00:12 dalek joined
dalek arVM: 18c8ca6 | Coke++ | build/ (2 files):
make build in dir containing space work
 00:34
00:37 retupmoca joined 00:57 travis-ci joined
travis-ci MoarVM build failed. Coke 'make build in dir containing space work' 00:57
travis-ci.org/MoarVM/MoarVM/builds/135485952 github.com/MoarVM/MoarVM/compare/7...c8ca63a8f6
00:57 travis-ci left
timotimo [Coke]: do you see that? 01:02
01:47 ilbot3 joined 01:53 FROGGS_ joined 02:23 colomon joined 02:38 colomon joined 02:42 colomon joined
nwc10 timotimo: I get that. "@moarshared@" is expanding to "" 06:10
instead of zero characters 06:11
and gcc doesn't like that empty argument
I don't know what the fix is, because I don't know where/when/why @moarshared@ is useful
and the main coffee machine at work is broken
so productivity today make take a hit 06:12
06:37 mtj_ joined 07:22 rurban joined 07:24 zakharyas joined
jnthn nwc10: The main one? So there's backups? :) 08:05
nwc10 I think only backup singular 08:06
the other one in the store cupboard appears to have been cleaned up and removed
and the "backup" (the manual one that I usually use) seemed to be playing up about 5 minutes ago
but yes, we're not yet shuffling around like zombies 08:07
but there might be a queue for the working machine
and a bunch of folks might need to learn how to use it
08:38 zakharyas joined 11:31 rurban joined
nwc10 I am pleased to report that the coffee machine engineer has finally been able to depart, and has explained proper TLC techniques to the admin staff 11:43
(this is actually slow news on my part - he left about 1h45 ago)
jnthn Oh, it wasn't cared for proper? 11:44
nwc10 it seems that the knowlege wasn't handed over from previous staff 11:45
at least, not well enough
nine If only writing had already been invented... 12:22
nwc10 I suspect that even "TFM" would help, if it could be found. 12:23
12:54 brrt joined
brrt good * 12:54
12:56 zakharyas joined
jnthn o/ brrt 13:10
lizmat jnthn brrt nwc10 o/
lizmat hopes for another blog post before tonight's P6W deadline 13:11
brrt \o lizmat jnthn nwc10 13:12
brrt is going to try to split node and value descriptor live ranges 13:14
theres quite a bit involved in that if you indulge me 13:21
we currently create and assign value descriptors in the live range calculation step 13:23
assign them to tiles 13:24
and then in the second step i assign a register to each value descriptor (as they are created) 13:25
however, what i want is to calculate just the live ranges and then create a value descriptor as we assign the register 13:27
finally, although not completely necessary, i'd like to get rid of tbe stack of 'active' value descriptors 13:29
and replace them with an array/heap of end ranges per registet 13:30
since the number of registers we have is small and fixed, a heap may be overkill 13:31
[Coke] Sorry, I hit commit and walked away. 13:39
is it complaining for anyone else, or just travis? 13:40
nwc10 me and travis. by "me", CentOS old-ish
(is there any other sort of CentOS?)
it's only "my" machine, not my machine. I don't get to choose the distro. 13:41
timotimo i can't build moar now either 13:43
well, of course i can by reverting the commit, but when i decided to check it out locally it asploded
[Coke] if it's just moarshared, I can move the "" logic into Configure.pl for that one, I think, to avoid the empty string.
nwc10 Without looking at the code (but with looking at too much work code) I hope that that means something like "we decide whether to use "" to wrap it more smartly at interpolation time" and not just "we start carrying around an escaped value internally" 13:45
13:45 brrt joined
nwc10 (work code is crazy and has far to many wrong ideas about when stuff should be escaped/unescaped) 13:45
however, really, please do whatever is easiest to make it work, and clear to a future maintainer
[Coke] No, I was just going to slap something together to get the build passing. ah well.
nwc10 ah OK.
