2015-02-02T00:02:53Z oleo quit (Read error: Connection reset by peer)
2015-02-02T00:03:40Z dougk_ joined #sbcl
2015-02-02T00:04:17Z oleo joined #sbcl
2015-02-02T00:06:52Z stassats quit (Ping timeout: 272 seconds)
2015-02-02T00:11:07Z gingerale quit (Ping timeout: 256 seconds)
2015-02-02T00:16:33Z jrm: I had another crash of stumpwm/sbcl, here is a backtrace: http://ftfl.ca/paste/stumpwm-crash.html
2015-02-02T00:17:18Z jrm: These crashes only happen after hours or even a day or two, so it I would like to get all relevant info out of it.
2015-02-02T00:17:52Z nyef: How on earth is THAT happening?
2015-02-02T00:18:31Z dougk_ quit (Ping timeout: 250 seconds)
2015-02-02T00:18:40Z jrm: Everything works fine with StuwmWM, but if I start with a (swank:create-server) I get lots of crashes.
2015-02-02T00:19:00Z jrm: Well, I should say lots, but it's not a matter of if, but when.
2015-02-02T00:19:47Z jrm: statssats mentioned something about restarting sbcl with sbcl with --lose-on-corruption
2015-02-02T00:21:30Z nyef: ... You're obviously on x86-64...
2015-02-02T00:22:38Z jrm: I'm on FreeBSD 10-STABLE AMD64.
2015-02-02T00:22:52Z nyef: Hrm.
2015-02-02T00:24:00Z nyef: What else were you doing at the time?
2015-02-02T00:24:32Z jrm: Nothing.  It often happens when I'm away from my computer and I come back.
2015-02-02T00:24:53Z jrm: ..and see that StumpWM crashed.
2015-02-02T00:26:13Z nyef: So... it might have something to do with your screensaver?
2015-02-02T00:26:36Z fe[nl]ix: IIRC on BSDs one has to explicitly disable SIGPIPE, otherwise you don't get EPIPE on a write(2)
2015-02-02T00:27:21Z nyef: Yeah, but this looks more like SIGTRAP.
2015-02-02T00:30:36Z jrm: It does also happen other times when I'm there, but there are no consistent patterns that I've recognized.
2015-02-02T00:30:49Z jrm: Is there anything else I can do to help track this down?
2015-02-02T00:31:57Z eudoxia quit (Quit: Leaving)
2015-02-02T00:33:48Z jrm: Or put another way, if I restart the crashed StumpWM/SBCL am I going to lose any helpful information.
2015-02-02T01:06:55Z DeadTrickster joined #sbcl
2015-02-02T01:12:02Z jrm: In case there is anything useful in the slime-events buffer, here it is: http://ftfl.ca/paste/slime-events.html
2015-02-02T01:14:01Z nyef: jrm: Can you attach with gdb and get a backtrace from that?
2015-02-02T01:15:32Z jrm: nyef: I don't see a core file and the StumpWM binary is still running, although the WM is dead.
2015-02-02T01:15:45Z nyef: Do you have a terminal open?
2015-02-02T01:15:53Z |3b|: stumpwm had been running for a while in that backtrace?
2015-02-02T01:15:53Z dougk_ joined #sbcl
2015-02-02T01:16:08Z |3b|: seems odd that it was in open-display if so
2015-02-02T01:16:25Z nyef: ... Good point.
2015-02-02T01:16:40Z jrm: |3b|: Yes, it was open for hours.
2015-02-02T01:16:58Z jrm: nyef: I got a terminal focused.
2015-02-02T01:17:32Z nyef: jrm: find the PID of the stumpwm process, gdb the binary, and use the "attach" command on the PID.
2015-02-02T01:18:34Z nyef: This is still completely weird, though. It's been running for ages, it's in open-display, and it's catching a bogus SIGTRAP...?
2015-02-02T01:18:58Z |3b|: was it run from a terminal or started directly from X?
2015-02-02T01:19:03Z nyef: Two of these things conflict, and the other just shouldn't be happening.
2015-02-02T01:19:50Z jrm: It was started via ~/.xsession which is symlinked to ~/.xinitrc which gets run from xdm.
2015-02-02T01:20:07Z |3b|: anything in .xsession-errors or whatever that file is called?
