00:12:12 stassats [~stassats@wikipedia/stassats] has joined #sbcl 00:18:59 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 01:03:17 -!- minion [~minion@common-lisp.net] has quit [Ping timeout: 240 seconds] 01:03:49 -!- lisppaste2 [~lisppaste@common-lisp.net] has quit [Ping timeout: 260 seconds] 01:04:01 -!- specbot [~specbot@common-lisp.net] has quit [Ping timeout: 246 seconds] 01:48:28 cpc26 [~cpc26@66-87-0-216.pools.spcsdns.net] has joined #sbcl 01:56:49 hi, is *load-truename* always nil outside of a load? 01:58:34 <|3b|> clhs says yes 01:58:51 3b: thanks! 01:59:35 <|3b|> though non-conformant programs might try to SETF it, and conformant CL implementations could allow them to 02:00:13 *|3b|* isn't sure if a conformant CL program could try to set it or not 02:01:29 I am having trouble with a macro (let ((truename *load-truename*)) ... 02:01:50 there is an if and error so it always error(s) 02:02:19 it is the define-module macro from GBBopen 02:02:48 seems *laod-truename* is used a bit and they are having issues with sbcl around this 02:03:14 <|3b|> possibly you want one or more of *compile-file-truename*, #., eval-when, or asdf:system-relative-pathname? 02:03:55 -!- homie [~levgue@xdsl-87-79-193-143.netcologne.de] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 02:04:02 *|3b|* guesses randomly due to not seeing actual code or actual errors 02:11:01 http://pastebin.com/pfbQ4Ppj 02:11:37 the error is on line 73 02:16:22 homie [~levgue@xdsl-87-79-193-143.netcologne.de] has joined #sbcl 02:27:42 *|3b|* meant something more like a backtrace than the call to ERROR itself 02:29:29 *|3b|* assumes you are actually calling LOAD and such (or otherwise doing whatever that code expects you to do to load it), not typing things at repl or interacting with individual forms in slime? 02:29:53 or using compile-file 02:32:50 <|3b|> ah, that is outside the ` isn't it, so yeah, might be an issue whether things are being compiled or loaded 02:34:59 <|3b|> cpc26: maybe try replacing the *load-truename* with (or *compile-file-truename* *load-truename*) ? 02:37:01 <|3b|> (or rewrite the whole thing to do the work in the expansion instead of while the macro is running) 02:37:38 http://pastebin.com/a03mu3PR 02:38:41 <|3b|> try C-c C-l instead of C-c C-k? 02:48:22 homie` [~levgue@xdsl-78-35-174-121.netcologne.de] has joined #sbcl 02:48:43 -!- fe[nl]ix [~lacedaemo@pdpc/supporter/professional/fenlix] has quit [Quit: Valete!] 02:48:54 -!- homie` [~levgue@xdsl-78-35-174-121.netcologne.de] has quit [Remote host closed the connection] 02:50:11 -!- homie [~levgue@xdsl-87-79-193-143.netcologne.de] has quit [Ping timeout: 252 seconds] 02:50:36 fe[nl]ix [~lacedaemo@pdpc/supporter/professional/fenlix] has joined #sbcl 02:52:31 -!- fe[nl]ix [~lacedaemo@pdpc/supporter/professional/fenlix] has quit [Client Quit] 02:57:01 |3b|: C-c C-l I get t but the :tutorial does not autoload 02:57:16 fe[nl]ix [~lacedaemo@pdpc/supporter/professional/fenlix] has joined #sbcl 02:57:41 -!- fe[nl]ix [~lacedaemo@pdpc/supporter/professional/fenlix] has quit [Client Quit] 02:58:27 |3b|: wait I had to define-repl-command 02:59:46 this all doesn't seem to be related to SBCL 03:01:45 fe[nl]ix [~lacedaemo@pdpc/supporter/professional/fenlix] has joined #sbcl 03:02:16 yeah I would think that but the original software is using *load-truename* and the author mentions it as an sbcl difference. 