2016-08-25T00:05:56Z cromachina joined #sbcl
2016-08-25T00:09:16Z milanj joined #sbcl
2016-08-25T00:30:07Z oleo_ joined #sbcl
2016-08-25T00:32:36Z oleo quit (Ping timeout: 265 seconds)
2016-08-25T00:32:46Z oleo_ quit (Changing host)
2016-08-25T00:32:47Z oleo_ joined #sbcl
2016-08-25T00:32:57Z oleo_ is now known as oleo
2016-08-25T01:04:31Z em1l joined #sbcl
2016-08-25T01:07:44Z em1l_ quit (Ping timeout: 250 seconds)
2016-08-25T01:10:12Z milanj quit (Quit: This computer has gone to sleep)
2016-08-25T01:28:06Z DeadTrickster quit (Ping timeout: 265 seconds)
2016-08-25T01:28:12Z scymtym quit (Ping timeout: 258 seconds)
2016-08-25T01:31:15Z ASau quit (Ping timeout: 276 seconds)
2016-08-25T01:31:57Z rumbler3_ joined #sbcl
2016-08-25T01:34:36Z rumbler31 quit (Ping timeout: 250 seconds)
2016-08-25T01:40:13Z whiteline quit (Remote host closed the connection)
2016-08-25T01:40:36Z whiteline joined #sbcl
2016-08-25T01:49:16Z whiteline quit (Remote host closed the connection)
2016-08-25T01:49:43Z whiteline joined #sbcl
2016-08-25T01:50:44Z DeadTrickster joined #sbcl
2016-08-25T02:13:07Z rumbler3_ quit
2016-08-25T03:13:57Z sjl quit (Ping timeout: 276 seconds)
2016-08-25T03:46:54Z sjl joined #sbcl
2016-08-25T04:05:57Z sjl quit (Ping timeout: 276 seconds)
2016-08-25T04:37:38Z milanj joined #sbcl
2016-08-25T04:40:25Z oleo quit (Quit: Leaving)
2016-08-25T05:26:56Z shka_ joined #sbcl
2016-08-25T06:11:39Z milanj quit (Quit: This computer has gone to sleep)
2016-08-25T06:28:52Z shka_ quit (Ping timeout: 240 seconds)
2016-08-25T07:31:11Z slyrus joined #sbcl
2016-08-25T07:43:49Z ASau joined #sbcl
2016-08-25T07:49:26Z scymtym joined #sbcl
2016-08-25T07:57:03Z PosterdatiMobile quit (Ping timeout: 240 seconds)
2016-08-25T07:57:46Z PosterdatiMobile joined #sbcl
2016-08-25T08:00:21Z scymtym quit (Ping timeout: 258 seconds)
2016-08-25T08:00:56Z scymtym joined #sbcl
2016-08-25T08:46:30Z Bike quit (Quit: inemabtale)
2016-08-25T09:17:50Z iskander quit (Ping timeout: 244 seconds)
2016-08-25T09:37:20Z milanj joined #sbcl
2016-08-25T09:38:19Z iskander joined #sbcl
2016-08-25T09:42:00Z ASau quit (Ping timeout: 276 seconds)
2016-08-25T10:00:08Z scymtym_ joined #sbcl
2016-08-25T10:04:33Z scymtym quit (Ping timeout: 258 seconds)
2016-08-25T10:10:22Z scymtym__ joined #sbcl
2016-08-25T10:14:31Z scymtym_ quit (Ping timeout: 258 seconds)
2016-08-25T10:33:36Z milanj quit (Quit: This computer has gone to sleep)
2016-08-25T10:40:34Z jackdaniel: bordeaux-threads+sbcl question: why impl-sbcl.lisp doesn't simply implement acquire-recursive-lock/release-recursive-lock with grab-mutex and release-mutex like impl-clozure.lisp does? as far as I understand, grab-mutex can handle both recursive and non-recursive, do we really need to do counting? (related paste http://paste.lisp.org/display/324197)
2016-08-25T10:59:11Z milanj joined #sbcl
2016-08-25T11:15:14Z sjl joined #sbcl
2016-08-25T11:15:45Z attila_lendvai joined #sbcl
2016-08-25T11:34:01Z rudolfochrist joined #sbcl
2016-08-25T11:37:56Z stassats joined #sbcl