2015-02-02T01:21:06Z |3b| suspects some other error happens first, then it tries to reconnect and gets confused
2015-02-02T01:22:05Z jrm: |3b|:http://ftfl.ca/paste/stumpwm-xession-errors.html
2015-02-02T01:22:50Z |3b|: is that entire file, or just part that looked interesting?
2015-02-02T01:22:53Z dougk_ quit (Ping timeout: 264 seconds)
2015-02-02T01:24:03Z nyef: Okay, THAT'S a bit more believable.
2015-02-02T01:24:23Z jrm: |3b|: Just the part that looked interesting.  I can paste the entire file, but it's quite large, because Conkeror constants spews warnings.
2015-02-02T01:24:54Z jrm: I'm attached to the binary/pid.
2015-02-02T01:24:59Z jrm: ...with gdb.
2015-02-02T01:25:29Z dougk_ joined #sbcl
2015-02-02T01:26:07Z jrm: http://ftfl.ca/paste/gdb-stumpwm.html
2015-02-02T01:27:06Z nyef: It's too late, the actual error is the one in the xsession log, and stumpwm trashes anything to do with it and tries to reconnect.
2015-02-02T01:27:35Z jrm: d'oh
2015-02-02T01:28:18Z nyef: How about this, have stumpwm open in a dedicated terminal window with --lose-on-corruption, and wait for it to happen again?
2015-02-02T01:29:10Z nyef: That way you might be able to get some postmortem information before it gets wiped out.
2015-02-02T01:29:10Z jrm: Sounds good.  I'll be back. :)
2015-02-02T01:29:13Z nikki93_ quit (Remote host closed the connection)
2015-02-02T01:29:16Z nyef: We'll be waiting.
2015-02-02T01:32:34Z |3b|: jrm: might also try setting stumpwm::*top-level-error-action* to :break
2015-02-02T01:32:43Z jrm: OK
2015-02-02T01:34:15Z DeadTrickster quit (Ping timeout: 264 seconds)
2015-02-02T01:37:48Z jrm: So, after (in-package :stumpwm) I put (setf *top-level-error-action* :break) ?
2015-02-02T01:38:49Z |3b| would just (setf stumpwm::*top-level-error-action* :break) after loading it, but in-package should work too
2015-02-02T01:39:30Z |3b|: (assuming you meant doing that after loading it and not in some random stumpwm source file, in which case it would depend on whether the variable had been defined yet)
2015-02-02T01:45:10Z echo-area joined #sbcl
2015-02-02T01:46:29Z jrm: In ~/.xinitrc instead of running StumpWM, I put exec xterm.  When xterm started, I did % stumpwm > stumpwm.out 2>&1.  I started emacs and did a slime-connect and did the (setf stumpwm::*top-level-error-action* :break).
2015-02-02T01:48:52Z DeadTrickster joined #sbcl
2015-02-02T01:59:49Z dougk_ quit (Ping timeout: 264 seconds)
2015-02-02T02:00:00Z LiamH joined #sbcl
2015-02-02T02:01:02Z dougk_ joined #sbcl
2015-02-02T02:09:52Z DeadTrickster quit (Ping timeout: 245 seconds)
2015-02-02T02:14:04Z echo-are` joined #sbcl
2015-02-02T02:14:48Z echo-area quit (Read error: Connection reset by peer)
2015-02-02T02:19:52Z echo-are` is now known as echo-area
2015-02-02T02:32:28Z dougk_ quit (Ping timeout: 244 seconds)
2015-02-02T02:44:28Z nikki93_ joined #sbcl
2015-02-02T03:08:47Z nikki93_ quit (Remote host closed the connection)
2015-02-02T03:11:50Z nikki93_ joined #sbcl
2015-02-02T03:22:27Z dougk_ joined #sbcl
2015-02-02T03:26:11Z nikki93_ quit (Remote host closed the connection)
2015-02-02T03:27:55Z nikki93_ joined #sbcl
2015-02-02T03:28:22Z nikki93_ quit (Remote host closed the connection)
2015-02-02T03:39:15Z christoph_debian quit (Ping timeout: 276 seconds)
2015-02-02T03:49:22Z nyef quit (Quit: G'night all)
2015-02-02T03:51:56Z christoph_debian joined #sbcl
2015-02-02T03:56:06Z ehaliewicz quit (Ping timeout: 246 seconds)
2015-02-02T04:03:12Z dougk_ quit (Ping timeout: 245 seconds)
2015-02-02T04:30:29Z LiamH quit (Quit: Leaving.)
