samcv hm i'm thinking of making the comments in pod 01:11
that actually looks pretty nice rendered in pod
rendering with p6doc: gist.github.com/samcv/4c17eb4d7a0b...5667adb1cf i think i'm gonna go for it
MasterDuke_ .ask masak the inconsistent examples in rt.perl.org/Ticket/Display.html?id=122990 now are errors. e.g., `say \(1) ~~ :(int as Str)` gives `===SORRY!=== Error while compiling -e Malformed parameter at -e:1 ------> say \(1) ~~ :(int⏏ as Str) expecting any of: constraint formal parameter`. is that ok or should the ticket be re-worded to make them work at all? 03:07
yoleaux MasterDuke_: I'll pass your message to masak.
lizmat Files=1192, Tests=56910, 207 wallclock secs (12.38 usr 4.53 sys + 1229.09 cusr 119.04 csys = 1365.04 CPU) 05:29
yoleaux 10 May 2017 09:29Z <nine> lizmat: I do in fact have a filter machine with a timer :) Guess what? In the sleepy state before going to bed, errors are made. Like...forgetting to put the pot in its place. Great way to ensure the next day starts as badly as possible ;)
brrt good * #perl6-dev 07:12
fun wished for all attendees at the perl toolchain summit
lizmat thanks 07:17
we're on #pts@irc.perl.org
brrt heh, might as well join :-) 07:33
chansen_ You may find this document <github.com/apple/swift/blob/master...sto.md> interesting, discusses strings in Swift. 07:51
[TuxCM] This is Rakudo version 2017.04.3-215-g6409ee582 built on MoarVM version 2017.04-62-g362277b7 08:07
csv-ip5xs 2.903
test 12.538
test-t 4.245 - 4.252
csv-parser 12.601
Geth rakudo/nom: 6ea2f12e1c | (Elizabeth Mattijsen)++ | src/core/Mixy.pm
Simplify Mixy.roll()

Don't need to check for definedness and elems twice.		 09:19
rakudo/nom: acb657fc06 | (Elizabeth Mattijsen)++ | 2 files
Exit R:Q.MIX-TOTAL-POSITIVE

It was made under the assumption it could be useful for Mixy.pick and Mixy.grab, but these methods are not supported on Mixy, so abstracting the functionality was not needed.
roast/master: 4 commits pushed by (Samantha McVey)++ 09:40
roast: 740ce8f3a9 | (Samantha McVey)++ | S15-nfg/GraphemeBreakTest.t
[unicode-grant] Add some usage info and clean up some incorrect comments

Test functions the same as previously.		 09:50
Zoffix m: my @a = ^10; say @a[none 3] 10:35
camelia none(3)
Zoffix Ah 10:36
NeuralAnomaly: status 10:52
NeuralAnomaly Zoffix, [✘] Next release will be in 1 week and 2 days. Since last release, there are 25 new still-open tickets (12 unreviewed and 0 blockers) and 242 unreviewed commits. See perl6.fail/release/stats for details
Zoffix NeuralAnomaly: blockers 10:58
NeuralAnomaly Zoffix, There is 1 release blocker. See perl6.fail/release/blockers
Zoffix, perl6.fail/131247 : )> in regex results in stuff being matched wrongly ( /)> . <(/ )
Zoffix The OP issue itself isn't a blocker, but I suspect there may be more capture-stuff broken. Gonna write more tests for it this weekened; marked ticket as blocker as a reminder 10:59
Geth rakudo/nom: c1e681958e | (Elizabeth Mattijsen)++ | src/core/Rakudo/QuantHash.pm
Fix copy-pasto in R:Q.MIX-ROLL

The last tested element would never be selected thanks to unnecessary Intifying.		 11:36
rakudo/nom: b9222061a3 | (Elizabeth Mattijsen)++ | 2 files
Make Mixy.roll(N) about 3x faster

