2016-09-13T15:10:20Z ccl-logbot joined #sbcl
2016-09-13T15:10:20Z 
2016-09-13T15:10:20Z names: ccl-logbot schjetne stassats scymtym DGASAU madbub gingerale angavrilov salva myrkraverk loke slyrus foom em1l irsol edgar-rft nyef brfennpocock DeadTrickster nzambe fe[nl]ix Blkt karswell` whiteline White_Flame drmeister christoph_debian leo_song angular_mike luis abruanese eschatologist pchrist mordocai nicdev brucem PuercoPop gko |3b| dougk_ minion carvite p_l jsnell jdz Xof les abbe chris2 mood Shinmera jzp fiddlerwoaroof sigjuice flip214 Posterdati
2016-09-13T15:10:20Z names: pkhuong xristos jackc dustinm`_ pootler_ trinque WojciechK iskander payphone jibanes Cthulhux _iwc antoszka ferada jackdaniel specbot
2016-09-13T15:11:02Z payphone quit (Read error: Connection reset by peer)
2016-09-13T15:11:16Z payphone joined #sbcl
2016-09-13T15:18:44Z Quadrescence joined #sbcl
2016-09-13T16:08:00Z slyrus quit (Ping timeout: 250 seconds)
2016-09-13T16:23:29Z slyrus joined #sbcl
2016-09-13T16:25:33Z Bike joined #sbcl
2016-09-13T17:27:27Z rumbler31 joined #sbcl
2016-09-13T17:52:07Z sjl joined #sbcl
2016-09-13T18:00:56Z slyrus quit (Ping timeout: 244 seconds)
2016-09-13T18:02:20Z karswell` quit (Ping timeout: 244 seconds)
2016-09-13T18:04:49Z scymtym quit (Ping timeout: 255 seconds)
2016-09-13T18:10:30Z des_consolado joined #sbcl
2016-09-13T18:19:51Z sjl quit (Read error: Connection reset by peer)
2016-09-13T18:24:06Z slyrus joined #sbcl
2016-09-13T18:24:29Z scymtym joined #sbcl
2016-09-13T18:30:33Z DGASAU quit (Ping timeout: 240 seconds)
2016-09-13T18:33:09Z DGASAU joined #sbcl
2016-09-13T18:58:17Z joshe joined #sbcl
2016-09-13T19:08:31Z slyrus quit (Ping timeout: 252 seconds)
2016-09-13T19:16:42Z payphone quit (*.net *.split)
2016-09-13T19:16:43Z salva quit (*.net *.split)
2016-09-13T19:16:43Z myrkraverk quit (*.net *.split)
2016-09-13T19:16:43Z eschatologist quit (*.net *.split)
2016-09-13T19:16:44Z abruanese quit (*.net *.split)
2016-09-13T19:16:44Z luis quit (*.net *.split)
2016-09-13T19:16:47Z em1l quit (*.net *.split)
2016-09-13T19:16:50Z rumbler31 quit (*.net *.split)
2016-09-13T19:16:51Z foom quit (*.net *.split)
2016-09-13T19:16:51Z DeadTrickster quit (*.net *.split)
2016-09-13T19:16:53Z dougk_ quit (*.net *.split)
2016-09-13T19:16:53Z |3b| quit (*.net *.split)
2016-09-13T19:16:53Z gko quit (*.net *.split)
2016-09-13T19:16:53Z PuercoPop quit (*.net *.split)
2016-09-13T19:16:56Z brucem quit (*.net *.split)
2016-09-13T19:16:56Z nicdev quit (*.net *.split)
2016-09-13T19:16:56Z pchrist quit (*.net *.split)
2016-09-13T19:16:56Z mordocai quit (*.net *.split)
2016-09-13T19:16:57Z joshe quit (*.net *.split)
2016-09-13T19:16:58Z gingerale quit (*.net *.split)
2016-09-13T19:16:59Z christoph_debian quit (*.net *.split)
2016-09-13T19:17:00Z jsnell quit (*.net *.split)
2016-09-13T19:17:00Z p_l quit (*.net *.split)
2016-09-13T19:17:01Z carvite quit (*.net *.split)
2016-09-13T19:17:01Z minion quit (*.net *.split)
2016-09-13T19:17:05Z Shinmera quit (*.net *.split)
2016-09-13T19:17:06Z mood quit (*.net *.split)
2016-09-13T19:17:06Z abbe quit (*.net *.split)
2016-09-13T19:17:07Z xristos quit (*.net *.split)
2016-09-13T19:17:10Z des_consolado quit (*.net *.split)
2016-09-13T19:17:10Z iskander quit (*.net *.split)
2016-09-13T19:17:10Z jibanes quit (*.net *.split)
2016-09-13T19:17:12Z angavrilov quit (*.net *.split)
2016-09-13T19:17:12Z brfennpocock quit (*.net *.split)
2016-09-13T19:17:12Z fe[nl]ix quit (*.net *.split)
2016-09-13T19:17:12Z Blkt quit (*.net *.split)
2016-09-13T19:17:13Z flip214 quit (*.net *.split)
2016-09-13T19:17:14Z Posterdati quit (*.net *.split)
2016-09-13T19:17:14Z madbub quit (*.net *.split)
2016-09-13T19:17:18Z pootler_ quit (*.net *.split)
2016-09-13T19:17:18Z trinque quit (*.net *.split)
2016-09-13T19:17:19Z DGASAU quit (*.net *.split)
2016-09-13T19:17:20Z leo_song quit (*.net *.split)
2016-09-13T19:17:22Z WojciechK quit (*.net *.split)
2016-09-13T19:17:23Z jackdaniel quit (*.net *.split)
2016-09-13T19:17:23Z ferada quit (*.net *.split)
2016-09-13T19:17:24Z antoszka quit (*.net *.split)
2016-09-13T19:17:30Z whiteline quit (*.net *.split)
2016-09-13T19:17:31Z jackc quit (*.net *.split)
2016-09-13T19:17:32Z Cthulhux quit (*.net *.split)
2016-09-13T19:17:32Z schjetne quit (*.net *.split)
2016-09-13T19:17:33Z nyef quit (*.net *.split)
2016-09-13T19:17:34Z White_Flame quit (*.net *.split)
2016-09-13T19:17:35Z dustinm`_ quit (*.net *.split)
2016-09-13T19:17:36Z _iwc quit (*.net *.split)
2016-09-13T19:17:36Z specbot quit (*.net *.split)
2016-09-13T19:17:38Z stassats quit (*.net *.split)
2016-09-13T19:17:39Z loke quit (*.net *.split)
2016-09-13T19:17:39Z nzambe quit (*.net *.split)
2016-09-13T19:17:40Z angular_mike quit (*.net *.split)
2016-09-13T19:17:42Z sigjuice quit (*.net *.split)
2016-09-13T19:17:45Z scymtym quit (*.net *.split)
2016-09-13T19:17:45Z Bike quit (*.net *.split)
2016-09-13T19:17:45Z irsol quit (*.net *.split)
2016-09-13T19:17:46Z edgar-rft quit (*.net *.split)
2016-09-13T19:17:46Z Xof quit (*.net *.split)
2016-09-13T19:17:46Z les quit (*.net *.split)
2016-09-13T19:17:47Z fiddlerwoaroof quit (*.net *.split)
2016-09-13T19:17:47Z pkhuong quit (*.net *.split)
2016-09-13T19:19:43Z White_Flame joined #sbcl
2016-09-13T19:19:43Z pchrist joined #sbcl
2016-09-13T19:19:43Z mordocai joined #sbcl
2016-09-13T19:19:43Z nicdev joined #sbcl
2016-09-13T19:19:43Z brucem joined #sbcl
2016-09-13T19:19:43Z PuercoPop joined #sbcl
2016-09-13T19:19:43Z gko joined #sbcl
2016-09-13T19:19:43Z |3b| joined #sbcl
2016-09-13T19:19:43Z dougk_ joined #sbcl
2016-09-13T19:19:43Z DeadTrickster joined #sbcl
2016-09-13T19:19:43Z foom joined #sbcl
2016-09-13T19:19:43Z rumbler31 joined #sbcl
2016-09-13T19:19:43Z dustinm` joined #sbcl
2016-09-13T19:19:43Z carvite joined #sbcl
2016-09-13T19:19:43Z jsnell joined #sbcl
2016-09-13T19:19:43Z christoph_debian joined #sbcl
2016-09-13T19:19:43Z gingerale joined #sbcl
2016-09-13T19:19:43Z joshe joined #sbcl
2016-09-13T19:19:43Z Cthulhux joined #sbcl
2016-09-13T19:19:51Z specbot joined #sbcl
2016-09-13T19:20:03Z DGASAU joined #sbcl
2016-09-13T19:20:03Z madbub joined #sbcl
2016-09-13T19:20:03Z pootler_ joined #sbcl
2016-09-13T19:20:03Z trinque joined #sbcl
2016-09-13T19:20:11Z whiteline joined #sbcl
2016-09-13T19:20:12Z pootler_ quit (Ping timeout: 250 seconds)
2016-09-13T19:20:18Z payphone joined #sbcl
2016-09-13T19:20:18Z salva joined #sbcl
2016-09-13T19:20:18Z myrkraverk joined #sbcl
2016-09-13T19:20:18Z eschatologist joined #sbcl
2016-09-13T19:20:18Z abruanese joined #sbcl
2016-09-13T19:20:18Z luis joined #sbcl
2016-09-13T19:20:29Z stassats joined #sbcl
2016-09-13T19:20:51Z sigjuice joined #sbcl
2016-09-13T19:20:58Z xristos joined #sbcl
2016-09-13T19:22:24Z leo_song joined #sbcl
2016-09-13T19:22:24Z WojciechK joined #sbcl
2016-09-13T19:22:24Z jackdaniel joined #sbcl
2016-09-13T19:22:24Z ferada joined #sbcl
2016-09-13T19:22:24Z antoszka joined #sbcl
2016-09-13T19:22:37Z nyef joined #sbcl
2016-09-13T19:22:37Z scymtym joined #sbcl
2016-09-13T19:22:37Z Bike joined #sbcl
2016-09-13T19:22:37Z irsol joined #sbcl
2016-09-13T19:22:37Z edgar-rft joined #sbcl
2016-09-13T19:22:37Z Xof joined #sbcl
2016-09-13T19:22:37Z les joined #sbcl
2016-09-13T19:22:37Z fiddlerwoaroof joined #sbcl
2016-09-13T19:22:37Z pkhuong joined #sbcl
2016-09-13T19:22:41Z jzp quit (Ping timeout: 265 seconds)
2016-09-13T19:22:56Z jackc joined #sbcl
2016-09-13T19:22:56Z loke joined #sbcl
2016-09-13T19:22:56Z jibanes joined #sbcl
2016-09-13T19:23:04Z p_l joined #sbcl
2016-09-13T19:23:21Z em1l joined #sbcl
2016-09-13T19:23:21Z schjetne joined #sbcl
2016-09-13T19:23:21Z abbe joined #sbcl
2016-09-13T19:23:21Z angavrilov joined #sbcl
2016-09-13T19:23:21Z brfennpocock joined #sbcl
2016-09-13T19:23:21Z fe[nl]ix joined #sbcl
2016-09-13T19:23:21Z Blkt joined #sbcl
2016-09-13T19:23:21Z flip214 joined #sbcl
2016-09-13T19:23:21Z Posterdati joined #sbcl
2016-09-13T19:23:22Z Shinmera joined #sbcl
2016-09-13T19:23:22Z mood joined #sbcl
2016-09-13T19:23:43Z xristos is now known as Guest47922
2016-09-13T19:24:01Z iskander joined #sbcl
2016-09-13T19:24:04Z p_l quit (Changing host)
2016-09-13T19:24:04Z p_l joined #sbcl
2016-09-13T19:24:10Z abbe is now known as Guest22211
2016-09-13T19:25:49Z minion joined #sbcl
2016-09-13T19:26:33Z drmeister quit (Ping timeout: 265 seconds)
2016-09-13T19:27:45Z des_consolado joined #sbcl
2016-09-13T19:28:18Z Guest47922 is now known as xristos
2016-09-13T19:28:25Z slyrus joined #sbcl
2016-09-13T19:28:43Z xristos quit (Changing host)
2016-09-13T19:28:43Z xristos joined #sbcl
2016-09-13T19:34:48Z _iwc joined #sbcl
2016-09-13T19:35:27Z _iwc quit (K-Lined)
2016-09-13T19:40:05Z rumbler31 quit (Remote host closed the connection)
2016-09-13T19:43:43Z p_l quit
2016-09-13T19:49:51Z SamSkulls joined #sbcl
2016-09-13T19:51:38Z nzambe joined #sbcl
2016-09-13T19:59:46Z slyrus quit (Ping timeout: 244 seconds)
2016-09-13T20:03:14Z pootler_ joined #sbcl
2016-09-13T20:04:51Z sjl joined #sbcl
2016-09-13T20:15:26Z drmeister joined #sbcl
2016-09-13T20:19:25Z angavrilov quit (Remote host closed the connection)
2016-09-13T20:20:37Z angular_mike_ joined #sbcl
2016-09-13T20:33:19Z angular_mike_ quit (Ping timeout: 260 seconds)
2016-09-13T20:34:25Z gingerale quit (Remote host closed the connection)
2016-09-13T20:35:55Z angular_mike_ joined #sbcl
2016-09-13T20:40:42Z rumbler31 joined #sbcl
2016-09-13T20:47:53Z rumbler31 quit (Ping timeout: 250 seconds)
2016-09-13T21:21:44Z nzambe quit (*.net *.split)
2016-09-13T21:27:49Z slyrus joined #sbcl
2016-09-13T21:27:55Z em1l quit (*.net *.split)
2016-09-13T21:27:55Z schjetne quit (*.net *.split)
2016-09-13T21:27:56Z Guest22211 quit (*.net *.split)
2016-09-13T21:27:57Z brfennpocock quit (*.net *.split)
2016-09-13T21:27:57Z Blkt quit (*.net *.split)
2016-09-13T21:27:57Z fe[nl]ix quit (*.net *.split)
2016-09-13T21:27:58Z flip214 quit (*.net *.split)
2016-09-13T21:27:58Z Posterdati quit (*.net *.split)
2016-09-13T21:27:59Z flip214_ joined #sbcl
2016-09-13T21:28:01Z em1l joined #sbcl
2016-09-13T21:28:04Z Blkt joined #sbcl
2016-09-13T21:28:06Z fe[nl]ix joined #sbcl
2016-09-13T21:28:17Z Posterdati joined #sbcl
2016-09-13T21:28:21Z abbe joined #sbcl
2016-09-13T21:28:24Z brfennpocock joined #sbcl
2016-09-13T21:29:06Z schjetne joined #sbcl
2016-09-13T21:31:39Z angular_mike_ quit (Ping timeout: 250 seconds)
2016-09-13T21:39:46Z angular_mike_ joined #sbcl
2016-09-13T21:47:42Z edgar-rft quit (Quit: edgar-rft)
2016-09-13T22:07:14Z karswell joined #sbcl
2016-09-13T22:30:39Z flip214_ quit (Ping timeout: 265 seconds)
2016-09-13T22:46:47Z ASau joined #sbcl
2016-09-13T22:50:01Z flip214 joined #sbcl
2016-09-13T22:50:01Z flip214 quit (Changing host)
2016-09-13T22:50:01Z flip214 joined #sbcl
2016-09-13T23:10:53Z DeadTrickster quit (Ping timeout: 250 seconds)
2016-09-13T23:17:43Z cromachina joined #sbcl
2016-09-13T23:37:12Z DeadTrickster joined #sbcl
2016-09-13T23:38:10Z DeadTrickster_ joined #sbcl
2016-09-13T23:43:01Z madbub quit (Ping timeout: 250 seconds)
2016-09-14T00:30:32Z slyrus quit (Ping timeout: 244 seconds)
2016-09-14T00:57:35Z sjl quit (Read error: Connection reset by peer)
2016-09-14T00:58:46Z salva quit (Ping timeout: 244 seconds)
2016-09-14T01:19:33Z stassats quit (Ping timeout: 240 seconds)
2016-09-14T01:34:51Z trinque quit (Remote host closed the connection)
2016-09-14T01:35:49Z trinque joined #sbcl
2016-09-14T01:39:14Z em1l_ joined #sbcl
2016-09-14T01:42:41Z em1l quit (Ping timeout: 244 seconds)
2016-09-14T01:44:46Z salva joined #sbcl
2016-09-14T01:50:51Z whiteline quit (Remote host closed the connection)
2016-09-14T01:52:03Z whiteline joined #sbcl
2016-09-14T02:12:31Z slyrus joined #sbcl
2016-09-14T02:19:46Z nzambe joined #sbcl
2016-09-14T02:28:49Z slyrus quit (Ping timeout: 255 seconds)
2016-09-14T02:29:20Z rumbler31 joined #sbcl
2016-09-14T02:31:37Z shcl joined #sbcl
2016-09-14T02:32:58Z shcl: Hey everyone!  It looks like load_core_file opens the core file but never closes it.  Is this intentional?
2016-09-14T02:33:45Z rumbler31 quit (Ping timeout: 250 seconds)
2016-09-14T02:34:09Z slyrus joined #sbcl
2016-09-14T03:43:45Z nyef: shcl: Maybe? I forget how posix mmap() semantics work if you close the file.
2016-09-14T03:54:27Z loke: closing the file that is mapped will not unmap
2016-09-14T03:54:46Z loke: normally you open→mmap→close...work...munmap
2016-09-14T03:55:35Z nyef: In that case, leaving it open might be a mistake... or a long-forgotten workaround for some odd unix or other.
2016-09-14T03:55:59Z nyef: After all, SBCL (or CMUCL) ran under Digital UNIX, IRIX, HPUX...
2016-09-14T03:56:06Z nyef: Okay, maybe not IRIX.
2016-09-14T03:56:22Z nyef: Definitely IRIX and very probably Digital UNIX.
2016-09-14T03:56:31Z nyef: Whatever the latter is/was called.
2016-09-14T03:56:40Z nyef: OSF/1 ?
2016-09-14T03:56:51Z loke: Did _anyone_ run OSF/1?
2016-09-14T03:57:03Z nyef: Wasn't it required for running the VLM?
2016-09-14T03:57:10Z loke: Does SBCL run on AIX?
2016-09-14T03:57:21Z loke: What is VLM?
2016-09-14T03:57:26Z slyrus quit (Ping timeout: 265 seconds)
2016-09-14T03:57:27Z nyef: Ask stassats about AIX. d-:
2016-09-14T03:57:34Z nyef: The Symbolics Virtual Lisp Machine.
2016-09-14T03:58:07Z edgar-rft joined #sbcl
2016-09-14T03:58:12Z nyef: ... A/UX might be an amusing port, if it weren't for the fact that the fastest A/UX machines ran at 40MHz.
2016-09-14T04:01:02Z loke: Ought to be enough for anybody
2016-09-14T04:01:36Z nyef: Eh. They had more than 640k of RAM. d-:
2016-09-14T05:14:22Z shcl: I tried closing the fd, and sbcl seems to run fine
2016-09-14T05:14:46Z shcl: I guess I'll file a bug
2016-09-14T05:17:41Z karswell quit (Remote host closed the connection)
2016-09-14T05:17:53Z karswell joined #sbcl
2016-09-14T05:24:01Z slyrus joined #sbcl
2016-09-14T05:27:30Z rumbler31 joined #sbcl
2016-09-14T05:31:58Z rumbler31 quit (Ping timeout: 255 seconds)
2016-09-14T05:34:54Z karswell quit (Remote host closed the connection)
2016-09-14T05:35:08Z karswell joined #sbcl
2016-09-14T06:00:10Z scymtym quit (Ping timeout: 244 seconds)
2016-09-14T06:05:05Z shcl quit (Remote host closed the connection)
2016-09-14T06:59:55Z p_l joined #sbcl
2016-09-14T07:12:29Z SamSkull` joined #sbcl
2016-09-14T07:12:37Z jibanes quit (Ping timeout: 265 seconds)
2016-09-14T07:14:50Z Bike quit (Quit: same)
2016-09-14T07:17:00Z SamSkulls quit (Ping timeout: 265 seconds)
2016-09-14T07:18:54Z jibanes joined #sbcl
2016-09-14T08:24:07Z angavrilov joined #sbcl
2016-09-14T08:56:03Z schjetne quit (Ping timeout: 240 seconds)
2016-09-14T08:56:05Z rudolfochrist joined #sbcl
2016-09-14T09:48:21Z ASau quit (Ping timeout: 244 seconds)
2016-09-14T09:57:48Z rudolfochrist quit (Ping timeout: 276 seconds)
2016-09-14T10:15:39Z stassats joined #sbcl
2016-09-14T10:19:50Z DavidGu joined #sbcl
2016-09-14T10:30:46Z DavidGu quit (Ping timeout: 250 seconds)
2016-09-14T10:36:55Z rudolfochrist joined #sbcl
2016-09-14T10:39:37Z myrkraverk: When I'm writing a daemon, and want to close all file descriptors, can that corrupt my SBCL somehow?
2016-09-14T10:39:59Z myrkraverk: I've seen example code with a loop, but I'm wondering if that's actually the best way to go about it.
2016-09-14T10:53:23Z rudolfochrist quit (Ping timeout: 250 seconds)
2016-09-14T10:55:07Z scymtym joined #sbcl
2016-09-14T11:07:07Z stassats quit (Ping timeout: 265 seconds)
2016-09-14T11:25:36Z edgar-rft quit (Quit: edgar-rft)
2016-09-14T12:03:26Z sjl joined #sbcl
2016-09-14T12:08:28Z flip214: myrkraverk: on linux you can also readdir(/proc/self/fd) and close only those...
2016-09-14T12:08:42Z flip214: not sure if it's faster - perhaps if you expect to only have a few open.
2016-09-14T12:09:39Z DGASAU quit (Ping timeout: 250 seconds)
2016-09-14T12:09:48Z flip214: perhaps select() and doing some binary search (to see whether any above RLIMIT_NOFILE/2 are open, RLIMIT_NOFILE/4, ...) would be faster and more portable
2016-09-14T12:10:02Z flip214: or looping, but that needs RLIMIT_NOFILE (minus a few) syscalls
2016-09-14T12:11:30Z DGASAU joined #sbcl
2016-09-14T12:14:48Z madbub joined #sbcl
2016-09-14T12:15:57Z schjetne joined #sbcl
2016-09-14T12:20:28Z DGASAU quit (Read error: Connection reset by peer)
2016-09-14T12:21:13Z DGASAU joined #sbcl
2016-09-14T12:27:34Z myrkraverk: flip214: yeah, I did consider implementing getrlimit with the alien API, but then I decided to "skip this step for now".
2016-09-14T12:27:56Z myrkraverk: And wait until it becomes a problem/issue/whatever, and deal with it then.
2016-09-14T12:31:28Z madbub_ joined #sbcl
2016-09-14T12:31:42Z madbub quit (Ping timeout: 265 seconds)
2016-09-14T12:32:31Z edgar-rft joined #sbcl
2016-09-14T12:38:12Z DGASAU quit (Read error: Connection reset by peer)
2016-09-14T12:40:48Z DGASAU joined #sbcl
2016-09-14T12:47:36Z DGASAU quit (Ping timeout: 265 seconds)
2016-09-14T12:48:44Z DGASAU joined #sbcl
2016-09-14T12:56:35Z slyrus quit (Ping timeout: 260 seconds)
2016-09-14T12:58:04Z slyrus joined #sbcl
2016-09-14T13:03:03Z leo_song quit (Read error: Connection reset by peer)
2016-09-14T13:03:33Z leo_song joined #sbcl
2016-09-14T13:33:10Z cromachina quit (Read error: Connection reset by peer)
2016-09-14T13:47:45Z slyrus quit (Ping timeout: 244 seconds)
2016-09-14T13:57:26Z Bike joined #sbcl
2016-09-14T14:25:28Z pkhuong quit (Ping timeout: 244 seconds)
2016-09-14T14:27:21Z pkhuong joined #sbcl
2016-09-14T14:27:45Z pkhuong is now known as Guest6162
2016-09-14T14:39:06Z madbub_ quit (Remote host closed the connection)
2016-09-14T14:43:26Z madbub joined #sbcl
2016-09-14T15:31:29Z edgar-rft quit (Quit: edgar-rft)
2016-09-14T16:11:52Z scymtym quit (Ping timeout: 255 seconds)
2016-09-14T16:14:50Z dougk_ quit (Ping timeout: 250 seconds)
2016-09-14T16:15:15Z scymtym joined #sbcl
2016-09-14T16:20:04Z trinque quit (Quit: trinque)
2016-09-14T16:26:44Z dougk_ joined #sbcl
2016-09-14T16:47:33Z gingerale joined #sbcl
2016-09-14T16:59:50Z sjl quit (Ping timeout: 260 seconds)
2016-09-14T17:29:17Z DGASAU quit (Read error: Connection reset by peer)
2016-09-14T17:30:29Z DGASAU joined #sbcl
2016-09-14T17:39:25Z brfennpocock quit (Ping timeout: 250 seconds)
2016-09-14T17:47:17Z DGASAU quit (Read error: Connection reset by peer)
2016-09-14T17:50:52Z DGASAU joined #sbcl
2016-09-14T17:53:22Z edgar-rft joined #sbcl
2016-09-14T17:57:21Z sjl joined #sbcl
2016-09-14T17:58:03Z scymtym quit (Ping timeout: 250 seconds)
2016-09-14T18:01:46Z DGASAU quit (Read error: Connection reset by peer)
2016-09-14T18:05:08Z DGASAU joined #sbcl
2016-09-14T18:10:37Z DGASAU quit (Read error: Connection reset by peer)
2016-09-14T18:12:00Z DGASAU joined #sbcl
2016-09-14T18:15:20Z Cthulhux quit (Changing host)
2016-09-14T18:15:20Z Cthulhux joined #sbcl
2016-09-14T18:22:57Z DGASAU quit (Ping timeout: 244 seconds)
2016-09-14T18:32:09Z Bike quit (Read error: Connection reset by peer)
2016-09-14T18:33:03Z Bike joined #sbcl
2016-09-14T18:33:17Z edgar-rft quit (Quit: edgar-rft)
2016-09-14T18:35:53Z DGASAU joined #sbcl
2016-09-14T18:40:03Z schjetne quit (Ping timeout: 248 seconds)
2016-09-14T18:45:44Z schjetne joined #sbcl
2016-09-14T18:52:02Z schjetne quit (Ping timeout: 265 seconds)
2016-09-14T18:53:00Z schjetne joined #sbcl
2016-09-14T19:00:40Z slyrus joined #sbcl
2016-09-14T19:27:53Z sjl quit (Ping timeout: 240 seconds)
2016-09-14T19:36:03Z slyrus quit (Ping timeout: 248 seconds)
2016-09-14T19:38:19Z slyrus joined #sbcl
2016-09-14T19:50:29Z Posterdati quit (Read error: Connection reset by peer)
2016-09-14T19:59:52Z gingerale quit (Remote host closed the connection)
2016-09-14T20:02:50Z scymtym joined #sbcl
2016-09-14T20:04:02Z Posterdati joined #sbcl
2016-09-14T20:37:57Z rumbler31 joined #sbcl
2016-09-14T21:01:23Z nyef: Should I be seeing a heap allocation sequence in (defun foo (x y) (let ((bar (list x x))) (declare (dynamic-extent bar)) (baz bar) (setf bar (list y y)) (baz bar))) ?
2016-09-14T21:03:28Z madbub quit (Remote host closed the connection)
2016-09-14T21:08:44Z nyef: I'm currently thinking that the answer is "no, but I'm seeing them because SBCL doesn't optimize DX quite as well as we might like it to".
2016-09-14T21:12:36Z nyef: Definitely shouldn't be seeing any for (defun foo (x y) (let ((a (when x (list x))) (b (when y (list y)))) (declare (dynamic-extent a b)) (bar a b))) though.
2016-09-14T21:19:33Z nyef: Preliminary conclusion: DX in SBCL is still not great.
2016-09-14T21:19:55Z nyef: Preliminary conclusion: DX in SBCL is worse than I thought it was.
2016-09-14T21:24:03Z nyef: Really, "at least it doesn't break the compiler or produce incorrect code anymore" is a weak defence at this point. (-:
2016-09-14T21:34:58Z angavrilov quit (Remote host closed the connection)
2016-09-14T21:38:24Z quadresce joined #sbcl
2016-09-14T21:46:25Z nyef: quadresce: Do I remember rightly that you have a microExplorer, but no software for it?
2016-09-14T21:46:42Z quadresce: nyef, I have lots of software
2016-09-14T21:46:49Z quadresce: nyef, a huge thing of floppies + CDs
2016-09-14T21:46:53Z nyef: Ah, okay.
2016-09-14T21:46:59Z quadresce: I just never got the software loaded onto any of them
2016-09-14T21:47:08Z quadresce: and one of the uX needs a pin repair
2016-09-14T21:47:34Z nyef: I found a download site with three versions of the microexplorer software.
2016-09-14T21:47:53Z quadresce: Wow, awesome. I know some folks who are trying to get this stuff backed up.
2016-09-14T21:48:06Z quadresce: It might be them, because they pinged me for my disks
2016-09-14T21:49:42Z quadresce: Pasting this question here too: hello folks, I was wondering if anyone had advice on producing binaries for SBCL. I know how to make them, of course, but what I'm interested in is removing some information that ideally shouldn't be there, like saved source forms, etc. (I'm not looking for/asking for a full tree-shaking solution.)
2016-09-14T21:50:47Z schjetne_ joined #sbcl
2016-09-14T21:51:41Z schjetne quit (Read error: Connection reset by peer)
2016-09-14T21:51:49Z nyef: Hrm. Lost the bloody link after I downloaded the files. /-:
2016-09-14T21:52:27Z nyef: You said it was a pin repair on the NuBus connector?
2016-09-14T21:55:45Z quadresce: Yes.
2016-09-14T21:56:06Z quadresce: Probably replacing the connector would be easiest, but I know nothing.
2016-09-14T21:57:48Z nyef: Probably. For which a donor card may need to be sacrificed, or a "new" connector purchased if such can be found.
2016-09-14T21:58:17Z quadresce: I think it's relatively easy to find a new connector. If not, there are billions of terrible NuBus cards that can be sacrificed.
2016-09-14T21:58:23Z quadresce: In any event I have two uX cards
2016-09-14T21:58:24Z nyef: Exactly.
2016-09-14T21:59:03Z fe[nl]ix: quadresce: compile the code with DEBUG 0
2016-09-14T21:59:52Z quadresce: How do you do this nicely with ASDF?
2016-09-14T22:00:31Z quadresce: (brb)
2016-09-14T22:00:48Z fe[nl]ix: quadresce: you can't, it's immplementation-specific
2016-09-14T22:01:15Z nyef: quadresce: At some point, I'm probably going to offer to take that broken uX board off your hands.
2016-09-14T22:01:30Z fe[nl]ix: quadresce: look into the implementation of sb-ext:restrict-compiler-policy
2016-09-14T22:01:44Z fe[nl]ix: that imposes a global minimum value on certain optimization types
2016-09-14T22:01:52Z fe[nl]ix: you should be able to impose a maximum too
2016-09-14T22:02:04Z fe[nl]ix: and set the max DEBUG 0
2016-09-14T22:02:37Z fe[nl]ix: otherwise you can look into the implementation of the swank-sbcl backend
2016-09-14T22:03:00Z fe[nl]ix: and see how SBCL stores source info, then write code to strip an image before save
2016-09-14T22:03:49Z fe[nl]ix: some parts of SBCL are stuck in the '70s :(
2016-09-14T22:04:10Z nyef: Possibly use map-allocated-objects to trash the debug info globally and force a full GC on save.
2016-09-14T22:05:06Z fe[nl]ix: I'll look into it
2016-09-14T22:05:29Z fe[nl]ix: this should reduce size of QPX binaries
2016-09-14T22:24:23Z myrkraverk quit (Remote host closed the connection)
2016-09-14T22:30:51Z quadresce: back
2016-09-14T22:32:17Z quadresce: fe[nl]ix, I'm happy to help. I would consider this semi-urgent for me for deployment.
2016-09-14T22:33:09Z fe[nl]ix: quadresce: in my case it's not urgent
2016-09-14T22:33:26Z fe[nl]ix: I might dedicate some time to it by the end of Q4
2016-09-14T22:33:49Z quadresce: I see :)
2016-09-14T22:33:58Z fe[nl]ix: we're all available for advice and review of new APIs, though :)
2016-09-14T22:36:03Z fe[nl]ix: my current focus is on ELF support
2016-09-14T22:36:32Z quadresce: fancy
2016-09-14T22:36:55Z fe[nl]ix: that will not help you unless there's a switch to save all debug info natively as DWARF
2016-09-14T22:37:02Z fe[nl]ix: and even that is tough
2016-09-14T23:00:40Z quadresce: nyef, this MAP-ALLOCATED-OBJECTS business is scary
2016-09-14T23:01:23Z fe[nl]ix: hahaha
2016-09-14T23:22:16Z schjetne_ quit (Ping timeout: 265 seconds)
2016-09-14T23:26:59Z quadresce: So I guess as a first order I want to (1) remove docstrings, (2) remove source forms, (3) remove source locations.
2016-09-14T23:27:08Z quadresce: I'm looking at sb-introspect for info about this stuff.
2016-09-14T23:29:59Z schjetne joined #sbcl
2016-09-14T23:32:50Z myrkraverk joined #sbcl
2016-09-14T23:35:31Z mordocai quit (Ping timeout: 250 seconds)
2016-09-14T23:41:23Z mordocai joined #sbcl
2016-09-14T23:45:55Z cromachina joined #sbcl
2016-09-15T00:19:23Z schjetne quit (Ping timeout: 250 seconds)
2016-09-15T00:24:55Z schjetne joined #sbcl
2016-09-15T00:24:59Z quadresce: man, the compiler's stored info is all over the place
2016-09-15T01:05:37Z slyrus quit (Ping timeout: 244 seconds)
2016-09-15T01:12:17Z quadresce: anyone down to do some weekend contracting work for this task? :)
2016-09-15T01:12:51Z nyef: Heh.
2016-09-15T01:13:05Z nyef: MAP-ALLOCATED-OBJECTS is a LOT less scary than it once was.
2016-09-15T01:13:26Z quadresce: It wasn't too bad to read I guess. It mostly made sense. It's still kind of brutal.
2016-09-15T01:13:41Z quadresce: Is there some sort of function to trap heap allocations?
2016-09-15T01:14:09Z quadresce: (with dynamic extent)
2016-09-15T01:14:36Z nyef: Ah, so that you'll catch an error to the face, or at least a warning, if something gets heap-allocated anyway?
2016-09-15T01:14:53Z quadresce: Error to the face.
2016-09-15T01:15:14Z nyef: Not aware of anything offhand, though if you want it for runtime I can think of a couple of angles.
2016-09-15T01:15:26Z quadresce: Yes for runtime.
2016-09-15T01:15:46Z quadresce: (That's not the contracting/bounty/whatever idea. I want some good info deleting code for that.)
2016-09-15T01:16:31Z nyef: Actually, I can think of a detector that will let you know if anything is heap-allocated within its body.
2016-09-15T01:16:51Z quadresce: nyef, the quiet sbcl genius
2016-09-15T01:17:42Z nyef: x86-64, I presume?
2016-09-15T01:19:21Z nyef: The basic idea is that for non-large-object allocations, the free-pointer in the allocation region will change if an allocation occurs.
2016-09-15T01:19:39Z nyef: And it is unlikely to return to precisely the same location post-GC anyway.
2016-09-15T01:20:21Z nyef: So "just" grab a copy of it beforehand (it'll be a fixnum) and compare it to the post-operation copy. If they are the same, great. If they're different, you allocated something to the heap.
2016-09-15T01:20:38Z nyef: Oh. Or you GCd. That'll confuse it as well.
2016-09-15T01:21:47Z nyef: Do we need to keep debug-info objects around for anything?
2016-09-15T01:22:12Z quadresce: nyef, arm64
2016-09-15T01:22:21Z quadresce: oh, the allocation trap yes
2016-09-15T01:22:26Z quadresce: x86_64
2016-09-15T01:31:34Z nyef: Nuking the code debug info only seems to save about 9.5 megs.
2016-09-15T01:31:41Z nyef: (sb-vm::map-allocated-objects (lambda (obj tag size) (declare (ignore size)) (when (= tag sb-vm:code-header-widetag) (print obj) (setf (sb-kernel:%code-debug-info obj) nil))) :dynamic)
2016-09-15T01:32:34Z nyef: I don't THINK that it's used anywhere critical outside of the debugger, but I really don't know for sure.
2016-09-15T01:32:49Z nyef: Use at your own risk, et cetera.
2016-09-15T01:33:44Z nyef: That's on a bare core, though. Nothing loaded.
2016-09-15T01:34:11Z nyef: Feel free to remove the call to PRINT in there, too. I only added it so that I knew that it was finding appropriate-looking objects.
2016-09-15T01:36:11Z nyef: Any more crazy internals hacks you might be interested in tonight? (-:
2016-09-15T01:37:58Z em1l joined #sbcl
2016-09-15T01:39:47Z nyef: Looks like docstrings are stashed in the globaldb.
2016-09-15T01:41:16Z em1l_ quit (Ping timeout: 244 seconds)
2016-09-15T01:41:57Z nyef: As are source-locations.
2016-09-15T01:41:59Z nyef: Hrm.
2016-09-15T01:42:42Z nyef: And some of this stuff might be stored in yet other places instead/as-well.
2016-09-15T01:45:03Z schjetne quit (Ping timeout: 240 seconds)
2016-09-15T01:45:23Z schjetne joined #sbcl
2016-09-15T01:48:07Z nyef: Okay, I have an angle for scrubbing out the globaldb a bit, too, though it's even more fragile/hacky than the thing above for killing code debug info.
2016-09-15T01:48:59Z nyef: Basically, call-with-each-globaldb-name, and pass it a function that uses (setf info) to kill out the various info-types that you want to have gone.
2016-09-15T01:49:40Z nyef: Ah. Or there's CLEAR-INFO, which would be a better choice than (SETF INFO).
2016-09-15T01:51:09Z karswell quit (Ping timeout: 250 seconds)
2016-09-15T02:02:20Z nyef: Wow, it's only been about three and a half years since the big map-allocate-objects rewrite.
2016-09-15T02:10:52Z nyef: Heh. "... has been completely rewritten, it may now exhibit different bugs than before."
2016-09-15T02:51:13Z nyef: Hrm. Okay, this somewhat works. Breaks loading some of the contribs, though, and doesn't manage to kill generic-function documentation.
2016-09-15T02:51:50Z nyef: Also, I'm not sure that killing the source-location and documentation stuff releases very much memory.
2016-09-15T02:53:36Z dougk_: hi nyef: are you there?
2016-09-15T02:53:42Z nyef: Hello dougk_.
2016-09-15T02:53:59Z dougk_: hey, does define-assembly-routine permit you to mess up the registers that you say are your :args ?
2016-09-15T02:54:04Z dougk_: or you do have to put them in temps ?
2016-09-15T02:54:17Z dougk_: i want to / need to mess them up. i can't remember how this works.
2016-09-15T02:54:30Z dougk_: the vop that wraps the routine does you the favor of declaring them, but i'm all confused
2016-09-15T02:55:18Z dougk_: you have precisely 5 minutes to answer or i will miss the last train home tonight :-)
2016-09-15T02:56:01Z nyef: "I don't remember offhand. Go home and ask me again tomorrow"?
2016-09-15T02:56:05Z dougk_: lolz
2016-09-15T02:56:08Z dougk_: i'll figure it out
2016-09-15T02:56:46Z dougk_: we don't usually mess them up except in the routines that are explicitly doing the fancy stuff such as unwinding
2016-09-15T02:57:26Z nyef: I think that HPPA TRUNCATE trashes them.
2016-09-15T02:57:35Z dougk_: ah nice example.
2016-09-15T02:57:58Z nyef: Yeah. A "nice" example, given how buggy it was.
2016-09-15T02:57:59Z dougk_: oh, serious i gotta run so don't bother with any more sage advice for the meanwhile
2016-09-15T02:58:03Z nyef: Heh.
2016-09-15T02:58:07Z nyef: Travel safe.
2016-09-15T03:00:46Z nyef: Hrm. How on earth do I have a stale fasl for asdf?
2016-09-15T03:00:57Z nyef: Clean build time, apparently.
2016-09-15T03:03:25Z nyef: quadresce: How does http://paste.lisp.org/display/326051 seem as a starting point?
2016-09-15T03:10:40Z nyef: Takes the core down from 42827824 to 33161264 on my system.
2016-09-15T03:34:54Z nyef: quadresce: Are you also wanting to purge xref data?
2016-09-15T03:50:26Z edgar-rft joined #sbcl
2016-09-15T04:06:03Z nyef: Killing the xref data and various CLOS documentation things netted another 170k, it seems.
2016-09-15T04:07:22Z nyef: quadresce: http://paste.lisp.org/display/326051#1 is the next version.
2016-09-15T04:07:39Z nyef: I think that I'm going to stop here for the moment.
2016-09-15T04:11:49Z nyef: Oh! And I found that microexplorer software link: http://www.vcfed.org/forum/showthread.php?45274-Macintosh-IIFX-with-TI-MicroExplorer-LISP-processor-card
2016-09-15T04:31:53Z ASau joined #sbcl
2016-09-15T04:38:44Z quadresce quit (Read error: Connection reset by peer)
2016-09-15T04:39:03Z quadresce joined #sbcl
2016-09-15T04:39:26Z quadresce is now known as Guest82013
2016-09-15T05:02:37Z rumbler31 quit (Remote host closed the connection)
2016-09-15T05:07:30Z DavidGu joined #sbcl
2016-09-15T05:36:19Z DavidGu quit (Quit: DavidGu)
2016-09-15T05:48:03Z scymtym quit (Ping timeout: 240 seconds)
2016-09-15T05:57:56Z gingerale joined #sbcl
2016-09-15T06:44:24Z angavrilov joined #sbcl
2016-09-15T06:45:58Z gingerale quit (Remote host closed the connection)
2016-09-15T07:03:11Z rumbler31 joined #sbcl
2016-09-15T07:08:34Z p_l quit (Read error: Connection reset by peer)
2016-09-15T07:08:48Z rumbler31 quit (Ping timeout: 244 seconds)
2016-09-15T07:09:57Z p_l joined #sbcl
2016-09-15T07:32:49Z nzambe quit (Quit: http://www.kiwiirc.com/ - A hand crafted IRC client)
2016-09-15T07:50:53Z scymtym joined #sbcl
2016-09-15T08:20:20Z Bike quit (Quit: death)
2016-09-15T08:36:39Z SamSkull` quit (Remote host closed the connection)
2016-09-15T08:38:22Z rumbler31 joined #sbcl
2016-09-15T08:40:12Z ASau quit (Ping timeout: 276 seconds)
2016-09-15T08:42:55Z rumbler31 quit (Ping timeout: 250 seconds)
2016-09-15T12:02:46Z sjl joined #sbcl
2016-09-15T12:11:42Z madbub joined #sbcl
2016-09-15T12:16:35Z schjetne quit (Ping timeout: 260 seconds)
2016-09-15T12:58:52Z rudolfochrist joined #sbcl
2016-09-15T13:16:41Z cromachina quit (Read error: Connection reset by peer)
2016-09-15T14:16:45Z rudolfochrist quit (Ping timeout: 260 seconds)
2016-09-15T14:28:53Z rumbler31 joined #sbcl
2016-09-15T14:39:25Z rumbler31 quit (Remote host closed the connection)
2016-09-15T14:42:44Z rumbler31 joined #sbcl
2016-09-15T14:47:21Z rumbler31 quit (Ping timeout: 250 seconds)
2016-09-15T14:51:15Z rumbler31 joined #sbcl
2016-09-15T14:55:37Z Guest6162 is now known as pkhuong
2016-09-15T14:55:50Z rumbler31 quit (Ping timeout: 260 seconds)
2016-09-15T14:58:18Z Bike joined #sbcl
2016-09-15T15:06:30Z rpg joined #sbcl
2016-09-15T15:06:30Z rpg quit (Client Quit)
2016-09-15T15:06:50Z rpg joined #sbcl
2016-09-15T15:09:07Z sjl quit (Ping timeout: 265 seconds)
2016-09-15T15:26:38Z Quadrescence: nyef, re "purge xref", yes
2016-09-15T15:28:03Z Quadrescence: nyef, this paste is scary. But maybe should be the start of an SB-DELIVERY contrib :)
2016-09-15T15:28:32Z Quadrescence: I'll try it out soon
2016-09-15T15:28:37Z nyef: Not doing it. This paste is scary for a reason. d-:
2016-09-15T15:29:21Z nyef: If you think of anything else that needs to be nuked, let me know, and I can try to add it.
2016-09-15T15:30:04Z nyef: Or any place that I missed where docstrings might hide.
2016-09-15T15:30:09Z Quadrescence: can we rename :RANDOM-DOCUMENTATION to :MISC-DOCUMENTATION
2016-09-15T15:30:48Z nyef: Probably, but not a call that I'm going to make.
2016-09-15T15:30:53Z Quadrescence: haha
2016-09-15T15:31:29Z Quadrescence: nyef, does this eliminate stored source forms
2016-09-15T15:31:46Z rpg_ joined #sbcl
2016-09-15T15:31:56Z Quadrescence: does %SIMPLE-FUN-INFO contain that
2016-09-15T15:34:54Z rpg quit (Ping timeout: 276 seconds)
2016-09-15T15:35:33Z jibanes quit (Quit: leaving)
2016-09-15T15:35:50Z jibanes joined #sbcl
2016-09-15T15:40:10Z nyef: I don't know, actually.
2016-09-15T15:40:21Z nyef: Is this for inline, or for something else?
2016-09-15T15:41:07Z nyef: Looks like it doesn't kill the inline expansion stuff.
2016-09-15T15:42:04Z nyef: And I don't know if there's any dependency between (:function :inline-expansion-designator) and (:function :inlinep). That'll take a small bit of digging.
2016-09-15T15:49:25Z nzambe joined #sbcl
2016-09-15T15:49:25Z klltkr joined #sbcl
2016-09-15T15:50:37Z scymtym quit (Ping timeout: 250 seconds)
2016-09-15T15:50:51Z scymtym joined #sbcl
2016-09-15T15:59:23Z nyef: Quadrescence: Losing the inline expansion designators saves an additional 300k.
2016-09-15T15:59:45Z nyef: "Just" add (sb-int:clear-info :function :inline-expansion-designator name) in the appropriate place.
2016-09-15T16:00:49Z nyef: Umm... I don't know if it works for setf-functions, if they even support inline expansion like this.
2016-09-15T16:15:47Z schjetne joined #sbcl
2016-09-15T16:16:09Z rpg_ quit (Ping timeout: 272 seconds)
2016-09-15T16:19:45Z sjl joined #sbcl
2016-09-15T16:24:13Z gingerale joined #sbcl
2016-09-15T16:27:35Z Intensity joined #sbcl
2016-09-15T16:35:55Z rpg joined #sbcl
2016-09-15T16:40:14Z edgar-rft quit (Quit: edgar-rft)
2016-09-15T16:46:01Z nyef: ... SBCL issues a compiler-notify for free D-X declarations?
2016-09-15T16:46:13Z nyef: Lovely.
2016-09-15T17:01:44Z klltkr quit (Ping timeout: 265 seconds)
2016-09-15T17:07:18Z klltkr joined #sbcl
2016-09-15T17:34:32Z scymtym quit (Ping timeout: 265 seconds)
2016-09-15T17:36:04Z slyrus joined #sbcl
2016-09-15T17:37:17Z rpg quit (Quit: Textual IRC Client: www.textualapp.com)
2016-09-15T18:36:19Z klltkr quit (Ping timeout: 255 seconds)
2016-09-15T18:46:20Z slyrus quit (Ping timeout: 244 seconds)
2016-09-15T18:54:29Z rpg joined #sbcl
2016-09-15T19:06:08Z slyrus joined #sbcl
2016-09-15T19:46:01Z prxq joined #sbcl
2016-09-15T19:57:51Z Guest82013 is now known as Quadresce
2016-09-15T19:57:54Z Quadresce quit (Changing host)
2016-09-15T19:57:54Z Quadresce joined #sbcl
2016-09-15T19:58:16Z Quadresce: nyef, btw I'm not trying to save space. Just trying to remove what might be considered somewhat sensitive info.
2016-09-15T19:58:55Z scymtym joined #sbcl
2016-09-15T20:00:32Z nyef: Quadresce: Okay, that's fair. And I don't know that some of this stuff might still be getting kept around for whatever reason, so you'll want to check for that anyway. Is there anything else that you can think of that needs to be deleted?
2016-09-15T20:01:48Z Quadresce: nyef, This is a side thought, but might it be possible to totally randomize SYMBOL-NAMEs? I don't think I'd actually do that, but it seems like it should be possible...
2016-09-15T20:02:41Z Quadresce: nyef, package docs?
2016-09-15T20:03:16Z rumbler31 joined #sbcl
2016-09-15T20:03:33Z nyef: Scrambling symbol names would affect their hash values, I think. Should be more-or-less doable, but there would be consequences.
2016-09-15T20:04:46Z Quadresce: nyef, are hash values cached? Also obviously this would affect FIND-SYMBOL. I think it's probably a bad idea to do in an uncontrolled fashion (:
2016-09-15T20:04:57Z nyef: Nuking package docstrings should be easy enough with a LIST-ALL-PACKAGES... And apparently there's a source-location there as well.
2016-09-15T20:05:23Z nyef: Hash values end up in hash tables... And you may well be right, I think that they ARE cached on the symbol.
2016-09-15T20:06:46Z Quadresce: Is there an easy way to see all of the strings in an image? I recall seeing something about looking at data in the constant section and whatnot. I guess you can map over everything
2016-09-15T20:09:51Z karswell joined #sbcl
2016-09-15T20:10:38Z nyef: Use map-allocated-objects and test for STRINGP?
2016-09-15T20:10:52Z nyef: You'd have to run it for all three spaces, but whatever.
2016-09-15T20:10:59Z Quadresce: Yeah that's what I mean
2016-09-15T20:11:20Z Quadresce: I guess I would in particular want to find notionally immutable strings
2016-09-15T20:38:46Z Quadresce quit (Read error: Connection reset by peer)
2016-09-15T20:39:03Z Quadresce joined #sbcl
2016-09-15T20:39:27Z Quadresce is now known as Guest19294
2016-09-15T20:47:00Z gingerale quit (Remote host closed the connection)
2016-09-15T20:54:38Z ASau joined #sbcl
2016-09-15T21:05:53Z rpg quit (Ping timeout: 240 seconds)
2016-09-15T21:16:08Z rumbler31 quit (Remote host closed the connection)
2016-09-15T21:16:40Z stassats joined #sbcl
2016-09-15T21:25:35Z rumbler31 joined #sbcl
2016-09-15T21:29:33Z rumbler31 quit (Ping timeout: 240 seconds)
2016-09-15T21:30:04Z prxq quit (Read error: Connection reset by peer)
2016-09-15T21:45:21Z nicdev` joined #sbcl
2016-09-15T21:46:06Z |3b|` joined #sbcl
2016-09-15T21:46:16Z PuercoPop quit (Ping timeout: 250 seconds)
2016-09-15T21:46:42Z foom quit (Ping timeout: 250 seconds)
2016-09-15T21:46:43Z |3b| quit (Ping timeout: 250 seconds)
2016-09-15T21:46:43Z gko quit (Ping timeout: 250 seconds)
2016-09-15T21:46:43Z brucem quit (Ping timeout: 250 seconds)
2016-09-15T21:47:10Z nicdev quit (Ping timeout: 250 seconds)
2016-09-15T21:47:10Z pchrist quit (Ping timeout: 250 seconds)
2016-09-15T21:49:22Z pchrist joined #sbcl
2016-09-15T21:50:36Z Guest19294 is now known as Quadresce
2016-09-15T21:50:39Z Quadresce quit (Changing host)
2016-09-15T21:50:39Z Quadresce joined #sbcl
2016-09-15T21:50:52Z Quadresce: Is there a way to make SBCL quiet when the heap exhausts?
