30T22:10:38Z nyef: At any rate, I don't know that those are the only three, but I do know that those three are affected. 2014-03-30T22:36:41Z ltbarcly joined #sbcl 2014-03-30T23:05:05Z irsol quit (Ping timeout: 252 seconds) 2014-03-30T23:07:43Z irsol joined #sbcl 2014-03-30T23:30:12Z ltbarcly quit (Quit: Computer has gone to sleep.) 2014-03-30T23:59:17Z edgar-rft quit (Quit: happening corrupted because everything burns) 2014-03-31T01:08:03Z kludge` quit (Disconnected by services) 2014-03-31T01:08:23Z kludge` joined #sbcl 2014-03-31T01:13:39Z scymtym quit (*.net *.split) 2014-03-31T01:13:39Z tmh_ quit (*.net *.split) 2014-03-31T01:13:39Z prxq quit (*.net *.split) 2014-03-31T01:13:39Z kanru quit (*.net *.split) 2014-03-31T01:14:26Z tmh_ joined #sbcl 2014-03-31T01:15:31Z xymox quit (Changing host) 2014-03-31T01:15:31Z xymox joined #sbcl 2014-03-31T01:31:42Z prxq_ joined #sbcl 2014-03-31T01:54:50Z kanru joined #sbcl 2014-03-31T02:34:14Z ehaliewicz joined #sbcl 2014-03-31T02:36:38Z ehaliewicz quit (Remote host closed the connection) 2014-03-31T02:36:55Z ehaliewicz joined #sbcl 2014-03-31T02:38:45Z christoph_debian quit (Ping timeout: 255 seconds) 2014-03-31T02:39:06Z ltbarcly joined #sbcl 2014-03-31T02:42:39Z prxq_ quit (Ping timeout: 268 seconds) 2014-03-31T02:51:42Z ehaliewicz quit (Remote host closed the connection) 2014-03-31T02:52:01Z christoph_debian joined #sbcl 2014-03-31T02:52:05Z ehaliewicz joined #sbcl 2014-03-31T02:53:11Z ehaliewicz quit (Remote host closed the connection) 2014-03-31T02:54:25Z ehaliewicz joined #sbcl 2014-03-31T02:54:57Z prxq_ joined #sbcl 2014-03-31T03:06:33Z LiamH quit (Quit: Leaving.) 2014-03-31T03:21:08Z nyef: Does SBCL have support for CAS on lexical variables? I'm thinking specifically the case of a closure being marshaled to another thread (or multiple threads), but still used for synchronization. 2014-03-31T03:23:41Z krzysz00: Is pvk's post effectively the proposal template now? 2014-03-31T04:09:10Z kanru quit (Read error: Connection reset by peer) 2014-03-31T04:10:20Z kanru joined #sbcl 2014-03-31T04:16:21Z pkhuong: nyef: no. 2014-03-31T04:16:45Z nyef: Mmm. Somehow, I didn't think so. 2014-03-31T04:17:02Z pkhuong: nyef: cons it is ;) 2014-03-31T04:17:49Z nyef: It's not my use-case, I'm more curious than anything else. 2014-03-31T04:18:15Z pkhuong: krzysz00: it contains my suggestions if you're not sure what to do. You're working on unicode right? You might want to reach out to people who have expressed interest in mentoring that project last year. 2014-03-31T04:18:16Z nyef: On the other hand, I might just add it to my list of things to look into at some point. 2014-03-31T04:21:09Z krzysz00: I wasn't around last year. I know Christophe does a lot of the Unicode work. Is he looking to mentor for GSoC? 2014-03-31T04:22:35Z slyrus joined #sbcl 2014-03-31T04:23:05Z pkhuong: krzysz00: http://search.gmane.org/?query=unicode+gsoc+2013&author=&group=gmane.lisp.steel-bank.devel&sort=relevance&DEFAULTOP=and&query= 2014-03-31T04:26:06Z krzysz00: Should have thought of looking there. 2014-03-31T04:26:28Z krzysz00: (Ow. SF has a quite un-fun archive system.) 2014-03-31T04:27:56Z krzysz00: Is Mr. Emerson on here often? 2014-03-31T04:27:58Z pranavrc joined #sbcl 2014-03-31T04:28:10Z pkhuong: on IRC? I don't think so. 2014-03-31T04:29:18Z krzysz00: Guess I'll send him an email then 2014-03-31T04:29:26Z krzysz00: And get back to work on these revisions. 