[00:20] *** leont left
[01:05] *** MasterDuke left
[01:58] *** pamplemousse left
[04:02] *** pamplemousse joined
[04:17] *** pamplemousse left
[05:04] <nwc10> good *, #moarvm 

[06:41] *** Altai-man joined
[06:45] *** brrt joined
[06:47] *** samcv left
[06:48] <nwc10> good *, brrt 

[06:50] <brrt> good * nwc10

[06:59] *** samcv joined
[07:04] *** samcv left
[07:08] *** samcv joined
[07:19] *** sena_kun joined
[07:21] *** Altai-man left
[07:21] *** samcv left
[07:46] *** samcv joined
[07:49] *** zakharyas joined
[08:02] *** leont joined
[08:54] *** samcv left
[09:02] *** samcv joined
[09:23] *** zakharyas1 joined
[09:23] *** brrt left
[09:24] *** zakharyas left
[09:25] <jnthn> morning

[09:26] <tellable6> 2020-09-10T23:36:21Z #raku-dev <Xliff> jnthn: How is work on RakuAST progressing?

[09:47] *** brrt joined
[09:51] <nwc10> correct!

[10:11] *** tbrowder left
[10:15] *** domidumont joined
[10:16] *** tbrowder joined
[10:37] *** zakharyas1 left
[10:41] <brrt> not anymore though

[10:47] *** zakharyas joined
[10:57] *** brrt left
[11:18] *** Altai-man joined
[11:21] *** sena_kun left
[11:25] *** sena_kun joined
[11:27] *** Altai-man left
[11:34] *** zakharyas left
[11:54] *** domidumont left
[12:15] *** domidumont joined
[12:25] *** brrt joined
[14:13] *** zakharyas joined
[14:15] *** zakharyas1 joined
[14:18] *** zakharyas left
[14:45] *** brrt left
[15:24] *** Altai-man joined
[15:27] *** sena_kun left
[16:08] <nine> MasterDuke: did you run the mvmroot checker with the full set of 6model types or only for MVMSpeshCandidate? In the latter case it may have missed references that are only typed MVMObject* or MVMCollectable&

[16:09] <tellable6> nine, I'll pass your message to MasterDuke

[16:12] <nine> frame->spesh_cand = chosen_cand needs replacing with MVM_ASSIGN_REF, too

[16:14] <nine> This would be more fun if rr wouldn't crap out on me half the time...

[16:14] <nine> Anyway, need to get off the train. Have a nice weekend!

[16:16] <nine> MVM_spesh_graph_mark misses the spesh_cand!

[17:05] *** MasterDuke joined
[17:08] <MasterDuke> nine++ nice catches. i actually had the same idea re check-roots.py while at $work today, will re-run this evening and look into those things you mentioned

[17:08] <tellable6> 2020-09-11T16:09:00Z #moarvm <nine> MasterDuke: did you run the mvmroot checker with the full set of 6model types or only for MVMSpeshCandidate? In the latter case it may have missed references that are only typed MVMObject* or MVMCollectable&

[17:10] <[Coke]> nine, you get to use the train? I haven't gone by train since... March (used to use it at least once a month for work)

[17:13] *** brrt joined
[17:48] *** zakharyas1 left
[17:52] *** brrt left
[18:15] *** vrurg left
[18:18] *** vrurg joined
[18:34] *** brrt joined
[18:43] <nwc10> [Coke]: I don't think that he needs to much, because when I went to visit him a couple of months ago, he forgot that we needed masks to go on a tram. Luckily I had two.

[18:44] <[Coke]> aha

[18:59] <nine> I ride my bicycle for 5 minutes to work ;) But I'm at my parents' over the weekend

[18:59] <[Coke]> I haven't gone anywhere for work since first week March. :)

[18:59] <[Coke]> so I envy you even the bike ride! :)

[19:00] <nwc10> first week of March - that was a world away: http://act.yapc.eu/gpw2020/

[19:00] <nwc10> oh yes, so Andrea was wearing the t-shirt from GPW2020, and one of her friends saw the logo and the date and asked if it actualyl had happened :-)

[19:01] <nwc10> I guess it will be one of the few "2020" branded things that happened in 2020, instead of 2021

[19:03] <brrt> heh

[19:03] <nine> Feels surreal to think about GPW2020. I remember lizmat checking corona statistics daily and it were about 30 cases in the Netherlands

[19:04] <timotimo> remember when we thought we would have an internals summit

[19:06] <lizmat> nine: yeah, but about 15km from where we live, Langbroich / Gangelt was one of the first hotspots in Germany, with 100s of cases *at that time*

[19:06] <lizmat> so in a way, I *was* already really surprised that GPW2020 went ahead as it did

[19:06] <MasterDuke> nine: i've  added `if (g->cand) MVM_gc_worklist_add(tc, worklist, &(g->cand));` to MVM_spesh_graph_mark and `MVM_ASSIGN_REF(tc, &(frame->header), frame->spesh_cand, chosen_cand);` to MVM_frame_invoke. nqp builds, but rakudo panics or segfaults

[19:06] <lizmat> it could have turned out pretty badly

[19:06] <MasterDuke> haven't re-run check-roots.py yet

[19:11] <MasterDuke> hm. tons more listed if i add MVMObject and MVMCollectable, but none look likely

[19:12] <nine> MasterDuke: you can ignore all in serialization related functions and the ones in the profiler

[19:14] <MasterDuke> bunch in src/6model/bootstrap.c: VMString, knowhow, type_obj, type

[19:15] <MasterDuke> bunch in src/6model/reprs/CPPStruct.c and src/6model/reprs/CUnion.c and src/6model/reprs/CStruct.c. i assume i can ignore those?

[19:15] <MasterDuke> Missing root for `thread_obj` in MVM_gc_mark_thread_blocked (child_tc); at src/core/threads.c:128 used in thread_obj.15_22 = thread_obj; at src/core/threads.c:149

[19:15] <MasterDuke> Missing root for `root` in rm = MVM_repr_alloc_init (tc, _2); at src/6model/reprs/MVMCompUnit.c:25 used in _3 = &root->header; at src/6model/reprs/MVMCompUnit.c:26

[19:16] <MasterDuke> those are the only other two besides the ones in bootstrap.c that look possibly relevant

[19:23] <MasterDuke> afk for a while

[19:25] *** sena_kun joined
[19:25] <nine> I did add a few comments where I was sure that the code is correct

[19:26] *** Altai-man left
[20:19] <MasterDuke> ah, both src/6model/reprs/MVMCompUnit.c:25 and src/core/threads.c:128 have comments either saying explicitly or implying MVMROOT is not needed

[20:20] *** brrt left
[20:20] *** patrickb joined
[20:37] *** domidumont left
[20:42] *** sena_kun left
[20:58] *** patrickb left
[22:22] <timotimo> nine: you don't happen to have a brilliant idea for how to turn an MVMROOT into the expanded version automatically for when i want to single-step through code in it?