»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, or /msg camelia p6: ... | irclog: irc.perl6.org or colabti.org/irclogger/irclogger_logs/perl6 | UTF-8 is our friend!
Set by moritz on 22 December 2015.
Zoffix I got some proptests failing, but they look iffy to me. A shift on `int` variable expects -170141183460469231731687303715884105728 as a value 00:02
m: use nqp; dd nqp::bitshiftr_i(-9223372036854775808, -64) 00:03
camelia -9223372036854775808
Zoffix m: use nqp; dd nqp::bitshiftr_I(-9223372036854775808, -64, Int)
camelia -170141183460469231731687303715884105728
Zoffix ZofBot: @expletives.pick 00:04
ZofBot Zoffix, m
MasterDuke i think so, but i didn't record what actually failed
MasterDuke maybe there was another file also? 00:04
Zoffix Some of the tests in that bunch are already fudged with RT#125466 00:05
synopsebot RT#125466 [open]: rt.perl.org/Ticket/Display.html?id=125466 [MATH][@LARRY] bitwise shift is inconsistent on int
Zoffix "There are three fudges where it's unknown what the proper behaviour should be (overflow or switch to Int?)." yeah, mine fall into that category too, so I'm just gonna fudge them (they're proptests, not spectests) 00:06
Zoffix votes "overflow with natives" while at it
.ask TimToady what's your opinion, should `use nqp; dd nqp::bitshiftr_i(-9223372036854775808, -64)` overflow and return -9223372036854775808 or switch to an Int instead -170141183460469231731687303715884105728? Right now we have this inconsistency between Int and int candidates, but I'm leaning towards: people might wants overflowing stuff 00:11
yoleaux Zoffix: I'll pass your message to TimToady.
Zoffix MasterDuke: rindolf the native shift candidates should now work, as of github.com/rakudo/rakudo/commit/3d73597556 00:17
cfa: are you still trying to build locally? Just commit 00:22
$forgiveness > $permission is how we like to say. `git revert` works, after all 00:23
lookatme :) 00:26
Geth doc: cb4a69905b | cfa++ | 2 files
Clarification on subscript adverbs.
 00:28
cfa hope that's okay?
i just wanted to test those links so waited on the full build, yeah
TEttinger I would vote for overflow as long as JS compatibility isn't a concern
my vote doesn't really count though
Zoffix TEttinger: why overflow? You couldn't for example do `my int64 @a = -9223372036854775808; @a[0] +>= -64` for example. It'd explode 00:32
cfa: looks fine to me 00:33
Zoffix m: my int64 @a = -9223372036854775808; @a[0] +>= -63; dd @a 00:34
camelia Cannot unbox 127 bit wide bigint into native integer
in block <unit> at <tmp> line 1
perlawhirl clickbaits www.0racle.info/articles/pick_and_choose_part1 00:35
happy for any grammar or logic corrections before I throw this on reddit 00:36
Zoffix TEttinger: ah, you can't do that anyway :( the meta-assign op gets boxed
m: my int64 @a = -9223372036854775808; @a[0] = @a[0] +> -63; dd @a
camelia Cannot unbox 127 bit wide bigint into native integer
in block <unit> at <tmp> line 1
Zoffix Ah, wait, we box on array access so far, don't we 00:37
Zoffix m: my int64 $a = -9223372036854775808; $a = $a +> -63; dd $a 00:37
camelia -4611686018427387904
Zoffix OK, yeah, I'm changing my vote on that the two should behave the same.
cfa hmm, it looks like Array tucks slice adverb documentation away under grep
TEttinger Zoffix: overflow matters for some algorithms and there should at the very least be an option
Zoffix TEttinger: oh, right I misread what you originally said. 00:38
TEttinger JS is just a mess regarding ints and overflow
Zoffix Yeah, that was my original reasoning too, but the above evals show this isn't a good method to rely on when it comes to controlling overflow. 00:39
TEttinger since Number doesn't actually support 64 bits of integer precision on JS
cfa actually, that's coming from list
MasterDuke perlawhirl: "is a branch mathematics", missing an "of" i assume 00:40
Zoffix .tell TimToady I changed my leaning; I think both HLL and native should behave the same as autoboxing makes stuff too unpredictable as can be seen here: irclog.perlgeek.de/perl6/2018-03-05#i_15883977
yoleaux Zoffix: I'll pass your message to TimToady.
Zoffix &
TEttinger I do like that GWT, when emulating Java's types with JS, will correctly handle long (64-bit signed integers) in the browser despite taking a performance hit to do so 00:41
TEttinger correctness is important 00:41
MasterDuke perlawhirl: "so it's can" 00:42
perlawhirl MasterDuke++ thanks, fixed 00:47
Zoffix MasterDuke: rindolf wait, so wtf is "run('./primesieve', $n, '-p1', :out)". Is that some program I need to install to run that example program? 00:50
Geth doc: ed980c7f69 | (Tom Browder)++ (committed using GitHub Web editor) | doc/Language/nativecall.pod6
tweak grammar, punctuation, spelling, format
 00:50
synopsebot Link: doc.perl6.org/language/nativecall
rindolf Zoffix: yes - primesieve.org/
Zoffix Thanks.
