github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
08:17 Kaypie left 08:19 Kaiepi left 08:54 patrickb joined 09:09 nebuchadnezzar joined 09:57 patrickb left 10:44 nine joined 10:45 camelia joined 11:28 domidumont joined 11:38 lucasb joined 12:06 MasterDuke left 12:49 MasterDuke joined 12:50 MasterDuke left, MasterDuke joined 13:12 patrickb joined
patrickb samcv: I've seen your PR. Great! I have just one more request, could you split it up into two files and add two entries to README.md? That'd be great! 13:13
timotimo how long until the project suggestions have to be finished? 13:14
AlexDaniel timotimo: summerofcode.withgoogle.com/how-it.../#timeline 13:19
so a day or two before Feb 6th?
timotimo thank you 13:20
patrickb 6th is the google deadline 13:26
4th is the one we set up to have some time to clean up things
timotimo OK
13:55 MasterDuke_ joined 13:56 patrickb left 13:58 MasterDuke left
timotimo brrt, you're not at fosdem, right? 14:51
brrt, the talk Right Now is "faster programs with guile 3", which i believe includes talk about a jit
live.fosdem.org/watch/k4201
15:28 lucasb left 16:00 domidumont left 18:12 MasterDuke_ is now known as MasterDuke, MasterDuke left, MasterDuke joined
Geth MoarVM: MasterDuke17++ created pull request #1043:
Jit (open|read|close)_dir
18:46
timotimo btw what more than an extra queue of reference addresses and a little bit of API do we need to actually have weak references in moarvm? 18:53
i'd imagine once we're through with the regular queue, we'd go through the list of weak refs and if the target has a forwarder update the pointer, otherwise redirect to VMNull for instance 18:54
19:07 zakharyas joined
timotimo maybe i should just prototype 19:26
20:26 zakharyas left 21:23 MasterDuke left 21:35 MasterDuke joined, MasterDuke left, MasterDuke joined
timotimo i guess it'd become problematic if you pass a weak reference to a sub and during binding the gc kicks in and what used to be some value is now null 21:41
though if there's only, for instance, a WeakHash, it'll only ever change from one nqp::atkey to another - though have fun with existskey, ugh 21:42
ehhh, too lazy to add a compose protocol to VMHash 22:04