2016-09-15T21:51:02Z Quadresce: Just bail out and not print anything.
2016-09-15T21:52:06Z PuercoPop joined #sbcl
2016-09-15T21:52:30Z brucem joined #sbcl
2016-09-15T21:52:47Z gko joined #sbcl
2016-09-15T21:59:13Z foom joined #sbcl
2016-09-15T22:07:41Z nyef: Not sure about "quiet", but you can at least get it to not fault to LDB via --disable-debugger, building without LDB support, or there's some magic thing you can hit from Lisp space to turn LDB off without killing the Lisp-side debugger.
2016-09-15T22:09:39Z fe[nl]ix: and also --lose-on-corruption
2016-09-15T22:35:12Z angavrilov quit (Remote host closed the connection)
2016-09-15T22:39:11Z ASau quit (Read error: Connection reset by peer)
2016-09-15T22:39:35Z ASau joined #sbcl
2016-09-15T23:05:23Z cromachina joined #sbcl
2016-09-15T23:18:23Z nicdev` is now known as nicdev
2016-09-15T23:22:23Z madbub quit (Remote host closed the connection)
2016-09-15T23:32:42Z scymtym quit (Remote host closed the connection)
2016-09-15T23:46:49Z stassats quit (Ping timeout: 248 seconds)
2016-09-15T23:54:28Z slyrus quit (Ping timeout: 255 seconds)
2016-09-16T00:19:08Z rpg joined #sbcl
2016-09-16T00:40:13Z sjl quit (Ping timeout: 244 seconds)
2016-09-16T01:01:16Z jrm joined #sbcl
2016-09-16T01:06:26Z jrm: In the platform table (http://www.sbcl.org/platform-table.html) I see older versions for FreeBSD under x86 and AMD64.  Is that just a matter of someone stepping up?  Can I just submit the binaries?
2016-09-16T01:32:57Z Quadresce: jrm, I'm not the one in charge of it but it almost surely is a matter of someone stepping up
2016-09-16T01:36:26Z jrm: I wonder if it's useful though?  Almost all users will (should) be installing the FreeBSD binary packages and if they install from source, it's probably easier to just do `pkg install sbcl` for bootstrapping.
2016-09-16T01:36:45Z em1l_ joined #sbcl
2016-09-16T01:37:11Z Quadresce quit (Remote host closed the connection)
2016-09-16T01:40:21Z em1l quit (Ping timeout: 265 seconds)
2016-09-16T02:26:17Z nyef: In IR1UTIL / LVAR-GOOD-FOR-DX-P, can anyone think of a good argument why we shouldn't change EVERY to SOME?
2016-09-16T02:40:26Z pkhuong: nyef: every seems correct to me.
2016-09-16T02:42:51Z pkhuong: if any of the value flowing in the lvar can't be DXed, DXing the lvar seems like a dodgy proposition.
2016-09-16T02:48:38Z nyef: Umm... Your two statements conflict.
2016-09-16T02:49:04Z ASau quit (Ping timeout: 255 seconds)
2016-09-16T02:49:09Z nyef: What I'm saying is that ANY DXable USE should allow the LVAR to be DXed.
2016-09-16T02:49:46Z nyef: The non-DXable USEs won't DX, and the DXable USEs will, and STACK analysis should be able to sort out the consequences... shouldn't it?
2016-09-16T02:50:27Z nyef: Currently, the EVERY is checking to make sure that ALL of the USEs are DXable.
2016-09-16T02:50:55Z nyef: Oh, wait, sorry. I didn't parse your first statement correctly.
2016-09-16T02:51:54Z nyef: Still, my line of argument stands: We should be able to mark an LVAR as DX even if not all of the USEs will DX.
2016-09-16T02:55:21Z nyef: Hrm. Might have found another DX bug. /-:
2016-09-16T03:05:01Z nyef: ... Okay, possibly not a DX bug. Odd borderline condition, though.
2016-09-16T03:07:50Z nyef: ... If this isn't a DX bug, then the specification for DX is broken.
2016-09-16T03:10:33Z nyef: Yeah, the specification for DX is broken.
2016-09-16T03:11:39Z nyef: (lambda (x) (let ((foo (cons x nil))) (let ((bar foo)) (declare (dynamic-extent bar)) (baz bar)) foo)) is illegal code: It attempts to return the CONS allocated as the value of FOO, which has dynamic extent.
2016-09-16T03:20:08Z nyef: The definition (in the glossary) of "otherwise inaccessible part" is broken.
2016-09-16T03:20:21Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2016-09-16T03:21:45Z nyef: The first paragraph of the description for the DYNAMIC-EXTENT declaration is difficult to interpret, at best.
2016-09-16T03:24:28Z nyef: This is dreadful. I think that I'm going to give up on DX for a while.
2016-09-16T04:17:05Z |3b|`: isn't that code invalid because the cons is accessible at the end of the LET BAR form, rather than from trying to return it?
2016-09-16T04:17:09Z |3b|` is now known as |3b|
2016-09-16T04:19:24Z nyef: Probably, yes.
2016-09-16T04:20:07Z nyef: But note that it's still a matter of a bound declaration affecting something in an outer scope.
2016-09-16T04:21:05Z |3b|: you mean causing FOO to be allocated on stack?
2016-09-16T04:21:09Z nyef: Yes!
2016-09-16T04:21:36Z |3b| thinks it is more "allowing it to be allocated on stack", and wouldn't complain if it wasn't
2016-09-16T04:22:05Z nyef: Would you complain if it was?
2016-09-16T04:22:07Z |3b|: being told the code was bad is of course nice in the given example though
2016-09-16T04:22:26Z nyef: It is a very surprising consequence of D-X.
2016-09-16T04:22:40Z |3b|: in the case of bad code as above, i can't really complain :)
2016-09-16T04:22:57Z nyef: Is it really "bad" code?
2016-09-16T04:22:58Z |3b| might be unhappy, but that applies to lots of things
2016-09-16T04:23:05Z |3b|: by my reading, it isn't conformant
2016-09-16T04:23:21Z nyef: By mine it isn't either.
2016-09-16T04:23:21Z |3b|: the declaration is wrong, same as if you declared BAR an integer
2016-09-16T04:23:36Z |3b|: so can't complain about anything the compiler does in response
2016-09-16T04:23:55Z nyef: It's only "wrong" in that FOO is still accessible.
2016-09-16T04:24:28Z |3b|: in that foo makes the cons accessible, but yes
2016-09-16T04:24:45Z |3b|: (if FOO were changed inside the inner LET, it could be valid code)
2016-09-16T04:26:07Z nyef: Yes, then it could be valid.
2016-09-16T04:26:25Z |3b|: making an effort to DX the outer binding, but not making the effort to see if it is safe to do so seems like possibly a bad design choice, but is permitted by the spec as far as i can tell
2016-09-16T04:26:30Z nyef: And here I thought that the free declaration semantics for DX were bad enough.
2016-09-16T04:27:19Z |3b|: declaring FOO DX in the inner let would be valid, in which case stack allocating FOO would be wrong
2016-09-16T04:27:55Z |3b| isn't sure what they expect compilers to do with that information
2016-09-16T04:28:07Z nyef: Indeed, (let ((foo (cons ...))) (let () (declare (dynamic-extent foo)) ...)) should not stack-allocate.
2016-09-16T04:28:29Z nyef: But if, in the inner let, we turn around and (setf foo (cons ...)), THAT should stack-allocate. (-:
2016-09-16T04:28:48Z |3b|: seems like if the compiler could tell it was safe to DX FOO in that case, it could do so without the declaration
2016-09-16T04:29:00Z |3b|: you mean the initial binding could have been stack allocated?
2016-09-16T04:29:32Z nyef: No, a free declaration permits stack allocation of mutations to the binding within the declaration scope.
2016-09-16T04:29:37Z |3b|: i guess if it is assigned twice, the first value could also be DX
2016-09-16T04:29:53Z nyef: But does not affect the initial binding.
2016-09-16T04:30:06Z |3b|: ah, right. the assigned value could be DX, but isn't allowed to escape that scope
2016-09-16T04:30:50Z |3b|: though the wording is a bit less clear in the case of modifications and a free declaration
2016-09-16T04:31:20Z nyef: The wording is clear as mud throughout, for all its attempts at precision.
2016-09-16T04:31:41Z |3b|: you need to interpret "still ... a part" as excluding new values
2016-09-16T04:32:01Z nyef: (let ((foo (intern ...))) (declare (dynamic-extent foo)) ...) is permitted to stack-allocate, for example.
2016-09-16T04:32:41Z |3b|: stack allocate what?
2016-09-16T04:32:45Z nyef: The symbol.
2016-09-16T04:33:12Z |3b|: isn't it accesible?
2016-09-16T04:33:16Z nyef: So?
2016-09-16T04:33:16Z |3b|: (through the package)
2016-09-16T04:33:20Z nyef: It's the value.
2016-09-16T04:33:27Z nyef: The value of INTERN is the symbol.
2016-09-16T04:33:48Z nyef: The symbol is, by definition, an otherwise inaccessible part of itself.
2016-09-16T04:33:56Z nyef: Therefore: It stack-allocates.
2016-09-16T04:34:54Z |3b|: it is a bound declaration, so needs to be inaccessible, rather than an otherwise inaccessible part
2016-09-16T04:35:38Z nyef: But it is the programmer's responsibility to render it inaccessible.
2016-09-16T04:35:47Z nyef: Not the compiler's responsibility to enforce it.
2016-09-16T04:36:17Z |3b|: ah, right... it is permitted under the "non-conforming code" part
2016-09-16T04:36:41Z |3b|: so it is permitted to unintern it at the end too, right? :)
2016-09-16T04:37:14Z nyef: Yeah, so long as it gets uninterned, and a reference isn't preserved elsewhere in the meantime, things are fine.
2016-09-16T04:37:37Z |3b|: well, 'fine' in the 'doing something random is permitted by the spec' sense
2016-09-16T04:37:59Z nyef: Yeah. Overall, this is horrifying.
2016-09-16T04:38:15Z |3b|: i mean that compiler actively removing all references to a DX object at the end of a bound declaration is as conforming as just returning stack garbage
2016-09-16T04:38:35Z |3b|: same way it is conforming to increment the symbol if you declared it an integer
2016-09-16T04:39:08Z |3b|: (or to error because it isn't an integer, as sbcl usually does)
2016-09-16T04:40:05Z |3b|: so from the perspective of sbcl rather than the spec, i'd expect it to avoid DX allocating anything it can't tell is either safe, or declared by conforming code (which pretty much means safe)
2016-09-16T04:40:24Z |3b|: (assuming it isn't SAFETY 0)
2016-09-16T04:40:51Z |3b|: and nice warnings/errors when the declarations are wrong, same as with types
2016-09-16T04:41:05Z nyef: ... does "free DX" declare a binding and everything referred to by that binding to be immutable
2016-09-16T04:41:06Z nyef: ?
2016-09-16T04:41:37Z |3b|: possibly, though i think that isn't intended
2016-09-16T04:42:02Z nyef: Oh my head.
2016-09-16T04:42:12Z |3b| suspects the "(if F established a binding for vari)" is overly restrictive
2016-09-16T04:42:19Z |3b|: if that weren't there, it wouldn't
2016-09-16T04:42:39Z nyef: Right.
2016-09-16T04:42:48Z nyef: Time to see if CLtL1 has anything useful to say here.
2016-09-16T04:43:11Z |3b|: actually, i guess you also need to say new things added to F need to be otherwise inaccessible, or my earlier argument about "still" falls apart and that needs fixed instead
2016-09-16T04:43:42Z |3b|: actually, maybe not
2016-09-16T04:44:02Z nyef: ... CLtL1 doesn't seem to have a DX declaration.
2016-09-16T04:44:14Z |3b|: it can't be DX if it escapes the form, so you can't assign anything DXable to F that leaves the body
2016-09-16T04:44:51Z schjetne quit (Ping timeout: 248 seconds)
2016-09-16T04:45:50Z |3b|: actually, i guess you can assign accessible things
2016-09-16T04:46:03Z |3b|: or maybe not
2016-09-16T04:47:24Z |3b|: ok, not immutable, but you can only permute the value it originally had?
2016-09-16T04:47:41Z |3b|: so (setf foo (nreverse foo)) and such
2016-09-16T04:49:26Z nyef: The issue writup, which is what was actually voted on, seems to have slightly different semantics.
2016-09-16T04:51:05Z nyef: Oh, lovely. "Free DX" requires that no references to the object can escape, either.
2016-09-16T04:53:50Z nyef: I believe that I have had enough of this dreadfulness for the time being.
2016-09-16T04:54:19Z edgar-rft joined #sbcl
2016-09-16T04:55:06Z |3b|: yeah, i think "otherwise inaccessible part" may be poorly defined, or poorly used
2016-09-16T04:55:35Z |3b|: it needs a separate "otherwise inaccessible" for the binding rather than the object
2016-09-16T04:55:54Z nyef: Right, that would help a lot.
2016-09-16T04:56:21Z |3b|: or 'part' needs to refer to the specific reference
2016-09-16T04:56:26Z nyef: Also more carefully explaining what happens with "free DX".
2016-09-16T04:56:26Z |3b|: (including bindings)
2016-09-16T04:57:44Z |3b|: ah, i guess you can POP the var in a free binding too
2016-09-16T04:57:52Z nyef: One of the (non-normative) examples is amusing, because a sufficiently smart compiler can infer (via escape analysis) that the value can be stack-allocated... or even not allocated at all.
2016-09-16T04:57:53Z |3b|: actually, nevermind
2016-09-16T04:58:06Z |3b|: can't pop it
2016-09-16T04:59:12Z |3b|: well, depending on whether the (if f established a binding) part is limiting that clause to only bound declarations
2016-09-16T04:59:22Z |3b|: if it isn't, then popping is allowed
2016-09-16T05:00:01Z nyef: That's the definition of a bound declaration, so yes it does limit.
2016-09-16T05:00:07Z |3b|: and if you fix the definition of 'otherwise inaccessible part' to mean removing the reference rather than removing /all/ references, you can set it to NIL
2016-09-16T05:00:17Z |3b|: no, i mean that clause needsto apply to both bound and free
2016-09-16T05:00:54Z nyef: I think I have this figured out. You can take additional references to the bound object in free DX, and you can mutate any part of the bound object /that has already been mutated since the binding was established/.
2016-09-16T05:01:00Z |3b|: if you POP the var in a free declaration, the POPped cons becomes inaccessible (assuming it had no other references) which should be allowed
2016-09-16T05:01:47Z nyef: We're not arguing about what the defined semantics should be, we're arguing about what they ARE. d-:
2016-09-16T05:01:51Z |3b|: but it is no longer "still ... part ...", so doesn't fall under the second clause
2016-09-16T05:02:25Z |3b|: right, what they /are/ doesn't allow POP by my reading
2016-09-16T05:02:36Z nyef: The second clause is for a free declaration only, it seems.
2016-09-16T05:02:39Z |3b|: unless you ignore the (if ...)
2016-09-16T05:03:03Z |3b|: right, but that is true without the (if ...) since it can't be a 'current value' of a variable that isn't in scope
2016-09-16T05:03:44Z |3b|: so if you treat the (if ...) as just supplemental rather than restrictive, you can POP
2016-09-16T05:04:05Z |3b|: but if it restricts the first clause to bound declarations, you can't
2016-09-16T05:04:15Z |3b|: and can only permute the value in place
2016-09-16T05:04:44Z |3b|: arbitrarily rearrange conses in a list, but not add or remove, or swap slots in a struct/class, but not add/remove values
2016-09-16T05:05:26Z |3b|: actually i guess you can add otherwise accessible values to the contents?
2016-09-16T05:06:13Z nyef: Heh. You can, and they can't be stack-allocated!
2016-09-16T05:06:38Z nyef: This is hilarious! (-:
2016-09-16T05:07:07Z |3b|: you just can't change the outer binding to any value that wasn't originally part of the object due to 'otherwise inaccessible part' being poorly specified in the case of the object itself
2016-09-16T05:08:12Z |3b|: hmm, i guess even if you change the definition of 'otherwise inaccessible part' to refer to references, multiple DX references to an object would still make things fun :)
2016-09-16T05:08:21Z |3b|: though probably still an improvement
2016-09-16T05:08:44Z nyef: There's also a bit about "the current value of" the variable.
2016-09-16T05:08:55Z nyef: Current at what time?
2016-09-16T05:09:08Z |3b| assumes when the declaration goes out of scope
2016-09-16T05:09:38Z nyef: Hah! You can PUSH to it!
2016-09-16T05:11:26Z |3b|: hmm, possibly so
2016-09-16T05:12:22Z |3b|: though i guess in that case, it is the compiler's responsibility to not stack allocate it
2016-09-16T05:13:16Z nyef: Yes, free DX is a bizarre set of constraints that actively inhibits stack allocation.
2016-09-16T05:13:17Z |3b|: since it is then a valid declaration, and the compiler has to maintain the semantics of CL
2016-09-16T05:13:39Z nyef: Hrm.
2016-09-16T05:13:39Z |3b|: yeah, not sure what it is supposed to do with free DX declarations
2016-09-16T05:13:59Z |3b|: i guess if you store the value and set it back at the end, you could DX intermediate assignments
2016-09-16T05:14:06Z nyef: clhs 3.3.1
2016-09-16T05:14:06Z specbot: Minimal Declaration Processing Requirements: http://www.lispworks.com/reference/HyperSpec/Body/03_ca.htm
2016-09-16T05:14:37Z |3b|: or if you clean up 'otherwise inaccessible part' you could DX intermediate values as long as you set it to an otherwise accessible value at the end
2016-09-16T05:16:32Z |3b|: (depending on whether making OIPs of the original value inaccessible is good enough or not)
2016-09-16T05:17:03Z nyef: Well, I think that I have a much better understanding of DX now, even if it IS a horribly broken mechanism. Thank you.
2016-09-16T05:19:21Z |3b|: actually, i guess the (intern ...) case falls under the "compiler's responsiblity to not DX it"?
2016-09-16T05:19:47Z |3b|: the symbol is otherwise inaccessible part of itself, so the declaration is valid, so compiler needsto maintain semantics?
2016-09-16T05:20:30Z nyef: Right. Declaration is valid: The symbol is no longer accessible once the scope closes.
2016-09-16T05:20:50Z nyef: If there are outside references, the programmer lied to the compiler.
2016-09-16T05:20:57Z |3b|: ah, right
2016-09-16T05:21:09Z |3b| is getting confused :)
2016-09-16T05:21:42Z nyef: Part of why you're getting confused is that you keep wanting bound DX to be sane.
2016-09-16T05:22:10Z |3b|: nah, i think that part is from trying to keep track of details of both free and bound at once
2016-09-16T05:23:05Z |3b|: though symbols are a complicated case anyway, since the way they and packages work internally isn't specified
2016-09-16T05:24:01Z |3b|: does the package have a reference to the name of the symbol? i don't think it has to, but it could (and is probably the most likely implementation)
2016-09-16T05:25:01Z |3b|: so hard to say if it is conforming to unintern that symbol and return its name
2016-09-16T05:27:11Z shka joined #sbcl
2016-09-16T05:37:27Z gingerale joined #sbcl
2016-09-16T06:36:12Z shka quit (Ping timeout: 244 seconds)
2016-09-16T07:03:16Z nicdev quit (Remote host closed the connection)
2016-09-16T07:23:54Z rudolfochrist joined #sbcl
2016-09-16T07:32:41Z scymtym joined #sbcl
2016-09-16T07:38:50Z angavrilov joined #sbcl
2016-09-16T07:45:54Z klltkr joined #sbcl
2016-09-16T07:54:43Z Bike quit (Quit: leaving)
2016-09-16T09:26:10Z sjl joined #sbcl
2016-09-16T09:38:51Z Quadrescence quit (Ping timeout: 265 seconds)
2016-09-16T09:43:01Z Quadrescence joined #sbcl
2016-09-16T09:43:16Z schjetne joined #sbcl
2016-09-16T09:44:36Z DeadTrickster_ quit (Ping timeout: 265 seconds)
2016-09-16T09:44:52Z DeadTrickster quit (Ping timeout: 255 seconds)
2016-09-16T09:58:08Z rudolfochrist quit (Ping timeout: 265 seconds)
2016-09-16T10:14:16Z DeadTrickster joined #sbcl
2016-09-16T10:14:28Z DeadTrickster_ joined #sbcl
2016-09-16T10:14:40Z sjl quit (Read error: Connection reset by peer)
2016-09-16T11:31:46Z madbub joined #sbcl
2016-09-16T11:34:17Z sjl joined #sbcl
2016-09-16T11:48:50Z scymtym: i have rewritten the xref packing one more (and hopefully final) time: https://github.com/scymtym/sbcl/commits/wip-compact-xref-and-repack . i'm not asking for another full review, but do TUNE-IMAGE-FOR-DUMP and the sb-introspect changes look OK?
2016-09-16T12:00:21Z rpg joined #sbcl
2016-09-16T12:25:27Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2016-09-16T12:54:05Z scymtym: dougk_: the SIMD FILL change breaks the x86 build, apparently because EMPTY_HT_SLOT is undefined: https://ci.cor-lab.org/job/sbcl-master/1749/featureset=1,label=ubuntu_trusty_32bit/console
2016-09-16T13:00:54Z rpg joined #sbcl
2016-09-16T13:11:29Z klltkr quit (Ping timeout: 244 seconds)
2016-09-16T13:31:07Z cromachina quit (Read error: Connection reset by peer)
2016-09-16T13:49:38Z klltkr joined #sbcl
2016-09-16T14:11:23Z rumbler31 joined #sbcl
2016-09-16T14:24:23Z madbub quit (Remote host closed the connection)
2016-09-16T14:43:51Z rumbler31 quit (Remote host closed the connection)
2016-09-16T14:46:10Z PosterdatiMobile joined #sbcl
2016-09-16T14:46:14Z _PosterdatiMobil joined #sbcl
2016-09-16T14:46:47Z _PosterdatiMobil quit (Client Quit)
2016-09-16T15:20:31Z PosterdatiMobile quit (Quit: KVIrc 4.2.0 Equilibrium http://www.kvirc.net/)
2016-09-16T15:36:50Z jdz quit (Quit: ZNC - http://znc.in)
2016-09-16T15:37:01Z jdz joined #sbcl
2016-09-16T15:51:09Z icerove joined #sbcl
2016-09-16T15:51:38Z icerove left #sbcl
2016-09-16T15:51:38Z madbub joined #sbcl
2016-09-16T15:52:58Z PosterdatiMobile joined #sbcl
2016-09-16T15:52:59Z _PosterdatiMobil joined #sbcl
2016-09-16T15:54:07Z PosterdatiMobile quit (Client Quit)
2016-09-16T15:54:07Z _PosterdatiMobil quit (Client Quit)
2016-09-16T16:07:56Z Bike joined #sbcl
2016-09-16T16:33:04Z sjl quit (Ping timeout: 265 seconds)
2016-09-16T16:43:33Z rumbler31 joined #sbcl
2016-09-16T16:52:37Z PosterdatiMobile joined #sbcl
2016-09-16T16:52:42Z _PosterdatiMobil joined #sbcl
2016-09-16T16:53:41Z _PosterdatiMobil quit (Client Quit)
2016-09-16T16:59:06Z klltkr quit (Ping timeout: 265 seconds)
2016-09-16T17:03:41Z rumbler31 quit (Remote host closed the connection)
2016-09-16T17:06:27Z sjl joined #sbcl
2016-09-16T17:09:06Z shka joined #sbcl
2016-09-16T17:13:06Z pkhuong quit (Ping timeout: 265 seconds)
2016-09-16T17:14:48Z pkhuong joined #sbcl
2016-09-16T17:15:34Z dougk_: scymtym: ok. i forgot that it wasn't a static symbol for every platform. probably might as well be
2016-09-16T17:15:38Z pkhuong is now known as Guest18921
2016-09-16T17:49:27Z scymtym quit (Ping timeout: 272 seconds)
2016-09-16T18:05:05Z slyrus joined #sbcl
2016-09-16T18:10:41Z scymtym joined #sbcl
2016-09-16T19:39:44Z rumbler31 joined #sbcl
2016-09-16T19:47:30Z shka quit (Ping timeout: 276 seconds)
2016-09-16T19:51:16Z _iwc joined #sbcl
2016-09-16T20:41:24Z rumbler31 quit (Remote host closed the connection)
2016-09-16T20:59:53Z stassats joined #sbcl
2016-09-16T21:02:15Z gingerale quit (Remote host closed the connection)
2016-09-16T21:09:30Z madbub quit (Ping timeout: 264 seconds)
2016-09-16T21:16:07Z edgar-rft quit (Quit: edgar-rft)
2016-09-16T21:56:30Z ASau joined #sbcl
2016-09-16T22:18:50Z fiddlerwoaroof quit (Read error: Connection reset by peer)
2016-09-16T23:36:52Z angavrilov quit (Remote host closed the connection)
2016-09-16T23:37:58Z cromachina joined #sbcl
2016-09-17T00:21:11Z sjl quit (Read error: Connection reset by peer)
2016-09-17T01:06:22Z pillton joined #sbcl
2016-09-17T01:17:21Z pillton: Does SBCL have any options for hinting that it is ok to move code to a better position? e.g. http://paste.lisp.org/display/326239
2016-09-17T01:18:02Z Guest18921 is now known as pkhuong
2016-09-17T01:18:09Z pkhuong: pillton: no
2016-09-17T01:19:14Z pillton: Ok. Thanks.
2016-09-17T01:35:34Z em1l joined #sbcl
2016-09-17T01:35:47Z nyef: ... Loop strength reduction + invariant lifting?
2016-09-17T01:38:52Z em1l_ quit (Ping timeout: 255 seconds)
2016-09-17T01:42:38Z pkhuong: and escape analysis because threads
2016-09-17T01:44:10Z pillton: I'm not 100% sure what these terms mean. I'm just trying to limit the number of operations performed inside a loop.
2016-09-17T01:44:50Z pillton: The loop is generated from a macro and it is possible to have nested loops.
2016-09-17T01:45:52Z pkhuong: I think you're more likely to get traction on specialised ways to let macros mark some sexps as CSE/invariant-able than having the compiler do all that in a transparent & emergent manner.
2016-09-17T01:46:00Z pkhuong: for that, we'd need a proposal (:
2016-09-17T01:47:03Z pillton: That is what I am contempating.
2016-09-17T01:48:54Z pillton: I already have the plumbing needed to avoid the call to the second make-array in http://paste.lisp.org/display/326239#1.
2016-09-17T01:49:01Z pkhuong: I think you could go very far if nested forms could somehow extend a surrounding "nested" scope.
2016-09-17T01:49:35Z pillton: I'm just pondering what I can do about nested loops.
2016-09-17T01:53:54Z pillton: I gotta go. I'll check back in later.
2016-09-17T02:01:25Z karswell quit (Remote host closed the connection)
2016-09-17T02:01:38Z pchrist quit (Remote host closed the connection)
2016-09-17T02:31:25Z stassats quit (Ping timeout: 260 seconds)
2016-09-17T02:39:29Z nyef: Probably should be a declaration, so user code can either condition it on #+sbcl or proclaim the declaration on #-sbcl.
2016-09-17T02:39:56Z nyef: (Or is it declaim the declaration? I always have a hard time remembering which is the macro and which the function.)
2016-09-17T02:40:29Z nyef: ... declaim.
2016-09-17T02:42:00Z nyef: Ah. declaim declaration looks to be idempotent?
2016-09-17T02:42:22Z nyef: So we could have a loop-invariant declaration for a binding?
2016-09-17T02:43:12Z nyef: (But, at that point, why not move the damned binding out of the loop in the first place?)
2016-09-17T03:29:33Z pkhuong: nyef: clos has some stuff like that
2016-09-17T03:29:37Z pkhuong: where we don't know if we need to compute a thing
2016-09-17T03:29:42Z pkhuong: but if we do, we want to cache it
2016-09-17T03:37:23Z nyef: Lovely.
2016-09-17T03:37:41Z karswell joined #sbcl
2016-09-17T04:12:24Z p_l quit (Ping timeout: 265 seconds)
2016-09-17T04:14:58Z p_l joined #sbcl
2016-09-17T04:15:23Z myrkraverk: Is there a documentation that explains how SBCL starts the RTE? Or is it just in the source code?
2016-09-17T04:15:38Z myrkraverk: As in, I'm curious about the boot process of SBCL.
2016-09-17T04:20:33Z nyef: The "RTE"?
2016-09-17T04:22:24Z pkhuong: runtime engine I assume.
2016-09-17T04:22:27Z pkhuong: *environment?
2016-09-17T04:23:52Z nyef: Yeah, "runtime environment" is my best guess, but it's a bit weak, hence my question.
2016-09-17T04:37:03Z nyef: It wouldn't surprise me if the only documentation on the matter is in the necronom^WEncyCMUCLopedia.
2016-09-17T04:56:33Z nyef: Now, if the question is "just" about the process of reloading a "warm core", that's easy enough to explain. If it's about the process of going from a pile of fasls to a "cold core" to a running system, that's less easy to explain.
2016-09-17T04:57:08Z nyef: Though I'm probably not going to make the attempt to explain either tonight, given the time.
2016-09-17T05:09:56Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2016-09-17T05:15:56Z myrkraverk: nyef: that'd be nice.
2016-09-17T05:16:02Z myrkraverk: And yes, I mean runtime environment.
2016-09-17T05:17:47Z edgar-rft joined #sbcl
2016-09-17T05:19:51Z nyef: myrkraverk: Which scenario are you asking about, though?
2016-09-17T05:46:21Z nyef: ... Time's up. Setting night-time running mode. Will check scrollback in the morning.
2016-09-17T05:46:22Z myrkraverk: I was wondering how SBCL goes from _start (or whatever the executable start symbol is called) to (toplevel) - and I'm not sure which scenario applies.
2016-09-17T05:46:29Z nyef: Oop.
2016-09-17T05:46:42Z nyef: Okay, the simple version, for a warm core, and then I crash.
2016-09-17T05:46:51Z myrkraverk: Ok.
2016-09-17T05:48:17Z nyef: The runtime sets up a few signal handlers, reads the core file header to find things like where the restart function is, the page table (on gencgc), and which pages need to be mapped where, mmap()s the various heap spaces, creates an initial thread structure, and then calls the initial function.
2016-09-17T05:50:15Z myrkraverk: Ok.
2016-09-17T05:50:56Z nyef: From there, we're at RESTART-LISP (a LABELS function) in SAVE-LISP-AND-DIE in code/save.lisp.
2016-09-17T05:51:25Z nyef: It reinitializes some things and then calls the toplevel.
2016-09-17T05:53:28Z myrkraverk: Right.
2016-09-17T05:53:37Z myrkraverk: I'll take a look at code/save.lisp now.
2016-09-17T05:54:54Z nyef: Good luck.
2016-09-17T05:59:09Z edgar-rft quit (Quit: edgar-rft)
2016-09-17T06:11:43Z gingerale joined #sbcl
2016-09-17T07:15:02Z benkard joined #sbcl
2016-09-17T07:21:51Z edgar-rft joined #sbcl
2016-09-17T07:33:06Z shka joined #sbcl
2016-09-17T07:37:37Z benkard quit (Quit: Textual IRC Client: www.textualapp.com)
2016-09-17T07:47:58Z benkard joined #sbcl
2016-09-17T08:44:25Z benkard quit (Ping timeout: 248 seconds)
2016-09-17T08:57:56Z Bike quit (Quit: leaving)
2016-09-17T08:57:58Z benkard joined #sbcl
2016-09-17T09:02:33Z benkard quit (Ping timeout: 248 seconds)
2016-09-17T09:13:00Z benkard joined #sbcl
2016-09-17T09:25:09Z angavrilov joined #sbcl
2016-09-17T09:45:14Z pchrist joined #sbcl
2016-09-17T10:02:11Z klltkr joined #sbcl
2016-09-17T10:07:47Z benkard quit (Quit: Textual IRC Client: www.textualapp.com)
2016-09-17T10:27:45Z sjl joined #sbcl
2016-09-17T10:33:07Z klltkr quit (Ping timeout: 248 seconds)
2016-09-17T10:44:36Z attila_lendvai joined #sbcl
2016-09-17T10:44:36Z attila_lendvai quit (Changing host)
2016-09-17T10:44:36Z attila_lendvai joined #sbcl
2016-09-17T10:59:47Z attila_lendvai quit (Ping timeout: 248 seconds)
2016-09-17T11:07:35Z scymtym quit (Ping timeout: 265 seconds)
2016-09-17T11:40:51Z scymtym joined #sbcl
2016-09-17T12:00:21Z rpg joined #sbcl
2016-09-17T12:06:25Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2016-09-17T12:09:04Z scymtym: we seem to have two regressions: 1. test failures on all platforms after "Prefer using a fixnum for layout-bitmap if possible" https://ci.cor-lab.org/job/sbcl-master/1750/ 2. everything fails to build since "Fix some disassembler cruft." except with clisp host of all things: https://ci.cor-lab.org/job/sbcl-master/1752/
2016-09-17T12:14:33Z ASau quit (Ping timeout: 240 seconds)
2016-09-17T12:27:04Z shka quit (Quit: Konversation terminated!)
2016-09-17T12:28:44Z shka joined #sbcl
2016-09-17T12:29:39Z shka quit (Client Quit)
2016-09-17T12:31:14Z shka joined #sbcl
2016-09-17T12:41:46Z shka quit (Quit: Konversation terminated!)
2016-09-17T12:42:54Z shka joined #sbcl
2016-09-17T12:45:05Z sjl quit (Ping timeout: 260 seconds)
2016-09-17T13:34:10Z sjl joined #sbcl
2016-09-17T14:17:26Z myrkraverk_ joined #sbcl
2016-09-17T14:17:39Z myrkraverk quit (Ping timeout: 248 seconds)
2016-09-17T14:17:58Z myrkraverk_ is now known as myrkraverk
2016-09-17T14:36:38Z DougNYC joined #sbcl
2016-09-17T14:54:20Z nicdev joined #sbcl
2016-09-17T15:07:56Z ASau joined #sbcl
2016-09-17T15:53:29Z sjl quit (Ping timeout: 244 seconds)
2016-09-17T16:06:19Z Bike joined #sbcl
2016-09-17T16:42:25Z edgar-rft quit (Quit: edgar-rft)
2016-09-17T16:43:37Z edgar-rft joined #sbcl
2016-09-17T17:15:33Z sjl joined #sbcl
2016-09-17T17:19:48Z klltkr joined #sbcl
2016-09-17T17:42:41Z scymtym quit (Ping timeout: 272 seconds)
2016-09-17T18:24:34Z scymtym joined #sbcl
2016-09-17T18:27:54Z edgar-rft quit (Quit: edgar-rft)
2016-09-17T18:37:25Z ASau quit (Ping timeout: 244 seconds)
2016-09-17T18:49:23Z ASau joined #sbcl
2016-09-17T18:59:01Z rpg joined #sbcl
2016-09-17T19:06:08Z klltkr quit (Ping timeout: 265 seconds)
2016-09-17T19:51:23Z edgar-rft joined #sbcl
2016-09-17T19:51:29Z shka quit (Ping timeout: 265 seconds)
2016-09-17T20:26:17Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2016-09-17T20:53:10Z edgar-rft quit (Quit: edgar-rft)
2016-09-17T21:03:28Z karswell` joined #sbcl
2016-09-17T21:07:51Z karswell quit (Ping timeout: 276 seconds)
2016-09-17T21:38:24Z karswell` quit (Remote host closed the connection)
2016-09-17T21:46:19Z rpg joined #sbcl
2016-09-17T22:54:45Z gingerale quit (Remote host closed the connection)
2016-09-18T00:50:24Z fiddlerwoaroof joined #sbcl
2016-09-18T00:54:03Z DougNYC quit (Remote host closed the connection)
2016-09-18T01:21:33Z sjl quit (Ping timeout: 240 seconds)
2016-09-18T01:26:15Z fiddlerwoaroof: I've noticed when strac-ing an sbcl process that sbcl periodically receives SIGSEGV signals
2016-09-18T01:26:17Z fiddlerwoaroof: http://paste.lisp.org/+6ZTZ
2016-09-18T01:26:52Z fiddlerwoaroof: Is this intended behavior?  Just by guessing, I suspect it has something to do with the way SBCL's gc works
2016-09-18T01:29:22Z nyef: Yes, SBCL uses SIGSEGV as part of its memory barrier system, among other things.
2016-09-18T01:30:44Z fiddlerwoaroof: ok, that's interesting
2016-09-18T01:31:07Z DougNYC joined #sbcl
2016-09-18T01:34:13Z fiddlerwoaroof: So, are the mprotect() calls used to setup a way to signal SBCL to run gc/allocate more memory?
2016-09-18T01:34:17Z em1l_ joined #sbcl
2016-09-18T01:34:55Z nyef: Somewhat. Sometimes it is used to say "next time that GC runs, scan this page for pointers".
2016-09-18T01:35:53Z DougNYC quit (Ping timeout: 240 seconds)
2016-09-18T01:37:04Z nyef: Or, rather, "next time that there is a write to this page, trap, so that we can mark it as needing to be scanned for pointers". Something like that.
2016-09-18T01:38:04Z em1l quit (Ping timeout: 265 seconds)
2016-09-18T01:43:36Z nyef: ... I find myself suddenly knowing how to make a version of SBCL with no C runtime.
2016-09-18T01:43:50Z nyef: It doesn't seem worth doing.
2016-09-18T01:44:30Z nyef: At least, not at this point.
2016-09-18T01:53:46Z karswell` joined #sbcl
2016-09-18T02:05:15Z DougNYC joined #sbcl
2016-09-18T02:07:07Z DougNYC quit (Remote host closed the connection)
2016-09-18T02:07:21Z DougNYC joined #sbcl
2016-09-18T02:08:39Z DougNYC quit (Remote host closed the connection)
2016-09-18T02:09:15Z DougNYC joined #sbcl
2016-09-18T02:13:41Z DougNYC quit (Ping timeout: 250 seconds)
2016-09-18T02:48:46Z rpg quit (Ping timeout: 244 seconds)
2016-09-18T03:20:17Z sigjuice quit (Quit: ZNC - http://znc.in)
2016-09-18T03:21:31Z sigjuice joined #sbcl
2016-09-18T05:22:35Z eschatologist quit (Ping timeout: 244 seconds)
2016-09-18T05:27:42Z eschatologist joined #sbcl
2016-09-18T05:38:50Z DougNYC joined #sbcl
2016-09-18T05:44:03Z DougNYC quit (Ping timeout: 240 seconds)
2016-09-18T06:32:19Z gingerale joined #sbcl
2016-09-18T07:05:17Z karswell` quit (Remote host closed the connection)
2016-09-18T07:05:31Z karswell` joined #sbcl
2016-09-18T07:12:19Z Intensity quit (Ping timeout: 255 seconds)
2016-09-18T07:13:21Z shka joined #sbcl
2016-09-18T07:17:18Z myrkraverk quit (Ping timeout: 264 seconds)
2016-09-18T07:17:50Z myrkraverk joined #sbcl
2016-09-18T07:47:23Z schjetne quit (Ping timeout: 265 seconds)
2016-09-18T07:57:12Z shrdlu68 joined #sbcl
2016-09-18T07:58:57Z shrdlu68: Hi guys, I'm getting mmap ensure-space errors on a machine, advising me to increase memory limits. I've tried setting ulimit -s unlimited simply to test whether it works, but no dice.
2016-09-18T07:59:35Z flip214: shrdlu68: I guess that SBCL's --dynamic-space-size argument is meant instead.
2016-09-18T08:00:42Z shrdlu68: flip214: Okay, let me try that.
2016-09-18T08:01:35Z shrdlu68: flip214: How do I tell the default --dynamic-space-size, to gauge by how much to increase it?
2016-09-18T08:02:00Z flip214: shrdlu68: "ulimit -s" is stack space only... perhaps you need to use -l, -d, -m, -v instead?
2016-09-18T08:03:55Z shrdlu68: flip214: -d -m and -v are unlimited, let me try -l
2016-09-18T08:04:09Z flip214: (print (sb-ext:dynamic-space-size))
2016-09-18T08:05:10Z flip214: is that in some vserver environment? I've seen problems with some of them, because the "big" virtual memory allocation breaks in some way.
2016-09-18T08:05:21Z shrdlu68: Still doesn't work.
2016-09-18T08:05:24Z flip214: shrdlu68: soft/hard limits checked?
2016-09-18T08:05:28Z flip214: what's the exact error message?
2016-09-18T08:05:41Z shrdlu68: mmap: wanted 1040384 bytes at 0x20000000, actually mapped at 0x36bf1c36c000
2016-09-18T08:05:41Z shrdlu68: ensure_space: failed to validate 1040384 bytes at 0x20000000
2016-09-18T08:05:47Z shrdlu68: (hint: Try "ulimit -a"; maybe you should increase memory limits.)
2016-09-18T08:06:14Z shrdlu68: It's not a virtual environment.
2016-09-18T08:07:13Z flip214: is that when starting SBCL, or when loading some ASDF system? which one?
2016-09-18T08:08:17Z flip214: shrdlu68: https://bugzilla.redhat.com/show_bug.cgi?id=879665 ?
2016-09-18T08:08:23Z shrdlu68: In one instance, when launching a saved sbcl executable image. In another, while trying to ./make.sh the sbcl git repo.
2016-09-18T08:09:06Z flip214: https://github.com/dimitri/pgloader/issues/112 says to reduce the dynsize as well...
2016-09-18T08:09:11Z flip214: even during building.
2016-09-18T08:09:48Z shrdlu68: ...reduce?
2016-09-18T08:14:30Z jsnell: that won't help, the error is not for the dynamic space
2016-09-18T08:15:33Z jsnell: you'd need to edit src/compiler/x86-64/parms.lisp, change the #x20000000 to something else, and recompile
2016-09-18T08:16:11Z jsnell: or alternatively figure out what is being mapped at that address by the system, and why
2016-09-18T08:16:18Z shrdlu68: The saved lisp image runs fine on my local machine, which has roughly the same ulimit values.
2016-09-18T08:16:40Z jsnell: it has nothing to do with resource limits
2016-09-18T08:17:29Z shrdlu68: Okay, let me try to upgrade my kernel and see whether the problem persists.
2016-09-18T08:17:39Z jsnell: there is already something mapped in the address space that conflicts with the memory mapping sbcl wants to do starting at 0x20000000
2016-09-18T08:18:35Z Bike quit (Quit: leaving)
2016-09-18T08:19:53Z jsnell: the simple first step is to do "cat /proc/self/maps", and see if another program ("cat" in this case) also has that conflicting mapping around
2016-09-18T08:27:05Z shrdlu68: jsnell: I don't know how to interpret the output, here: http://vhbin.net/k3xmwpfusb4e
2016-09-18T08:34:32Z edgar-rft joined #sbcl
2016-09-18T08:35:10Z jsnell: the first column is what you want to look at. those are the address ranges that are mapped for that process. what you were hoping to see is one that overlaps with the range 0x20000000-0x20100000
2016-09-18T08:36:05Z jsnell: (but there isn't one, so whatever is getting mapped there isn't happening for all programs on your system, just sbcl)
2016-09-18T08:38:11Z ASau quit (Remote host closed the connection)
2016-09-18T08:38:49Z ASau joined #sbcl
2016-09-18T09:15:19Z DougNYC joined #sbcl
2016-09-18T09:17:12Z shrdlu68: Upgraded my kernel and all is fine now.
2016-09-18T09:20:37Z DougNYC quit (Ping timeout: 265 seconds)
2016-09-18T09:38:57Z Intensity joined #sbcl
2016-09-18T09:53:35Z p_l quit (Ping timeout: 272 seconds)
2016-09-18T09:54:54Z p_l joined #sbcl
2016-09-18T10:10:01Z shrdlu68 quit (Quit: leaving)
2016-09-18T10:10:41Z p_l quit (Ping timeout: 272 seconds)
2016-09-18T10:12:43Z p_l joined #sbcl
2016-09-18T10:21:29Z p_l quit (Ping timeout: 248 seconds)
2016-09-18T10:24:04Z p_l joined #sbcl
2016-09-18T11:07:55Z ASau quit (Ping timeout: 265 seconds)
2016-09-18T11:21:45Z sjl joined #sbcl
2016-09-18T11:26:18Z scymtym quit (Ping timeout: 264 seconds)
2016-09-18T11:31:03Z PosterdatiMobile quit (Ping timeout: 240 seconds)
2016-09-18T11:41:11Z schjetne joined #sbcl
2016-09-18T11:48:29Z Intensity quit (Remote host closed the connection)
2016-09-18T11:53:24Z p_l quit (Ping timeout: 265 seconds)
2016-09-18T11:56:23Z p_l joined #sbcl
2016-09-18T12:07:13Z scymtym joined #sbcl
2016-09-18T12:59:28Z klltkr joined #sbcl
2016-09-18T13:49:22Z Cthulhux: sbcl's launchpad bug tracker does not let me in. "Bad bot, go away! Request aborted."
2016-09-18T13:49:32Z Cthulhux: lovely canonical quality software.
2016-09-18T14:03:22Z scymtym: Cthulhux: what's the bug?
2016-09-18T14:03:55Z Cthulhux: scymtym, how can i report bugs with your bugtracker? :-)
2016-09-18T14:04:15Z Cthulhux: beside that, i only have a feature request (zlib for windows support :D)
2016-09-18T14:04:19Z Cthulhux: nothing important
2016-09-18T14:04:59Z ASau joined #sbcl
2016-09-18T14:08:24Z stassats joined #sbcl
2016-09-18T14:10:04Z scymtym: Cthulhux: i see. unfortunately, looking into that would require having a usable win32 development environment for sbcl
2016-09-18T14:10:50Z Cthulhux: adding zlib works just like other platforms here..
2016-09-18T14:13:41Z scymtym: are you suggesting adding it to the prebuilt binaries, then?
2016-09-18T14:14:27Z Shinmera: Cthulhux: this may be roundabout, but Portacle does include an SBCL with zlib included. https://github.com/Shinmera/portacle/releases/tag/0.0.7-2
2016-09-18T14:14:44Z Cthulhux: yes, i do. having it "in the source" would not make much sense. people who know how to build sbcl are probably capable of adding zlib themselbes
2016-09-18T14:14:47Z Cthulhux: *v
2016-09-18T14:15:23Z stassats: Cthulhux: my response would be "wontfix"
2016-09-18T14:15:47Z Cthulhux: Shinmera, nice but just a workaround. (i didn't know portacle but i prefer having it added to my environment anyway.)
2016-09-18T14:15:51Z Cthulhux: stassats, why?
2016-09-18T14:17:03Z DougNYC joined #sbcl
2016-09-18T14:18:09Z stassats: because it wouldn't work without zlib
2016-09-18T14:18:42Z Cthulhux: statically linked zlib wouldn't care.
2016-09-18T14:19:00Z stassats: but sbcl doesn't link it statically
2016-09-18T14:19:17Z Cthulhux: not yet.
2016-09-18T14:19:27Z Cthulhux: but it really should :-)
2016-09-18T14:19:32Z stassats: well, then it's really has nothing to do with win32
2016-09-18T14:19:52Z stassats: we don't build with zlib for any other platform either
2016-09-18T14:20:21Z Cthulhux: why not?
2016-09-18T14:20:35Z stassats: because it wouldn't work without zlib
2016-09-18T14:21:13Z Cthulhux: why would it have to?
2016-09-18T14:26:39Z cromachina: sbcl doesn't seem to use it, it's not part of the standard, and it can be provided as a library that the user can pick the version of
2016-09-18T14:28:18Z Cthulhux: yes it does. https://github.com/sbcl/sbcl/blob/master/src/code/save.lisp#L173
2016-09-18T14:29:26Z cromachina: ah i see
2016-09-18T14:31:15Z myrkraverk: Cthulhux: Well, what typically happens in the free software world is this: you get to fix it yourself.  If you cannot, you get to fund it.  There's no other easy way to make sure it happens.
2016-09-18T14:32:22Z Cthulhux: myrkraverk, there's nothing to "fix" because it's not "broken", technically. that's why it would be a "feature request", not a bug report. but i can't release official binaries myself.
2016-09-18T14:34:11Z cromachina: and the goal is so image compression can be used?
2016-09-18T14:34:28Z scymtym_ joined #sbcl
2016-09-18T14:34:33Z myrkraverk: No, you cannot -- and I can't either.  All I'm saying is that it usually takes effort to get stuff done, and not just wishes.
2016-09-18T14:35:13Z Cthulhux: cromachina, yes.
2016-09-18T14:35:58Z myrkraverk: Cthulhux: I'd actually check what's involved, by building my own SBCL with zlib on windows (I don't have windows) and only then propose patches, or changes.  But that's me.
2016-09-18T14:36:53Z Cthulhux: myrkraverk, i can't officially propose things because launchpad doesn't let me in. but no patches are really needed because sbcl does everything. it just needs to default to using it.
2016-09-18T14:37:48Z Cthulhux: cromachina, (this might be just me and my users but obviously small binary sizes are still a thing in 2016. one of my applications had bad reviews because i wrote it in electron, 70 mib of webkit for its runtime..)
2016-09-18T14:38:23Z scymtym quit (Remote host closed the connection)
2016-09-18T14:39:14Z myrkraverk: Cthulhux: It might be adding or changing the detection logic in the build scripts.
2016-09-18T14:39:32Z myrkraverk: That's why I said it's good to know what you're actually asking for -- in code.
2016-09-18T14:40:18Z stassats: Cthulhux: well, not really
2016-09-18T14:40:32Z stassats: it can't default to using it because it would need the headers to be present
2016-09-18T14:40:46Z stassats: it can't default to it, because it doesn't statically link it
2016-09-18T14:41:31Z scymtym_: stassats: to make arrays which are displaced to non-simple arrays non-horrible, i tried the following and it seems to work well: http://paste.lisp.org/display/326437 do you see any reason we cannot do something like this?
2016-09-18T14:42:29Z stassats: an overview?
2016-09-18T14:42:52Z scymtym_: manage backpointers in a weak hash-table
2016-09-18T14:43:42Z scymtym_: currently, there is linear for broken weak pointers each time a backpointer is added
2016-09-18T14:43:57Z scymtym_: *linear scan
2016-09-18T14:44:36Z stassats: ok, what about concurrent access?
2016-09-18T14:45:37Z scymtym_: didn't think of that. i will think about it now
2016-09-18T14:46:33Z DougNYC quit (Ping timeout: 248 seconds)
2016-09-18T14:46:57Z scymtym_: isn't that already a problem with the current problem?
2016-09-18T14:47:14Z scymtym_: it just sets the backpointer list, no cas
2016-09-18T14:48:05Z scymtym_: *the current code
2016-09-18T14:51:19Z stassats: then it'd be probably a good idea to resolve that
2016-09-18T14:51:34Z stassats: but i actually never liked the business with those back links
2016-09-18T14:52:26Z scymtym_: it's a bit silly. just to allow incompatible adjustments of the backing to invalidate the displaced arrays
2016-09-18T14:55:25Z scymtym_: maybe as a first step, we could avoid the thing if the backing array has a header but is not adjustable. isn't that the case for rank > 1 arrays?
