00:30:09 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 276 seconds] 00:41:52 -!- edgar-rft [~GOD@HSI-KBW-149-172-63-75.hsi13.kabel-badenwuerttemberg.de] has quit [Quit: nuclear meltdown] 01:00:25 momo-reina [~user@173.244.199.112] has joined #sbcl 01:02:14 -!- foreignFunction [~niksaak@ip-4761.sunline.net.ua] has quit [Quit: Leaving.] 01:04:15 -!- momo-reina [~user@173.244.199.112] has quit [Remote host closed the connection] 01:23:52 momo-reina [~user@173.244.199.112] has joined #sbcl 01:28:57 -!- wbooze [~wbooze@xdsl-78-35-144-234.netcologne.de] has quit [Read error: Connection reset by peer] 02:27:52 Bike_ [~Glossina@65-102-1-250.ptld.qwest.net] has joined #sbcl 02:28:21 -!- Bike [~Glossina@67-5-235-38.ptld.qwest.net] has quit [Ping timeout: 248 seconds] 02:31:03 -!- Bike_ is now known as Bike 02:32:30 -!- momo-reina [~user@173.244.199.112] has quit [Remote host closed the connection] 02:48:58 Bike_ [~Glossina@67-5-235-75.ptld.qwest.net] has joined #sbcl 02:49:06 -!- davazp [~user@178.167.157.176.threembb.ie] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 02:51:48 -!- Bike [~Glossina@65-102-1-250.ptld.qwest.net] has quit [Ping timeout: 264 seconds] 02:52:22 -!- Bike_ is now known as Bike 03:08:41 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 04:28:49 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Read error: Connection reset by peer] 04:33:10 -!- drmeister [~drmeister@pool-173-59-25-70.phlapa.fios.verizon.net] has quit [Remote host closed the connection] 05:19:29 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 05:37:06 sdemarre [~serge@45.146-66-87.adsl-dyn.isp.belgacom.be] has joined #sbcl 05:57:30 -!- Snamich [~Snamich@71-9-62-86.dhcp.snlo.ca.charter.com] has quit [Quit: Snamich] 06:14:56 -!- drdo [~user@85.207.54.77.rev.vodafone.pt] has quit [Ping timeout: 260 seconds] 06:27:08 prxq [~mommer@mnhm-590c24dc.pool.mediaWays.net] has joined #sbcl 06:34:35 pranavrc [~pranavrc@122.164.134.47] has joined #sbcl 06:34:35 -!- pranavrc [~pranavrc@122.164.134.47] has quit [Changing host] 06:34:35 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 06:46:52 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 06:50:24 attila_lendvai [~attila_le@92.47.255.58] has joined #sbcl 06:50:25 -!- attila_lendvai [~attila_le@92.47.255.58] has quit [Changing host] 06:50:25 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 07:15:07 I believe that the next GSoC step is for us as an organization to request "awesome" and "blue-sky" slot counnts 07:15:37 that is, what is the minimum number that will not make us cry, and what is the maximum number that will not make us fall over from the weight of mentoring responsibilities 07:18:30 I'll listen to opinions from mentor-types all day :) 07:21:15 wbooze [~wbooze@xdsl-87-79-195-248.netcologne.de] has joined #sbcl 07:23:17 inklesspen [~jon-freen@inklesspen.com] has joined #sbcl 07:24:25 I'm using debian and the version of sbcl in the package manager is 1.0.57; I see on the website that the latest is 1.1.7 07:24:51 are there enough improvements and bugfixes that it's worth my time to get sbcl from outside the package manager? 07:27:12 I seem to be using a 32-bit version of debian as well 07:28:11 -!