| IRC logs at
Set by AlexDaniel on 12 June 2018.
00:42 leont left 01:13 elcaro left 01:20 elcaro joined 01:36 elcaro left 01:50 elcaro joined 04:06 rba left 04:07 rba_ joined, rba_ is now known as rba 05:26 dogbert17 joined 05:28 dogbert11 left
nwc10 good *, #moarvm 05:57
nine *yawn* 06:04
nwc10 I feel like that too 06:05
06:31 ggoebel left 06:33 patrickb joined 06:46 Altai-man joined 07:03 sena_kun joined 07:05 leont joined, Altai-man left
MasterDuke full on cold, kids waking up in the night, yep 07:24
08:20 patrickb left 08:22 Altai-man joined 08:24 sena_kun left 08:39 brrt joined
brrt good * 08:45
timotimo good * 08:46
nwc10 good *
jnthn morning o/ 09:27
nwc10 \o 09:30
MasterDuke aloha 09:33
brrt not yet everyone here 09:37
way way way back, when I did something for parrot, we had a parrot weekly meeting
timotimo did you see my appimages? 09:39
jnthn brrt: Do you think there'd be some value in some kind of scheduled MoarVM team meetups? 09:41
Raku is getting a steering committee, so maybe it would be a fitting time to do something here too. 09:42
nwc10 I don't know what I think about that part 09:43
timotimo having "meeting minutes" in some central place could be good to help me not forget what everybody is/was doing - but also i sometimes just completely forget what i was doing
jnthn Though I don't think we need anything so "heavyweight"; more just a committers get-together every so often
nwc10 but I will state that I felt that a key useful "innovation" of the parrot meeting ("parrotsketch") was to get people to type up their short reports *in advance* and then just paste them into IRC 09:44
this avoided delays
jnthn oh, yes, that's worth stealing :)
nwc10 this is only important if the plan involves reports
we failed to do this at work, but our meetings moved off IRC soon enough that I didn't push for it
jnthn Well, given what timotimo said they'd find useful, maybe it should :) It also, if published somewhere, is good communication that things are happening. 09:45
Anyway, I'm up for some kind of regular sync-up of what we're all doing, as well as an opportunity to discuss blockers/ideas/whatever. 09:46
brrt jnthn: yeah, I think something like that 09:47
I mean, the #moarvm community, such as it is, is fairly light-touch, that might be a good thing to presever
jnthn Yes; that's worked well for us, imo. 09:48
brrt but I do agree that a sense of 'what is happening' would be nice, too 09:50
jnthn So maybe we pick an interval, a time, and simply say all those with a commit bit are welcome? Then everyone can pre-submit a "what I've been doing" and perhaps "things I'd like to discuss". 09:52
And then we see if it's useful and course-correct it as we go.
brrt ++ from me. (now let's bikeshed about the time) 09:54
MasterDuke in this channel? or using some sort of virtual meeting thing? 09:55
jnthn If we had a dedicated channel for it we can have it logged and that solves minutes? I'm open to other suggestions.
timotimo that's a good idea 09:58
jnthn OK, so: interval, time :) 10:10
nwc10 well, "morning", obviosly 10:12
timotimo let's not forget finding a good pun for the channel name
nwc10 "morning virtual meeting" is a bit long
timotimo Meeting Of All Relevant Virtual Members 10:14
not finding a good word for the V
nwc10 that's more a backronym than a pun. Actually, as was mine 10:15
.oO( moarons )
timotimo MoarMerrier
10:29 brrt left
MasterDuke fewer concrete people 10:37
12:00 evalable6 left, linkable6 left 12:01 linkable6 joined 12:02 evalable6 joined
[Coke] moaryouknow 12:02
jnthn metamoarphic 12:04
MasterDuke i just removed the `rounds` parameter from `nqp::isprime_I` (and the various places in moarvm and ran tools/update_ops.p6) and nqp segfaults in `MVM_cu_string` when calling `nqp::isprime_I`. it's trying to index 3670016 into the string heap, when num_strings is only 255 12:16
12:17 ggoebel joined
MasterDuke this seems like the sort of thing that would have happened if i hadn't run tools/update_ops.p6 12:18
12:20 brrt joined 12:21 nebuchadnezzar left 12:23 ggoebel left, sena_kun joined
Geth MoarVM: b243d01789 | (Patrick Böker)++ | src/io/fileops.c
Fix MVM_file_isexecutable() when being root

Different from the readability and writeability tests, executability is not affected by the user being root. As long as the `x` flag isn't set, not even root can execute a file.
Fixes Raku/nqp#652
linkable6 NQP#652 [closed]: `fileexecutable` returns incorrect value in Linux systems (Alpine, Debian) when the user is root.
Geth MoarVM: 96d32a1ad8 | (Patrick Böker)++ | src/io/fileops.c
Don't duplicate file existence checks

When calling `uv_fs_stat()` it's possible to determine whether the file is missing by looking at the return code. So there is no need to check that beforehand. Doing it with a single `uv_fs_stat()` call also prevents a race condition.
MoarVM: 738024f9c2 | (Jonathan Worthington)++ (committed using GitHub Web editor) | src/io/fileops.c
Merge pull request #1341 from patrickbkr/fix-root-file-executable

Fix MVM_file_isexecutable() when being root
12:25 Altai-man left
Geth MoarVM: 124fe2be99 | (Nicholas Clark)++ | 2 files
copy_to should call MVM_gc_write_barrier with the *new* key's address.