2016-09-18T14:57:06Z stassats: right
2016-09-18T15:19:13Z Intensity joined #sbcl
2016-09-18T15:31:17Z sjl quit (Read error: Connection reset by peer)
2016-09-18T15:52:06Z slyrus quit (Ping timeout: 264 seconds)
2016-09-18T15:53:49Z jackdaniel: hey! is it me, or SBCL's sb-ext:run-program can't run "fc-match" even if it's in the PATH? (giving absolute pathname works fine)
2016-09-18T15:54:14Z jackdaniel: (sb-ext:run-program "fc-match" '("-v")) ; -> Couldn't execute fc-match: no such file or directory
2016-09-18T15:54:34Z stassats: you've never used sb-ext:run-program before?
2016-09-18T15:54:54Z jackdaniel: no, not really
2016-09-18T15:55:22Z stassats: well, it needs :search
2016-09-18T15:56:46Z jackdaniel: ah, OK, when I give the pathname to :search it seems to work, thanks. So basically I should get environment variable PATH, splice it and feed :search with it?
2016-09-18T15:57:03Z stassats: huh?
2016-09-18T15:57:53Z jackdaniel: if I want to run applications found in paths put in PATH environment variable
2016-09-18T15:58:45Z jackdaniel: ah, search is a boolean, sorry! and thanks (I've called it like (sb-ext:run-program "fc-match" '("-v") :search "/usr/bin/")
2016-09-18T15:59:07Z Bike joined #sbcl
2016-09-18T15:59:09Z jackdaniel: and assumed you put there paths, nvm. Thank you
2016-09-18T15:59:26Z stassats: well, there's documentation
2016-09-18T15:59:27Z stassats: http://sbcl.org/manual/index.html#Running-external-programs
2016-09-18T15:59:37Z stassats: if you don't want to guess, that is
2016-09-18T15:59:53Z jackdaniel: stassats: yes, my bad, sorry for bothering you
2016-09-18T16:00:04Z stassats: no problem
2016-09-18T16:00:04Z jackdaniel: (and thank you for help)
2016-09-18T16:02:52Z scymtym_: checking for adjustablility in the backing array makes (time (let ((to (make-array '(10 10)))) (loop :repeat 10000 :collect (make-array 100 :displaced-to to)) nil)) go from 1s/800MB to 1ms/1MB
2016-09-18T16:02:52Z stassats: well, the build is broken
2016-09-18T16:03:03Z stassats: The bounding indices 0 and 256 are bad for a sequence of length 0.
2016-09-18T16:03:42Z scymtym_: which commit?
2016-09-18T16:04:17Z stassats: 800MB? strange
2016-09-18T16:05:35Z scymtym_: there can be a few thousand backpointers, then REMOVE some
2016-09-18T16:05:45Z stassats: scymtym_: the current one
2016-09-18T16:06:48Z scymtym_: yeah, i see it on the CI server now. sorry. i built that stuff a million times to make sure i didn't break anything
2016-09-18T16:06:58Z scymtym_ sighs
2016-09-18T16:07:44Z scymtym_: ah, because i always build with --fancy
2016-09-18T16:27:34Z scymtym_: stassats: i pushed a fix
2016-09-18T16:28:48Z scymtym_: btw is raw-slots-interleaved.impure.lisp failing for anybody else?
2016-09-18T16:29:26Z scymtym_: (from before me breaking the build)
2016-09-18T16:31:33Z stassats: sure does
2016-09-18T16:32:54Z scymtym_: ok, thanks. i sometimes mess up patching the test runner or something similar and get false positives
2016-09-18T16:38:49Z stassats: an idea, what if we give the job of fixing up shrunk displaced targets to the GC?
2016-09-18T16:39:27Z stassats: shrink-vector doesn't remove the space before the GC is run
2016-09-18T16:40:16Z stassats: and since the consequences are undefined, it will see old values, but not garbage or getting onto protected pages, or anything like that
2016-09-18T16:40:31Z scymtym_: how is SHRINK-VECTOR involved?
2016-09-18T16:40:47Z stassats: shrink-vector is the reason this is done
2016-09-18T16:41:11Z scymtym_: is that what ADJUST-ARRAY does when the size decreases?
2016-09-18T16:41:34Z stassats: yes
2016-09-18T16:41:52Z scymtym_: ok, i see
2016-09-18T16:44:12Z scymtym_: so, keep the backpointers, managing them with cas for thread-safety and have the gc invalidate displaced arrays without the wasted effort of scanning through weak pointers on MAKE-ARRAY?
2016-09-18T16:44:26Z stassats: no backpointers, actually
2016-09-18T16:44:48Z stassats: when the GC encounters a displaced array it checks that the displaced is still valid
2016-09-18T16:44:57Z scymtym_: ah, because we scan the displaced arrays
2016-09-18T16:45:52Z scymtym_: that sounds much better than any lisp-side improvement we have discussed
2016-09-18T17:02:14Z attila_lendvai joined #sbcl
2016-09-18T17:10:40Z edgar-rft quit (Quit: edgar-rft)
2016-09-18T17:13:22Z ASau` joined #sbcl
2016-09-18T17:14:44Z ASau quit (Ping timeout: 244 seconds)
2016-09-18T17:18:15Z des_consolado quit (Quit: leaving)
2016-09-18T17:27:02Z White_Flame quit (*.net *.split)
2016-09-18T17:27:02Z dustinm` quit (*.net *.split)
2016-09-18T17:29:12Z White_Flame joined #sbcl
2016-09-18T17:29:12Z dustinm` joined #sbcl
2016-09-18T17:35:13Z stassats: but adjust-array has other cases than just shrink-vector
2016-09-18T17:35:22Z stassats: like going from a displaced vector to a non displaced one
2016-09-18T17:35:28Z stassats: or to a different displacement
2016-09-18T17:41:36Z stassats: i think that can be handled in with-array-data
2016-09-18T17:42:31Z stassats: by saving some sort of a token, to check if the arrays are synchronized
2016-09-18T17:44:54Z scymtym_ goes afk for a bit
2016-09-18T17:45:48Z stassats: actually, not a toke, but saving length+offset and then checking if the array it goes after can accommodate it
2016-09-18T17:46:20Z stassats: though, that makes AREF slower, it's not really fast as it is
2016-09-18T17:50:29Z scymtym_ quit (Ping timeout: 272 seconds)
2016-09-18T17:53:59Z DougNYC joined #sbcl
2016-09-18T17:55:46Z edgar-rft joined #sbcl
2016-09-18T17:58:33Z DougNYC quit (Ping timeout: 240 seconds)
2016-09-18T18:12:36Z stassats: i think i can actually arrange things so that it doesn't need any magic
2016-09-18T18:12:48Z stassats: just check the bounds after all the displacements are unraveled
2016-09-18T18:16:07Z ASau` is now known as ASau
2016-09-18T18:24:02Z scymtym joined #sbcl
2016-09-18T19:22:39Z edgar-rft quit (Quit: edgar-rft)
2016-09-18T19:23:04Z White_Flame quit (Ping timeout: 250 seconds)
2016-09-18T19:23:16Z White_Flame joined #sbcl
2016-09-18T19:29:07Z gingerale quit (Remote host closed the connection)
2016-09-18T19:40:19Z attila_lendvai quit (Ping timeout: 248 seconds)
2016-09-18T20:25:58Z attila_lendvai joined #sbcl
2016-09-18T20:36:33Z pipping joined #sbcl
2016-09-18T20:46:34Z les quit (Ping timeout: 244 seconds)
2016-09-18T20:48:41Z les joined #sbcl
2016-09-18T21:18:39Z shka quit (Ping timeout: 276 seconds)
2016-09-18T21:19:38Z nyef quit (Ping timeout: 244 seconds)
2016-09-18T21:42:24Z attila_lendvai quit (Quit: Leaving.)
2016-09-18T22:29:23Z klltkr quit (Ping timeout: 248 seconds)
2016-09-18T23:34:51Z nyef joined #sbcl
2016-09-19T00:05:18Z karswell` quit (Ping timeout: 244 seconds)
2016-09-19T01:25:19Z stassats quit (Ping timeout: 265 seconds)
2016-09-19T01:33:04Z em1l joined #sbcl
2016-09-19T01:36:37Z em1l_ quit (Ping timeout: 272 seconds)
2016-09-19T02:55:15Z DougNYC joined #sbcl
2016-09-19T02:59:52Z DougNYC quit (Ping timeout: 255 seconds)
2016-09-19T02:59:54Z slyrus joined #sbcl
2016-09-19T03:53:07Z karswell joined #sbcl
2016-09-19T04:25:01Z nzambe quit (Quit: http://www.kiwiirc.com/ - A hand crafted IRC client)
2016-09-19T04:33:44Z shka joined #sbcl
2016-09-19T05:36:02Z DougNYC joined #sbcl
2016-09-19T05:41:03Z DougNYC quit (Ping timeout: 240 seconds)
2016-09-19T05:59:24Z edgar-rft joined #sbcl
2016-09-19T06:00:24Z scymtym quit (Ping timeout: 244 seconds)
2016-09-19T06:04:12Z slyrus quit (Ping timeout: 265 seconds)
2016-09-19T06:08:17Z shka quit (Ping timeout: 240 seconds)
2016-09-19T06:10:24Z slyrus joined #sbcl
2016-09-19T06:39:34Z scymtym joined #sbcl
2016-09-19T06:39:50Z ASau quit (Ping timeout: 260 seconds)
2016-09-19T07:24:54Z DougNYC joined #sbcl
2016-09-19T07:29:18Z DougNYC quit (Ping timeout: 244 seconds)
2016-09-19T07:34:34Z Bike quit (Quit: fall into)
2016-09-19T08:12:42Z Xof quit (Ping timeout: 244 seconds)
2016-09-19T09:15:12Z klltkr joined #sbcl
2016-09-19T10:03:38Z Xof joined #sbcl
2016-09-19T11:00:40Z DougNYC joined #sbcl
2016-09-19T11:05:33Z DougNYC quit (Ping timeout: 240 seconds)
2016-09-19T11:49:58Z stassats joined #sbcl
2016-09-19T11:50:16Z stassats: good news, i can reproduce that osx thing
2016-09-19T11:53:25Z stassats: and i see what's going on
2016-09-19T12:23:43Z madbub joined #sbcl
2016-09-19T12:49:20Z DougNYC joined #sbcl
2016-09-19T12:52:58Z sjl joined #sbcl
2016-09-19T12:53:33Z DougNYC quit (Ping timeout: 240 seconds)
2016-09-19T12:55:50Z rpg joined #sbcl
2016-09-19T13:07:27Z nzambe joined #sbcl
2016-09-19T13:16:48Z stassats: got some failure on the first test run, but can't catch it again
2016-09-19T13:16:53Z stassats: damn
2016-09-19T13:28:20Z cromachina quit (Read error: Connection reset by peer)
2016-09-19T14:12:50Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2016-09-19T14:27:52Z stassats: ok, can reproduce it now
2016-09-19T14:28:07Z karswell quit (Remote host closed the connection)
2016-09-19T14:28:22Z karswell joined #sbcl
2016-09-19T14:30:37Z sjl quit (Ping timeout: 255 seconds)
2016-09-19T14:37:33Z DougNYC joined #sbcl
2016-09-19T14:41:55Z rtmpdavid joined #sbcl
2016-09-19T14:59:03Z DougNYC quit (Ping timeout: 240 seconds)
2016-09-19T15:00:50Z klltkr quit (Quit: Lost terminal)
2016-09-19T15:01:20Z rtmpdavid quit (Remote host closed the connection)
2016-09-19T15:18:31Z rpg joined #sbcl
2016-09-19T15:26:00Z Bike joined #sbcl
2016-09-19T15:59:03Z rpg quit (Ping timeout: 240 seconds)
2016-09-19T16:10:21Z stassats: os_sem_init(0xdf72040): Operation timed out
2016-09-19T16:10:22Z stassats: huh
2016-09-19T16:10:33Z stassats: i can't imagine that all this is caused by my changes
2016-09-19T16:14:23Z stassats: but i can't really isolate that because without the changes it fails to start up
2016-09-19T16:15:49Z stassats: but x86 does start, ok
2016-09-19T16:18:55Z sjl joined #sbcl
2016-09-19T16:20:46Z slyrus: stassats: any idea why this is just showing up now?
2016-09-19T16:20:58Z stassats: slyrus: new xcode
2016-09-19T16:21:38Z stassats: this was the problem https://bugs.launchpad.net/sbcl/+bug/1470996
2016-09-19T16:21:42Z slyrus: ah right... that must mean I haven't built (outside of dockerized linux builds) on this box in a few days...
2016-09-19T16:22:01Z stassats: so, i did all that and it starts up
2016-09-19T16:22:07Z stassats: but sometimes something goes wrong
2016-09-19T16:22:23Z slyrus: fun!
2016-09-19T16:22:26Z stassats: pthread_create writes into some unsafe place
2016-09-19T16:22:45Z stassats: that shouldn't even be related
2016-09-19T16:23:39Z stassats: ok, with my changes, x86 fails the test, trying without them, but the test is long and not exactly deterministic
2016-09-19T16:24:10Z stassats: maybe it runs out of mach ports or something
2016-09-19T16:25:57Z stassats: (loop (sb-thread:make-thread (lambda () (sleep 1000000000000)))) seem to be bailing out sooner than before
2016-09-19T16:26:29Z gingerale joined #sbcl
2016-09-19T16:28:01Z stassats: ok, x86 fails with or without my changes, but built with the latest clang
2016-09-19T16:28:32Z stassats: well, (progn (defparameter *x* 0) (handler-case (loop (sb-thread:make-thread (lambda () (sleep 1000000000000))) (incf *x*)) (error () *x*))) is actually 2046 in both cases
2016-09-19T16:29:32Z stassats: but now it takes 0.157 seconds of real time as opposed to 37.337 seconds of real time
2016-09-19T16:29:42Z stassats: so, thread creation is now faster exposing a race condition of sorts?
2016-09-19T16:38:23Z stassats: now getting Cannot set thread_exception_port context, that's a new assertion
2016-09-19T16:46:29Z Quadresce joined #sbcl
2016-09-19T16:47:19Z shka joined #sbcl
2016-09-19T17:02:36Z rpg joined #sbcl
2016-09-19T17:03:16Z nzambe quit (Quit: http://www.kiwiirc.com/ - A hand crafted IRC client)
2016-09-19T17:43:05Z stassats: it starts to look more like a kernel bug
2016-09-19T17:49:08Z nzambe joined #sbcl
2016-09-19T18:10:12Z stassats: the failure mode does not make any sense
2016-09-19T18:11:30Z rumbler31 joined #sbcl
2016-09-19T18:11:53Z slyrus quit (Ping timeout: 248 seconds)
2016-09-19T18:13:28Z stassats: it allocates a pthread but for some reason in protected space
2016-09-19T18:15:23Z DougNYC joined #sbcl
2016-09-19T18:20:18Z DougNYC quit (Ping timeout: 264 seconds)
2016-09-19T18:37:33Z sjl quit (Ping timeout: 272 seconds)
2016-09-19T18:39:22Z attila_lendvai joined #sbcl
2016-09-19T18:40:04Z rumbler31 quit (Remote host closed the connection)
2016-09-19T18:41:32Z rumbler31 joined #sbcl
2016-09-19T18:42:00Z Intensity quit (Remote host closed the connection)
2016-09-19T18:42:46Z Intensity joined #sbcl
2016-09-19T18:49:24Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-19T18:57:39Z attila_lendvai quit (Ping timeout: 248 seconds)
2016-09-19T18:59:03Z attila_lendvai joined #sbcl
2016-09-19T19:06:08Z rumbler31 quit (Read error: No route to host)
2016-09-19T19:06:48Z rumbler31 joined #sbcl
2016-09-19T19:13:25Z Quadresce joined #sbcl
2016-09-19T19:16:41Z rumbler31 quit (Remote host closed the connection)
2016-09-19T19:26:59Z attila_lendvai quit (Ping timeout: 248 seconds)
2016-09-19T19:38:12Z attila_lendvai joined #sbcl
2016-09-19T19:42:27Z attila_lendvai quit (Ping timeout: 248 seconds)
2016-09-19T19:44:39Z attila_lendvai joined #sbcl
2016-09-19T19:45:07Z rumbler31 joined #sbcl
2016-09-19T19:49:49Z rumbler31 quit (Ping timeout: 265 seconds)
2016-09-19T19:59:13Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-19T19:59:22Z gingerale quit (Remote host closed the connection)
2016-09-19T20:03:59Z DougNYC joined #sbcl
2016-09-19T20:08:03Z DougNYC quit (Ping timeout: 240 seconds)
2016-09-19T20:08:40Z slyrus joined #sbcl
2016-09-19T20:22:03Z shka quit (Ping timeout: 240 seconds)
2016-09-19T20:28:34Z pipping: hi. i'm trying to wrap my head around run-program's :status-hook. in particular, what I can and can't do from inside
2016-09-19T20:29:48Z pipping: i first thought I had to use locks from e.g. bordeaux-threads in order to access variables but that seems to be nonesense (causing deadlock through recursive lock acquisition, if anything) because status-hooks don't run in separate threads
2016-09-19T20:30:27Z pipping: nonsense, even
2016-09-19T20:31:34Z stassats: what are you trying to achieve?
2016-09-19T20:32:52Z pipping: i'd like to spawn a number of asynchronous processes, sleep and be notified as soon as the first one finishes, and find out which one it was
2016-09-19T20:34:19Z stassats: and what's the problem?
2016-09-19T20:36:21Z christoph_debian quit (Quit: WeeChat 1.0.1)
2016-09-19T20:36:36Z pipping: I'm wondering what happens if while one status hook is running
2016-09-19T20:36:52Z stassats: anything can happen
2016-09-19T20:36:55Z pipping:  - either another process changes status
2016-09-19T20:37:07Z pipping:  - or the same process does
2016-09-19T20:37:22Z nzambe quit (Quit: http://www.kiwiirc.com/ - A hand crafted IRC client)
2016-09-19T20:39:44Z christoph_debian joined #sbcl
2016-09-19T20:42:06Z Quadresce joined #sbcl
2016-09-19T20:46:31Z pipping: so e.g. if a SIGCHLD is received, get-processes-status-changes is run, which calls the status hook for each process that has exited since the last update to *active-processes*
2016-09-19T20:48:33Z pipping: all of these hooks are run in sequence, orderly. I just don't understand how receiving another SIGCHLD while the old sigchld-handler is still inside get-processes-status-changes, would be handled. would that have to wait?
2016-09-19T20:54:31Z madbub quit (Remote host closed the connection)
2016-09-19T20:54:39Z stassats: it get be triggered at any time
2016-09-19T20:58:58Z pipping: but if get-processes-status-changes was interrupted by another sigchld-handler calling with-active-processes-lock while the first one is inside the with-active-processes-lock block, the latter would end up waiting for that lock to be released, forever...? I must be misunderstanding something still, sorry :/
2016-09-19T20:59:43Z stassats: with-active-processes-lock won't be interrupted
2016-09-19T21:09:04Z pipping: Okay. That makes a lot of sense. But the part of get-processes-status-changes that calls the status hooks is not protected by the active-processes-lock and thus can be interrupted(?).
2016-09-19T21:09:21Z stassats: yes
2016-09-19T21:12:26Z pipping: So I think then sb-ext:compare-and-swap will solve my problem.
2016-09-19T21:13:32Z rumbler31 joined #sbcl
2016-09-19T21:14:25Z slyrus quit (Ping timeout: 244 seconds)
2016-09-19T21:19:41Z rumbler31 quit (Ping timeout: 272 seconds)
2016-09-19T21:20:17Z edgar-rft quit (Quit: edgar-rft)
2016-09-19T21:20:19Z slyrus joined #sbcl
2016-09-19T21:31:54Z ASau joined #sbcl
2016-09-19T21:36:33Z rumbler31 joined #sbcl
2016-09-19T21:42:31Z sjl joined #sbcl
2016-09-19T21:45:03Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-19T21:51:08Z ASau quit (*.net *.split)
2016-09-19T21:51:08Z attila_lendvai quit (*.net *.split)
2016-09-19T21:51:10Z jdz quit (*.net *.split)
2016-09-19T21:51:12Z salva quit (*.net *.split)
2016-09-19T21:51:32Z salva joined #sbcl
2016-09-19T21:51:37Z ASau joined #sbcl
2016-09-19T21:51:55Z DougNYC joined #sbcl
2016-09-19T21:52:05Z jdz joined #sbcl
2016-09-19T21:54:09Z slyrus quit (Ping timeout: 276 seconds)
2016-09-19T21:56:33Z DougNYC quit (Ping timeout: 240 seconds)
2016-09-19T21:58:05Z rumbler31 quit (Remote host closed the connection)
2016-09-19T22:00:49Z slyrus joined #sbcl
2016-09-19T22:16:58Z attila_lendvai joined #sbcl
2016-09-19T22:16:58Z attila_lendvai quit (Changing host)
2016-09-19T22:16:58Z attila_lendvai joined #sbcl
2016-09-19T22:28:53Z rumbler31 joined #sbcl
2016-09-19T22:33:55Z rumbler31 quit (Ping timeout: 255 seconds)
2016-09-19T22:48:36Z rpg quit (Ping timeout: 265 seconds)
2016-09-19T22:59:20Z stassats: i think i see what's going on
2016-09-19T23:00:13Z stassats: the thread gets deallocated by sbcl before it pthread_create returns
2016-09-19T23:00:35Z stassats: that's why it writes into unampped memory
2016-09-19T23:00:53Z stassats: thread termination gets delayed by 5 on darwin
2016-09-19T23:01:14Z stassats: but now threads are created faster and the condition when 5 threads die is encountered more often
2016-09-19T23:02:32Z stassats: i never liked that DELAY_THREAD_POST_MORTEM stuff
2016-09-19T23:02:53Z stassats: looks like another one of those "solutions" that actually make it harder to debug things down the line
2016-09-19T23:23:08Z cromachina joined #sbcl
2016-09-19T23:40:11Z DougNYC joined #sbcl
2016-09-19T23:45:12Z DougNYC quit (Ping timeout: 265 seconds)
2016-09-19T23:50:21Z eschatologist quit (Quit: ZNC 1.6.3+deb1 - http://znc.in)
2016-09-19T23:53:33Z irsol quit (Ping timeout: 244 seconds)
2016-09-19T23:54:48Z irsol joined #sbcl
2016-09-19T23:56:32Z eschatologist joined #sbcl
2016-09-20T00:10:22Z nyef: stassats: I think that we have had (possibly still have) a "pragmatist" doing maintenance. One whose goal is to get things running
2016-09-20T00:10:47Z nyef: "for the time being", without necessarily making sure that the result is correct and will stay that way.
2016-09-20T00:12:53Z rumbler31 joined #sbcl
2016-09-20T00:13:11Z stassats: "the time being" being a decade or so
2016-09-20T00:14:35Z stassats: ok, i think we need to use pthread_key_create with a destructor
2016-09-20T00:14:46Z stassats: this schedule_thread_post_mortem is really racy, no mater how you slice it
2016-09-20T00:15:09Z stassats: delaying 5 thread deaths, killing in another thread, etc.
2016-09-20T00:16:51Z nyef: I've never really understood all of that stuff, and mostly just been glad that it's never applied to any system I've wanted to use.
2016-09-20T00:17:21Z nyef: ... And, given that I'm supposedly getting another OSX box, this time for personal use, it's time for me to be glad that you're tackling the issues now. (-:
2016-09-20T00:17:25Z rumbler31 quit (Ping timeout: 255 seconds)
2016-09-20T00:17:41Z stassats: (on another note, it's surprising how responsive the OSX UI is under load)
2016-09-20T00:18:09Z stassats: even though all the actions are actually slowed down
2016-09-20T00:18:18Z nyef: I'll take your word for that, given how unresponsive the OSX UI is for me at the best of times.
2016-09-20T00:18:58Z stassats: i guess it just has high priority
2016-09-20T00:21:23Z stassats: CCL doesn't bother with any of that stuff, it leaves thread stack allocation to pthread_create
2016-09-20T00:21:47Z nyef: ... Why aren't we doing that?
2016-09-20T00:22:01Z nyef: Oh, right, guard pages.
2016-09-20T00:24:38Z attila_lendvai quit (Ping timeout: 252 seconds)
2016-09-20T00:24:57Z stassats: knowing pthreads and different OSes, pthread_key_create will probably work differently
2016-09-20T00:25:03Z stassats: but i can test all the OSes
2016-09-20T00:25:14Z stassats: except that all the tests suck
2016-09-20T00:26:34Z stassats: and foreign code can have its own pthread_key_create
2016-09-20T00:26:51Z stassats: which can call back into lisp
2016-09-20T00:27:39Z stassats: (that can be solved, kinda, since the destructor can request to be called again)
2016-09-20T00:27:48Z stassats: anway, needs some think through
2016-09-20T00:36:05Z sjl quit (Ping timeout: 260 seconds)
2016-09-20T00:56:41Z slyrus quit (Ping timeout: 265 seconds)
2016-09-20T01:17:24Z rpg joined #sbcl
2016-09-20T01:31:53Z em1l_ joined #sbcl
2016-09-20T01:35:03Z em1l quit (Ping timeout: 240 seconds)
2016-09-20T01:57:24Z DougNYC joined #sbcl
2016-09-20T02:01:33Z DougNYC quit (Ping timeout: 240 seconds)
2016-09-20T02:03:13Z stassats quit (Ping timeout: 250 seconds)
2016-09-20T03:29:37Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2016-09-20T03:35:28Z rpg joined #sbcl
2016-09-20T03:46:49Z irsol quit (Remote host closed the connection)
2016-09-20T03:50:21Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2016-09-20T03:54:46Z rpg joined #sbcl
2016-09-20T03:55:06Z rpg quit (Client Quit)
2016-09-20T03:57:12Z irsol joined #sbcl
2016-09-20T04:55:19Z shka joined #sbcl
2016-09-20T04:56:41Z irsol quit (Ping timeout: 240 seconds)
2016-09-20T05:04:21Z nyef quit (Remote host closed the connection)
2016-09-20T05:04:37Z nyef joined #sbcl
2016-09-20T05:27:44Z irsol joined #sbcl
2016-09-20T05:33:35Z DougNYC joined #sbcl
2016-09-20T05:38:03Z DougNYC quit (Ping timeout: 240 seconds)
2016-09-20T05:54:16Z gingerale joined #sbcl
2016-09-20T06:05:30Z shka quit (Ping timeout: 265 seconds)
2016-09-20T06:11:31Z rumbler31 joined #sbcl
2016-09-20T06:15:51Z rumbler31 quit (Ping timeout: 250 seconds)
2016-09-20T06:41:50Z ASau quit (Ping timeout: 260 seconds)
2016-09-20T06:50:40Z edgar-rft joined #sbcl
2016-09-20T06:53:05Z Quadresce joined #sbcl
2016-09-20T06:53:20Z Quadresce quit (Client Quit)
2016-09-20T07:09:53Z igajsin1 joined #sbcl
2016-09-20T07:10:45Z igajsin1 left #sbcl
2016-09-20T07:24:58Z irsol quit (Ping timeout: 248 seconds)
2016-09-20T07:45:17Z DeadTrickster_ quit (Ping timeout: 248 seconds)
2016-09-20T07:53:21Z nyef` joined #sbcl
2016-09-20T07:55:18Z nyef quit (Ping timeout: 265 seconds)
2016-09-20T08:22:51Z irsol joined #sbcl
2016-09-20T08:31:06Z stassats joined #sbcl
2016-09-20T09:04:19Z Bike quit (Quit: daijoubu?)
2016-09-20T09:09:34Z DougNYC joined #sbcl
2016-09-20T09:14:33Z DougNYC quit (Ping timeout: 240 seconds)
2016-09-20T09:29:36Z scymtym quit (Ping timeout: 244 seconds)
2016-09-20T09:39:05Z schjetne quit (Ping timeout: 240 seconds)
2016-09-20T09:45:30Z rtmpdavid joined #sbcl
2016-09-20T10:01:54Z scymtym joined #sbcl
2016-09-20T10:16:26Z sjl joined #sbcl
2016-09-20T10:32:18Z jibanes quit (Ping timeout: 264 seconds)
2016-09-20T10:32:34Z jibanes joined #sbcl
2016-09-20T10:39:24Z karswell quit (Remote host closed the connection)
2016-09-20T10:39:36Z karswell joined #sbcl
2016-09-20T10:48:41Z hajovonta joined #sbcl
2016-09-20T10:48:44Z hajovonta: hello
2016-09-20T10:49:18Z hajovonta: can I build SBCL on Intel Atom architecture?
2016-09-20T10:50:52Z hajovonta: more specifically, I'm thinking about the Intel Compute Stick
2016-09-20T10:58:24Z DougNYC joined #sbcl
2016-09-20T10:58:42Z christoph_debian: atom? sure
2016-09-20T11:00:00Z sjl quit (Ping timeout: 276 seconds)
2016-09-20T11:02:33Z DougNYC quit (Ping timeout: 240 seconds)
2016-09-20T11:02:41Z hajovonta: oh, that's good news
2016-09-20T11:03:05Z hajovonta: there is an intel computing stick option with ubuntu preinstalled
2016-09-20T11:04:27Z hajovonta: i can't seem to find sbcl-related material for raspberry or these new devices recently
2016-09-20T11:05:06Z hajovonta: i found this: http://temporal.pr0.pl/devblog/2015/03/07/lisp-on-raspberry-pi-2-part-1-setting-up/ but seems to be a little outdated
2016-09-20T11:05:27Z hajovonta: it says he chose ccl over sbcl because of lack of multithreading support in sbcl
2016-09-20T11:06:38Z Shinmera: if I remember correctly SBCL still does not have threads on ARM.
2016-09-20T11:11:35Z sjl joined #sbcl
2016-09-20T11:16:42Z hajovonta: hm
2016-09-20T11:17:23Z hajovonta: download page says sbcl 1.2.7-1.3.9 is available for arm
2016-09-20T11:17:43Z hajovonta: but says nothing about threading support
2016-09-20T11:17:59Z Xof: Shinmera's recollection is correct
2016-09-20T11:18:12Z Xof: I'd like threads on arm, and have a quarter-finished port, but I am now out of time until December
2016-09-20T11:19:00Z Shinmera: Is that only for ARM64 or both?
2016-09-20T11:19:07Z Xof: no, arm64 has threads already
2016-09-20T11:19:13Z Shinmera: Oh, then I remembered wrong.
2016-09-20T11:19:51Z Shinmera: What was the issue with ARM again? Something about not having enough registers?
2016-09-20T11:19:58Z Xof: it has enough registers, just
2016-09-20T11:20:05Z Xof: the issue is just finding someone who cares enough to do the work
2016-09-20T11:20:12Z Shinmera: I see.
2016-09-20T11:21:47Z hajovonta: arm64? i think the raspberry pi 3 has a 64bit cpu
2016-09-20T11:22:10Z myrkraverk quit (Read error: Connection reset by peer)
2016-09-20T11:23:24Z myrkraverk joined #sbcl
2016-09-20T11:25:29Z hajovonta: ahh "64-bit Raspberry Pi 3 kernel status is in its early days as of this writing (May 2016)."
2016-09-20T11:26:55Z sjl quit (Ping timeout: 265 seconds)
2016-09-20T11:32:22Z stassats: ok, good news (not really) even the old sbcl succumbed to pthread_create death, just very rarely
2016-09-20T11:32:41Z stassats: so, i'm pushing the fix that allows to start on osx and working on better thread destruction
2016-09-20T11:35:08Z stassats: the fix is also an optimization, actually
2016-09-20T11:35:16Z stassats: much faster thread creation
2016-09-20T11:37:15Z stassats: hajovonta: i would highly recommend using the arm64 port
2016-09-20T11:37:31Z stassats: but the rpi situation is weird
2016-09-20T11:39:40Z hajovonta: hi stassats
2016-09-20T11:39:41Z hajovonta: why?
2016-09-20T11:39:50Z stassats: which part?
2016-09-20T11:39:56Z hajovonta: why is it weird?
2016-09-20T11:40:57Z stassats: well, no suitable 64 bit OS
2016-09-20T11:41:00Z stassats: is that still the case?
2016-09-20T11:42:31Z hajovonta: i just read this: https://wiki.ubuntu.com/ARM/RaspberryPi/RaspberryPi3
2016-09-20T11:45:41Z stassats: what were your plans for rpi/compute stick?
2016-09-20T11:45:45Z flip214: http://www.heise.de/ct/artikel/Die-Neuerungen-von-Linux-4-8-3283402.html#nav__a__9
2016-09-20T11:45:49Z flip214: sorry, is in german
2016-09-20T11:48:39Z DougNYC joined #sbcl
2016-09-20T11:49:46Z hajovonta: stassats: i have a project that
2016-09-20T11:50:03Z hajovonta: that can utilize several nodes for computing
2016-09-20T11:50:11Z hajovonta: a kind of distributed computing thing
2016-09-20T11:50:21Z stassats: compute stick sounds a bit expensive for that
2016-09-20T11:50:40Z hajovonta: yep, i checked the price and i came to the same conclusion in the meantime :)
2016-09-20T11:50:49Z hajovonta: but the rpi would be feasible
2016-09-20T11:51:14Z hajovonta: i have two laptops and a vps for testing
2016-09-20T11:51:28Z stassats: i guess you don't really need threads then
2016-09-20T11:52:08Z hajovonta: for now, you are right. but later it would be cool if nodes could utilize all of the available cores
2016-09-20T11:52:26Z stassats: well, just launch several instances of sbcl
2016-09-20T11:52:33Z DougNYC quit (Ping timeout: 240 seconds)
2016-09-20T11:52:59Z hajovonta: good idea
2016-09-20T11:53:51Z hajovonta: i can already test it without an rpi
2016-09-20T11:53:56Z hajovonta: on one of my laptops
2016-09-20T11:54:19Z sjl joined #sbcl
2016-09-20T12:06:53Z madbub joined #sbcl
2016-09-20T12:26:45Z madbub_ joined #sbcl
2016-09-20T12:27:15Z madbub quit (Ping timeout: 272 seconds)
2016-09-20T12:37:41Z nzambe joined #sbcl
2016-09-20T13:10:28Z cromachina quit (Read error: Connection reset by peer)
2016-09-20T13:14:46Z rpg joined #sbcl
2016-09-20T13:19:10Z hajovonta quit (Remote host closed the connection)
2016-09-20T13:25:15Z rpg_ joined #sbcl
2016-09-20T13:26:36Z rpg quit (Ping timeout: 244 seconds)
2016-09-20T13:28:05Z rpg_ is now known as rpg
2016-09-20T13:37:33Z scymtym quit (Ping timeout: 272 seconds)
2016-09-20T13:53:22Z sjl__ joined #sbcl
2016-09-20T13:54:20Z sjl quit (Ping timeout: 265 seconds)
2016-09-20T13:58:41Z sjl__ quit (Ping timeout: 240 seconds)
2016-09-20T13:59:24Z sjl__ joined #sbcl
2016-09-20T14:00:38Z rumbler31 joined #sbcl
2016-09-20T14:01:21Z scymtym joined #sbcl
2016-09-20T14:04:50Z rumbler31 quit (Ping timeout: 244 seconds)
2016-09-20T14:37:58Z schjetne joined #sbcl
2016-09-20T14:59:10Z sjl__ is now known as sjl
2016-09-20T15:04:16Z DGASAU quit (Read error: Connection reset by peer)
2016-09-20T15:04:19Z slyrus joined #sbcl
2016-09-20T15:05:54Z DGASAU joined #sbcl
2016-09-20T15:07:20Z slyrus: stassats: around?
2016-09-20T15:07:27Z stassats: yes
2016-09-20T15:08:32Z slyrus: where does mach_port_get_context live?
2016-09-20T15:08:49Z stassats: what do you mean?
2016-09-20T15:10:37Z slyrus: d'oh. never mind. I thought that was one of our functions, not a Mac OS thing.
2016-09-20T15:11:20Z stassats: so, if you saw it, the memory fault was happening with the old code as well, just more rarely
2016-09-20T15:12:02Z slyrus: right. and now I get ldb on make-target-2.sh :(
2016-09-20T15:12:20Z stassats: what does it say?
2016-09-20T15:13:02Z slyrus: let me try again and make sure that I was building the latest source
2016-09-20T15:14:39Z scymtym_ joined #sbcl
2016-09-20T15:16:03Z scymtym quit (Ping timeout: 240 seconds)
2016-09-20T15:18:17Z slyrus: CORRUPTION WARNING in SBCL pid 20980(tid 140735174086656):
2016-09-20T15:18:23Z slyrus: Memory fault at 0x11aea0 (pc=0x11aea0, sp=0xe777eb8)
2016-09-20T15:19:04Z stassats: so, it faults at PC
2016-09-20T15:19:21Z stassats: that doesn't sound like my thing, there shouldn't be threads anyway
2016-09-20T15:19:33Z stassats: except for the initial thread
2016-09-20T15:24:26Z DougNYC joined #sbcl
2016-09-20T15:27:59Z stassats: slyrus: reproduced that, without actually using sb-thread
2016-09-20T15:28:11Z stassats: though Memory fault at 0x129e20 (pc=0x1000fff390, sp=0xd42bec8)
2016-09-20T15:28:14Z stassats: not at PC
2016-09-20T15:33:10Z stassats: well, without my change it doesn't start at all, even without any threads
2016-09-20T15:34:30Z stassats: and sb-dynamic-core doesn't work without sb-thread, damn it
2016-09-20T15:35:02Z Bike joined #sbcl
2016-09-20T15:35:20Z nyef`: OSX is the platform that we invented that horrid arrange_return_to_lisp_function() thing for, isn't it?
2016-09-20T15:35:39Z nyef`: Or, at least, the platform that requires crazy thunks for it.
2016-09-20T15:36:43Z stassats: well, it doesn't use signals
2016-09-20T15:41:59Z stassats: ok, only non threaded x86-64 dies in cold init
2016-09-20T15:43:29Z slyrus: stassats: I was building a threaded x86-64 build here
2016-09-20T15:43:50Z stassats: fault at 0x129e20 (pc=0x1000fff390, which is movq   0x129e20, %r11
2016-09-20T15:43:52Z stassats: huh what?
2016-09-20T15:45:15Z nyef`: Is that supposed to be runtime space, heap space, linkage-table, or something else?
2016-09-20T15:46:03Z DougNYC quit (Ping timeout: 240 seconds)
2016-09-20T15:47:28Z stassats: the next instruction is leaq   0x90(%r11), %rdx
2016-09-20T15:47:41Z stassats: very strange sequence
2016-09-20T15:48:51Z stassats: or is lldb just stupid?
2016-09-20T15:49:23Z nyef`: Looks like the destination is on the right side?
2016-09-20T15:49:23Z stassats: i disassemble a working function, it's MOV R11, [R12+24] LEA RDX, [R11+144]
2016-09-20T15:49:26Z stassats: though that's on threads
2016-09-20T15:50:19Z nyef`: That's an allocation sequence, isn't it?\
2016-09-20T15:50:52Z stassats: found an unthreaded older sbcl
2016-09-20T15:50:56Z stassats: MOV R11, [#x126D50]
2016-09-20T15:50:59Z stassats: LEA RDI, [R11+272]
2016-09-20T15:51:27Z stassats: so, dunno, is 0x129e20 the intel syntax for an ea?
2016-09-20T15:51:35Z nyef`: Allocation sequence. [R12+24] is the free pointer for the current allocation region.
2016-09-20T15:51:36Z stassats: err, at&t
2016-09-20T15:51:56Z stassats: will try to switch to intel syntax
2016-09-20T15:52:04Z stassats: (and i'm not used to lldb at all, sigh)
2016-09-20T15:52:42Z nyef`: I think that a bare constant like that is a direct memory access, yes. There's some tag to say "immediate", and some further tag to say "indirect" for the few instructions that allow an indirect there.
2016-09-20T15:53:01Z stassats: ok, back to sanity
2016-09-20T15:53:02Z stassats: mov    r11, qword ptr [0x129e20]
2016-09-20T15:53:08Z stassats: so, 0x129e20 is boxed_region
2016-09-20T15:53:24Z stassats: nyef`: right, e.g. movl   $0x48000001, %edi         ; imm = 0x48000001
2016-09-20T15:54:08Z stassats: do we fault on boxed_region or something?
2016-09-20T15:54:12Z stassats: to grow it?
2016-09-20T15:54:17Z stassats: or is that from cheneygc?
2016-09-20T15:54:23Z nyef`: That's cheneygc, IIRC.
2016-09-20T15:54:49Z nyef`: The gencgc version is a bounds-check against the end address, which is the second slot in the alloc_region structure.
2016-09-20T15:57:35Z stassats: p &boxed_region
2016-09-20T15:57:35Z stassats: (alloc_region *) $25 = 0x00000000086f4e20
2016-09-20T15:59:39Z stassats: that's not linkage-table stuff, is it?
2016-09-20T16:01:21Z stassats: there's no such mapping
2016-09-20T16:03:00Z stassats: doesn't make any sense
2016-09-20T16:03:22Z nyef`: Oh, FFS. I declare a parameter to be UNSIGNED-BYTE, it gets tag-tested to be sure that it's a FIXNUM or a BIGNUM, and then a call to GENERIC-< is issued.
2016-09-20T16:03:33Z stassats: sbcl.nm: 0000000000129e20 S _boxed_region
2016-09-20T16:03:35Z stassats: huh
2016-09-20T16:03:56Z nyef`: ASR for executables?
2016-09-20T16:04:14Z stassats: well, it works for threaded builds
2016-09-20T16:04:25Z stassats: but wait! i have sb-dynamic-core for threaded builds
2016-09-20T16:04:29Z nyef`: Yeah.
2016-09-20T16:04:42Z stassats: that explains slyrus's failure
2016-09-20T16:10:52Z stassats: but the man page says Do not make a position independent executable (PIE).  This is the default, when targeting 10.6 and earlier.
2016-09-20T16:11:33Z stassats: yet -mmacosx-version-min=10.6
2016-09-20T16:11:41Z slyrus: so do we just make sb-dynamic-core the default on darwin? :)
2016-09-20T16:12:03Z stassats: slyrus: doesn't work without threads
2016-09-20T16:12:13Z stassats: will just disable pie
2016-09-20T16:12:18Z slyrus: ok, so make threads the default too? :)
2016-09-20T16:12:44Z slyrus: at least your fix works for the sb-dynamic-core builds. progress!
2016-09-20T16:12:45Z stassats: well, threads are currently broken
2016-09-20T16:12:49Z stassats: (ha-ha)
2016-09-20T16:13:53Z slyrus: awesome
2016-09-20T16:14:56Z nyef`: How long before we just give up and move the runtime stuff over to the Lisp side? (-:
2016-09-20T16:15:08Z stassats: for even more headaches?
2016-09-20T16:15:26Z nyef`: Something like that, yes.
2016-09-20T16:15:54Z slyrus: what's wrong with threads (and sb-dynamic-core) now?
2016-09-20T16:16:30Z stassats: slyrus: a race condition when the thread exits and deallocates memory and pthread_create hasn't returned yet
2016-09-20T16:16:56Z slyrus: that sounds familiar-ish
2016-09-20T16:17:23Z stassats: i'm thinking of attaching a destructor to pthread_key_create
2016-09-20T16:17:37Z stassats: and destroying the thread stuff there
2016-09-20T16:17:37Z slyrus: running threads tests, things seem ok
2016-09-20T16:18:12Z slyrus: does the thread_post_mortem stuff help?
2016-09-20T16:18:25Z stassats: not really
2016-09-20T16:18:31Z slyrus: OK
2016-09-20T16:18:45Z stassats: 5 threads can die before pthread_key_create returns
2016-09-20T16:19:18Z stassats: especially when there is a lot of threads and things are scheduled just right
2016-09-20T16:20:11Z stassats: ok, pushed no-pie stuff
2016-09-20T16:21:45Z slyrus: thanks!
2016-09-20T16:33:28Z shka joined #sbcl
2016-09-20T16:41:40Z stassats: pthread_key destructor will probably run on the same thread as well
2016-09-20T16:41:50Z stassats: so, can't free the stack there
2016-09-20T16:42:05Z stassats: but, there's a lock around create_os_thread
2016-09-20T16:42:12Z stassats: can probably just grab it on darwin
2016-09-20T16:42:44Z stassats: or everywhere
2016-09-20T16:51:35Z stassats: will have to leave the test running overnight
2016-09-20T16:51:39Z stassats: but that seems like a solution
2016-09-20T16:51:45Z stassats: and getting rid of DELAY_THREAD_POST_MORTEM
2016-09-20T16:53:57Z nyef`: Does this also cover the... what was it... FreeBSD case?
2016-09-20T16:54:04Z nyef`: I think it was FreeBSD, at least.
2016-09-20T16:54:19Z stassats: there's some freebsd stuff for locking around pthread_create
2016-09-20T16:54:51Z stassats: but i'm not sure whether it needs to be locked in perform_thread_post_mortem
2016-09-20T16:56:12Z stassats: this is probably because pthread_join on darwin doesn't grab a lock or something
2016-09-20T16:59:06Z stassats: rather, pthread_join does grab a lock but pthread_create doesn't
2016-09-20T17:28:05Z scymtym_ quit (Ping timeout: 272 seconds)
2016-09-20T17:28:41Z Quadresce joined #sbcl
2016-09-20T17:34:15Z madbub_ quit (Ping timeout: 265 seconds)
2016-09-20T17:36:49Z madbub joined #sbcl
2016-09-20T17:54:55Z rumbler31 joined #sbcl
2016-09-20T18:10:14Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-20T18:10:56Z Quadresce joined #sbcl
2016-09-20T18:26:23Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-20T18:30:21Z Quadresce joined #sbcl
2016-09-20T18:30:53Z sjl quit (Ping timeout: 265 seconds)
2016-09-20T18:38:03Z rumbler31 quit (Remote host closed the connection)
2016-09-20T18:38:56Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-20T18:39:28Z Quadresce joined #sbcl
2016-09-20T18:39:53Z rumbler31 joined #sbcl
2016-09-20T18:42:10Z slyrus quit (Ping timeout: 255 seconds)
2016-09-20T18:56:31Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-20T19:02:31Z DougNYC joined #sbcl
2016-09-20T19:07:33Z DougNYC quit (Ping timeout: 240 seconds)
2016-09-20T19:29:59Z sjl joined #sbcl
2016-09-20T19:38:43Z rumbler31 quit (Remote host closed the connection)
2016-09-20T19:43:23Z nzambe quit (Quit: http://www.kiwiirc.com/ - A hand crafted IRC client)
2016-09-20T19:44:23Z nzambe joined #sbcl
2016-09-20T19:59:44Z attila_lendvai joined #sbcl
2016-09-20T19:59:44Z attila_lendvai quit (Changing host)
2016-09-20T19:59:44Z attila_lendvai joined #sbcl
2016-09-20T20:01:09Z gingerale quit (Read error: Connection reset by peer)
2016-09-20T20:02:38Z shka quit (Ping timeout: 265 seconds)
2016-09-20T20:28:58Z rumbler31 joined #sbcl
2016-09-20T20:30:03Z rumbler31 quit (Remote host closed the connection)
2016-09-20T20:34:52Z shka joined #sbcl
2016-09-20T20:42:13Z Quadresce joined #sbcl
2016-09-20T20:51:18Z DougNYC joined #sbcl
2016-09-20T20:54:03Z madbub quit (Remote host closed the connection)
2016-09-20T20:55:33Z DougNYC quit (Ping timeout: 240 seconds)
2016-09-20T20:56:09Z pootler_ quit (Quit: Connection closed for inactivity)
2016-09-20T20:56:15Z ASau joined #sbcl
2016-09-20T21:06:55Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-20T21:13:07Z Quadresce joined #sbcl
2016-09-20T21:13:29Z stassats quit (Ping timeout: 240 seconds)
2016-09-20T21:23:16Z stassats joined #sbcl
2016-09-20T21:29:21Z irsol quit (Ping timeout: 244 seconds)
2016-09-20T21:29:51Z stassats quit (Ping timeout: 276 seconds)
2016-09-20T21:30:22Z stassats joined #sbcl
2016-09-20T21:32:14Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-20T21:32:23Z stassats: got a memory fault on an interrupt, but quit before debugging it
2016-09-20T21:32:24Z Quadresce joined #sbcl
2016-09-20T21:32:31Z stassats: good luck reproducing it...
2016-09-20T21:34:46Z stassats: i can deadlock it, though
2016-09-20T21:35:07Z stassats: sigh, threads are not really robust on osx
2016-09-20T21:36:48Z stassats: have one sbcl with (let ((nthreads 3)) (loop (mapc #'sb-thread:join-thread (loop repeat nthreads collect (sb-thread:make-thread (lambda () (loop repeat 1000 do (sleep 0.00001))))))))
2016-09-20T21:37:05Z stassats: then from another (loop (sb-posix:kill 7611 sb-unix:sigchld))
2016-09-20T21:37:17Z stassats: and after a while it manages to deadlock
2016-09-20T21:40:02Z scymtym joined #sbcl
2016-09-20T21:40:19Z stassats: (SB-THREAD::%WAIT-FOR-MUTEX #<SB-THREAD:MUTEX "thread result lock" owner: #<SB-THREAD:THREAD waiting on: #<MUTEX "Lock for active processes." (free)> {10031A3753}>> #<SB-THREAD:THREAD "main thread" RUNNING {1002E3E983}> NIL NIL NIL NIL NIL NIL)
2016-09-20T21:43:56Z attila_lendvai quit (Ping timeout: 252 seconds)
2016-09-20T21:44:48Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-20T21:45:42Z Quadresce joined #sbcl
2016-09-20T21:50:38Z stassats:  actually, i can get that to lock on linux too
2016-09-20T21:51:24Z stassats: interestingly, it comes back to life after some time
2016-09-20T21:53:27Z stassats: but not on osx
2016-09-20T21:53:28Z stassats: weird stuff
2016-09-20T21:56:02Z shka quit (Ping timeout: 244 seconds)
2016-09-20T22:11:39Z stassats: so a thread is waiting on a free mutex
2016-09-20T22:13:52Z pootler_ joined #sbcl
2016-09-20T22:26:48Z pootler_ is now known as mrh
2016-09-20T22:27:32Z irsol joined #sbcl
2016-09-20T22:27:51Z mrh is now known as merv
2016-09-20T22:29:11Z merv is now known as Merv
2016-09-20T22:30:44Z stassats: Received signal 10 in non-lisp thread 123145302839296, resignalling to a lisp thread.
2016-09-20T22:30:47Z stassats: great
2016-09-20T22:39:39Z DougNYC joined #sbcl
2016-09-20T22:40:55Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-20T22:41:01Z cromachina joined #sbcl
2016-09-20T22:44:03Z DougNYC quit (Ping timeout: 240 seconds)
2016-09-20T22:47:18Z cromachina quit (Ping timeout: 264 seconds)
2016-09-20T22:47:55Z cromachina joined #sbcl
2016-09-20T22:48:14Z Quadresce joined #sbcl
2016-09-20T22:52:16Z stassats: can't figure why a thread is waiting on a free mutex
2016-09-20T22:53:12Z |3b|: freed while you were trying to look at it?
2016-09-20T22:53:45Z stassats: well, that's the whole idea of mutexes, someone has to free it
2016-09-20T22:54:16Z |3b|: right, i mean it wasn't free when it waited
2016-09-20T22:54:37Z stassats: it wouldn't deadlock then
2016-09-20T22:58:05Z stassats: and i can't interrupt it to see what's going on
2016-09-20T22:58:11Z stassats: lldb says it's in nanosleep
2016-09-20T23:05:27Z stassats: and now emacs crashed
2016-09-20T23:05:30Z stassats: fun
2016-09-20T23:06:19Z stassats: and again
2016-09-20T23:20:26Z stassats: maybe it sleeps too much
2016-09-20T23:20:55Z stassats: or never gets scheduled
2016-09-20T23:27:31Z |3b|: could it be printing the wrong lock if it gets interrupted in the middle of sigchld handler grabbing the lock?
