github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
timotimo jnthn: in order to make Tap more easily subclassable, the slow path may want to also take *%foo and pass it on to self.bless 10:37
jnthn timotimo: Feel free to tweak it as you need :) 11:46
Kaypie what am i doing wrong here? hastebin.com/sarusikuco.php 12:02
Kaypie i'm trying to make it so the file descriptor can be passed back to the callback that nqp::asyncconnect's given, but i get Too few positionals passed; expected 7 arguments but got 6 when i try to run it 12:02
jnthn Kaypie: That looks correct, unless you missed a place somewhere, or updated the wrong thing in Rakudo...what does your Rakudo diff look like? 12:04
Kaypie: Also, did you make install MoarVM so that Rakudo is using your changed one?
I don't see anything wrong with the patch 12:05
Kaypie it looks like this hastebin.com/abumoyehuv.rb
and it should be installed, but i'll double check
ah i forgot the prefix when i configured moar 12:10
jnthn aha :) 12:12
Yeah, the Rakudo side of the patch looks sensible too
Kaypie the fd being -1 when i test isn't very helpful, though
oh wait i was still using nqp::filenofh in the test file 12:13
Geth MoarVM: MasterDuke17++ created pull request #915:
JIT neg_I and abs_I
12:23
timotimo whoops 12:31
on JVM the order of peerhost/peerport and sockethost/socketport is switched, i.e. what you think is the remote address is now your local address 12:32
also, the failed callback for async listen doesn't push the four arguments for the hosts/ports, so it would probably not work 12:34
Geth MoarVM: 2e94688175 | (Timo Paulssen)++ | src/io/asyncsocket.c
asyncsocket_listen shall expose listen host&port

to be used by rakudo to report the actually-in-use port when binding to port 0.
Also includes a fix for a potentially invalidated pointer.
13:09
dogbert17 jnthn: MoarVM panic: Corrupt multi dispatch cache: cur_node != 0, re-check == (nil)
jnthn dogbert17: Thanks 13:12
timotimo i'm glad i made this 13:15
that should go in the changelog! 13:17
and it should go in the docs 13:18
and maybe even tests ...
and could even be used by testing code ... 13:19
travis-ci MoarVM build failed. Timo Paulssen 'asyncsocket_listen shall expose listen host&port 13:19
travis-ci.org/MoarVM/MoarVM/builds/406588120 github.com/MoarVM/MoarVM/compare/4...9468817580
timotimo oh crap 13:19
whoops
Geth MoarVM: fda0d8e2d3 | (Timo Paulssen)++ | src/io/asyncsocket.c
fix botched up commit
13:20
timotimo well, travis? 13:26
good sign; success takes a bit longer 13:29
travis-ci MoarVM build passed. Timo Paulssen 'fix botched up commit' 13:41
travis-ci.org/MoarVM/MoarVM/builds/406590695 github.com/MoarVM/MoarVM/compare/2...a0d8e2d3cc
brrt the windows failure is nonobvious 15:31
that is what I have to say about it today
i hope we're all not under a big release pressure
jnthn For those eagerly awaiting the follow-up blog post promised in my one on usage tracking, here it is: 6guts.wordpress.com/2018/07/21/mor...-tracking/ 16:00
dogbert17 jnthn++, now I have two posts to read :) 16:16
timotimo oooh that was quick 17:05
lovely
jnthn: maybe for the BB 3 put in a little explanation for the litui32(72) to guardconc, and maybe say a word or two about the fact that two "INS Deopt One" are in the same spot? 17:12
good post, jnthn++ 17:20
japhb jnthn++ # much posting! :-) 19:39
Geth MoarVM: jstuder-gh++ created pull request #916:
Fix Multiple Statevar Initialization Bugs
19:44
Geth MoarVM: 428ed0e2aa | MasterDuke17++ (committed using GitHub Web editor) | src/math/bigintops.c
Fix some typos in comments
20:52
jnthn pdfs.semanticscholar.org/d68e/b3e2...b89f8c.pdf # somewhat interesting read 23:44