[Coke] fixed. 13:46
dalek arVM: 0364d75 | coke++ | build/ (2 files):
Revert "make build in dir containing space work"

This reverts commit 18c8ca63a8f64fd0f6f2dbeda0f2b330dca253ce.
Fixes the build on OS X in a dir with a space, but breaks travis.
brrt just heard two very, very, very loud bangs in the city 13:47
timotimo uh oh
[Coke] hurls github.com/MoarVM/MoarVM/issues/374 instead. 13:48
[Coke] will try out any future updates on a branch instead.
brrt yes, and followed by a siren 13:49
no news though
timotimo very loud bangs, could always be car collisions. in this case maybe trucks rather than regular cars?
brrt louder than that 13:51
and two of them
the news page was just updated 13:52
'what was it'
journalism in 2016
it was, an airplane 13:55
timotimo o_O
how the hell did that happen
brrt it was expected
and it was announced an hour ago 13:56
timotimo "we're going to be exploding an airplane, so plug your ears please"
nine sonic boom?
timotimo ah, that could be
brrt sonic boom, yes
nine Was actually my instant first idea when I read two very loud bangs :) But I'm an aviation nerd ;) 13:57
brrt folks were already assuming this was an earthquake
i was assuming gas explosion
oh well :-) 13:59
jnthn nine: Is there being two of them significant/a clue in some way? :) 14:00
brrt apparantly the front and the back of the airplane make different bangs each
jnthn Ah :)
timotimo that's interesting to know 14:01
14:10 travis-ci joined
travis-ci MoarVM build passed. Will "Coke" Coleda 'Revert "make build in dir containing space work" 14:10
travis-ci.org/MoarVM/MoarVM/builds/135596029 github.com/MoarVM/MoarVM/compare/1...64d75f28a5
14:10 travis-ci left
dalek arVM/smoke-me/spaceybuild: 092d5da | Coke++ | build/ (2 files):
make build in dir containing space work
 14:54
[Coke] there we go, can leave it there until I iron it out. 14:55
dalek arVM/smoke-me/spaceybuild: 71be0f0 | coke++ | build/ (2 files):
moarshared isn't a simple path.

only quote the path-y parts		 15:04
[Coke] that might do it. 15:05
wonder why travis doesn't report the failure as soon as it knows a build failed (rather than waiting for all configurations to finish) 15:12
timotimo good question 15:13
especially annoying with moar and clang, because it's so slow
15:17 travis-ci joined
travis-ci MoarVM build failed. Coke 'make build in dir containing space work' 15:17
travis-ci.org/MoarVM/MoarVM/builds/135614759 github.com/MoarVM/MoarVM/commit/092d5daecf65
15:17 travis-ci left
[Coke] That's the original commit moved to the smoke-me branch... 15:19
15:40 travis-ci joined
travis-ci MoarVM build passed. Will "Coke" Coleda 'moarshared isn't a simple path. 15:40
travis-ci.org/MoarVM/MoarVM/builds/135617713 github.com/MoarVM/MoarVM/compare/0...be0f07f0ca
15:40 travis-ci left
[Coke] there we go. 15:51
TimToady one also notes that the booms usually come about a quarter second apart, which is longer than you'd expect taking the plane's length into account 15:52
[Coke] create a PR from a branch that was already tested by travis... get another travis check. wtf. 15:53
TimToady so either the second shock goes slower than the first, or it's from some phenomenon where the air meets up some place behind the plane
I suppose one could look it up these days...