  - this reverts acb657fc0607bc74253 , as we do need R:Q.MIX-TOTAL-POSITIVE
  - since we need to iterate over the hash anyway, why not use iterator
  - so use R:Q.MIX-ROLL always
  - also get rid of now unneeded R:Q.WeightedRoll class		 11:46
dogbert11 .ask Zoffix any comments about Prey? 12:13
yoleaux dogbert11: I'll pass your message to Zoffix.
Zoffix . 12:14
yoleaux 12:13Z <dogbert11> Zoffix: any comments about Prey?
Zoffix dogbert11: I didn't buy it yet. I looked through steam reviews and some people having issues with their save files being corrupted after 20 hours. I figured I got other games ATM and just placed Prey on wishlist; maybe I'll snag it on sale some time in the future 12:15
dogbert11 Zoffix: interesting, saw some tests on youtube where they said that the game performs really well (fps wise) even on 'budget' hardware 12:28
Zoffix The only game that performed poorly on my hardware so far is Mass Effect Andromefail
Well, it performed OK after I shut down my VM to free more RAM, but it's the only game that required doing so 12:30
timotimo Zoffix: you may really want to update your weechat 12:31
Zoffix Why?
timotimo there's been a security vulnerability in versions up to 1.7
also, there's been versions between your 1.0.1 and 1.7.1 :) 12:32
Zoffix Which one? The one freenode msged about? I don't care about it.
timotimo i didn't know freenode msged about that
oh, yeah
there
dogbert11 runs WeeChat 1.2 12:33
timotimo has 1.7 still
Zoffix runs whatever's in the repo
timotimo would just have to /upgrade, though
Zoffix Setting up weechat (1.0.1-1+deb8u1) ... 12:34
m: class { has Str() $.host = '0.0.0.0' } 13:05
camelia 5===SORRY!5=== Error while compiling <tmp>
Coercion Str(Any) is insufficiently type-like to qualify a variable
at <tmp>:1
------> 3class { has Str() $.host7⏏5 = '0.0.0.0' }
expecting any of:
constraint
Zoffix waaat
TWEAK it is, I guess 13:06
jnthn Coercion types are only supported on parameters
Zoffix That's too bad 13:07
Zoffix removes `()` parts and stops careing :)
jnthn They were added to Perl 6 fairly late on, replacing the parameter syntax `Foo $x as Bar`
So they were implemented simply by deconstructing them to use the same internals that `as` did 13:08
And `as` was tossed before 6.c
At some point we'll have to try and figure out a good way to generalize them 13:09
But it's a bit icky
Zoffix cool
dogbert11 jnthn: have you seen Zoffix report github.com/MoarVM/MoarVM/issues/594 ? 13:11
jnthn Glanced it last night 13:18
dogbert11 if you have a hint, given that there's some kind of cast missing, we could do a PR 13:19
Geth rakudo/nom: 87a95fc135 | (Elizabeth Mattijsen)++ | 2 files
Make BagHash.grab(N) about 10% faster

  - also make it lazy: only items grabbed are actually removed
  - abstract grabbing logic into R:Q.BAG-GRAB
  - give another 20% improvement when moving to containerless BagHashes		 13:31
tadzik timotimo: oi, does your profiler-tui require some additional rakudo patches? 14:35
says no such column: callee_id" when I try it 14:38
Zoffix tadzik: are you sure you got latest nqp? That sounds like you might be missing one of the recent updates to its SQL profiler 14:51
timotimo tadzik: it does not 14:52
oh
there was just a rename
perhaps the tui needs to be updated for latest nqp? 14:53
MasterDuke_ timotimo: is your tui operational? src somewhere? 14:59
timotimo it's not interactive yet
it only shows a little part of the call graph
MasterDuke_ ah 15:00
timotimo github.com/timo/p6profiler-tools
i've decided against making arrow keys and similar things work and just let you navigate with hjkl
MasterDuke_ heh, taking a hard line on the vim/emacs debate i see 15:02
tadzik timotimo: nah, I have the fresh stuff; renames did help, yes
timotimo: gist.github.com/tadzik/80060834fa0...53cfcbe94e 15:03
timotimo EZ
tadzik but can I config it to use hnei on my colemak? :(
timotimo only if i'm allowed to use liae
tadzik you have my blessing 15:04
also, how did you make it so that when I do SIGINT it prints "Buf[uint8].new(3)" :D 15:05
timotimo makeraw means i get ctrl-c before the shell does ... or something?
tadzik hm, mebbe 15:07
timotimo perhaps all i need is noecho, though 15:10
tadzik now that I look at it, valgrind.org/docs/manual/cl-format.html doesn't look that scary 15:12
it won't be useful for all of the profiling data we get, but being able to just use qcachegrind would be sick
timotimo feel free to write something 15:13
tadzik yep, mite do 15:14
timotimo cool
did you see the graphviz output for the sqlite data?
tadzik I saw the file :P Haven't run it yet 15:17
timotimo then allow me to ...
give you this link: hack.p6c.org/~timo/forestfire_callgraph_4.svg
Geth rakudo/nom: 788e6de6dc | (Elizabeth Mattijsen)++ | 8 files
Bind all values in Baggys