This is a regression introduced since the last release by commit 9d0f1b326d8d:
   Replace MVM_str_hash_bind_nt with MVM_str_hash_insert_nt.
MoarVM: 39743de418 | (Jonathan Worthington)++ (committed using GitHub Web editor) | 2 files
Merge pull request #1347 from MoarVM/MVMHash-copy_to-marking-bug

copy_to should call MVM_gc_write_barrier with the *new* key's address.
MasterDuke what the current diff looks like if anyone is curious
and i did a couple rounds of make && make m-bootstrap-files && make just to see if that was it, no change 12:39
Geth MoarVM: 87ad486f30 | (Nicholas Clark)++ | 5 files
sparc32 does not support unaligned 32 bit reads or writes.

Our bytecode is 2-byte aligned, but we need to read 4-byte values. sparc will SIGBUS if asked to perform a misaligned read.
We also need to add the same hack as s390 to get a definition of AO_fetch_compare_and_swap_full ... (9 more lines)
MoarVM: 0772199939 | (Nicholas Clark)++ | src/6model/reprs/P6opaque.c
Correctly align object references (ie pointers) in P6opaque object bodies.

Previously these were written at whatever offset the composer happened to be at, which meant that some were only 4-byte aligned. On systems with 8-byte pointers which care strictly about alignment, these leads to SIGBUS.
This is sufficient to get MoarVM, NQP and most Rakudo tests working on sparc64.
MoarVM: 34e0fb55e3 | (Jonathan Worthington)++ (committed using GitHub Web editor) | 6 files
Merge pull request #1343 from MoarVM/sparc

sparc64 support
jnthn MasterDuke: Did you update interp.c? Does it calculate the next instruction offset correctly? 12:41
MasterDuke argh! i did update it, but not the offset calculation
and of course that makes perfect sense
jnthn :) 12:42
MasterDuke yep, nqp is good now 12:43
jnthn :)
I'll return to the spesh candidate thing some time, but I'm not sure I'm feeling any more clever today than I was yesterday...
Plus quite keen to get a bit of work in on new-disp, at last...
MasterDuke i'm all stuffed up and blowing my nose every 2-3min, in no shape for anything complicated either 12:44
jnthn I slept badly on Sunday night. Much better last night, but it seems that doesn't quite make up for it... 12:46
MasterDuke i thought simply removing a parameter from an op would be within reach, but we see where that ended up...
12:50 ggoebel joined
brrt I've started seriously considering buying an apple if they do follow through with the their apple laptops 12:50
jnthn Considering buying one, or not buying one? :) 12:51
What the heck, I just build new-disp without any asan stuff turned on and somehow it ends up building it as if I passed --asan... 12:56
Build leftovers I guess
brrt considering buying one. 13:22
I haven't bought one ever
yet 13:23
13:31 domidumont joined
[Coke] They are shiny. 13:35
timotimo you're interested in apple's own "silicon"?
brrt I'm interested in the arm64 yes 14:19
looks quite promising
14:30 brrt left 14:53 nebuchadnezzar joined
Kaiepi how do you guys debug c on windows? 14:59
timotimo first step, move over to linux
just kidding. i hear msvcpp is very good?
Kaiepi not an option in this case :(
timotimo vscode and qt creator also have GDB front-ends
Kaiepi re linux
ahhh 15:00
didn't realize vscode came with that
timotimo oh
right, it doesn't come with it
there's a c and cpp development plugin from microsoft that i have installed 15:01
Kaiepi i have it fsr 15:03
timotimo "for some reason"? 15:12
Kaiepi yeah 15:13
15:14 MasterDuke left 16:00 MasterDuke joined 16:20 ggoebel left 16:22 Altai-man joined 16:25 sena_kun left 16:39 domidumont left 16:43 zakharyas joined 17:03 patrickb joined 18:02 reportable6 left 18:06 reportable6 joined 18:18 zakharyas left 19:43 brrt joined 20:02 Kaiepi left 20:23 sena_kun joined 20:24 Altai-man left 20:31 ggoebel joined 20:50 sena_kun left 21:10 patrickb left 22:04 SmokeMachine left, chansen_ left, chansen_ joined 22:05 SmokeMachine joined 22:49 brrt left