- Bike [~Glossina@67-5-235-75.ptld.qwest.net] has quit [Ping timeout: 260 seconds] 07:33:01 edgar-rft [~GOD@HSI-KBW-149-172-63-75.hsi13.kabel-badenwuerttemberg.de] has joined #sbcl 07:55:45 inklesspen: it depends on what you want to do 07:55:56 write a relatively simple webapp 07:56:29 then I would say that as long as the libraries/frameworks/whatever that you're using work OK, and you're comfortable with what you've got, don't worry 07:57:49 ok 07:59:57 the first time you get a truly weird bug that looks like it is a compiler error, you might want to cheeck whether you see the same problem with the latest upstream. Until then, have fun 08:10:44 thanks 08:10:53 -!- inklesspen [~jon-freen@inklesspen.com] has left #sbcl 08:36:59 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Quit: Ping timeout: ] 08:39:11 -!- pchrist [~spirit@gentoo/developer/pchrist] has quit [Quit: leaving] 09:12:57 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 09:15:50 pchrist [~spirit@gentoo/developer/pchrist] has joined #sbcl 09:53:03 drmeister [~drmeister@pool-173-59-25-70.phlapa.fios.verizon.net] has joined #sbcl 09:56:56 -!- wbooze [~wbooze@xdsl-87-79-195-248.netcologne.de] has quit [Quit: none] 09:58:23 -!- drmeister [~drmeister@pool-173-59-25-70.phlapa.fios.verizon.net] has quit [Remote host closed the connection] 09:58:47 wbooze [~wbooze@xdsl-87-79-195-248.netcologne.de] has joined #sbcl 10:09:39 foreignFunction [~niksaak@ip-4761.sunline.net.ua] has joined #sbcl 10:17:30 Strigoides [~owen@114-134-0-102.lightwire.co.nz] has joined #sbcl 10:35:21 stassats [~stassats@wikipedia/stassats] has joined #sbcl 10:38:33 -!- Strigoides [~owen@114-134-0-102.lightwire.co.nz] has quit [Quit: leaving] 11:00:57 pranavrc [~pranavrc@122.174.17.101] has joined #sbcl 11:00:58 -!- pranavrc [~pranavrc@122.174.17.101] has quit [Changing host] 11:00:59 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 11:04:30 Krystof: how many mentors? (: 11:07:49 that's definitely one limiting factor 11:08:08 there are 8ish signed up -- but whether they are suitable for the student projects we have on offer is another question 11:19:46 -!- sdemarre [~serge@45.146-66-87.adsl-dyn.isp.belgacom.be] has quit [Ping timeout: 256 seconds] 11:32:02 -!- yacks [~py@180.151.36.168] has quit [Quit: Leaving] 11:40:13 davazp [~user@178.167.241.9.threembb.ie] has joined #sbcl 11:41:37 I'm trying to load a script with sbcl that uses sb:posix. Cannot do this without sudo, though. How I can put this command as a daemon into supervisord? 11:42:28 I'm pretty sure you can use sb-posix without root access. 11:43:20 I get this error on trying to load the script without sudo: Error in SB-POSIX:INITGROUPS: Operation not permitted (1) 11:43:23 Any ideas? 11:44:30 -!- antoszka [~antoszka@unaffiliated/antoszka] has quit [Ping timeout: 264 seconds] 11:44:34 The backtrace shows a syscall-error followed by the change-user function 11:44:48 antoszka [~antoszka@unaffiliated/antoszka] has joined #sbcl 11:46:19 sb-posix has nothing to do with it 11:46:36 just don't execute things to which you don't have permission 11:48:13 and sb-posix:initgroups? where's that coming from? 11:52:02 It's in the daemon script. I'm trying to deploy a restas app. https://github.com/archimag/restas/blob/master/contrib/restas-daemon.lisp#L202 11:52:09 Also, the user has complete root access 11:52:24 well then, you should ask the author of that script 11:53:00 Guess I'll do that. Thanks :) 11:59:41 -!- wbooze [~wbooze@xdsl-87-79-195-248.netcologne.de] has quit [Remote host closed the connection] 12:05:04 wbooze [~wbooze@xdsl-87-79-195-248.netcologne.de] has joined #sbcl 12:29:40 Krystof: I'm not sure how we can give a number before potential mentors check in on the projects they're interested in. 12:36:01 yacks [~py@180.151.36.168] has joined #sbcl 12:40:33 sdemarre [~serge@45.146-66-87.adsl-dyn.isp.belgacom.be] has joined #sbcl 13:05:44 -!