03:02:48 wondering a way to use something in place of *load-truename* at this point 03:03:03 they do recommend loading 03:04:20 also has examples for other cl that use it and, I assume, work - like allegro and clisp 03:04:57 <|3b|> well, if *load-truename* behaves differently on allegro or clisp, that sounds like a bug 03:05:15 <|3b|> or if they don't expand macros during compilation, also a bug 03:05:45 <|3b|> if you are compiling things you should be loading, that sounds like user error (or poor documentation if it didn't tell you that) 03:05:55 i imagine that this project you are looking at is just written by someone who isn't particularly competent in CL 03:06:15 *|3b|* suspects it is more just ancient code 03:06:29 maybe, he started in the 70's and has been using it since 03:08:11 so basically sbcl is to the spec and the others are working outside the spec by some older convention? 03:08:37 or you're drawing the wrong conclusions 03:08:54 *|3b|* suspects doing the same thing would break the same way on the others, but something is subtly 'not the same thing' in the other examples 03:10:02 ok here is something from the docs - "SBCL has a very strict interpretation of *load-truename* semantics" 03:10:24 so this is some older pre-ansi convention? 03:10:58 <|3b|> no, that (if i found the right context) just seems to mean that SBCL doesn't LOAD .sbclrc, nothing to do with modules 03:11:21 right you did but I think that is part of the auto laod issue? 03:11:37 the macros use it a lot 03:11:45 <|3b|> in .sbclrc? 03:11:52 no is the rest of the app 03:12:08 the .sbclrc is just to get the init started 03:12:15 <|3b|> so what does how SBCL interprets .sbclrc have to do with stuff that isn't in .sbclrc? 03:12:31 <|3b|> note that the .sbclrc code calls LOAD, which should set that variable 03:12:50 no I think it is how *load-truname* has to be in a load 03:12:59 nothing to do with .sbclrc 03:13:16 I think maybe the other lisps are different 03:13:42 i don't think so 03:13:43 <|3b|> if you mean 'other lisps are buggy', you should make a test case and file bug reports :) 03:14:16 homie [~levgue@xdsl-78-35-174-121.netcologne.de] has joined #sbcl 03:16:03 <|3b|> as far as i can see, any CL where that macro works when compiled is buggy 03:16:30 right 03:16:39 <|3b|> since the only requirement of compilation is to expand macros, so the evaluation of *load-truename* happens before the .fasl is loaded 03:17:56 it should use #.(or *compile-file-truename* *load-truename*) instead 03:18:05 *|3b|* supposes there is an edge case of calling compile-file during the loading of another file, not sure if it is specified to be reset there... relying on it is probably a bad idea anyway though, since i'd assume you want the current file's path not some random file that happened to load this one 03:18:33 <|3b|> stassats: without the #. i think, since it is in a macro 03:19:14 <|3b|> and if i understand correctly, it wants the values at the call site, not when the macro is defined 03:21:55 so l-t-v 03:23:34 *|3b|* thinks that might fall under 'fix the macro' option suggested earlier :) 03:24:14 <|3b|> (as in 'don't try to do everything during macroexpansion') 03:24:26 or anything, really... 