00:25 boomteam joined 03:52 eternaleye joined 04:01 eternaleye joined 05:28 JimmyZ joined 05:33 JimmyZ joined
JimmyZ /home/jimmy/rakudo/../MoarVM/bin/moar --libpath="/home/jimmy/rakudo/../MoarVM/languages/nqp/lib" perl6.moarg 07:44
Stage start : 0.000
Heap corruption detected: pointer 0x7faa75ca2880 to past fromspace
jnthn: ^^
timotimo moarg :o 07:45
07:46 FROGGS joined
jnthn JimmyZ: If we're lucky, it'll be one of the already-found-but-not-fixed bugs. 07:47
And even if it isn't, I suspect the auto-whacker will get tthere
timotimo i looked into capsicum yesterday (just watched the presentation, actually), so i was wondering if moarvm or rakudo in general could have some super nice capability management facilities 07:49
that would be pretty darn cool
otherwise, i suppose it'll be doable enough to just use the native syscalls
JimmyZ good morning, jnthn 08:03
FROGGS timotimo: reconfiguring helped, thanks :o) 08:05
timotimo there really ought to be something that automatically tells you to do that, would get rid of so much confusion 08:06
JimmyZ auto-whacker?
timotimo nwc10++ made a tool to automatically try re-building with ever-decreasing nursery sizes 08:09
JimmyZ oh, I saw it. nice tool. 08:11
it's nicer to add to MoarVM by some configure options 08:13
nwc10 jnthn: I have another, but not diagnosed it yet. I guess this is stating the obvious, but if Win32 lets one do the same memory permissions tricks, your laptop could join in the fun. 08:27
jnthn teaching, bbl 08:29
nwc10: I probably *could*, but scaling things beyond me is always helpful :) 09:43
nwc10: And at the moment we seem to be achieving that ;)
nwc10 only sort of. I was hoping that we'd scale beyond me too (in that, I think that a couple of others on the channel understood where to fix the problems it can spot) 09:50
I was also starting to be suspicious that the handle SEGV is not a GC bug 09:52
09:54 FROGGS joined, lizmat joined, japhb_ joined, hoelzro joined, rurban joined
jnthn nwc10: The one I fixed last night? 09:58
nwc10 yes 10:21
as best I can tell, the GC bugs are deterministic, for deterministic input 10:22
the handle bug is only really fixed on Win32? Because it needs something that libuv doesn't yet have for *nix? 11:24
FROGGS or unix does not suffer from it 11:25
nwc10 (gdb) p handle
$1 = (uv_handle_t *) 0xa1e8964aa0964c80
moarvm at d3212c7860619ed957ed890e55b984ebcfeb1473 -- Unbust non-Windows build.
FROGGS I know that commit 11:26
11:58 tgt joined 13:20 cognominal joined
[Coke] do we want reports of build warnings? 13:32
anyone building nqp-m on os x? I just got a build failure. (seems to work fine on diakopter's linux box, though) 13:43
jnthn [Coke]: Build failure in NQP or in Moar? 13:58
13:59 jnap joined 14:05 lue joined 14:15 colomon_ joined 14:24 nickalexeyev joined 14:31 diakopter joined, ingy joined
nwc10 "to whom it may concern", I'm not currently finding (more) GC bugs, because I'm not able to fix the ones I did find. I was hoping someone else (and someone not jnthn) would be provide input there 14:36
so if you think you're waiting for me, you're deadlocked :-)
[Coke] jnthn: nqp, I think. 14:46
jnthn: gist.github.com/coke/7988126
15:20 camelia joined, flussence joined
FROGGS I'll have time tonight in $hotel 15:21
nwc10 cool
just curious - how far from $HOME is $hotel?