  - (Bag|Mix)Hash no longer have containers, so easier on memory
  - updates are handled by special iterators or proxies
  - simplifies Baggy creation/handling significantly
  - removes !PAIR specific methods, just use Pair.new always
timotimo (it'll start you out on the top left where there's nothing)
tadzik mm, quite a graph :) 15:18
sweet
timotimo isn't it
it's amusing that almost every routine has an add_to_cache call coming off of it 15:19
might want to have a filter option for those
Zoffix timotimo: what code is that graph for?
Curious to see ->[Numeric]->[Numeric] in it 15:20
tadzik Zoffix: forestfire, I guess :)
timotimo yup
Zoffix What's that?
timotimo hover your mouse over the boxes to see the filename
it's the forest fire cellular automaton you can find on rosettacode
Zoffix Ah, those are Numerics of different types
timotimo yup 15:21
MasterDuke_ can anyone else profile forestfire? it's segfaulting for me 15:34
Zoffix wooo \o/ Tardigrade just took its first breath :) An app with two static pages: temp.perl6.party/tg.gif 15:45
And the code for it is just `use TG; TG.new.get('/' => 'index').get('/hello').start`
:)
timotimo MasterDuke_: throw out the signal tap 15:46
Zoffix Well, I was hesitant in the past year, but I guess I *am* writing a web frame work in Perl 6, 'cause I need it for $work and current options don't make me smile
MasterDuke_ timotimo: yeah, thanks, just figured that out 15:47
timotimo i've got some other local changes somewhere 15:48
Zoffix: one-line web apps, the wave of the future? 15:54
Zoffix Yeah, Serverless is currently the hot trend, I'm predicting Codeless will be next :p
llfourn Userless is the way of the future. 15:55
MasterDuke_ timotimo: are blocks, i.e., { }, free in nqp? i don't think they are, so `if $foo { ... }` is slower than `... if $foo` even in nqp, correct? 16:00
timotimo nqp is a lot better at inlining stuff
so it could very well be that such a block is free 16:01
MasterDuke_ but not guaranteed?
timotimo correct
Geth rakudo/nom: 3fb3c27ed3 | (Elizabeth Mattijsen)++ | 3 files
Remove a lot of superstitious decont()s

Probably some case of copy-pasto/cargo culting		 16:03
MasterDuke_ are they more expensive even if the block is never descended into? 16:04
timotimo i expect a containing block would start off with a "takeclosure" op for each inner block
that allocates a BOOTCode usually
MasterDuke_ i understand some of those words 16:05
so `if $foo { ... }` could make things slower even if $foo is never true? 16:06
timotimo possibly 16:07
if the block is inlined fully, it won't make things slower
perlpilot since the grant for replacing the engine behind blogs.perl.org stalled, it would be neat if someone could put together a Perl 6 solution. :) 16:42
stmuk or just got "november" working 16:47
Zoffix is perfectly happy with perl6.party
In fact, I think the repo even has a beta-quality Perl 6 version of it, back when I gave up due to memory leaks last year 16:48
stmuk s/got/ge 16:53
t
tadzik tadzik.net/pub/VVqTQ082Wm.png 16:59
how's that \o/
MasterDuke_ !! nice 17:00
tadzik the inclusive time (at least) is off, but the call structure seems to be preserved 17:02
lizmat cool 17:11
Zoffix Interesting that none of the DateTime::Format modules did the obvious thing and just provide routines one could give to DateTime.format attribute 17:58
I meant formatter attr
Not very comfortable that attr is 17:59
m: DateTime.now.clone(:formatter{sprintf "%02d/%02d/%04d", .day, .month, .year}).say 18:00
camelia 11/05/2017
Zoffix
.oO( DateTime.now.format({sprintf "%02d/%02d/%04d", .day, .month, .year}).say )
meh
m: sprintf "%02d/%02d/%04d", .day, .month, .year with DateTime.now 18:05
camelia ( no output )
Zoffix So I guess printing DateTime objects you got from somewhere isn't reliable, since they could have custom formatters 18:06
m: Date.today.clone(:formatter{ .day }).say 18:07
camelia 2017-05-11
Zoffix ZOFFLOP: t/spec/S11-modules/nested.t 18:17
Geth rakudo/nom: a9a161aef4 | (Zoffix Znet)++ | src/core/Date.pm
Make Date.clone take a formatter

  - It's one of the easiest ways to alter the formatter
   on an existing Date
  - DateTime.clone does take a formatter already		 18:22
roast: f304b9910a | (Zoffix Znet)++ | S32-temporal/Date.t
Test Date.clone can take a formatter