03:24:33 <|3b|> right 03:25:01 macroexpansion-time IO and side effects usually end up being subtly wrong anyway 03:25:38 <|3b|> not quite that bad, since it just builds a path i think 03:26:38 <|3b|> i think the main thing that would move would be the error check (but not really user visibly since you can't compile it currently anyway) 03:27:06 well, there's no guarantee that I won't move fasls post-compile :| 03:28:48 i don't see a reason to know where files are located at all 03:28:49 <|3b|> yeah, not sure what the intended behavior is there 03:29:05 s/files/fasls/ 03:29:36 because i always store them outside of the source tree 03:30:02 <|3b|> i think this app does something like that too, or at least in a separate part of the source tree or something 03:30:19 *|3b|* didn't actually read the details 04:01:08 -!- cpc26 [~cpc26@66-87-0-216.pools.spcsdns.net] has quit [Quit: WeeChat 0.3.3] 04:36:18 slyrus_ [~chatzilla@adsl-76-254-45-145.dsl.pltn13.sbcglobal.net] has joined #sbcl 04:53:38 -!- Phooodus [~foo@68.107.217.139] has quit [Ping timeout: 260 seconds] 05:05:29 -!- gor[e] [~svr@gw1.masterhost.ru] has quit [Ping timeout: 240 seconds] 05:18:45 tcr1 [~tcr@217.162.131.235] has joined #sbcl 05:56:18 -!- tcr1 [~tcr@217.162.131.235] has quit [Quit: Leaving.] 07:13:12 tcr1 [~tcr@217-162-131-235.dynamic.hispeed.ch] has joined #sbcl 07:22:43 -!- homie [~levgue@xdsl-78-35-174-121.netcologne.de] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 07:26:55 homie [~levgue@xdsl-78-35-174-121.netcologne.de] has joined #sbcl 07:38:21 nikodemus [~nikodemus@cs181199216.pp.htv.fi] has joined #sbcl 07:38:21 -!- ChanServ has set mode +o nikodemus 07:41:41 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 07:53:36 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 08:01:39 -!- pchrist [~spirit@gentoo/developer/pchrist] has quit [Read error: Operation timed out] 08:16:23 pchrist [~spirit@gentoo/developer/pchrist] has joined #sbcl 08:16:37 -!- homie [~levgue@xdsl-78-35-174-121.netcologne.de] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 08:33:11 attila_lendvai [~attila_le@catv-80-98-24-21.catv.broadband.hu] has joined #sbcl 08:33:11 -!- attila_lendvai [~attila_le@catv-80-98-24-21.catv.broadband.hu] has quit [Changing host] 08:33:11 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 08:33:22 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has left #sbcl 08:48:39 homie [~levgue@xdsl-78-35-174-121.netcologne.de] has joined #sbcl 09:17:54 -!- Krystof [~csr21@csrhodes.plus.com] has quit [Ping timeout: 240 seconds] 09:28:19 udzinari [~user@ip-89-102-12-6.net.upcbroadband.cz] has joined #sbcl 09:31:07 Krystof [~csr21@csrhodes.plus.com] has joined #sbcl 09:31:07 -!- ChanServ has set mode +o Krystof 10:26:20 -!- nikodemus [~nikodemus@cs181199216.pp.htv.fi] has quit [Quit: This computer has gone to sleep] 10:35:56 nikodemus [~nikodemus@cs181199216.pp.htv.fi] has joined #sbcl 10:35:56 -!- ChanServ has set mode +o nikodemus 10:46:42 superjudge [~mjl@c83-250-110-188.bredband.comhem.se] has joined #sbcl 11:20:16 -!- superjudge [~mjl@c83-250-110-188.bredband.comhem.se] has quit [Ping timeout: 276 seconds] 11:29:41 -!- nikodemus [~nikodemus@cs181199216.pp.htv.fi] has quit [Quit: This computer has gone to sleep] 11:36:18 superjudge [~mjl@83.250.110.188] has joined #sbcl 11:41:05 -!- superjudge [~mjl@83.250.110.188] has quit [Ping timeout: 260 seconds] 12:10:17 cpc26 [~cpc26@66-87-2-18.pools.spcsdns.net] has joined #sbcl 12:25:30 -!