2016-08-25T11:46:32Z stassats: jackdaniel: recursive locks are different from non recursive locks
2016-08-25T11:49:58Z DGASAU joined #sbcl
2016-08-25T12:16:30Z cromachina_ joined #sbcl
2016-08-25T12:17:51Z schjetne quit (Ping timeout: 264 seconds)
2016-08-25T12:19:52Z cromachina quit (Ping timeout: 258 seconds)
2016-08-25T12:20:15Z DeadTrickster quit (Ping timeout: 258 seconds)
2016-08-25T12:23:42Z payphone quit (Ping timeout: 258 seconds)
2016-08-25T12:25:15Z payphone joined #sbcl
2016-08-25T12:25:54Z leo_song quit (Ping timeout: 250 seconds)
2016-08-25T12:26:16Z leo_song joined #sbcl
2016-08-25T12:27:10Z DeadTrickster joined #sbcl
2016-08-25T12:45:18Z rudolfochrist quit (Ping timeout: 276 seconds)
2016-08-25T12:45:31Z rudolfoc` joined #sbcl
2016-08-25T12:47:27Z rudolfoc` is now known as rudolfochrist
2016-08-25T12:55:38Z DavidGu joined #sbcl
2016-08-25T13:00:49Z DavidGu1 joined #sbcl
2016-08-25T13:02:51Z DavidGu quit (Ping timeout: 264 seconds)
2016-08-25T13:02:52Z DavidGu1 is now known as DavidGu
2016-08-25T13:03:04Z jackdaniel: so it's not simple grab/release in case of SBCL? OK, thanks.
2016-08-25T13:06:39Z attila_lendvai quit (Read error: Connection reset by peer)
2016-08-25T13:08:43Z attila_lendvai joined #sbcl
2016-08-25T13:08:54Z attila_lendvai quit (Changing host)
2016-08-25T13:08:54Z attila_lendvai joined #sbcl
2016-08-25T13:10:31Z cromachina_ quit (Read error: Connection reset by peer)
2016-08-25T13:24:07Z oleo joined #sbcl
2016-08-25T13:30:01Z scymtym__ quit (Ping timeout: 258 seconds)
2016-08-25T13:35:51Z rudolfochrist quit (Ping timeout: 264 seconds)
2016-08-25T13:41:08Z rudolfochrist joined #sbcl
2016-08-25T13:52:41Z rudolfochrist quit (Quit: ERC (IRC client for Emacs 25.1.1))
2016-08-25T14:22:14Z stassats quit (Ping timeout: 260 seconds)
2016-08-25T14:24:15Z stassats joined #sbcl
2016-08-25T14:28:41Z stassats quit (Ping timeout: 265 seconds)
2016-08-25T14:44:55Z scymtym joined #sbcl
2016-08-25T14:58:28Z DavidGu quit (Ping timeout: 244 seconds)
2016-08-25T15:07:08Z gingerale joined #sbcl
2016-08-25T16:19:07Z karswell` is now known as karswell
2016-08-25T16:26:02Z rpg joined #sbcl
2016-08-25T16:46:40Z Bike joined #sbcl
2016-08-25T16:47:25Z DavidGu joined #sbcl
2016-08-25T16:49:14Z attila_lendvai quit (Ping timeout: 260 seconds)
2016-08-25T16:54:54Z Posterdati quit (Ping timeout: 276 seconds)
2016-08-25T16:55:36Z |3b| quit (Read error: Connection reset by peer)
2016-08-25T16:55:40Z PosterdatiMobile quit (Ping timeout: 252 seconds)
2016-08-25T16:57:18Z |3b| joined #sbcl
2016-08-25T17:07:03Z DavidGu quit (Ping timeout: 264 seconds)
2016-08-25T17:07:32Z Posterdati joined #sbcl
2016-08-25T17:08:24Z PosterdatiMobile joined #sbcl
2016-08-25T17:18:04Z shka_ joined #sbcl
2016-08-25T17:56:13Z slyrus quit (Ping timeout: 250 seconds)
2016-08-25T17:58:33Z sjl quit (Ping timeout: 240 seconds)
2016-08-25T18:46:39Z les quit (Ping timeout: 258 seconds)
2016-08-25T18:49:20Z les joined #sbcl
2016-08-25T18:58:24Z DavidGu joined #sbcl
2016-08-25T18:58:28Z DavidGu quit (Client Quit)
2016-08-25T19:00:43Z rpg: anyone there know about sbcl-devel?
