[00:11] *** Ven joined
[00:33] *** Ven joined
[00:47] *** Ven joined
[01:01] *** Ven joined
[01:13] *** Ven joined
[01:29] *** Ven joined
[01:44] *** Ven joined
[01:51] *** tokuhiro_ joined
[01:59] *** Ven joined
[02:13] *** Ven joined
[02:29] *** Ven joined
[02:47] *** ilbot3 joined
[03:53] *** tokuhiro_ joined
[07:57] *** TimToady joined
[07:58] <timotimo> JimmyZ: i'm looking through moarvm issues on github and i've found your issue 150; a script to generate a sphere as a pgm. i don't understand the script, though. it has "until start { ... } { write output data; sleep 1 }" in it, that's quite weird, IMO

[08:00] <timotimo> holy crap, Blob.new: @pixels is CRAZY slow

[08:02] <timotimo> ah, that's why.

[08:13] <JimmyZ> timotimo: the code is not mine, I saw it in the irc log

[08:14] <timotimo> mhm

[08:14] <timotimo> ah, i'm not sure if i can actually get away with making Blob.new: @foo faster using nqp::splice

[08:51] *** tokuhiro_ joined
[09:03] <dalek> MoarVM: 4a35f2c | timotimo++ | src/io/syncfile.c:

[09:03] <dalek> MoarVM: handle files where stat falseley reports 0 bytes size

[09:03] <dalek> MoarVM: review: https://github.com/MoarVM/MoarVM/commit/4a35f2cc3d

[09:48] <arnsholt> What's the bottleneck, OOC? Getting data out of @pixels, or into the Blob?

[10:10] <timotimo> oof, good question

[10:11] <timotimo> the thing is, the original script i was looking at could benefit a whole lot just by using a Buf[int8] instead of a regular array for this

[10:11] <timotimo> in that case, there was no need to create a blob in the first place

[10:52] *** tokuhiro_ joined
[11:01] *** leont joined
[11:21] <arnsholt> Right, right

[11:21] <arnsholt> Once again, it turns out that choosing the right data structure is kind of important =)

[11:43] *** Ven_ joined
[11:43] *** vendethiel joined
[11:53] *** tokuhiro_ joined
[12:01] *** Ven_ joined
[12:14] *** Ven_ joined
[12:30] *** Ven_ joined
[12:45] *** Ven__ joined
[13:54] *** tokuhiro_ joined
[14:06] *** vendethiel joined
[14:41] *** Ven joined
[14:44] *** Ven joined
[15:20] *** Ven joined
[15:30] *** Ven joined
[15:45] *** Ven joined
[16:00] *** Ven joined
[16:16] *** Ven joined
[16:29] *** diakopter joined
[17:04] <timotimo> jnthn: something has been bugging me: we have utf8-c8 to allow things to roundtrip through Str that potentially contains binary garbage, because filesystems may allow filenames with "anything but slash or null" in them ... but how do we roundtrip given that we do NFC and then NFG in order to create a Str?

[18:14] *** brrt joined
[18:14] <brrt> good *

[18:14] <brrt> is jnthn available?

[18:22] <brrt> i'm.. having a nasty issue

[18:23] <brrt> conceptually, at least

[18:24] <brrt> in OSR, and deopt, and various other purposes, i may want to put a label directly before of after an instruction

[18:26] <brrt> hmm

[18:26] <brrt> maybe it isn't a problem after all

[18:28] <brrt> it just, *sigh*, requires me to rework some of the funky decisions i made in the original jit labelling code

[18:29] <brrt> thereby demonstrating that i'm an idiot

[18:30] <arnsholt> Probably not =)

[18:30] <arnsholt> Just that compilers are hard

[18:51] <brrt> yeah, and people naive :-)

[18:52] <brrt> except that i have a vested interest in people believing that compilers are not, in fact, hard

[18:52] <brrt> and they aren't

[18:52] <brrt> you just have to do not the stupid things i did

[18:52] <brrt> :-P

[18:54] <brrt> anyway, i'd like everyone to believe it is really simple, so they all feel they can help out

[18:54] <brrt> obviously ending with me doing nothing

[18:55] * brrt afk now

[19:00] *** vendethiel joined
[19:03] *** lizmat joined
[19:13] *** lizmat_ joined
[19:31] *** lizmat joined
[19:47] *** tokuhiro_ joined
[20:33] *** Ven joined
[20:40] *** ShimmerFairy joined
[20:53] *** tokuhiro_ joined
[21:02] *** Ven_ joined
[21:03] *** Ven joined
[21:32] *** Ven joined
[21:43] *** zakharyas joined
[21:52] *** Ven joined
[22:02] *** Ven joined
[22:20] *** Ven joined
[22:54] *** tokuhiro_ joined