2015-02-02T05:01:53Z selat joined #sbcl
2015-02-02T05:48:11Z nikki93_ joined #sbcl
2015-02-02T06:07:49Z oleo quit (Quit: Verlassend)
2015-02-02T06:11:23Z pranavrc joined #sbcl
2015-02-02T06:11:24Z pranavrc quit (Changing host)
2015-02-02T06:11:24Z pranavrc joined #sbcl
2015-02-02T06:30:32Z scymtym_ joined #sbcl
2015-02-02T06:39:33Z edgar-rft quit (Ping timeout: 245 seconds)
2015-02-02T06:42:27Z nikki93_ quit (Remote host closed the connection)
2015-02-02T06:42:59Z nikki93_ joined #sbcl
2015-02-02T06:47:22Z nikki93_ quit (Ping timeout: 245 seconds)
2015-02-02T07:14:03Z gingerale joined #sbcl
2015-02-02T07:36:41Z Shinmera joined #sbcl
2015-02-02T07:50:55Z edgar-rft joined #sbcl
2015-02-02T07:52:53Z Quadrescence quit (Quit: This computer has gone to sleep)
2015-02-02T08:05:18Z angavrilov joined #sbcl
2015-02-02T08:07:59Z selat quit (Quit: Lost terminal)
2015-02-02T08:32:05Z adlai quit (Ping timeout: 250 seconds)
2015-02-02T08:44:41Z Cymew joined #sbcl
2015-02-02T08:46:12Z adlai joined #sbcl
2015-02-02T09:48:24Z ASau joined #sbcl
2015-02-02T10:23:04Z flip214: now that there's a quad-core raspberry pi coming, having threads on ARM would be even nicer...
2015-02-02T10:26:29Z selat joined #sbcl
2015-02-02T10:32:45Z selat quit (Ping timeout: 256 seconds)
2015-02-02T10:54:13Z momo-reina joined #sbcl
2015-02-02T10:56:07Z echo-area quit (Read error: Connection reset by peer)
2015-02-02T11:01:23Z scymtym_ quit (Ping timeout: 244 seconds)
2015-02-02T11:23:26Z eudoxia joined #sbcl
2015-02-02T11:54:40Z stassats joined #sbcl
2015-02-02T11:55:31Z stassats quit (Remote host closed the connection)
2015-02-02T11:56:42Z stassats joined #sbcl
2015-02-02T12:05:32Z attila_lendvai joined #sbcl
2015-02-02T12:05:32Z attila_lendvai quit (Changing host)
2015-02-02T12:05:32Z attila_lendvai joined #sbcl
2015-02-02T12:24:46Z pranavrc quit (Remote host closed the connection)
2015-02-02T12:31:05Z pranavrc joined #sbcl
2015-02-02T12:31:05Z pranavrc quit (Changing host)
2015-02-02T12:31:05Z pranavrc joined #sbcl
2015-02-02T12:33:33Z pranavrc quit (Remote host closed the connection)
2015-02-02T12:35:40Z pranavrc joined #sbcl
2015-02-02T12:52:12Z nyef joined #sbcl
2015-02-02T12:56:08Z stassats: solved the local-tail-call after call confusing the debugger
2015-02-02T12:56:19Z stassats: just making < into <= if the frame is not escaped
2015-02-02T13:01:25Z stassats: just to explain: (flet ((test () (break "a") (flet ((tail ())) (declare (notinline tail)) (tail)))) (test))
2015-02-02T13:01:42Z stassats: the debugger now is 0: ((FLET TEST)) instead of  0: ((FLET TAIL :IN TEST))
2015-02-02T13:01:53Z hlavaty joined #sbcl
2015-02-02T13:12:06Z momo-reina quit (Remote host closed the connection)
2015-02-02T13:13:05Z Cymew quit (Quit: Konversation terminated!)