2014-03-31T04:30:05Z krzysz00: What's the right thing to do about a patch stuck in limbo? 2014-03-31T04:30:40Z pkhuong: Make sure it's on launchpad and ping the -devel mailing list. 2014-03-31T04:32:38Z pranavrc_ joined #sbcl 2014-03-31T04:32:51Z krzysz00: It's on Launchpad. I'll ping for the non-controversial one, at least 2014-03-31T04:36:44Z ltbarcly_ joined #sbcl 2014-03-31T04:39:53Z prxq_ quit (*.net *.split) 2014-03-31T04:39:54Z pranavrc quit (*.net *.split) 2014-03-31T04:39:54Z ltbarcly quit (*.net *.split) 2014-03-31T04:46:07Z prxq_ joined #sbcl 2014-03-31T04:46:56Z drmeister quit (Remote host closed the connection) 2014-03-31T04:48:18Z pranavrc_ quit (Read error: Connection reset by peer) 2014-03-31T04:49:41Z pranavrc joined #sbcl 2014-03-31T04:49:45Z pranavrc quit (Changing host) 2014-03-31T04:49:45Z pranavrc joined #sbcl 2014-03-31T04:54:15Z ltbarcly_ is now known as ltbarcly 2014-03-31T04:56:11Z krzysz00: Apperently, there's `sb-impl::char-decomposition`, which no one ever calls or anything. 2014-03-31T04:57:30Z ltbarcly quit (Quit: Computer has gone to sleep.) 2014-03-31T04:59:28Z nyef quit (Quit: G'night all.) 2014-03-31T05:10:59Z prxq_ is now known as prxq 2014-03-31T05:16:09Z yacks joined #sbcl 2014-03-31T05:29:15Z oleo quit (Quit: Leaving) 2014-03-31T05:51:23Z michael_lee joined #sbcl 2014-03-31T06:03:38Z krzysz00 quit (Ping timeout: 265 seconds) 2014-03-31T06:10:57Z angavrilov joined #sbcl 2014-03-31T06:10:58Z sdemarre joined #sbcl 2014-03-31T06:33:00Z sdemarre quit (Quit: Leaving.) 2014-03-31T06:35:45Z drmeister joined #sbcl 2014-03-31T06:37:07Z ASau quit (Ping timeout: 264 seconds) 2014-03-31T06:40:23Z drmeister quit (Ping timeout: 265 seconds) 2014-03-31T06:59:02Z edgar-rft joined #sbcl 2014-03-31T07:15:59Z harovali joined #sbcl 2014-03-31T07:17:35Z harovali: hi, I've just downloaded sbcl-1.1.17-source.tar.bz2, untarred it, and tried to install it as root, sourcing install.sh. It gives this error signaling a fila which in theory should be in the tarball, but it's not: bash: output/prefix.def: No such file or directory 2014-03-31T07:31:01Z Krystof: if you've downloaded the source, you need to build it first 2014-03-31T07:31:08Z Krystof: ./make.sh is your friend 2014-03-31T07:32:40Z Krystof: minion: memo for krzysz00: why do you say char-decomposition is never called? It's used in implementing normalization forms 2014-03-31T07:32:41Z minion: Remembered. I'll tell krzysz00 when he/she/it next speaks. 2014-03-31T07:35:18Z yacks quit (Ping timeout: 255 seconds) 2014-03-31T07:52:21Z harovali: thanks Krystof 2014-03-31T07:57:37Z pranavrc quit (Remote host closed the connection) 2014-03-31T08:04:42Z yacks joined #sbcl 2014-03-31T08:35:47Z edgar-rft quit (Quit: lifeform experiment lost because of unknown reasons) 2014-03-31T08:49:33Z irsol quit (Read error: Operation timed out) 2014-03-31T08:50:49Z irsol joined #sbcl 2014-03-31T08:55:24Z nicdev quit (Ping timeout: 252 seconds) 2014-03-31T08:56:06Z nicdev joined #sbcl 2014-03-31T08:57:14Z luis` quit (Ping timeout: 252 seconds) 2014-03-31T08:58:35Z pranavrc joined #sbcl 2014-03-31T09:00:02Z luis joined #sbcl 2014-03-31T09:03:31Z pranavrc quit (Ping timeout: 264 seconds) 2014-03-31T09:04:54Z kanru quit (Ping timeout: 265 seconds) 2014-03-31T09:13:44Z brucem: How do y'all handle symbol versioning stuff when calling libc stuff? Like on OS X, you should call readdir$INODE64() rather than just plain readdir() if you want access to the 64 bit inodes? 