I see P5 version has them hardcoded: github.com/shlomif/project-euler/b..._618_v1.pl 00:51
I think I'll that that version and try to convert it to P6 and see how it compares to yall's versions
:)
Zoffix 0m17.789s for p5 version... I can beat it :P 00:52
cfa okay, closed off the issue---i think the operator link from signatures and the hash clarification are sufficient 00:53
otherwise this'll get way too noisy
Geth doc: 5c99779397 | (Richard Hainsworth)++ (committed using GitHub Web editor) | doc/Language/nativecall.pod6
Minor bit logic error
 01:48
synopsebot Link: doc.perl6.org/language/nativecall
Zoffix Well, I failed. Got 33s to P5's 17.7s 02:28
m: say 33/17.7
camelia 1.864407
Zoffix Still 86% slower, despite being NQP gibberish 02:29
.tell rindolf managed to get P6 version to 1.86x the time of P5 version by rewriting it in NQP, if you were curious: gist.github.com/zoffixznet/cb19c86...19f7ee9383
yoleaux Zoffix: I'll pass your message to rindolf.
MasterDuke where does a profile say the time is being spent?
rindolf Zoffix: hi
yoleaux 02:29Z <Zoffix> rindolf: managed to get P6 version to 1.86x the time of P5 version by rewriting it in NQP, if you were curious: gist.github.com/zoffixznet/cb19c86...19f7ee9383
Zoffix Hmmmm 02:30
Actually, I think that might be 'cause we don't buffer TTYs
rindolf Zoffix: holy shit
rindolf Zoffix: that looks really cluttered 02:31
Zoffix hm, no nevermind, 0m33.353s with the say commented out (I was piping to /dev/null anyway)
rindolf Zoffix: ah
Zoffix It looks really nice and even :) 02:32
.tell MasterDuke FWIW rindolf's program in NQP version; still 1.86x slower. Maybe some of these ops don't get JITted? No idea how to check: gist.github.com/zoffixznet/cb19c86...19f7ee9383 02:33
yoleaux Zoffix: I'll pass your message to MasterDuke.
Zoffix goes to bed
jdv79 m: {}<sadf>.sort.head(2).perl.say 03:25
camelia (Any,).Seq
jdv79 why does that have an element?
or, why isn't that ().Seq? 03:26
ZzZombo Because it does have an element. 03:27
m: dd {}<a>
camelia Any % = Any
jdv79 oh, default. meh... 03:28
ZzZombo You just added it.
Geth doc: 1b6453970e | cfa++ | doc/Type/Signature.pod6
A slightly gentler walk through :D/:U constraint examples
 04:22
synopsebot Link: doc.perl6.org/type/Signature
cfa hopefully that's okay; i found the existing documentation a little dense as it dives straight into a code block 04:25
xiaoyafeng hello , I found min function just return one elem. 05:36
xiaoyafeng but look this code: my %dd = ( a => 3, b => 3, c =>4); (keys %dd).min({%dd{$_}}); 05:38
I expect return 2 elems:a and b
do I miss something? 05:39
shinobi-cl m: my %dd = ( a => 3, b => 3, c =>4); say %dd.minpairs; 05:40
camelia (a => 3 b => 3)
shinobi-cl maybe minpairs is what you need
m: my %dd = ( a => 3, b => 3, c =>4); my %smallest = %dd.minpairs; say %smallest.keys; 05:42
camelia (a b)
xiaoyafeng Thanks! let me check minpairs document;) 05:43
ZzZombo `min` returns values 05:47
shinobi-cl sure. BTW, if you don't want to use a temporary hash (%smallest) you have to use this:
m: my %dd = ( a => 3, b => 3, c =>4); say (%dd.minpairs.hash).keys;
camelia (a b)
ZzZombo why the parentheses in `say`? 05:48
shinobi-cl oh, sorry those were a leftover of what i was testing... you do not need those :P 05:49
xiaoyafeng so .hash means explicit conversion to hash type? 05:52
shinobi-cl yes, for some reason, minpairs returns a Seq, not another hash. but using .hash fixes that 05:54
m: my %dd = ( a => 3, b => 3, c =>4); say %dd.minpairs.perl;
camelia (:a(3), :b(3)).Seq
xiaoyafeng one more question, is there a count like function in perl6? I find classify, similar but not equal. how translate below perl5 code to perl6? count_by {$_}, (a,a, a, b,b,b c,c) 05:56
will return a => 3, b => 3, c => 2;
ZzZombo m: say <a a a b b c c>.Bag 06:06
camelia Bag(a(3), b(2), c(2))
atweiden-air my latest innocuous commit seems to freeze rakudo compilation of Config::TOML github.com/atweiden/config-toml/co...22e967be5c 06:07
can anyone with a high perf comp try and get Config::TOML to compile?
atweiden-air fixed: freeze was caused by a missing `.Array()` in method table-header-text, uncovered after reverting hypers 06:48
Geth whateverable: a401e1da8d | (Aleks-Daniel Jakimenko-Aleksejev)++ | lib/Whateverable.pm6
Support older zstd version, complain if not installed

Tackles issue #122 a little bit. This *should* make it work for older zstd versions like 0.5.1 (but I haven't tested it just yet).
