|
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. |
|||
|
06:41
sena_kun joined
|
|||
| lizmat | Q: can you thread-safely ask for the nqp::elems(hash) while it is being changed by another thread ? | 12:24 | |
| jdv | how would that make sense? my naive guess is no unless the hash is one of those fancy concurrency safe ones which nqp/moarvm hashes are not. | 13:14 | |
| lizmat | it would make sense for a progress indicator | 13:16 | |
| and it would be helpful if a lock wouldn't be needed | 13:17 | ||
| ugexe | reads must be protected if writes can occur | 15:53 | |
| lizmat | but is that also right for the "elems" value of a hash? | 16:01 | |
| or is that not an int in a struct somewhere ? | |||
| ugexe | m: use nqp; my %hash; my $p = start { for 1..Inf { %hash{$_} = 1 } }; for 1..Inf { say nqp::elems(%hash) } | 16:09 | |
| camelia | (signal SEGV) | ||
| ugexe | m: use nqp; my %hash; my $p = start { for 1..Inf { %hash{$_} = 1 } }; sleep 1; for 1..Inf { $ = nqp::elems(%hash) } | 16:10 | |
| camelia | MoarVM oops: MVM_str_hash_count called with a stale hashtable pointer at <tmp>:1 (<ephemeral file>:) from <tmp>:1 (<ephemeral file>:<unit>) from <tmp>:1 (<ephemeral file>:<unit-outer>) from NQP::src/HLL/Compiler.nqp:196 (/home/cameli… |
||
| lizmat | I guess that answers that question, thanks :-) | 16:11 | |
|
21:17
rypervenche left
21:39
rypervenche joined
22:41
sena_kun left
|
|||