github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
nine Oh, repl.t has also a time out through doesn't-hang. That's just 15 seconds. 06:15
15 seconds sound rather short for a test where the failure condition seems to be that it won't finish at all. Why not wait 3 minutes or so? If a bug so tested reappears, it will still tell you. 06:17
Kaiepi idk if anyone saw last time i asked but is there anything i need to do before pr 1166 can be merged? 14:51
timotimo M#1166 15:22
synopsebot M#1166 [open]: github.com/MoarVM/MoarVM/pull/1166 [IP6NS Grant] Improve hostname resolution
timotimo i haven't looked again since nine asked if you could make it more mechanism than policy 15:38
nine timotimo: I'm also fine with a merge of the functionality now and moving of the policiy to upper layers later on. As we already have a different policy hard coded in MoarVM it'd not really be a regression in that regard. 15:48
timotimo that's true 15:49
nine The only thing I don't want is to get us locked into one particular policy by some part of the API design. But if Kaiepi is positive that the policy part can be moved without problem later, that's fine with me.
timotimo i haven't reviewed the code, but i'd be +1 on the merge then
Kaiepi i'm fairly certain it can 15:51
without being breaking
nine So...when I come across what's obviously a Backtrace's frame list in a serialization issue. How would I find out where that Backtrace comes from? I'd guess that there's some Failure that shouldn't be there 16:32
timotimo we need something like the heap snapshot profiler but for serialization 16:35
something something gdb python script? 16:48
lizmat nine: should a Backtrace / Failure / Exception *ever* be serialized ?? 18:06
nine lizmat: no. And AFAIK we can't even serialize exceptions 18:22
nine Here I am, wondering where that backtrace comes from.... the backtrace...which is literally the thing that is only there for showing where something happened 18:23
lizmat m: dd Backtrace.new
camelia Backtrace.new
lizmat m: dd Backtrace.new.list
camelia (Backtrace::Frame.new(file => "SETTING::src/core.c/Backtrace.pm6", line => 94, code => { ... }, subname => ""), Backtrace::Frame.new(file => "SETTING::src/core.c/Backtrace.pm6", line => 94, code => multi method new (Backtrace: *%_) { #`(Method|7890018…
lizmat are there actually Backtrace::Frames in that list? 18:24
nine Ooook....I now know that it's the require Inline::Perl5; in CompUnit::Repository::Perl5 which is throwing. Progress! 18:38
Intriguingly that's a failure in INDIRECT_NAME_LOOKUP. The program however continues happily until the serialization phase 18:40
nine And Inline::Perl5 gets loaded just fine 18:40

Powered by Cro and the Raku® Programming Language.

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

Thank you!