github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
01:27 lizmat left 01:33 lucasb left 02:11 AlexDani` joined, p6bannerbot sets mode: +v AlexDani` 02:15 AlexDaniel left 02:48 AlexDani` is now known as AlexDaniel 06:54 reportable6 left 06:55 reportable6 joined 06:56 p6bannerbot sets mode: +v reportable6 07:05 domidumont joined, p6bannerbot sets mode: +v domidumont 08:03 brrt joined 08:04 p6bannerbot sets mode: +v brrt
brrt \o 08:04
yoleaux 16 Dec 2018 17:58Z <samcv> brrt: then we don't need to worry about libraries on the users computer changing?
brrt .seen samcv
yoleaux I saw samcv 16 Dec 2018 17:58Z in #moarvm: <samcv> .tell brrt then we don't need to worry about libraries on the users computer changing?
brrt .tell samcv checkout commit github.com/MoarVM/MoarVM/commit/66...1ba4b2e789
yoleaux brrt: I'll pass your message to samcv.
08:04 domidumont left 08:19 domidumont joined 08:20 p6bannerbot sets mode: +v domidumont 08:22 zakharyas joined 08:23 p6bannerbot sets mode: +v zakharyas 08:57 lizmat joined, p6bannerbot sets mode: +v lizmat 09:35 brrt left 09:59 brrt joined, p6bannerbot sets mode: +v brrt 10:08 brrt left 10:42 dogbert2_ left 11:03 domidumont left 11:54 dogbert2_ joined 11:55 p6bannerbot sets mode: +v dogbert2_ 11:58 lucasb joined, p6bannerbot sets mode: +v lucasb 12:39 zakharyas left 12:52 domidumont joined 12:53 p6bannerbot sets mode: +v domidumont
dogbert2_ github.com/libuv/libuv/blob/v1.x/ChangeLog # libuv 1.24.1 13:45
14:27 zakharyas joined 14:28 p6bannerbot sets mode: +v zakharyas 14:58 brrt joined 14:59 p6bannerbot sets mode: +v brrt
brrt \o 15:00
jnthn o/ brrt 15:06
brrt ohai jnthn 15:07
I forwarded an invite to a VM conference to your public email address. Did you receive that? 15:08
jnthn I did...then meant to look at it and reply over the weekend, then the weekend was busier than expected. :) 15:09
From the glance I had, it looks interesting though
Were you considering going?
brrt A bit
But I consider myself a bit undercredentialed 15:10
jnthn Ah, 'cus it's an academic-y conference? 15:12
I don't think I'm much more credentialed :P
brrt on the other hand, might be still very educational 15:13
also, it's Italy 15:15
jnthn :) 15:18
15:52 brrt left 16:09 zakharyas left
japhb (backlogging) timotimo, why do you say "int-to-num-to-int is safe"? Are the ints in question constrained to <= 53 bits, so as to fit in the num mantissa without loss of lower bits? 16:32
jnthn, brrt: Design and implementation of a real-world multi-threaded spesh/JIT VM (in active use!) seems to me to be a pretty strong credential. :-) 16:36
lizmat and another Perl 6 Weekly hits the Net: p6weekly.wordpress.com/2018/12/17/...-designed/ 16:46
16:57 domidumont left
timotimo japhb: i say it because using num for those was wrong in the first place and if the numbers get F'ed by being put into nums that'd already be a bug 17:08
japhb: but i'm not entirely happy with that thing, hence not having pushed it yet
japhb Ah, so this is in particular about the existing domain of the data, not a general statement about an available opt that will always work 17:10
timotimo what i totally can do is limit it to some situations where i know the op that writes the initial int is extremely unlikely to be wanted in num format 17:13
such as the elems op
iterating over an array until you're past elems(@foo) is probably a bit funky if you're numifying it 17:14
17:27 domidumont joined 17:28 p6bannerbot sets mode: +v domidumont
timotimo index_s r57(4), r45(3), r18(25), r57(3) 17:30
coerce_in r28(10), r57(4)
const_i64_16 r57(5), liti16(0)
coerce_in r27(9), r57(5)
lt_n r57(6), r28(10), r27(9)
like this for example
though if you have a string with 2**53 characters, maybe something has already gone wrong :)
m: say 2 ** 53 17:31
camelia 9007199254740992
timotimo probably won't fit the average user's ram
17:40 zakharyas joined 17:41 p6bannerbot sets mode: +v zakharyas
Geth MoarVM: 987269ea9f | (Jonathan Worthington)++ | src/core/interp.c
Get readnum working, try for 32-bit num support

This might need tweaking for big endian platforms, but is at least closer to being useful by working properly on little-endian ones.
18:16
18:53 zakharyas left 19:43 domidumont left 21:28 lucasb left
Geth MoarVM: d69e87b7fc | (Stefan Seifert)++ | src/core/interp.c
Use write_buf also when switching endianness

Calling write_buf once instead of multiple at_pos calls should be faster and also give the same error behavior whether we need to switch bytes before or not.
21:48