2014-03-31T09:46:46Z sdemarre joined #sbcl 2014-03-31T09:49:57Z pranavrc joined #sbcl 2014-03-31T10:05:53Z ams: brucem: on gnu it is handled by the linker. 2014-03-31T10:07:05Z brucem: ams: entirely? is there some magic linker script that tells the linker what to do? I know that some of it is set up in the libc headers, so some stuff happens at compile time in gcc / clang. 2014-03-31T10:07:49Z ams: brucem: entierly, though in the case of readdir(); that is wraped using a macro in one of the headers based on if you are using _LARGE_FILE or not. 2014-03-31T10:08:11Z ams: brucem: glibc has version specific information that the linker uses 2014-03-31T10:08:31Z brucem: so how does sbcl handle that for stuff in the headers? 2014-03-31T10:08:50Z brucem: I just ran into a fun instance of this breaking user code in Open Dylan which is why I wonder how sbcl handles it. 2014-03-31T10:08:58Z ams: why would sbcl handle it at all? 2014-03-31T10:09:43Z ams: if sbcl needs to call readdir(), it calls readdir ... not __readdir64, readdir$INODE64 or some other crazy thing 2014-03-31T10:10:10Z brucem: then on OS X, it would do the wrong thing if you wanted 64 bit inodes. 2014-03-31T10:10:35Z brucem: since you need the magic in the headers to have the compiler DTRT 2014-03-31T10:11:06Z ams: it is up to osx to use the right version of readdir if you want 64-bit file support or not; or you explicitly specify that you want 64-bit via some _LARGE_FILE macro 2014-03-31T10:11:34Z brucem: okay, you clearly haven't worked on this. 2014-03-31T10:11:40Z ams: uhm, yeah i have 2014-03-31T10:11:49Z brucem: I'll wait for an sbcl maintainer to come forward. 2014-03-31T10:11:57Z ams: this has nothing to do with sbcl 2014-03-31T10:13:25Z drmeister joined #sbcl 2014-03-31T10:13:28Z ams: it is the operating system libraries and compilers task to decide on wether to use 64-bit or whatever representation for file offests, and what not, you as a application developer do not need to care. 2014-03-31T10:13:40Z brucem: ams: sbcl is a compiler. 2014-03-31T10:14:10Z ams: brucem: which is calling C functions. 2014-03-31T10:14:29Z ams: if you compile sbcl on a system where readdir happens to be readdir64; that is what you will use. 2014-03-31T10:14:40Z ams: if you compile it on a system that uses readdir7, then that is what it will use. 2014-03-31T10:15:00Z brucem: okay, I go back to my earlier statement and will wait. 2014-03-31T10:15:35Z ams: or just read the source code 2014-03-31T10:15:55Z ams: or read the large file support document, or posix. 2014-03-31T10:17:32Z brucem: There's no C involved. There's just raw machine code being generated by the compiler (be it SBCL or Open Dylan). There aren't headers being parsed. There aren't macros being enabled to set up the right things when parsing the headers. 2014-03-31T10:17:36Z drmeister quit (Ping timeout: 240 seconds) 2014-03-31T10:19:34Z ams: of course there is C invovled 2014-03-31T10:20:16Z ams: just look at src/runtime/wrap.c 2014-03-31T10:21:15Z brucem: ahh, that's precisely the answer I was looking for. They do wrap everything. 