Also, it should now die with a more helpful message if zstd or lrzip is not installed. ... (7 more lines)		 08:22
AlexDaniel actually maybe someone can suggest a better approach for github.com/perl6/whateverable/commit/a401e1da8d 08:28
yoleaux 00:17Z <jnthn> AlexDaniel: zstd 0.5.1, lrzip 0.621
AlexDaniel how else can I check that the command does not exist 08:29
AlexDaniel `run` just exits with “exit code: 1”, and that doesn't tell enough 08:29
e: run ‘oeunohunteohuntoahuonathu’
evalable6 (exit code 1) The spawned command 'oeunohunteohuntoahuonathu' exited unsuccessfully (exit code: 1)
in block <unit> at /tmp/ki3cOsjZr7 line 1
AlexDaniel 6c: run ‘oeunohunteohuntoahuonathu’
committable6 AlexDaniel, gist.github.com/aa42b6eb50f114c31e...0a59b26a53 08:30
AlexDaniel oh ya see, it was -1 at some point
which is kinda useful-ish
bisect: old=2016.09 new=2016.10 run ‘oeunohunteohuntoahuonathu’
bisectable6 AlexDaniel, Bisecting by output (old=2016.09 new=2016.10) because on both starting points the exit code is 1
AlexDaniel, bisect log: gist.github.com/17d512674b22c5d7dc...a7a994d1fb 08:31
AlexDaniel, (2016-09-17) github.com/rakudo/rakudo/commit/c4...608041eb9b
AlexDaniel no way
c: c4fd9f515d49d^,c4fd9f515d49d run ‘oeunohunteohuntoahuonathu’
committable6 AlexDaniel, gist.github.com/af495ec00f05dcaf0a...aa0af9a0b2
AlexDaniel bisect: old=c4fd9f515d49d new=2016.10 run ‘oeunohunteohuntoahuonathu’
bisectable6 AlexDaniel, On both starting points (old=c4fd9f5 new=2016.10) the exit code is 1 and the output is identical as well 08:31
AlexDaniel, Output on both points: «The spawned command 'oeunohunteohuntoahuonathu' exited unsuccessfully (exit code: -1)␤ in block <unit> at /tmp/69ysAF8JdZ line 1␤␤»
AlexDaniel ah!! I'm blind! 08:32
bisect: old=2017.05 new=2017.06 run ‘oeunohunteohuntoahuonathu’
bisectable6 AlexDaniel, Bisecting by output (old=2017.05 new=2017.06) because on both starting points the exit code is 1
AlexDaniel, bisect log: gist.github.com/70f4e99d23827941ba...e5c69604c3
AlexDaniel, (2017-06-12) github.com/rakudo/rakudo/commit/92...a08fb98bd2
AlexDaniel c: 92bd7e4f54a^,92bd7e4f54a run ‘oeunohunteohuntoahuonathu’
committable6 AlexDaniel, gist.github.com/0b57fda07c31cf2e7a...3029b13bde
AlexDaniel yeah… 08:33
TEttinger is oeunohunteohuntoahuonathu an inside joke? 08:36
AlexDaniel TEttinger: no, it's like asldfhlgdsahf;ldsafj;alfdsj but better 08:37
c: 92bd7e4f54a^,92bd7e4f54a run(‘oeunohunteohuntoahuonathu’).signal.say 08:38
committable6 AlexDaniel, ¦92bd7e4f54a^: «254␤» ¦92bd7e4: «0␤»
TEttinger weird 08:39
AlexDaniel I filed R#1590 08:40
synopsebot R#1590 [open]: github.com/rakudo/rakudo/issues/1590 [regression] Procs should not get back with “exit code: 1” if the program does not exist at all
moritz moin moin 10:15
Zoffix m: {}<sadf>:v.sort.head(2).perl.say 11:26
camelia ().Seq
Zoffix jdv79: ^ just ask for :v if you don't want default-default of autovivification
AlexDaniel perl 12:13
oops
.tell perlawhirl if python has combinations_with_replacement in itertools, why don't we have a more readable alternative? ;) Maybe file a ticket 12:14
yoleaux AlexDaniel: I'll pass your message to perlawhirl.