2016-09-20T23:28:10Z |3b|: or even getting a false positive due to multiple threads getting interrupted at once
2016-09-20T23:28:12Z stassats: it can't get interrupted
2016-09-20T23:28:49Z |3b|: am i misinterpreting (with-interrupts (check-deadlock))?
2016-09-20T23:30:28Z stassats: it's safe to interrupt that
2016-09-20T23:35:29Z stassats: but i can't seem to be able to interrupt the nanosleep
2016-09-20T23:41:02Z stassats: uhm
2016-09-20T23:41:08Z stassats: i see
2016-09-20T23:41:30Z stassats: nanosleep is not interrupt safe on darwin
2016-09-20T23:42:19Z stassats: libsystem_c.dylib`nanosleep => libsystem_pthread.dylib`_pthread_testcancel =>  libsystem_platform.dylib`OSSpinLockLock => interrupt
2016-09-20T23:42:44Z stassats: in the handler: libsystem_c.dylib`nanosleep => libsystem_pthread.dylib`_pthread_testcancel => libsystem_platform.dylib`_OSSpinLockLockSlow
2016-09-20T23:43:07Z stassats: so, yeah, nanosleep is not reentrant
2016-09-20T23:49:36Z sjl quit (Ping timeout: 276 seconds)
2016-09-20T23:50:27Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-20T23:52:44Z stassats: which sucks
2016-09-20T23:55:14Z stassats: fix one bug, discover another
2016-09-20T23:55:17Z stassats: rinse and repeat
2016-09-20T23:59:30Z stassats: why isn't it just a recursive lock then?
2016-09-20T23:59:32Z stassats: sigh, apple
2016-09-21T00:00:27Z stassats: it actually gets interrupted here while locking
2016-09-21T00:15:05Z edgar-rft quit (Quit: edgar-rft)
2016-09-21T00:25:21Z ASau quit (Ping timeout: 244 seconds)
2016-09-21T00:25:24Z ASau` joined #sbcl
2016-09-21T00:47:56Z irsol quit (Remote host closed the connection)
2016-09-21T00:48:34Z irsol joined #sbcl
2016-09-21T00:57:33Z DougNYC joined #sbcl
2016-09-21T01:02:33Z DougNYC quit (Ping timeout: 240 seconds)
2016-09-21T01:05:09Z irsol quit (Ping timeout: 250 seconds)
2016-09-21T01:11:34Z irsol joined #sbcl
2016-09-21T01:30:45Z em1l joined #sbcl
2016-09-21T01:34:05Z em1l_ quit (Ping timeout: 260 seconds)
2016-09-21T01:46:09Z Merv quit (Quit: Connection closed for inactivity)
2016-09-21T01:56:57Z stassats quit (Ping timeout: 248 seconds)
2016-09-21T02:46:20Z DougNYC joined #sbcl
2016-09-21T02:50:33Z DougNYC quit (Ping timeout: 240 seconds)
2016-09-21T03:25:58Z DavidGu joined #sbcl
2016-09-21T04:34:40Z DougNYC joined #sbcl
2016-09-21T04:39:03Z DougNYC quit (Ping timeout: 240 seconds)
2016-09-21T05:09:52Z cromachina quit (Read error: Connection reset by peer)
2016-09-21T05:09:58Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2016-09-21T05:38:44Z DavidGu quit (Quit: DavidGu)
2016-09-21T05:50:27Z shka joined #sbcl
2016-09-21T06:20:54Z shka quit (Ping timeout: 264 seconds)
2016-09-21T06:40:21Z DavidGu joined #sbcl
2016-09-21T07:38:17Z DavidGu quit (Quit: DavidGu)
2016-09-21T07:40:25Z Bike quit (Quit: sicken)
2016-09-21T08:00:15Z edgar-rft joined #sbcl
2016-09-21T08:10:33Z DougNYC joined #sbcl
2016-09-21T08:15:03Z DougNYC quit (Ping timeout: 240 seconds)
2016-09-21T09:23:47Z edgar-rft quit (Quit: edgar-rft)
2016-09-21T09:25:06Z ASau` quit (Ping timeout: 264 seconds)
2016-09-21T09:32:39Z p_l quit (Ping timeout: 272 seconds)
2016-09-21T09:34:53Z p_l joined #sbcl
2016-09-21T09:59:06Z DougNYC joined #sbcl
2016-09-21T10:03:33Z DougNYC quit (Ping timeout: 240 seconds)
2016-09-21T10:19:31Z scymtym quit (Ping timeout: 255 seconds)
2016-09-21T10:27:19Z DavidGu joined #sbcl
2016-09-21T10:31:01Z edgar-rft joined #sbcl
2016-09-21T10:36:30Z Merv joined #sbcl
2016-09-21T10:50:37Z Merv` joined #sbcl
2016-09-21T10:54:00Z Merv` quit (Client Quit)
2016-09-21T11:01:00Z scymtym joined #sbcl
2016-09-21T11:07:46Z scymtym quit (Remote host closed the connection)
2016-09-21T11:12:43Z scymtym joined #sbcl
2016-09-21T11:17:37Z DougNYC joined #sbcl
2016-09-21T11:21:33Z DougNYC quit (Ping timeout: 240 seconds)
2016-09-21T12:00:17Z rpg joined #sbcl
2016-09-21T12:01:18Z madbub joined #sbcl
2016-09-21T12:02:03Z sjl joined #sbcl
2016-09-21T12:04:43Z stassats joined #sbcl
2016-09-21T12:05:16Z rpg quit (Ping timeout: 255 seconds)
2016-09-21T12:10:12Z pootler joined #sbcl
2016-09-21T12:39:39Z cromachina joined #sbcl
2016-09-21T12:41:07Z edgar-rft quit (Quit: edgar-rft)
2016-09-21T12:43:28Z Merv` joined #sbcl
2016-09-21T12:44:03Z Merv` quit (Remote host closed the connection)
2016-09-21T12:46:06Z DavidGu quit (Ping timeout: 264 seconds)
2016-09-21T12:52:54Z Merv` joined #sbcl
2016-09-21T12:56:07Z madbub quit (Ping timeout: 255 seconds)
2016-09-21T12:56:21Z madbub joined #sbcl
2016-09-21T13:02:28Z madbub quit (Ping timeout: 265 seconds)
2016-09-21T13:02:35Z madbub_ joined #sbcl
2016-09-21T13:25:50Z rpg joined #sbcl
2016-09-21T13:54:03Z cromachina quit (Read error: Connection reset by peer)
2016-09-21T14:52:33Z schjetne quit (Ping timeout: 240 seconds)
2016-09-21T14:53:36Z DougNYC joined #sbcl
2016-09-21T15:04:29Z dougk_ quit (Ping timeout: 250 seconds)
2016-09-21T15:15:03Z DougNYC quit (Ping timeout: 240 seconds)
2016-09-21T15:17:27Z dougk_ joined #sbcl
2016-09-21T15:42:32Z dougk_ quit (Remote host closed the connection)
2016-09-21T15:47:28Z Bike joined #sbcl
2016-09-21T16:00:37Z gingerale joined #sbcl
2016-09-21T16:28:20Z dougk_ joined #sbcl
2016-09-21T16:31:54Z stassats: i wonder if the nanosleep reentrancy problem can somehow be worked around
2016-09-21T16:32:05Z stassats: like not calling it from sigchld-handler
2016-09-21T16:32:13Z stassats: that leaves stop-for-gc
2016-09-21T16:32:55Z nyef`: Dare I suggest using safepoints instead? d-:
2016-09-21T16:33:36Z stassats: and leave sleep uninterruptable?
2016-09-21T16:34:12Z stassats: well, clearly nanosleep should just be fixed, but this being apple i don't have my hopes up
2016-09-21T16:34:33Z nyef`: I still think that we shouldn't have to GC-stop threads that are in alien code, just arrange that they GC-stop themselves if they try to re-enter Lisp while a GC is still going on.
2016-09-21T16:34:41Z gingerale quit (Remote host closed the connection)
2016-09-21T16:35:07Z stassats: safepoints do something like that
2016-09-21T16:35:12Z stassats: except that safepoints are broken too
2016-09-21T16:35:13Z stassats: so...
2016-09-21T16:35:23Z _mjl joined #sbcl
2016-09-21T16:35:24Z nyef`: Right.
2016-09-21T16:35:37Z nyef`: Hence the "d-:" when I suggested them.
2016-09-21T16:35:56Z pootler quit (Remote host closed the connection)
2016-09-21T16:36:20Z stassats: stuff like that is probably the only reason i like open source
2016-09-21T16:36:27Z stassats: because i'm able to fix their bugs
2016-09-21T16:40:29Z Merv` quit (Ping timeout: 244 seconds)
2016-09-21T16:49:13Z stassats: sem_wait on darwin calls __pthread_testcancel, but it's different from _pthread_testcancel
2016-09-21T16:49:24Z stassats: it doesn't do any locking
2016-09-21T16:51:42Z gingerale joined #sbcl
2016-09-21T16:53:02Z stassats: i wonder if nanosleep is unwind safe
2016-09-21T17:06:07Z madbub_ quit (Ping timeout: 272 seconds)
2016-09-21T17:06:45Z madbub joined #sbcl
2016-09-21T17:16:41Z stassats: i can actually reimplement nanosleep using a syscall
2016-09-21T17:17:40Z logrus joined #sbcl
2016-09-21T17:26:55Z stassats: i can just use the code from apple, but it's under Apple Public Source License
2016-09-21T17:27:10Z stassats: which is DFSG incompatible
2016-09-21T17:27:18Z stassats: but i already read the code
2016-09-21T17:28:23Z pootler joined #sbcl
2016-09-21T17:31:45Z shka joined #sbcl
2016-09-21T17:36:04Z stassats: i'll rewrite it in lisp
2016-09-21T17:38:41Z stassats: or rather, i'll use the semantics of SB-UNIX:NANOSLEEP, restarting when interrupted
2016-09-21T17:38:53Z stassats: that should make it novel enough
2016-09-21T17:44:45Z scymtym quit (Ping timeout: 272 seconds)
2016-09-21T17:59:03Z edgar-rft joined #sbcl
2016-09-21T18:04:49Z attila_lendvai joined #sbcl
2016-09-21T18:04:49Z attila_lendvai quit (Changing host)
2016-09-21T18:04:49Z attila_lendvai joined #sbcl
2016-09-21T18:31:47Z DougNYC joined #sbcl
2016-09-21T18:33:04Z pootler quit (Read error: Connection reset by peer)
2016-09-21T18:36:33Z DougNYC quit (Ping timeout: 240 seconds)
2016-09-21T18:45:48Z gingerale quit (Remote host closed the connection)
2016-09-21T18:54:28Z rumbler31 joined #sbcl
2016-09-21T19:00:08Z sjl quit (Ping timeout: 272 seconds)
2016-09-21T19:17:51Z _mjl quit (Ping timeout: 272 seconds)
2016-09-21T19:31:43Z fe[nl]ix: nyef`: how 'about fixing safepoints and making them mandatory ?
2016-09-21T19:31:51Z pootler joined #sbcl
2016-09-21T19:33:37Z stassats: sleep ought to be interruptable, no matter what is used
2016-09-21T19:33:53Z nyef`: I don't understand safepoints to begin with, and I'm not the one working on this bug.
2016-09-21T19:34:02Z stassats: and even when we fix the known safepoint issue, i'm not comfortable replacing a battle tested strategy
2016-09-21T19:34:11Z stassats: to see them fail in another place
2016-09-21T19:34:40Z nyef`: Mmm. That too.
2016-09-21T19:34:56Z stassats: i think if i just replace nanosleep in darwin it should make things better
2016-09-21T19:35:13Z stassats: even though interrupting arbitrary code is a bad idea, but nanosleep is a prime target for interrupts
2016-09-21T20:13:36Z shka quit (Ping timeout: 265 seconds)
2016-09-21T20:15:28Z schjetne joined #sbcl
2016-09-21T20:20:06Z DougNYC joined #sbcl
2016-09-21T20:24:33Z DougNYC quit (Ping timeout: 240 seconds)
2016-09-21T20:32:21Z ASau joined #sbcl
2016-09-21T20:40:56Z rumbler31 quit
2016-09-21T20:43:11Z karswell quit (Remote host closed the connection)
2016-09-21T20:44:33Z rumbler31 joined #sbcl
2016-09-21T20:50:19Z prxq joined #sbcl
2016-09-21T21:05:22Z Blkt quit (Remote host closed the connection)
2016-09-21T21:05:22Z fe[nl]ix quit (Read error: Connection reset by peer)
2016-09-21T21:05:37Z Blkt joined #sbcl
2016-09-21T21:05:38Z fe[nl]ix joined #sbcl
2016-09-21T21:09:33Z madbub quit (Ping timeout: 240 seconds)
2016-09-21T21:31:55Z edgar-rft quit (Quit: edgar-rft)
2016-09-21T21:34:39Z jsnell quit (Quit: leaving)
2016-09-21T21:34:48Z jsnell joined #sbcl
2016-09-21T21:41:09Z Quadrescence quit (Quit: Leaving)
2016-09-21T21:56:41Z rpg quit (Read error: Connection reset by peer)
2016-09-21T22:00:29Z eschatologist quit (Ping timeout: 250 seconds)
2016-09-21T22:01:44Z eschatologist joined #sbcl
2016-09-21T22:05:14Z rumbler31 quit (Remote host closed the connection)
2016-09-21T22:06:46Z rumbler31 joined #sbcl
2016-09-21T22:08:18Z rpg joined #sbcl
2016-09-21T22:08:41Z DougNYC joined #sbcl
2016-09-21T22:12:47Z Quadrescence joined #sbcl
2016-09-21T22:13:03Z DougNYC quit (Ping timeout: 240 seconds)
2016-09-21T22:14:33Z rumbler31 quit (Remote host closed the connection)
2016-09-21T22:24:09Z stassats quit (Ping timeout: 248 seconds)
2016-09-21T22:29:30Z stassats joined #sbcl
2016-09-21T22:36:17Z prxq quit (Read error: Connection reset by peer)
2016-09-21T22:44:59Z DougNYC joined #sbcl
2016-09-21T23:13:29Z attila_lendvai quit (Ping timeout: 252 seconds)
2016-09-21T23:34:29Z cromachina joined #sbcl
2016-09-21T23:48:26Z scymtym joined #sbcl
2016-09-21T23:50:40Z DougNYC quit
2016-09-22T00:03:26Z sjl joined #sbcl
2016-09-22T00:06:51Z edgar-rft joined #sbcl
2016-09-22T00:15:24Z rumbler31 joined #sbcl
2016-09-22T00:20:07Z rumbler31 quit (Ping timeout: 255 seconds)
2016-09-22T00:35:16Z edgar-rft quit (Quit: edgar-rft)
2016-09-22T00:38:24Z sjl quit (Ping timeout: 244 seconds)
2016-09-22T00:42:17Z foom quit (Ping timeout: 248 seconds)
2016-09-22T00:55:17Z foom joined #sbcl
2016-09-22T01:18:39Z hydan joined #sbcl
2016-09-22T01:29:37Z hydan quit (Remote host closed the connection)
2016-09-22T01:30:33Z em1l_ joined #sbcl
2016-09-22T01:33:33Z em1l quit (Ping timeout: 240 seconds)
2016-09-22T01:51:35Z rumbler31 joined #sbcl
2016-09-22T01:56:30Z rumbler31 quit (Ping timeout: 260 seconds)
2016-09-22T02:50:25Z rpg quit (Ping timeout: 265 seconds)
2016-09-22T05:18:03Z jrm quit (Read error: Connection reset by peer)
2016-09-22T05:19:35Z jrm joined #sbcl
2016-09-22T05:29:43Z shka joined #sbcl
2016-09-22T05:51:11Z DavidGu joined #sbcl
2016-09-22T05:54:08Z DavidGu quit (Client Quit)
2016-09-22T06:04:16Z gingerale joined #sbcl
2016-09-22T06:29:54Z shka quit (Ping timeout: 264 seconds)
2016-09-22T07:12:48Z Bike quit (Quit: irssi security problems yeeeeeah)
2016-09-22T07:12:57Z Bike joined #sbcl
2016-09-22T07:26:49Z ASau quit (Ping timeout: 272 seconds)
2016-09-22T07:30:56Z gargaml joined #sbcl
2016-09-22T07:38:09Z edgar-rft joined #sbcl
2016-09-22T07:58:14Z gingerale quit (Remote host closed the connection)
2016-09-22T08:37:05Z Bike quit (Quit: leaving)
2016-09-22T09:30:23Z scymtym quit (Ping timeout: 244 seconds)
2016-09-22T09:50:41Z DeadTrickster quit (Ping timeout: 240 seconds)
2016-09-22T09:57:50Z DeadTrickster joined #sbcl
2016-09-22T09:58:10Z scymtym joined #sbcl
2016-09-22T11:18:01Z sjl joined #sbcl
2016-09-22T11:24:03Z pootler quit (Ping timeout: 240 seconds)
2016-09-22T11:33:33Z pootler joined #sbcl
2016-09-22T11:54:15Z chris2 quit (Ping timeout: 265 seconds)
2016-09-22T11:54:54Z madbub joined #sbcl
2016-09-22T11:57:31Z chris2 joined #sbcl
2016-09-22T12:02:07Z chris2 quit (Ping timeout: 255 seconds)
2016-09-22T12:04:12Z chris2 joined #sbcl
2016-09-22T12:29:19Z gargaml quit (Ping timeout: 250 seconds)
2016-09-22T12:32:42Z gargaml joined #sbcl
2016-09-22T12:37:39Z _mjl joined #sbcl
2016-09-22T12:38:08Z rumbler31 joined #sbcl
2016-09-22T12:38:41Z DavidGu joined #sbcl
2016-09-22T12:43:04Z rumbler31 quit (Ping timeout: 255 seconds)
2016-09-22T12:43:28Z DavidGu quit (Ping timeout: 265 seconds)
2016-09-22T12:45:16Z Merv_ joined #sbcl
2016-09-22T12:48:42Z pootler quit (Ping timeout: 276 seconds)
2016-09-22T12:48:49Z Merv_ quit (Client Quit)
2016-09-22T12:59:25Z rumbler31 joined #sbcl
2016-09-22T13:04:23Z rumbler31 quit (Ping timeout: 272 seconds)
2016-09-22T13:08:17Z pootler joined #sbcl
2016-09-22T13:14:07Z gargaml quit (Ping timeout: 255 seconds)
2016-09-22T13:14:37Z DavidGu joined #sbcl
2016-09-22T13:20:00Z DavidGu quit (Quit: DavidGu)
2016-09-22T13:21:48Z gargaml joined #sbcl
2016-09-22T13:21:54Z DavidGu joined #sbcl
2016-09-22T13:26:48Z DavidGu quit (Quit: DavidGu)
2016-09-22T13:30:10Z DavidGu joined #sbcl
2016-09-22T13:46:08Z sjl quit (Ping timeout: 244 seconds)
2016-09-22T13:48:07Z cromachina quit (Read error: Connection reset by peer)
2016-09-22T13:51:10Z DavidGu quit (Ping timeout: 265 seconds)
2016-09-22T13:54:02Z DavidGu joined #sbcl
2016-09-22T14:01:14Z rumbler31 joined #sbcl
2016-09-22T14:05:52Z rumbler31 quit (Ping timeout: 255 seconds)
2016-09-22T14:08:05Z DavidGu quit (Ping timeout: 265 seconds)
2016-09-22T14:24:53Z DavidGu joined #sbcl
2016-09-22T14:31:49Z DavidGu quit (Ping timeout: 265 seconds)
2016-09-22T14:41:51Z attila_lendvai joined #sbcl
2016-09-22T14:54:55Z DavidGu joined #sbcl
2016-09-22T14:56:57Z DavidGu quit (Read error: Connection reset by peer)
2016-09-22T14:57:18Z DavidGu joined #sbcl
2016-09-22T14:58:54Z DavidGu quit (Read error: Connection reset by peer)
2016-09-22T14:59:42Z DavidGu joined #sbcl
2016-09-22T15:03:28Z DavidGu quit (Read error: Connection reset by peer)
2016-09-22T15:22:45Z rumbler31 joined #sbcl
2016-09-22T15:30:33Z schjetne quit (Ping timeout: 276 seconds)
2016-09-22T15:34:12Z DavidGu joined #sbcl
2016-09-22T15:35:21Z DavidGu quit (Client Quit)
2016-09-22T15:36:06Z schjetne joined #sbcl
2016-09-22T15:42:28Z Merv_ joined #sbcl
2016-09-22T15:42:30Z schjetne quit (Ping timeout: 260 seconds)
2016-09-22T15:44:45Z pootler quit (Ping timeout: 265 seconds)
2016-09-22T15:48:34Z schjetne joined #sbcl
2016-09-22T16:05:22Z edgar-rft quit (Quit: edgar-rft)
2016-09-22T16:32:51Z rumbler31 quit (Remote host closed the connection)
2016-09-22T16:41:37Z scymtym quit (Ping timeout: 272 seconds)
2016-09-22T16:55:01Z rtmpdavid quit (Remote host closed the connection)
2016-09-22T16:57:35Z rtmpdavid joined #sbcl
2016-09-22T17:03:49Z rumbler31 joined #sbcl
2016-09-22T17:07:17Z rtmpdavid quit (Remote host closed the connection)
2016-09-22T17:13:30Z rpg joined #sbcl
2016-09-22T17:13:41Z attila_lendvai quit (Ping timeout: 252 seconds)
2016-09-22T17:21:03Z pipping: Is it possible to dump an image from within sbcl without dying? e.g. through something like fork-and-then-call-"save-lisp-and-die"?
2016-09-22T17:24:41Z chris2 quit (Ping timeout: 272 seconds)
2016-09-22T17:26:45Z nyef`: Sure. Go ahead.
2016-09-22T17:27:13Z nyef`: Have a look at SB-POSIX:FORK.
2016-09-22T17:27:51Z attila_lendvai joined #sbcl
2016-09-22T17:27:51Z attila_lendvai quit (Changing host)
2016-09-22T17:27:51Z attila_lendvai joined #sbcl
2016-09-22T17:28:30Z stassats: if you're saving in the middle of something the image is probably not in a good state to be started up again
2016-09-22T17:28:51Z stassats: threads, sockets, shared libraries, what have you
2016-09-22T17:35:33Z rumbler31 quit (Read error: Connection reset by peer)
2016-09-22T17:36:18Z pipping: stassats: but saving in the middle of something not easily resumable is a bad idea even without forking before s-l-a-d, no?
2016-09-22T17:36:30Z rumbler31 joined #sbcl
2016-09-22T17:36:55Z stassats: forking implies that it's in the middle of something
2016-09-22T17:37:06Z rumbler31 quit (Read error: Connection reset by peer)
2016-09-22T17:37:46Z rumbler31 joined #sbcl
2016-09-22T17:42:00Z Merv_ quit (Remote host closed the connection)
2016-09-22T17:42:11Z rumbler31 quit (Ping timeout: 250 seconds)
2016-09-22T17:42:18Z nyef`: Think of save-lisp-and-die as suspend-to-disk, but without erasing the suspended system image when you restart.
2016-09-22T17:48:24Z pipping: heh, a suspend-to-disk is considerably less flexible, though. with save-lisp-and-die, I get to choose the entry point. Say I wanted to use save-lisp-and-die to create two different executables from some code that I've loaded, which only differ by entry point... If I don't fork (which I guess I cannot with sbcl on windows), I need to start up sbcl, s-l-a-d, start up sbcl again, s-l-a-d again...
2016-09-22T17:48:47Z pipping: s/I wanted/If I wanted/
2016-09-22T17:49:22Z stassats: well, why not do that then?
2016-09-22T17:49:24Z stassats: why overthink it?
2016-09-22T17:52:02Z pipping: I'm not concerned with an actual use case that I have but the design of uiop:dump-image and how it fits into asdf:program-op. What the interface of it is. Currently, afaict, dump-image (which calls s-l-a-d on sbcl) will sometimes exit the repl and sometimes not (I think on cmu cl e.g. it wouldn't). Just wonder if the "your repl may die" really needs to be part of the interface
2016-09-22T17:52:22Z shka joined #sbcl
2016-09-22T17:52:50Z rpg: stassats: Another reason is that anything that requires starting multiple lisp processes cannot be done in ASDF.
2016-09-22T17:53:21Z chris2 joined #sbcl
2016-09-22T17:53:37Z rpg: So I have a lisp system that builds 4 executables (because of different entry points).  I can express each one in the .asd file, but to build them, I have to add a Makefile (so I can start up multiple lisp jobs).
2016-09-22T17:53:48Z rpg: Not tragic, by any means, but somehow unappealing.
2016-09-22T17:55:02Z nyef`: What, can RUN-PROGRAM not invoke SBCL now?
2016-09-22T17:55:31Z rpg: nyef`: It's not an ASDF thing.
2016-09-22T17:55:42Z rpg: ASDF is an "in a single process" build manager.
2016-09-22T17:56:19Z rpg: nyef`: I could fix it to invoke RUN-PROGRAM but compared to make, CL is a really crappy process automation tool.
2016-09-22T17:56:44Z nyef`: So you can't use ASDF to invoke SB-GROVEL to find various structure layouts and symbol values from the C compiler?
2016-09-22T17:56:46Z rumbler31 joined #sbcl
2016-09-22T17:57:12Z rpg: nyef`: ???
2016-09-22T17:57:35Z rumbler31 quit (Remote host closed the connection)
2016-09-22T17:57:36Z nyef`: I'll provisionally not argue the "really crappy process automation tool" bit, but the rest of your argument doesn't hold water.
2016-09-22T17:57:46Z nyef`: ... Unless said water is frozen first.
2016-09-22T17:57:58Z rpg: nyef`: what's the rest?
2016-09-22T17:58:24Z rpg: ASDF does a lot of stuff to maintain the integrity of the current running process.  That's a big part of what it does.
2016-09-22T17:58:33Z rpg: Is that the part you don't agree with?
2016-09-22T17:58:59Z nyef`: No, it's the strong implication that ASDF can't be used with RUN-PROGRAM to do things with processes other than the one running ASDF.
2016-09-22T17:59:01Z rpg: Trying to extend it to run separate processes is a lot of work to do something that make already does really well.
2016-09-22T17:59:27Z rpg: ASDF reasons about build processes by looking at lots of in-process cached information.
2016-09-22T17:59:40Z rpg: It's not like make, which essentially uses only the filesystem.
2016-09-22T17:59:54Z Bike joined #sbcl
2016-09-22T18:00:18Z rpg: nyef`: I'm sure it would be "a mere matter of programming" to get it to reason properly about out-of-process stuff, but that's a lot more work than writing a 10-line makefile.
2016-09-22T18:01:53Z nyef`: And looking to see if anyone has already done the work is also more work than writing the makefile, I'm sure.
2016-09-22T18:02:56Z rpg: nyef`: There's POIU and all sorts of other stuff, but yes, you're right.  And ASDF has already bloated considerably; I wouldn't want to try to turn it into a multi-process system as well. As Fare is fond of saying "that's for ASDF 4"
2016-09-22T18:04:49Z stassats: please, don't change asdf any more
2016-09-22T18:04:55Z stassats: i don't want to deal with the fallout
2016-09-22T18:05:17Z rpg: stassats: Believe me, I deal with the fallout as much as anyone, so I'm motivated for stability.
2016-09-22T18:06:09Z nyef`: Given sb-grovel's grovel-constants-file thing, it should be doable as a separate extension, not a change to ASDF itself.
2016-09-22T18:06:38Z stassats: or cffi-grovel
2016-09-22T18:06:50Z rpg: nyef`: Sorry, you've lost me. I really don't use either SB-GROVEL or CFFI-GROVEL.
2016-09-22T18:06:55Z rpg: Why are those relevant?
2016-09-22T18:19:44Z scymtym joined #sbcl
2016-09-22T18:20:06Z nyef`: Because they have an ASDF integration for a custom component, and that component calls out to a C compiler to do the damage.
2016-09-22T18:21:55Z nyef`: At least SB-GROVEL does, I don't know about CFFI-GROVEL.
2016-09-22T18:26:53Z pipping: But these are two very different issues, no? One is about calling out to a c compiler (which can be difficult enough) which won't have to start from any good 'state'. The other is: You're in your lisp image and it took you hours to get where you are (loading and processing tons of data). Now you want to dump two images without leaving lisp (because you want a solution that's platform-independent so you can't
2016-09-22T18:26:55Z pipping: rely on make, shell, etc.). What do you do?
2016-09-22T18:27:26Z stassats: you change the way you do things
2016-09-22T18:27:54Z stassats: why even save two images in the first place?
2016-09-22T18:28:38Z pipping: Maybe you're creating two executables with different entry points.
2016-09-22T18:28:52Z rpg: stassats: An example of why I thought about this was a compression library.  It could have a "squish" and "unsquish" executable.
2016-09-22T18:29:21Z stassats: pipping: you to have two huge binaries just for different entry points?
2016-09-22T18:29:25Z stassats: want
2016-09-22T18:29:55Z stassats: have you considered command line arguments?
2016-09-22T18:30:01Z stassats: executable names?
2016-09-22T18:30:08Z stassats: like (if (equal (pathname-name (sb-ext:parse-native-namestring (car sb-ext:*posix-argv*))) "foo") (foo) (bar)) ?
2016-09-22T18:30:12Z stassats: one is a symlink
2016-09-22T18:30:12Z rpg: stassats: I actually just checked and I have both /usr/bin/zip and /usr/bin/unzip
2016-09-22T18:30:34Z rpg: And actually I bumped into this issue in a case that had much more radically differing command line arguments.
2016-09-22T18:30:40Z pipping: maybe one is my-product-professional-edition and the other is my-product-intentionally-crippled-edition
2016-09-22T18:30:46Z rpg: Imagine you were building git in CL....
2016-09-22T18:31:00Z stassats: pipping: you can afford to load things twice
2016-09-22T18:31:22Z rpg: Like I said, it's not a huge deal to fire up lisp multiple times, and it's what I do now, but it seems a little off.
2016-09-22T18:31:26Z stassats: and having different entry points? sorry, i can change it in a minute
2016-09-22T18:32:45Z rpg: stassats: Here's an example http://pastebin.com/hyfrrCNt
2016-09-22T18:35:19Z rpg: This was really just an idle conversation that pipping  and I were having -- it seemed somehow not aesthetically pleasing to have to have a build process smeared across ASDF and make instead of all in ASDF, but it's obviously not a crisis!
2016-09-22T18:35:36Z rpg: make handles it fine!
2016-09-22T18:39:11Z nyef` imagines that someone is building git in CL...
2016-09-22T18:39:17Z nyef`: ... Where do I find this thing?
2016-09-22T18:39:18Z shka quit (Quit: Konversation terminated!)
2016-09-22T18:39:29Z shka joined #sbcl
2016-09-22T18:41:30Z stassats: and it's called "clit"?
2016-09-22T18:46:03Z rumbler31 joined #sbcl
2016-09-22T18:51:10Z rumbler31 quit (Ping timeout: 255 seconds)
2016-09-22T18:51:27Z rumbler31 joined #sbcl
2016-09-22T18:59:04Z Shinmera: nyef`: I think there have been VCS done in Lisp before. Can't remember the name of any of them though.
2016-09-22T19:09:28Z fe[nl]ix: rpg: I agree with enforcing that. just make sure that SLIME now dispatches C-c C-c on a .asd to asdf:load-asd instead of the usual cl:compile+cl:load
2016-09-22T19:09:39Z fe[nl]ix: or people will get upset
2016-09-22T19:10:42Z fe[nl]ix: not many but, as usual, those who do are vocal about it
2016-09-22T19:12:05Z rpg: fe[nl]ix: As ASDF has become more complex, there are more and more assumptions baked into it, and we cannot expect that users will even read the manual, much less the code.
2016-09-22T19:12:21Z rpg: So if we are going to make assumptions, we need to enforce those assumptions.
2016-09-22T19:12:36Z rpg: We should probably move this to #asdf....
2016-09-22T19:12:56Z fe[nl]ix: #lisp is better, for a wider audience
2016-09-22T19:13:12Z fe[nl]ix: as long as we're not too prolix
2016-09-22T19:13:14Z fe[nl]ix: :)
2016-09-22T19:14:00Z rpg: #lisp is fine, but this needs to leave #sbcl, I think
2016-09-22T19:23:20Z fe[nl]ix: hahaha
2016-09-22T19:24:10Z rumbler31 quit (Ping timeout: 260 seconds)
2016-09-22T19:28:50Z shka quit (Ping timeout: 272 seconds)
2016-09-22T19:30:27Z slyrus joined #sbcl
2016-09-22T19:31:03Z slyrus: bah. sb-concurrency test error building on macos sierra :(
2016-09-22T19:34:57Z stassats: slyrus: frlock?
2016-09-22T19:38:54Z rpg quit (Ping timeout: 264 seconds)
2016-09-22T19:47:55Z _mjl quit (Ping timeout: 265 seconds)
2016-09-22T20:05:57Z rumbler31 joined #sbcl
2016-09-22T20:09:46Z Quadresce joined #sbcl
2016-09-22T20:20:06Z pootler joined #sbcl
2016-09-22T20:55:07Z madbub quit (Remote host closed the connection)
2016-09-22T21:04:11Z slyrus: bah. worked fine the second time through :)
2016-09-22T21:04:16Z slyrus: I think so
2016-09-22T21:05:02Z edgar-rft joined #sbcl
2016-09-22T21:23:15Z rumbler31 quit (Remote host closed the connection)
2016-09-22T21:36:59Z DeadTrickster quit (*.net *.split)
2016-09-22T21:37:00Z nzambe quit (*.net *.split)
2016-09-22T21:37:02Z pillton quit (*.net *.split)
2016-09-22T21:37:05Z les quit (*.net *.split)
2016-09-22T21:37:05Z leo_song quit (*.net *.split)
2016-09-22T21:37:06Z attila_lendvai quit (*.net *.split)
2016-09-22T21:37:07Z dustinm` quit (*.net *.split)
2016-09-22T21:37:09Z Xof quit (*.net *.split)
2016-09-22T21:37:10Z White_Flame quit (*.net *.split)
2016-09-22T21:37:10Z fiddlerwoaroof quit (*.net *.split)
2016-09-22T21:37:11Z minion quit (*.net *.split)
2016-09-22T21:37:11Z pipping quit (*.net *.split)
2016-09-22T21:37:12Z Shinmera quit (*.net *.split)
2016-09-22T21:37:13Z mood quit (*.net *.split)
2016-09-22T21:37:13Z scymtym quit (*.net *.split)
2016-09-22T21:37:16Z angavrilov quit (*.net *.split)
2016-09-22T21:37:17Z iskander quit (*.net *.split)
2016-09-22T21:37:17Z jackc quit (*.net *.split)
2016-09-22T21:37:17Z loke quit (*.net *.split)
2016-09-22T21:37:19Z p_l quit (*.net *.split)
2016-09-22T21:37:22Z jrm quit (*.net *.split)
2016-09-22T21:37:22Z stassats quit (*.net *.split)
2016-09-22T21:37:26Z WojciechK quit (*.net *.split)
2016-09-22T21:37:27Z jackdaniel quit (*.net *.split)
2016-09-22T21:37:27Z ferada quit (*.net *.split)
2016-09-22T21:37:27Z antoszka quit (*.net *.split)
2016-09-22T21:37:28Z xristos quit (*.net *.split)
2016-09-22T21:37:29Z em1l_ quit (*.net *.split)
2016-09-22T21:37:30Z nyef` quit (*.net *.split)
2016-09-22T21:37:30Z jdz quit (*.net *.split)
2016-09-22T21:37:31Z pkhuong quit (*.net *.split)
2016-09-22T21:37:33Z angular_mike_ quit (*.net *.split)
2016-09-22T21:37:34Z drmeister quit (*.net *.split)
2016-09-22T21:37:35Z logrus quit (*.net *.split)
2016-09-22T21:37:35Z dougk_ quit (*.net *.split)
2016-09-22T21:37:42Z sigjuice quit (*.net *.split)
2016-09-22T21:37:42Z |3b| quit (*.net *.split)
2016-09-22T21:37:42Z edgar-rft quit (*.net *.split)
2016-09-22T21:37:42Z pootler quit (*.net *.split)
2016-09-22T21:37:42Z Quadresce quit (*.net *.split)
2016-09-22T21:37:42Z schjetne quit (*.net *.split)
2016-09-22T21:37:42Z christoph_debian quit (*.net *.split)
2016-09-22T21:37:42Z gko quit (*.net *.split)
2016-09-22T21:37:42Z PuercoPop quit (*.net *.split)
2016-09-22T21:37:42Z abbe quit (*.net *.split)
2016-09-22T21:37:43Z slyrus quit (*.net *.split)
2016-09-22T21:37:48Z irsol quit (*.net *.split)
2016-09-22T21:37:49Z salva quit (*.net *.split)
2016-09-22T21:37:50Z _iwc quit (*.net *.split)
2016-09-22T21:37:51Z flip214 quit (*.net *.split)
2016-09-22T21:37:51Z payphone quit (*.net *.split)
2016-09-22T21:37:52Z abruanese quit (*.net *.split)
2016-09-22T21:37:52Z luis quit (*.net *.split)
2016-09-22T21:37:55Z gargaml quit (*.net *.split)
2016-09-22T21:37:55Z jsnell quit (*.net *.split)
2016-09-22T21:37:57Z nicdev quit (*.net *.split)
2016-09-22T21:37:57Z pchrist quit (*.net *.split)
2016-09-22T21:37:58Z mordocai quit (*.net *.split)
2016-09-22T21:37:58Z Posterdati quit (*.net *.split)
2016-09-22T21:37:59Z whiteline quit (*.net *.split)
2016-09-22T21:38:04Z Quadrescence quit (*.net *.split)
2016-09-22T21:38:10Z Bike quit (*.net *.split)
2016-09-22T21:38:10Z chris2 quit (*.net *.split)
2016-09-22T21:38:11Z foom quit (*.net *.split)
2016-09-22T21:38:11Z eschatologist quit (*.net *.split)
2016-09-22T21:38:11Z Merv quit (*.net *.split)
2016-09-22T21:38:12Z DGASAU quit (*.net *.split)
2016-09-22T21:38:12Z myrkraverk quit (*.net *.split)
2016-09-22T21:38:12Z jibanes quit (*.net *.split)
2016-09-22T21:38:13Z Intensity quit (*.net *.split)
2016-09-22T21:38:14Z brucem quit (*.net *.split)
2016-09-22T21:38:17Z specbot quit (*.net *.split)
2016-09-22T21:38:17Z Cthulhux quit (*.net *.split)
2016-09-22T21:38:18Z joshe quit (*.net *.split)
2016-09-22T21:38:19Z carvite quit (*.net *.split)
2016-09-22T21:38:55Z edgar-rft joined #sbcl
2016-09-22T21:38:55Z pootler joined #sbcl
2016-09-22T21:38:55Z Quadresce joined #sbcl
2016-09-22T21:38:55Z slyrus joined #sbcl
2016-09-22T21:38:55Z scymtym joined #sbcl
2016-09-22T21:38:55Z attila_lendvai joined #sbcl
2016-09-22T21:38:55Z schjetne joined #sbcl
2016-09-22T21:38:55Z gargaml joined #sbcl
2016-09-22T21:38:55Z jrm joined #sbcl
2016-09-22T21:38:55Z em1l_ joined #sbcl
2016-09-22T21:38:55Z stassats joined #sbcl
2016-09-22T21:38:55Z jsnell joined #sbcl
2016-09-22T21:38:55Z logrus joined #sbcl
2016-09-22T21:38:55Z dougk_ joined #sbcl
2016-09-22T21:38:55Z p_l joined #sbcl
2016-09-22T21:38:55Z irsol joined #sbcl
2016-09-22T21:38:55Z nyef` joined #sbcl
2016-09-22T21:38:55Z jdz joined #sbcl
2016-09-22T21:38:55Z salva joined #sbcl
2016-09-22T21:38:55Z christoph_debian joined #sbcl
2016-09-22T21:38:55Z Xof joined #sbcl
2016-09-22T21:38:55Z pipping joined #sbcl
2016-09-22T21:38:55Z White_Flame joined #sbcl
2016-09-22T21:38:55Z dustinm` joined #sbcl
2016-09-22T21:38:55Z sigjuice joined #sbcl
2016-09-22T21:38:55Z fiddlerwoaroof joined #sbcl
2016-09-22T21:38:55Z nicdev joined #sbcl
2016-09-22T21:38:55Z pchrist joined #sbcl
2016-09-22T21:38:55Z angavrilov joined #sbcl
2016-09-22T21:38:55Z _iwc joined #sbcl
2016-09-22T21:38:55Z pkhuong joined #sbcl
2016-09-22T21:38:55Z gko joined #sbcl
2016-09-22T21:38:55Z PuercoPop joined #sbcl
2016-09-22T21:38:55Z |3b| joined #sbcl
2016-09-22T21:38:55Z mordocai joined #sbcl
2016-09-22T21:38:55Z Posterdati joined #sbcl
2016-09-22T21:38:55Z whiteline joined #sbcl
2016-09-22T21:38:55Z flip214 joined #sbcl
2016-09-22T21:38:55Z angular_mike_ joined #sbcl
2016-09-22T21:38:55Z abbe joined #sbcl
2016-09-22T21:38:55Z drmeister joined #sbcl
2016-09-22T21:38:55Z minion joined #sbcl
2016-09-22T21:38:55Z iskander joined #sbcl
2016-09-22T21:38:55Z mood joined #sbcl
2016-09-22T21:38:55Z Shinmera joined #sbcl
2016-09-22T21:38:55Z loke joined #sbcl
2016-09-22T21:38:55Z jackc joined #sbcl
2016-09-22T21:38:55Z antoszka joined #sbcl
2016-09-22T21:38:55Z ferada joined #sbcl
2016-09-22T21:38:55Z jackdaniel joined #sbcl
2016-09-22T21:38:55Z WojciechK joined #sbcl
2016-09-22T21:38:55Z xristos joined #sbcl
2016-09-22T21:38:55Z luis joined #sbcl
2016-09-22T21:38:55Z abruanese joined #sbcl
2016-09-22T21:38:55Z payphone joined #sbcl
2016-09-22T21:40:40Z Bike joined #sbcl
2016-09-22T21:40:40Z chris2 joined #sbcl
2016-09-22T21:40:40Z foom joined #sbcl
2016-09-22T21:40:40Z eschatologist joined #sbcl
2016-09-22T21:40:40Z Merv joined #sbcl
2016-09-22T21:40:40Z DGASAU joined #sbcl
2016-09-22T21:40:40Z myrkraverk joined #sbcl
2016-09-22T21:40:40Z jibanes joined #sbcl
2016-09-22T21:40:40Z Intensity joined #sbcl
2016-09-22T21:40:40Z brucem joined #sbcl
2016-09-22T21:40:40Z specbot joined #sbcl
2016-09-22T21:40:40Z Cthulhux joined #sbcl
2016-09-22T21:40:40Z joshe joined #sbcl
2016-09-22T21:40:40Z carvite joined #sbcl
2016-09-22T21:40:57Z DeadTrickster joined #sbcl
2016-09-22T21:40:57Z nzambe joined #sbcl
2016-09-22T21:41:00Z angular_mike_ quit (Ping timeout: 265 seconds)
2016-09-22T21:41:10Z pillton joined #sbcl
2016-09-22T21:41:24Z Merv quit (Ping timeout: 255 seconds)
2016-09-22T21:41:28Z les joined #sbcl
2016-09-22T21:41:28Z leo_song joined #sbcl
2016-09-22T21:42:43Z Quadrescence joined #sbcl
2016-09-22T21:49:42Z rpg joined #sbcl
2016-09-22T22:07:26Z Merv joined #sbcl
2016-09-22T22:09:29Z angular_mike_ joined #sbcl
2016-09-22T22:20:43Z edgar-rft quit (Quit: edgar-rft)
2016-09-22T22:25:15Z stassats quit (Ping timeout: 276 seconds)
2016-09-22T22:29:54Z stassats joined #sbcl
2016-09-22T22:32:34Z rumbler31 joined #sbcl
2016-09-22T22:36:57Z rumbler31 quit (Ping timeout: 248 seconds)
2016-09-22T22:42:32Z Quadresce: Are there any regex libraries that operate on streams?
2016-09-22T22:48:42Z slyrus quit (Remote host closed the connection)
2016-09-22T22:52:22Z stassats quit (Ping timeout: 255 seconds)
2016-09-22T23:08:26Z attila_lendvai quit (Ping timeout: 252 seconds)
2016-09-22T23:19:53Z cromachina joined #sbcl
2016-09-22T23:22:07Z Xof: what kind of cat is sierra?
2016-09-22T23:29:16Z stassats joined #sbcl
2016-09-22T23:35:50Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2016-09-22T23:50:57Z rpg joined #sbcl
2016-09-22T23:56:43Z pipping: Xof: they switched from cats to national parks / rock formations or something like that
2016-09-22T23:56:55Z pipping: Xof: hence the el capitan
2016-09-23T00:04:27Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2016-09-23T00:17:59Z karswell joined #sbcl
2016-09-23T00:35:07Z rpg joined #sbcl
2016-09-23T00:45:19Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2016-09-23T00:45:45Z rpg joined #sbcl
2016-09-23T00:58:27Z karswell quit (Read error: Connection reset by peer)
2016-09-23T01:21:26Z nyef`: Clearly, they ran out of cat breeds.
2016-09-23T01:23:11Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2016-09-23T01:29:23Z em1l joined #sbcl
2016-09-23T01:32:57Z em1l_ quit (Ping timeout: 265 seconds)
2016-09-23T01:35:44Z rpg joined #sbcl
2016-09-23T02:19:46Z joshe: there's plenty more cats, I think they only wanted to use 9 of them though
2016-09-23T02:20:54Z stassats: joshe: and how many years can a mountain exist before it's washed to the sea?
2016-09-23T02:22:53Z joshe: a long, long time
2016-09-23T02:23:03Z joshe: longer than lisp standardization cycles even
2016-09-23T02:33:27Z DGASAU quit (Read error: Connection reset by peer)
2016-09-23T02:34:37Z eschatologist quit (Quit: ZNC 1.6.3+deb1 - http://znc.in)
2016-09-23T02:35:36Z pchrist quit (Ping timeout: 250 seconds)
2016-09-23T02:37:14Z eschatologist joined #sbcl
2016-09-23T02:39:10Z DGASAU joined #sbcl
2016-09-23T02:40:12Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-23T02:40:41Z Quadresce joined #sbcl
2016-09-23T02:44:12Z pchrist joined #sbcl
2016-09-23T02:46:36Z rpg quit (Quit: Textual IRC Client: www.textualapp.com)
2016-09-23T02:59:52Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-23T03:10:37Z karswell joined #sbcl
2016-09-23T04:15:01Z Quadresce joined #sbcl
2016-09-23T04:51:21Z schjetne quit (Ping timeout: 248 seconds)
2016-09-23T04:56:01Z _mjl joined #sbcl
2016-09-23T04:58:53Z edgar-rft joined #sbcl
2016-09-23T05:09:30Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-23T05:10:51Z _mjl quit (Ping timeout: 276 seconds)
2016-09-23T05:13:56Z shka joined #sbcl
2016-09-23T05:16:56Z shka quit (Read error: No route to host)
2016-09-23T05:17:28Z shka joined #sbcl
2016-09-23T05:50:22Z pootler quit (Remote host closed the connection)
2016-09-23T05:52:33Z Quadresce joined #sbcl
2016-09-23T05:52:38Z Quadresce quit (Changing host)
2016-09-23T05:52:38Z Quadresce joined #sbcl
2016-09-23T06:04:48Z DavidGu joined #sbcl
2016-09-23T06:06:38Z DavidGu quit (Read error: Connection reset by peer)
2016-09-23T06:07:09Z DavidGu joined #sbcl
2016-09-23T06:09:06Z DavidGu quit (Read error: Connection reset by peer)
2016-09-23T06:09:31Z DavidGu joined #sbcl
2016-09-23T06:11:34Z DavidGu quit (Read error: Connection reset by peer)
2016-09-23T06:13:36Z DavidGu joined #sbcl
2016-09-23T06:20:47Z shka quit (Quit: Konversation terminated!)