(we used to get sonic booms regularly in CA when they were landing shuttles at Edwards) 15:54
I suppose distance and altitide could also have an effect on timing 15:55
TimToady wonders how much the first shockwave heats the air...
brrt hmm, that probably depends on the effective surface area and the speed of accleration 16:11
also, in this case, there were two of them 16:12
intercepting a passenger plane that had lost radio contact.. which has since been restored
TimToady well, was just wondering how much of the slowdown of the 2nd shock was thermal
one can scarcely blame them for going super for that... 16:13
plus you don't have as much or ocean to do it over there than we do here in CA :) 16:14
desert or ocean
well, don't want to OT this too much in a working IRC channel...
jnthn Hm, passenger planes losing radio contact for a while seems kinda commonish. There was a case here in .cz within the last month, and I'm pretty sure one in the UK in the last couple of months too... 16:22
brrt well, defense panicked in .nl, at least this time 16:37
obviously i have no idea how serious it was
it was just amazingly loud 16:38
dalek arVM/even-moar-jit: c217b9a | brrt++ | src/jit/ (2 files):
Split live ranges from value descriptors

At least conceptually, a value descriptor can have a quite different
  (typically shorter) lifespan than a live range does. This allows
multiple value descriptors per node, and in the future to manage spilled nodes.		 16:42
brrt ta-da
timotimo yay 16:44
brrt (wow, USA used 31.5% /less/ coal in march 2016 than in 2015 :-o) 16:45
timotimo is that a comparison between march 2016 and 2015 16:46
or is it a comparison between march 2016 and march 2015?
brrt last of these 16:48
timotimo OK :)
saving 31.5% on a 12x increase isn't really saving much :D
brrt right :-) 16:49
anyway, i want to split size out of value
brrt afk 16:50
timotimo :)
17:22 zakharyas joined 17:25 domidumont joined
TimToady guesses the coal numbers are a combo of El Niño and cheap oil effects 17:27
with a few regulatory monkey bars as well 17:28
er, I guess those'd be monkey wrenches
TimToady should reserve MONKEY-BAR and MONKEY-WRENCH for pragmas too... 17:29
timotimo MONKEY-FOO?
TimToady TYPING SEE-NO-EVAL BRAINS GUTS BUSINESS TRAP SHINE are current recognized 17:32
though I suspect it's always plural in common usage: monkey bars... 17:33
timotimo do -TRAP and -SHINE already have a meaning? i must have missed those
TimToady no, they're just suggested future ones
I think GUTS maps to 'use nqp' 17:34
timotimo right, i remember MONKEY-GUTS 17:35
TimToady so that 'use MONKEY' also turns that on
timotimo oh, i didn't know about MONKEY yet, either 17:36
TimToady we don't advertise it heavily, in case it turns out to be a bad idea :)
TimToady smiles every time someone takes the loving care to type MONKEY-SEE-NO-EVAL, when just MONKEY would do 17:37
but I suppose this is a bit OT for #moarvm as well... 17:38
jnthn wonders if use MONKEY-GUTS coulda been use PRIMATIVES; :P 17:40
TimToady that would not automatically turn on with 'use MONKEY;' 17:43
mst MONKEY-SPANK - from the same naming geniuses who brought you STD.pm 18:07
TimToady
.oO(this discussion is going down a monkey hole rapidly...)
 18:08
ilmari apologises for giving mst the idea in /msg
I was trying to keep it out of the gutter, but had to vent it somewhere
nine ilmari: and you tell it to someone named MST? ;) 18:09
TimToady to their credit, most monkeys stay out of most gutters
ilmari nine: I wasn't expecting him to leak it back when I said I was refraining from saying it in the channel 18:10
mst I didn't so much leak as ooze 18:11
ilmari
.oO( splurge )
TimToady ilmari: as long as it's officially off the record, that's okay :)
mst I don't even own a turntable
ilmari leaves the channel in the gutter and pubtimes 18:12
TimToady turntables have gone out of fashion for most railroads these days...
ilmari used to live next to the roundhouse, but it hasn't had a turnable in it for 150 years 18:13
TimToady this discussion seems to have gone off the monkey tracks... 18:14
ilmari well, lately it has had lots, since it's now a music venue
TimToady
.oO(use MONKEY-LIVE;)
 18:19
18:31 FROGGS joined 18:45 zakharyas joined 20:15 rurban_ joined 20:51 patrickz joined 22:29 tomboy64 joined

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!