- udzinari [~user@ip-89-102-12-6.net.upcbroadband.cz] has quit [Remote host closed the connection] 12:38:06 udzinari [~user@ip-89-102-12-6.net.upcbroadband.cz] has joined #sbcl 12:59:15 nikodemus [~nikodemus@cs181199216.pp.htv.fi] has joined #sbcl 12:59:29 -!- ChanServ has set mode +o nikodemus 13:03:37 -!- peddie [~peddie@XVM-107.MIT.EDU] has quit [Ping timeout: 276 seconds] 13:22:37 peddie [~peddie@XVM-107.MIT.EDU] has joined #sbcl 13:23:34 rmarynch [~roman@88.135.194.233] has joined #sbcl 13:28:50 Are COMPILE and LOAD really thread-safe? Or they just grab some mutex until the compilation/load is finished? 13:35:28 multi-threaded load of pure tests into the same SBCL is only 20%-30% faster than single-threaded. It cannot even eat two CPU cores with three threads. Luckily, there are impure tests, which are parallel processes... But I haven't finished that part yet. 13:45:11 rmarynch: fasl loading and the compiler are essentially single threaded 13:46:08 i would advice against running tests in multiple threads in the same image, as it makes it harder to figure out where a failure comes from -- not all tests are necessarily thread safe 13:46:36 nikodemus: so, it makes sense to spawn threads for impure tests only? 13:46:49 i would not spawn threads at all 13:47:14 for parallelizing the tests i would stricky keep everything in separate processes 13:48:00 but... it turns out that multiple threads aren't a problem in practice, then ... it isn't a problem :) 13:48:48 each thread runs run-impure-in-child-sbcl, so the test is in a separate process 13:50:15 that's fine then 13:50:38 threads are needed to manage the processes, while supervisor runs pure tests at the same time... I will submit the code to LP once it is finished, and we can review it then 13:51:51 -!- peddie [~peddie@XVM-107.MIT.EDU] has quit [Ping timeout: 258 seconds] 13:53:05 strictly speaking you don't need threads to manage processes -- there's the sb-ext:process api you can use 13:53:26 but if threads make it easier, that doesn't sound like a problem 13:53:57 superjudge [~mjl@c83-250-110-188.bredband.comhem.se] has joined #sbcl 13:54:59 btw, why all .sh tests are #-win32? They fail there, or msys fails to run them? 13:55:35 these tests seem to be related to pathnames 13:55:47 (not all) 13:57:23 not sure 13:58:28 I will run them tomorrow in the office, I have Windows box there. 14:08:08 peddie [~peddie@XVM-107.MIT.EDU] has joined #sbcl 14:14:37 -!- peddie [~peddie@XVM-107.MIT.EDU] has quit [Ping timeout: 240 seconds] 14:20:47 peddie [~peddie@XVM-107.MIT.EDU] has joined #sbcl 14:29:38 -!- rmarynch [~roman@88.135.194.233] has quit [Quit: Leaving] 14:47:29 -!- loke [~elias@bb119-74-155-99.singnet.com.sg] has quit [Read error: Connection reset by peer] 14:47:48 loke [~elias@bb121-6-7-182.singnet.com.sg] has joined #sbcl 15:31:19 homie` [~levgue@xdsl-78-35-143-207.netcologne.de] has joined #sbcl 15:31:31 -!- peddie [~peddie@XVM-107.MIT.EDU] has quit [Ping timeout: 258 seconds] 15:33:31 -!- homie [~levgue@xdsl-78-35-174-121.netcologne.de] has quit [Ping timeout: 248 seconds] 15:49:59 -!- udzinari [~user@ip-89-102-12-6.net.upcbroadband.cz] has quit [Remote host closed the connection] 16:00:59 peddie [~peddie@XVM-107.MIT.EDU] has joined #sbcl 16:01:44 -!