2016-09-23T06:25:09Z rumbler31 joined #sbcl
2016-09-23T06:29:05Z DavidGu quit (Read error: Connection reset by peer)
2016-09-23T06:29:45Z DavidGu joined #sbcl
2016-09-23T06:30:03Z rumbler31 quit (Ping timeout: 250 seconds)
2016-09-23T06:44:02Z pootler joined #sbcl
2016-09-23T06:45:22Z pootler quit (Remote host closed the connection)
2016-09-23T06:46:34Z gingerale joined #sbcl
2016-09-23T06:57:38Z _mjl joined #sbcl
2016-09-23T07:11:45Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-23T07:20:33Z _mjl quit (Ping timeout: 240 seconds)
2016-09-23T08:01:57Z pootler joined #sbcl
2016-09-23T08:04:25Z attila_lendvai joined #sbcl
2016-09-23T08:16:58Z Bike quit (Quit: elemen)
2016-09-23T08:26:29Z rumbler31 joined #sbcl
2016-09-23T08:30:56Z rumbler31 quit (Ping timeout: 250 seconds)
2016-09-23T08:55:37Z pootler quit (Read error: Connection reset by peer)
2016-09-23T09:38:44Z attila_lendvai quit (Ping timeout: 252 seconds)
2016-09-23T09:47:30Z pootler joined #sbcl
2016-09-23T09:54:06Z _mjl joined #sbcl
2016-09-23T10:09:11Z Merv_ joined #sbcl
2016-09-23T10:12:57Z pootler quit (Ping timeout: 248 seconds)
2016-09-23T10:26:11Z DavidGu quit (Quit: DavidGu)
2016-09-23T10:40:34Z madbub joined #sbcl
2016-09-23T11:06:26Z schjetne joined #sbcl
2016-09-23T11:16:09Z _mjl quit (Ping timeout: 276 seconds)
2016-09-23T11:17:17Z scymtym quit (Ping timeout: 265 seconds)
2016-09-23T11:50:28Z scymtym joined #sbcl
2016-09-23T12:10:35Z rtmpdavid joined #sbcl
2016-09-23T12:26:00Z salva quit (Ping timeout: 244 seconds)
2016-09-23T13:04:42Z cromachina quit (Read error: Connection reset by peer)
2016-09-23T13:09:38Z _mjl joined #sbcl
2016-09-23T13:19:54Z edgar-rft quit (Quit: edgar-rft)
2016-09-23T14:34:00Z DGASAU quit (Read error: Connection reset by peer)
2016-09-23T14:36:32Z DGASAU joined #sbcl
2016-09-23T14:43:16Z rumbler31 joined #sbcl
2016-09-23T15:43:34Z schjetne quit (Ping timeout: 240 seconds)
2016-09-23T15:46:04Z sjl joined #sbcl
2016-09-23T15:57:34Z Bike joined #sbcl
2016-09-23T16:10:09Z rumbler31 quit (Ping timeout: 272 seconds)
2016-09-23T16:22:49Z slyrus joined #sbcl
2016-09-23T16:26:03Z rpg joined #sbcl
2016-09-23T16:32:32Z edgar-rft joined #sbcl
2016-09-23T16:44:37Z madbub quit (Read error: Connection reset by peer)
2016-09-23T16:46:56Z rpg quit (Ping timeout: 265 seconds)
2016-09-23T16:46:59Z madbub joined #sbcl
2016-09-23T16:54:30Z rpg joined #sbcl
2016-09-23T16:59:54Z rtmpdavid quit (Ping timeout: 264 seconds)
2016-09-23T17:00:08Z Merv_ quit (Remote host closed the connection)
2016-09-23T17:02:54Z schjetne joined #sbcl
2016-09-23T17:26:47Z sjl quit (Ping timeout: 272 seconds)
2016-09-23T17:44:18Z pootler joined #sbcl
2016-09-23T17:49:18Z rumbler31 joined #sbcl
2016-09-23T17:59:43Z rpg quit (Ping timeout: 272 seconds)
2016-09-23T18:13:01Z scymtym quit (Ping timeout: 272 seconds)
2016-09-23T18:26:38Z rpg joined #sbcl
2016-09-23T18:27:06Z slyrus quit (Remote host closed the connection)
2016-09-23T18:31:23Z rpg quit (Ping timeout: 272 seconds)
2016-09-23T18:31:24Z pootler quit (Read error: Connection reset by peer)
2016-09-23T18:52:08Z Quadresce joined #sbcl
2016-09-23T19:08:07Z rpg joined #sbcl
2016-09-23T19:10:47Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-23T19:20:27Z gingerale quit (Remote host closed the connection)
2016-09-23T19:34:24Z Quadresce joined #sbcl
2016-09-23T19:45:02Z scymtym joined #sbcl
2016-09-23T20:02:23Z rpg quit (Ping timeout: 244 seconds)
2016-09-23T20:02:25Z pootler joined #sbcl
2016-09-23T20:02:43Z rpg joined #sbcl
2016-09-23T20:03:27Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-23T20:13:11Z Quadresce joined #sbcl
2016-09-23T20:22:48Z _mjl quit (Ping timeout: 276 seconds)
2016-09-23T20:59:34Z irsol quit (Ping timeout: 244 seconds)
2016-09-23T21:01:24Z irsol joined #sbcl
2016-09-23T21:01:52Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-23T21:05:24Z rumbler31 quit (Remote host closed the connection)
2016-09-23T21:17:45Z Quadresce joined #sbcl
2016-09-23T21:31:58Z edgar-rft quit (Quit: edgar-rft)
2016-09-23T21:46:00Z rpg quit (Ping timeout: 276 seconds)
2016-09-23T22:13:14Z rpg joined #sbcl
2016-09-23T22:15:22Z rumbler31 joined #sbcl
2016-09-23T22:20:01Z rumbler31 quit (Ping timeout: 265 seconds)
2016-09-23T22:21:41Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-23T22:23:56Z Quadresce joined #sbcl
2016-09-23T22:24:40Z stassats quit (Ping timeout: 260 seconds)
2016-09-23T22:28:34Z DGASAU quit (Ping timeout: 250 seconds)
2016-09-23T22:30:50Z stassats joined #sbcl
2016-09-23T22:37:03Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-23T22:41:33Z Quadresce joined #sbcl
2016-09-23T22:52:30Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-23T23:22:20Z Quadresce joined #sbcl
2016-09-23T23:24:48Z rpg quit (Ping timeout: 276 seconds)
2016-09-23T23:25:55Z Quadresce quit (Client Quit)
2016-09-23T23:39:07Z rpg joined #sbcl
2016-09-23T23:50:42Z Quadresce joined #sbcl
2016-09-23T23:58:12Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-24T00:05:12Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2016-09-24T00:21:47Z cromachina joined #sbcl
2016-09-24T00:42:01Z Quadresce joined #sbcl
2016-09-24T00:46:48Z Quadresce quit (Client Quit)
2016-09-24T01:28:08Z em1l_ joined #sbcl
2016-09-24T01:31:49Z em1l quit (Ping timeout: 265 seconds)
2016-09-24T02:06:33Z rumbler31 joined #sbcl
2016-09-24T02:10:49Z rumbler31 quit (Ping timeout: 248 seconds)
2016-09-24T03:23:59Z |3b| quit (Remote host closed the connection)
2016-09-24T03:25:11Z |3b| joined #sbcl
2016-09-24T04:01:01Z ccl-logbot joined #sbcl
2016-09-24T04:01:01Z 
2016-09-24T04:01:01Z names: ccl-logbot angavrilov_ madbub pkhuong_ scymtym brucem_ carvite pchrist_ |3b|` schjetne__ myrkraverk leo_song luis Quadrescence specbot les Cthulhux iskander sigjuice minion payphone nyef`` karswell` Merv_ em1l_ cromachina stassats irsol Bike eschatologist pillton nzambe DeadTrickster Intensity jibanes foom chris2 gargaml jrm jsnell logrus dougk_ p_l jdz christoph_debian Xof pipping White_Flame dustinm` fiddlerwoaroof nicdev _iwc gko PuercoPop mordocai
2016-09-24T04:01:01Z names: Posterdati whiteline flip214 abbe drmeister mood Shinmera loke jackc antoszka ferada jackdaniel WojciechK xristos abruanese
2016-09-24T04:14:24Z Merv joined #sbcl
2016-09-24T04:25:04Z ccl-logbot joined #sbcl
2016-09-24T04:25:04Z 
2016-09-24T04:25:04Z names: ccl-logbot irsol schjetne foom Bike_ |3b|`` myrkraverk Cthulhux` leo_song Merv angavrilov_ madbub pkhuong_ scymtym brucem_ carvite pchrist_ luis Quadrescence specbot les iskander sigjuice minion payphone karswell` Merv_ cromachina stassats eschatologist pillton DeadTrickster Intensity jibanes chris2 gargaml jrm jsnell logrus dougk_ p_l jdz christoph_debian Xof pipping White_Flame dustinm` fiddlerwoaroof nicdev _iwc gko PuercoPop mordocai Posterdati
2016-09-24T04:25:05Z names: whiteline flip214 abbe drmeister mood Shinmera loke jackc antoszka ferada jackdaniel WojciechK xristos
2016-09-24T04:26:04Z abruanese joined #sbcl
2016-09-24T04:27:42Z leo_song quit (Ping timeout: 244 seconds)
2016-09-24T04:30:01Z leo_song joined #sbcl
2016-09-24T04:34:40Z em1l joined #sbcl
2016-09-24T04:38:57Z nyef joined #sbcl
2016-09-24T04:39:22Z karswell` quit (Read error: Connection reset by peer)
2016-09-24T04:41:17Z slyrus joined #sbcl
2016-09-24T05:12:38Z minion quit (Ping timeout: 244 seconds)
2016-09-24T05:13:05Z minion joined #sbcl
2016-09-24T05:39:24Z Bike_ is now known as Bike
2016-09-24T06:37:31Z gingerale joined #sbcl
2016-09-24T06:46:11Z brucem_ is now known as brucem
2016-09-24T06:47:50Z brucem quit (Changing host)
2016-09-24T06:47:50Z brucem joined #sbcl
2016-09-24T07:35:30Z joshe joined #sbcl
2016-09-24T07:35:30Z angular_mike_ joined #sbcl
2016-09-24T07:35:30Z fe[nl]ix joined #sbcl
2016-09-24T07:35:30Z Blkt joined #sbcl
2016-09-24T07:35:31Z angular_mike_ quit (Changing host)
2016-09-24T07:35:31Z angular_mike_ joined #sbcl
2016-09-24T07:35:51Z edgar-rft joined #sbcl
2016-09-24T07:35:58Z angular_mike_ quit (Max SendQ exceeded)
2016-09-24T07:44:38Z angular_mike_ joined #sbcl
2016-09-24T07:50:18Z ASau joined #sbcl
2016-09-24T08:05:51Z pkhuong_ quit (Ping timeout: 272 seconds)
2016-09-24T08:07:16Z pkhuong joined #sbcl
2016-09-24T08:07:49Z pkhuong is now known as Guest70538
2016-09-24T08:20:13Z gingerale quit (Remote host closed the connection)
2016-09-24T08:25:06Z pootler joined #sbcl
2016-09-24T08:27:33Z Merv_ quit (Ping timeout: 276 seconds)
2016-09-24T08:37:46Z Bike quit (Quit: megalomanic)
2016-09-24T09:10:56Z _mjl joined #sbcl
2016-09-24T09:14:44Z igajsin1 joined #sbcl
2016-09-24T09:15:00Z igajsin1 left #sbcl
2016-09-24T10:01:32Z _mjl quit (Remote host closed the connection)
2016-09-24T10:06:17Z pootler quit (Remote host closed the connection)
2016-09-24T10:22:16Z specbot quit (Remote host closed the connection)
2016-09-24T10:22:16Z minion quit (Remote host closed the connection)
2016-09-24T10:26:34Z iskander: hello #sbcl
2016-09-24T10:27:05Z minion joined #sbcl
2016-09-24T10:27:05Z specbot joined #sbcl
2016-09-24T10:27:19Z iskander: when is the next release ? 1.3.9 has issues on macos sierra, i saw it is already fixed
2016-09-24T10:27:40Z iskander: a new release would be nice
2016-09-24T10:31:55Z scymtym: iskander: sbcl releases are usually made at the end of each month. the upcoming release is planned for the end of the next week
2016-09-24T10:32:11Z iskander: thanks
2016-09-24T10:36:20Z _mjl joined #sbcl
2016-09-24T10:38:12Z scymtym quit (Ping timeout: 276 seconds)
2016-09-24T10:44:11Z DavidGu joined #sbcl
2016-09-24T11:00:58Z madbub quit (Remote host closed the connection)
2016-09-24T11:07:17Z scymtym joined #sbcl
2016-09-24T11:38:36Z Guest70538 quit (Ping timeout: 244 seconds)
2016-09-24T11:40:12Z pkhuong joined #sbcl
2016-09-24T11:40:36Z pkhuong is now known as Guest86789
2016-09-24T11:41:40Z ASau` joined #sbcl
2016-09-24T11:41:47Z ASau quit (Remote host closed the connection)
2016-09-24T11:45:07Z sjl joined #sbcl
2016-09-24T11:51:44Z pipping: Is it expected and intended that calling `sb-posix:chdir` will affect the result of `sb-posix:getcwd` yet not what "./" refers to?
2016-09-24T11:51:55Z stassats: yes
2016-09-24T11:52:54Z pipping: okay. Is "./" something well-defined that I should be using on SBCL at all?
2016-09-24T11:53:09Z stassats: yes
2016-09-24T11:54:14Z pipping: What is it? The directory sbcl was started in? Never to change?
2016-09-24T11:54:24Z stassats: clhs *d-p-d*
2016-09-24T11:54:24Z specbot: *default-pathname-defaults*: http://www.lispworks.com/reference/HyperSpec/Body/v_defaul.htm
2016-09-24T11:56:06Z pipping: Right. That makes sense. Thanks.
2016-09-24T12:00:30Z rpg joined #sbcl
2016-09-24T12:04:28Z hydan joined #sbcl
2016-09-24T12:08:18Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2016-09-24T12:53:16Z |3b|`` is now known as |3b|
2016-09-24T12:58:46Z DeadTrickster_ joined #sbcl
2016-09-24T13:04:23Z karswell` joined #sbcl
2016-09-24T13:07:21Z DeadTrickster_ quit (Ping timeout: 265 seconds)
2016-09-24T13:16:02Z hydan quit (Remote host closed the connection)
2016-09-24T13:17:37Z karswell` quit (Read error: Connection reset by peer)
2016-09-24T13:28:11Z Cthulhux` quit (Changing host)
2016-09-24T13:28:11Z Cthulhux` joined #sbcl
2016-09-24T13:56:11Z nyef quit (Remote host closed the connection)
2016-09-24T13:56:24Z nyef joined #sbcl
2016-09-24T14:26:02Z DavidGu quit (Quit: DavidGu)
2016-09-24T14:27:09Z gingerale joined #sbcl
2016-09-24T14:47:18Z sjl quit (Ping timeout: 264 seconds)
2016-09-24T14:47:18Z schjetne quit (Ping timeout: 264 seconds)
2016-09-24T14:50:52Z slyrus quit (Ping timeout: 255 seconds)
2016-09-24T14:54:39Z rpg joined #sbcl
2016-09-24T15:19:45Z ASau` is now known as ASau
2016-09-24T15:21:01Z attila_lendvai joined #sbcl
2016-09-24T15:21:01Z attila_lendvai quit (Changing host)
2016-09-24T15:21:01Z attila_lendvai joined #sbcl
2016-09-24T15:25:35Z sjl joined #sbcl
2016-09-24T15:34:07Z rpg quit (Quit: Textual IRC Client: www.textualapp.com)
2016-09-24T16:11:22Z karswell` joined #sbcl
2016-09-24T16:14:01Z pootler joined #sbcl
2016-09-24T16:30:05Z Bike joined #sbcl
2016-09-24T17:35:51Z scymtym quit (Ping timeout: 272 seconds)
2016-09-24T17:59:54Z schjetne joined #sbcl
2016-09-24T18:08:19Z Merv_ joined #sbcl
2016-09-24T18:08:55Z specbot quit (Disconnected by services)
2016-09-24T18:08:57Z specbot joined #sbcl
2016-09-24T18:09:24Z jdz_ joined #sbcl
2016-09-24T18:10:25Z jdz quit (Ping timeout: 265 seconds)
2016-09-24T18:10:26Z pootler quit (Ping timeout: 265 seconds)
2016-09-24T18:10:54Z pchrist_ quit (Ping timeout: 265 seconds)
2016-09-24T18:10:54Z les quit (Ping timeout: 265 seconds)
2016-09-24T18:11:36Z pchrist joined #sbcl
2016-09-24T18:11:37Z les joined #sbcl
2016-09-24T18:25:37Z karswell` quit (Read error: Connection reset by peer)
2016-09-24T18:27:48Z scymtym joined #sbcl
2016-09-24T18:46:01Z ccl-logbot joined #sbcl
2016-09-24T18:46:01Z 
2016-09-24T18:46:01Z names: ccl-logbot foom2 Posterdati angavrilov dougk_ edgar-rft whiteline abruanese_ Cthulhux irsol Bike_ myrkraverk gingerale- abbe_ schjetne payphone_ pootler scymtym les pchrist jdz_ attila_lendvai nyef ASau Guest86789 minion Blkt fe[nl]ix joshe leo_song Merv Quadrescence eschatologist pillton gargaml jrm p_l Xof pipping White_Flame dustinm` fiddlerwoaroof nicdev drmeister xristos WojciechK jackdaniel ferada antoszka Shinmera mood
2016-09-24T18:46:34Z jackdani1l joined #sbcl
2016-09-24T18:46:44Z stassats` joined #sbcl
2016-09-24T18:46:59Z jsnell joined #sbcl
2016-09-24T18:47:05Z luis joined #sbcl
2016-09-24T18:47:06Z christoph_debian joined #sbcl
2016-09-24T18:47:19Z sigjuice joined #sbcl
2016-09-24T18:47:24Z specbot joined #sbcl
2016-09-24T18:47:27Z nyef quit (Ping timeout: 356 seconds)
2016-09-24T18:47:30Z drmeister quit (Ping timeout: 312 seconds)
2016-09-24T18:47:34Z DeadTrickster_ joined #sbcl
2016-09-24T18:47:36Z carvite joined #sbcl
2016-09-24T18:47:38Z jackdaniel quit (Disconnected by services)
2016-09-24T18:47:40Z jackdani1l is now known as jackdaniel
2016-09-24T18:47:42Z loke joined #sbcl
2016-09-24T18:47:47Z |3b|` joined #sbcl
2016-09-24T18:47:52Z logrus joined #sbcl
2016-09-24T18:47:52Z iskander joined #sbcl
2016-09-24T18:47:53Z chris2 joined #sbcl
2016-09-24T18:48:33Z mordocai joined #sbcl
2016-09-24T18:49:03Z Intensity joined #sbcl
2016-09-24T18:49:03Z jibanes joined #sbcl
2016-09-24T18:49:06Z Cthulhux quit (Changing host)
2016-09-24T18:49:06Z Cthulhux joined #sbcl
2016-09-24T18:49:12Z Cthulhux is now known as Cthulhux`
2016-09-24T18:49:17Z brucem joined #sbcl
2016-09-24T18:49:39Z flip214 joined #sbcl
2016-09-24T18:49:39Z flip214 quit (Changing host)
2016-09-24T18:49:39Z flip214 joined #sbcl
2016-09-24T18:50:36Z _mjl joined #sbcl
2016-09-24T18:50:41Z em1l joined #sbcl
2016-09-24T18:50:56Z nyef joined #sbcl
2016-09-24T18:51:01Z PuercoPop joined #sbcl
2016-09-24T18:51:06Z gko joined #sbcl
2016-09-24T18:51:50Z jackc joined #sbcl
2016-09-24T19:20:58Z ccl-logbot joined #sbcl
2016-09-24T19:20:58Z 
2016-09-24T19:20:58Z names: ccl-logbot mood carvite jibanes DeadTrickster mordocai christoph_debian _iwc_ edgar-rft Intensity ASau _mjl stassats pipping foom2 White_Flame PuercoPop chris2 brucem angavrilov fiddlerwoaroof sigjuice Bike Guest97090 dougk__ sjl gingerale myrkraverk nyef` igajsin3 luis` schjetne_ jsnell_ Shinmera- specbot jackc gko em1l flip214 iskander |3b|` loke jackdaniel abbe_ payphone_ pootler scymtym les pchrist jdz_ attila_lendvai minion Blkt fe[nl]ix joshe
2016-09-24T19:20:58Z names: leo_song Merv Quadrescence eschatologist pillton gargaml jrm p_l dustinm` nicdev xristos WojciechK ferada antoszka
2016-09-24T19:22:12Z _iwc_ is now known as _iwc
2016-09-24T19:22:34Z p_l quit (Ping timeout: 244 seconds)
2016-09-24T19:23:05Z payphone_ quit (Ping timeout: 244 seconds)
2016-09-24T19:23:54Z p_l joined #sbcl
2016-09-24T19:26:53Z payphone joined #sbcl
2016-09-24T19:30:13Z abruanese joined #sbcl
2016-09-24T19:32:29Z angular_mike_ joined #sbcl
2016-09-24T19:33:42Z whiteline joined #sbcl
2016-09-24T19:40:09Z ccl-logbot joined #sbcl
2016-09-24T19:40:09Z 
2016-09-24T19:40:09Z names: ccl-logbot scymtym_ irsol ShinmerARGH nyef`` jackc_ Merv_ Quadresce` whiteline angular_mike_ abruanese payphone p_l mood carvite jibanes DeadTrickster _iwc edgar-rft Intensity _mjl stassats pipping foom2 White_Flame PuercoPop chris2 brucem angavrilov fiddlerwoaroof sigjuice Bike Guest97090 dougk__ sjl gingerale igajsin3 luis` schjetne_ jsnell_ specbot gko flip214 iskander |3b|` loke jackdaniel abbe_ les pchrist jdz_ attila_lendvai Blkt fe[nl]ix joshe
2016-09-24T19:40:09Z names: Merv eschatologist pillton gargaml dustinm` nicdev xristos
2016-09-24T19:40:19Z em1l joined #sbcl
2016-09-24T19:40:46Z leo_song joined #sbcl
2016-09-24T19:42:32Z angular_mike_ quit (Ping timeout: 270 seconds)
2016-09-24T19:44:10Z mordocai joined #sbcl
2016-09-24T19:45:03Z DeadTrickster quit (Ping timeout: 272 seconds)
2016-09-24T19:47:45Z |3b|` is now known as |3b|
2016-09-24T19:50:55Z ASau joined #sbcl
2016-09-24T20:06:09Z ccl-logbot joined #sbcl
2016-09-24T20:06:09Z 
2016-09-24T20:06:09Z names: ccl-logbot DeadTrickster angular_mike_ chris2 payphone Quadrescence whiteline jrm ASau mordocai leo_song em1l ShinmerARGH nyef`` Merv_ abruanese p_l carvite jibanes _iwc edgar-rft Intensity _mjl stassats pipping foom2 White_Flame PuercoPop brucem angavrilov fiddlerwoaroof sigjuice Bike dougk__ sjl gingerale igajsin3 luis` schjetne_ jsnell_ specbot Guest97090 gko flip214 iskander |3b| loke jackdaniel abbe_ les pchrist jdz_ attila_lendvai Blkt fe[nl]ix joshe
2016-09-24T20:06:09Z names: Merv eschatologist pillton gargaml dustinm` nicdev xristos
2016-09-24T20:07:41Z irsol joined #sbcl
2016-09-24T20:07:44Z irsol quit (Changing host)
2016-09-24T20:07:44Z irsol joined #sbcl
2016-09-24T20:10:23Z drmeister joined #sbcl
2016-09-24T20:11:00Z Wojciech_K joined #sbcl
2016-09-24T20:11:05Z christoph_debian joined #sbcl
2016-09-24T20:21:35Z ccl-logbot joined #sbcl
2016-09-24T20:21:35Z 
2016-09-24T20:21:35Z names: ccl-logbot ferada sigjuice scymtym fiddlerwoaroof Shinmera White_Flame Bike brucem_ p_l PuercoPop luis foom2 Merv Posterdati WojciechK carvite payphone_ Quadresce` igajsin em1l_ nicdev` angavrilov_ gargaml ASau` minion irsol angular_mike_ Intensity dougk__ sjl gingerale schjetne_ jsnell_ specbot gko iskander |3b| jackdaniel abbe_ les pchrist jdz_ attila_lendvai Blkt fe[nl]ix joshe eschatologist pillton dustinm` xristos
2016-09-24T20:21:49Z leo_song joined #sbcl
2016-09-24T20:22:18Z edgar-rft joined #sbcl
2016-09-24T20:22:22Z mood joined #sbcl
2016-09-24T20:22:33Z loke joined #sbcl
2016-09-24T20:33:06Z ccl-logbot joined #sbcl
2016-09-24T20:33:06Z 
2016-09-24T20:33:06Z names: ccl-logbot jrm PuercoPop scymtym_ chris2 sigjuice brucem_ p_l luis Merv Posterdati WojciechK carvite payphone_ Quadresce` igajsin em1l_ nicdev` angavrilov_ gargaml ASau` irsol angular_mike_ Intensity dougk__ sjl gingerale schjetne_ jsnell_ specbot gko |3b| jackdaniel abbe_ pchrist jdz_ attila_lendvai Blkt fe[nl]ix joshe pillton dustinm` xristos
2016-09-24T20:33:14Z edgar-rft joined #sbcl
2016-09-24T20:34:03Z les joined #sbcl
2016-09-24T20:34:03Z abruanese joined #sbcl
2016-09-24T20:34:03Z eschatologist joined #sbcl
2016-09-24T20:34:03Z iskander joined #sbcl
2016-09-24T20:34:03Z White_Flame joined #sbcl
2016-09-24T20:34:12Z foom2 joined #sbcl
2016-09-24T20:34:56Z mood joined #sbcl
2016-09-24T20:35:20Z minion joined #sbcl
2016-09-24T20:35:20Z stassats joined #sbcl
2016-09-24T20:35:45Z loke joined #sbcl
2016-09-24T20:36:16Z flip214 joined #sbcl
2016-09-24T20:36:16Z flip214 quit (Changing host)
2016-09-24T20:36:16Z flip214 joined #sbcl
2016-09-24T20:36:38Z jackc joined #sbcl
2016-09-24T20:37:00Z christoph_debian joined #sbcl
2016-09-24T20:37:16Z pkhuong joined #sbcl
2016-09-24T20:37:32Z attila_lendvai quit (Ping timeout: 252 seconds)
2016-09-24T20:37:41Z pkhuong is now known as Guest37840
2016-09-24T20:37:48Z fiddlerwoaroof joined #sbcl
2016-09-24T20:37:50Z antoszka joined #sbcl
2016-09-24T20:38:20Z leo_song joined #sbcl
2016-09-24T20:39:24Z Shinmera- joined #sbcl
2016-09-24T20:40:11Z ferada joined #sbcl
2016-09-24T20:40:12Z mordocai joined #sbcl
2016-09-24T20:40:33Z jibanes joined #sbcl
2016-09-24T20:42:17Z joshe quit (Quit: Server buffer killed)
2016-09-24T20:54:04Z ccl-logbot joined #sbcl
2016-09-24T20:54:04Z 
2016-09-24T20:54:04Z names: ccl-logbot loke Wojciech_K luis` edgar-rft Quadresce_ sigjuice iskander_ _iwc ShinmerARGH igajsin payphone joshe Bike White_Flame carvite flip214_ jackc_ pootler jibanes mordocai ferada leo_song antoszka fiddlerwoaroof Guest37840 christoph_debian minion mood foom2 eschatologist abruanese jrm PuercoPop scymtym_ chris2 brucem_ p_l em1l_ nicdev` angavrilov_ gargaml ASau` irsol angular_mike_ dougk__ sjl gingerale schjetne_ jsnell_ specbot gko |3b| jackdaniel
2016-09-24T20:54:04Z names: abbe_ pchrist jdz_ Blkt fe[nl]ix pillton dustinm` xristos
2016-09-24T20:55:17Z les joined #sbcl
2016-09-24T20:55:42Z angular_mike_ quit (Ping timeout: 265 seconds)
2016-09-24T20:56:34Z Intensity joined #sbcl
2016-09-24T20:59:11Z payphone quit (Ping timeout: 244 seconds)
2016-09-24T21:00:00Z payphone joined #sbcl
2016-09-24T21:00:59Z Posterdati joined #sbcl
2016-09-24T21:00:59Z stassats joined #sbcl
2016-09-24T21:05:44Z pipping joined #sbcl
2016-09-24T21:09:20Z Merv joined #sbcl
2016-09-24T21:21:37Z angular_mike_ joined #sbcl
2016-09-24T21:24:13Z Cthulhux joined #sbcl
2016-09-24T21:29:06Z drmeister joined #sbcl
2016-09-24T21:40:00Z Guest37840 quit (Ping timeout: 244 seconds)
2016-09-24T21:40:31Z leo_song quit (Ping timeout: 244 seconds)
2016-09-24T21:45:07Z pkhuong joined #sbcl
2016-09-24T21:45:31Z pkhuong is now known as Guest89203
2016-09-24T21:45:55Z leo_song joined #sbcl
2016-09-24T21:48:00Z Quadrescence joined #sbcl
2016-09-24T21:50:50Z Guest89203 quit (Ping timeout: 244 seconds)
2016-09-24T21:51:53Z mordocai quit (Ping timeout: 244 seconds)
2016-09-24T21:52:39Z drmeister quit (*.net *.split)
2016-09-24T21:52:43Z payphone quit (*.net *.split)
2016-09-24T21:52:45Z les quit (*.net *.split)
2016-09-24T21:52:46Z loke quit (*.net *.split)
2016-09-24T21:52:48Z Quadresce_ quit (*.net *.split)
2016-09-24T21:52:48Z luis` quit (*.net *.split)
2016-09-24T21:58:06Z pkhuong_ joined #sbcl
2016-09-24T21:59:19Z mordocai joined #sbcl
2016-09-24T22:12:10Z cromachina joined #sbcl
2016-09-24T22:13:25Z drmeister joined #sbcl
2016-09-24T22:13:25Z payphone joined #sbcl
2016-09-24T22:13:25Z les joined #sbcl
2016-09-24T22:13:25Z loke joined #sbcl
2016-09-24T22:13:25Z luis` joined #sbcl
2016-09-24T22:17:17Z angular_mike_ quit (Ping timeout: 250 seconds)
2016-09-24T22:18:32Z prxq joined #sbcl
2016-09-24T22:25:37Z stassats quit (Ping timeout: 250 seconds)
2016-09-24T22:28:02Z angular_mike_ joined #sbcl
2016-09-24T22:32:11Z stassats joined #sbcl
2016-09-24T22:47:36Z gingerale quit (Remote host closed the connection)
2016-09-24T22:54:44Z edgar-rft quit (Quit: edgar-rft)
2016-09-24T22:58:00Z prxq quit (Ping timeout: 244 seconds)
2016-09-24T22:58:07Z prxq joined #sbcl
2016-09-24T22:58:22Z nyef`` joined #sbcl
2016-09-24T23:04:40Z chris2 quit (Quit: trotz alledem!)
2016-09-24T23:06:51Z chris2 joined #sbcl
2016-09-24T23:27:00Z slyrus joined #sbcl
2016-09-24T23:36:35Z sjl quit (Ping timeout: 250 seconds)
2016-09-24T23:39:32Z prxq quit (Remote host closed the connection)
2016-09-25T00:28:59Z brucem_ is now known as brucem
2016-09-25T00:29:13Z brucem quit (Changing host)
2016-09-25T00:29:13Z brucem joined #sbcl
2016-09-25T00:56:57Z chris2 quit (Ping timeout: 265 seconds)
2016-09-25T01:00:49Z slyrus quit (Ping timeout: 265 seconds)
2016-09-25T01:01:45Z chris2 joined #sbcl
2016-09-25T01:04:06Z slyrus joined #sbcl
2016-09-25T01:26:23Z em1l joined #sbcl
2016-09-25T01:29:40Z em1l_ quit (Ping timeout: 250 seconds)
2016-09-25T01:43:21Z karswell` joined #sbcl
2016-09-25T01:45:36Z karswell` quit (Remote host closed the connection)
2016-09-25T01:46:21Z karswell` joined #sbcl
2016-09-25T04:51:38Z whiteline joined #sbcl
2016-09-25T04:57:04Z DavidGu joined #sbcl
2016-09-25T05:34:56Z DavidGu quit (Read error: Connection reset by peer)
2016-09-25T05:35:10Z DavidGu joined #sbcl
2016-09-25T05:39:27Z DavidGu quit (Ping timeout: 244 seconds)
2016-09-25T05:47:29Z whiteline quit (Remote host closed the connection)
2016-09-25T05:48:25Z whiteline joined #sbcl
2016-09-25T05:58:58Z _mjl joined #sbcl
2016-09-25T06:00:16Z DavidGu joined #sbcl
2016-09-25T06:04:15Z DavidGu quit (Ping timeout: 244 seconds)
2016-09-25T06:17:15Z DavidGu joined #sbcl
2016-09-25T06:17:50Z p_l quit (Ping timeout: 250 seconds)
2016-09-25T06:19:02Z p_l joined #sbcl
2016-09-25T06:41:02Z Quadresce joined #sbcl
2016-09-25T06:45:20Z Quadresce quit (Client Quit)
2016-09-25T06:51:03Z DavidGu quit (Ping timeout: 240 seconds)
2016-09-25T06:55:02Z DavidGu joined #sbcl
2016-09-25T07:23:20Z DavidGu quit (Ping timeout: 260 seconds)
2016-09-25T07:27:00Z ShinmerARGH is now known as Shinmera
2016-09-25T07:27:53Z Xof joined #sbcl
2016-09-25T07:35:21Z DavidGu joined #sbcl
2016-09-25T07:37:56Z schjetne_ is now known as schjetne
2016-09-25T07:43:48Z gingerale joined #sbcl
2016-09-25T09:03:12Z Bike quit (Quit: kill living beings)
2016-09-25T09:17:31Z ASau`` joined #sbcl
2016-09-25T09:21:08Z ASau` quit (Ping timeout: 250 seconds)
2016-09-25T09:35:36Z edgar-rft joined #sbcl
2016-09-25T09:38:24Z ASau`` is now known as ASau
2016-09-25T09:40:32Z shka joined #sbcl
2016-09-25T11:01:41Z DavidGu quit (Quit: DavidGu)
2016-09-25T12:31:15Z sjl joined #sbcl
2016-09-25T12:32:53Z ASau quit (Ping timeout: 265 seconds)
2016-09-25T12:51:54Z angular_mike_ quit (Read error: Connection reset by peer)
2016-09-25T12:52:19Z les quit (Ping timeout: 272 seconds)
2016-09-25T12:54:52Z drmeister quit (Ping timeout: 272 seconds)
2016-09-25T12:57:03Z angular_mike_ joined #sbcl
2016-09-25T12:58:02Z drmeister joined #sbcl
2016-09-25T12:58:10Z les joined #sbcl
2016-09-25T13:37:14Z sjl quit (Ping timeout: 265 seconds)
2016-09-25T14:08:07Z Posterdati quit (Ping timeout: 250 seconds)
2016-09-25T14:22:13Z Posterdati joined #sbcl
2016-09-25T14:29:43Z Posterdati quit (Ping timeout: 255 seconds)
2016-09-25T14:35:31Z abbe_ is now known as abbe
2016-09-25T14:43:14Z Posterdati joined #sbcl
2016-09-25T15:20:31Z nyef`` quit (Ping timeout: 272 seconds)
2016-09-25T16:00:49Z DGASAU joined #sbcl
2016-09-25T16:05:08Z iskander_ quit (Ping timeout: 250 seconds)
2016-09-25T16:11:15Z iskander joined #sbcl
2016-09-25T16:44:30Z iskander quit (Ping timeout: 260 seconds)
2016-09-25T16:50:00Z iskander joined #sbcl
2016-09-25T17:02:32Z cromachina_ joined #sbcl
2016-09-25T17:06:00Z cromachina quit (Ping timeout: 276 seconds)
2016-09-25T17:19:00Z rpg joined #sbcl
2016-09-25T17:42:58Z Bike joined #sbcl
2016-09-25T17:56:13Z attila_lendvai joined #sbcl
2016-09-25T17:56:13Z attila_lendvai quit (Changing host)
2016-09-25T17:56:13Z attila_lendvai joined #sbcl
2016-09-25T18:14:24Z DavidGu joined #sbcl
2016-09-25T18:16:24Z sjl joined #sbcl
2016-09-25T18:20:40Z DavidGu quit (Quit: DavidGu)
2016-09-25T18:22:44Z rumbler31 joined #sbcl
2016-09-25T18:53:34Z eschatologist quit (Ping timeout: 250 seconds)
2016-09-25T18:57:10Z rumbler31 quit (Remote host closed the connection)
2016-09-25T19:02:01Z rumbler31 joined #sbcl
2016-09-25T19:02:11Z rumbler31 quit (Remote host closed the connection)
2016-09-25T19:02:24Z DavidGu joined #sbcl
2016-09-25T19:02:59Z eschatologist joined #sbcl
2016-09-25T19:11:42Z eschatologist quit (Ping timeout: 265 seconds)
2016-09-25T19:11:57Z rumbler31 joined #sbcl
2016-09-25T19:16:17Z DavidGu quit (Quit: DavidGu)
2016-09-25T19:16:43Z DavidGu joined #sbcl
2016-09-25T19:36:30Z gingerale quit (Remote host closed the connection)
2016-09-25T19:59:25Z rumbler31 quit (Remote host closed the connection)
2016-09-25T20:02:47Z attila_lendvai quit (Ping timeout: 252 seconds)
2016-09-25T20:02:52Z DavidGu quit (Ping timeout: 265 seconds)
2016-09-25T20:13:05Z _mjl quit (Ping timeout: 265 seconds)
2016-09-25T20:46:56Z ASau joined #sbcl
2016-09-25T20:55:27Z DeadTrickster joined #sbcl
2016-09-25T21:06:17Z igajsin quit (Ping timeout: 250 seconds)
2016-09-25T21:10:15Z igajsin joined #sbcl
2016-09-25T21:27:01Z shka quit (Ping timeout: 244 seconds)
2016-09-25T21:33:00Z eschatologist joined #sbcl
2016-09-25T22:26:26Z stassats quit (Ping timeout: 244 seconds)
2016-09-25T22:33:13Z stassats joined #sbcl
2016-09-25T22:39:26Z nyef joined #sbcl
2016-09-25T22:47:59Z nyef` joined #sbcl
2016-09-25T22:51:03Z nyef quit (Ping timeout: 265 seconds)
2016-09-25T23:45:01Z eschatologist quit (Ping timeout: 255 seconds)
2016-09-25T23:47:56Z edgar-rft quit (Quit: edgar-rft)
2016-09-25T23:51:29Z sjl quit (Ping timeout: 240 seconds)
2016-09-25T23:58:19Z eschatologist joined #sbcl
2016-09-26T00:39:33Z nyef`: Okay, which platforms need check-builds this time? I'm already trying an HPPA check-build, but as I'm not sure where I left off with the HPPA backend, only that it still isn't in great shape, it isn't likely to tell us much either way.
2016-09-26T00:40:12Z nyef`: SPARC, MIPS, ARM, and maybe Alpha?
2016-09-26T00:40:39Z nyef`: ... And PPC, if I can swing that?
2016-09-26T00:49:45Z stassats: nyef`: ppc and arm should be okay
2016-09-26T00:50:37Z nyef`: So, just SPARC and MIPS for me at this point?
2016-09-26T00:51:06Z stassats: well, i can launch a mips build as well, but it's le
2016-09-26T00:52:00Z stassats: i launched it anyhow
2016-09-26T00:52:09Z stassats: but i'll get the results in the morning
2016-09-26T00:52:44Z stassats: i guess i'll test windows builds tomorrow as well
2016-09-26T00:52:44Z nyef`: I'll do a MIPS BE build or two this week, possibly tomorrow.
2016-09-26T00:53:04Z nyef`: And I may as well do an Alpha build.
2016-09-26T00:53:56Z stassats: though i only really care about arm64 and windows being buildable, since i make binaries for them and not having a buildable realease tag is a bad idea
2016-09-26T00:54:31Z stassats: for everything else there are users who can do the testing
2016-09-26T00:54:44Z stassats: (since they don't test before the release)
2016-09-26T01:00:03Z nyef`: Testing before the release is too proactive to bother with. d-:
2016-09-26T01:00:19Z stassats: that's why we should break things more often
2016-09-26T01:00:26Z stassats: to condition them to test more often
2016-09-26T01:00:39Z stassats: a cunning plan
2016-09-26T01:00:50Z nyef`: Very well, how about I break the debugger?
2016-09-26T01:01:11Z stassats: "i put a break in your BREAK"
2016-09-26T01:02:47Z nyef`: Actually, we need to talk about the debugger anyway.
2016-09-26T01:03:03Z nyef`: I can fix breakpoints, but there's some overhead.
2016-09-26T01:03:34Z stassats: i never trust them, so i never use them, never have an incentive to fix them
2016-09-26T01:03:46Z stassats: it's rare that i can't redefine a function
2016-09-26T01:04:35Z nyef`: Either a small pad (2-3 bytes on x86oids, an instruction or two on most/all other platforms) that holds a NOP, or adding the concept of a "restart address" for each program location that can have a breakpoint installed on it, that includes an equivalent to the instruction that gets overwritten by the breakpoint followed by a jump to the next instruction.
2016-09-26T01:05:11Z stassats:  breakpoints on function entry, right?
2016-09-26T01:05:25Z stassats: or in callers?
2016-09-26T01:05:30Z stassats: or was that stepping?
2016-09-26T01:05:57Z stassats: or both?
2016-09-26T01:06:09Z nyef`: Breakpoints at any code-location in the debug-info for an object, and on function entry.
2016-09-26T01:06:41Z stassats: that's what, error traps?
2016-09-26T01:06:54Z stassats: beside function calls
2016-09-26T01:07:13Z stassats: error traps are already kinda break pointy
2016-09-26T01:07:52Z nyef`: There are a few places that count as code-locations.  Error traps are one, many function calls wind up being smoe.
2016-09-26T01:07:55Z nyef`: Err... some.
2016-09-26T01:08:50Z nyef`: Breakpoint based stepping is faster when nobody is stepping, slower for a function being stepped for any thread that isn't stepping it, and for a thread being stepped the bottleneck is the user anyway.
2016-09-26T01:09:21Z stassats: i wouldn't really worry about other threads
2016-09-26T01:09:22Z nyef`: Also, trace :encapsulate nil uses breakpoints, so that could be made to work reliably everywhere as well.
2016-09-26T01:10:38Z stassats: we could add more locations, like for each form, but form translation is pretty murky
2016-09-26T01:11:13Z nyef`: Basically, I can make breakpoints work reliably and not break anything else, but there are two mechanisms that I can use to do so, and I'm undecided as to which one to use.
2016-09-26T01:11:17Z stassats: so, i guess stepping around function calls is the most clear and simple point
2016-09-26T01:13:03Z stassats: well, instruction rewriting needs to be 'atomic'
2016-09-26T01:13:18Z stassats: on x86, cc is atomic, but we can't encode the trap kinda
2016-09-26T01:13:58Z nyef`: We can still use a LOCK + 16-bit write to splat down the instruction + trap code.
2016-09-26T01:14:23Z stassats: but on arm64 a trap and and trap code fits into a single instruction
2016-09-26T01:14:25Z nyef`: OR we can switch to UD2-breakpoints everywhere, which frees up INT3 / SIGTRAP for use in the debugger.
2016-09-26T01:14:42Z nyef`: And then we go down the list of ports.
2016-09-26T01:14:45Z stassats: there's a reason darwin uses ud2-breakpoints
2016-09-26T01:15:00Z stassats: because int3 is not reliable
2016-09-26T01:15:18Z nyef`: Right, how is the darwin / OSX port doing these days? I'm suddenly finding that I want to try using it.
2016-09-26T01:15:29Z stassats: it builds
2016-09-26T01:15:31Z stassats: (ha-ha)
2016-09-26T01:15:36Z nyef`: Oh dear.
2016-09-26T01:16:05Z stassats: well, i'm rewriting nanosleep in terms of syscalls, didn't get into this release
2016-09-26T01:16:12Z stassats: other than that, it's ok
2016-09-26T01:16:30Z stassats: (nanosleep happens to be is nonreentrant)
2016-09-26T01:16:38Z nyef`: So, might be worth checking out in a couple of weeks?
2016-09-26T01:16:57Z stassats: you can check it now, the nanosleep problem rarely manifests itself
2016-09-26T01:17:37Z nyef`: Hrm. Might have to install xcode or something first, but okay. I'll add it to my list of things to look into "soon".
2016-09-26T01:18:19Z stassats: i could just plop the nanosleep variant from darwin verbatim, but it's license might be problematic
2016-09-26T01:18:27Z stassats: so have to rewrite it in a different way
2016-09-26T01:19:57Z stassats: and who knows what else is nonreentrant, but nanosleep is pretty likely to be interrupted, and we use unblock interrupts around nanosleep in locks
2016-09-26T01:21:01Z stassats: (and sigchld-handler grabs a lock, there's your reentry)
2016-09-26T01:25:10Z em1l_ joined #sbcl
2016-09-26T01:29:09Z em1l quit (Ping timeout: 272 seconds)
2016-09-26T02:00:02Z karswell` quit (Remote host closed the connection)
2016-09-26T02:31:36Z nyef`: Hrm.
2016-09-26T02:31:53Z nyef`: Looks like something changed that breaks some stuff in the HPPA backend.
2016-09-26T02:32:19Z nyef`: Slowly working my way through the test suite, but...  Eesh.
2016-09-26T02:43:37Z nyef`: ... I thought that we specifically didn't use LOOP anywhere that the host compiler could see?
2016-09-26T02:43:50Z nyef`: (see ENCODE-INTERNAL-ERROR-ARGS)
2016-09-26T03:09:17Z nyef`: Okay, HPPA check-build failed. Something is fairly dreadfully wrong there.
2016-09-26T03:10:22Z nyef`: Might be my local changes, but I somehow don't think so.
2016-09-26T03:25:29Z rpg quit (Ping timeout: 240 seconds)
2016-09-26T03:51:40Z rumbler31 joined #sbcl
2016-09-26T03:58:15Z rumbler31 quit (Remote host closed the connection)
2016-09-26T04:09:29Z karswell joined #sbcl
2016-09-26T04:13:36Z rpg joined #sbcl
2016-09-26T04:17:53Z rpg quit (Ping timeout: 240 seconds)
2016-09-26T04:24:49Z shka joined #sbcl
2016-09-26T05:11:48Z Quadresce` joined #sbcl
2016-09-26T05:13:46Z Merv_ joined #sbcl
2016-09-26T05:14:28Z dustinm`_ joined #sbcl
2016-09-26T05:14:54Z _mjl joined #sbcl
2016-09-26T05:15:28Z shka quit (Ping timeout: 244 seconds)
2016-09-26T05:16:08Z WojciechK joined #sbcl
2016-09-26T05:16:11Z Merv__ joined #sbcl
2016-09-26T05:16:28Z pkhuong joined #sbcl
2016-09-26T05:16:31Z carvite_ joined #sbcl
2016-09-26T05:16:52Z pkhuong is now known as Guest92513
2016-09-26T05:20:07Z sigjuice_ joined #sbcl
2016-09-26T05:20:07Z mordocai_ joined #sbcl
2016-09-26T05:20:58Z mordocai quit (*.net *.split)
2016-09-26T05:20:58Z pkhuong_ quit (*.net *.split)
2016-09-26T05:20:59Z Quadrescence quit (*.net *.split)
2016-09-26T05:21:02Z Merv quit (*.net *.split)
2016-09-26T05:21:03Z pipping quit (*.net *.split)
2016-09-26T05:21:06Z Wojciech_K quit (*.net *.split)
2016-09-26T05:21:07Z sigjuice quit (*.net *.split)
2016-09-26T05:21:10Z carvite quit (*.net *.split)
2016-09-26T05:21:12Z pootler quit (*.net *.split)
2016-09-26T05:21:13Z _iwc quit (*.net *.split)
2016-09-26T05:21:15Z mood quit (*.net *.split)
2016-09-26T05:21:16Z foom2 quit (*.net *.split)
2016-09-26T05:21:18Z dustinm` quit (*.net *.split)
2016-09-26T05:21:24Z carvite_ quit (Changing host)
2016-09-26T05:21:24Z carvite_ joined #sbcl
2016-09-26T05:21:26Z carvite_ is now known as carvite
2016-09-26T05:22:27Z eschatologist quit (Ping timeout: 276 seconds)
2016-09-26T05:23:15Z Merv__ is now known as Merv
2016-09-26T05:24:01Z mood joined #sbcl
2016-09-26T05:25:50Z eschatologist joined #sbcl
2016-09-26T05:27:29Z foom2 joined #sbcl
2016-09-26T05:28:37Z pipping joined #sbcl
2016-09-26T05:29:24Z _mjl quit (Ping timeout: 265 seconds)
2016-09-26T05:30:42Z _iwc_ joined #sbcl
2016-09-26T05:35:12Z _iwc_ is now known as _iwc
2016-09-26T06:03:29Z ASau quit (Ping timeout: 240 seconds)
2016-09-26T06:10:40Z DeadTrickster quit (Quit: Leaving)
2016-09-26T06:10:47Z DeadTrickster joined #sbcl
2016-09-26T06:19:37Z gingerale joined #sbcl
2016-09-26T06:45:48Z flip214_ is now known as flip214
2016-09-26T06:45:53Z flip214 quit (Changing host)
2016-09-26T06:45:53Z flip214 joined #sbcl
2016-09-26T06:54:19Z salva joined #sbcl
2016-09-26T06:55:03Z edgar-rft joined #sbcl
2016-09-26T07:12:06Z attila_lendvai joined #sbcl
2016-09-26T07:12:06Z attila_lendvai quit (Changing host)
2016-09-26T07:12:06Z attila_lendvai joined #sbcl
2016-09-26T07:41:27Z _mjl joined #sbcl
2016-09-26T07:50:54Z igajsin quit (Quit: Leaving.)
2016-09-26T07:59:54Z Guest92513 quit (Ping timeout: 264 seconds)
2016-09-26T08:16:29Z whiteline quit (Remote host closed the connection)
2016-09-26T08:17:31Z whiteline joined #sbcl
2016-09-26T08:28:02Z Bike quit (Quit: leaving)
2016-09-26T08:29:56Z pkhuong joined #sbcl
2016-09-26T08:30:19Z pkhuong is now known as Guest66752
2016-09-26T08:35:02Z gargaml quit (Quit: WeeChat 1.5)
2016-09-26T10:35:11Z ktt9 joined #sbcl
2016-09-26T10:41:09Z ktt9: Hi. Does there any secret lore (in human language) on how sbcl compilation works exist? I'm strongly interested in understanding underlying concepts as well, and my only background is Stanford MOOC on compilers (including writing one), and some private experiments with cl-lex and yacc systems. I'm reading sbcl sources for a week now, and this is quite interesting, I can't cease to wonder if there's something more errr... User-friendly
2016-09-26T10:41:09Z ktt9: and welcoming, and I just didn't happen to find.
2016-09-26T10:42:29Z Shinmera: Legend has it that after the first week or month of Python development nobody understood it completely anymore.
2016-09-26T10:45:00Z jackdaniel: ktt9: find an easy issue and try to solve it, I think its the best way to start understanding a system
2016-09-26T10:45:20Z jackdaniel: ktt9: https://www.pvk.ca/Blog/2013/04/13/starting-to-hack-on-sbcl/ - this is a very good read
2016-09-26T10:52:37Z ktt9: Oh. Thanks!
2016-09-26T10:52:45Z jackdaniel: sure
2016-09-26T10:55:12Z ktt9: Shinmera: but Python is not so big, as far as I know. Why?
2016-09-26T10:56:51Z Shinmera: It's not hard for me to imagine that it becomes difficult to grasp the whole system quickly the smarter it becomes and the more tricks are applied.
2016-09-26T10:57:14Z Shinmera: But I don't have any actual experience with it. I can only tell you what I've heard.
2016-09-26T10:59:47Z ktt9: Oh. I see.
2016-09-26T11:00:37Z ktt9: Quickly, well, yes.
2016-09-26T11:02:07Z ktt9: And completely understand how every part works. But on at least at black box level, that is, how control flows and data propagates, I think anyone who aims to work at the system should understand.
2016-09-26T11:02:25Z Shinmera: Sure.
2016-09-26T11:12:03Z stassats: ktt9: it just takes time
2016-09-26T11:12:15Z stassats: i've been at it for years and there are still gray areas
2016-09-26T11:12:57Z ktt9: Yeah, I know. It's just that a good advice and guidance can reduce that time.
2016-09-26T11:13:10Z flip214: YEARS UPON YEARS OF HARD LABOUR.... and that's only to get the README.1st.
2016-09-26T11:13:27Z stassats: ktt9: my advice is that you want to have a goal
2016-09-26T11:13:30Z stassats: a bug or a feature
2016-09-26T11:13:35Z stassats: otherwise it's futile
2016-09-26T11:13:45Z stassats: flip214: which says "There's no README"
2016-09-26T11:13:57Z flip214: ;)
2016-09-26T11:14:06Z ktt9: stassats: Oh, but I have c:
2016-09-26T11:14:19Z stassats: ktt9: world dominance doesn't count
2016-09-26T11:14:23Z ktt9: :D
2016-09-26T11:14:32Z jackdaniel: damn it
2016-09-26T11:14:40Z jackdaniel leaves CL community
2016-09-26T11:14:43Z jackdaniel: :D
2016-09-26T11:14:46Z ktt9: Why so? I't a good goal to my opinion!
2016-09-26T11:15:18Z stassats: jackdaniel: but where did you manage to find CL "community"?
2016-09-26T11:15:30Z scymtym_ quit (Ping timeout: 250 seconds)
2016-09-26T11:15:48Z jackdaniel: well, hrm, something
2016-09-26T11:15:51Z jackdaniel: ;)
2016-09-26T11:16:06Z whiteline quit (Remote host closed the connection)
2016-09-26T11:16:34Z whiteline joined #sbcl
2016-09-26T11:25:33Z sjl joined #sbcl
2016-09-26T11:28:47Z ktt9: Internals CLiki still returns error 500.
2016-09-26T11:29:09Z stassats: "still", did you expect it to magically resurface?
2016-09-26T11:29:19Z stassats: after years of being down
2016-09-26T11:31:04Z flip214: well, full moon is long gone, so yes
2016-09-26T11:32:29Z ktt9: To be honest, yes. Or at least link to it will be removed from the Links section of sbcl.org
2016-09-26T11:33:50Z ktt9: But what prevents whoever can fix it from fixing it?
