|
github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm Set by AlexDaniel on 12 June 2018. |
|||
|
00:00
reportable6 left
00:01
reportable6 joined
01:39
fake_space_whale joined
03:45
fake_space_whale left
06:00
reportable6 left
06:04
reportable6 joined
09:49
domidumont joined
10:00
chloekek joined
10:26
fake_space_whale joined
10:28
Kaiepi left
10:29
Kaiepi joined
11:52
chloekek left
12:00
reportable6 left
12:05
reportable6 joined
12:16
vesper11 joined
12:20
vesper joined
12:21
vesper11 left
12:42
kiwi_21 joined
12:44
Ven`` joined
12:53
chloekek joined
12:54
kiwi_21 left
13:00
Kaiepi left,
Kaiepi joined
13:03
Kaiepi left,
Kaiepi joined
13:09
pamplemousse joined
13:32
tellable6 joined
13:35
lizmat joined
13:48
evalable6 left
13:49
evalable6 joined
14:04
Ven`` left
14:17
Ven`` joined
14:22
lucasb joined
|
|||
| Geth | MoarVM: b26638ebc1 | cygx++ | 2 files add MVM_vm_run_bytecode() as alternative to MVM_vm_run_file() |
14:26 | |
| MoarVM: ff4f6031d2 | cygx++ | src/moar.c factor out run_deserialization_frame() to avoid code repetition |
|||
| MoarVM: dfa5c30d02 | (Jonathan Worthington)++ (committed using GitHub Web editor) | 2 files Merge pull request #1159 from cygx/run-bytecode add MVM_vm_run_bytecode() as alternative to MVM_vm_run_file() |
|||
|
14:40
chloekek left
14:49
evalable6 left
14:52
evalable6 joined
14:55
brrt joined
|
|||
| brrt | \o | 14:56 | |
| pamplemousse | o/ | ||
| timotimo | \o | 14:57 | |
| brrt | ohai | 15:11 | |
| brrt is wondering, whether to tackle linear IR now | |||
|
15:25
zakharyas joined
|
|||
| Geth | MoarVM/jit-expr-float: 29 commits pushed by (Bart Wiegmans)++ review: github.com/MoarVM/MoarVM/compare/0...56991bb600 |
15:35 | |
| brrt | I kind of want to merge the floating point support. Any objections? | ||
| timotimo | not from me! | 15:40 | |
| brrt | lets see if we get objections from CI | ||
|
15:52
evalable6 left
15:56
evalable6 joined
16:11
dogbert17 joined
16:15
Ven`` left
|
|||
| brrt | timotimo: do we have a stupid debug printer for MVMObject? | 16:15 | |
| tellable6 | brrt, I'll pass your message to timotimo | ||
| timotimo | we have a stupid debug printer for P6opaque | 16:16 | |
| tellable6 | 2019-08-18T16:15:41Z #moarvm <brrt> timotimo: do we have a stupid debug printer for MVMObject? | ||
| timotimo | but not for MVMObject in general | ||
| brrt | hmmm | ||
| timotimo | oh, do you mean a prettyprinter for gdb? | ||
| a plugin? | |||
| brrt | nah, just something I can do instead of fprintf | ||
| timotimo | oh, like inside your own code | 16:17 | |
| brrt | yeah | ||
| timotimo | there's surely something that could be done based on the code that the debugserver uses to tell the debug client what an object is | ||
| brrt | hmmm | 16:26 | |
| I can use the debug_name | |||
| also, is there a way to make a typed array in nqp? | 16:27 | ||
| an array with slot_type = MVM_ARRAY_I64 | |||
| timotimo | sure, the nqp test suite has some of those | ||
| also any code that tries to encode/decode string data | 16:28 | ||
| brrt | yay | 16:35 | |
| I found it | |||
| nqp::list_i() does the trick :-) | |||
| timotimo | oh! of course! | ||
| brrt | ok, I'm afk for a bit | ||
| I stole your devirt code from graph.c timotimo :-) | |||
| timotimo | i have been robbed! | 16:36 | |
| brrt | and am now integrating it into the epxr jit | ||
.oO( Intellectual property is intellectual theft? ) |
|||
| timotimo | i'm glad that exprjit gets a bit of the devirt love :) | ||
| Geth | MoarVM: 5c787f3782 | (Timo Paulssen)++ | src/spesh/optimize.c Only getlexperinvtype ā constant if only one type record exists It only appeared in profiling, and only rarely, but somehow a getlexperinvtype had gotten multiple wrong types, the first of which turns out to be NQPMu, which is definitely wrong. This gets rid of that crash, but the underlying question why those types were recorded in the first place isn't solved. |