2014-03-31T10:21:19Z brucem: We don't. 2014-03-31T10:21:47Z ams: sbcl sure wraps; i don't know who we is. 2014-03-31T10:22:10Z brucem: We lean on our C-FFI much harder. (Clearly: we is Open Dylan as I said above) 2014-03-31T10:22:17Z ams: and then in src/code/unix.lisp you export the interface to lisp 2014-03-31T10:22:36Z ams: brucem: same thing with sbcl; see code/unix.lisp 2014-03-31T10:22:51Z kanru joined #sbcl 2014-03-31T10:24:00Z ams: the wrap stuff is just useful to work around portability issues (like look at one of the functions above made to work with windows) 2014-03-31T10:24:03Z brucem: ams: right, they FFI to their C wrappers. We lacked the C wrappers for many things, which led to the problem that I found today and the question of how they handle symbol versioning / decoration issues. They wrap it all in C to let it happen there. 2014-03-31T10:24:20Z ams: brucem: they could ffi directly to the standard libc 2014-03-31T10:24:32Z ams: s/they/sbcl/ 2014-03-31T10:24:40Z ams: there is nothing that stops that 2014-03-31T10:24:53Z brucem: Sure. They'll just get old functionality. 2014-03-31T10:25:00Z ams: (at least in this case; for some other functions it might since they are not guaranteed to be functions; but can be macros and so) 2014-03-31T10:26:46Z brucem: on OS X, by FFI'ing directly to readdir(), you'll get the old 32 bit inodes. (Because this is what happened to me and I tracked it down to this.) 2014-03-31T10:26:57Z brucem: but wrap it in C and all is well with the world. 2014-03-31T10:27:03Z ams: did you set any large_file stuff? 2014-03-31T10:27:44Z brucem: Don't need that for readdir() on OS X. just build for 10.6 or later (we already require 10.7) 2014-03-31T10:27:45Z ams: you needed to do that on gnu back in the day ... _LARGE_FILE_ or something .. 2014-03-31T10:29:32Z brucem: ams: from /usr/include/dirent.h, what happens is that if you compile with C, you'll see this prototype: struct dirent *readdir(DIR *) __DARWIN_INODE64(readdir); ... that aliases it to readdir$INODE64, but there's still an old-style readdir() around that is 32 bit inodes. So if you just FFI to the libc, you'll get the old behavior. 2014-03-31T10:30:17Z ams: i would just write a wrapper 2014-03-31T10:31:00Z brucem: Yeah, that's what I'm doing ... I just wanted to know how SBCL dealt with it. 2014-03-31T10:31:06Z brucem: (And clearly now: they wrap it) 2014-03-31T10:31:10Z ams: it just calls readdir :-) 2014-03-31T10:31:21Z ams: not really wrapping it either... 2014-03-31T10:31:32Z ams: it just calls readdir() and lets the system C compiler figure out if it is 64/32 2014-03-31T11:05:21Z krzysz00 joined #sbcl 2014-03-31T11:15:53Z ehaliewicz quit (Ping timeout: 255 seconds) 2014-03-31T12:02:09Z drmeister joined #sbcl 2014-03-31T12:06:54Z drmeister quit (Ping timeout: 268 seconds) 2014-03-31T12:10:47Z segv- joined #sbcl 2014-03-31T12:27:18Z drmeister joined #sbcl 2014-03-31T12:28:20Z jdz joined #sbcl 2014-03-31T12:34:47Z drmeister quit (Remote host closed the connection) 2014-03-31T12:36:41Z psilord quit (Quit: Leaving.) 