2015-02-02T13:14:06Z Cymew joined #sbcl
2015-02-02T13:15:45Z pranavrc quit (Remote host closed the connection)
2015-02-02T13:17:45Z pranavrc joined #sbcl
2015-02-02T13:29:25Z scymtym: stassats: does debug.impure.lisp:local-tail-call-variables require backend changes to pass on x86?
2015-02-02T13:29:37Z stassats: i don't think so
2015-02-02T13:29:45Z stassats: but i didn't test it
2015-02-02T13:29:49Z scymtym: something is wrong then
2015-02-02T13:29:53Z scymtym: it currently fails
2015-02-02T13:30:39Z scymtym: can i test something on the x86 machine to narrow down the problem?
2015-02-02T13:31:05Z stassats: no, i'll deal with it
2015-02-02T13:31:53Z scymtym: thanks
2015-02-02T13:32:37Z nyef: Hooray for multilib systems?
2015-02-02T13:35:03Z stassats: scymtym: doesn't fail here
2015-02-02T13:36:35Z pranavrc quit
2015-02-02T13:39:37Z scymtym: stassats: i saw failures on the x86 build slave and a second x86 machine; build slave is here: https://ci.cor-lab.org/job/sbcl-master/839/ (rows are feature combinations, so it seems to be consistent across features)
2015-02-02T13:40:26Z stassats: oh, you're not talking about the latest commit?
2015-02-02T13:41:40Z scymtym: the link is to the earliest build (containing multiple commits) in which the test failed
2015-02-02T13:42:02Z stassats: when it tests today HEAD tell me if it fails
2015-02-02T13:42:13Z scymtym: oh, maybe i'm misunderstanding completely
2015-02-02T13:42:44Z scymtym: yeah, it is currently building
2015-02-02T13:51:05Z psilord1 quit (Quit: Leaving.)
2015-02-02T14:01:57Z LiamH joined #sbcl
2015-02-02T14:03:56Z scymtym: stassats: seems to be fixed. is it plausible that the "Show the right function in the backtrace from local tail calls." commit fixed it?
2015-02-02T14:04:06Z stassats: yes
2015-02-02T14:04:20Z stassats: see the new comment in the local-tail-call-variables test
2015-02-02T14:05:49Z scymtym: i see, sorry for the noise
2015-02-02T14:06:09Z stassats: more testing is always welcome
2015-02-02T14:06:15Z scymtym: still interesting that there has been a difference between x86 and x86_64, though
2015-02-02T14:08:19Z dougk_ joined #sbcl
2015-02-02T14:20:06Z dougk_ quit (Ping timeout: 252 seconds)
2015-02-02T14:20:43Z stassats: gah, inlined lambdas causing problems
2015-02-02T14:20:57Z stassats: lp 1412881
2015-02-02T14:20:57Z specbot: https://bugs.launchpad.net/bugs/1412881
2015-02-02T14:21:23Z stassats: this is caused by Y not being visible inside the inlined lambda
2015-02-02T14:21:50Z stassats: which is sure enough right, Y can't be referenced by the inlined lambda
2015-02-02T14:22:07Z stassats: but to the debugger it should look like it's alive
2015-02-02T14:24:47Z edgar-rft quit (Quit: experience expired by memory leak)
2015-02-02T14:25:43Z stassats: curiously enough, the lambda-home vars are visible
2015-02-02T14:25:49Z stassats: i.e. X and Z
2015-02-02T14:29:48Z stassats: which may actually explain another bug
2015-02-02T14:39:12Z oleo joined #sbcl
2015-02-02T14:42:09Z dougk_ joined #sbcl
2015-02-02T14:45:10Z attila_lendvai quit (Quit: Leaving.)