2016-08-25T19:01:15Z rpg: about how it's administered?
2016-08-25T19:12:48Z milanj quit (Quit: This computer has gone to sleep)
2016-08-25T19:49:01Z attila_lendvai joined #sbcl
2016-08-25T19:49:01Z attila_lendvai quit (Changing host)
2016-08-25T19:49:01Z attila_lendvai joined #sbcl
2016-08-25T19:54:30Z scymtym quit (Ping timeout: 258 seconds)
2016-08-25T19:56:42Z jackc- quit (Quit: Reconnecting)
2016-08-25T19:56:49Z jackc- joined #sbcl
2016-08-25T20:10:40Z milanj joined #sbcl
2016-08-25T20:22:49Z gingerale quit (Remote host closed the connection)
2016-08-25T20:26:01Z sjl joined #sbcl
2016-08-25T20:53:10Z rpg quit (Quit: Textual IRC Client: www.textualapp.com)
2016-08-25T20:57:32Z scymtym joined #sbcl
2016-08-25T20:59:10Z stassats joined #sbcl
2016-08-25T21:05:30Z PosterdatiMobile quit (Ping timeout: 265 seconds)
2016-08-25T21:06:33Z Posterdati quit (Ping timeout: 240 seconds)
2016-08-25T21:12:39Z PosterdatiMobile joined #sbcl
2016-08-25T21:17:26Z slyrus joined #sbcl
2016-08-25T21:21:09Z Posterdati joined #sbcl
2016-08-25T21:56:35Z rszeno joined #sbcl
2016-08-25T22:01:36Z rszeno quit (Quit: Leaving.)
2016-08-25T22:03:01Z veckon joined #sbcl
2016-08-25T22:04:57Z shka_ quit (Ping timeout: 276 seconds)
2016-08-25T22:05:25Z veckon: Hello, has anyone gotten SBCL to build from source on mac os 10.12? Everything seems to go fine until warm init when the heap explodes (it just keeps rising until you run out of swap).
2016-08-25T22:10:37Z stassats quit (Read error: Connection reset by peer)
2016-08-25T22:10:44Z stassats joined #sbcl
2016-08-25T22:11:17Z dougk: veckon:  I use mac os 10.11.6 with no problem. do you have a build log?
2016-08-25T22:12:50Z veckon: I suspect that it's something to do with 10.12. I've never had an issue before.
2016-08-25T22:13:13Z veckon: Not sure, I had to kill the process.
2016-08-25T22:15:00Z veckon: Where would a log be? (Is there a flag I should add to when running make.sh?)
2016-08-25T22:16:14Z dougk: no flag, just whatever the stdout/stderr say. What's the last thing printed?
2016-08-25T22:17:09Z veckon: //entering make-target-2.sh
2016-08-25T22:17:10Z veckon: //doing warm init - compilation phase
2016-08-25T22:17:10Z veckon: This is SBCL 1.3.8.46-4d9f789, an implementation of ANSI Common Lisp.
2016-08-25T22:17:10Z veckon: More information about SBCL is available at <http://www.sbcl.org/>.
2016-08-25T22:17:13Z veckon:  
2016-08-25T22:17:16Z veckon: SBCL is free software, provided as is, with absolutely no warranty.
2016-08-25T22:17:19Z veckon: It is mostly in the public domain; some portions are provided under
2016-08-25T22:17:22Z veckon: BSD-style licenses.  See the CREDITS and COPYING files in the
2016-08-25T22:17:25Z veckon: distribution for more information.
2016-08-25T22:17:28Z dougk: don't paste it here, paste it somewhere
2016-08-25T22:17:28Z veckon: Just the standard boilerplate when you startup the image.