AlexDaniel “One final observation about Perl 6 and math: although Perl 6 has all the usual functions from math.h, it could certainly use a few more.” 12:15
www.evanmiller.org/a-review-of-perl-6.html
↑ very similar issue :) 12:18
Zoffix I propose an better idea: a module. 13:02
People keep saying how small our ecosystem is, yet everyone's so keen to add stuff to core. 13:03
And often protest when suggestion to add a module is. And stuff that's implemented in core then doesn't get neither tests nor documentation. Just shows how insignificant the added feature really is. 13:04
Zoffix checks whether IO::Path.parent(Int) was ever documented or tested
Nope. 13:06
Not to mention, if you're adding a method to List, you essentially must add it to 99% of Perl 6 classes, to maintain the everything-is-a-1-el-list semantics. 13:08
Like why were both .classify and .classify-list added :/
</rant>
DrForr It worked so well with PHP... 13:15
Zoffix Yeah, I remember 10 years ago #perl cracking jokes about inconsistencies in routine names in PHP... In Perl 6 we now have worse inconsistencies. That's just Karma. 13:16
DrForr I'm not sure if I was trying to be ironic, sarcastic or depressed there. 13:17
Zoffix :) 13:19
huf oh it'll be fine. it's not like "ironic", "sarcastic" and "depressed" fit any pattern 13:21
yet we handle them fine
DrForr Speak for yourself. :/
(sorry, just playing the waiting game today and not doing a wonderful job of it.) 13:22
OTOH App::Mi6 does help with releasing. 13:23
Geth doc: 3d7974671f | (Zoffix Znet)++ | doc/Type/IO/Path.pod6
Document IO::Path.parent(UInt)

Spec: github.com/perl6/roast/commit/bf3f347e11
   github.com/perl6/roast/commit/1b39fb200c
Rakudo impl: github.com/rakudo/rakudo/commit/7bea3a2dddcfc11 Closes github.com/perl6/doc/issues/1616 D#1616		 13:28
synopsebot Link: doc.perl6.org/type/IO::Path
synopsebot D#1616 [closed]: github.com/perl6/doc/issues/1616 [NOTSPECCED][docs] IO::Path.parent(Int)
DrForr (more accurately, sending out notes to recruiters that haven't bothered to respond and waiting on those non-replies while looking for others to send email to that will never get read.)
"Wonderful! Great job with the interview!" ... "Huh? Who are you?" 13:30
rindolf lizmat: hi 13:31
lizmat: github.com/shlomif/project-euler/t...-euler/618 - here is the py and pl5 versions
lizmat rindolf: will have a look at it after I finish the P6W today
rindolf lizmat: this is the p6 code - github.com/perl6/perl6-examples/bl...imotimo.p6 13:32
lizmat: thanks
lizmat: this is the nqp version - gist.github.com/zoffixznet/cb19c86...19f7ee9383 13:34
cfa morning folks
El_Che DrForr: I got 2 job offers from the same guy on the same day. 1 was non-really-technical multi-language helpdesk, the other was CTO :)
worthless :)
rindolf lizmat: by Zoffix++
DrForr Thanks, it's a gray day here and I need to keep my mood up. 13:35
rindolf El_Che: heh 13:36
lizmat just moved the curtain a bit to block out the sun 13:38
DrForr Hrm, I *do* have cheap bus fares to Amsterdam from here... :) 13:39
lizmat DrForr: I'm not in Amsterdam :-) 13:45
~190km south-east of it 13:46
El_Che lizmat: .nl is a big suburb of Amsterdam 13:47
rindolf El_Che: heh
DrForr I know, one step at a time :)
Geth doc: titsuki++ created pull request #1821:
Add Range.rand		 14:34
doc: c3e251d4f4 | (Itsuki Toyota)++ | doc/Type/Range.pod6
Add Range.rand
 14:35
doc: 68cade4271 | (Itsuki Toyota)++ (committed using GitHub Web editor) | doc/Type/Range.pod6
Merge pull request #1821 from titsuki/range-rand

Add Range.rand
synopsebot Link: doc.perl6.org/type/Range
Geth doc: ef09dec099 | (Zoffix Znet)++ (committed using GitHub Web editor) | doc/Type/Range.pod6
fix typo
 14:39
titsuki Zoffix: Thanks! 14:44
El_Che .tell Zoffix: isn't the d release supposed to clean up some of the inconsistencies? 14:52
yoleaux El_Che: What kind of a name is "Zoffix:"?!
El_Che .tell Zoffix isn't the d release supposed to clean up some of the inconsistencies?
yoleaux El_Che: I'll pass your message to Zoffix.
Zoffix El_Che: not that I'm aware of. 15:04
El_Che: and I think the ship has sailed for many of them. Fixing all of them at this point in time would have too much impact on users. 15:10
El_Che Zoffix: I thought the versioning of the lang was to have a license to break stuff, with a compiler/implementation targetting x versions 15:16
wishful thinking, then :) 15:17
Zoffix El_Che: not at all what I'd call it. People wouldn't be too keen on using a language that keeps changing everything from under them all the time. Even though you can specify a language version, if you do so, it means you cannot use latest features in that file and I can't imagine latest compilers supporting ancient language versions indefinitely. More-over, the language-versioning doesn't help much with 15:20
objects and method implementations (what does a 6.c version file meant to do when it's given a 6.d Str object?)