2015-02-02T15:05:15Z irsol quit (Ping timeout: 244 seconds)
2015-02-02T15:17:04Z psilord1 joined #sbcl
2015-02-02T15:27:28Z dougk__ joined #sbcl
2015-02-02T15:29:49Z dougk_ quit (Ping timeout: 264 seconds)
2015-02-02T15:30:48Z foom quit (Ping timeout: 245 seconds)
2015-02-02T15:31:57Z LiamH quit (Ping timeout: 256 seconds)
2015-02-02T15:38:39Z irsol joined #sbcl
2015-02-02T15:43:00Z foom joined #sbcl
2015-02-02T15:44:13Z irsol quit (Ping timeout: 255 seconds)
2015-02-02T15:45:28Z stassats: git a fix for that
2015-02-02T15:46:04Z stassats: many debugger improvements this release
2015-02-02T15:51:45Z DeadTrickster joined #sbcl
2015-02-02T15:53:47Z irsol joined #sbcl
2015-02-02T15:57:14Z selat joined #sbcl
2015-02-02T16:02:49Z Cymew quit (Ping timeout: 264 seconds)
2015-02-02T16:03:42Z irsol quit (Ping timeout: 264 seconds)
2015-02-02T16:08:20Z eudoxia quit (Quit: Leaving)
2015-02-02T16:15:35Z LiamH joined #sbcl
2015-02-02T16:16:33Z irsol joined #sbcl
2015-02-02T16:21:39Z irsol quit (Ping timeout: 264 seconds)
2015-02-02T16:39:06Z DeadTrickster quit (Ping timeout: 264 seconds)
2015-02-02T16:48:39Z irsol joined #sbcl
2015-02-02T16:54:02Z irsol quit (Ping timeout: 245 seconds)
2015-02-02T16:54:30Z DeadTrickster joined #sbcl
2015-02-02T16:58:00Z fitzsim quit (Ping timeout: 265 seconds)
2015-02-02T17:09:46Z fitzsim joined #sbcl
2015-02-02T17:24:15Z fitzsim quit (Read error: Connection reset by peer)
2015-02-02T17:25:30Z fitzsim joined #sbcl
2015-02-02T17:29:26Z fitzsim quit (Read error: Connection reset by peer)
2015-02-02T17:29:40Z fitzsim joined #sbcl
2015-02-02T17:38:22Z DeadTrickster quit (Ping timeout: 240 seconds)
2015-02-02T17:39:16Z fitzsim quit (Ping timeout: 245 seconds)
2015-02-02T17:40:10Z fitzsim joined #sbcl
2015-02-02T17:51:40Z fitzsim quit (Read error: Connection reset by peer)
2015-02-02T17:51:55Z fitzsim joined #sbcl
2015-02-02T17:54:15Z DeadTrickster joined #sbcl
2015-02-02T18:06:44Z Bicyclidine joined #sbcl
2015-02-02T18:09:38Z irsol joined #sbcl
2015-02-02T18:12:25Z andrewrorer117 joined #sbcl
2015-02-02T18:14:49Z irsol quit (Ping timeout: 264 seconds)
2015-02-02T18:27:40Z nikki93 joined #sbcl
2015-02-02T18:29:00Z psy_ quit (Read error: Connection reset by peer)
2015-02-02T18:29:09Z nikki93 quit (Remote host closed the connection)
2015-02-02T18:29:18Z psy_ joined #sbcl
2015-02-02T18:33:36Z nikki93 joined #sbcl
2015-02-02T18:34:01Z fridim_ joined #sbcl
2015-02-02T18:36:36Z nikki93 quit (Remote host closed the connection)
2015-02-02T18:37:51Z soggybre1d quit (Ping timeout: 264 seconds)
2015-02-02T18:38:30Z soggybread joined #sbcl
2015-02-02T18:39:31Z DeadTrickster quit (Ping timeout: 256 seconds)
2015-02-02T18:49:07Z irsol joined #sbcl
2015-02-02T18:54:37Z DeadTrickster joined #sbcl
2015-02-02T18:55:39Z irsol quit (Ping timeout: 250 seconds)
2015-02-02T19:07:37Z Bicyclidine quit (Ping timeout: 264 seconds)
2015-02-02T19:09:24Z irsol joined #sbcl
2015-02-02T19:18:57Z Bicyclidine joined #sbcl
2015-02-02T19:19:23Z eudoxia joined #sbcl
2015-02-02T19:20:13Z irsol quit (Ping timeout: 250 seconds)
2015-02-02T19:37:30Z nikki93 joined #sbcl
2015-02-02T19:38:29Z nikki93 quit (Read error: Connection reset by peer)
2015-02-02T19:38:59Z nikki93 joined #sbcl
2015-02-02T19:39:35Z DeadTrickster quit (Ping timeout: 256 seconds)
2015-02-02T19:40:18Z nikki93 quit (Remote host closed the connection)
