»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 Man, I originally tested this with Rat/FatRat allomorph and it was working, but the change was made to ditch the FatRat :P 00:00
b2gills .tell tyil If you have two multis that differ in just a named parameter, it should be marked as being required `:$named!` www.reddit.com/r/ProgrammingLangua...s/dz0mmxn/ 00:06
yoleaux b2gills: I'll pass your message to tyil.
lucasb m: sub f(Empty) { 'ok' }; say f(Slip) 00:12
camelia Use of uninitialized value of type Slip in numeric context
ok
in sub f at <tmp> line 1
Zoffix lucasb: file those as bugs 00:14
lucasb m: say Slip ~~ Empty # what should this evaluate to? 00:15
camelia Use of uninitialized value of type Slip in numeric context
True
in block <unit> at <tmp> line 1
lucasb nevermind, evaluates to True 00:16
lucasb is this warning legit or not? 00:17
Zoffix Doesn't look like it to me
lucasb R#1833 00:26
synopsebot R#1833 [open]: github.com/rakudo/rakudo/issues/1833 Spurious warning when smartmatching Slip ~~ Empty
Zoffix m: constant MidRat = do { my $mr := Rat.HOW.new_type: :name<MidRat>; $mr.^add_parent: class :: does Rational[UInt, UInt] {}; $mr.^add_parent: Rat; $mr.^compose; $mr }; dd MidRat.new(1, 2) 00:35
camelia Type check failed in binding to parameter '<anon>'; expected Any but got DeT (?)
in block <unit> at <tmp> line 1
Zoffix man, almost works. 00:36
m: class Z does Rational[uint64, uint64] {}.new: (my uint64 $ = 42), my uint64 $ = 2 00:38
camelia Type check failed in binding to parameter 'nu'; expected uint64 but got Int (42)
in block <unit> at <tmp> line 1
Zoffix
.oO( what did I get myself into... )
m: role Z[::P1] { method new(P1 $x) { dd $x} }; class :: does Z[int] {}.new: my int $ = 42 00:40
camelia Type check failed in binding to parameter '$x'; expected int but got Int (42)
in method new at <tmp> line 1
in block <unit> at <tmp> line 1
Zoffix .ask jnthn is there some trick to resolve ambiguity with `Numeric` being available twice in `class MidRat does Rational[UInt, UInt] is Rat {}.new.Numeric`... Creating a temp class that `does Rational[UInt, UInt]` and inheriting from that seems to avoid ambiguity but (*) inherits from a temp class (*) hits some bug with parametarization 01:14
yoleaux Zoffix: I'll pass your message to jnthn.
Zoffix .ask thundergnat in FatRat.Str: "speed increase, 16 digits would work fine; but it isn't spec"... Are you sure those tests were actually spec (as in 6.c-errata branch tests) and not just proptests (new tests in master branch)? 01:18
yoleaux Zoffix: I'll pass your message to thundergnat.
Zoffix .tell jnthn more obvious example: `class A does Numeric {}; class B is A does Numeric {}.new.Numeric` I want the B's `does Numeric` to replace the effects of `A` doing Numeric. And in my case it needs to replace parametarization types. Is that doable? 01:21
yoleaux Zoffix: I'll pass your message to jnthn.
Zoffix m: say "Will I regret MidRat 10 years from now? {<Yes No>.pick}" 01:23
camelia Will I regret MidRat 10 years from now? Yes
Zoffix I'm starting to think so too...
K, I'm cancelling MidRat. In leu of making Rat type's denominator parametarized with Int instead of uint64. All the features of a MidRat to be merged with Rat: that is, Rat can be created with a denominator of 64+ bits, but it will degrade to a Num, like a MidRat would. 01:26
And that leaves us open to exploring the MidRat concept in the future, whereas if we go MidRat route now, we lock ourselves into it. 01:27
Zoffix realizes that annuls a large portion of the grant since native-typed Rats were another item along with MidRat. 01:30
I guess I'll just keep improving Rats and see where that lands me… if it falls short of the grant's plan, then oh well 01:31
ktown my _homework_ has me thinking about scoping. 02:02
as a perl5 guy by day, I have not used 'use strict'. so vars are global unless I use 'my' 02:03
timotimo i hear "no strict" in perl5 is frowned upon 02:04
Zoffix yeah, it's a terrible thing
ktown so my first pass perl6 stuff has many error because I do not add the 'my' everywhere 02:05
Zoffix ktown: there's a way to avoid those errors in Perl 6, but it's a really bad idea. I recommend you unlearn that bad habbit and learn proper scoping. It'll save you time hunting bugs. 02:06
ktown maybe a better way is for a var to be local to the block first, then 'my' if local to an inner block
timotimo that'll also let you pepper the code with type restrictions on your variables to catch some thinkos
ktown Zoffix: it's not about proper scoping it's about default scoping doing the 'right' thing 02:08
Zoffix ktown: that doesn't protect you from typos. "hello = 42; add_to_database(helo + 42)" That's a fatal error in proper code, but with your "local to block by default", you have a silent bug in your code where you lose the value of a variable because you typoed its name
AlexDaniel m: no strict; $x = 42; say $x² 02:09
camelia 1764
ktown default to global is bad
I was manually transpiling phython to perl6, and wondering about the need to add 'my' everywhere 02:10
Zoffix m: no strict; $hello = 42; add_to_database(++$helo); sub add_to_database { "Adding $^v to the database. Boy, I sure hope that value is correct.".say }
camelia Adding 1 to the database. Boy, I sure hope that value is correct.