El_Che true 15:21
Zoffix Even with the things we *are* able to change, I'd never call it "license to break stuff", but "diminishing impact from fixing stuff"
rakudo.party/post/Perl6-On-Specs-V...d-Breakage
I imagine some driven person will eventually create a fork of Perl 6. It'd not have "Perl" in the name and it won't have billion methods most people won't need in core. It'd also polish all the inconsistencies before releasing the stable version (or even before releasing a publicly available version). 15:22
And would have full tests and documentation of everything. 15:24
AlexDaniel well good luck to them, but I'm pretty sure we'll fix the name before that happens :) 15:25
El_Che hehe
Zoffix + puritanical standards of development: mandatory tests, docs, and justification for all features, and consistency review for how the feature and its behaviour fits with the rest of the language.
Anyway, toast is done 15:27
buggable: toast
buggable Zoffix, Between 2018.02-1-118ga5a6c77 and 2018.02: 12 (1.21%) modules got burnt; 17 (1.72%) got unsucced; 258 (26.09%) out of 989 modules appear unusable. See toast.perl6.party/ for details.
AlexDaniel did JSON::Fast fail because of the .comb change? 15:29
“The iterator of this Seq is already in use/consumed by another Seq”
sounds related
Zoffix AlexDaniel: yeah, and it should already be fixed.
AlexDaniel: where do you see it fail? It says "Succ" for me 15:30
AlexDaniel oh…
toast.perl6.party/module?module=Di...18ga5a6c77
and also: toast.perl6.party/module?module=Ap...18ga5a6c77 15:31
Zoffix AlexDaniel: well, the reason is that module locked their version: modules.perl6.org/dist/Dist::Helpe...6.json#L11
And App::Cpan6 is by the same author, so I bet the same is the case there.
AlexDaniel awesome 15:31
Zoffix I wonder if the author meant those locks as "at least this version" rather than precicely this version. 15:32
AlexDaniel tyil: ↑ ping? :) 15:32
AlexDaniel but that looks better than I expected actually! 15:32
so I think we're good with the change generally 15:33
AlexDaniel runs away for the rest of the evening
rindolf lizmat: here? 15:35
lizmat rindolf: yes
rindolf lizmat: i want to go for a walk
lizmat please do: I'm working on the Perl 6 weekly, and I said I'd look at the Euler question *after* I finished the Perl 6 Weekly 15:36
rindolf lizmat: ok
lizmat which will still be at least 5 hours or so
rindolf lizmat: ah 15:38
[Coke] FYI: news.perlfoundation.org/2018/03/jan...votes.html 15:41
lizmat I was just about to paste that link here as well :-)
Zoffix wait, wtf... both Dist::Helper and App::Cpan6 both have dist-helper as repo URL
lizmat [Coke]: would it be difficult to put a link to the original proposals in the Grant Votes post ? 15:42
Zoffix .tell tyil FWIW, both Dist::Helper and App::Cpan6 appear to be in the ecosystem, despite apparently App::Cpan6's repo being renamed to Dist::Helper's repo
yoleaux Zoffix: I'll pass your message to tyil.
lizmat would be helpful to those that will only see the Grant Votes post in the P6W 15:43
El_Che rperl got a lot of rejections
DrForr A touch disapponted that my +1s for JJ's project didn't seem to get published, but it looked like it went through.
JJ++
AlexDaniel Oh! So JJ's work is going to be funded! 15:44
that's wonderful news!
Zoffix :) 15:45
DrForr Yep, 5:1 1 abstention, the rest were all solid 'no's.
[Coke] lizmat: sure, I can do an update.
DrForr: I didn't see any of your comments held up in moderation 15:46
or in spam
DrForr They were as anonymous as I couldn't get the password recovery to work. 15:47
El_Che are comments counted as votes?
DrForr And at least one was done on a tablet so f*ck knows if it actually got posted.
Zoffix El_Che: no, it's community feedback. TPF votes. 15:48
DrForr El_Che: The committee votes, they take the comments into consideration.
[Coke] lizmat: done, will take a minute to update the cache 15:49
lizmat [Coke]++
[Coke] thanks for the suggestion. (I really need to script the generation of these posts.) 15:49
(not the actual posting itself, but the content)
Zoffix I'll also likely submit a TPF grant proposal for next round. Draft is here: (from last draft I posted, I bumped the amount requested slightly, based on feedback): temp.perl6.party/Constant-Rat-TPF-Grant.html
[Coke] I will note that having a dollar off the price is probably not going to help anything on the GC end. :) 15:50
but I appreciate you saving the TPF a dollar. 15:51
Zoffix :D
[Coke] (assuming funding, etc. I have no idea how people are going to vote on things)
jkramer This is weird: dpaste.com/02K9QVB 15:54
Why is it only working when it's on the same line?
