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.
japhb Welp, Rakudo 11c8f82a0 (HEAD~1) works with Cairo, but HEAD does not. HEAD is: "* fb909efce - (HEAD -> master, origin/master) Bump NQP to get the latest MoarVM fixes (2 days ago) <Elizabeth Mattijsen> 01:50
Nicholas good *, #moarvm 07:01
nine vrurg: most of my PRs fail CI already due to that branch thing 07:45
lizmat and yet another Rakudo Weekly News hits the Net: rakudoweekly.blog/2022/01/17/2022-...kucon-how/ 13:01
MasterDuke anybody have a suggestion for what could be happening with that segv uncovered by my new ternary warning in the optimizer? 15:14
m: my $v = 3; my $sign = $v +> 31 ?? -1 !! 1; say $sign          # this one
camelia (signal SEGV)
MasterDuke if i'm stopped at dispatch_o in the interpreter, is there a good way to see what all the arguments are? 15:28
nine MasterDuke: yes, look at callsite and args 16:19
Or just MVM_dump_bytecode(tc) and look at the registers. That's even easier 16:20
MasterDuke the problem with dumping the bytecode is that it's large (in this case especially so) and looking back and forth for that `->` to see where i am is annoying 16:28
but maybe i'm in the right place randomly. i see 'lang-hllize', and then '$!descriptor', and then 'raku-rv-decont' 16:30
so maybe this is where the VMHash is getting upgraded into a P6opaque
oh, and a callsite doesn't know the argument values 16:32
nine callsite knows how many and what kind of arguments to expect. args holds the register numbers of registers holding these arguments. tc->cur_frame->work holds the registers holding the arguments 16:33
MasterDuke might be nice to create some debug function to prettily print them 16:35
interesting, i 16:39
'm here github.com/rakudo/rakudo/blob/mast...de.pm6#L25
well, i noticed that `method prec` is `is raw` in traits.pm6, but no change if i make the one in Code.pm6 `is raw` also 16:44