2016-09-26T11:34:08Z ktt9: SBCL benchmarks link is also dead.
2016-09-26T11:34:20Z stassats: do we even know who can fix it?
2016-09-26T11:40:04Z flip214: hmmm, I've got a nearly empty coffee cup, let me take a look.
2016-09-26T11:42:28Z ktt9: Hmm. Is sbcl.org under control of someone from this room?
2016-09-26T11:42:40Z stassats: yes
2016-09-26T11:42:41Z flip214: not me
2016-09-26T11:44:15Z ktt9: That's good.
2016-09-26T11:45:11Z fe[nl]ix: ktt9: Python is the name of the compiler in SBCL
2016-09-26T11:53:35Z ktt9: fe[nl]ix: oh my
2016-09-26T11:54:08Z stassats: to be exact, the name of the CMUCL compiler
2016-09-26T11:54:19Z ktt9: I thought it was about that one of Guido van Rosuum.
2016-09-26T11:56:33Z Shinmera: Python the programming language came later.
2016-09-26T11:57:41Z ktt9: Yeah.
2016-09-26T12:02:59Z madbub joined #sbcl
2016-09-26T12:03:41Z flip214: so, are you interested about Python-the-sbcl-compiler or Python-the-programming language, really?
2016-09-26T12:04:20Z ktt9: I'm interested in Python-the-sbcl-compiler, of course.
2016-09-26T12:05:34Z ktt9: I just thought that Shinmera's comment was on big systems in general. And Python language as an example. Thus, it was a misunderstanding on my side. Sorry.
2016-09-26T12:11:03Z madbub quit (Remote host closed the connection)
2016-09-26T12:11:50Z madbub joined #sbcl
2016-09-26T12:12:35Z madbub quit (Remote host closed the connection)
2016-09-26T12:13:24Z madbub joined #sbcl
2016-09-26T12:14:32Z madbub quit (Remote host closed the connection)
2016-09-26T12:16:14Z madbub joined #sbcl
2016-09-26T12:17:13Z `lain joined #sbcl
2016-09-26T12:18:03Z myrkraverk joined #sbcl
2016-09-26T12:22:09Z edgar-rft quit (Quit: edgar-rft)
2016-09-26T12:37:57Z nzambe joined #sbcl
2016-09-26T12:41:32Z nzambe quit (Client Quit)
2016-09-26T12:41:52Z nzambe joined #sbcl
2016-09-26T13:17:22Z edgar-rft joined #sbcl
2016-09-26T13:20:10Z pootler joined #sbcl
2016-09-26T13:22:41Z Merv_ quit (Ping timeout: 240 seconds)
2016-09-26T13:43:09Z rpg joined #sbcl
2016-09-26T13:46:13Z rumbler31 joined #sbcl
2016-09-26T13:53:28Z cromachina_ quit (Read error: Connection reset by peer)
2016-09-26T14:07:07Z _mjl quit (Ping timeout: 244 seconds)
2016-09-26T14:22:31Z pootler quit (Ping timeout: 265 seconds)
2016-09-26T14:26:31Z pootler joined #sbcl
2016-09-26T14:32:02Z ASau joined #sbcl
2016-09-26T14:34:12Z pootler quit (Read error: No route to host)
2016-09-26T14:34:58Z pootler joined #sbcl
2016-09-26T14:44:41Z Guest66752 quit (Ping timeout: 240 seconds)
2016-09-26T14:48:25Z salva quit (Ping timeout: 272 seconds)
2016-09-26T14:52:14Z pkhuong joined #sbcl
2016-09-26T14:52:42Z pkhuong is now known as Guest89469
2016-09-26T14:59:54Z pootler quit
2016-09-26T15:17:35Z rumbler31 quit (Remote host closed the connection)
2016-09-26T15:21:34Z stassats: ok, on mips i have Signal 10 received (PC: 0x40f83c) when building sb-gmp
2016-09-26T15:21:44Z mordocai_ is now known as mordocai
2016-09-26T15:21:49Z stassats: and Signal 10 received (PC: 0x104dd2c) when testing sb-sprof
2016-09-26T15:35:07Z nyef`: And I have an UNBOUND-VARIABLE SB-DISASSEM::SIZE in sb-sprof on SPARC.
2016-09-26T15:35:22Z nyef`: Running the test suite now.
2016-09-26T15:35:34Z nyef`: My first mips check-build is still running.
2016-09-26T15:38:29Z scymtym joined #sbcl
2016-09-26T15:39:55Z ktt9: What mips hardware do you use? Just curious.
2016-09-26T15:40:28Z nyef`: I'm currently doing a build on an EtherLite-3, but I'll also be running one on an Origin 350 (running Linux) later on.
2016-09-26T15:48:03Z nyef`: SAP-REF-INT (compiler/target-disassem) is busted on hppa, mips, sparc, and alpha.
2016-09-26T15:48:51Z nyef`: dougk__: I think this one's on you.
2016-09-26T15:49:54Z Bike joined #sbcl
2016-09-26T15:52:21Z nyef`: Further check-builds are being deferred at this point: Every single platform I have slated for check-builds is affected by this one.
2016-09-26T15:55:25Z Quadresce joined #sbcl
2016-09-26T15:55:26Z nyef`: stassats: Signal 10 is SIGBUS. Not sure what's up there.
2016-09-26T15:57:39Z dougk__: nyef`: looking
2016-09-26T15:58:45Z nyef`: dougk__: Thank you.
2016-09-26T16:00:17Z ktt9: nyef`: oh wow. That's some exotic hardware.
2016-09-26T16:01:51Z nyef`: I'm thinking that it might be amusing to do a 680x0 port one of these days, just to have a new contender for least-useful-SBCL-port to go with the Alpha and HPPA backends.
2016-09-26T16:02:17Z stassats: vax?
2016-09-26T16:02:31Z sjl quit (Ping timeout: 272 seconds)
2016-09-26T16:02:35Z stassats: i imagine 68k is still being used
2016-09-26T16:02:37Z nyef`: Or that. Hey, maybe we can also do a VMS port as well?
2016-09-26T16:03:00Z stassats: so, sb-sprof is broken because it disassembles, why is sb-gmp broken then?
2016-09-26T16:04:16Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-26T16:04:56Z nyef`: That's probably the use of allocate-vector-on-heap not smacking the end of the vector to force a GC-trap.
2016-09-26T16:05:51Z nyef`: A long-known issue, and on my list to clean up one of these days.
2016-09-26T16:05:56Z stassats: sb-gmp test are hella weird
2016-09-26T16:06:15Z stassats: nyef`: it didn't happen before
2016-09-26T16:06:26Z stassats: but that might be explained by libgmp not being installed
2016-09-26T16:07:06Z nyef`: It depends on the alignment of the new allocation with respect to the GC trigger.
2016-09-26T16:07:28Z stassats: which would be unlikely as well
2016-09-26T16:07:38Z stassats: (libgmp)
2016-09-26T16:07:45Z nyef`: SPARC test results are... better than anticipated, given the current situation.
2016-09-26T16:08:28Z stassats: sb-gmp was broken wrt cheneygc
2016-09-26T16:08:42Z stassats: so i guess it's not a new problem and i can stop bothering
2016-09-26T16:08:46Z nyef`: Right.
2016-09-26T16:09:16Z nyef`: And it's not so much broken wrt cheneygc, it's that some platforms are broken wrt cheneygc and alien code and direct access to freshly-allocated heap data.
2016-09-26T16:11:40Z Quadresce joined #sbcl
2016-09-26T16:11:55Z nyef`: Which IS on my list of things to audit for, but it's a longstanding issue and people don't seem to notice it very often. (-:
2016-09-26T16:12:11Z stassats: well, nobody uses cheneygc targets
2016-09-26T16:13:17Z nyef`: Right, because they're MIPS, ALPHA, and HPPA. Every other target that supports cheneygc also supports gencgc.
2016-09-26T16:14:28Z nyef`: Again, another thing on my list of things to fix, but sitting somewhere after getting the various backends up to a certain level of "maintainedness".
2016-09-26T16:16:21Z nyef`: Should I add a NEWS entry for commit 09a7ad10afff6e64111d97fef1eb2f2a3c9a7ddb or is it not worth the bother?
2016-09-26T16:18:09Z nyef`: ... Actually, that's another one that I should audit for. /-:
2016-09-26T16:24:56Z nyef`: Okay, new SPARC check-build running. Thanks, dougk__!
2016-09-26T16:25:53Z nyef`: And MIPS check-build on the ERLite-3 restarted.
2016-09-26T16:26:36Z stassats: i'll set off a mips-le build as well
2016-09-26T16:37:58Z jrm quit (Quit: ciao)
2016-09-26T16:39:47Z jrm joined #sbcl
2016-09-26T17:13:30Z shka joined #sbcl
2016-09-26T17:20:40Z nyef`: Running test suite on SPARC.
2016-09-26T17:21:06Z nyef`: Clean build this time, at least.
2016-09-26T17:30:19Z sjl joined #sbcl
2016-09-26T17:31:57Z foom2 is now known as foom
2016-09-26T17:44:29Z scymtym quit (Ping timeout: 272 seconds)
2016-09-26T17:56:02Z nyef`: SPARC test suite results unchanged from my last recording, whenever it was.
2016-09-26T17:56:19Z stassats: well, sb-gmp no longer fails
2016-09-26T17:56:24Z stassats: sb-sprof still does
2016-09-26T17:57:22Z stassats: not what i expected
2016-09-26T17:59:10Z nyef`: Still waiting on my current MIPS check-build.
2016-09-26T17:59:53Z stassats: and sp-sprof doesn't crash when i repeat the stuff manually
2016-09-26T18:01:54Z stassats: interrupted something, Memory fault at 0x4 (pc=0x30b56974, sp=(nil))
2016-09-26T18:01:56Z stassats: fun stuff
2016-09-26T18:02:12Z stassats: somebody is hogging both cores of the machine, rude
2016-09-26T18:05:03Z stassats: but (load "contrib/sb-sprof/test.lisp") does crash
2016-09-26T18:05:22Z stassats:   12: SB-DI::HANDLE-SINGLE-STEP-BEFORE-TRAP 0x1f69717 fp = 0x76c20a9c LRA = 0x1f697bf pc = 0x70
2016-09-26T18:05:25Z stassats: huh-huh
2016-09-26T18:06:12Z stassats: http://paste.lisp.org/display/327166
2016-09-26T18:07:03Z nyef`: Iiinteresting.
2016-09-26T18:07:13Z nyef`: Especially given the HPPA damage that I was seeing yesterday.
2016-09-26T18:07:33Z stassats: i love how SB-DI::SINGLE-STEP-INFO-FROM-CONTEXT calls SB-ALIEN::ALIEN-TYPE-CLASS-OR-LOSE
2016-09-26T18:07:38Z stassats: so cheap
2016-09-26T18:07:58Z stassats: but there's no stepping being performed, bogus stepping checks again?
2016-09-26T18:09:53Z stassats: but nothing should have changed around that
2016-09-26T18:10:33Z stassats: p *STEPPING* => value: 0x00000000: 0
2016-09-26T18:10:37Z stassats: so, no stepping is going on
2016-09-26T18:12:22Z stassats: don't see anything wrong and it worked before
2016-09-26T18:13:29Z pootler joined #sbcl
2016-09-26T18:15:27Z stassats: but looking at disassembly, there's some instruction scheduling going on, so i can't really make out heads or tails
2016-09-26T18:16:20Z stassats: like, completely
2016-09-26T18:16:34Z stassats: yet it works, apparently, but not under sprof
2016-09-26T18:17:02Z mjl joined #sbcl
2016-09-26T18:17:55Z nyef`: Contribs building now on this end.
2016-09-26T18:18:41Z stassats: i don't get how stepping is getting tested here http://paste.lisp.org/display/327166#1
2016-09-26T18:19:12Z stassats: considering that i wrote the current incarnation of stepping tests for mips
2016-09-26T18:19:58Z stassats: BGEZ $NL4, L2, but OR $NL4, $ZERO, 8
2016-09-26T18:20:58Z nyef`: That's not the single-step trap!
2016-09-26T18:21:08Z nyef`: That's a PA exit sequence.
2016-09-26T18:21:33Z stassats: well, right
2016-09-26T18:21:38Z stassats: i did change trap numbers
2016-09-26T18:21:42Z stassats: huh what?
2016-09-26T18:21:55Z stassats: does it hard code or something?
2016-09-26T18:22:04Z nyef`: Not sure what's going on there yet.
2016-09-26T18:22:28Z nyef`: But look at the actual logic: It's monkeying $ALLOC, it's clearly an allocation sequence, and there's a conditional trap.
2016-09-26T18:23:17Z nyef`: MIPS P-A uses a FLAG-TN of $NL4.
2016-09-26T18:23:33Z nyef`: Yup!
2016-09-26T18:23:44Z nyef`: compiler/mips/macros, have a look at PSEUDO-ATOMIC.
2016-09-26T18:23:47Z nyef`: The BREAK form there.
2016-09-26T18:24:00Z stassats: (inst break 0 16)
2016-09-26T18:24:03Z stassats: goddamn it
2016-09-26T18:24:19Z nyef`: This kind of insanity is ALL OVER these backends.
2016-09-26T18:24:59Z scymtym joined #sbcl
2016-09-26T18:25:15Z stassats: well, at least it's an easy fix
2016-09-26T18:25:22Z nyef`: That it is.
2016-09-26T18:26:54Z sjl quit (Ping timeout: 264 seconds)
2016-09-26T18:26:58Z stassats: ha, i even deleted pseudo-atomic-trap
2016-09-26T18:27:15Z stassats: "Remove pseudo-atomic-trap on mips, unused."
2016-09-26T18:27:19Z stassats: well, duh
2016-09-26T18:27:52Z stassats: well, if it's unused and hard coded, the part that receives it must be hardcoding it too
2016-09-26T18:30:07Z nyef`: ... You renumbered trap codes?
2016-09-26T18:33:30Z stassats: i removed the unused ones
2016-09-26T18:33:42Z nyef`: MIPS-only, or more generally?
2016-09-26T18:33:43Z stassats: well, pseudo-atomic-trap happened to be used but harcodedly
2016-09-26T18:34:00Z stassats: more generally
2016-09-26T18:34:06Z stassats: but most of them didn't get renumbered
2016-09-26T18:34:10Z nyef`: Hrm. I wonder if that's what broke HPPA?
2016-09-26T18:34:28Z stassats: see e01d089efd4f7e36cc3e9c1498d111a2b8ccd4b4
2016-09-26T18:35:03Z stassats: didn't touch hppa
2016-09-26T18:37:43Z slyrus quit (Ping timeout: 265 seconds)
2016-09-26T18:47:11Z stassats: but if 16 is hardcoded in another place, then how did past that and into HANDLE-SINGLE-STEP-BEFORE-TRAP
2016-09-26T18:50:21Z stassats: break	0x0, 0x10 in mips-assem as well
2016-09-26T18:50:23Z stassats: sigh
2016-09-26T18:54:17Z nyef`: Am I still on?
2016-09-26T18:54:28Z stassats: on where?
2016-09-26T18:54:33Z nyef`: Ah, good.
2016-09-26T18:54:42Z nyef`: My computer just took a suspend-to-ram for no apparent reason.
2016-09-26T18:54:54Z nyef`: Didn't know if it had killed my network connections as well.
2016-09-26T18:56:22Z stassats: well, i still don't get how was pseudo-atomic-trap handled before
2016-09-26T18:57:55Z nyef`: You don't? It was 16 in the P-A macro, 0x10 in the runtime, including in the trap logic that you removed from mips-arch.c.
2016-09-26T18:58:22Z stassats: well, yes i'm thinking that it was something i removed indeed
2016-09-26T18:58:39Z nyef`: 510d909a6a7d9a16355d4fba132156293f83c45d
2016-09-26T19:00:08Z stassats: huh, strange
2016-09-26T19:00:41Z stassats: why would i remove that?
2016-09-26T19:00:57Z nyef`: Because it was "unused"?
2016-09-26T19:01:15Z stassats: no, the  interrupt_handle_pending(context); and stuff
2016-09-26T19:02:05Z nyef`: Because it was the pseudo-atomic trap, which is what you were getting rid of.
2016-09-26T19:02:22Z stassats: ok, i think i see why, because it should actually be pending-interrupt-trap
2016-09-26T19:02:33Z stassats: then it'll go into handle_trap and do the right thing
2016-09-26T19:03:02Z nyef`: ... Should it really?
2016-09-26T19:03:28Z stassats: well, it has the same code except for arch_clear_pseudo_atomic_interrupted(context);
2016-09-26T19:03:46Z nyef`: Which is important.
2016-09-26T19:04:19Z stassats: only for mips?
2016-09-26T19:04:48Z nyef`: More generally, isn't it?
2016-09-26T19:05:40Z stassats: interrupt_handle_pending calls arch_clear_pseudo_atomic_interrupted(context);
2016-09-26T19:07:06Z stassats: only hppa and mips call it
2016-09-26T19:07:16Z stassats: and ppc
2016-09-26T19:07:27Z stassats: ok, sparc too
2016-09-26T19:07:30Z stassats: but that's it
2016-09-26T19:08:06Z nyef`: So, most of the cheneygc-enabled ports?
2016-09-26T19:08:17Z stassats: is it the way they do PA?
2016-09-26T19:08:43Z stassats: via reg_ALLOC
2016-09-26T19:09:51Z nyef`: I don't know at this point.
2016-09-26T19:10:30Z nyef`: Giving up on the MIPS test suite run, it stopped in (:SLEEP :PRETTY-MUCH-FOREVER).
2016-09-26T19:10:42Z stassats: happened to me once
2016-09-26T19:15:01Z stassats: well, i don't want to be figuring out whether arch_clear_pseudo_atomic_interrupted is actually needed and will special case
2016-09-26T19:15:05Z stassats: not happy about that, though
2016-09-26T19:16:08Z nyef`: We can put it on the list for next month.
2016-09-26T19:21:16Z stassats: ok, testing a fix
2016-09-26T19:22:21Z stassats: it's strange that i deleted that sigtrap_handler special cases without it really being equivalent
2016-09-26T19:33:24Z nyef`: The point is that it didn't look like it was being used, so it didn't need to be kept.
2016-09-26T19:33:50Z stassats: but still, my first question upon seeing the diff was "huh, why was that deleted?"
2016-09-26T19:34:14Z stassats: now, is single stepping really broken on mips?
2016-09-26T19:34:36Z stassats: or was it combination with PA that broke it?
2016-09-26T19:35:45Z nyef`: Well, if the PA traps were triggering single-step handling...?
2016-09-26T19:41:53Z pootler quit (Read error: Connection reset by peer)
2016-09-26T20:19:29Z pootler joined #sbcl
2016-09-26T20:24:41Z stassats: ok, the windows build is broken
2016-09-26T20:36:32Z stassats: a runtime compilation problem resolved, i have a cold init hiccup
2016-09-26T20:36:41Z stassats: great, just what i need this evening
2016-09-26T20:37:00Z specbot quit (Remote host closed the connection)
2016-09-26T20:37:07Z specbot joined #sbcl
2016-09-26T20:37:53Z stassats: the final assertion in scavenge
2016-09-26T20:38:06Z `lain quit (Remote host closed the connection)
2016-09-26T20:38:35Z stassats: Final object pointer 00000694A6BC79F0, start 0000001002947830, end 0000001002947A00
2016-09-26T20:40:00Z nyef`: Lovely.
2016-09-26T20:40:35Z nyef`: Wait, this is 64-bit. No current 64-bit platform is important, you can let this one go. d-:
2016-09-26T20:40:53Z stassats: true, win32-x86 builds fine
2016-09-26T20:41:21Z stassats: i have no clue, i guess i gotta bisect it
2016-09-26T20:42:46Z stassats: except that i'll have to patch the build failure by hand
2016-09-26T20:42:47Z stassats: sigh
2016-09-26T20:43:05Z stassats: but maybe it was introduced before
2016-09-26T20:44:36Z leo_song quit (Ping timeout: 244 seconds)
2016-09-26T20:46:13Z Xof: stassats: the git bisect man page has an example for patching in a hot-fix while bisecting
2016-09-26T20:46:52Z Xof: (may be more trouble than it's worth)
2016-09-26T20:46:52Z leo_song joined #sbcl
2016-09-26T20:47:19Z stassats: the build failure was introduced pretty late, so maybe i'll get lucky
2016-09-26T20:48:49Z stassats: how the hell can it be so wildly off? 00000694A6BC79F0 and 0000001002947A00
2016-09-26T20:49:45Z stassats: and i'm in luck, it's before the build failure
2016-09-26T20:50:25Z nyef`: Corrupted length of some sort?
2016-09-26T20:50:28Z gingerale quit (Remote host closed the connection)
2016-09-26T20:50:38Z stassats: there were changes to interleaved strctures
2016-09-26T20:55:15Z stassats: well
2016-09-26T20:55:30Z stassats: traded for another build failure, the -std=c99 one
2016-09-26T20:56:24Z stassats: good thing i already narrowed it down to before interleaved structures changes
2016-09-26T20:57:05Z slyrus joined #sbcl
2016-09-26T20:58:04Z eschatologist quit (Ping timeout: 255 seconds)
2016-09-26T21:03:30Z eschatologist joined #sbcl
2016-09-26T21:04:20Z shka quit (Ping timeout: 260 seconds)
2016-09-26T21:04:26Z stassats: well, i think i got it
2016-09-26T21:04:28Z stassats: i see a "long"
2016-09-26T21:05:02Z nyef`: Heh! And that varies by target.
2016-09-26T21:05:12Z stassats: "think hard" beat git bisect
2016-09-26T21:08:30Z nyef`: "All the world's a VAX" syndrome strikes again.
2016-09-26T21:08:57Z stassats: but there are a lot of longs lurking around
2016-09-26T21:10:07Z nyef`: Yet more reason to rewrite as much of the runtime in Lisp as possible. d-:
2016-09-26T21:14:19Z stassats: and without involving libc
2016-09-26T21:14:38Z stassats: which would solve the nanosleep deadlock problem on darwin, at least
2016-09-26T21:17:30Z stassats: well, actually, git bisect is showing a different commit
2016-09-26T21:17:51Z stassats: but the long was problematic too
2016-09-26T21:18:20Z stassats: b66a010ee5c9cfae7d2f4181c437508e05febc34 is what git bisect shows
2016-09-26T21:21:13Z madbub quit (Remote host closed the connection)
2016-09-26T21:25:30Z sjl joined #sbcl
2016-09-26T21:28:27Z stassats: can't figure out what's wrong with it
2016-09-26T21:32:14Z nyef`: It's fairly opaque to me, too.
2016-09-26T21:32:47Z stassats: i'm going to have to actually debug it
2016-09-26T21:45:31Z stassats: ok, one of the scavtab functions returns an outrages value
2016-09-26T21:45:35Z stassats: widetag 25
2016-09-26T21:45:40Z stassats: 0x
2016-09-26T21:46:31Z stassats: complex-single-float-widetag? in cold init?
2016-09-26T21:58:46Z stassats: "profound swearing"
2016-09-26T21:58:50Z stassats: 1LU => 1LLU
2016-09-26T22:01:05Z stassats: how do i avoid an ifdef here?
2016-09-26T22:03:32Z stassats: well, who cares, i don't want to spend a minute longer on this
2016-09-26T22:04:59Z Guest89469 is now known as pkhuong
2016-09-26T22:07:37Z stassats: (uword_t)1 seems like the right choice
2016-09-26T22:12:28Z stassats: release is a go from me
2016-09-26T22:13:04Z nyef`: I've still got platforms to test, but definite progress so far. (-:
2016-09-26T22:16:40Z rszeno joined #sbcl
2016-09-26T22:27:30Z stassats quit (Ping timeout: 276 seconds)
2016-09-26T22:31:34Z stassats joined #sbcl
2016-09-26T22:47:41Z attila_lendvai quit (Ping timeout: 252 seconds)
2016-09-26T22:48:00Z slyrus quit (Ping timeout: 265 seconds)
2016-09-26T23:06:26Z attila_lendvai joined #sbcl
2016-09-26T23:16:43Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-26T23:21:28Z cromachina joined #sbcl
2016-09-26T23:23:35Z Quadresce joined #sbcl
2016-09-26T23:48:59Z sjl quit (Ping timeout: 265 seconds)
2016-09-27T00:14:22Z rumbler31 joined #sbcl
2016-09-27T00:32:17Z myrkraverk quit (Ping timeout: 244 seconds)
2016-09-27T00:33:08Z nyef`: MIPS BE test suite run, dead in fake_foreign_function_call.
2016-09-27T00:33:32Z nyef`: In float.pure, just after :COMPLEX-FLOAT-STACK.
2016-09-27T00:33:34Z stassats: have you updated after the trap changes?
2016-09-27T00:33:46Z nyef`: Current commit is "Fix windows builds."
2016-09-27T00:33:50Z stassats: ok
2016-09-27T00:34:10Z nyef`: So, IIRC, that's a blown P-A bits in $ALLOC.
2016-09-27T00:41:17Z stassats: when just running float.pure.lisp?
2016-09-27T00:41:36Z nyef`: It was a full test suite run.
2016-09-27T00:45:21Z nyef`: Okay, here's one: You can't use arch_clear_pseudo_atomic_interrupted() for a pending-interrupt trap, because it affects $NL4.
2016-09-27T00:46:37Z nyef`: And receive-pending-interrupts doesn't grab NL4, does it?
2016-09-27T00:46:48Z nyef`: No, it doesn't.
2016-09-27T00:46:54Z nyef`: So that's one badness right there.
2016-09-27T00:47:39Z nyef`: Hrm. This test is for PAA, not PAI.
2016-09-27T00:47:59Z stassats: well, pending-interrupt eventually calls arch_clear_pseudo_atomic_interrupted anyway
2016-09-27T00:48:03Z stassats: just not always
2016-09-27T00:50:56Z rszeno quit (Quit: Leaving.)
2016-09-27T00:51:04Z nyef`: Under the circumstances, this might be a random failure, subject to the whims of who-knows-what for actually triggering.
2016-09-27T00:51:19Z nyef`: I'm re-running the test suite now.
2016-09-27T00:54:07Z nyef`: Okay, um... wow. Does interrupt_handle_pending() clear PAI if a pending GC is blocked by without WITHOUT-GCING when leaving the WITHOUT-GCING?
2016-09-27T00:56:01Z nyef`: Oh, ouch. Yes, it does.
2016-09-27T00:56:08Z nyef`: Hrm.
2016-09-27T00:56:10Z nyef`: Does it?
2016-09-27T00:57:10Z stassats: there's a thing called :RESERVE-LOCATIONS
2016-09-27T00:57:17Z stassats: nl4 is one of reserve-non-descriptor-regs
2016-09-27T00:57:22Z stassats: no idea what that means exactly
2016-09-27T00:58:36Z nyef`: Okay, it does. So, that's already badness right there for us.
2016-09-27T00:58:50Z nyef`: Also means Not A Regression This Month.
2016-09-27T00:59:15Z nyef`: ... I think.
2016-09-27T01:04:04Z nyef`: Blam! Test suite died in the same place.
2016-09-27T01:05:18Z nyef`: I don't think that I'm up for trying to track this down tonight.
2016-09-27T01:08:45Z stassats: trying to break NL4, the disassembly looks strange
2016-09-27T01:09:00Z stassats:  SW $NL4, $NFP[4] LW $NL4, $NFP[4]
2016-09-27T01:09:13Z stassats: like it always tries to store NL4
2016-09-27T01:18:59Z stassats: well, NL4 appears to be spilled pretty regularly
2016-09-27T01:19:14Z stassats: can't say that that's a great thing, performance wise
2016-09-27T01:22:04Z stassats: so, sb-unix::receive-pending-interrupt is enough to get it spilled
2016-09-27T01:22:11Z stassats: no idea why, though
2016-09-27T01:24:02Z em1l joined #sbcl
2016-09-27T01:26:05Z attila_lendvai quit (Ping timeout: 252 seconds)
2016-09-27T01:27:25Z em1l_ quit (Ping timeout: 260 seconds)
2016-09-27T01:33:56Z Xof quit (Ping timeout: 244 seconds)
2016-09-27T01:34:27Z stassats: mips seems to reorder array stores and loads
2016-09-27T01:34:38Z stassats: i wonder how that will square with barriers and stuffs
2016-09-27T01:36:09Z stassats: sb-thread:barrier (:compiler) actually does work
2016-09-27T01:36:23Z pkhuong: mips and arm need hardware barriers almost everywhere
2016-09-27T01:36:27Z stassats: for some reason results in lots of vops
2016-09-27T01:37:01Z stassats: pkhuong: i meant the backend reorders stuff
2016-09-27T01:37:35Z nyef`: We can probably defeat that by sticking a WITHOUT-SCHEDULING in the barrier VOPs.
2016-09-27T01:37:51Z nyef`: At the same time, until we have threads, we don't really need the barriers.
2016-09-27T01:37:56Z stassats: %compiler-barrier seems to do enough
2016-09-27T01:38:01Z stassats: even now
2016-09-27T01:38:11Z stassats: but adds a NOP
2016-09-27T01:39:04Z nyef`: Doesn't look like there's a VOP for it on MIPS?
2016-09-27T01:39:26Z stassats: an inlined (defun sb!vm:%compiler-barrier () #!+memory-barrier-vops (sb!vm:%compiler-barrier) (values))
2016-09-27T01:40:06Z nyef`: Right.
2016-09-27T01:41:06Z stassats: is that the instruction scheduler reordering stuff?
2016-09-27T01:42:02Z stassats: or something earlier?
2016-09-27T01:42:14Z stassats: because it doesn't get reordered on x86
2016-09-27T01:43:46Z nyef`: Woudln't surprise me if it were the scheduler.
2016-09-27T01:44:20Z stassats: though the rational for reordering sequential array loads is unclear
2016-09-27T01:45:08Z stassats: and the way it spills NL4 is very strange as well
2016-09-27T01:45:42Z edgar-rft quit (Quit: edgar-rft)
2016-09-27T01:47:09Z stassats: the vops are actually reordered, does the scheduler affect vops too?
2016-09-27T01:48:59Z nyef`: Probably works in terms of basic blocks, to a point.
2016-09-27T01:49:15Z nyef`: You could try doing a build with the scheduler disabled, to see what happens.
2016-09-27T01:49:28Z stassats: well, that'll take an hour
2016-09-27T01:52:24Z stassats: setting *assembly-optimize* to NIL doesn't reorder stuff
2016-09-27T01:53:39Z nyef`: There you go, then.
2016-09-27T01:54:49Z stassats: and seems like there are load delay slots
2016-09-27T01:54:53Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-27T01:54:56Z stassats: still why reorder stuff needlessly
2016-09-27T01:55:48Z nyef`: Probably something nondeterministic in the implementation of the scheduler.
2016-09-27T01:56:11Z stassats: do load delay slots even matter for mips nowadays?
2016-09-27T01:56:44Z nyef`: Not on superscalar chips, but I don't know if modern MIPS chips are superscalar very often.
2016-09-27T01:57:43Z nyef`: I know that most of my MIPS hardware is MIPS IV, which IS superscalar, but some of it is broadcom and cavium SOCs, and I have no idea about those.
2016-09-27T01:58:46Z stassats: that's probably why debugging the mips backend isn't that easy
2016-09-27T01:58:56Z stassats: all those reorders, are they even right?
2016-09-27T01:59:27Z stassats: i have no new test failures on mips-le
2016-09-27T02:00:35Z nyef`: I think that we're at the point of one, possibly longstanding, observed failure on MIPS, and two very-plausible issues found via desk-checking.
2016-09-27T02:04:27Z stassats: the never ending battle with bugs
2016-09-27T02:04:38Z stassats: compiler, os, hardware
2016-09-27T02:04:46Z stassats: i'm amazed that anything works at all
2016-09-27T02:05:10Z nyef`: Now, see, this is why I want a verified compiler. Where at least we know that the compiler works, generates correct code, and so on.
2016-09-27T02:11:07Z nyef`: Ugh. Changing the TEST_SBCL_ARGS or whatever it is to try and get LDB on lossage masks the issue.
2016-09-27T02:11:26Z nyef`: Increases the chance of this being a longstanding lossage, at least.
2016-09-27T02:42:25Z Quadresce joined #sbcl
2016-09-27T03:21:15Z loke quit (Remote host closed the connection)
2016-09-27T03:23:03Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-27T03:25:09Z nyef`: I'm giving MIPS a tentative pass at this point. There's clearly some deep brokenness, but I can't point to it being a recent change.
2016-09-27T03:25:09Z loke joined #sbcl
2016-09-27T04:00:14Z Quadresce joined #sbcl
2016-09-27T04:15:59Z shka joined #sbcl
2016-09-27T04:25:16Z shka quit (Quit: Konversation terminated!)
2016-09-27T04:54:50Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-27T05:05:05Z shka joined #sbcl
2016-09-27T05:12:27Z ASau quit (Ping timeout: 276 seconds)
2016-09-27T05:14:24Z rpg quit (Ping timeout: 276 seconds)
2016-09-27T05:29:29Z shka quit (Quit: Konversation terminated!)
2016-09-27T05:46:28Z pootler quit (Remote host closed the connection)
2016-09-27T06:03:00Z scymtym quit (Ping timeout: 265 seconds)
2016-09-27T06:09:30Z jdz_ is now known as jdz
2016-09-27T06:24:37Z rumbler31 quit (Ping timeout: 255 seconds)
2016-09-27T06:27:18Z rumbler31 joined #sbcl
2016-09-27T06:28:01Z salva joined #sbcl
2016-09-27T06:29:27Z _mjl joined #sbcl
2016-09-27T06:39:24Z myrkraverk joined #sbcl
2016-09-27T06:51:16Z gingerale joined #sbcl
2016-09-27T07:10:59Z edgar-rft joined #sbcl
2016-09-27T07:43:47Z scymtym joined #sbcl
2016-09-27T07:51:15Z myrkraverk quit (Ping timeout: 272 seconds)
2016-09-27T08:00:34Z Bike quit (Quit: coward)
2016-09-27T08:23:15Z attila_lendvai joined #sbcl
2016-09-27T08:35:29Z pootler joined #sbcl
2016-09-27T09:04:48Z DeadTrickster_ joined #sbcl
2016-09-27T09:31:24Z myrkraverk joined #sbcl
2016-09-27T09:32:37Z scymtym: i ran into a strange problem with the package system and/or reader: after resolving a name conflict caused by USE-PACKAGE (or rather the :use DEFPACKAGE option), both symbols appear in DO-SYMBOLS and the reader seems to just choose one of the two https://www.techfak.uni-bielefeld.de/~jmoringe/reader-problem.png (the PARAMETERS symbols are the problem). slightly changing the forms can flip the "chosen" symbols. i will try
2016-09-27T09:32:37Z scymtym: reduced test case
2016-09-27T09:33:56Z scymtym: here is a paste: http://paste.lisp.org/display/327210
2016-09-27T09:34:03Z scymtym: i don't know why i made a screenshot
2016-09-27T09:57:42Z payphone_ joined #sbcl
2016-09-27T09:57:55Z payphone quit (Ping timeout: 272 seconds)
2016-09-27T10:02:48Z scymtym: apparently, the TAKE-NEW restart does not work if both conflicting symbols are inherited. and the randomness is probably due to MRU-TABLE-INDEX
2016-09-27T10:46:00Z schjetne quit (Ping timeout: 250 seconds)
2016-09-27T10:59:35Z sjl joined #sbcl
2016-09-27T11:15:06Z rumbler31 quit (Remote host closed the connection)
2016-09-27T11:15:44Z rumbler31 joined #sbcl
2016-09-27T11:17:12Z angavrilov_ is now known as angavrilov
2016-09-27T11:20:16Z rumbler31 quit (Ping timeout: 255 seconds)
2016-09-27T11:20:39Z rumbler31 joined #sbcl
2016-09-27T11:23:09Z `lain joined #sbcl
2016-09-27T12:00:36Z attila_lendvai quit (Read error: No route to host)
2016-09-27T12:15:36Z attila_lendvai joined #sbcl
2016-09-27T12:37:49Z madbub joined #sbcl
2016-09-27T12:37:50Z `lain quit (Read error: Connection reset by peer)
2016-09-27T12:38:48Z `lain joined #sbcl
2016-09-27T12:43:46Z nyef`: Okay, good news. I'm getting basically the same test results on 1.3.9 MIPS, so if we're looking at a regression, it's either not this month or really subtle.
2016-09-27T12:44:50Z nyef`: I still need to do the O350 build...  Probably two, so that I have a 1.3.9 baseline to work from there as well, and the Alpha build, and figure out WTF is up with HPPA, but yay, progress!
2016-09-27T12:48:05Z `lain quit (Read error: Connection reset by peer)
2016-09-27T12:48:54Z `lain joined #sbcl
2016-09-27T13:04:16Z scymtym quit (Quit: rebooting)
2016-09-27T13:06:43Z scymtym joined #sbcl
2016-09-27T13:19:35Z ktt9: That's cool! Also what Alpha hardware do you have?
2016-09-27T13:19:36Z madbub quit (Read error: Connection reset by peer)
2016-09-27T13:19:44Z madbub_ joined #sbcl
2016-09-27T13:23:55Z nyef`: I have a half-set-up DS10L, a more-than-half-dead DS10L (as a parts donor), and a not-set-up CS20 (with no drive slide, for some reason, WTF?)
2016-09-27T13:29:03Z rpg joined #sbcl
2016-09-27T13:38:02Z scymtym_ joined #sbcl
2016-09-27T13:38:19Z scymtym quit (Ping timeout: 272 seconds)
2016-09-27T13:42:02Z scymtym joined #sbcl
2016-09-27T13:45:55Z scymtym_ quit (Ping timeout: 272 seconds)
2016-09-27T13:49:14Z schjetne joined #sbcl
2016-09-27T13:51:00Z cromachina quit (Read error: Connection reset by peer)
2016-09-27T14:40:48Z scymtym_ joined #sbcl
2016-09-27T14:41:12Z scymtym quit (Ping timeout: 276 seconds)
2016-09-27T14:44:35Z rpg quit (Quit: Textual IRC Client: www.textualapp.com)
2016-09-27T14:50:57Z stassats quit (Ping timeout: 276 seconds)
2016-09-27T14:53:09Z flip214: (TRACE "DBUS") and then running the notify-example make my SBCL go into "Heap exhausted, game over."...
2016-09-27T14:53:32Z flip214: although it got a 3GB heap ;/
2016-09-27T14:53:59Z flip214: looks like some argument or return value can't be pretty printed.
2016-09-27T15:09:19Z _mjl quit (Ping timeout: 255 seconds)
2016-09-27T15:18:09Z nyef`: Recursive structure, and no *print-pretty*?
2016-09-27T15:22:11Z nyef`: ... There's a DBUS library?
2016-09-27T15:24:06Z pootler quit (Ping timeout: 276 seconds)
2016-09-27T15:26:28Z pootler joined #sbcl
2016-09-27T15:34:49Z rumbler31 quit (Remote host closed the connection)
2016-09-27T15:41:08Z dougk__ quit (Remote host closed the connection)
2016-09-27T15:44:09Z ktt9: nyef`: just wow. Never ever heard of that stuff.
2016-09-27T15:49:02Z dougk_ joined #sbcl
2016-09-27T15:58:37Z gargaml joined #sbcl
2016-09-27T16:00:02Z salva quit (Ping timeout: 252 seconds)
2016-09-27T16:01:03Z Bike joined #sbcl
2016-09-27T16:02:31Z gargaml quit (Read error: Connection reset by peer)
2016-09-27T16:03:22Z gargaml joined #sbcl
2016-09-27T16:08:40Z gargaml quit (Read error: Connection reset by peer)
2016-09-27T16:08:50Z attila_lendvai quit (Ping timeout: 252 seconds)
2016-09-27T16:10:34Z gargaml joined #sbcl
2016-09-27T16:11:12Z gargaml quit (Client Quit)
2016-09-27T16:27:18Z Quadresce joined #sbcl
2016-09-27T16:27:26Z eschatologist quit (Quit: ZNC 1.6.3+deb1 - http://znc.in)
2016-09-27T16:28:03Z eschatologist joined #sbcl
2016-09-27T16:35:08Z specbot quit (Disconnected by services)
2016-09-27T16:35:43Z Quadrescence joined #sbcl
2016-09-27T16:35:54Z specbot joined #sbcl
2016-09-27T16:36:44Z myrkraverk_ joined #sbcl
2016-09-27T16:37:04Z gingerale quit (Ping timeout: 255 seconds)
2016-09-27T16:37:05Z mood quit (Ping timeout: 255 seconds)
2016-09-27T16:37:08Z Quadresce` quit (Ping timeout: 240 seconds)
2016-09-27T16:37:11Z sigjuice_ quit (Ping timeout: 265 seconds)
2016-09-27T16:37:13Z myrkraverk quit (Ping timeout: 265 seconds)
2016-09-27T16:37:18Z angular_mike_ quit (Ping timeout: 272 seconds)
2016-09-27T16:37:28Z DeadTrickster_ quit (Ping timeout: 260 seconds)
2016-09-27T16:37:29Z edgar-rft quit (Ping timeout: 260 seconds)
2016-09-27T16:37:29Z em1l quit (Ping timeout: 260 seconds)
2016-09-27T16:37:30Z carvite quit (Ping timeout: 260 seconds)
2016-09-27T16:37:30Z WojciechK quit (Ping timeout: 260 seconds)
2016-09-27T16:37:31Z DGASAU quit (Ping timeout: 260 seconds)
2016-09-27T16:37:32Z dougk_ quit (Ping timeout: 255 seconds)
2016-09-27T16:37:37Z DeadTrickster quit (Ping timeout: 272 seconds)
2016-09-27T16:37:37Z foom quit (Ping timeout: 272 seconds)
2016-09-27T16:37:40Z schjetne quit (Ping timeout: 265 seconds)
2016-09-27T16:37:41Z Quadresce quit (Ping timeout: 265 seconds)
2016-09-27T16:37:43Z myrkraverk_ is now known as myrkraverk
2016-09-27T16:37:56Z dougk_ joined #sbcl
2016-09-27T16:37:59Z foom joined #sbcl
2016-09-27T16:38:00Z ktt9` joined #sbcl
2016-09-27T16:38:02Z `lain` joined #sbcl
2016-09-27T16:38:06Z _iwc quit (Ping timeout: 260 seconds)
2016-09-27T16:38:10Z pkhuong quit (Ping timeout: 265 seconds)
2016-09-27T16:38:11Z pipping quit (Ping timeout: 265 seconds)
2016-09-27T16:38:14Z schjetne joined #sbcl
2016-09-27T16:38:16Z Quadresce joined #sbcl
2016-09-27T16:38:23Z jrm quit (Ping timeout: 244 seconds)
2016-09-27T16:38:25Z gingerale- joined #sbcl
2016-09-27T16:38:28Z ktt9 quit (Ping timeout: 244 seconds)
2016-09-27T16:38:30Z `lain quit (Write error: Connection reset by peer)
2016-09-27T16:38:43Z DeadTrickster joined #sbcl
2016-09-27T16:39:06Z carvite joined #sbcl
2016-09-27T16:39:38Z nzambe quit (Ping timeout: 257 seconds)
2016-09-27T16:40:33Z mjl quit (Ping timeout: 265 seconds)
2016-09-27T16:41:08Z sigjuice joined #sbcl
2016-09-27T16:42:23Z pootler quit (Ping timeout: 252 seconds)
2016-09-27T16:46:08Z gingerale joined #sbcl
2016-09-27T16:46:14Z myrkraverk quit (Ping timeout: 252 seconds)
2016-09-27T16:46:56Z dustinm`_ quit (Ping timeout: 240 seconds)
2016-09-27T16:46:58Z ktt9`` joined #sbcl
2016-09-27T16:47:06Z madbub_ quit (Ping timeout: 240 seconds)
2016-09-27T16:47:16Z sigjuice quit (Ping timeout: 272 seconds)
2016-09-27T16:47:18Z dustinm` joined #sbcl
2016-09-27T16:47:20Z iskander quit (Ping timeout: 272 seconds)
2016-09-27T16:47:20Z p_l quit (Ping timeout: 272 seconds)
2016-09-27T16:47:21Z scymtym__ joined #sbcl
2016-09-27T16:47:26Z madbub joined #sbcl
2016-09-27T16:47:31Z sigjuice joined #sbcl
2016-09-27T16:47:38Z eschatologist_ joined #sbcl
2016-09-27T16:47:39Z DGASAU joined #sbcl
2016-09-27T16:47:42Z schjetne quit (Ping timeout: 272 seconds)
2016-09-27T16:47:42Z dougk_ quit (Ping timeout: 272 seconds)
2016-09-27T16:47:43Z sjl quit (Ping timeout: 272 seconds)
2016-09-27T16:47:53Z schjetne joined #sbcl
2016-09-27T16:48:01Z myrkraverk_ joined #sbcl
2016-09-27T16:48:03Z foom quit (Ping timeout: 255 seconds)
2016-09-27T16:48:04Z eschatologist quit (Excess Flood)
2016-09-27T16:48:04Z foom joined #sbcl
2016-09-27T16:48:07Z Bike_ joined #sbcl
2016-09-27T16:48:08Z leo_song quit (Ping timeout: 244 seconds)
2016-09-27T16:48:13Z gingerale- quit (Ping timeout: 244 seconds)
2016-09-27T16:48:13Z Bike quit (Ping timeout: 244 seconds)
2016-09-27T16:48:14Z `lain`` joined #sbcl
2016-09-27T16:48:16Z dougk__ joined #sbcl
2016-09-27T16:48:17Z `lain` quit (Read error: Connection reset by peer)
2016-09-27T16:48:22Z ktt9` quit (Ping timeout: 264 seconds)
2016-09-27T16:48:24Z myrkraverk_ is now known as myrkraverk
2016-09-27T16:48:24Z scymtym_ quit (Ping timeout: 272 seconds)
2016-09-27T16:48:24Z angular_mike_ joined #sbcl
2016-09-27T16:48:35Z jrm joined #sbcl
2016-09-27T16:48:46Z pootler joined #sbcl
2016-09-27T16:48:50Z nyef`: Hrm. Test gc.impure.lisp / :bug-981106 is taking its own sweet time at 100% CPU.
2016-09-27T16:49:35Z p_l joined #sbcl
2016-09-27T16:49:53Z Wojciech_K joined #sbcl
2016-09-27T16:50:33Z pipping joined #sbcl
2016-09-27T16:51:17Z em1l joined #sbcl
2016-09-27T16:51:17Z iskander joined #sbcl
2016-09-27T16:51:19Z sjl joined #sbcl
2016-09-27T16:51:22Z edgar-rft joined #sbcl
2016-09-27T16:53:44Z pkhuong joined #sbcl
2016-09-27T16:54:03Z leo_song joined #sbcl
2016-09-27T16:54:07Z pkhuong is now known as Guest25046
2016-09-27T16:55:17Z nyef`: Ah, just slow, it seems.
2016-09-27T16:55:23Z mood joined #sbcl
2016-09-27T17:05:15Z _iwc joined #sbcl
2016-09-27T17:06:02Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-27T17:07:49Z rumbler31 joined #sbcl
2016-09-27T17:09:01Z Merv_ joined #sbcl
2016-09-27T17:09:04Z payphone joined #sbcl
2016-09-27T17:10:54Z carvite quit (Ping timeout: 240 seconds)
2016-09-27T17:10:54Z pootler quit (Read error: Connection reset by peer)
2016-09-27T17:11:07Z ktt9`` quit (Ping timeout: 240 seconds)
2016-09-27T17:11:08Z nyef`: Okay, MIPS is good to go. I still need to do Alpha and then try to figure out what's going on with HPPA.
2016-09-27T17:11:18Z leo_song_ joined #sbcl
2016-09-27T17:11:19Z mood quit (Ping timeout: 272 seconds)
2016-09-27T17:11:19Z Wojciech_K quit (Ping timeout: 272 seconds)
2016-09-27T17:11:20Z scymtym__ quit (Ping timeout: 272 seconds)
2016-09-27T17:11:20Z dustinm` quit (Ping timeout: 272 seconds)
2016-09-27T17:11:21Z luis` quit (Ping timeout: 272 seconds)
2016-09-27T17:11:22Z _iwc quit (Excess Flood)
2016-09-27T17:11:23Z scymtym__ joined #sbcl
2016-09-27T17:11:28Z Wojciech_K joined #sbcl
2016-09-27T17:11:41Z dougk_ joined #sbcl
2016-09-27T17:11:43Z luis` joined #sbcl
2016-09-27T17:11:56Z `lain`` quit (Ping timeout: 244 seconds)
2016-09-27T17:11:57Z payphone_ quit (Ping timeout: 244 seconds)
2016-09-27T17:11:59Z p_l quit (Ping timeout: 255 seconds)
2016-09-27T17:12:02Z dougk__ quit (Ping timeout: 255 seconds)
2016-09-27T17:12:02Z Bike_ quit (Ping timeout: 255 seconds)
2016-09-27T17:12:13Z dustinm` joined #sbcl
2016-09-27T17:12:18Z leo_song quit (Ping timeout: 264 seconds)
2016-09-27T17:12:20Z ktt9``` joined #sbcl
2016-09-27T17:12:22Z em1l quit (Ping timeout: 264 seconds)
2016-09-27T17:12:23Z carvite_ joined #sbcl
2016-09-27T17:12:27Z leo_song_ is now known as leo_song
2016-09-27T17:12:28Z carvite_ quit (Changing host)
2016-09-27T17:12:28Z carvite_ joined #sbcl
2016-09-27T17:12:29Z Bike joined #sbcl
2016-09-27T17:12:35Z em1l_ joined #sbcl
2016-09-27T17:12:43Z carvite_ is now known as carvite
2016-09-27T17:14:55Z angular_mike_ quit (Ping timeout: 265 seconds)
2016-09-27T17:16:30Z p_l joined #sbcl
2016-09-27T17:18:35Z rumbler31 quit (Remote host closed the connection)
2016-09-27T17:20:03Z ktt9```` joined #sbcl
2016-09-27T17:20:56Z DGASAU` joined #sbcl
2016-09-27T17:21:48Z sjl quit (Read error: Connection reset by peer)
2016-09-27T17:21:53Z sjl_ joined #sbcl
2016-09-27T17:21:53Z iskander quit (Ping timeout: 240 seconds)
2016-09-27T17:21:55Z _iwc joined #sbcl
2016-09-27T17:21:56Z scymtym__ quit (Ping timeout: 272 seconds)
2016-09-27T17:22:03Z foom2 joined #sbcl
2016-09-27T17:22:07Z scymtym__ joined #sbcl
2016-09-27T17:22:12Z DGASAU quit (Read error: Connection reset by peer)
2016-09-27T17:22:30Z myrkraverk quit (Ping timeout: 255 seconds)
2016-09-27T17:22:31Z foom quit (Ping timeout: 255 seconds)
2016-09-27T17:22:35Z dougk_ quit (Ping timeout: 272 seconds)
2016-09-27T17:22:55Z ktt9``` quit (Ping timeout: 264 seconds)
2016-09-27T17:22:55Z carvite quit (Ping timeout: 264 seconds)
2016-09-27T17:22:59Z dougk_ joined #sbcl
2016-09-27T17:23:03Z iskander joined #sbcl
2016-09-27T17:28:04Z DeadTrickster quit (*.net *.split)
2016-09-27T17:29:07Z Merv_ quit (Remote host closed the connection)
2016-09-27T17:29:55Z DGASAU` is now known as DGASAU
2016-09-27T17:31:51Z sjl_ quit (Ping timeout: 244 seconds)
2016-09-27T17:32:52Z carvite joined #sbcl
2016-09-27T17:32:53Z iskander quit (Ping timeout: 272 seconds)
2016-09-27T17:32:53Z leo_song quit (Ping timeout: 272 seconds)
2016-09-27T17:32:56Z madbub quit (Remote host closed the connection)
2016-09-27T17:32:59Z Guest25046 quit (Ping timeout: 272 seconds)
2016-09-27T17:32:59Z edgar-rft quit (Ping timeout: 272 seconds)
2016-09-27T17:33:19Z dougk_ quit (Ping timeout: 272 seconds)
2016-09-27T17:33:19Z foom2 quit (Ping timeout: 272 seconds)
2016-09-27T17:33:21Z edgar-rft joined #sbcl
2016-09-27T17:33:42Z dougk_ joined #sbcl
2016-09-27T17:33:58Z leo_song_ joined #sbcl
2016-09-27T17:34:05Z madbub joined #sbcl
2016-09-27T17:34:10Z leo_song_ is now known as leo_song
2016-09-27T17:34:22Z foom2 joined #sbcl
2016-09-27T17:34:48Z nyef`: Alpha checkbuild running.