- sdemarre [~serge@45.146-66-87.adsl-dyn.isp.belgacom.be] has quit [Ping timeout: 256 seconds] 13:06:28 -!- yacks [~py@180.151.36.168] has quit [Read error: Operation timed out] 13:23:15 pkhuong: I've pinged the mps people to request feedback 13:23:26 nice. 13:24:13 so, I'm hunting down an IR1 optimisation issue at ITA, and I'm looking at specialising if/if conversion when the NULL test can be optimised away. 13:24:58 *stassats* doesn't get why the values-list vop has to manually do fixnum tagging 13:25:17 if i change the result to unsigned-reg, it says # is not valid as the second result to VOP: VALUES-LIST Primitive type: FIXNUM SC restrictions: (SB-VM::UNSIGNED-REG) 13:26:13 stassats: my guess would be that it's because the vop is called by name rather than through any kind of translate mechanism 13:26:22 it is 13:27:02 but i don't know why is that restricted to fixnum 13:27:07 and at this point, we're doing The Trick. Do we also want to do it for foldable combinations? What if the destination is known, but some arguments are variable still? 13:28:47 do we think that the PRNG project is leuler's? Or mine? 13:28:49 actually, it does tagging manually because it also performs division 13:28:55 (or both)? 13:29:16 and there probably isn't a good way to express it to optimize SHR RCX, 2 CMP RCX, 2 away 13:29:55 though, value-list can return start and end, instead of start and count, and then the comparisons could be simpler 13:29:57 Krystof: leuler can't be the primary mentor but is interested in being CCed and helping when he can 13:30:42 stassats: CMP RCX, 11? 13:30:54 pkhuong: they are in different vops 13:31:41 re values list, I think you'd have to create temporary TNs and explicitly insert a MOVE VOP. 13:34:52 stassats: for code size, ISTM moving values-list's error sequence to elsewhere would be pretty effective. 13:40:22 from the way that the gdocs are framing the issue, asking for more than 2 "awesome projects" is likely to be seen as naïve 13:40:32 as it happens there are two fairly clear standouts 13:41:47 I think so too 13:44:03 *stassats* is having a hard time finding the sources of the return-multiple assembly-routine 13:45:07 stassats: src/assembly/x86-64/assem-rtns.lisp 13:45:22 it seems to be conditionalized out 13:46:13 no, #+sb-assembling is a horrible hack 13:46:18 it is. read the condition. 13:46:58 "It's temporarily pushed onto the *FEATURES*" 13:46:59 i see 13:47:28 why, though? to save space? 13:48:15 yacks [~py@180.151.36.168] has joined #sbcl 13:48:33 probably for some reason that made sense in 1990 13:49:28 sbcl-internals cliki is still down... 13:49:47 *stassats* should find another place to put all these little bits 13:49:54 ok, here's another way of thinking about it: the people who can really plausibly be active sbcl mentors in the google sense over the summer are, as far as I can tell, pkhuong, me and the MPS team 13:50:03 (am I forgetting anything?) 13:50:20 guilting jsnell or froydnj? 13:50:42 possible 13:51:01 though they have Real Jobs 13:51:03 but yeah, ISTM the binding constraint is the number of mentors. 