- homie` [~levgue@xdsl-78-35-143-207.netcologne.de] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 16:03:44 homie [~levgue@xdsl-78-35-143-207.netcologne.de] has joined #sbcl 16:41:30 -!- Quadrescence [~Quad@unaffiliated/quadrescence] has quit [Quit: omghaahhahaohwow] 16:50:52 -!- nikodemus [~nikodemus@cs181199216.pp.htv.fi] has quit [Ping timeout: 258 seconds] 17:09:18 -!- cmm [~cmm@bzq-79-183-205-247.red.bezeqint.net] has quit [Ping timeout: 246 seconds] 17:10:20 cmm [~cmm@bzq-79-183-205-247.red.bezeqint.net] has joined #sbcl 17:12:41 -!- slyrus_ [~chatzilla@adsl-76-254-45-145.dsl.pltn13.sbcglobal.net] has quit [Ping timeout: 240 seconds] 17:14:39 nikodemus [~nikodemus@cs181199216.pp.htv.fi] has joined #sbcl 17:14:39 -!- ChanServ has set mode +o nikodemus 17:19:44 slyrus_ [~chatzilla@adsl-76-254-45-145.dsl.pltn13.sbcglobal.net] has joined #sbcl 17:32:07 rmarynch [~roman@88.135.194.233] has joined #sbcl 17:36:19 I have tested the launcher prototype with Phenom X4 920, the original one runs 7 minutes, the threaded one runs less than 3 minutes. 17:36:32 there is one more possibility to make it faster 17:37:37 split threads.impure.lisp on several small files. The current file runs almost 3 minutes in the separate thread 17:39:42 other threads are done, but this one still goes. The file has many tests, so it is possible. Right? 17:40:15 can you break them down logically? 17:40:58 I think yes, hash-table, classes, locks - all these are not very related 17:41:14 but they use some common helpers 17:41:52 which should be moved to threads-tests-utils.lisp or something like that 17:46:04 -!- peddie [~peddie@XVM-107.MIT.EDU] has quit [Ping timeout: 258 seconds] 17:49:24 peddie [~peddie@XVM-107.MIT.EDU] has joined #sbcl 18:01:57 -!- cmm [~cmm@bzq-79-183-205-247.red.bezeqint.net] has quit [Quit: leaving] 18:10:36 -!- nikodemus [~nikodemus@cs181199216.pp.htv.fi] has quit [Ping timeout: 258 seconds] 18:36:50 nikodemus [~nikodemus@62-241-240-154.bb.dnainternet.fi] has joined #sbcl 18:36:50 -!- ChanServ has set mode +o nikodemus 18:49:28 -!- rmarynch [~roman@88.135.194.233] has quit [Quit: Leaving] 20:09:49 -!- superjudge [~mjl@c83-250-110-188.bredband.comhem.se] has quit [Ping timeout: 240 seconds] 20:12:59 Phoodus [~foo@ip68-231-47-70.ph.ph.cox.net] has joined #sbcl 20:32:23 -!- slyrus [~slyrus@adsl-76-254-45-145.dsl.pltn13.sbcglobal.net] has quit [Remote host closed the connection] 20:35:03 -!- Phoodus [~foo@ip68-231-47-70.ph.ph.cox.net] has quit [Ping timeout: 252 seconds] 20:47:29 -!- stassats [~stassats@wikipedia/stassats] has quit [Read error: Connection reset by peer] 21:11:16 -!- tcr1 [~tcr@217-162-131-235.dynamic.hispeed.ch] has quit [Quit: Leaving.] 21:23:41 prxq [~mommer@91.143.109.12] has joined #sbcl 22:01:40 -!- prxq [~mommer@91.143.109.12] has quit [Quit: Leaving] 22:05:03 -!- Krystof [~csr21@csrhodes.plus.com] has quit [Ping timeout: 246 seconds] 22:18:31 Krystof [~csr21@csrhodes.plus.com] has joined #sbcl 22:18:31 -!- ChanServ has set mode +o Krystof 22:32:50 -!- nikodemus [~nikodemus@62-241-240-154.bb.dnainternet.fi] has quit [Ping timeout: 276 seconds] 23:04:10 Phoodus [~foo@ip68-231-47-70.ph.ph.cox.net] has joined #sbcl 23:13:07 attila_lendvai [~attila_le@catv-80-98-24-21.catv.broadband.hu] has joined #sbcl 23:13:07 -!- attila_lendvai [~attila_le@catv-80-98-24-21.catv.broadband.hu] has quit [Changing host] 23:13:07 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl