Welcome to the main channel on the development of MoarVM, a virtual machine for NQP and Rakudo (moarvm.org). This channel is being logged for historical purposes.
Set by lizmat on 24 May 2021.
00:02 reportable6 left 00:03 reportable6 joined 00:37 nine left, nine joined 01:37 quotable6 left, benchable6 left, linkable6 left, committable6 left, notable6 left, squashable6 left, nativecallable6 left, unicodable6 left, reportable6 left, coverable6 left, bloatable6 left, shareable6 left, releasable6 left, tellable6 left, sourceable6 left, greppable6 left, statisfiable6 left, bisectable6 left, evalable6 left 01:38 sourceable6 joined, committable6 joined, unicodable6 joined, evalable6 joined, notable6 joined 01:39 quotable6 joined, reportable6 joined, tellable6 joined, nativecallable6 joined, benchable6 joined 01:40 bisectable6 joined, statisfiable6 joined, linkable6 joined, shareable6 joined, bloatable6 joined, squashable6 joined 01:41 releasable6 joined, coverable6 joined, greppable6 joined 02:11 nine left, camelia left 02:12 camelia joined, nine joined 03:43 linkable6 left, evalable6 left 03:44 evalable6 joined 03:45 linkable6 joined 04:00 frost joined 06:03 reportable6 left 06:04 reportable6 joined
dogbert17 nine: let me know if you need more information and if you want me to report the bug 08:12
09:13 evalable6 left, linkable6 left 09:14 evalable6 joined, linkable6 joined
MasterDuke [Coke]: i remember in college wondering if that would work if we got some of the kits to make them yourselves, like our parents made for us with fruit juice as kids. we thought you'd have to have so little alcohol as to not be worth it 09:41
what's the ABV of the ones you found?
10:36 evalable6 left, linkable6 left 10:38 linkable6 joined 10:39 evalable6 joined
lizmat and yet another Rakudo Weekly News hits the Net: rakudoweekly.blog/2021/06/22/2021-...all-steps/ 11:57
with apologies for being 1 day late 11:58
12:02 reportable6 left
MasterDuke lizmat++ (and fyi, a couple entries in "Core Developments" have that `style="color:initial;"` on their `<em>` again) 12:03
12:05 reportable6 joined
lizmat argh WP again 12:07
I think I got all of them? 12:09
12:10 frost left
lizmat MasterDuke++ 12:10
MasterDuke looks like 12:11
13:54 frost joined 14:20 frost left
Geth MoarVM/new-disp: 5bcb87516f | (Jonathan Worthington)++ | 2 files
Add HLL configuration for dispatchers

We'll allow languages so specify their standard call and method call dispatchers, in order to facilitate inter-language calls (between NQP and Raku for now, although in principle anything).
15:45
Nicholas jnthnwrthngtn: MoarVM panic: Fixed size allocator: wrong size in free: expected 64, got 8 15:50
Aborted
that's NQP on bfcce46e597a5620bf8b64836e36c2c99869b038
MoarVM on 5bcb87516f1e218769be411395c2f5eaa5a2c167
+#define FSA_SIZE_DEBUG 1 15:51
(etc)
MasterDuke "although in principle anything" ... oooh, MoarVM eventually gains GraalVM-like multi-lang interoperability... 15:55
Nicholas but with less startup overhead? 15:57
MasterDuke hopefully. though the native images you can build with it start up incredibly quickly 15:59
jnthnwrthngtn Nicholas: Regression, or just how it is? 16:00
Will try and repro in a moment 16:01
Nicholas "not know" - I think that it used to pass, but it's a different machine from yesterday (1 debian version earlier, but otherwise I think identical setup) 16:07
but might be a regression since yesterday
jnthnwrthngtn Interop is really hard even when one is designing a language with it in mind from the start. I recently tracked down a resource leak that occurred in some Kotlin code because it was passing a Kotlin lambda, stored in a field, to a Java function. Since the lambda was in a field it should pass the same object each time, no? Apparently not; it wraps it on the interop boundary. The Java code was using the object
identity of the thing passed to in order to keep track of handlers it needed to add/remove, but the interop wrappings were distinct objects. Oops.
nine Ah, yes, interop. Always fun :D 16:12
18:03 reportable6 left 18:04 reportable6 joined
[Coke] MasterDuke: 6% ABV 18:34
MasterDuke interesting, surprised it's that high 18:58
Nicholas jnthnwrthngtn: I am getting different results on different machines. I believe that the only thing that differs is debian version. So clearly I have something wrong, as this can't explain the panic. 21:13
jnthnwrthngtn Nicholas: No, hmm, that one I'd expect to be consistent 21:14
Nicholas: I did a rebase and force push of all the branches yesterday; I don't suppose it could be some accident around that?
Nicholas no, I have the same commit hash for NQP and MoarVM on both. The same diff for MoarVM on both (all the debug stuff) and I think the same MVM_ exports in the environment 21:16
jnthnwrthngtn OK, then it's weird. 21:19
Nicholas Yes, totally agree. I need to go to sleep, and look at this with a fresh head.
jnthnwrthngtn Rest well 21:20
Nicholas Um, it's not yet actually "good morning", but 21:22
==6252== Conditional jump or move depends on uninitialised value(s)
==6252== at 0x4A4DDCB: MVM_spesh_log_is_logging (log.h:30)
==6252== by 0x4A65ABF: MVM_interp_run (interp.c:3154)
==6252== by 0x4C164BF: MVM_vm_run_file (moar.c:513)
==6252== by 0x109A43: main (main.c:305)
MasterDuke jnthnwrthngtn: now might not be the best time, but github.com/MoarVM/MoarVM/pull/1426 could use a final review at some point
Nicholas that was
valgrind '/home/nick/Sandpit/moar-g/bin/moar' --libpath=src/vm/moar/stage0 src/vm/moar/stage0/nqp.moarvm --bootstrap --setting-path=gen/moar/stage1 --module-path=gen/moar/stage1 --no-regex-lib --target=mbc --setting=NQPCORE --stable-sc=stage1 --output=gen/moar/stage1/QRegex.moarvm gen/moar/stage1/QRegex.nqp
that doesn't seem great. 21:23
jnthnwrthngtn grmbl, my ability to reproduce it ain't either 21:26
Nicholas I still fail to go to bed. I should. You should too, when you're ready. Let me figure this one out, but it might take a day or two. I'm not yet "shot" of the other job. 21:28
jnthnwrthngtn Oh, I managed it. 21:30
With blocking + nodelay
I'm mildly confused how this happens in an NQP build where new-disp isn't used yet and so in theory "nothing changed" 21:32
Though fair warning: very soon everything is about to change, probably in a "breaking stuff for a while" way (I really need to get NQP shoved over to new-disp before I make more progress on that for Rakudo) 21:33
23:25 m6502 joined