|
02:26
FROGGS_ joined
02:33
cognominal joined
|
|||
| dalek | arVM/tommy-hash: adfdd67 | diakopter++ | / (2 files): build it righter |
02:41 | |
| arVM/tommy-hash: 6381236 | diakopter++ | build/Makefile.in: even more righter |
|||
| arVM/tommy-hash: 41d30df | diakopter++ | 3rdparty/tommy/tommytypes.h: MSVC needs this... |
|||
| arVM/tommy-hash: 627db6d | diakopter++ | src/6model/reprs/KnowHOWREPR.c: hash needs to do some initialization now, but luckily it doesn't allocate anything else |
|||
|
02:42
vendethiel joined
02:47
ilbot3 joined
|
|||
| diakopter | jnthn: any idea how to make moarvm dump a HLL callstack while debugging in Visual Studio | 02:59 | |
|
04:32
stmuk_ joined
|
|||
| nwc10 | good *, #moarvm | 07:29 | |
|
07:55
llfourn joined
08:26
timotimo joined
|
|||
| timotimo | diakopter: print MVM_dump_stacktrace(tc) in any frame that has a tc | 09:09 | |
|
09:39
virtualsue joined
10:18
kjs_ joined
10:49
kjs_ joined
11:52
kjs_ joined
12:14
cognominal joined
12:45
kjs_ joined
13:15
kjs_ joined
13:19
kjs_ joined
13:36
leont joined
15:00
llfourn joined
16:01
llfourn joined
16:22
virtualsue joined
|
|||
| dalek | arVM: be41cae | jnthn++ | docs/ChangeLog: ChangeLog for 2015.12. |
16:26 | |
| nwc10 | jnthn: ++ Copy callsites when inliing | 16:30 | |
| also, arguably, s/unicode/Unicode/g | 16:31 | ||
| dalek | arVM: 892d580 | jnthn++ | docs/ChangeLog: ChangeLog tweaks; nwc10++. |
16:32 | |
| nwc10 | ++ Add missing unsigned operand types; clean up unsinged int ops | ||
| :-) | |||
| jnthn | heh | ||
| timotimo | good-looking changelog :) | 16:33 | |
| nwc10 | ++ Somewhat cheating unsinged parameter ops. | ||
| timotimo | inliing is a bit like skiiing? | ||
| jnthn | Change log was mostly derived directly from git log --oneline --reverse :) | ||
| nwc10 | I think you need to neuralise your spellchecker excpetions | ||
| jnthn | So all these errors are those of the original authors :P | ||
| dalek | arVM: 64d5884 | jnthn++ | docs/ChangeLog: Correct spellos; nwc10++. |
16:34 | |
| nwc10 | I can't spot any more | ||
| hoelzro | o/ #moarvm | 16:40 | |
| jnthn | o/ hoelzro | 16:41 | |
| hoelzro | o/ jnthn | ||
| timotimo | yo ro | ||
| hoelzro | o/ timotimo | 16:43 | |
| nwc10 | \o \o \o \o | 16:44 | |
| we weem to be out of balance | |||
| hoelzro | o/ nwc10 \o | 16:49 | |
| Moar has no guaranteed GC order, right? even if A has a pointer to B, you don't know which will be collected first? | 17:16 | ||
| I did some sleuthing and think I determined what the implemented order is, but I assume that's subject to change | |||
| JimmyZ | I think so | 17:17 | |
| jnthn | Yeah, you can't rely on such ordering | 17:18 | |
| hoelzro | k, thanks for confirming | 17:20 | |
|
17:20
Begi joined,
Begi left
|
|||
| hoelzro | speaking of the GC, I seem to remember there was something that causes a nursery GC or something upon a frame pop, but I couldn't find that in the code when I looked. Did I dream that? | 17:21 | |
| jnthn | Surely | 17:22 | |
| That'd be hideously expensive. | |||
| hoelzro | ok, it was a dream then =) | 17:27 | |
|
18:28
uruwi joined
19:05
virtualsue joined
|
|||
| uruwi | Hello? | 19:17 | |
| timotimo | hi | 19:18 | |
| uruwi | Hi again | ||
| Just lurking around about the mutex crash. | 19:19 | ||
| timotimo | OK | 19:21 | |
| yeah, i've got nothin' | |||
| uruwi | Yeah, some bugs aren't easy to fix. | 19:37 | |
| timotimo | like the run_alt method not found problem | 19:38 | |
|
20:48
virtualsue joined
|
|||
| TimToady | can we notice we didn't find run-alt, and re-install it somehow? | 20:59 | |
| as a workaround? | |||
| or is that just the tip of the iceberg? | |||
| jnthn | TimToady: Tip of the iceberg, I fear | 21:00 | |
| And we'd not be in a place to really | |||
| TimToady | well, full speed ahead then :) | ||
| jnthn | Got the final bit of travel tomorrow, which will hopefully make me less sick than the bit I did a couple of days ago. So mebbe I'll be able to find it tomorow once that's done. | 21:01 | |
| TimToady | well, Merry Christmas Eve Eve | 21:02 | |
| jnthn | :) | ||
| timotimo | TimToady: i think what nine has discovered is that it's not grabbing an NFA object at all, but a "random NQPArray with always different numbers of values" | 21:03 | |
| jnthn | Yeah...it's a corruption issue | ||
| From the description | 21:04 | ||
| I failed after an hour's trying to recreate it locally; nine++ left me a new hint in the backlog, so I can try that tomorrow | |||
| TimToady: Did the "as Foo" alias for coercion syntax go away, and if not should it? | 21:11 | ||
| Seems we still got it... | |||
| lizmat | it was recently removed from panda, afaik | 21:18 | |
| jnthn | OK. Not sure I want to carry it forward... | 21:24 | |
| lizmat | fwiw, there is nu use of "as" in the core anymore | ||
| jnthn | aye | 21:26 | |
| TimToady | as far as I'm concerned, it can go | ||
| jnthn | k | 21:27 | |
| lizmat | double checked panda, no " as " in there | 21:31 | |
| jnthn | ok | 21:33 | |
| TimToady | so what could possibly go worng? | ||
| jnthn | I've got a patch locally to rip it out | ||
| And it's gone | 21:46 | ||
| rest time :) | |||
| diakopter | timotimo: my problem is: something is corrupting the call stack and related, such that debugging isn't working | 22:46 | |
| bah | |||
| timotimo | something? it's probably just the jit | 22:47 | |
| the jit pretty much always makes the stack unreadable to gdb | |||
| diakopter | jit is disabled | 22:50 | |
| this is MSVC | |||
| timotimo | oh | 22:54 | |
| -fno-omit-frame-pointer? | |||