Zoffix jkramer: RT#131900 15:55
synopsebot RT#131900 [open]: rt.perl.org/Ticket/Display.html?id=131900 [REPL] REPL loses custom operators
jkramer Oh, ok 15:56
tyil Zoffix: Dist::Helper is indeed the correct name, I'll drop the App::Cpan6 from PAUSE 16:04
yoleaux 15:42Z <Zoffix> tyil: FWIW, both Dist::Helper and App::Cpan6 appear to be in the ecosystem, despite apparently App::Cpan6's repo being renamed to Dist::Helper's repo
tyil also it should've been "at least this version" checks, I did ask around here a couple times for the correct notation but got no replies, and I had to at least get the versions mentioned in the META6, so I went with this for now 16:05
Zoffix eco: Test::META 16:07
buggable Zoffix, Test::META 'Test a distributions META file': github.com/jonathanstowe/Test-META
Zoffix hm. I thought that module had the "at least this version" specified, but I think it got removed because zef had trouble understanding them 16:08
Zoffix m: say v0.9.6 ~~ v0.9.6+ 16:08
camelia True
Zoffix m: say v0.9.10 ~~ v0.9.6+
camelia True
Zoffix Not an expert on it, but I think :ver<v0.9.6+> is how you'd specify it, but I don't know whether zef supports that, even if it were correct format 16:09
timotimo m: say v0.9.6 ~~ v0.9.6^
camelia 5===SORRY!5=== Error while compiling <tmp>
Missing required term after infix
at <tmp>:1
------> 3say v0.9.6 ~~ v0.9.6^7⏏5<EOL>
expecting any of:
prefix
term
timotimo m: say v0.9.6 ~~ v0.9.6~
camelia 5===SORRY!5=== Error while compiling <tmp>
Missing required term after infix
at <tmp>:1
------> 3say v0.9.6 ~~ v0.9.6~7⏏5<EOL>
expecting any of:
prefix
term
timotimo hm, so + really is allowed there? 16:10
m: say v0.9.6 ~~ v0.9.6-
camelia 5===SORRY!5=== Error while compiling <tmp>
Malformed postfix call
at <tmp>:1
------> 3say v0.9.6 ~~ v0.9.7⏏056-
Zoffix Yeah, it's part of the version spec.
And * too 16:11
m: say v0.9.10 ~~ v0.9.*
camelia True
Zoffix github.com/rakudo/rakudo/blob/mast...on.pm6#L78 (for +) and github.com/rakudo/rakudo/blob/mast...on.pm6#L40 (for *)
aw :( no history due to .pm -> .pm6 rename: github.com/rakudo/rakudo/commits/m...ersion.pm6 16:12
Zoffix More readabable version of the version parser: github.com/rakudo/rakudo/blob/46a7...pm#L38-L60 16:13
rindolf hi all 16:17
Zoffix: hi
Zoffix: sup?
Zoffix: thanks again for your help
jkramer Is it possible to perform modifiy two LHVs at the same time with some operation? For example I'd like to do: %foo{$a, $b} += 2; (add 2 to the values of both $a and $b in that hash)
I'm getting "Cannot assign to an immutable value" with that 16:18
ZzZombo m: say <a b c><a c> 16:19
camelia Type List does not support associative indexing.
in block <unit> at <tmp> line 1
ZzZombo m: say %(<a b c>)<a c>
camelia Odd number of elements found where hash initializer expected:
Found 3 (implicit) elements:
Last element seen: "c"
in block <unit> at <tmp> line 1
timotimo jkramer: you can >>+=>> for example 16:20
jkramer Ah nice, I keep forgetting about that synta
x 16:21
rindolf ZzZombo: are you asleep? ;)
ZzZombo Yes, why?
rindolf ZzZombo: never knew someone who could chat on irc when sleeping 16:22
ZzZombo You never seen someone who needs to finish work before going to bed? Wow.
rindolf ZzZombo: i did 16:23
Zoffix m: my %foo; my $a = "meows"; my $b = "foos"; %foo{$a, $b}»++; dd %foo 16:25
camelia Hash %foo = {:foos(1), :meows(1)}
Zoffix m: my %foo; my $a = "meows"; my $b = "foos"; %foo{$a, $b} »+=» 2; dd %foo
camelia Hash %foo = {:foos(2), :meows(2)}
Zoffix my %foo; my $a = "meows"; my $b = "foos"; %foo{$a, $b} = 2 xx 2; dd %foo
m: my %foo; my $a = "meows"; my $b = "foos"; %foo{$a, $b} = 2 xx 2; dd %foo
camelia Hash %foo = {:foos(2), :meows(2)}
Zoffix m: my %foo; my $a = "meows"; my $b = "foos"; %foo{$a, $b}».:<++>; dd %foo 16:28
camelia Cannot find method 'Any' on object of type List
in block <unit> at <tmp> line 1
Zoffix booo
c: 2017.01 my %foo; my $a = "meows"; my $b = "foos"; %foo{$a, $b}».:<++>; dd %foo 16:30
committable6 Zoffix, ¦2017.01: «Cannot find method 'Any' on object of type List␤ in block <unit> at /tmp/TDLmLNCG6V line 1␤␤ «exit code = 1»»
Zoffix I guess that's one case that was missed during the Great Method Hypering work 16:31
Zoffix m: my %foo; my $a = "meows"; my $b = "foos"; %foo{$a, $b}».&[+=](2); dd %foo 16:36
camelia Hash %foo = {:foos(2), :meows(2)}
Zoffix m: my %foo; my $a = "meows"; my $b = "foos"; %foo{$a, $b}».&[RRRRRRRRRRRRRRRRRR+=](2); dd %foo
camelia Hash %foo = {:foos(2), :meows(2)}
Zoffix hehe, there're literally infinite ways to do what you want, jkramer :) 16:37
Geth doc: 5d73d9a99c | (JJ Merelo)++ | 3 files
Removing a couple of 404s

Mainly related to #561		 20:25
AlexDaniel heh, Amazon Go. Cool, cool. Except that you'd still have to go to the damn store. Outdated concept :P :P 20:30
dylanwh Amazon Rust is safer. 20:33
lizmat and another Perl 6 Weekly hits the Net: p6weekly.wordpress.com/2018/03/05/...atic-perl/ 20:36
comborico1611 lizmat: I've been messing around A BUNCH with the Lisp community lately, and I can tell you your weekly newsletter really adds a lot of important connectivity to the Perl 6 community. 20:45
(By comparison) 20:46
lizmat thank you :-)
comborico1611 You bet!