2015-02-02T19:42:07Z nikki93 joined #sbcl
2015-02-02T19:46:46Z nikki93 quit (Ping timeout: 255 seconds)
2015-02-02T19:54:58Z DeadTrickster joined #sbcl
2015-02-02T20:05:20Z stassats: it seems like half of the debug info is not used
2015-02-02T20:09:43Z irsol joined #sbcl
2015-02-02T20:11:12Z nikki93 joined #sbcl
2015-02-02T20:12:26Z prxq joined #sbcl
2015-02-02T20:12:51Z nikki93 quit (Read error: Connection reset by peer)
2015-02-02T20:12:59Z nikki93 joined #sbcl
2015-02-02T20:17:30Z nikki93 quit (Remote host closed the connection)
2015-02-02T20:19:35Z irsol quit (Ping timeout: 246 seconds)
2015-02-02T20:19:37Z alchemis7 quit (Ping timeout: 255 seconds)
2015-02-02T20:27:47Z nikki93 joined #sbcl
2015-02-02T20:29:12Z nikki93 quit (Remote host closed the connection)
2015-02-02T20:39:23Z DeadTrickster quit (Ping timeout: 240 seconds)
2015-02-02T20:52:48Z irsol joined #sbcl
2015-02-02T20:55:30Z DeadTrickster joined #sbcl
2015-02-02T21:02:49Z irsol quit (Ping timeout: 264 seconds)
2015-02-02T21:06:56Z edgar-rft joined #sbcl
2015-02-02T21:12:41Z selat quit (Quit: Lost terminal)
2015-02-02T21:17:22Z irsol joined #sbcl
2015-02-02T21:23:27Z irsol quit (Ping timeout: 264 seconds)
2015-02-02T21:24:24Z irsol joined #sbcl
2015-02-02T21:27:52Z nikki93 joined #sbcl
2015-02-02T21:31:22Z stassats: managed to pry out the successors thing from the debug info
2015-02-02T21:31:28Z stassats: which wasn't used since 0.8.something
2015-02-02T21:33:04Z angavrilov quit (Remote host closed the connection)
2015-02-02T21:39:24Z DeadTrickster quit (Ping timeout: 246 seconds)
2015-02-02T21:42:05Z nikki93 quit (Remote host closed the connection)
2015-02-02T21:48:11Z nikki93 joined #sbcl
2015-02-02T21:54:58Z DeadTrickster joined #sbcl
2015-02-02T21:58:11Z stassats: saving about 50K on 3.7M of compiled fasls
2015-02-02T21:58:18Z stassats: with debug 2, that is
2015-02-02T21:59:29Z nikki93 quit (Remote host closed the connection)
2015-02-02T22:00:05Z nyef: So... about a percent and a half?
2015-02-02T22:00:22Z stassats: also makes for simpler code
2015-02-02T22:00:52Z nyef: Okay, that's a win.
2015-02-02T22:01:23Z stassats: i should update some debug info comments
2015-02-02T22:01:54Z stassats: or add some
2015-02-02T22:04:37Z stassats: and it seems that we do not use location kinds
2015-02-02T22:05:09Z stassats: all we encode is where do find variables, and when they are visible
2015-02-02T22:05:19Z stassats: and to which forms vops may belong
2015-02-02T22:05:49Z stassats: the actual type of locations doesn't matter
2015-02-02T22:06:40Z stassats: but i want to introduce a new kind of locations, float-exception-location, which encodes from which registers an exception may come
2015-02-02T22:07:52Z fridim_ quit (Ping timeout: 240 seconds)
2015-02-02T22:12:03Z stassats: clhs arithmetic-error-operands
2015-02-02T22:12:03Z specbot: http://www.lispworks.com/reference/HyperSpec/Body/f_arithm.htm
2015-02-02T22:12:04Z stassats: for this
2015-02-02T22:12:40Z stassats: but i will need to decode the debug-info when the error is created, which may slow things down
2015-02-02T22:13:21Z stassats: but i also think i don't need to create the usual debug structures
2015-02-02T22:16:06Z prxq quit (Remote host closed the connection)
2015-02-02T22:16:57Z stassats: what if we encode all trap information in debug functions?
