| IRC logs at
Set by AlexDaniel on 12 June 2018.
00:02 reportable6 left 00:05 reportable6 joined 01:05 benchable6 left, nativecallable6 left, linkable6 left, sourceable6 left, evalable6 left, quotable6 left, coverable6 left, greppable6 left, notable6 left, unicodable6 left, statisfiable6 left, committable6 left, bisectable6 left, squashable6 left, tellable6 left, releasable6 left, bloatable6 left, shareable6 left, statisfiable6 joined, coverable6 joined, bisectable6 joined 01:06 notable6 joined, linkable6 joined, nativecallable6 joined, evalable6 joined, tellable6 joined, unicodable6 joined 01:07 releasable6 joined, shareable6 joined, benchable6 joined, committable6 joined, quotable6 joined 01:08 squashable6 joined, bloatable6 joined, sourceable6 joined, greppable6 joined
[Coke] - didn't someone do this? 01:30
01:52 sxmx1 left, Voldenet left 01:53 sxmx1 joined, Voldenet joined, tobs joined, BinGOs joined, jraspass joined
[Coke] does some ticket maintenance on MoarVM but can't actually fix anything. 02:11
nwc10: would it be helpful to have a ticket label "porting" or some such?
03:05 lucasb left 03:57 frost-lab joined 05:54 geth_ joined
geth_ MoarVM/master: 13 commits pushed by (Nicholas Clark)++
05:57 Geth left 06:02 reportable6 left 06:03 reportable6 joined
japhb Are there any GC tunables available from up at the Raku level other than `VM.request-garbage-collection` (AKA nqp::force_gc)? 06:25
I'm facing occasional GC hiccups in my server that jump from 1-5ms to 100-150ms, which is well into the territory of "consciously noticeable by humans". I'm wondering if there's anything I can do to alter that for latency apps. 06:27
(Testing generally indicates humans prefer medium-but-steady performance to high-but-unpredictable performance for anything involving human reaction times. Of course, truly steadiness won't make up for truly poor performance, but that's a whole different problem.) 06:29
*Of course, steadiness 06:30
It would be nice to be able to tune steadiness versus throughput for the GC behavior, but I understand that's a whole pile of research of its own. 06:31
06:51 domidumont joined
MasterDuke japhb: i don't think there's anything else available from raku, and not much in moarvm either that could be exposed. definitely not at the maturity of the jvm in that regard 07:30
japhb Dang. Well, I knew it was a long shot, but I figured I'd ask in case I missed something. 07:34
07:34 zakharyas joined 07:57 dogbert17 left 08:03 dogbert17 joined 08:06 brrt joined
nwc10 good *, brrt 08:15
08:17 sena_kun left 08:18 sena_kun joined
brrt good * nwc10 08:23
08:43 geth_ left, Geth joined
nine A small step for mankind, but a big hop for a water tank :) 10:33
10:34 evalable6 left
lizmat yup... now only to not have this fire after landing :-) 10:34
not an issue on Mars or the Moon, but pesky when getting back to Earth, I'd say 10:35
10:36 evalable6 joined
brrt (what rocket are we talking about today?) 10:51
nwc10 -- In general, this has been followed by a dramatic explosion as one aspect or another of the incredibly complex series of events required doesn't work quite right. The biggest exception was one case where that explosion waited for several minutes after the rocket's landing. Today's launch followed the script right 10:52
up to the landing, at which point everything changed. The landing went off without a hitch this time, and the hardware stayed intact—albeit on fire—well after the landing.
brrt ooh, nice 10:55
11:23 zakharyas left
timotimo japhb: have you considered grabbing a heap snapshot or two, or looking at allocation statistics in the profiler's output? 11:23
11:26 MasterDuke left
timotimo most helpful is that the profiler will count up for each type how many of them were freed by a minor or by a major collection 11:28
types that get collected majorily in minor collections contribute to other objects becoming considered "older" at shorter lifetimes 11:29
11:36 brrt left 12:02 reportable6 left 12:04 reportable6 joined 13:32 dogbert11 joined 13:35 dogbert17 left 13:36 frost-lab left 14:29 dogbert17 joined 14:30 brrt joined 14:32 dogbert12 joined 14:33 dogbert11 left 14:34 dogbert17 left 14:45 dogbert17 joined 14:48 dogbert12 left 14:52 dogbert11 joined 14:55 dogbert17 left 15:42 zakharyas joined
tbrowder [Coke]: ref issue #1492, I added a partial log showing the coderef warnings and line ref 15:44
16:15 Kaiepi left 16:17 Kaiepi joined, Kaiepi left 16:18 Kaiepi joined 16:25 domidumont left 16:39 Kaiepi left 17:12 brrt left 17:45 patrickb joined 17:51 brrt joined 17:59 zakharyas left 18:01 reportable6 left 18:04 reportable6 joined 18:21 brrt left
japhb timotimo: I'll look into both of those. Yesterday I needed to feel more success and less yak shaving, so I did actual MUGS-level work. :-) 18:58
timotimo i think i already asked you before but i've forgotten what MUGS is 19:06
19:07 cog_ joined 19:10 cog left
dogbert11 there's a flapper lurking in t/spec/S03-operators/repeat.t 19:12
test 56 the second part: "" 19:13
19:20 MasterDuke joined 19:38 dogbert17 joined
japhb timotimo: and #mugs here on Freenode 19:38
Happy to answer any questions. :-)
19:42 dogbert11 left 19:46 dogbert17 left 20:03 zakharyas joined 20:09 dogbert17 joined, dogbert11 joined 20:13 dogbert17 left 20:19 dogbert11 left
timotimo what games are implemented there? 20:19
20:20 dogbert11 joined 20:26 brrt joined 20:27 zakharyas left 20:29 dogbert11 left, MasterDuke left 20:32 dogbert11 joined, patrickb left
japhb timotimo: So far, pretty minor stuff, just to feel out the APIs. Guessing games, tic-tac-toe, that sort of thing. One of the other Rakoons (raydiak?) asked if I would put some more energy into action games, to fill out more of those pieces of the API. 20:53
That's when I discovered that JSON wasn't going to cut it any more and wrote CBOR::Simple (and TinyFloats, and Cro::CBOR) to have a more efficient (and faster to validate on message receipt) encoding. 20:54
So while I wait for the next Cro release, I've been working on server-side generated particle effects that are rendered client-side, in order to test sending game state updates with hundreds of updates per frame. 20:55
21:24 Kaiepi joined 21:28 brrt left
[Coke] tbrowder: thanks 21:47
22:39 dogbert17 joined 22:42 dogbert11 left
tbrowder [Coke]: you're welcome. and i just recompiled moarvm separately and i found a couple of warnings in 3rd party libs. i'll add those to the issue in a bit. 23:21
fixing them is beyond my current c fu 23:22
[Coke] might be worth keeping the 3rd party lib warnings separate since we can't as easily fix them and may have to report them upstream 23:29
tbrowder yepper, i thought about that, then passed the buck for now :-( 23:51
i'll be fresher tomorrow...