El_Che comborico1611: the Lisp community can do a weekly like that. It has side effects.
comborico1611 El_Che: Is that a joke? 20:47
El_Che comborico1611: It can be proofed it's a joke
comborico1611 I'm still new to Lisp, so I'm not sure what people say about Lisp and so forth.
:-) I'm confused, but happy.
El_Che en.wikipedia.org/wiki/Side_effect_...r_science) 20:48
dylanwh it's a joke depending on you interpreter it.
El_Che dylanwh++
AlexDaniel Question! How can I avoid stuff like this: github.com/perl6/whateverable/blob...p6#L51-L53 20:57
that is, a type constraint spanning across multiple lines
OK I can have a regex defined earlier or something like that, but it doesn't make things much clearer
is there something like `also` but for constraints? 20:58
Zoffix AlexDaniel: I'd just define a subset or a regex github.com/zoffixznet/perl6-ZofBot...er.pm6#L13 21:07
Zoffix There's no 'also' for parametets and given `returns` is on the chopping block unlikely to be 21:08
Zoffix or just put params one per line and stuff and make em more like attributes github.com/zoffixznet/perl6-ZofBot...er.pm6#L59 21:09
AlexDaniel hmm ok, thanks! 21:11
Zoffix or just process the regex in the body and return $.NEXT if the event doesn't match 21:12
timotimo stmuk: can you add wakelift.de/rss to pl6anet? 21:13
Zoffix timotimo: you'll need to PR your feed addition to this file: github.com/stmuk/pl6anet.org/blob/...anetrc#L93 21:16
timotimo oh, cool 21:17
El_Che lizmat: it's always a pleasure to read the weekly! 21:34
El_Che lizmat: you should curate all the stuff that interest me :) 21:35
timotimo github.com/arangodb/linenoise-ng - this may be interesting for us 22:27
timotimo i'd like a line editing library that lets me take control of the terminal at arbitrary points so i can do miscellaneous screen operations 22:36
timotimo huh, i wonder why titsuki and krunen both have a Term::termios which are almost identical, but each has only themselves in the author and authors field? 22:49
TEttinger are they the same person? 22:58
timotimo i don't think so 22:59
titsuki_ timotime: Should I delete mine ASAP? I received an answer that there are no problem if there are two modules that have the same name but have the different authors. 23:08
TEttinger maybe only delete yours if they're identical? if yours is bugfree but the other isn't, yours should stay for sure 23:10
timotimo no problem to keep it; users can disambiguate by the :auth anyway, it's just confusing :) 23:11
i'd prefer one on cpan over one on github+ecosystem, fwiw
titsuki_ timotime: OK, I see. But I'll delete mine sometime soon. 23:13
timotimo but yours is the one on cpan :D 23:14
m: say 5.7 / 7.15 23:15
camelia 0.797203
timotimo zoffix, you can make it about 20% faster by making the \Token parameter required - that prevented the jit from jitting that frame.
Zoffix timotimo: well, wow! With that change I get it nearly 50% faster. 18s vs P5's 17s 23:18
timotimo oh my
Zoffix timotimo: so JIT can't jit anything with optional params? Should we look through core code and see if anything could have optionals eliminated? Also, how did you figure out that that wasn't being jitted?
timotimo first i ran it with --profile to see that calc_S, which was at the top of the routines list by "exclusive time", was 100% red on the right side 23:19
then i used MVM_JIT_LOG to find the line that starts with "BAIL:" after the line that has "calc_S" in it
Zoffix m: say 33.298/18.634
camelia 1.786949 23:20
Zoffix Yeah, changing Token to mandatory param makes the program 78% faster
timotimo it was param_op_o, which is "get a parameter that's optional, positional, and object typed"
you probably got a better improvement because i used a much smaller workload
jnthn Most of the time, optional parameter ops are rewritten and never make it do the JIT 23:21
So something must have blocked that rewriting
timotimo potentially OSR-related 23:22
because calc_S was only called a single time and relied on OSR to get jitted at all
Zoffix I don't see calc_S line at all in jit log 23:23
timotimo somehow p6box_I comes out at 14.5% inclusive time in a perf report
Constructing JIT graph (cuuid: 1, name: 'calc_S')
this is what it ought to look like
Zoffix $ G calc_S j.log
2018.02.1 zoffix@VirtualBox~/R/rakudo (master)$
timotimo that's strange, can you send the jitlog file and maybe the shell history over? 23:24
Zoffix Ah, I don't have debug options in moarvm. Does that matter?