2016-08-25T22:17:43Z dougk: and literally nothing else after that?
2016-08-25T22:17:59Z veckon: Nope, then the heap just goes up.
2016-08-25T22:18:38Z dougk: ok, in that case try adding :sb-show to customize-target-features.lisp-expr or --with-sb-show to the make.sh invocation
2016-08-25T22:18:53Z veckon: okay, I'll report back. Thanks!
2016-08-25T22:25:33Z stassats quit (Ping timeout: 240 seconds)
2016-08-25T22:28:11Z veckon: I couldn't get everything, but here's a decent amount of the scrollback before it explodes. https://gist.github.com/aamedina/d032300463ece466cf3265ecd89f300a
2016-08-25T22:29:53Z veckon: Oops, forgot to kill the process. Was up to 60 GB of memory already =P
2016-08-25T22:30:37Z dougk: wow. that's doing absolutely nothing after entering make-target-2. This will require either annotating the C code with "hi" messages at various points, or lldb
2016-08-25T22:30:44Z dougk: what's your pleasure?
2016-08-25T22:31:01Z dougk: it's not even gotten to calling Lisp, most likely
2016-08-25T22:31:05Z veckon: Whatever you recommend?
2016-08-25T22:33:31Z dougk: well, i'm not very competent with lldb and would not be able to tell you what to type. So if you don't mind, the first place I'd put a "hi" message is in "ensure_space" just to see if it got that far
2016-08-25T22:33:48Z veckon: sure
2016-08-25T22:35:22Z veckon: I'm decently comfortable with lldb fwiw if you have a debugging strategy in mind, I'm just ignorant of the sbcl internals.
2016-08-25T22:37:00Z scymtym quit (Remote host closed the connection)
2016-08-25T22:37:13Z scymtym joined #sbcl
2016-08-25T22:37:47Z dougk: ah, sure, stop it in main and let's what happens
2016-08-25T22:38:23Z dougk: you want to invoke lldb on src/runtime/sbcl and tell it the arguments of --core output/cold-sbcl.core
2016-08-25T22:40:05Z veckon: k
2016-08-25T22:41:12Z veckon: So I'm there. What next?
2016-08-25T22:41:46Z dougk: i'd guess we should binary search "main". It's not that bad. Put a breakpoint at gc_init
2016-08-25T22:42:06Z veckon: kk
2016-08-25T22:42:08Z dougk: and run that. If it fails, move the breakpoint back to the call to os_init
2016-08-25T22:43:01Z dougk: i believe you'll want to tell lldb to ignore sigusr1
2016-08-25T22:43:32Z dougk: what's the equivalent of gdb's "handle sigusr1 nostop pass" ?
2016-08-25T22:44:36Z veckon: process handle -p BOOL -s BOOL SIGUSR1
2016-08-25T22:44:42Z veckon: pass/stop for -p / -s
2016-08-25T22:45:16Z dougk: works for me. Thanks
2016-08-25T22:45:55Z veckon: so it looks like everything up to gc_init is running
2016-08-25T22:46:18Z dougk: ok how about continue to load_core_file
2016-08-25T22:47:59Z attila_lendvai quit (Ping timeout: 260 seconds)
2016-08-25T22:48:14Z veckon: yep
2016-08-25T22:48:42Z dougk: yep meaning ok? then continue to create_initial_thread
2016-08-25T22:49:10Z veckon: Yes, okay.
2016-08-25T22:49:59Z dougk: hm. how about put a breakpoint at initial_thread_trampoline and continue
2016-08-25T22:51:14Z dougk: and is this threads or no threads?
2016-08-25T22:51:26Z veckon: I've tried both but this one is threads
2016-08-25T22:52:53Z veckon: Now I'm just falling through the repl
2016-08-25T22:53:07Z dougk: so it's magically working?
2016-08-25T22:53:17Z veckon: I can't explain it
2016-08-25T22:54:05Z veckon: actually don't we need to dump the core?
2016-08-25T22:54:15Z veckon: to output/sbcl.core
2016-08-25T22:54:15Z dougk: and so it printed something like "COLD-INIT... " (PACKAGE="") and "Length(TLFs)" ?
