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 00:05 reportable6 joined 01:06 lizmat left, lizmat joined 01:34 frost joined 04:51 linkable6 left, evalable6 left 04:52 evalable6 joined 05:22 frost left 05:53 linkable6 joined 06:02 linkable6 left, reportable6 left
Nicholas good *, #moarvm 07:08
08:05 reportable6 joined 08:23 frost joined 09:02 linkable6 joined 10:12 discord-raku-bot left 10:13 discord-raku-bot joined
jnthnwrthngtn moarning o/ 10:13
Nicholas \o 10:18
lizmat o/o 10:23
MasterDuke ahoy 10:50
jnthnwrthngtn: if you backlog, there's some further discussion and information about that large difference between `is default(0)` and `= 0` 10:56
nine jnthnwrthngtn: do I understand your message correctly that you will ponder a bit how to handle continuations and MVM_callstack_first_frame_in_region? So no action required by me? 11:25
jnthnwrthngtn nine: Well, my first thought is that we might need directional forms, e.g. the continuation root would want to look at the *next* region after the continuation mark. tbh, I assumed that a continuation record + a frame record would always fit in a single region 11:27
I didn't really anticipate wanting to support a region so small that'd not work out :)
But looking in the previous region would find a frame that is not part of the continuation 11:28
nine If it just were crazy small regions, I'd not even have fixed it. But there's another way of causing us to exceed the one region: a crazy large frame. And that's quite possible with work being part of the frame now. 11:29
What I don't understand: we expect the continuation record to be eventually followed by a frame. And that first frame we find is the one we're looking for. What difference does it make if that frame is in the same region or in the next one? IIUC the continuation takes hold of the whole rest of the region chain anyway. 11:38
Is it possible that the frame we're looking for does not appear on the callstack at all and that we find some different frame further down the chain instead? 11:39
jnthnwrthngtn Oh hang on, I misread it. :/ You really are looking in the *next* regiion. Somehow I'd read it as looking at the previous one.
nine Aah...yes that next/prev had me confused as well at some point :) 11:40
Darn stacks insisting on growing upwards and all
jnthnwrthngtn Not sure how I managed that. Well, I am, I was already really tired by the time I got to the third PR. :)
I'll take a closer look at it after lunch. 11:41
nine Well...downwards actually. But that way it demonstrates the confusion even better :D
jnthnwrthngtn Which I'll disappear for nwo; bbiab :)
nine enjoy!
dogbert17 MoarVM panic: Adding pointer 0x555d62d3b158 to past fromspace to GC worklist # hmm 11:59
11:59 linkable6 left
dogbert17 when running: MVM_SPESH_BLOCKING=1 ./rakudo-m -Ilib t/spec/S32-str/indent.rakudo.moar , nursery size is 11k 11:59
perhaps something for nine :) 12:00
12:02 reportable6 left 13:00 linkable6 joined 13:04 reportable6 joined
dogbert17 h, I forgot the gist, gist.github.com/dogbert17/9bf6d86d...ad6b3dafee 13:10
18:02 reportable6 left 18:03 reportable6 joined 22:28 evalable6 left, linkable6 left 23:29 evalable6 joined 23:31 linkable6 joined