13:52:06 I guess you and I could at a stretch do two each? (If we did have other more available mentors I think I'd want to spread students around rather than take more on) 13:52:28 no. I don't see myself doing more than one and a half. 13:53:55 redefining return-multiple at run-time is not a good idea... 13:54:40 probably not (: 13:54:54 ah, happy memories of redefining %coerce-to-function 13:56:09 sb-eval's got me out of pretty deep holes though. Who knew the main benefit of having an interpreter would be easier compiler hacking? :) 13:56:31 ok, so 1½ (pkhuong) + 1½ (Krystof) + 1 (mps) + 1 (jsnell/froydnj/reb/ANOther) with strong preferences to reduce the halves to other ANOthers -- I still can't see how we could possibly support more than 5 13:57:04 so the only question then is if we should request fewer than 5 "blue sky" slots. I guess if the MPS people say that they don't want to take any of the students on 13:57:54 or if we don't have five viable projects 13:58:16 there's that (: 13:58:35 is tagging needed when calling to assembly routines? 13:58:43 did you just change a score, pkhuong? I could have sworn there was a 4 in there somewhere 13:58:46 sdemarre [~serge@45.146-66-87.adsl-dyn.isp.belgacom.be] has joined #sbcl 13:59:15 yes, I did. 13:59:37 Krystof: I think it was one of the first score I gave. 13:59:57 right-oh 14:00:56 (I'm not missing a useful button like "show all of these proposals in one big page", am I?) 14:00:57 or put another way, is the number of return values has to be tagged? 14:02:00 Krystof: unfortunately, no. I made a tarball of the pdfs. 14:02:18 return-multiple seems to jump through a lot of code to return 1 value, which, i assume, is the most frequent case 14:03:07 stassats: why would that be? 14:03:27 would what be? 14:03:54 the most frequent case 14:04:21 because it's just 1 value 14:06:00 stassats: ok? but known single-value returns don't go through return-multiple. 14:07:51 without tail calls and on user defined functions, it'll go through return-multiple 14:08:37 no? 14:08:46 yes? 14:09:37 you mean tail calls with cleanups? 14:10:18 i use debug 2 all the time, so no tail calls for me, unless SPEED is high 14:12:00 ah sure. 14:12:18 and making debuggable code faster without compromising debuggability is a nice goal 14:12:26 I don't think they had that in mind when they wrote return-multiple 14:13:42 Change (policy node (> space speed)) ? 14:14:12 maybe something like separate debugging symbols would be beneficial 14:16:52 it would take more space, but needed only when debugging 14:17:12 though it's probably quite a complicated task 14:18:29 the VOP I see here for return-multiple checks for single-value return before calling to the assembly routine 14:19:14 forgot about that part, right 14:19:51 so, then return-multiple shouldn't check for one? although that check is only for speed >= space 14:21:06 -!- wbooze [~wbooze@xdsl-87-79-195-248.netcologne.de] has quit [Ping timeout: 256 seconds] 14:21:58 returning single value takes so little code, it could be emitted unconditionally 14:22:20 if they're really adamant about saving space... 14:23:51 leuler [~user@p548FC232.dip0.t-ipconnect.de] has joined #sbcl 14:34:42 wbooze [~wbooze@xdsl-78-35-161-109.netcologne.de] has joined #sbcl 14:41:43 do unused registers for values passing have to be set with NIL? 14:46:07 yes 14:46:20 for the gc? 14:46:44 i would assume that the count would be enough for other code 14:47:44 we assume that registers are defaulted to NIL 14:48:14 (in non-single value return contexts) 15:50:43 Bike [~Glossina@63-230-172-89.ptld.qwest.net] has joined #sbcl 16:08:35 Krystof: if you see a mentor application from edwardk, it was a mistake (: 16:13:12 ok! Interesting mistake 16:13:14 Krystof: not a mistake, actually. 16:13:21 ok, interesting not mistake 16:13:25 now I am totally confused :-) 16:13:56 He's done gsoc a couple times, and might be able to help out. 16:14:59 does he have any sbcl experience? 