2014-03-31T13:01:36Z pranavrc quit 2014-03-31T13:04:59Z eudoxia joined #sbcl 2014-03-31T13:11:07Z fe[nl]ix: brucem: I wrote a library that wraps the correct functions, https://github.com/sionescu/libfixposix 2014-03-31T13:14:06Z ams: i would recommend gnulib or similar if one really wants sanity across non-posix compliant systems; or where you want to use a specific version of posix 2014-03-31T13:18:56Z fe[nl]ix: libfixposix is similar to gnulib, just not gpl 2014-03-31T13:29:33Z yacks quit (Ping timeout: 255 seconds) 2014-03-31T13:38:03Z cmack joined #sbcl 2014-03-31T13:38:24Z drmeister joined #sbcl 2014-03-31T13:38:52Z ams: parts of gnulib can be used under non-gpl licenses. 2014-03-31T13:38:55Z ams: or lgpl 2014-03-31T13:39:58Z attila_lendvai joined #sbcl 2014-03-31T13:39:58Z attila_lendvai quit (Changing host) 2014-03-31T13:39:58Z attila_lendvai joined #sbcl 2014-03-31T13:43:27Z asedeno joined #sbcl 2014-03-31T13:47:19Z drmeister quit (Remote host closed the connection) 2014-03-31T13:52:52Z pkhuong: brucem: yeah, there's a lot of strange backward compatiblity stuff. I've also seen struct fields defined with CPP macros (still not sure that's posix). Letting a real C compiler figure it out is the way to go, imo. 2014-03-31T14:01:18Z yacks joined #sbcl 2014-03-31T14:01:27Z psilord joined #sbcl 2014-03-31T14:04:46Z attila_lendvai quit (Quit: Leaving.) 2014-03-31T14:07:18Z oleo joined #sbcl 2014-03-31T14:18:20Z brucem: fe[nl]ix: in this case, it looks like lfp doesn't wrap enough actually to avoid the issue that I had :) 2014-03-31T14:18:46Z brucem: pkhuong: yeah ... and in some cases, gnu libc made poor decisions ... alas. 2014-03-31T14:19:05Z brucem: pkhuong: I've worked a fair bit with musl, but I can't go and replace the libc. :) And I have a fix for my issues to date... 2014-03-31T14:24:15Z tmh_ quit (Changing host) 2014-03-31T14:24:15Z tmh_ joined #sbcl 2014-03-31T14:39:24Z foom: musl will eventually have the same behavior, even if they don't today. 2014-03-31T14:39:47Z foom: changing the function's linker symbol name is how you keep backwards compatibility with old programs 2014-03-31T14:41:26Z brucem: foom: yeah ... they just made a few better decisions about things like large files. 2014-03-31T14:41:44Z brucem: foom: and I mainly just used their headers... I replaced the libc headers in emscripten with them on contract last year. 2014-03-31T14:42:18Z foom: Well, they made *later* decisions about such things, which sure makes it easier. ;) 2014-03-31T14:42:27Z brucem: foom: btw, haven't gotten a response out of jrb about the Goo CVS history. :( 2014-03-31T14:42:53Z nyef joined #sbcl 2014-03-31T14:43:06Z nyef: G'morning all. 2014-03-31T14:43:16Z foom: ah. too bad dvcsen weren't popular yet. 2014-03-31T14:45:33Z drmeister joined #sbcl 2014-03-31T14:50:19Z drmeister quit (Ping timeout: 264 seconds) 2014-03-31T15:27:51Z drmeister joined #sbcl 2014-03-31T15:32:53Z krzysz00 quit (Quit: leaving) 2014-03-31T15:41:14Z drmeister quit (Remote host closed the connection) 2014-03-31T15:59:47Z attila_lendvai joined #sbcl 2014-03-31T15:59:47Z attila_lendvai quit (Changing host) 2014-03-31T15:59:47Z attila_lendvai joined #sbcl 2014-03-31T16:02:58Z harovali quit (Ping timeout: 265 seconds) 2014-03-31T16:03:53Z DGASAU quit (Ping timeout: 252 seconds) 2014-03-31T16:04:15Z harovali joined #sbcl 2014-03-31T16:09:06Z ebrasca joined #sbcl 2014-03-31T16:09:29Z drmeister joined #sbcl 2014-03-31T16:21:02Z drmeiste_ joined #sbcl 2014-03-31T16:22:12Z drmeister quit (Ping timeout: 268 seconds) 2014-03-31T16:35:30Z hlavaty joined #sbcl 2014-03-31T16:36:24Z ltbarcly joined #sbcl 