github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
nwc10 good *, #moarvm 06:55
samcv dogbert17, what commit id is that log from? 07:47
samcv I think you maybe need to investigate each one. Though this page on most common threadsanitizer data races may be helpful github.com/google/sanitizers/wiki/...rDataRaces I'm reading it now 07:59
dogbert11 samcv: This is Rakudo version 2020.07-64-g9d6d8dd7a built on MoarVM version 2020.07-19-g9a5203395 10:43
dogbert11 could it be a problem if two threads are trying to write to the same address here: github.com/MoarVM/MoarVM/blob/mast...aque.c#L21 14:40
jnthn Not from the point of view of VM integrity, no 14:41
dogbert11 so they can't clobber something
nwc10 jnthn: I hope that your beer fridge operative is doing better than mine - I am now drinking the only cold beer that was in the fridge. 14:42
next release is in 4 days? 14:43
dogbert11 jnthn, nwc10: here's a gist showing the (non existing ?) problem: gist.github.com/dogbert17/36d5c6e2...d8ea243bdd 14:44
nwc10 I still don't feel at all comfortable with "what tsan is trying to tell me" 14:46
but I saw a lot of output from ubsan that I think I can address.
dogbert11 alignment stuff?
nwc10 yes 14:47
dogbert11 perhaps tsan is one big false positive :-)
nwc10 my face doesn't fit into ASCII
I suspect that mostly it is not
dogbert11 the output above is from on old example written by jnthn: perl6-m -e "await (^80).map({ start { EVAL '/0|1|2/' } }); say 'ok'" 14:49
it has a tendency to fail 14:50
samcv dogbert11, what arguments do you compile tsan with? I tried "./Configure.pl --prefix=$HOME/raku --compiler=clang --tsan --debug" but I get this error on starting nqp/raku "MoarVM: Initialization of GC start condition variable failed" 14:51
dogbert11 samcv: perl Configure.pl --debug --tsan --prefix=../../install/ # I'm using gcc 14:52
jnthn nwc10: Hm, I fear there is only one in it at home too, though there's plenty of others I can add :) 15:01
jnthn uh-oh, getting home might become interesting, though 15:17
(huge storm just started)
samcv jnthn, same here 16:15
jnthn By now it's "light rain" but I'm still not so excited to walk in it... 16:21
nwc10 still blue skies and fluffy clouds here, but we can hear thunder 16:29
and the final section of the lawn is now seeded. 16:30

Powered by Cro and the Raku® Programming Language.

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

Thank you!