16:16:36 Not that much... not as much as ghc experience, certainly. 16:16:55 (he also suggests that having more way more mentors on board than slots is a good signal ;) 16:19:05 well, OK, let's game the system a little bit, then 16:23:33 danlentz [~danlentz@c-68-37-70-235.hsd1.nj.comcast.net] has joined #sbcl 17:08:09 -!- yacks [~py@180.151.36.168] has quit [Quit: Leaving] 17:33:16 Snamich [~Snamich@71-9-62-86.dhcp.snlo.ca.charter.com] has joined #sbcl 17:47:24 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Ping timeout: 258 seconds] 17:56:39 gabnet [~gabnet@ACaen-652-1-284-216.w92-154.abo.wanadoo.fr] has joined #sbcl 17:59:35 yacks [~py@180.151.36.168] has joined #sbcl 18:32:49 -!- wbooze [~wbooze@xdsl-78-35-161-109.netcologne.de] has quit [Quit: none] 18:33:38 -!- Bike [~Glossina@63-230-172-89.ptld.qwest.net] has quit [Quit: Reconnecting] 18:33:55 Bike [~Glossina@63-230-172-89.ptld.qwest.net] has joined #sbcl 18:48:37 -!- gabnet [~gabnet@ACaen-652-1-284-216.w92-154.abo.wanadoo.fr] has quit [Quit: Ex-Chat] 19:16:16 can someone confirm whether "raito"s sbcl-devel report on C-c happens on non-ancient Windows? Unfortunately, recent SBCL on recent Windows is the only combination he doesn't seem to have tested. 19:18:56 i seem to remember C-c exiting sbcl for me, on windows 7 19:19:21 *stassats* is not too anxious to reboot to check 19:20:20 -!- robgssp [~user@2620:8d:8000:e50:21c:c0ff:fea3:7cc5] has quit [Remote host closed the connection] 19:21:11 robgssp [~user@2620:8d:8000:e50:f2de:f1ff:fe98:8a47] has joined #sbcl 19:26:13 if its exits, it's just too old an SBCL 19:51:02 wbooze [~wbooze@xdsl-78-35-161-109.netcologne.de] has joined #sbcl 19:55:21 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 20:06:12 -!- sdemarre [~serge@45.146-66-87.adsl-dyn.isp.belgacom.be] has quit [Ping timeout: 256 seconds] 20:23:06 drmeister [~drmeister@mobile-198-228-194-066.mycingular.net] has joined #sbcl 20:25:03 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 20:38:00 -!- drmeister [~drmeister@mobile-198-228-194-066.mycingular.net] has quit [Remote host closed the connection] 20:41:27 %nip-values says "we cannot move stack-allocated DX objects", what to do with such objects then? 20:53:02 *stassats* is not sure how the stack got into the state that nip-values can't work on it 20:53:53 *stassats* is banging his head against lp#533930 20:54:18 lp 533930 20:54:19 https://bugs.launchpad.net/bugs/533930 20:55:34 maybe the catch is messing within return-from is messing with the stack 20:57:15 *stassats* should open a channel #stassats-talking-about-not-getting-sbcl-internals 20:58:48 LiamH [~none@96.231.225.69] has joined #sbcl 21:16:40 why does nip-values even exist? when will the stack enter such a state when you have to remove things from the middle of it? 21:16:58 i have the code which makes nip-values, but i don't understand why 21:17:49 multiple-value return-from? 21:17:56 leoc [~leoc.git@p5DDB8930.dip0.t-ipconnect.de] has joined #sbcl 21:19:22 the code i'm looking at doesn't have return-from, just m-v and d-x 21:19:50 this one https://github.com/sbcl/sbcl/blob/master/tests/dynamic-extent.impure.lisp#L72 21:21:26 it inserts nip-values to remove the dx value 21:24:43 ok, i guess i see how that happens, the multiple-values are used after the call to bar, while the dx is not 21:24:56 so it has to leave the m-v on the stack, but remove dx 21:26:25 then i'm not sure what's the fundamental problem if dx is left on the stack and uv is removed 21:27:46 nip-values was included to reduce a vast number of weird corruption bugs that pfdietz' random tester used to throw up 21:28:01 trawling misc.lsp for them might be worthwhile 21:28:12 though that sounds a bit strange, since dx is supposed to be dynamic-extent and shouldn't be left 21:29:04 stassats: we already have problems with loops around some stack-allocated stuff. 