2016-09-27T17:36:26Z myrkraverk joined #sbcl
2016-09-27T17:40:00Z iskander joined #sbcl
2016-09-27T17:41:31Z mood joined #sbcl
2016-09-27T17:44:55Z pkhuong joined #sbcl
2016-09-27T17:45:19Z pkhuong is now known as Guest34542
2016-09-27T17:48:14Z mjl joined #sbcl
2016-09-27T17:48:19Z eschatologist_ is now known as eschatologist
2016-09-27T17:53:20Z angular_mike_ joined #sbcl
2016-09-27T17:55:52Z nzambe joined #sbcl
2016-09-27T17:57:41Z ktt9````` joined #sbcl
2016-09-27T18:01:21Z ktt9```` quit (Ping timeout: 240 seconds)
2016-09-27T18:02:54Z ktt9````` quit (Ping timeout: 264 seconds)
2016-09-27T18:04:13Z Quadresce joined #sbcl
2016-09-27T18:10:15Z leo_song_ joined #sbcl
2016-09-27T18:10:42Z foom2 quit (Ping timeout: 272 seconds)
2016-09-27T18:10:42Z dougk_ quit (Ping timeout: 272 seconds)
2016-09-27T18:10:43Z edgar-rft quit (Ping timeout: 272 seconds)
2016-09-27T18:10:43Z carvite quit (Ping timeout: 272 seconds)
2016-09-27T18:10:44Z scymtym__ quit (Ping timeout: 272 seconds)
2016-09-27T18:10:57Z scymtym__ joined #sbcl
2016-09-27T18:11:03Z dougk_ joined #sbcl
2016-09-27T18:11:29Z Guest34542 quit (Ping timeout: 264 seconds)
2016-09-27T18:11:30Z leo_song quit (Ping timeout: 264 seconds)
2016-09-27T18:11:36Z foom2 joined #sbcl
2016-09-27T18:11:36Z carvite joined #sbcl
2016-09-27T18:11:37Z leo_song_ is now known as leo_song
2016-09-27T18:12:16Z edgar-rft joined #sbcl
2016-09-27T18:13:39Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-27T18:17:22Z pkhuong_ joined #sbcl
2016-09-27T18:20:58Z p_l quit (Ping timeout: 272 seconds)
2016-09-27T18:21:01Z mjl_ joined #sbcl
2016-09-27T18:21:04Z mjl_ quit (Changing host)
2016-09-27T18:21:04Z mjl_ joined #sbcl
2016-09-27T18:21:34Z Quadresce joined #sbcl
2016-09-27T18:21:53Z mjl quit (Ping timeout: 264 seconds)
2016-09-27T18:23:46Z p_l joined #sbcl
2016-09-27T18:23:52Z pootler joined #sbcl
2016-09-27T18:24:22Z mjl_ is now known as mjl
2016-09-27T18:32:37Z rumbler31 joined #sbcl
2016-09-27T18:39:31Z nyef`: "fatal error encountered in SBCL pid 12005: hash_table table!=this table 32a183b7"
2016-09-27T18:40:05Z nyef`: Seventeen(!) contribs failed to build.
2016-09-27T18:41:03Z rpg joined #sbcl
2016-09-27T18:58:13Z eschatologist quit (Ping timeout: 272 seconds)
2016-09-27T19:00:45Z rpg quit (Read error: Connection reset by peer)
2016-09-27T19:03:57Z nyef`: Looks like something went wrong trying to compile uiop, and produced a short fasl, and then almost everything else choked on that.
2016-09-27T19:04:30Z nyef`: Repeatable, too.
2016-09-27T19:05:42Z nyef`: Right, and the short fasl for uiop is caused by the hash_table table!=this table thing.
2016-09-27T19:05:44Z nyef`: Lovely.
2016-09-27T19:11:25Z madbub quit (Remote host closed the connection)
2016-09-27T19:11:37Z eschatologist joined #sbcl
2016-09-27T19:14:08Z madbub joined #sbcl
2016-09-27T19:16:31Z eschatologist quit (Ping timeout: 272 seconds)
2016-09-27T19:18:59Z _mjl joined #sbcl
2016-09-27T19:19:13Z gingerale quit (Remote host closed the connection)
2016-09-27T19:19:44Z pootler quit (Read error: Connection reset by peer)
2016-09-27T19:20:13Z eschatologist joined #sbcl
2016-09-27T19:20:17Z pootler joined #sbcl
2016-09-27T19:21:28Z pootler quit (Read error: Connection reset by peer)
2016-09-27T19:26:03Z rumbler3_ joined #sbcl
2016-09-27T19:26:33Z les quit (Ping timeout: 244 seconds)
2016-09-27T19:26:54Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-27T19:26:57Z p_l quit (Ping timeout: 272 seconds)
2016-09-27T19:27:01Z angular_mike_ quit (Ping timeout: 272 seconds)
2016-09-27T19:27:07Z Bike quit (Ping timeout: 272 seconds)
2016-09-27T19:27:20Z foom2 quit (Ping timeout: 272 seconds)
2016-09-27T19:27:20Z dougk_ quit (Ping timeout: 272 seconds)
2016-09-27T19:27:20Z scymtym__ quit (Ping timeout: 272 seconds)
2016-09-27T19:27:21Z rumbler31 quit (Ping timeout: 265 seconds)
2016-09-27T19:27:36Z p_l joined #sbcl
2016-09-27T19:27:43Z foom2 joined #sbcl
2016-09-27T19:27:43Z dougk_ joined #sbcl
2016-09-27T19:27:59Z Bike_ joined #sbcl
2016-09-27T19:29:37Z les joined #sbcl
2016-09-27T19:34:03Z specbot quit (Remote host closed the connection)
2016-09-27T19:34:51Z minion quit (Disconnected by services)
2016-09-27T19:34:55Z minion joined #sbcl
2016-09-27T19:35:12Z irsol quit (Ping timeout: 250 seconds)
2016-09-27T19:35:13Z jsnell_ quit (Ping timeout: 250 seconds)
2016-09-27T19:35:24Z angavrilov quit (Read error: Connection reset by peer)
2016-09-27T19:35:30Z specbot joined #sbcl
2016-09-27T19:35:34Z Shinmera quit (Ping timeout: 250 seconds)
2016-09-27T19:35:34Z christoph_debian quit (Ping timeout: 250 seconds)
2016-09-27T19:35:35Z abruanese quit (Ping timeout: 250 seconds)
2016-09-27T19:35:47Z mjl quit (Ping timeout: 272 seconds)
2016-09-27T19:35:53Z foom2 quit (Ping timeout: 272 seconds)
2016-09-27T19:36:02Z foom2 joined #sbcl
2016-09-27T19:36:02Z abruanese joined #sbcl
2016-09-27T19:36:04Z angavrilov joined #sbcl
2016-09-27T19:36:11Z dougk_ quit (Ping timeout: 272 seconds)
2016-09-27T19:36:18Z christop1 joined #sbcl
2016-09-27T19:36:24Z iskander quit (Ping timeout: 264 seconds)
2016-09-27T19:36:26Z Bike_ quit (Ping timeout: 264 seconds)
2016-09-27T19:36:37Z dougk_ joined #sbcl
2016-09-27T19:36:44Z iskander_ joined #sbcl
2016-09-27T19:36:46Z Shinmera joined #sbcl
2016-09-27T19:39:04Z Quadresce joined #sbcl
2016-09-27T19:39:51Z eschatologist quit (Ping timeout: 265 seconds)
2016-09-27T19:40:24Z angular_mike_ joined #sbcl
2016-09-27T19:42:07Z Quadresce quit (Client Quit)
2016-09-27T19:44:46Z iskander_ quit (Ping timeout: 272 seconds)
2016-09-27T19:44:53Z Shinmera- joined #sbcl
2016-09-27T19:45:05Z dougk_ quit (Ping timeout: 272 seconds)
2016-09-27T19:45:06Z christop1 quit (Ping timeout: 272 seconds)
2016-09-27T19:45:07Z foom2 quit (Ping timeout: 272 seconds)
2016-09-27T19:45:28Z foom2 joined #sbcl
2016-09-27T19:45:28Z dougk_ joined #sbcl
2016-09-27T19:45:34Z eschatologist joined #sbcl
2016-09-27T19:45:40Z angular_mike_ quit (Ping timeout: 272 seconds)
2016-09-27T19:45:49Z Shinmera quit (Ping timeout: 264 seconds)
2016-09-27T19:46:00Z christoph_debian joined #sbcl
2016-09-27T19:46:22Z iskander joined #sbcl
2016-09-27T19:48:14Z irsol joined #sbcl
2016-09-27T19:48:14Z Quadresce joined #sbcl
2016-09-27T19:49:40Z Bike joined #sbcl
2016-09-27T19:51:21Z eschatologist quit (Ping timeout: 272 seconds)
2016-09-27T19:55:34Z slyrus joined #sbcl
2016-09-27T19:56:21Z eschatologist joined #sbcl
2016-09-27T20:00:03Z angular_mike_ joined #sbcl
2016-09-27T20:00:29Z mjl joined #sbcl
2016-09-27T20:02:30Z jsnell joined #sbcl
2016-09-27T20:04:42Z eschatologist quit (Ping timeout: 264 seconds)
2016-09-27T20:06:22Z scymtym joined #sbcl
2016-09-27T20:12:58Z eschatologist joined #sbcl
2016-09-27T20:20:41Z _mjl quit (Ping timeout: 240 seconds)
2016-09-27T20:21:55Z Cthulhux quit (Ping timeout: 250 seconds)
2016-09-27T20:22:21Z flip214 quit (Ping timeout: 250 seconds)
2016-09-27T20:23:07Z scymtym quit (Ping timeout: 272 seconds)
2016-09-27T20:23:09Z irsol quit (Ping timeout: 272 seconds)
2016-09-27T20:23:10Z foom2 quit (Ping timeout: 272 seconds)
2016-09-27T20:23:15Z _iwc quit (Excess Flood)
2016-09-27T20:23:17Z pootler joined #sbcl
2016-09-27T20:23:32Z scymtym joined #sbcl
2016-09-27T20:23:43Z dougk_ quit (Ping timeout: 272 seconds)
2016-09-27T20:23:46Z Cthulhux joined #sbcl
2016-09-27T20:24:00Z slyrus quit (Ping timeout: 264 seconds)
2016-09-27T20:24:07Z slyrus joined #sbcl
2016-09-27T20:24:13Z dougk_ joined #sbcl
2016-09-27T20:24:15Z ShinmerARGH joined #sbcl
2016-09-27T20:24:16Z foom2 joined #sbcl
2016-09-27T20:24:19Z Shinmera- quit (Ping timeout: 272 seconds)
2016-09-27T20:24:52Z irsol joined #sbcl
2016-09-27T20:27:13Z eschatologist quit (Ping timeout: 265 seconds)
2016-09-27T20:27:27Z attila_lendvai joined #sbcl
2016-09-27T20:27:27Z attila_lendvai quit (Changing host)
2016-09-27T20:27:27Z attila_lendvai joined #sbcl
2016-09-27T20:27:29Z flip214 joined #sbcl
2016-09-27T20:27:29Z flip214 quit (Changing host)
2016-09-27T20:27:29Z flip214 joined #sbcl
2016-09-27T20:28:00Z eschatologist joined #sbcl
2016-09-27T20:32:39Z pootler quit (Ping timeout: 272 seconds)
2016-09-27T20:32:42Z angular_mike_ quit (Ping timeout: 272 seconds)
2016-09-27T20:32:46Z mjl quit (Ping timeout: 272 seconds)
2016-09-27T20:32:47Z Bike quit (Ping timeout: 272 seconds)
2016-09-27T20:32:51Z Bike joined #sbcl
2016-09-27T20:33:04Z slyrus quit (Ping timeout: 264 seconds)
2016-09-27T20:33:13Z foom2 quit (Ping timeout: 272 seconds)
2016-09-27T20:33:13Z dougk_ quit (Ping timeout: 272 seconds)
2016-09-27T20:33:33Z dougk_ joined #sbcl
2016-09-27T20:33:33Z foom2 joined #sbcl
2016-09-27T20:33:35Z flip214 quit (Read error: Connection reset by peer)
2016-09-27T20:33:41Z rpg joined #sbcl
2016-09-27T20:33:42Z flip214 joined #sbcl
2016-09-27T20:33:42Z flip214 quit (Changing host)
2016-09-27T20:33:42Z flip214 joined #sbcl
2016-09-27T20:33:44Z slyrus_ joined #sbcl
2016-09-27T20:33:52Z slyrus_ is now known as slyrus
2016-09-27T20:43:11Z Shinmera joined #sbcl
2016-09-27T20:44:01Z _iwc joined #sbcl
2016-09-27T20:45:12Z ShinmerARGH quit (Ping timeout: 264 seconds)
2016-09-27T20:45:13Z christoph_debian quit (Ping timeout: 264 seconds)
2016-09-27T20:45:44Z mjl joined #sbcl
2016-09-27T20:46:30Z foom2 quit (Ping timeout: 272 seconds)
2016-09-27T20:46:38Z dougk_ quit (Ping timeout: 272 seconds)
2016-09-27T20:46:39Z eschatologist quit (Ping timeout: 272 seconds)
2016-09-27T20:46:44Z dougk_ joined #sbcl
2016-09-27T20:46:46Z foom2 joined #sbcl
2016-09-27T20:46:56Z christoph_debian joined #sbcl
2016-09-27T20:47:03Z angular_mike_ joined #sbcl
2016-09-27T20:48:37Z eschatologist joined #sbcl
2016-09-27T20:51:48Z Wojciech_K quit (Ping timeout: 244 seconds)
2016-09-27T20:52:31Z Wojciech_K joined #sbcl
2016-09-27T21:14:06Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-27T21:24:43Z pootler joined #sbcl
2016-09-27T21:30:03Z pootler quit (Ping timeout: 276 seconds)
2016-09-27T21:31:23Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2016-09-27T21:32:56Z Quadresce joined #sbcl
2016-09-27T21:34:16Z Quadresce: what's sbcl dev team's attitude toward doing things like simplifying macroexpansions
2016-09-27T21:35:04Z Quadresce: like simplifying (let ((x x) ...) to just ... when appropriate? Is the add'l complexity not worth the purely readability gain?
2016-09-27T21:36:01Z madbub quit (Ping timeout: 240 seconds)
2016-09-27T21:36:27Z Quadresce: Particular example:
2016-09-27T21:36:28Z Quadresce:    (LET ((#:RESULT0 RESULT))
2016-09-27T21:36:28Z Quadresce:      (LET ((#:NEW1 ELEMENT))
2016-09-27T21:36:28Z Quadresce:        (FUNCALL #'(SETF AREF) #:NEW1 #:RESULT0 0)))
2016-09-27T21:36:58Z nyef`: "The compiler will already do the work, so why bother?"
2016-09-27T21:37:22Z dougk_: and symbol-macros forbid naive techniques for deciding that something is side-effect-free
2016-09-27T21:37:31Z nyef`: Oh, there you go.
2016-09-27T21:37:38Z nyef`: And I've DONE tricky things with symbol-macros, too.
2016-09-27T21:37:39Z fe[nl]ix: OTOH I really like how readable macroexpansions are on Allegro
2016-09-27T21:38:15Z dougk_: maybe that compiler is worse at eliminating single-use lets so they had to make a trade-off
2016-09-27T21:38:40Z fe[nl]ix: or even on Clozure
2016-09-27T21:39:44Z fe[nl]ix: one nice thing would be for cl:macroexpand to always rebind *gensym-counter* to 0
2016-09-27T21:39:48Z dougk_: to properly simplify you have to look inside macros all the way down. people might be surprised by that
2016-09-27T21:40:19Z dougk_: i think gensyms are hooey.  just use make-symbol, otherwise your image ends up with 47000 different spellings of #:STREAM-nnnn for no reason
2016-09-27T21:40:23Z ASau joined #sbcl
2016-09-27T21:40:24Z nyef`: ... Have to expand the inner macros in order to determine if the outer macros can use the simpler form?
2016-09-27T21:40:33Z dougk_: yes
2016-09-27T21:41:13Z nyef`: And then, in order for macroexpand-1 to make sense, you have to use the un-expanded form in your expansion.
2016-09-27T21:41:25Z dougk_: yup
2016-09-27T21:42:03Z fe[nl]ix: dougk_: right, but many people do use it and it's nice to have readable expansions with at 1-2 digits instead of 5
2016-09-27T21:42:14Z nyef`: And there are scenarios where this might be appropriate, but it's a lot of work both on the part of the macro writer and at expansion time.
2016-09-27T21:42:52Z Quadresce: Maybe there's a good way to add only a little bit of extra compiler code to improve the 90% case
2016-09-27T21:43:05Z dougk_: it doesn't help the compiler at all
2016-09-27T21:43:17Z nyef`: It helps the user experience, though.
2016-09-27T21:44:30Z Quadresce: This goes without saying, but unlike say GCC, Lisp compilers are really central to the entire development process and there's more surface area that the user touches, so thinking of it as just a translator isn't probably the best thing to do.
2016-09-27T21:47:12Z Quadresce: (the bad UX here is just coming from numerical code that has tons of array sets and gets, and the setf expanders are just awful with double lets. I wish that it would translate after macroexpansion into (BLAH-VECTOR-SET A N X)... )
2016-09-27T22:02:21Z pkhuong_: if you're looking at your own code, you could use inline functions to block recursive expansion
2016-09-27T22:18:59Z nicdev` is now known as nicdev
2016-09-27T22:21:36Z nyef`: Looks like we have another regression this month.
2016-09-27T22:22:19Z nyef`: I'm building last months release, and it is building contribs.
2016-09-27T22:22:43Z nyef`: ... Which means that I should be doing the same on HPPA as well.
2016-09-27T22:25:40Z pootler joined #sbcl
2016-09-27T22:29:41Z pootler quit (Ping timeout: 240 seconds)
2016-09-27T22:48:49Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-27T22:49:44Z DGASAU` joined #sbcl
2016-09-27T22:50:38Z DGASAU quit (Ping timeout: 244 seconds)
2016-09-27T22:52:42Z Posterdati quit (Ping timeout: 244 seconds)
2016-09-27T23:04:03Z Quadresce joined #sbcl
2016-09-27T23:04:40Z cromachina joined #sbcl
2016-09-27T23:06:15Z Posterdati joined #sbcl
2016-09-27T23:14:44Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-27T23:16:47Z Quadresce joined #sbcl
2016-09-27T23:16:53Z Quadresce: I'm very surprised. On x64 I get almost no perf benefit from using single vs double-float
2016-09-27T23:19:04Z nyef`: Are the floats getting boxed?
2016-09-27T23:21:26Z Quadresce: I tried pretty hard to avoid it
2016-09-27T23:21:30Z scymtym quit (Remote host closed the connection)
2016-09-27T23:21:32Z Quadresce: (it's for a big application)
2016-09-27T23:21:45Z scymtym joined #sbcl
2016-09-27T23:22:33Z nyef`: Well, unboxed floats, of either sort, aren't going to have the same sort of impact as boxed floats, and with a 64-bit host platform it's the different box styles that gives the massive disparity between single and double floats.
2016-09-27T23:26:18Z pootler joined #sbcl
2016-09-27T23:31:36Z pootler quit (Ping timeout: 276 seconds)
2016-09-27T23:42:59Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-27T23:44:21Z Quadresce joined #sbcl
2016-09-27T23:45:40Z Quadresce quit (Client Quit)
2016-09-28T00:12:28Z sjl joined #sbcl
2016-09-28T00:17:00Z Quadresce joined #sbcl
2016-09-28T00:22:22Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-28T00:27:12Z pootler joined #sbcl
2016-09-28T00:32:03Z pootler quit (Ping timeout: 276 seconds)
2016-09-28T00:34:10Z rumbler3_ quit (Remote host closed the connection)
2016-09-28T00:43:16Z Quadrescence quit (Ping timeout: 244 seconds)
2016-09-28T00:43:47Z luis` quit (Ping timeout: 244 seconds)
2016-09-28T00:45:55Z luis` joined #sbcl
2016-09-28T00:52:45Z Quadrescence joined #sbcl
2016-09-28T00:54:33Z stassats joined #sbcl
2016-09-28T01:12:21Z sjl quit (Ping timeout: 240 seconds)
2016-09-28T01:13:10Z sjl joined #sbcl
2016-09-28T01:22:47Z em1l joined #sbcl
2016-09-28T01:26:39Z em1l_ quit (Ping timeout: 276 seconds)
2016-09-28T01:27:49Z pootler joined #sbcl
2016-09-28T01:32:54Z pootler quit (Ping timeout: 264 seconds)
2016-09-28T01:46:53Z attila_lendvai quit (Ping timeout: 252 seconds)
2016-09-28T01:50:52Z rumbler31 joined #sbcl
2016-09-28T01:53:21Z nyef`: Preliminary result for HPPA is that the failures that I was noticing earlier were from my local changes.
2016-09-28T01:55:31Z rumbler31 quit (Ping timeout: 272 seconds)
2016-09-28T02:07:58Z nyef`: Yeah, okay, HPPA failures were my local changes.
2016-09-28T02:08:04Z nyef`: Still tracking down the ALPHA thing.
2016-09-28T02:12:48Z sjl quit (Read error: Connection reset by peer)
2016-09-28T02:28:32Z pootler joined #sbcl
2016-09-28T02:33:31Z pootler quit (Ping timeout: 272 seconds)
2016-09-28T02:38:48Z stassats quit (Ping timeout: 276 seconds)
2016-09-28T02:49:57Z rpg joined #sbcl
2016-09-28T03:29:18Z pootler joined #sbcl
2016-09-28T03:33:45Z pootler quit (Ping timeout: 244 seconds)
2016-09-28T03:39:09Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2016-09-28T04:19:07Z nyef`: Bisection still isn't complete, but I'm thinking that this "change layout bitmap" thing is going to be the culprit.
2016-09-28T04:21:23Z nyef`: There are a couple of places in the runtime where it casts to a 32-bit signed integer to initialize a 64-bit signed integer, if I'm reading this right.
2016-09-28T04:30:03Z pootler joined #sbcl
2016-09-28T04:35:07Z pootler quit (Ping timeout: 272 seconds)
2016-09-28T04:51:15Z karswell quit (Ping timeout: 244 seconds)
2016-09-28T05:28:20Z nyef`: ... And the "obvious" change is one that stassats put in for win32-x86-64 already. Right.
2016-09-28T05:28:36Z nyef`: Seeing if that cleared up the damage.
2016-09-28T05:30:50Z pootler joined #sbcl
2016-09-28T05:35:36Z pootler quit (Ping timeout: 276 seconds)
2016-09-28T05:43:03Z slyrus quit (Ping timeout: 265 seconds)
2016-09-28T06:10:42Z scymtym quit (Ping timeout: 264 seconds)
2016-09-28T06:30:55Z salva joined #sbcl
2016-09-28T06:31:41Z pootler joined #sbcl
2016-09-28T06:32:04Z _mjl joined #sbcl
2016-09-28T06:32:57Z gingerale joined #sbcl
2016-09-28T06:35:41Z pootler quit (Ping timeout: 240 seconds)
2016-09-28T07:07:54Z ASau quit (Ping timeout: 276 seconds)
2016-09-28T07:26:19Z scymtym joined #sbcl
2016-09-28T07:32:19Z pootler joined #sbcl
2016-09-28T07:36:35Z pootler quit (Ping timeout: 244 seconds)
2016-09-28T07:44:27Z shka joined #sbcl
2016-09-28T07:48:15Z pootler joined #sbcl
2016-09-28T07:56:52Z Bike quit (Quit: one)
2016-09-28T09:15:47Z pkhuong_ quit (Ping timeout: 244 seconds)
2016-09-28T10:10:46Z DGASAU` is now known as DGASAU
2016-09-28T10:36:56Z attila_lendvai joined #sbcl
2016-09-28T11:35:35Z sjl joined #sbcl
2016-09-28T11:44:44Z attila_lendvai quit (Ping timeout: 252 seconds)
2016-09-28T11:45:10Z attila_lendvai joined #sbcl
2016-09-28T12:00:27Z DeadTrickster joined #sbcl
2016-09-28T12:00:28Z ktt9 joined #sbcl
2016-09-28T12:00:35Z rpg joined #sbcl
2016-09-28T12:04:41Z rpg quit (Ping timeout: 240 seconds)
2016-09-28T12:04:42Z madbub joined #sbcl
2016-09-28T12:15:11Z flip214: shouldn't (TRACE) output have such sane defaults as to *not* run into heap-exhausted?
2016-09-28T12:25:34Z stassats joined #sbcl
2016-09-28T12:27:20Z sjl quit (Read error: Connection reset by peer)
2016-09-28T12:32:53Z scymtym: maybe TRACE should have something like SB-EXT:*DEBUG-PRINT-VARIABLE-ALIST*
2016-09-28T12:38:34Z ktt9` joined #sbcl
2016-09-28T12:40:25Z nyef`: Okay, something still wrong on ALPHA, introduced with that "change layout bitmap" thing.
2016-09-28T12:41:13Z stassats: did you update? because i had to fix it already
2016-09-28T12:41:46Z nyef`: Current commit is 0aaa6417437d4ea844aad587ac466a5f020c23ad
2016-09-28T12:41:55Z nyef`: "Fix comment error".
2016-09-28T12:42:10Z stassats: that's a yes then
2016-09-28T12:42:39Z ktt9 quit (Ping timeout: 276 seconds)
2016-09-28T12:43:55Z stassats: how does it manifest?
2016-09-28T12:44:20Z nyef`: fatal error encountered in SBCL pid 20170: hash_table table!=this table 3281120f
2016-09-28T12:45:27Z nyef`: While building contribs.
2016-09-28T12:48:06Z stassats: why does it print only one table?
2016-09-28T12:48:47Z nyef`: Don't know.
2016-09-28T12:49:10Z stassats: that was a rhetorical question
2016-09-28T12:50:26Z nyef`: Haven't had coffee yet this morning. (-:
2016-09-28T12:57:29Z nyef`: Geh! sword_t is a pointer-sized-integer.
2016-09-28T12:57:35Z nyef`: That'd do it.
2016-09-28T13:00:42Z stassats: but it should be enough, shouldn't it?
2016-09-28T13:00:54Z stassats: or is the problem with sign extension?
2016-09-28T13:01:02Z nyef`: The sign extension.
2016-09-28T13:01:24Z nyef`: We cast an unsigned, 32-bit quantity to a signed 64-bit quantity, and then expect that it sign extended.
2016-09-28T13:01:31Z stassats: (digits :rest-p t :c-type #!-alpha "sword_t" #!+alpha "u32")
2016-09-28T13:01:35Z stassats: ...
2016-09-28T13:01:57Z nyef`: Yeah, buried treasure.
2016-09-28T13:02:44Z stassats: so why is one signed and another unsigned?
2016-09-28T13:02:51Z stassats: i suppose it doesn't really matter
2016-09-28T13:05:48Z nyef`: Okay, why don't we just make sword_t 32 bits on alpha?
2016-09-28T13:06:06Z stassats: you know what would happen
2016-09-28T13:06:32Z nyef`: Yeah, the core header parsing would break. But other than that, it looks plausible.
2016-09-28T13:07:26Z stassats: just don't fall on your sword_t
2016-09-28T13:07:33Z nyef`: Heh!
2016-09-28T13:08:44Z nyef`: Well, it's this, or we take the argument that "the alpha backend doesn't work anyway at this point, so let's not bother keeping it from getting any worse while we're in code-freeze for fixing regressions".
2016-09-28T13:09:55Z stassats: anything that only touches alpha is fair game at any stage in the release cycle
2016-09-28T13:09:57Z stassats: or hppa
2016-09-28T13:10:08Z stassats: (because nobody uses those anyway)
2016-09-28T13:10:42Z stassats: hppa should probably be reinstated on http://sbcl.org/platform-table.html
2016-09-28T13:10:54Z stassats: can you make a build or something?
2016-09-28T13:11:08Z stassats: "Historically SBCL also ran on HP PA-RISC processors under Linux and on Alpha processors under Tru64, but this support has not been maintained."
2016-09-28T13:11:16Z stassats: i guess that's not entirely true for the moment
2016-09-28T13:12:07Z nyef`: Well, the PA-RISC stuff is still somewhat broken.
2016-09-28T13:12:31Z stassats: so, you're telling me we have a single port that's not "somewhat broken"?
2016-09-28T13:12:47Z pipping: speaking on that page and the binaries...
2016-09-28T13:13:01Z pipping: the link http://prdownloads.sourceforge.net/sbcl/sbcl-1.2.7-x86-64-openbsd56-binary.tar.bz2 leads to nowhere
2016-09-28T13:13:10Z nyef`: Okay, the PA-RISC stuff is still rather more broken than the other backends.
2016-09-28T13:15:05Z stassats: openbsd is a bitch to make binaries for
2016-09-28T13:15:16Z stassats: they constantly brake ABI
2016-09-28T13:15:20Z stassats: or break
2016-09-28T13:15:47Z pipping: so... you basically need a new binary for each OS release? uh :/
2016-09-28T13:16:26Z stassats: something like that, and it'd be a good case that it'll be backwards compatible
2016-09-28T13:16:29Z stassats: which is not always the case
2016-09-28T13:17:03Z stassats: so, i have no idea for which version i actually built sbcl-1.2.7-x86-64-openbsd-binary.tar.bz2
2016-09-28T13:17:54Z stassats: well, probably 56 is right
2016-09-28T13:20:16Z stassats: pipping: i renamed the file
2016-09-28T13:20:47Z pipping: stassats: great, thanks
2016-09-28T13:22:34Z stassats: i don't want to make another "it'd be a good idea to change" statement but then never actually do that, but the download page should be extended
2016-09-28T13:22:45Z stassats: to easily list past binaries for a target
2016-09-28T13:22:47Z nyef`: Did I say that it'd break core file parsing? Bloody hell, we should have done this ages ago!
2016-09-28T13:23:06Z stassats: and list what they have for features
2016-09-28T13:23:13Z stassats: (and i'll never do that)
2016-09-28T13:23:42Z rumbler31 joined #sbcl
2016-09-28T13:23:52Z stassats: so, where are all the non technical contributors to SBCL?
2016-09-28T13:24:53Z flip214: if you find some, tell me where. I'd need a dozen, too.
2016-09-28T13:26:40Z nyef`: Okay, building with the new change.
2016-09-28T13:29:18Z pipping: If there's any value in running a few simple enough scripts and uploading the results, I could do that for (amd64) openbsd 5.9/6.0 or freebsd 10/11
2016-09-28T13:31:10Z stassats: no, i want the download page to be changed
2016-09-28T13:40:59Z pipping looks at https://sourceforge.net/projects/sbcl/files/sbcl/
2016-09-28T13:41:29Z pipping notices that the common format is just the version yet there's also sbcl-1.0.58, sbcl-1.0.27, sbcl-0.6.9
2016-09-28T13:41:44Z stassats: don't look there
2016-09-28T13:42:05Z pipping: is there a list of all binaries?
2016-09-28T13:44:14Z stassats: can be gotten through ssh
2016-09-28T13:48:02Z stassats: pipping: http://stassats.users.sourceforge.net/binaries.txt
2016-09-28T13:48:03Z stassats: for example
2016-09-28T13:51:07Z Xof joined #sbcl
2016-09-28T13:52:43Z nyef`: Good morning Xof.
2016-09-28T13:52:47Z Xof: hi
2016-09-28T13:53:42Z shka quit (Quit: Konversation terminated!)
2016-09-28T13:53:51Z nyef`: I'm building a possible fix for the regression on alpha now.
2016-09-28T14:00:59Z slyrus joined #sbcl
2016-09-28T14:05:23Z Xof: excellent
2016-09-28T14:05:36Z Xof: You come with solutions, not just problems :-)
2016-09-28T14:06:46Z nyef`: Last night, all I had was the problem.
2016-09-28T14:10:59Z Xof: I'm glad I didn't come then, then!
2016-09-28T14:12:15Z rpg joined #sbcl
2016-09-28T14:13:13Z cromachina quit (Read error: Connection reset by peer)
2016-09-28T14:14:38Z sjl joined #sbcl
2016-09-28T14:15:16Z whiteline quit (Read error: Connection reset by peer)
2016-09-28T14:18:48Z whiteline joined #sbcl
2016-09-28T14:31:03Z nyef`: Hrm. Okay, situation vastly improved, but there's now an issue with get-universal-time. /-:
2016-09-28T14:32:18Z reb joined #sbcl
2016-09-28T14:33:50Z reb: I just ran SBCL tests on Linux x86-64 and got an unexpected failure in the LURKING-THREADS test.  Here's the error: http://paste.lisp.org/+70JN
2016-09-28T14:35:41Z slyrus quit (Ping timeout: 240 seconds)
2016-09-28T14:40:21Z reb quit (Remote host closed the connection)
2016-09-28T14:40:34Z reb joined #sbcl
2016-09-28T14:42:29Z scymtym quit (Ping timeout: 272 seconds)
2016-09-28T14:43:25Z pipping: yikes. all the inconsistencies in the file names
2016-09-28T14:43:53Z pipping: sbcl-0.9.13-powerpc-darwin-binary.tar.bz2
2016-09-28T14:43:57Z pipping: sbcl-0.9.13-Mac-Intel-binary.tar.bz2
2016-09-28T14:44:01Z pipping: ...
2016-09-28T14:46:40Z nyef`: Hrm. Rebuilding just the contribs, and I get a different fault. I think I'll chalk this one up to the DX SAP thing, commit the fix that I have on hand, and call it a day.
2016-09-28T14:47:52Z stassats: reb: it's not that unexpected
2016-09-28T14:48:21Z reb: stassats: I see now looking at the test that it could be flaky.
2016-09-28T14:50:05Z _mjl quit (Ping timeout: 272 seconds)
2016-09-28T14:51:44Z nyef`: Fix pushed.
2016-09-28T14:51:57Z nyef`: So, alpha is now no-less-broken than it was a month ago.
2016-09-28T14:52:31Z nyef`: And there is a small pile of cleanups that we can now do.
2016-09-28T14:53:20Z Xof: pipping: don't worry too much about 0.9.13
2016-09-28T14:53:54Z nyef`: I don't think I'm running anything older than 1.0.23 as a build host, FWIW.
2016-09-28T14:54:21Z nyef`: Admittedly, I need a patch to the current tree in order to build on 1.0.23, but that's another matter.
2016-09-28T14:56:00Z nyef`: (Have to define FDEFN-P in src/code/cross-misc.lisp, in order to avoid an unknown function reference from globaldb, probably because of the definition of FDEFN as (SATISFIES FDEFN-P).)
2016-09-28T14:56:49Z nyef`: Xof: Actually, since you're here, and I'm thinking about it, is this "fix" something that I should commit? Probably not for this release, but early next release?
2016-09-28T14:58:02Z Xof: probably, given that some of our release binaries are that old
2016-09-28T14:58:50Z nyef`: Okay, this month or next?
2016-09-28T15:00:27Z Xof: next; people can always cross-compile if necessary
2016-09-28T15:01:04Z nyef`: That's what I figured. It's not a recent regression, anyway.
2016-09-28T15:15:18Z pootler quit
2016-09-28T15:27:55Z whiteline quit (Remote host closed the connection)
2016-09-28T15:28:25Z whiteline joined #sbcl
2016-09-28T15:33:31Z edgar-rft quit (Quit: edgar-rft)
2016-09-28T15:52:40Z slyrus joined #sbcl
2016-09-28T15:58:42Z Bike joined #sbcl
2016-09-28T16:12:48Z pipping: stassats: was something like http://139.162.161.212/asdf/output.html what you have in mind? (i've spent zero thought so far on cleaning up duplicates (ppc vs powerpc) and style)
2016-09-28T16:13:08Z stassats: well, not really
2016-09-28T16:13:17Z stassats: that's not really possible to read
2016-09-28T16:15:32Z scymtym joined #sbcl
2016-09-28T16:21:36Z ktt9`: >where are all the non technical contributors to SBCL?
2016-09-28T16:22:00Z ktt9`: I volunteer!
2016-09-28T16:23:00Z joshe: I'm not a huge fan of the existing platform download table myself
2016-09-28T16:23:10Z ktt9`: But of course I also want to work on some technical problems too.
2016-09-28T16:23:32Z joshe: openbsd doesn't fit well in there as each release is ABI incompatible
2016-09-28T16:23:53Z joshe: it would be nice to have download links for the current and previous releases
2016-09-28T16:24:17Z joshe: but then I'm not even building openbsd release binaries anymore so it's just talk
2016-09-28T16:25:05Z joshe: I just tell people to pkg_add sbcl and use that to build the version they want
2016-09-28T16:47:08Z pipping: stassats: I agree. What would you like to see instead then?
2016-09-28T16:47:41Z Shinmera: Maybe something that has just a row for each implementation listing the latest version with a drop-down/roll-out menu of all past versions?
2016-09-28T16:48:43Z stassats: pipping: bah, i didn't want to lift a finger to accomplish this
2016-09-28T16:49:07Z stassats: so, something readable and updatable for future releases, etc.
2016-09-28T16:49:25Z stassats: have to figure that out yourself
2016-09-28T16:52:26Z pipping will tabe a stab at what Shinmera said
2016-09-28T16:52:37Z pipping: take, even
2016-09-28T16:56:10Z rpg_ joined #sbcl
2016-09-28T16:58:43Z rpg quit (Ping timeout: 244 seconds)
2016-09-28T17:03:05Z slyrus quit (Ping timeout: 272 seconds)
2016-09-28T17:23:34Z karswell joined #sbcl
2016-09-28T17:29:41Z sjl quit (Ping timeout: 272 seconds)
2016-09-28T17:33:23Z foom2 is now known as foom
2016-09-28T17:41:14Z nyef`: Did... I not push the alpha fix?
2016-09-28T17:42:05Z nyef`: No, that definitely sent. WTF?
2016-09-28T17:42:20Z nyef`: Oh, I see. There's another commit on top.
2016-09-28T17:57:09Z Cthulhux quit (Changing host)
2016-09-28T17:57:09Z Cthulhux joined #sbcl
2016-09-28T18:46:19Z _mjl joined #sbcl
2016-09-28T18:52:03Z Quadresce joined #sbcl
2016-09-28T18:58:07Z rpg_ is now known as rpg
2016-09-28T19:01:53Z pipping: Shinmera: think you can http://139.162.161.212/asdf/output.html the final polish? html/css is not my area of expertise. as you can see, the hovering is broken and ugly
2016-09-28T19:01:59Z pipping: +give
2016-09-28T19:02:17Z pipping: stassats: (http://139.162.161.212/asdf/output.html is now probably closer to what you had in mind, right?)
2016-09-28T19:02:40Z stassats: that looks better
2016-09-28T19:03:18Z pipping: as I said, the looks are considerably worse than what you'd expect from a project as mature as sbcl but I won't be able to solve that.
2016-09-28T19:04:30Z Shinmera: pipping: I had something else in mind. If you want I can give it a shot myself tomorrow. Don't have the time tonight
2016-09-28T19:05:33Z pipping: Shinmera: I wouldn't be surprised if you could put something together that's 10 times better in 1/10 the time.
2016-09-28T19:05:46Z slyrus joined #sbcl
2016-09-28T19:06:15Z pipping: Shinmera: What I did might still save you some time though
2016-09-28T19:06:25Z pipping: Shinmera: here's the data I worked with: http://dpaste.com/3V1TM1S.txt
2016-09-28T19:06:36Z Shinmera: Thanks!
2016-09-28T19:07:03Z pipping: Shinmera: and here's the code: http://dpaste.com/2S8J715
2016-09-28T19:07:33Z stassats: it'd be better if it weren't a hover but after a click
2016-09-28T19:08:09Z pipping: stassats: I tried that but gave up after a while in frustration
2016-09-28T19:08:25Z stassats: and the table splits in half revealing the information in the middle
2016-09-28T19:10:56Z edgar-rft joined #sbcl
2016-09-28T19:28:51Z ktt9 joined #sbcl
2016-09-28T19:29:08Z ktt9: Hi all. Do we accept pull-requests via github?
2016-09-28T19:29:23Z stassats: ktt9: nope
2016-09-28T19:40:15Z ktt9: Hm... I'm sure that I'm not the first who asks, but whatever. Why not?
2016-09-28T19:41:08Z stassats: should we accept contributions from all those newfangled web sites now?
2016-09-28T19:41:19Z stassats: there are established channels already
2016-09-28T19:42:44Z Quadresce: The dedicated review workflow of something like GitHub is of course nice. But, as projects like Linux shows, plaintext email seems to work too...
2016-09-28T19:43:59Z ktt9: Well, I think the more channels the better, actually. Especially when there's demand. Also Linux does accept contributions via Github.
2016-09-28T19:44:19Z Quadresce: ktt9: More channels is definitely not better. One place to make contributions is better.
2016-09-28T19:44:33Z Quadresce: ktt9: I said Linux shows that email can work.
2016-09-28T19:45:56Z ktt9: Oh.
2016-09-28T19:46:09Z PuercoPop: also git format-patch is not much different than a PR. And the steps are  documented in the HACKING file of SBCL
2016-09-28T19:46:20Z ktt9: Why, of course it can work. But the question is, is it most effective way.
2016-09-28T19:47:03Z jackdaniel: I find applying patches from mail less troublesome than opening firefox entering github and clicking
2016-09-28T19:47:37Z Quadresce: ktt9: I mean, I think the whole SourceForge/launchpad combo is a mess, but I understand that changing the direction of the inertia can be difficult or not completely beneficial.
2016-09-28T19:48:24Z Quadresce: jackdaniel, email patches are very coarse-grained. It's sort of wholesale acceptance/rejection. It is difficult to do line-by-line review and commenting.
2016-09-28T19:48:56Z PuercoPop: If someone finds it to be a mayor inconvenience, they can always set up a bot that takes the github PR's and send them as a patch to the mailing list.
2016-09-28T19:49:34Z jackdaniel: Quadresce: yes, I see benefits of PR too, just saying that merging patches sent by e-mail is more efficient (if we measure in keystrokes / mouseclicks)
2016-09-28T19:50:06Z nyef`: Is that the metric that we should be using, though?
2016-09-28T19:50:08Z Quadresce: I find that optimizing review and communication to be better than optimizing clicks-to-merge. :)
2016-09-28T19:50:24Z jackdaniel: right
2016-09-28T19:51:14Z nyef`: Using github, right now, is absolutely not optimizing for review and communication: Basically none of the SBCL maintainers watch github for this stuff, so there's essentially no review or communication.
2016-09-28T19:52:02Z ktt9: Sigh.
2016-09-28T19:52:12Z ktt9: Understood.
2016-09-28T19:52:36Z ktt9: Than we at least should add README.md to explain the way we accept contributions.
2016-09-28T19:52:55Z Quadresce: GitHub probably doesn't work on the SBCL devs' 2400 baud modems ;)
2016-09-28T19:53:12Z ktt9: So there will not be unansvered PR's pending since forever on Github. That's bad for image.
2016-09-28T19:53:18Z stassats: yes, those pesky SBCL developers
2016-09-28T19:53:21Z ktt9: And in general.
2016-09-28T19:53:26Z Quadresce: stassats, <3
2016-09-28T19:53:39Z stassats: ktt9: people don't read things
2016-09-28T19:53:47Z ktt9: Why, they do.
2016-09-28T19:53:48Z PuercoPop: ktt9: there is a HACKING file already
2016-09-28T19:53:52Z stassats: case in point, README actually does direct you to proper contribution channels
2016-09-28T19:54:00Z nyef` points out that SBCL already has a README, and it points to the HACKING file for contributions...
2016-09-28T19:54:21Z stassats: ktt9: but you didn't read it, so...
2016-09-28T19:54:39Z ktt9: But I did D:
2016-09-28T19:55:01Z nyef`: You know, I'm still somewhat annoyed that SBCL was patched to work around a longstanding known bug in Wine.
2016-09-28T19:55:20Z stassats: nyef`: was it?
2016-09-28T19:55:34Z nyef`: Yes. And I'll give good odds that the Wine bug STILL isn't fixed.
2016-09-28T19:57:54Z ktt9: Was it done by some outsider contributor?
2016-09-28T19:57:54Z ktt9: And why on earth SBCL needed patch for Wine bug?
2016-09-28T19:58:10Z ktt9: Also my point is that at least README should say in large font that nobody reviews PRs and if one wants to contribute, they must send patches somewhere else.
2016-09-28T19:58:29Z ktt9: That's what you expect from project residing on GH.
2016-09-28T19:58:41Z nyef`: Umm... But SBCL resides on sourceforge.
2016-09-28T19:59:11Z ktt9: n_n
2016-09-28T19:59:21Z pipping: if you reside on gh long enough, gh resides on you.
2016-09-28T19:59:35Z stassats: thank you for proving my "people don't read things" again
2016-09-28T19:59:47Z ktt9: I meant the mirror.
2016-09-28T19:59:48Z stassats: since github says "Mirror"
2016-09-28T20:00:06Z ktt9: Does that imply ignoring PRs?
2016-09-28T20:00:21Z stassats: i'm so sick of this "github this or github that" that i'm inclined to just remove the github mirror
2016-09-28T20:00:29Z stassats: is that your goal?
2016-09-28T20:00:33Z ktt9: No.
2016-09-28T20:00:47Z ktt9: Sorry.
2016-09-28T20:01:16Z ktt9: I didn't mean to anger you.
2016-09-28T20:01:28Z stassats: well, you didn't anger me
2016-09-28T20:01:38Z stassats: github did, if anything
2016-09-28T20:02:14Z nyef`: "This service is provided as an inconvenience for our developers."
2016-09-28T20:02:40Z ktt9: It's just that Github is my home, and I wanted mirror of my favorite compiler of my favorite language to be more friendly towards all GH users.
2016-09-28T20:02:42Z ktt9: That's all.
2016-09-28T20:05:29Z _mjl quit (Ping timeout: 272 seconds)
2016-09-28T20:06:28Z Quadresce: ktt9, Maybe if you begin to make lots of great contributions to SBCL, there might be room for the other devs to compromise their workflow to optimize for your development workflow for it. :)
2016-09-28T20:09:31Z pkhuong joined #sbcl
2016-09-28T20:09:54Z pkhuong is now known as Guest3864
2016-09-28T20:10:52Z ktt9: Quadresce: Does it work like this?
2016-09-28T20:12:36Z Quadresce: ktt9: Well, there are people---the SBCL devs---who spend a lot of time and effort on their SBCL work. And ultimately they move the project forward. It would be nonsense to attenuate their productivity so some random internet folks can send a couple typo fixes here and there, maybe once or twice during the course of said random people's involvement with SBCL.
2016-09-28T20:14:16Z Quadresce: However, if someone is added to this list of folks who help move the project forward, and if their productivity would increase by modifying the overall developer workflow, then I'm sure the group of developers would be willing to listen to such requests. But until then, it's not worth it.
2016-09-28T20:15:51Z nyef`: Ultimately, you'd probably need someone with access to both workflows to decide that having a bridge between the two workflows is worth having /and to build and maintain that bridge/.
2016-09-28T20:20:14Z ktt9: Okay, I understand.
2016-09-28T20:22:28Z ktt9: I apologize again for causing somewhat heated discussion.
2016-09-28T20:22:45Z ktt9: It comes from my lack of experience, as it seems.
2016-09-28T20:23:35Z ktt9: So, the bug tracker is Launchpad and main repo is hosted on SF, right?
2016-09-28T20:23:42Z Quadresce: Yes.
2016-09-28T20:23:50Z Quadresce: (With that said, we would be happy and excited to see some patches to improve SBCL!)
2016-09-28T20:25:13Z ktt9: How should I (or should at all) notificate others of what bug I'm currently working on?
2016-09-28T20:26:51Z Quadresce: Find a bug, check if it has been assigned. If not, maybe write a comment on LP that you're going to work on it and you have an approach, and just solve it.
2016-09-28T20:28:50Z ktt9: Understood.
2016-09-28T20:29:23Z nyef`: Or even just solve it, and attach a patch to the issue.
2016-09-28T20:29:54Z ktt9: So clashes are unlikely.
2016-09-28T20:31:05Z nyef`: At this point, probably unlikely, yes.
2016-09-28T20:32:05Z Quadresce: ktt9: I wouldn
2016-09-28T20:32:09Z Quadresce: t think about it too hard. :)
2016-09-28T20:37:00Z ktt9: Okay.
2016-09-28T20:38:43Z edgar-rft quit (Quit: edgar-rft)
2016-09-28T20:54:22Z gingerale quit (Remote host closed the connection)
2016-09-28T21:03:02Z madbub quit (Remote host closed the connection)
2016-09-28T21:07:18Z ASau joined #sbcl
2016-09-28T21:13:16Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-28T21:17:30Z fe[nl]ix quit (Remote host closed the connection)
2016-09-28T21:17:30Z Blkt quit (Read error: Connection reset by peer)
2016-09-28T21:17:45Z Blkt joined #sbcl
2016-09-28T21:17:45Z fe[nl]ix joined #sbcl
2016-09-28T21:37:45Z Quadresce joined #sbcl
2016-09-28T21:56:21Z rumbler31 quit (Remote host closed the connection)
2016-09-28T22:58:23Z rumbler31 joined #sbcl
2016-09-28T23:02:41Z rumbler31 quit (Ping timeout: 240 seconds)
2016-09-28T23:03:18Z edgar-rft joined #sbcl
2016-09-28T23:31:05Z cojy joined #sbcl
2016-09-28T23:32:23Z Arathnim joined #sbcl
2016-09-28T23:34:06Z slyrus quit (Ping timeout: 264 seconds)
2016-09-28T23:42:16Z Arathnim: How does sbcl deal with compiling block/return-from? Particularly in cases like (block a (foo (lambda () (return-from a 'bar)))). I tried examining the disassemble output, but it's three screens of code about threading.
2016-09-28T23:43:36Z nyef`: Hrm. This is the closure case of non-local exit.
2016-09-28T23:44:50Z Arathnim: I've been working on a toy lisp compiler, but I'm having problems with lexical closures and this kind of non-local exit.
2016-09-28T23:45:33Z Quadresce: Arathnim, those have dynamic extent
2016-09-28T23:48:50Z cromachina joined #sbcl
2016-09-28T23:48:56Z Arathnim: It's allocated at runtime, or in the static space before the stack?
2016-09-28T23:49:46Z nyef`: So, the exit itself is only valid while the stack for BLOCK A still exists.
2016-09-28T23:50:24Z nyef`: That's true for both EXIT-EXTENT:MINIMAL (the actual spec) and the other EXIT-EXTENT proposal (the name of which escapes me).
