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.
MasterDuke is it really possible to get something that can't `$foo.HOW.name($foo)`? 09:13
patrickb nine: Then my guess of the op being reversed again was actually correct. (Not that such guesses would buy me anything but a grin :-) 09:38
nine patrickb: definitely a good guess! 11:22
MasterDuke: just compare the output of rakudo --target=optimize -e 'my $a = 0; $a += 1;' and rakudo --target=optimize -e 'my $a = 0; $a = $a + 1;' 11:33
MasterDuke yeah, i saw that the += version has an extra `definite($a) ?? $a !! 0`. i wonder how possible/easy it would be to check if there's a compile-time value and if so, re-write it into the `$a = $a + 1` version (or remove that check) 11:36
nine Worth a try I guess. 11:38
MasterDuke could `self.find_single_symbol('Mu', :setting-only)` in World ever actually die (i.e., not find 'Mu')? 11:51
nine nqp::die("Cannot find symbol $name in $setting_name"); 11:58
That would indicate so
Oh, 'Mu' specifically? 11:59
lizmat I guess very early in the bootstrap, Mu may not be a thing yet ? 12:00
nine I dare say if we make it through settings compilation without that try, we don't need it 12:01
MasterDuke yeah, there are a couple places where it tries to get hardcoded values, like 'Mu', and i'm wondering if those *actually* need any checks 12:02
MasterDuke interesting. in windows, when using an nqp with github.com/Raku/nqp/pull/764 creating even a minimal profile dies with `KnowHOW methods must be called on object instance with REPR KnowHOWREPR 16:20
   at gen\moar\stage2\NQPHLL.nqp:245  (C:\Users\Dan\Source\nqp/NQPHLL.moarvm:get_remapped_type_id)`
lizmat fwiw, some of my modules currently fail testing in weird ways on Windows 16:33
[Coke] recent failure mode? 17:03
or just-now-testing on windows? 17:04
MasterDuke just now testing the pr
for me
nine michaelheap.com/git-ignore-rev/ 19:04
lizmat interesting 19:06
MasterDuke yeah, just found out about that a couple months ago. good timing as we were discussing doing a repo-wide linting/formatting commit, but one of the cons was the impact on git blame. dramatically minimized that particular con 19:07
lizmat feels there should be a trigger in a commit message to make that commit automatically be added to the ignorable commits ? 19:08
feels there should be a trigger in a commit message to make that commit automatically be added to the ignorable commits ? 19:09
Geth MoarVM: 418711f010 | (Daniel Green)++ | src/io/syncfile.c
Whoops, fix a wrong format string change

It slipped in accidentally when fixing some others for MinGW.
22:38