2016-08-25T22:54:19Z veckon: Yes
2016-08-25T22:54:21Z dougk: no, it hasn't finished building
2016-08-25T22:54:39Z dougk: next thing is to emulate the contents of make-target-2.sh by hand
2016-08-25T22:54:58Z dougk: try (load "make-target-2.lisp") and that might die right there if it's reproducible
2016-08-25T22:55:23Z veckon: That appeared to work
2016-08-25T22:56:00Z stassats joined #sbcl
2016-08-25T22:56:11Z dougk: wth. so the next next thing is get back to the shell and just do the lines from make-target-2.sh starting at "echo // doing warm init - load and dump phase"
2016-08-25T22:56:29Z dougk: but not the echo, just the src/runtime/sbcl obviously
2016-08-25T22:56:36Z veckon: yeah
2016-08-25T22:56:48Z veckon: should I load make-target-2-load.lisp?
2016-08-25T22:56:57Z dougk: it exits before reloading
2016-08-25T22:57:01Z dougk: so that wouldn't be the same
2016-08-25T22:57:06Z veckon: Ah okay
2016-08-25T22:57:48Z milanj quit (Quit: This computer has gone to sleep)
2016-08-25T22:59:40Z veckon: That does not work
2016-08-25T23:00:05Z veckon: Should I attempt to load it through lldb again?
2016-08-25T23:00:10Z dougk: and what did it do - suck up all memory again before printing so much as "COLD-INIT..." ?
2016-08-25T23:00:13Z veckon: yes
2016-08-25T23:00:29Z veckon: It just explodes after the welcome message
2016-08-25T23:00:36Z veckon: nothing else, no COLD-INIT
2016-08-25T23:00:41Z dougk: wow. so I'd go back to printf-style debugging, because clearly lldb is affecting the thing under examination
2016-08-25T23:00:50Z veckon: It gets further in lldb somehow
2016-08-25T23:01:49Z dougk: open up runtime.c and sprinkle it with fprintf(stderr, "Got here 1\n"); ... "Got here 2\n";  Apply liberally.
2016-08-25T23:02:07Z DavidGu joined #sbcl
2016-08-25T23:02:12Z veckon: sure
2016-08-25T23:02:42Z dougk: then to repeat the build as quickly as possible, do at the shell "XC_LISP=/path/to/your/bootstrap/lisp sh slam.sh"
2016-08-25T23:03:01Z dougk: what's the bootstrap lisp - also sbcl ?
2016-08-25T23:03:13Z veckon: yes
2016-08-25T23:03:41Z veckon: thanks for your help by the way!
2016-08-25T23:04:00Z veckon: sprinking liberally now
2016-08-25T23:04:00Z dougk: oh, forgot one thing. If you didn't add ':sb-after-xc-core' to your features you can't slam it.
2016-08-25T23:04:02Z dougk: you're welcome
2016-08-25T23:04:18Z dougk: so don't bother sprinkling until rebuilding from top if you didn't have that feature which I suspect you didn't
2016-08-25T23:04:26Z veckon: Alright
2016-08-25T23:04:38Z dougk: see explanation in 'base-target-features.lisp-expr'
2016-08-25T23:05:08Z veckon: kk I'll just rebuild it with that feature first
2016-08-25T23:05:32Z veckon: would fancy include that feature?
2016-08-25T23:06:07Z dougk: not sure
2016-08-25T23:06:07Z dougk: i don't think so
2016-08-25T23:06:19Z dougk: oh, it does
2016-08-25T23:06:20Z dougk: sorry
2016-08-25T23:06:27Z veckon: ah I might be able to slam then
2016-08-25T23:09:14Z veckon: didn't work so I'm rebuilding.
2016-08-25T23:09:19Z veckon: with the feature explicitly there
2016-08-25T23:09:40Z dougk: what did it say? no after-xc core or something else ?
