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.
Geth MoarVM: 55be3aa27f | (Nicholas Clark)++ | src/moar.h
#include <assert.h> explicitly.

It turns out that atomic_ops.h was pulling in assert.h for us.		 08:48
Nicholas I guess that means 08:49
assert(UGT == Good);
nine assert(moarning) 08:52
MasterDuke random idea of the day, would it be good to use something like metacpan.org/pod/TAP::Formatter::JUnit so our test runs in CI can show up nicely in the "Test" tab of azure? 13:24
docs.microsoft.com/en-us/azure/dev...-tests-tab
nine These build failures make me nervous: build.opensuse.org/package/live_bu...RM/aarch64 build.opensuse.org/package/live_bu...ARM/armv7l 14:26
At least the URI one seems quite reproducible (haven't tried the other one)
MasterDuke are those with master? 14:29
nine Yes. But I don't know when it started. I've gotten lots of build failures since about new-disp landed but never took the time to have a closer look. Especially because I have no idea to investigate a build hang.
MasterDuke can you log into the building system? 14:30
nine no 14:32
[Coke] getting nativecall failures on windows rakudo 'nmake test' at HEAD 14:37
looks similar to new-disp original failures that were resolved 14:38
rakudo at f77fda7d1
(test fails in nmake test, but works when run standalone) 14:40
e.g. t\04-nativecall\07-writearrier.t 14:41
nine Does nmake test run tests in parallel? 15:11
[Coke] yes 15:25
(I have mine configured to, anyway)
nine JWST fully deployed! 15:33
Nicholas you're ahead of SpaceXMasterrace on this 15:34
nine Watching the NASA live stream: www.youtube.com/watch?v=tlGTem8vkB0 15:35
timo is it just my machine or is that stream kinda stuttery 15:59
nine same here 16:00
Good news: looks like fix_unsigned just needs one more bootstrap, so it's 2 in total 16:10
MasterDuke a friend of mine works on the jwst, took me to visit it years ago. haven't heard from him recently, i assume he's pretty busy right now 16:18
timo: up for questions about profiling? 16:20
Nicholas nine: before I forget, and I've not done my homework on this 17:42
but removing the (old) disp code means that most bytecode from 2013-mumble to "mid last year" no longer validates
so effectively all or nearly all deprecated ops can be re-used
I wasn't early enough to say "so they can be used for unsigned" 17:43

Powered by Cro and the Raku® Programming Language.

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

Thank you!