Rakudo fix: github.com/rakudo/rakudo/commit/a9a161aef4
MasterDuke_ jnthn: are you around? did you happen to see my question yesterday about the change in error message from `my %h{Any} of Str; %h<a>=2` caused by github.com/rakudo/rakudo/commit/06...7ee518b737 18:44
c: 06729d9~1 my %h{Any} of Str; %h<a>=2 18:48
committable6 MasterDuke_, ¦06729d9~1: «Type check failed in assignment to %h; expected Str but got Int␤ in block <unit> at /tmp/WQi_IzXz30:1␤ «exit code = 1»»
MasterDuke_ c: 06729d9 my %h{Any} of Str; %h<a>=2 18:49
committable6 MasterDuke_, ¦06729d9: «Type check failed in binding assignval; expected Str but got Int␤ in block <unit> at /tmp/poGJVjlhFd:1␤ «exit code = 1»»
Zoffix ZOFVM: Files=1229, Tests=135468, 114 wallclock secs (22.04 usr 3.25 sys + 2329.11 cusr 172.40 csys = 2526.80 CPU)
MasterDuke_ jnthn: ^^^ lost the name of the variable
timotimo: i noticed that a profile of forestfire had bind_one_param as the top by exclusive time, so i tried making a bunch of its ifs postfix, but that actually resulted in 2million more frames created. any idea why? here's the diff gist.github.com/MasterDuke17/bc1c7...228154956c 19:12
Zoffix One thing I miss from Perl 5 is sub import... Can't for the life of me make `use ... ` line take the args I want it to take 19:21
Even passing `[42, :bar<70>]` as positionals make it complain about there being no :bar tag :S 19:22
timotimo MasterDuke_: i have changes locally that make bind_one_param almost completely disappear 19:36
making the lists flat instead of two-elements for all the cases where it has $x, $y makes that work
MasterDuke_ timotimo: nice. i've seen bind_one_param in other profiles though, so making it faster would also be good 19:38
don't understand why my patch made it worse
Zoffix Random fact: Rakudo's version is NOT `before` v42 19:42
Zoffix found it out while writing some tests :)
that failed :()
timotimo MasterDuke_: yeah, we hit it when we can't lower a signature for some reason (there are multiple things the lowerer just refuses to try) 19:47
jnthn c: 06729d9~1 my Str %h{Any}; %h<a> = 2 20:26
committable6 jnthn, ¦06729d9~1: «Type check failed in binding assignval; expected Str but got Int␤ in block <unit> at /tmp/SMorwcYVNR:1␤ «exit code = 1»»
jnthn c: 06729d9 my Str %h{Any}; %h<a> = 2
committable6 jnthn, ¦06729d9: «Type check failed in binding assignval; expected Str but got Int␤ in block <unit> at /tmp/1z8Y0hmleW:1␤ «exit code = 1»»
jnthn MasterDuke_: You fixed the bug that means, I guess, that the TypedHash[::TValue, ::TKey] didn't get mixed in
MasterDuke_: Now that it was been, the type check in the signautre at github.com/rakudo/rakudo/blob/eb1c...sh.pm#L582 fails 20:27
And so it gives the same error as the case without the "of"
So we got the better error due to a bug prior to this :P 20:28
MasterDuke_ i fixed something?
jnthn You made `my %h{Any} of Str` behave the same as `my Int %h{Any}` 20:31
See my evals above
MasterDuke_ you mean github.com/rakudo/rakudo/commit/06...7ee518b737 ?
that was you 20:32
jnthn oh wow 20:33
I presumed you were asking me about a commit you'd done :P
Didn't even look at who authored it :D
But yeah, it's a fix
MasterDuke_ c
c: all my Str %h{Any}; %h<a> = 2 20:34
jnthn That made things better, except in the case it made it works :)
*worse
committable6 MasterDuke_, gist.github.com/387791da733e822f6d...0675f95c60
MasterDuke_ definition of an upgrade...old bugs out, new bugs in :)
huh, last releases that show '%h' were 2014.02,2014.03.01 20:35
then a bunch of `cannot stringify this` until 2014.09 20:37
jnthn m: dd :p[1] 20:41
camelia block <unit>
jnthn m: dd (:p[1])
camelia :p([1])
jnthn Zoffix: rt.perl.org/Ticket/Display.html?id=131279 doesn't work because :p[1] is the syntax for constructing an adverb with a value that is the array [1] ;) 20:42
timotimo m: dd 20:43
camelia block <unit>
timotimo ^- i still find this kind of a puzzlin feature
Zoffix doh :) 20:44
Totally slipped my mind :)
m: sub foo { dd }()
camelia sub foo
Zoffix m: use Test; cmp-ok "foo", '~~', *.contains('bar') 21:35
camelia not ok 1 -

# Failed test at <tmp> line 1
WhateverCode object coerced to string (please use .gist or .perl to do that)
in sub cmp-ok at /home/camelia/rakudo-m-inst-1/share/perl6/sources/C712FE6969F786C9380D643DF17E85D06868219E (Test) line 243…
Zoffix wonders if there's a good way to improve that... 21:36

Powered by Cro and the Raku® Programming Language.

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

Thank you!