2016-08-25T23:09:43Z veckon: es
2016-08-25T23:09:45Z veckon: yes*
2016-08-25T23:09:48Z veckon: no after-xc core
2016-08-25T23:10:38Z veckon: okay finished, and I already started sprinking the printfs
2016-08-25T23:10:38Z veckon: so
2016-08-25T23:10:42Z veckon: before the header gets printed
2016-08-25T23:10:45Z veckon: I see a few things
2016-08-25T23:10:59Z veckon: https://gist.github.com/aamedina/8c4048b950d31f949673f81a574f8839
2016-08-25T23:11:24Z dougk: ok, go crazy with that. I have to catch a train
2016-08-25T23:11:27Z veckon: kk
2016-08-25T23:11:48Z veckon: thanks again!
2016-08-25T23:12:42Z stassats: could you get a backtrace? when the heap starts growing
2016-08-25T23:13:02Z veckon: No, I can't. I can't do anything interactively with the process.
2016-08-25T23:13:12Z stassats: not even attach gdb?
2016-08-25T23:13:16Z stassats:  /lldb
2016-08-25T23:13:19Z veckon: Yes I can do that
2016-08-25T23:13:30Z veckon: but it looks like I can get further with LLDB
2016-08-25T23:13:38Z veckon: so it might be a bit of a red herring
2016-08-25T23:13:45Z veckon: Not sure, if you have ideas I'll do whatever you want :)
2016-08-25T23:14:54Z stassats: once you attach *db, do call ldb_monitor()
2016-08-25T23:15:00Z veckon: okay
2016-08-25T23:15:32Z stassats: there is away to get backtraces, but i don't remember the incantations
2016-08-25T23:15:48Z stassats: so just do it from ldb_monitor
2016-08-25T23:16:05Z stassats: it'll probably crash anyhow, but worth a try
2016-08-25T23:19:08Z veckon: can't seem to run it
2016-08-25T23:20:15Z stassats: but does an already built binary run?
2016-08-25T23:20:26Z veckon: At least 1.2.1 does
2016-08-25T23:22:13Z stassats: can you build 1.2.1?
2016-08-25T23:22:22Z veckon: I'll check
2016-08-25T23:22:40Z stassats: it's strange that warm init fails yet a build image works
2016-08-25T23:23:52Z veckon: compiling it now.
2016-08-25T23:27:07Z veckon: Looks successful
2016-08-25T23:27:17Z stassats: well, if you could bisect that
2016-08-25T23:27:19Z veckon: Maybe I'll try compiling another version
2016-08-25T23:27:23Z veckon: and then bisecting a closer one
2016-08-25T23:28:12Z stassats: don't think i want to upgrade to macOS just yet
2016-08-25T23:28:16Z veckon: hehe
2016-08-25T23:28:24Z veckon: It broke my favorite Lisp :(
2016-08-25T23:28:36Z veckon: But I needed to for work >_<
2016-08-25T23:28:53Z stassats: i wonder if it can be launched in a vm
2016-08-25T23:29:50Z stassats: too much hassle, it seems
2016-08-25T23:30:22Z veckon: thanks for your help btw, going to do a bisect quickly and if its useless then I'll try compiling something more recent
2016-08-25T23:30:25Z veckon: and then trying again
2016-08-25T23:30:29Z nyef: I'm hoping to upgrade to MacOS 8. Maybe 9. d-:
2016-08-25T23:30:47Z stassats: nyef: whoa, slow down there!
2016-08-25T23:31:28Z nyef: System 6.0.7, with multifinder?
2016-08-25T23:31:41Z stassats: veckon: i can send a SBCL 1.3.8.31-ed664e9 binary your way to see if it can start
2016-08-25T23:31:48Z veckon: Sure
2016-08-25T23:31:53Z veckon: yeah 1249 revisions to test
2016-08-25T23:31:55Z veckon: hehe
2016-08-25T23:39:31Z stassats: ok, on your half built image can you do ./src/runtime/sbcl --core output/cold-sbcl.core
2016-08-25T23:40:08Z veckon: currently building a different release, will try soon.
2016-08-25T23:42:01Z DougNYC joined #sbcl
2016-08-25T23:44:14Z veckon: well, looks like I can build up to 1.2.16, but not 1.3.0.
2016-08-25T23:45:58Z stassats quit (Ping timeout: 265 seconds)
2016-08-25T23:55:09Z stassats joined #sbcl