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.
nine Ha! Did that last week, when I last played 05:55
[Coke] I need to get up to date on these new-fangled games. 15:26
I got a PS5 and have used it mainly to play PS4/Fallout4
timo wow you got a ps5. did you have to hurt anybody for it? 15:27
MasterDuke58 horizon zero dawn is pretty much the newest game i have, but it isn't all that recent, 2017 15:28
[Coke] timo: I had to hit refresh SO MANY TIMES
fallout 4 was 2015. :)
timo i think the newest game i own is boyfriend dungeon, which came out the other day but i havent played it yet. if i limit to games ive played that would probably be subnautica below zero 15:30
MasterDuke58 i just got in the virtual queue on amd.com a hour or so ago with a 45 min wait time, but got a notice 10 min in saying they were all out of video cards
heh, i just started playing the original subnautica a month or two ago because my daughter likes watching the "alien fish game" 15:31
timo its super good 15:33
snarky pda lady is a very good character
MasterDuke58 changing topics, anyone taken a look at github.com/MoarVM/MoarVM/pull/1528 ? 15:35
MasterDuke58 afk for a while 15:39
timo lgtm 15:47
Geth MoarVM: 7355f744ed | (Daniel Green)++ | 4 files
Free filename if exception when loading bytecode

Because sometimes the given filename is malloced and sometimes it's from argv, we also have to add a flag for whether it should be freed.		 18:40
MoarVM: 343d0f37f2 | MasterDuke17++ (committed using GitHub Web editor) | 4 files
Merge pull request #1528 from MasterDuke17/free_filename_if_exception_thrown_when_loading_bytecode

Free filename if exception when loading bytecode
MasterDuke58 thanks. ^^^ got rid of an 'XXX', which i always enjoy doing 18:41
Geth MoarVM/new-disp: 42abdc5b59 | (Stefan Seifert)++ | src/core/frame.c
Fix segfault caused by GC processing of half-initialized frame

If logging the frame entry in the spesh log triggered a send_log and if that happened precisely when some thread wanted start a GC run, the GC may have come across a frame that had an uninitialized params sub-struct. This could cause us to dereference uninitialized pointes resulting in segfaults.
Fix by initializing params before logging the frame entry with spesh, i.e. doing something that could potentially enter the GC.		 19:08
MoarVM/new-disp: 14ff58f32b | (Stefan Seifert)++ | src/spesh/worker.c
Fix potential dead lock when running with MVM_SPESH_BLOCKING=1

When running with spesh blocking, send_log creates a new spesh log entry with an already locked mutex, then marks the current thread blocked for GC (which might enter the GC right away). The spesh worker thread picks up that spesh log and waits for the mutex. The spesh thread never joins the GC, so everyone ends up waiting.
Fix by marking the spesh worker thread blocked for GC before acquiring the spesh log's mutex.
MasterDuke58 nice. making dogbert17 work hard for those segvs 19:30
[Coke] I'm getting a failure trying to do a rakudo build on new-disp. 21:35
gist.github.com/coke/86e97dd0690df...046155ac07 21:36
(started working in a fresher clone) 21:42
dogbert17 oh no, some has fixed all the new-disp bugs, now we'll have to find more :) 22:04
nine++
s/some/someone/
it seems that there are still things to do, e.g. 'MoarVM oops: Unknown handler in inline cache entry' 22:05
to repro run 'while ./rakudo-m -Ilib t/02-rakudo/15-gh_1202.t; do :; done' and wait, no special changes needed 22:06

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!