16:48 | |
| MoarVM: 24e2986a1d | (Timo Paulssen)++ | src/gc/orchestrate.c Do heapsnapshot/instrumented data gathering earlier in GC The other place it was in allowed other threads to continue running while the main thread was doing the data gathering for instrumented profiles or the heapsnapshot profiler, which caused quite some confusion and some crashes. |
|||
| timotimo | old-ish work, finally committed | ||
|
16:56
evalable6 left
16:59
evalable6 joined
|
|||
| dogbert17 | timotimo++ | 17:03 | |
| timotimo | okay, just a few days old | 17:05 | |
| dogbert17 | vrurg might be interested, I remember him mentioning a SEGV when trying to heap profile something | 17:06 | |
| MasterDuke | timotimo: does that fix profiling of the mqtt tests? | 17:11 | |
| tellable6 | MasterDuke, I'll pass your message to timotimo | ||
| timotimo | it fixes one bug that only sometimes happened | 17:14 | |
| tellable6 | 2019-08-18T17:11:59Z #moarvm <MasterDuke> timotimo: does that fix profiling of the mqtt tests? | ||
| timotimo | how did the mqtt tests usually fail? | 17:15 | |
| MasterDuke | SEGV after the tests finished, but before it wrote out the profile, `MoarVM panic: Internal error: invalid thread ID 60 in GC work pass` during the tests, or that NQPMu thing which i think you just fixed | 17:16 | |
| it's always one of those, i've never seen a profile actually completed | 17:18 | ||
| oh, i did just get a segv during the tests | 17:19 | ||
|
17:37
brrt left
17:45
camelia left
17:47
camelia joined
|
|||
| dogbert17 | timotimo: could the profiling problem with mqtt also be spesh related, i.e. that the profiling have pointer to spesh related stuff which are no longer valid? | 17:51 | |
| tellable6 | dogbert17, I'll pass your message to timotimo | ||
| timotimo | doesn't sound too likely, i don't think | ||
| tellable6 | 2019-08-18T17:51:02Z #moarvm <dogbert17> timotimo: could the profiling problem with mqtt also be spesh related, i.e. that the profiling have pointer to spesh related stuff which are no longer valid? | ||
| dogbert17 | there tends to be an 'MVM_spesh_deopt_one' in the backtrace | 17:52 | |
| timotimo | with jit turned off? | ||
| dogbert17 | yes | 17:53 | |
| it looks like this: gist.github.com/dogbert17/8ad7bc49...43ad399bc5 | 17:56 | ||
| timotimo | OK, that's in the middle of running i believe? | 17:57 | |
| like, that test file has like 500 tests | |||
| dogbert17 | yes, the nursery has been made smaller and MVM_GC_DEBUG=1 | 17:58 | |
|
18:00
reportable6 left
18:03
reportable6 joined
|
|||
| MasterDuke | timotimo: in case you have notifications disabled, i added a bunch of new backtraces here github.com/MoarVM/MoarVM/issues/1023 | 18:34 | |
| tellable6 | MasterDuke, I'll pass your message to timotimo | ||
|
18:36
domidumont left,
evalable6 left
18:37
evalable6 joined
18:56
evalable6 left
|
|||
| timotimo | urgh | 19:07 | |
| tellable6 | 2019-08-18T18:34:23Z #moarvm <MasterDuke> timotimo: in case you have notifications disabled, i added a bunch of new backtraces here github.com/MoarVM/MoarVM/issues/1023 | ||
| MasterDuke | useful at all? | 19:10 | |
| timotimo | not sure, but perhaps not very | ||
|
19:31
zakharyas left
|
|||
| dogbert17 | turning off spesh does indeed generate a profile ... 24 megs in size | 20:38 | |
|
20:40
Kaiepi left
20:47
Kaiepi joined
20:48
Kaypie joined
21:10
sena_kun joined,
pamplemousse left
21:12
sena_kun left
22:03
fake_space_whale left
22:19
coverable6 left
22:22
coverable6 joined
22:48
coverable6 left
22:55
coverable6 joined
22:57
coverable6 left
22:59
coverable6 joined
23:00
coverable6 left
23:06
coverable6 joined
23:08
coverable6 left
23:12
coverable6 joined,
coverable6 left
23:13
coverable6 joined
23:15
coverable6 left
23:17
coverable6 joined
23:27
coverable6 left
23:28
coverable6 joined
|
|||