2016-09-28T23:50:45Z Arathnim: Well, yes, but where is the information stored that allows it to traverse the stack backward to the context of the block?
2016-09-28T23:51:28Z nyef`: Roughly, the closure closes over the stack frame pointer of BLOCK A, and the "current catch block" for BLOCK A.
2016-09-28T23:52:37Z nyef`: Umm... And there's a landing pad of some sort in BLOCK A that does an unbind-to-here and then returns.
2016-09-28T23:54:01Z Arathnim: That's what I thought. Looks like I need to read more on compiling closures.
2016-09-28T23:58:12Z Arathnim: Thanks for the help.
2016-09-29T00:02:18Z sjl joined #sbcl
2016-09-29T00:06:59Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-29T00:09:38Z stassats: more interesting is how not to blow the lid off when the closure is called out of extent
2016-09-29T00:10:02Z nyef`: Right, I think that that's where we currently fall down.
2016-09-29T00:10:41Z myrkraverk quit (Ping timeout: 240 seconds)
2016-09-29T00:11:03Z stassats: it's the first bug in the tracker
2016-09-29T00:11:20Z stassats: actually, might be interesting to poke around it
2016-09-29T00:12:07Z nyef`: It's on my list to look at "at some point".
2016-09-29T00:12:17Z nyef`: It's one of the last two core compiler bugs, IIRC.
2016-09-29T00:17:42Z nyef`: I've probably touched, if not made major changes, to just about everything involved in that NLX issue.
2016-09-29T00:17:51Z Quadresce joined #sbcl
2016-09-29T00:21:14Z stassats: ok, i see, we don't zero out the closed over uwp block
2016-09-29T00:24:11Z nyef`: Hrm. Maybe three core compiler bugs. There's that check-consistency one as well.
2016-09-29T00:24:30Z stassats: check-consistency is easy
2016-09-29T00:24:42Z stassats: there's a proposed solution already
2016-09-29T00:24:55Z stassats: just do a cast instead of changing derived types
2016-09-29T00:25:20Z stassats: and i don't think it causes any problems
2016-09-29T00:25:23Z nyef`: Yeah, we just haven't done it yet. (-:
2016-09-29T00:26:33Z nyef`: Right, hasn't caused problems because the issue is caused by dataflow analysis ignoring what amounts to a dead node, but the node not getting deleted before the inlining, and then the inlining causes this type conflict...
2016-09-29T00:26:47Z nyef`: ... but then the conflict is due to dead code, which then gets snapped out.
2016-09-29T00:27:19Z nyef`: I've been fixing too much stuff with CAST nodes, though, you know?
2016-09-29T00:27:44Z nyef`: What was that other compiler bug that you were looking at?
2016-09-29T00:28:06Z nyef`: Oh, right, "XEPs, let conversion, and inlining".
2016-09-29T00:28:49Z stassats: yeah, that involves copying IR
2016-09-29T00:28:57Z stassats: i made some progress but then i kinda got bored by it
2016-09-29T00:31:38Z nyef`: I might take a look at it at some point.
2016-09-29T00:31:52Z stassats: don't
2016-09-29T00:31:57Z nyef`: Oh?
2016-09-29T00:31:57Z stassats: i intend on finishing it
2016-09-29T00:32:00Z nyef`: Okay then.
2016-09-29T00:32:07Z stassats: it's just grunt work
2016-09-29T00:32:56Z stassats: basically looking at ir1 conversion routines and replicate that, but with ir1 as a sourc
2016-09-29T00:32:56Z stassats: e
2016-09-29T00:40:14Z edgar-rft quit (Quit: edgar-rft)
2016-09-29T00:42:20Z nyef`: Do we not have the "original" S-expression form?
2016-09-29T00:42:33Z stassats: we do, but it's of no use
2016-09-29T00:42:37Z nyef`: Why not?
2016-09-29T00:42:50Z nyef`: Did the lexenv get destroyed or something?
2016-09-29T00:42:54Z stassats: yes
2016-09-29T00:42:57Z nyef`: Ah.
2016-09-29T00:43:30Z stassats: it gets out of sync
2016-09-29T00:44:18Z nyef`: Damn.
2016-09-29T00:44:22Z nyef`: Fair enough, then.
2016-09-29T00:45:57Z stassats: ok, for some reason there's not enough %lexical-exit-breakup
2016-09-29T00:48:29Z stassats: but for a working function there are two %lexical-exit-breakups
2016-09-29T01:04:13Z cojy left #sbcl
2016-09-29T01:10:07Z stassats: but that's my first real foray into nlx
2016-09-29T01:10:20Z stassats: so i don't feel local
2016-09-29T01:14:20Z attila_lendvai quit (Ping timeout: 252 seconds)
2016-09-29T01:19:28Z stassats: huh, "recursion in known function definition"
2016-09-29T01:19:31Z stassats: that's weird
2016-09-29T01:20:56Z stassats: doing (multiple-value-call #'print (lambda ()))
2016-09-29T01:21:10Z stassats: or (multiple-value-call #'print 1)
2016-09-29T01:22:31Z em1l_ joined #sbcl
2016-09-29T01:25:12Z nyef`: I'm... not seeing that effect with M-V-C?
2016-09-29T01:25:34Z em1l quit (Ping timeout: 244 seconds)
2016-09-29T01:25:42Z stassats: DEBUG
2016-09-29T01:25:58Z stassats: which makes it even weirder
2016-09-29T01:28:27Z stassats: (locally (declare (optimize (sb-c::let-conversion 0))) (multiple-value-call #'print 1))
2016-09-29T01:28:29Z stassats: to be exact
2016-09-29T01:28:38Z nyef`: Ah.
2016-09-29T01:28:54Z stassats: probably something involve the XEP
2016-09-29T01:29:32Z nyef`: Disabling let-conversion worries me, really.
2016-09-29T01:31:50Z stassats: ok, just (block nil (throw t (lambda () (return)))) is clearer
2016-09-29T01:31:50Z stassats:  
2016-09-29T01:32:24Z stassats: it looks like it doesn't get to the cleanup because it gets thrown
2016-09-29T01:32:48Z nyef`: ... Riiight. Yeah, I can see that now.
2016-09-29T01:33:01Z nyef`: The cleanup needs to be converted to an "implicit" UWP.
2016-09-29T01:33:38Z stassats: and a good way to test this snippet is (sb-kernel:%closure-index-ref (catch t (foo)) 0)
2016-09-29T01:33:59Z nyef`: Or something needs to be placed on the binding stack that UNBIND-TO-HERE knows how to deal with and trash the closed-over exit information for.
2016-09-29T01:36:56Z nyef`: Good catch, though. (-:
2016-09-29T01:37:11Z stassats: and it's not unique to throw
2016-09-29T01:37:38Z stassats: (sb-kernel:%closure-index-ref (car (simple-condition-format-arguments (nth-value 1 (ignore-errors (block nil (error "~a" (lambda () (return)))))))) 0)
2016-09-29T01:37:42Z stassats: => #<value cell 155538648 {10052E69FF}>
2016-09-29T01:41:41Z nyef`: Right, any "non-local" NLX can do it. That's not surprising.
2016-09-29T01:42:09Z nyef`: Sounds like you can at least write the test cases now.
2016-09-29T01:43:38Z stassats: there's already a uwp block in there somewhere
2016-09-29T01:43:53Z stassats: or maybe not
2016-09-29T01:44:25Z stassats: yeah, there is
2016-09-29T01:50:49Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-29T01:55:10Z stassats: the unwind is for when the block is returned from
2016-09-29T01:55:31Z stassats: so for (defun bar (x) (funcall x x)) and (block nil (bar (lambda (x) (return x))))
2016-09-29T01:55:34Z stassats: it gets zeroed
2016-09-29T01:55:51Z stassats: but (defun bar (x) (throw t x)) does not
2016-09-29T01:58:21Z stassats: and it's the unwind that gets closed over
2016-09-29T02:11:02Z stassats: i think somehow it can be reused
2016-09-29T02:11:12Z stassats: but i can't quite wrap my head around it
2016-09-29T02:15:36Z stassats: sometimes the can be merged, sometimes they can't
2016-09-29T02:20:58Z stassats: well, for safe code it wouldn't really matter
2016-09-29T02:21:09Z stassats: but i'd prefer an efficient and general solution
2016-09-29T02:26:53Z stassats: but what if it actually used a special variable instead of a closure?
2016-09-29T02:27:20Z stassats: probably will blow out tls indexes
2016-09-29T03:32:09Z stassats quit (Ping timeout: 244 seconds)
2016-09-29T03:45:37Z rumbler31 joined #sbcl
2016-09-29T03:47:12Z karswell quit (Ping timeout: 272 seconds)
2016-09-29T03:47:13Z salva quit (Ping timeout: 272 seconds)
2016-09-29T03:47:21Z Wojciech_K quit (Ping timeout: 240 seconds)
2016-09-29T03:47:45Z karswell joined #sbcl
2016-09-29T03:47:49Z cromachina quit (Ping timeout: 272 seconds)
2016-09-29T03:49:58Z Wojciech_K joined #sbcl
2016-09-29T03:50:21Z rumbler31 quit (Ping timeout: 272 seconds)
2016-09-29T03:59:51Z salva joined #sbcl
2016-09-29T04:41:31Z Quadresce joined #sbcl
2016-09-29T04:51:02Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-29T05:09:58Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2016-09-29T05:13:08Z slyrus joined #sbcl
2016-09-29T05:13:47Z karswell quit (Remote host closed the connection)
2016-09-29T05:13:59Z karswell joined #sbcl
2016-09-29T05:25:19Z _mjl joined #sbcl
2016-09-29T05:39:30Z scymtym quit (Ping timeout: 264 seconds)
2016-09-29T06:06:37Z Quadresce joined #sbcl
2016-09-29T06:08:07Z edgar-rft joined #sbcl
2016-09-29T06:15:48Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-29T06:36:37Z gingerale joined #sbcl
2016-09-29T06:58:11Z ASau` joined #sbcl
2016-09-29T06:59:20Z ASau quit (Ping timeout: 244 seconds)
2016-09-29T07:05:21Z sjl quit (Read error: Connection reset by peer)
2016-09-29T07:33:46Z scymtym joined #sbcl
2016-09-29T07:56:57Z ASau` quit (Remote host closed the connection)
2016-09-29T07:57:22Z gingerale quit (Remote host closed the connection)
2016-09-29T07:57:23Z ASau` joined #sbcl
2016-09-29T08:13:18Z ASau` quit (Ping timeout: 276 seconds)
2016-09-29T08:30:04Z Bike quit (Quit: dreadful eiri)
2016-09-29T08:30:40Z attila_lendvai joined #sbcl
2016-09-29T08:40:29Z ktt9: Hello.
2016-09-29T08:40:41Z ktt9: What's wrong with patch to this bug: https://bugs.launchpad.net/sbcl/+bug/659107 ?
2016-09-29T09:55:31Z scymtym: somebody suggested nominating sbcl for http://osb-alliance.de/osbar to me (in private communication because "[my] email address was the first one in the git log with a de tld")
2016-09-29T09:59:46Z rumbler31 joined #sbcl
2016-09-29T09:59:48Z scymtym: even if we were interested in doing this, i don't think sbcl would be eligible because there is a prerequisite saying "[the project] must contain at least one innovation for businesses or public institutions" (my translation)
2016-09-29T10:00:34Z scymtym: so i lean towards just replying "no thanks", but i wanted to at least bring the issue up here first
2016-09-29T10:02:07Z pipping: is public visibility a problem of sbcl? I would assume everyone who knows common lisp knows sbcl. so the issue would rather be a lack of visibility of common lisp.
2016-09-29T10:03:47Z rumbler31 quit (Ping timeout: 244 seconds)
2016-09-29T10:04:00Z pipping: (publicity is what they offer, right? I see no mention of funding anywhere)
2016-09-29T10:04:23Z scymtym: that's my understanding as well, from skimming the contents
2016-09-29T10:20:40Z cpape joined #sbcl
2016-09-29T10:23:05Z jackdaniel: hrm, providing high quality public domain implementation of the programming language serves both business and public institutions (more as a social innovation though)
2016-09-29T10:23:19Z jackdaniel: just saying :)
2016-09-29T10:34:10Z pipping: jackdaniel: yes. but if you were not even interested in winning the prize that would save you the trouble of debating the chances/effort ratio
2016-09-29T10:36:37Z jackdaniel: I was replaying to scymtym's "so I lean towards just replaying 'no thanks'" in context of 'must contain at least one innovation', not if its worthwhile
2016-09-29T10:42:08Z pipping: if you take a look at the list of the 3 winners from 2015 you'll see a red thread: they have something rather different in mind than SBCL, lisp, or compilers (keywords/quotes: smart home solution, java library to read/write meta data in "your invoice PDFs", providing an overview over when your software licenses expire)
2016-09-29T10:43:53Z pipping: my prediction for the first price this year is: a tool that makes you 10% more efficient at editing spreadsheets.
2016-09-29T10:43:59Z Xof: end-business apps
2016-09-29T10:44:21Z Xof: since we're not in the selling market, publicity has very little value
2016-09-29T10:44:29Z jackdaniel: pipping: I'm still not sure how does it invalidate what I said, but mkay ;)
2016-09-29T11:02:06Z pipping puts on his annoying hat and wonders:
2016-09-29T11:02:45Z pipping: is there something I can do to move https://bugs.launchpad.net/sbcl/+bug/1114689 forward?
2016-09-29T11:45:33Z milanj joined #sbcl
2016-09-29T11:55:22Z madbub joined #sbcl
2016-09-29T12:00:27Z rpg joined #sbcl
2016-09-29T12:05:20Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2016-09-29T12:48:04Z cromachina joined #sbcl
2016-09-29T12:54:00Z stassats joined #sbcl
2016-09-29T12:56:55Z stassats: ktt9: re the patch, --external-format doesn't actually set the external format, but locale
2016-09-29T12:58:01Z stassats: basically, entirely misses the point
2016-09-29T12:59:26Z attila_lendvai quit (Ping timeout: 252 seconds)
2016-09-29T13:15:28Z attila_lendvai joined #sbcl
2016-09-29T13:22:20Z francogrex joined #sbcl
2016-09-29T13:22:38Z francogrex: Hi I run sbcl as an inferior lisp process in emacs
2016-09-29T13:23:10Z cromachina quit (Read error: Connection reset by peer)
2016-09-29T13:23:34Z francogrex: I have two sbcl(s) one 32 and one 64 bits. the one of 32 bits when launched as such it detects and loads the sbclrc. In the same manner sbcl-64 is launched but does not load the sbclrc!
2016-09-29T13:47:54Z slyrus quit (Ping timeout: 264 seconds)
2016-09-29T13:54:57Z ktt9`: stassats, oh, I see.
2016-09-29T13:55:26Z stassats: and why i didn't comment there? well, i don't really care for --external-format to be added
2016-09-29T14:11:22Z rpg joined #sbcl
2016-09-29T14:11:34Z ktt9` quit (Remote host closed the connection)
2016-09-29T14:17:11Z atgreen joined #sbcl
2016-09-29T14:17:26Z ktt9` joined #sbcl
2016-09-29T14:17:28Z atgreen: does anybody have details on potential conflict between SBCL and address space layout randomization (ASLR)?  The most promising references on the web link to gmane archives for details, and those are gone.
2016-09-29T14:18:04Z m00natic` joined #sbcl
2016-09-29T14:18:18Z stassats: it just has to be disabled
2016-09-29T14:18:20Z stassats: which it is
2016-09-29T14:19:57Z francogrex quit (Quit: Lost terminal)
2016-09-29T14:23:35Z atgreen: ? I'm running sbcl in a container.  I don't think it is disabling it.
2016-09-29T14:23:49Z atgreen: so I'm trying to figure out what the risk is.
2016-09-29T14:24:08Z stassats: sbcl just won't work
2016-09-29T14:24:10Z stassats: that's the risk
2016-09-29T14:26:03Z atgreen: well, it's running a little.  can you share a little more insight?  Like sbcl may try to use an address range that the OS has mapped something else into?  or?
2016-09-29T14:26:37Z stassats: sbcl automatically disables randomization when it has to
2016-09-29T14:27:19Z atgreen: I don't think it can do that in a locked down container environment
2016-09-29T14:28:06Z atgreen:     WARNING:
2016-09-29T14:28:06Z atgreen:     Couldn't re-execute SBCL with proper personality flags (/proc isn't mounted? setuid?)
2016-09-29T14:28:07Z atgreen:     Trying to continue anyway.
2016-09-29T14:29:14Z atgreen: My guess is that it's trying to run again with ASLR turned off, but it can't - so it's just going ahead anyways despite the risk
2016-09-29T14:29:18Z atgreen: is that right?
2016-09-29T14:29:29Z stassats: consult your container documentation then
2016-09-29T14:29:52Z rpg quit (Ping timeout: 244 seconds)
2016-09-29T14:32:42Z atgreen: that's not what I came here for.  I was hoping to understand what was happening inside of SBCL
2016-09-29T14:45:13Z scymtym quit (Ping timeout: 272 seconds)
2016-09-29T14:46:21Z stassats: sbcl doesn't work in the presence of randomization, that's what's going on
2016-09-29T14:52:39Z foom: stassats: sometimes it works.
2016-09-29T14:52:50Z stassats: then it's not random enough
2016-09-29T14:52:53Z foom: I'm not actually sure, but I think it might actually reliably work in 64bit.
2016-09-29T14:53:21Z foom: The kernel doesn't spread the memory out over the entire address space of a 64bit process.
2016-09-29T14:53:57Z stassats: if there's no executable randomization
2016-09-29T14:54:01Z rpg joined #sbcl
2016-09-29T14:54:26Z foom: hmm?
2016-09-29T14:55:17Z nyef`: atgreen: Roughly, SBCL's core files are memory images similar to an executable file, but without full relocation information, so some spaces simply can't be moved, and then SBCL doesn't have a mechanism for relocating the spaces that CAN (in theory) be moved, and SBCL's garbage collector doesn't handle the sort of gaps that would be caused by things being in the middle of a heap space.
2016-09-29T14:57:12Z nyef`: Basically, it's a lot of tricky work to do, to benefit a very small number of users, but that number of users is growing, so it may yet happen.
2016-09-29T14:57:34Z foom: I'm certain that there is a location that aslr wouldn't ever get in the way on x86-64. I'm not certain if the current location is such a location.
2016-09-29T14:58:42Z stassats: interestingly, (multiple-value-call #'+ 'a 'b) doesn't say a word
2016-09-29T14:59:07Z stassats: neither does (multiple-value-call #'print 'a 'b 'c 'd)
2016-09-29T14:59:21Z stassats: but (multiple-value-call #'print) complains
2016-09-29T15:00:49Z jsnell: so there isn't really a hard dependency on /proc; that's just the mechanism we use to re-execute the same binary
2016-09-29T15:01:11Z jsnell: you can just run it with a setarch wrapper manually, and guarantee no ASLR
2016-09-29T15:02:00Z jsnell: (and avoid that warning as well)
2016-09-29T15:04:14Z stassats: what if we fork and call main()?
2016-09-29T15:04:33Z foom: Hm, I think we're actually good. It always puts the binary either at around 0x55xx'xxxx'xxxx or 0x7Fxx'xxxx'xxxx.
2016-09-29T15:04:34Z atgreen: jsnell.. except I'm wanting to run in a hosted docker container environment, where I believe the system calls for disabling ASLR are not available
2016-09-29T15:04:58Z foom: Which is well out of the way of our memory ranges.
2016-09-29T15:05:20Z stassats: foom: what about heap locations?
2016-09-29T15:07:27Z stassats: or just fork?
2016-09-29T15:08:28Z atgreen: foom: what memory ranges does sbcl use on 64-bit x86 Linux?
2016-09-29T15:09:29Z jsnell: atgreen: I have never heard of setpersonality being restricted. have you verified it?
2016-09-29T15:10:41Z stassats: and we already find the executable location for appended cores, in os_get_runtime_executable_path(0);
2016-09-29T15:11:04Z _mjl quit (Quit: Leaving)
2016-09-29T15:11:44Z jsnell: forking wouldn't work
2016-09-29T15:12:44Z atgreen: jsnell, yes - personality() is not on the seccomp whitelist in this environment
2016-09-29T15:13:04Z stassats: os_get_runtime_executable_path uses /proc/self/exe too
2016-09-29T15:13:07Z foom: atgreen: Hm, #x2000'0000 and #x10'0000'0000 upstream, apparently, although we have a local patch changing it to #1100'0000'0000,
2016-09-29T15:13:18Z jsnell: atgreen: wow, harsh
2016-09-29T15:13:23Z atgreen: https://docs.docker.com/engine/security/seccomp/
2016-09-29T15:14:55Z stassats: atgreen: can't you change it?
2016-09-29T15:15:00Z slyrus joined #sbcl
2016-09-29T15:15:50Z atgreen: stassats, you can change it if you are a cluster admin.  but I'm talking about an environment where I am a guest of a multi-tenanted hosted environment with no admin credentials
2016-09-29T15:15:59Z atgreen: openshift.com, for instance
2016-09-29T15:16:12Z stassats: then you're out of luck
2016-09-29T15:16:38Z foom: stassats: if one ignored this entire conversation, then, yes, he's out of luck.
2016-09-29T15:16:42Z foom: stassats: otherwise, ...
2016-09-29T15:16:45Z stassats: atgreen: make sure you build with sb-dynamic-core then
2016-09-29T15:16:47Z atgreen: exactly :)
2016-09-29T15:18:30Z atgreen: so, foom, what I'm understanding is that the ASLR implementation in 64-bit x86 Linux may be limited enough that it does not conflict with the memory range SBCL expects to use.  Is that right?
2016-09-29T15:18:41Z foom: atgreen: Yes.
2016-09-29T15:18:55Z stassats: well, the executable can be mapped randomly as well
2016-09-29T15:19:07Z stassats: and it's normally expected to be constant
2016-09-29T15:19:11Z stassats: unless you use sb-dynamic-core
2016-09-29T15:19:19Z stassats: for a small indirection penalty
2016-09-29T15:19:21Z foom: That will not happen unless it's built with PIE
2016-09-29T15:19:55Z foom: *cannot* happen, I should say.
2016-09-29T15:20:02Z atgreen: I can check to see how it is built for CentOS
2016-09-29T15:20:12Z foom: atgreen: if it works at all, this isn't an issue
2016-09-29T15:21:05Z atgreen: well, it's working.  I was just afraid that  after some indeterminate amount of time SBCL would stomp over something critical
2016-09-29T15:21:27Z stassats: atgreen: if it starts, then it's ok
2016-09-29T15:21:35Z atgreen: ok, great!
2016-09-29T15:22:13Z stassats: foom: PIE gets pushed everywhere nowadays
2016-09-29T15:22:19Z stassats: like the latest OSX
2016-09-29T15:22:29Z stassats: or Xcode, rather
2016-09-29T15:25:38Z stassats: weirdly, the (multiple-value-call #'print 1) warning only surfaces with one argument
2016-09-29T15:28:34Z stassats:  multiple-value-call ir1 translator acts differently in that case
2016-09-29T15:29:12Z stassats: or not really
2016-09-29T15:30:19Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2016-09-29T15:32:17Z rpg joined #sbcl
2016-09-29T15:32:23Z sjl joined #sbcl
2016-09-29T15:37:49Z stassats: ok, i see the problem in (multiple-value-call #'print 1)
2016-09-29T15:38:15Z stassats: the conversion from mv-combination to combination gives the lambda the same name as the target function
2016-09-29T15:38:41Z stassats: and ltn-analyze-known-call gets confused by this
2016-09-29T15:45:02Z stassats: an easy fix, just don't give it any source-name
2016-09-29T15:57:00Z stassats: solved (multiple-value-call #'print 1 2 3) as well
2016-09-29T15:57:11Z stassats: but not the type error
2016-09-29T16:13:29Z scymtym joined #sbcl
2016-09-29T16:14:18Z sjl quit (Ping timeout: 264 seconds)
2016-09-29T16:15:40Z scymtym: 2|5jh,m.,khjkjj,."~/code/cor-lab/rsb/rsb-cpp/build//rsb/Version.cpp""~/code/cor-lab/rsb/rsb-cpp/build//rsb/Version.cpp""~/code/cor-lab/rsb/rsb-cpp/build//rsb/Version.cpp""~/code/cor-lab/rsb/rsb-cpp/build//rsb/Version.cpp"1111111111!11111111111kjk4
2016-09-29T16:15:41Z scymtym: <,k
2016-09-29T16:16:04Z stassats: right
2016-09-29T16:16:21Z scymtym: sorry, child <-> keyboard activity
2016-09-29T16:16:38Z ktt9`: My first guess was cat.
2016-09-29T16:16:47Z ktt9`: Happens all the time.
2016-09-29T16:16:56Z stassats: ktt9`: cats are not smart enough to be writing C++
2016-09-29T16:17:00Z sjl joined #sbcl
2016-09-29T16:17:49Z ktt9`: I'm a cat. Meow.
2016-09-29T16:24:21Z milanj quit (Quit: Leaving)
2016-09-29T16:27:49Z sjl quit (Ping timeout: 272 seconds)
2016-09-29T16:29:30Z stassats: ok, (multiple-value-call #'known-fun more-than-one-argument) doesn't get optimized
2016-09-29T16:30:46Z stassats: and solving that touches a lot of stuff
2016-09-29T16:31:39Z sjl joined #sbcl
2016-09-29T16:39:20Z ktt9` quit (Quit: ERC Version 5.3 (IRC client for Emacs))
2016-09-29T16:44:30Z whiteline quit (Quit: Leaving)
2016-09-29T16:44:50Z whiteline joined #sbcl
2016-09-29T16:46:06Z m00natic` quit (Ping timeout: 264 seconds)
2016-09-29T16:47:38Z m00natic` joined #sbcl
2016-09-29T16:49:40Z chris2 quit (Ping timeout: 255 seconds)
2016-09-29T16:52:01Z stassats: and (multiple-value-call #'bar 1) is not optimized as well
2016-09-29T16:52:09Z stassats: yeah, m-v-call needs some work
2016-09-29T16:58:51Z slyrus quit (Ping timeout: 272 seconds)
2016-09-29T17:09:03Z m00natic` quit (Remote host closed the connection)
2016-09-29T17:17:19Z Quadresce joined #sbcl
2016-09-29T17:18:20Z chris2 joined #sbcl
2016-09-29T17:21:46Z rumbler31 joined #sbcl
2016-09-29T17:25:42Z sjl quit (Ping timeout: 264 seconds)
2016-09-29T17:26:40Z rpg quit (Read error: Connection reset by peer)
2016-09-29T17:34:42Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-29T17:35:00Z Quadresce joined #sbcl
2016-09-29T17:36:00Z Bike joined #sbcl
2016-09-29T17:37:55Z stassats: Fault @ 0x1001521830, page 676 not marked as write-protected:
2016-09-29T17:47:52Z sjl joined #sbcl
2016-09-29T17:50:03Z rpg joined #sbcl
2016-09-29T17:50:40Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-29T18:00:26Z sjl quit (Read error: Connection timed out)
2016-09-29T18:01:07Z sjl joined #sbcl
2016-09-29T18:11:41Z Arathnim quit (Ping timeout: 240 seconds)
2016-09-29T18:47:48Z stassats: managed to get (multiple-value-call #'print 1 t) let converted
2016-09-29T18:49:43Z stassats: hacky and incomplete, but i got the idea how to accomplish it in general
2016-09-29T18:58:52Z gingerale joined #sbcl
2016-09-29T19:17:58Z Quadresce joined #sbcl
2016-09-29T19:31:57Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-29T19:42:52Z jdz_ joined #sbcl
2016-09-29T19:43:21Z abbe quit (Ping timeout: 252 seconds)
2016-09-29T19:43:21Z jdz quit (Ping timeout: 252 seconds)
2016-09-29T19:44:35Z abbe joined #sbcl
2016-09-29T19:46:35Z rpg quit (Ping timeout: 244 seconds)
2016-09-29T19:47:00Z pchrist_ joined #sbcl
2016-09-29T19:50:01Z pchrist quit (Ping timeout: 252 seconds)
2016-09-29T19:50:04Z sigjuice quit (Quit: ZNC - http://znc.in)
2016-09-29T19:51:10Z sigjuice joined #sbcl
2016-09-29T20:03:07Z DeadTrickster quit (Ping timeout: 244 seconds)
2016-09-29T20:03:48Z dougk_ quit (Ping timeout: 272 seconds)
2016-09-29T20:04:20Z DeadTrickster joined #sbcl
2016-09-29T20:05:13Z karswell quit (Remote host closed the connection)
2016-09-29T20:05:59Z karswell joined #sbcl
2016-09-29T20:17:58Z dougk_ joined #sbcl
2016-09-29T20:18:29Z Quadresce joined #sbcl
2016-09-29T20:33:26Z rumbler31 quit (Remote host closed the connection)
2016-09-29T20:38:13Z rpg joined #sbcl
2016-09-29T20:39:16Z Quadresce: Has anybody thought about implementing a concurrent collector?
2016-09-29T20:39:30Z stassats: no, never
2016-09-29T20:40:00Z gingerale quit (Remote host closed the connection)
2016-09-29T20:43:30Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-29T20:50:52Z joshe: heh
2016-09-29T20:57:16Z Quadresce joined #sbcl
2016-09-29T20:58:32Z Quadresce quit (Client Quit)
2016-09-29T21:05:37Z emartenson__ joined #sbcl
2016-09-29T21:05:52Z loke quit (Ping timeout: 252 seconds)
2016-09-29T21:05:56Z Quadresce joined #sbcl
2016-09-29T21:06:45Z Quadresce: stassats, I of course am attempting to ask for an answer with more substance, because it is something that is becoming increasingly important.
2016-09-29T21:07:18Z stassats: is it?
2016-09-29T21:07:59Z Quadresce: Possibly, but maybe not if a particular application can be re-designed to be process-oriented instead of thread-oriented.
2016-09-29T21:08:13Z Quadresce: I'm interested in 200+ GB heaps with many threads.
2016-09-29T21:19:48Z schjetne quit (Ping timeout: 276 seconds)
2016-09-29T21:20:37Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2016-09-29T21:23:04Z pipping left #sbcl
2016-09-29T21:24:23Z schjetne joined #sbcl
2016-09-29T21:26:24Z nyef`: Quadresce: I've occasionally thought about trying to implement different GC schemes in SBCL, but the answer I keep coming to is that it's a lot of fiddly detail, and it'd be time taken away from other things.
2016-09-29T21:33:02Z rumbler31 joined #sbcl
2016-09-29T21:37:10Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-29T21:38:09Z rumbler31 quit (Ping timeout: 272 seconds)
2016-09-29T21:39:14Z Quadresce joined #sbcl
2016-09-29T21:47:02Z ASau joined #sbcl
2016-09-29T21:48:11Z atgreen quit (Read error: Connection reset by peer)
2016-09-29T21:55:41Z madbub quit (Ping timeout: 240 seconds)
2016-09-29T21:56:42Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-29T22:06:37Z rpg joined #sbcl
2016-09-29T22:39:09Z Quadresce joined #sbcl
2016-09-29T22:40:59Z slyrus joined #sbcl
2016-09-29T22:43:09Z Quadresce quit (Client Quit)
2016-09-29T22:45:19Z Quadresce joined #sbcl
2016-09-29T22:47:58Z Quadresce quit (Client Quit)
2016-09-29T22:59:53Z atgreen joined #sbcl
2016-09-29T23:15:51Z Quadresce joined #sbcl
2016-09-29T23:21:36Z rpg_ joined #sbcl
2016-09-29T23:21:41Z rpg quit (Read error: Connection reset by peer)
2016-09-29T23:25:58Z attila_lendvai: I think having first-class heaps would help much more by allowing the programmers to formally tell the system what they know about allocation... i.e. an UWP could free up a big chunk of memory without even scavenging it
2016-09-29T23:36:30Z ktt9 quit (Ping timeout: 244 seconds)
2016-09-29T23:40:07Z slyrus quit (Ping timeout: 244 seconds)
2016-09-30T00:04:19Z White_Flame: first class heaps would be great for an application server settings, to allow hard maximums on memory usage for subcomponent
2016-09-30T00:04:20Z White_Flame: s
2016-09-30T00:08:00Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-30T00:08:55Z cromachina joined #sbcl
2016-09-30T00:11:35Z stassats: i can list "would be great" all night
2016-09-30T00:24:23Z attila_lendvai: a rewording: first class heaps are probably simpler to implement than threaded GC's
2016-09-30T00:28:22Z nyef`: There is that, more or less, although that's not necessarily easy to get right either.
2016-09-30T00:49:06Z sjl quit (Ping timeout: 264 seconds)
2016-09-30T01:14:47Z attila_lendvai quit (Ping timeout: 252 seconds)
2016-09-30T01:21:21Z em1l joined #sbcl
2016-09-30T01:24:29Z em1l_ quit (Ping timeout: 244 seconds)
2016-09-30T01:34:18Z rumbler31 joined #sbcl
2016-09-30T01:39:27Z rumbler31 quit (Ping timeout: 272 seconds)
2016-09-30T01:44:39Z rpg_ quit (Quit: Textual IRC Client: www.textualapp.com)
2016-09-30T02:19:15Z emartenson__ is now known as loke
2016-09-30T03:29:39Z stassats quit (Ping timeout: 276 seconds)
2016-09-30T05:28:58Z flip214: After booting Linux 4.7.0 I now get "ensure_space: failed to validate X bytes at Y"
2016-09-30T05:29:10Z flip214: ulimit is okay
2016-09-30T05:32:00Z flip214: Hmmm, "locked memory" is now _required_ to be big enough... it wasn't with 4.6.
2016-09-30T06:11:29Z scymtym quit (Remote host closed the connection)
2016-09-30T06:28:57Z m00natic joined #sbcl
2016-09-30T06:45:21Z jdz_ is now known as jdz
2016-09-30T07:15:30Z ASau quit (Ping timeout: 264 seconds)
2016-09-30T07:39:54Z scymtym joined #sbcl
2016-09-30T08:26:47Z gingerale joined #sbcl
2016-09-30T08:35:26Z rumbler31 joined #sbcl
2016-09-30T08:40:06Z rumbler31 quit (Ping timeout: 264 seconds)
2016-09-30T08:42:12Z Bike quit (Quit: no)
2016-09-30T09:26:07Z attila_lendvai joined #sbcl
2016-09-30T10:06:38Z attila_lendvai quit (Ping timeout: 252 seconds)
2016-09-30T10:26:27Z rudolfochrist joined #sbcl
2016-09-30T10:57:24Z rumbler31 joined #sbcl
2016-09-30T11:00:30Z rudolfochrist quit (Ping timeout: 264 seconds)
2016-09-30T11:13:07Z sjl joined #sbcl
2016-09-30T11:15:42Z edgar-rft quit (Ping timeout: 276 seconds)
2016-09-30T11:26:47Z rudolfochrist joined #sbcl
2016-09-30T11:31:34Z rudolfochrist quit (Ping timeout: 244 seconds)
2016-09-30T11:33:07Z rumbler31 quit (Remote host closed the connection)
2016-09-30T11:46:02Z atgreen quit (Ping timeout: 244 seconds)
2016-09-30T12:09:53Z madbub joined #sbcl
2016-09-30T12:10:18Z madbub quit (Remote host closed the connection)
2016-09-30T12:11:51Z madbub joined #sbcl
2016-09-30T12:17:18Z sjl quit (Ping timeout: 264 seconds)
2016-09-30T12:35:19Z rumbler31 joined #sbcl
2016-09-30T12:41:37Z rumbler31 quit (Remote host closed the connection)
2016-09-30T12:51:19Z cromachina quit (Quit: Leaving)
2016-09-30T12:59:59Z atgreen joined #sbcl
2016-09-30T13:01:38Z stassats joined #sbcl
2016-09-30T13:10:33Z edgar-rft joined #sbcl
2016-09-30T13:35:37Z m00natic quit (Remote host closed the connection)
2016-09-30T14:05:33Z sjl joined #sbcl
2016-09-30T14:49:11Z scymtym quit (Remote host closed the connection)
2016-09-30T14:49:47Z scymtym joined #sbcl
2016-09-30T14:49:51Z sjl quit (Ping timeout: 272 seconds)
2016-09-30T15:20:02Z slyrus joined #sbcl
2016-09-30T15:34:59Z rumbler31 joined #sbcl
2016-09-30T15:49:33Z Xof:   The value #P"obj/from-xc/src/code/early-source-location.lisp-obj"
2016-09-30T15:49:34Z Xof:   is not of type
2016-09-30T15:49:34Z Xof:     (OR (VECTOR CHARACTER) (VECTOR NIL) BASE-STRING SYMBOL ...).
2016-09-30T15:49:37Z Xof: ...
2016-09-30T15:49:54Z Xof: dougk_ ?
2016-09-30T15:52:00Z Xof: is :key 'namestring enough?
2016-09-30T15:53:04Z Bike joined #sbcl
2016-09-30T15:59:21Z Xof: <tap> <tap> is this thing on?
2016-09-30T16:00:08Z karswell` joined #sbcl
2016-09-30T16:00:33Z scymtym: Xof: we can hear you
2016-09-30T16:00:56Z karswell quit (Remote host closed the connection)
2016-09-30T16:01:26Z Xof: ok
2016-09-30T16:01:54Z Xof: if anyone can look at dougk_'s commit and confirm that adding :key 'namestring is reasonably safe, I can release
2016-09-30T16:02:04Z Xof: as it is, the system doesn't build right now
2016-09-30T16:02:07Z scymtym: and the thing seems to be happening in all configurations: https://ci.cor-lab.org/job/sbcl-master/1770/
2016-09-30T16:02:10Z nyef`: The change to genesis?
2016-09-30T16:02:24Z Xof: (which makes the release dance a bit difficult)
2016-09-30T16:02:37Z dougk_: hi
2016-09-30T16:02:53Z nyef`: Why is this even necessary?
2016-09-30T16:03:06Z dougk_: because you had two different addresses for undefined tramp
2016-09-30T16:03:15Z dougk_: this was totally insane until i figured it out
2016-09-30T16:03:24Z nyef`: Err?
2016-09-30T16:03:32Z dougk_: once in the preload and once in the normal build-order
2016-09-30T16:03:43Z nyef`: It's not supposed to BE in the normal build-order.
2016-09-30T16:03:56Z dougk_: but it is
2016-09-30T16:03:58Z nyef`: That's the point of the :not-genesis flag.
2016-09-30T16:04:37Z dougk_: can you see if your output/cold-sbcl.map has all the tramps listed twice? mine does
2016-09-30T16:05:08Z nyef`: On x86-64/Linux? Mine doesn't.
2016-09-30T16:05:23Z nyef`: "1.3.9.125-9efb8c0-dirty"
2016-09-30T16:06:33Z nyef`: You're doing a parallel build, aren't you?
2016-09-30T16:06:44Z dougk_: yes, i wonder if that matters
2016-09-30T16:06:53Z nyef`: It almost-certainly does.
2016-09-30T16:07:20Z nyef`: Have a look at src/cold/compile-cold-sbcl.lisp
2016-09-30T16:07:42Z dougk_: oh, then i'll revert that commit
2016-09-30T16:08:06Z Xof: thanks
2016-09-30T16:08:22Z Xof: obviously include it early in 1.3.10 with whatever it needs to not type-error :-)
2016-09-30T16:08:35Z nyef`: I think that the original read-only-tramps work predates the parallel build work and wasn't updated when it was forward-ported, because almost nobody DOES a parallel build yet.
2016-09-30T16:08:56Z dougk_: right
2016-09-30T16:09:01Z stassats: is the build broken or i'm just late to the party?
2016-09-30T16:09:23Z Xof: it's broken
2016-09-30T16:09:39Z stassats: cool
2016-09-30T16:09:44Z Xof: we know why
2016-09-30T16:09:46Z dougk_: reverted.
2016-09-30T16:09:48Z Xof: thanks
2016-09-30T16:10:20Z dougk_: you're welcome
2016-09-30T16:10:44Z stassats: wouldn't #"equal work better?
2016-09-30T16:11:49Z nyef`: Doesn't matter, the root problem was in compile-cold-sbcl for parallel builds, not in genesis at all.
2016-09-30T16:11:54Z Xof: judging by the discussion about how this happened, there's probably a better fix in the ... yes
2016-09-30T16:13:27Z nyef`: Basically, the parallel-build and non-parallel-build logic for dealing with file flags diverged a bit around the :not-genesis flag for read-only-tramps.
2016-09-30T16:40:33Z sjl joined #sbcl
2016-09-30T16:49:29Z scymtym: would it make sense to add a "parallel" build with N=1 to the continuous integration setup? (i don't want to block more than one core at a time on the machine)
2016-09-30T16:52:57Z Quadresce joined #sbcl
2016-09-30T16:57:26Z rumbler31 quit (Remote host closed the connection)
2016-09-30T17:09:29Z karswell` quit (Remote host closed the connection)
2016-09-30T17:09:56Z karswell` joined #sbcl
2016-09-30T17:18:46Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-30T17:21:53Z nyef`: ... Looks like we're out of code freeze?
2016-09-30T17:23:40Z stassats: yeah
2016-09-30T17:26:59Z nyef`: Why is there an (UNLESS (FIND :NOT-TARGET FLAGS) ...) in the (UNLESS (POSITION :NOT-TARGET FLAGS) ...) in PARALLEL-MAKE-HOST-2 ?
2016-09-30T17:27:47Z nyef`: Or am I mis-reading what's going on there?
2016-09-30T17:29:09Z stassats: i was merging it, may have mismerged it
2016-09-30T17:29:33Z nyef`: So it should be :NOT-GENESIS for the inner case?
2016-09-30T17:33:57Z rumbler31 joined #sbcl
2016-09-30T17:48:16Z nyef`: Ah, yes, it should. I see the change that introduced it.
2016-09-30T17:56:01Z jrm2 joined #sbcl
2016-09-30T17:58:42Z jrm quit (Ping timeout: 276 seconds)
2016-09-30T17:58:42Z jrm2 is now known as jrm
2016-09-30T18:34:03Z sjl quit (Ping timeout: 272 seconds)
2016-09-30T18:50:17Z jasom joined #sbcl
2016-09-30T18:51:15Z jasom: For any places that are CASable, can one assume that writes to them are atomic (i.e. if many threads SETF a place simultaneously, at any given point the place will have a value equal to one of the values SETF to that place)?
2016-09-30T18:57:47Z nyef`: They might be atomic, but mind your barrier semantics!
2016-09-30T18:59:05Z jasom: but this also means I can safely setf two different slots of a standard-object simultaneously, no?
2016-09-30T18:59:49Z stassats: you can
2016-09-30T19:00:14Z stassats: even the same one
2016-09-30T19:02:02Z jasom: meh, I hate low-level threaading.
2016-09-30T19:02:08Z rumbler31 quit (Remote host closed the connection)
2016-09-30T19:02:38Z jasom: I can keep a *lot* of state in my brain at once, and yet it is still hard.  I don't know why anyone ever thought it was a good idea to use mutexes for thread safety in everyday code.
2016-09-30T19:03:09Z stassats: you have a better solution?
2016-09-30T19:04:22Z jasom: stassats: multiprocessing, queues, &c. like we use cond instead of goto, despite them both being jmp at the machine level.
2016-09-30T19:04:54Z jasom: implement a queue with a mutex, and then move state through queues rather than simultaneously accessing things at your high level code.
2016-09-30T19:06:21Z stassats: i don't quite follow, but ok
2016-09-30T19:06:42Z jasom: basically the less of my code relies on me correctly following the low-level semantics of multithreaded code on the particular platform, the more likely my code is to be correct.
2016-09-30T19:07:15Z stassats: mutexes aren't really low level
2016-09-30T19:07:27Z jasom: when to use a mutex is low level
2016-09-30T19:07:44Z jasom: it requires knowing if the value(s) you are updating are updated atomically or not.
2016-09-30T19:08:21Z stassats: that's what the mutex is for
2016-09-30T19:09:37Z stassats: you can't really optimize things without know what the things are
2016-09-30T19:09:58Z stassats: so you just want things to be equally slow?
2016-09-30T19:10:02Z stassats: well, no one's stopping you
2016-09-30T19:10:53Z jasom: If I don't write to an object that might be accessed from other threads I don't need to remember if slot-value is atomic and I don't need to use a mutex.
2016-09-30T19:12:20Z jasom: it's using shared state as inter-thread communcation that is error prone, not the mutexes per-se.
2016-09-30T19:13:00Z stassats: well, mutexes usually ensure atomicity and serialization
2016-09-30T19:16:42Z schjetne quit (Ping timeout: 264 seconds)
2016-09-30T19:29:15Z schjetne joined #sbcl
2016-09-30T19:33:17Z rumbler31 joined #sbcl
2016-09-30T19:54:48Z scymtym quit (Ping timeout: 244 seconds)
2016-09-30T20:07:12Z attila_lendvai joined #sbcl
2016-09-30T20:08:20Z Quadresce joined #sbcl
2016-09-30T20:21:10Z reb quit (Remote host closed the connection)
2016-09-30T20:45:43Z atgreen quit (Read error: Connection reset by peer)
2016-09-30T20:46:01Z slyrus quit (Remote host closed the connection)
2016-09-30T21:01:12Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-30T21:02:41Z atgreen joined #sbcl
2016-09-30T21:03:48Z Quadresce joined #sbcl
2016-09-30T21:18:30Z madbub quit (Ping timeout: 244 seconds)
2016-09-30T21:25:11Z prxq joined #sbcl
2016-09-30T21:30:41Z atgreen quit (Read error: Connection reset by peer)
2016-09-30T21:31:34Z ASau joined #sbcl
2016-09-30T21:34:31Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-30T21:34:47Z rumbler31 quit (Remote host closed the connection)
2016-09-30T21:37:16Z Quadresce joined #sbcl
2016-09-30T21:37:19Z prxq quit (Remote host closed the connection)
2016-09-30T21:45:38Z scymtym joined #sbcl
2016-09-30T21:49:43Z Baggers joined #sbcl
2016-09-30T21:57:14Z Baggers: I'm sorry for the variation on a old question, but given (defstruct bar (ptr (sb-sys:int-sap 0) :type sb-sys:system-area-pointer)) is there a way to avoid the sap->ptr cost here:
2016-09-30T21:57:15Z Baggers: (setf (bar-ptr foo) (sb-sys:sap+ (bar-ptr foo) 1))
2016-09-30T21:57:18Z Baggers: I have read https://sourceforge.net/p/sbcl/mailman/sbcl-help/thread/AANLkTikbdzyFeMg7nDWUllF8aCbLCLxAEOFXOcsdH0SM@mail.gmail.com/ and inlining bar-ptr & (setf bar-ptr) have no effect. I would like to know if there is any way to modify a pointer in a struct in-place. (I'm open to hairy solutions)
2016-09-30T21:58:22Z Baggers: I'm asking as updating pointers is the #1 cause of consing in my programs
2016-09-30T21:58:38Z stassats: Baggers: store integers
2016-09-30T21:58:48Z stassats: we don't have support unboxed saps in structures
2016-09-30T21:59:41Z stassats: the type would be sb-ext:word
2016-09-30T22:00:21Z Baggers: stassats: thanks, that helps a lot. Is there a sensible way to pass these to sbcl's ffi where a pointer is expected?
2016-09-30T22:00:45Z stassats: you'll have to int-sap them
2016-09-30T22:00:59Z stassats: but if everything is inlined where you do that it won't be a problem
2016-09-30T22:01:22Z Baggers: stassats: that is a massive relief to hear, thankyou
2016-09-30T22:02:21Z stassats: and sb-ext:word would be the wrong thing on Alpha, but you don't care about that
2016-09-30T22:02:23Z stassats: (nobody does)
2016-09-30T22:05:02Z Baggers: agreed, I wont be targeting alpha
2016-09-30T22:05:18Z stassats: but defstruct should really start supporting unboxed system-area-pointer
2016-09-30T22:09:33Z nyef`: stassats: Word is a poor type for SAPs, they can be wider than heap words on some platforms.
2016-09-30T22:09:41Z nyef`: Ah, so you pointed out.
2016-09-30T22:09:56Z stassats: "some platform"
2016-09-30T22:09:59Z nyef`: As nobody, I care about alpha. d-:
2016-09-30T22:10:22Z stassats: i doubt you care about running third party programs on it
2016-09-30T22:12:40Z nyef`: Right. That would be unlikely to work, given the DX SAP thing in SBCL's runtime. I just care about the situation not getting any worse, since it should be our exemplar platform for barriers for multi-thread synchronization.
2016-09-30T22:13:04Z stassats: how are threads coming up on alpha?
2016-09-30T22:13:05Z gingerale quit (Remote host closed the connection)
2016-09-30T22:13:48Z nyef`: Threads on alpha won't happen until we've sorted that DX SAP thing (or just made it a true 64-bit port), gotten gencgc running, and so on.
2016-09-30T22:14:02Z nyef`: It sits behind bringing SPARC up to scratch in terms of my priorities.
2016-09-30T22:14:35Z nyef`: (And SPARC sits behind HPPA, because some platform other than HPPA should have the "least maintained backend" card for a little while, as far as I'm concerned.)
2016-09-30T22:15:02Z nyef`: And HPPA seems to sit behind a few other things on my queue.
2016-09-30T22:17:17Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-30T22:18:15Z DeadTrickster quit (Ping timeout: 272 seconds)
2016-09-30T22:18:20Z Quadresce joined #sbcl
2016-09-30T22:22:54Z slyrus joined #sbcl
2016-09-30T22:26:25Z Quadresce quit (Quit: This computer has gone to sleep)
2016-09-30T22:40:35Z stassats: made saps raw on x86-64, but there is a question of alien-value
2016-09-30T22:41:00Z stassats: alien-value-sap wasn't boxed before
2016-09-30T22:41:25Z stassats: wasn't raw before
2016-09-30T22:41:38Z stassats: but if it's raw now then the constructor needs to be inlined
2016-09-30T22:41:52Z rumbler31 joined #sbcl
2016-09-30T22:41:59Z stassats: so, it'll take up less space, good, but now each time alien-value-sap is called it's going to cons
2016-09-30T22:46:42Z rumbler31 quit (Ping timeout: 264 seconds)
2016-09-30T22:46:55Z Baggers: goodnight all, thanks again for the help
2016-09-30T22:46:56Z Baggers left #sbcl
2016-09-30T22:47:55Z cromachina joined #sbcl
2016-09-30T22:56:08Z stassats: sb-bsd-sockets uses a lot of alien conversion, so, if it improves then i'm pushing raw saps on x86-64
2016-09-30T22:58:56Z stassats: and putting inline on %sap-alien gets   The function SB!C::WARN-IF-INLINE-FAILED/PROCLAIM is undefined.
2016-09-30T23:06:28Z stassats: and inline %sap-alien has no effect
2016-09-30T23:06:39Z stassats: "could not be inlined because its source code was not saved."
2016-09-30T23:06:47Z stassats: during xc
2016-09-30T23:11:15Z stassats: well, i don't want to fiddle with that, i think alien-value should just become a primitive object
2016-09-30T23:12:47Z stassats: though, probably won't save any space
2016-09-30T23:15:33Z stassats: it can, if alien-value-type is stored in the header, but that's more involved
2016-09-30T23:15:48Z stassats: ok, i still need to figure out how the hell to inline %sap-alien
2016-09-30T23:17:54Z atgreen joined #sbcl
2016-09-30T23:21:49Z sjl joined #sbcl
2016-09-30T23:32:22Z slyrus quit (Remote host closed the connection)
2016-09-30T23:32:50Z slyrus joined #sbcl
2016-09-30T23:50:31Z stassats: and what is https://github.com/sbcl/sbcl/blob/master/src/compiler/aliencomp.lisp#L469 supposed to do?