Perl 6 - the future is here, just unevenly distributed

IRC log for #moarvm, 2016-11-25

| Channels | #moarvm index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:05 timotimo i'll try sleep again, maybe with more success than yesterday
00:08 jnthn Rest well :)
06:30 xtt joined #moarvm
06:35 domidumont joined #moarvm
06:37 xtt2 joined #moarvm
07:53 Ven joined #moarvm
07:55 Ven joined #moarvm
07:59 xtt joined #moarvm
08:10 geekosaur joined #moarvm
08:11 timotimo well, i'm up
08:15 nine But did you find some sleep?
08:18 timotimo yeah, there was some
08:21 xtt2 joined #moarvm
08:25 zakharyas joined #moarvm
08:32 Ven joined #moarvm
08:33 domidumont joined #moarvm
08:39 domidumont joined #moarvm
09:53 zakharyas1 joined #moarvm
09:56 Ven joined #moarvm
11:03 * viki clears throat
11:03 viki https://github.com/MoarVM/MoarVM/pull/442
11:13 * brrt checks out
11:14 brrt i have no idea why that should do what you mean :-)
11:16 viki You mean why 1 / 2**1020 should be a proper number and not 0?
11:16 viki r: use nqp; nqp::div_In(1, 2**1020)
11:16 camelia rakudo-moar 64343d, rakudo-jvm 76b061: ( no output )
11:16 viki r: use nqp; say nqp::div_In(1, 2**1020)
11:16 camelia rakudo-jvm 76b061: OUTPUT«8.9E-308␤»
11:16 camelia ..rakudo-moar 64343d: OUTPUT«0␤»
11:16 viki r: use nqp; say nqp::div_In(2**1020, 1)
11:16 camelia rakudo-moar 64343d: OUTPUT«Inf␤»
11:16 camelia ..rakudo-jvm 76b061: OUTPUT«1.1235582092889474E307␤»
11:17 viki And since that stuff is used by Perl 6's Rats, we get:
11:17 viki r: say 1/2**1020
11:17 camelia rakudo-moar 64343d: OUTPUT«0␤»
11:17 camelia ..rakudo-jvm 76b061: OUTPUT«8.9E-308␤»
11:17 viki r: say 1e0/2e0**1020
11:17 camelia rakudo-moar 64343d: OUTPUT«8.90029543402881e-308␤»
11:17 camelia ..rakudo-jvm 76b061: OUTPUT«8.900295434028806E-308␤»
11:17 viki etc
11:46 Ven joined #moarvm
11:52 viki .ask MasterDuke can you look at this ticket? IIRC we decided this was a DITHWIDT and wasn't worth a fix, right? https://rt.perl.org/Ticket/Display.html?id=128551
11:52 yoleaux2 viki: I'll pass your message to MasterDuke.
11:53 viki oh whoops, wc
12:08 Ven joined #moarvm
12:12 Ven joined #moarvm
12:21 Ven joined #moarvm
13:44 Ven joined #moarvm
14:02 Ven joined #moarvm
14:02 domidumont joined #moarvm
15:01 brrt viki: I mean, I have no idea why that patch works
15:01 brrt although I believe that it works
15:02 viki brrt: it works because when we do MAX, in the case of 1/2**1020 you end up with 1 / 2 ** 57, which is less than zero and so the numerator ends up being zero.
15:05 domidumont joined #moarvm
15:06 viki brrt: and the fix avoids that case, but I'm not 100% sure it's a good fix
15:15 Ven joined #moarvm
15:17 Ven_ joined #moarvm
15:33 brrt hmmm
15:34 brrt honestly, I have no idea how IEEE floating point works at all
15:47 Ven joined #moarvm
17:51 zakharyas joined #moarvm
19:35 domidumont joined #moarvm
19:49 timotimo i don't have an idea why IEEE floating points works at all ;)
19:50 moritz I don't hae any idea if IEEE floating points work at all ;)
21:27 zakharyas joined #moarvm
22:39 TimToady joined #moarvm

| Channels | #moarvm index | Today | | Search | Google Search | Plain-Text | summary