21:30:01 ehaliewicz [~user@50-0-51-11.dsl.static.sonic.net] has joined #sbcl 21:36:38 can we get rid of ir2 and put something comprehensible instead? 21:43:31 hm, the test case in dynamic-extent.impure succeeds, but if i modify it a bit it returns garbage 21:44:24 ah, of course, i shouldn't be returned dxed values 21:45:01 one of the down sides of being open for business is a marked increase in really bad bug reports 21:45:18 better than no bug reports 21:45:39 (psychologically, maybe it's an up side; at least now there are bug reports that are easier to deal with than yet another modular arithmetic problem) 21:54:09 -!- ehaliewicz [~user@50-0-51-11.dsl.static.sonic.net] has quit [Ping timeout: 256 seconds] 21:54:30 drmeister [~drmeister@mobile-198-228-227-216.mycingular.net] has joined #sbcl 21:54:59 *stassats* finds a bug where pop-values is wrong, maybe that will shed some light 21:55:02 -!- drmeister [~drmeister@mobile-198-228-227-216.mycingular.net] has quit [Remote host closed the connection] 21:56:08 drmeister [~drmeister@mobile-198-228-227-216.mycingular.net] has joined #sbcl 22:19:59 -!- wbooze [~wbooze@xdsl-78-35-161-109.netcologne.de] has quit [Ping timeout: 255 seconds] 22:36:20 wbooze [~wbooze@xdsl-78-35-161-109.netcologne.de] has joined #sbcl 22:36:50 -!- wbooze [~wbooze@xdsl-78-35-161-109.netcologne.de] has quit [Read error: Connection reset by peer] 22:38:47 wbooze [~wbooze@xdsl-78-35-161-109.netcologne.de] has joined #sbcl 23:08:17 Bike_ [~Glossina@75-164-175-225.ptld.qwest.net] has joined #sbcl 23:10:18 -!- Bike [~Glossina@63-230-172-89.ptld.qwest.net] has quit [Ping timeout: 264 seconds] 23:10:36 -!- Bike_ is now known as Bike 23:11:04 -!- leuler [~user@p548FC232.dip0.t-ipconnect.de] has quit [Quit: ERC Version 5.1.2 $Revision: 1.796.2.6 $ (IRC client for Emacs)] 23:15:17 -!- LiamH [~none@96.231.225.69] has quit [Ping timeout: 248 seconds] 23:34:06 deriving a type of NIL doesn't seem to produce any error message 23:34:31 is there any way to signal a type error from inside defoptimizer derive-type? 23:44:50 sometimes that's just how we get stronger dead code elimination 23:45:25 but it should be possible to do something for combination nodes (and other stuff that isn't subject to constraint propagation) 23:46:18 i'm deriving types for append/nconc, and if it finds (append fixnum x y) it returns NIL, but an error message would be nice 23:48:39 by it you mean the type deriver? 23:48:51 yes 23:49:26 I don't think that the best way to handle this. 23:49:32 ASau``` [~user@p4FF96CD6.dip0.t-ipconnect.de] has joined #sbcl 23:49:34 can you assert-lvar-type the arguments isntead? 23:50:36 defknown on append would also work. 23:50:54 defknown wouldn't really work 23:51:01 (append fixnum) is legal 23:51:13 (append x y fixnum) is too 23:51:25 ah, yes. 23:53:08 -!- ASau`` [~user@p5797F028.dip0.t-ipconnect.de] has quit [Ping timeout: 255 seconds] 23:56:59 doing both appears to work 23:58:57 doing assert-lvar-type on all arguments insert type checks, even though append will perform them itself 23:59:11 would doing assert-lvar-type only when i detect a type mismatch be better?