2015-02-02T22:17:59Z stassats: would have been real nice with conditional traps
2015-02-02T22:19:11Z nyef: We should HAVE conditional traps on some platforms.
2015-02-02T22:19:34Z nyef: Hey, call site and return information that the runtime can use. Please?
2015-02-02T22:19:44Z nyef: Pretty please?
2015-02-02T22:19:50Z stassats: for what?
2015-02-02T22:20:06Z nyef: Pretty please with precise stack-scavenging and table-based unwind on top?
2015-02-02T22:20:36Z stassats: right now locations are only written with debug >= 2
2015-02-02T22:21:05Z stassats: isn't table unwind quite slow?
2015-02-02T22:21:16Z stassats: or is it just gcc?
2015-02-02T22:21:18Z pkhuong_: stassats: slow to unwind
2015-02-02T22:21:22Z pkhuong_: 0 overhead to setup
2015-02-02T22:21:29Z stassats: yes, that's what i meant
2015-02-02T22:21:42Z pkhuong_: it makes sense for unwind protect
2015-02-02T22:21:52Z pkhuong_: perhaps not so much for throw/return-from
2015-02-02T22:21:55Z nyef: Slower to unwind, but once we have it we can lose the binding stack.
2015-02-02T22:22:37Z nyef: Oh, there's an idea. Table-based locations for UWP targets only?
2015-02-02T22:23:01Z stassats: there used to be a thing called trace table
2015-02-02T22:23:10Z stassats: are a trace of trace table
2015-02-02T22:23:18Z stassats: it was supposed to be used for something like that
2015-02-02T22:23:51Z nyef: Yeah, had I gotten to the point of actual code for table-based unwind, I was planning on sticking the extra info in the trace table.
2015-02-02T22:24:59Z stassats: it can go into debug-fun
2015-02-02T22:27:11Z nyef: Just hope that nobody comes along with a post-process tool to torch all of the debug-funs.
2015-02-02T22:29:57Z stassats: it'll be their own fault
2015-02-02T22:30:13Z nyef: True.
2015-02-02T22:34:24Z dougk__ quit (Ping timeout: 265 seconds)
2015-02-02T22:34:31Z stassats: can't think of a solution to eslewhere-trap-is-actually-before-start-pc
2015-02-02T22:35:01Z stassats: chances of an interrupt there before the CC are low
2015-02-02T22:35:25Z stassats: so i guess i can just invalidate :debug-environment vars when the error is arg-count-error
2015-02-02T22:35:39Z stassats: not bullet-proof, but better than nothing
2015-02-02T22:39:47Z DeadTrickster quit (Ping timeout: 256 seconds)
2015-02-02T22:50:59Z eudoxia quit (Quit: Leaving)
2015-02-02T22:55:40Z DeadTrickster joined #sbcl
2015-02-02T23:01:04Z DeadTrickster quit (Ping timeout: 245 seconds)
2015-02-02T23:03:43Z dougk_ joined #sbcl
2015-02-02T23:04:59Z psilord1 quit (Quit: Leaving.)
2015-02-02T23:13:06Z dougk_: hey. surely TYPE= should be commutative, right?
2015-02-02T23:13:44Z dougk_: (type= (specifier-type '(not (satisfies flummox))) (specifier-type '(real 1d0 1d0))) => NIL and T but flip them and the answer is NIL and NIL
2015-02-02T23:15:00Z dougk_: the latter answer is right, because the type machinery can't be certain about anything pertaining to 'flummox'
2015-02-02T23:24:42Z Bicyclidine quit (Ping timeout: 264 seconds)
2015-02-02T23:33:23Z pkhuong_: dougk_: yes, it should be :\
2015-02-02T23:34:08Z dougk_ quit (Ping timeout: 245 seconds)
2015-02-02T23:36:07Z Shinmera quit (Quit: しつれいしなければならないんです。)
2015-02-02T23:36:11Z Bicyclidine joined #sbcl
2015-02-02T23:42:43Z nyef: Ugh. MORE type-system trouble? /-:
2015-02-02T23:54:09Z gingerale quit (Ping timeout: 265 seconds)
2015-02-02T23:58:55Z dougk_ joined #sbcl
2015-02-02T23:59:03Z Bicyclidine quit (Quit: quit)