15:28 ingy joined
jnthn nwc10: Others are very welcome, though I may get a chance too...depends how late my trip to oelrepubliken goes... 15:30
15:38 lizmat_ joined 15:42 FROGGS joined 15:57 _ilbot joined 15:58 colomon joined 16:02 diakopte1 joined 16:09 colomon_ joined 16:12 dagurval_ joined 16:18 sorear_ joined, lee_ joined 16:21 moritz joined, camelia joined, ggoebel117 joined, rurban joined, hoelzro joined, japhb_ joined, lizmat_ joined, wsri joined, jnthn joined, krunen joined 16:25 lee__ joined 16:33 moritz_ joined 17:23 crimclub joined 18:36 FROGGS joined 18:47 mamaevao joined 18:49 cognominal__ joined 19:42 ssutch joined 19:59 ssutch joined
[Coke] Segmentation fault 20:27
make: *** [CORE.setting.moarvm] Error 139
(from today's rakudo-m build - I was getting to perl6-m on previous days. 20:28
(on host06)
nwc10 do you have a core dump? 20:30
(either way, it's probably a garbage pointer in ...->handle, which seems to be a non-deterministic bug, I think caused by thread timing races) 20:31
[Coke] nwc10: I don't see a core file, nope. 20:32
lue nwc10: I'm having a similar segfault issue, or at least did. Not sure how it stopped, all I know is it stopped after using gdb. 20:36
nwc10 I'm seeing the crash sometimes, and not other times 20:38
20:54 FROGGS joined
FROGGS nwc10: it would be 573km by car 20:56
jnthn: I sort of implemented $*OS and $*OSVER, but now I get "Cannot assign to immutable value" or some such... is that known? 20:58
that is when compiling Test.pm I think
the tests itself complain about prostcircumfix:<( )> or so 20:59
jnthn: btw, would it be okay if we take perl's $^O and put it in our generated config file? 21:11
hmmmm, or we do what parrot does..., trying to execute things like `/sbin/sysctl -n kernel.osrelease` 21:13
dunno how we get Windows's version
[Coke] `ver` ? 21:14
that gives me "Microsoft Windows [Version 6.1.7600]" on this win 7 box.
FROGGS and what do we put in $*OSVER ? 21:16
6.1.7600 ?
[Coke] I think that's not unreasonable.
since the other option is "Windows 7" or "Windows Server 2008 R2"
[Coke] points at wikipedia's Ver_(command) 21:17
FROGGS I think the latter is specced as DISTRO.name
ethercalc.org/wrwsbhvuim 21:18
I created that spreadsheet to imagine how these objects might look like from reading the spec
21:48 benabik joined 21:55 tgt joined
jnthn back 22:09
FROGGS: we can rely on a simple thing (shove $^O in config) for now. 22:11
22:19 FROGGS joined
jnthn So I spent the evening in oelrepubliken instead of fixing GC bugs. 22:27
FROGGS: See backlog if you missed my comment 22:29
FROGGS k
jnthn BTW, one of the sanity tests also produces a "could not assign to immutable value" error 22:31
FROGGS k, so it is not my fault :o) 22:32
(in theory)
jnthn No, doubt it 22:33
I didn't get chance to track it down yet
And tonight I...went to a place with a beer menu in the hundreds. :) 22:34
diakopter gah
jnthn Should have time tomorrow evening or in next couple of days, though :)
Tomorrow is last day of teaching for the year.
diakopter wonderful time to download next mac os x update, Apple... while I'm on plane wifi.
timotimo :D 22:35
jnthn In fact, tomorrow is last day of teaching until some time in 20th-ish of Jan. :)
timotimo just hope it doesn't crash during the installation
recovery could be a bit harder than usual in that case!
jnthn ...the plane, or the OSX update? :D
diakopter -_-
^ a tongue-in-cheek grimace ;) 22:36
jnthn :) 22:37
diakopter -_-p
FROGGS .oO( Operator, I need an exit! ) 22:38
diakopter draw one on the wall!
FROGGS hmmm... what movie was that? 22:39
diakopter Beetlejuice, and probably others 22:40
FROGGS looks like I need to watch it again :o)
diakopter or 3 times..
FROGGS yeah, I remember that 22:41
diakopter 2.8Mbps isn't bad.. 22:42
er, 3.2! 22:43
jnthn It's fast, but it's not full fart...
timotimo :D
FROGGS that is an awesome page: nadeausoftware.com/articles/2012/01...ing_system 22:44
jnthn useful ) 22:48
OK, sleep here...teaching tomorrow :) 22:53
'night
diakopter 'n
FROGGS gnight 22:55
./perl6-m -e 'say nqp::backendconfig()'
Unhandled exception: Cannot find method 'postcircumfix:<( )>'

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!