timotimo shouldn't matter
the --debug option in moarvm's Configure.pl just instructs C compiler and linker to keep debug symbols around, i.e. which memory locations in the binary correspond to what functions and variables and such
it's important for debugging with gdb 23:25
Zoffix timotimo: wget temp.perl6.party/j.log; and I made it with `MVM_JIT_LOG=j.log ./perl6 ../z.p6 > /dev/null` where `./perl6` is 2018.02.1-118-ga5a6c7786 built on MoarVM version 2018.02-27-g760b0913e and ../z.p6 is gist.github.com/zoffixznet/3e3d31a...fe59704470 23:27
Zoffix chuckles at `686320:BAIL: op <setmethcache>`
timotimo that's certainly interesting
Zoffix ZofBot: where'd you put the meth cache? 23:28
ZofBot Zoffix, And he said, “Yes, it’s the -to-
timotimo why it doesn't show up at all ... no clue :( 23:29
but now it's fast, so ... :)
Zoffix It shows up if I also set MVM_SPESH_BLOCKING=1; maybe program exits before it get a chance to write that out? 23:30
mhm, if I add `sleep 2` at the end of the program, it shows up even without SPESH_BLOCKING 23:31
timotimo a speshlog would tell us if it got to specialize it before it quit
it shouldn't take long enough to spesh that so that it could fail to install :o
Zoffix Maybe it installs but quits before writing to the file? It's like a 20MB file 23:32
timotimo oh, that could be, yeah 23:33
so, if we had an op that turns a native int into a str without going via Int, that'd make this thing faster 23:34
at least i think so
Zoffix nice
timotimo nope, barely a difference at all 23:35
Zoffix :) 23:35
timotimo what allocates all these Int objects ...
7757486 Int objects
Zoffix I got `nqp::say(nqp::concat(nqp::tostr_I(nqp::box_i(Token, Int)), 23:36
nqp::concat(' ', nqp::tostr_I(nqp::box_i($ip, Int))))),
there. Could that be it?
timotimo i commented that out and it still allocates that much
Zoffix m: say 18.287/17.735 23:37
camelia 1.031125
timotimo it's barely a difference at all
Zoffix .tell rindolf I got P6(NQP) version fast enough to be to just 3% slower than P5: gist.github.com/zoffixznet/6082e29...4d7426791c
yoleaux Zoffix: I'll pass your message to rindolf.
timotimo not getting line numbers because nqp code :S 23:38
rindolf Zoffix: hi
yoleaux 23:37Z <Zoffix> rindolf: I got P6(NQP) version fast enough to be to just 3% slower than P5: gist.github.com/zoffixznet/6082e29...4d7426791c
Zoffix And 2s is what it takes to compiler it, so I'd say my original challenge completed: I got P5 beat (despite the final result being NQP soup appreciated by few:P) 23:39
s/compile <( r// 23:40
m: use nqp; my int $x = 42; for ^10 { say nqp:add_i($x, 1) } 23:41
camelia ===SORRY!===
Lexical with name '$x' has wrong type
Zoffix why u say dat?
oh right, nqp int != P6 native int 23:42
right?
timotimo no, that should be compatible
Zoffix Ah lol 23:43
m: say 'nqp:add_i'.comb(':').elems
timotimo um yo
camelia 1
timotimo check what happens when you turn $ii into int, too
Zoffix damn, I thought it were an int alread y:o 23:44
timotimo it may surprise you
Zoffix Woooohoo!
Actually, this is what I originally was expecting when I started on this journey :) 'cause that other program I rewrote in NQP was a bunch faster than P5 23:45
m: say 17.735/7.168
camelia 2.474191
timotimo that's more like it. 23:46
Zoffix rindolf: timotimo++ made the program 2.5x FASTER than Perl 5 :D gist.github.com/zoffixznet/4df7a30...fb65bbb0a4
timotimo can you measure with MVM_JIT_EXPR_DISABLE=yes ?
Zoffix 0m6.904s 23:47
timotimo so a bit faster, yeah?
Zoffix Yeah
timotimo how stable is that measurement?
Zoffix 0m6.813s 0m6.976s 0m6.874s 23:47
timotimo OK
well, brrt has lots of ideas for making the exprjit smarter :) 23:48
Zoffix Sweet :)
timotimo with $fkE at 25, the program runs the GC a single time 23:50
rindolf Zoffix: nice 23:51
timotimo given ~2 seconds of that are parsing the code - no doubt because of the huge array of primes at the start - it's pretty good 23:56

Powered by Cro and the Raku® Programming Language.

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

Thank you!