2014-03-31T16:39:30Z ltbarcly_ joined #sbcl 2014-03-31T16:41:04Z ltbarcly quit (Ping timeout: 252 seconds) 2014-03-31T17:03:50Z yacks quit (Remote host closed the connection) 2014-03-31T17:05:29Z edgar-rft joined #sbcl 2014-03-31T17:11:13Z drmeister joined #sbcl 2014-03-31T17:14:15Z drmeiste_ quit (Ping timeout: 240 seconds) 2014-03-31T17:24:53Z DGASAU joined #sbcl 2014-03-31T17:33:01Z ltbarcly joined #sbcl 2014-03-31T17:34:46Z ltbarcly_ quit (Ping timeout: 245 seconds) 2014-03-31T17:42:37Z drmeister quit (Read error: Connection reset by peer) 2014-03-31T17:42:47Z drmeister joined #sbcl 2014-03-31T17:43:01Z eudoxia quit (Quit: leaving) 2014-03-31T17:43:12Z eudoxia joined #sbcl 2014-03-31T17:48:56Z eudoxia quit (Quit: leaving) 2014-03-31T17:51:48Z kanru quit (Ping timeout: 255 seconds) 2014-03-31T17:52:59Z ltbarcly quit (Quit: Computer has gone to sleep.) 2014-03-31T17:58:48Z ltbarcly joined #sbcl 2014-03-31T17:59:48Z ltbarcly quit (Client Quit) 2014-03-31T18:02:13Z eudoxia joined #sbcl 2014-03-31T18:06:44Z drmeiste_ joined #sbcl 2014-03-31T18:09:55Z drmeister quit (Ping timeout: 240 seconds) 2014-03-31T18:12:58Z redline6561 quit (Ping timeout: 240 seconds) 2014-03-31T18:13:03Z ltbarcly joined #sbcl 2014-03-31T18:26:52Z ASau joined #sbcl 2014-03-31T18:57:16Z eudoxia quit (Quit: leaving) 2014-03-31T19:13:06Z Bicyclidine joined #sbcl 2014-03-31T19:13:13Z angavrilov quit (Remote host closed the connection) 2014-03-31T19:18:50Z redline6561 joined #sbcl 2014-03-31T19:33:38Z harovali quit (Ping timeout: 246 seconds) 2014-03-31T19:55:33Z drmeister joined #sbcl 2014-03-31T19:59:08Z sdemarre quit (Ping timeout: 240 seconds) 2014-03-31T19:59:16Z drmeiste_ quit (Ping timeout: 268 seconds) 2014-03-31T20:29:23Z ltbarcly quit (Quit: Computer has gone to sleep.) 2014-03-31T20:50:02Z harovali joined #sbcl 2014-03-31T20:50:22Z cmack quit (Remote host closed the connection) 2014-03-31T21:27:08Z attila_lendvai quit (Ping timeout: 240 seconds) 2014-03-31T21:28:16Z attila_lendvai joined #sbcl 2014-03-31T21:28:25Z ltbarcly joined #sbcl 2014-03-31T21:31:32Z oleo quit (Read error: Operation timed out) 2014-03-31T21:31:53Z oleo joined #sbcl 2014-03-31T21:32:03Z attila_lendvai1 joined #sbcl 2014-03-31T21:32:03Z attila_lendvai quit (Disconnected by services) 2014-03-31T21:32:03Z attila_lendvai1 quit (Changing host) 2014-03-31T21:32:03Z attila_lendvai1 joined #sbcl 2014-03-31T21:50:06Z ehaliewicz joined #sbcl 2014-03-31T21:55:48Z attila_lendvai1 quit (Read error: Operation timed out) 2014-03-31T21:56:41Z edgar-rft quit (Quit: continuation lost because no existence expected) 2014-03-31T21:58:52Z drmeister quit (Remote host closed the connection) 2014-03-31T22:00:09Z psilord quit (Quit: Leaving.) 2014-03-31T22:10:11Z michael_lee quit (Remote host closed the connection) 2014-03-31T22:16:07Z slyrus quit (Ping timeout: 264 seconds) 2014-03-31T22:22:42Z segv- quit (Ping timeout: 255 seconds) 2014-03-31T22:23:39Z drmeister joined #sbcl 2014-03-31T22:24:17Z ltbarcly quit (Quit: Computer has gone to sleep.) 2014-03-31T22:28:11Z drmeister quit (Read error: Connection reset by peer) 2014-03-31T22:35:11Z drmeister joined #sbcl 2014-03-31T22:43:00Z eudoxia joined #sbcl 2014-03-31T22:45:39Z drmeister quit (Read error: Connection reset by peer) 2014-03-31T23:01:30Z drmeister joined #sbcl 2014-03-31T23:20:13Z psilord joined #sbcl