Zoffix m: my $hello = 42; add_to_database(++$helo); sub add_to_database { "Adding $^v to the database. Boy, I sure hope that value is correct.".say } 02:11
camelia 5===SORRY!5=== Error while compiling <tmp>
Variable '$helo' is not declared. Did you mean '$hello'?
at <tmp>:1
------> 3my $hello = 42; add_to_database(++7⏏5$helo); sub add_to_database { "Adding $^
Zoffix It even tells you what the right variable name likely is!
AlexDaniel why do we even have ‘no strict’ in p6 though… 02:12
timotimo yes.
AlexDaniel greppable6: no strict
greppable6 AlexDaniel, 72 lines, 5 modules: gist.github.com/bdb024c50f6a74673c...d1f7260cd8
AlexDaniel omg
oh ok 02:13
docs and false positives
ktown imho dynamic typed langs are easier because they are not hard typed. and vars exist on the fly 02:14
perlawhirl another nice thing i didn't think about until recently is the perl6 parser will protect you from undeclared subs 02:15
ie. `if 1 < 0: print(foo())` will not error in python (or eqv in perl or ruby)
but eqv in perl6 will
of course, linters should pick this sort of thing up, but it's nice to have the compiler do it for you 02:16
ktown in perl5 is was alwasy easier then javascript or php because you can use a var without having to test existance, then go =New array if not
timotimo well, there's ||= and //= for that
and autoviv works inside of structures as well 02:17
which is worth a whole lot
AlexDaniel perlawhirl: btw even though there's no compile-time error for methods, rakudo does give you recommendations in “Did you mean” 02:18
m: class Foo { }; say Foo.gst
camelia No such method 'gst' for invocant of type 'Foo'. Did you mean 'gist'?
in block <unit> at <tmp> line 1
AlexDaniel m: class Foo { }; say Foo.gint
camelia No such method 'gint' for invocant of type 'Foo'. Did you mean any of these?
List
gist
list
min

in block <unit> at <tmp> line 1
ktown rocky error, with out testing myself. how does python scope? 02:20
ktown googles
local to block 02:23
timotimo fortunately they've introduced "nonlocal" a few years ago
ktown I know that declaring every var and hard typeing it the 'right' way but not the easy way 02:26
ktown the all or nothing of p5 'use strict' forcing declariation made me avoid it. 02:27
Zoffix ktown: `use strict` in Perl 5 is lexical. 02:28
$ perl -e '$x = 42; { use strict; my $y; }; print $x' 02:29
42
ktown m: no strict; use localscope; $a =1 ; say $a ; { my $a = 2 say $a} 02:32
camelia ===SORRY!===
Could not find localscope at line 1 in:
/home/camelia/.perl6
/home/camelia/rakudo-m-inst-2/share/perl6/site
/home/camelia/rakudo-m-inst-2/share/perl6/vendor
/home/camelia/rakudo-m-inst-2/share/perl6
CompUni…
ktown m: no strict; use localscope; $a =1 ; say $a ; { my $a = 2 ; say $a}
camelia ===SORRY!===
Could not find localscope at line 1 in:
/home/camelia/.perl6
/home/camelia/rakudo-m-inst-2/share/perl6/site
/home/camelia/rakudo-m-inst-2/share/perl6/vendor
/home/camelia/rakudo-m-inst-2/share/perl6
CompUni…
Zoffix "localscope" is not a thing.
You could do implement it with a slang, since we have support for local variables in QAST 02:33
ktown I know :(
Zoffix (slangs are an unsupported experimental feature tho)
or maybe it's possible with a macro
oh, nm, I thought you could write QAST directly in macros 02:35
ktown use strict was always to strict and got in the way of doing things 02:37
Zoffix Well, there's always PHP, eh? 02:37
ktown no, there is node.js 02:38
Zoffix I thought `'use strict';` was now in vogue in JS
ktown in JS you use lint to tell you to not use global vars 02:39
ktown i'm thinking through the where to use perl6 question. I want the answer to be perl6 > python and perl6 > node.js and perl6 > POE 02:51
Zoffix ktown: I don't code much, but I P6 increasingly useful in churning out quick command line automation tools. The multi-dispatch + auto-command-line flags generation with sub MAIN + auto --help generation from comments above those MAIN candidates really makes it a quick work. 02:53
timotimo POE is one of the perl5 async io things?
Zoffix timotimo: yeah 02:54
timotimo wouldn't know
Zoffix not just IO, there's a whole bunch of plugins. Including IRC clients and everything
It's an async loop and I always found its interface clunky AF
ktown: here's one of my script automating a bunch of build process commands. I think the first version of it took me less than half an hour to write and automated typing a bunch of commands to typing a couple of letters: github.com/zoffixznet/z/blob/master/bin/z 02:55
I plan to automate a lot of my currently-manual processes at $work next and hoping to write some module to make the process even easier. 02:56
m: say .1 + .2 == .3
camelia True
ktown Zoffix: true true the sub main and multi dispatch is AWESOME! the stuff i put on rosetta code as an eg
Zoffix ktown: ^ try that in python :P
ktown do not get me wrong... I hire devs then teach them perl5 becasue it's tje most effective way to get stuff done 02:59
I do not think python is better, but have to ask why it's so popular 03:00
Zoffix perl6 > POE hands down. I always found POE ass-backwards, while in perl 6 async/parallism is built in the language from scratch and is part of the syntax; not just some module. perl6 > python... Well, we currently lose on perf and modules, since Python is 28 years old and we're only 2 years old. I guess I'd put decent async/parallelism and good Unicode support without needing some library. And while I don't 03:01
know Python, the code I've seen looks like line noise compared to Perl 6: perl6.party/post/Python-is-The-New...e-Language en.wikipedia.org/wiki/C3_lineariza..._Python_3. perl6 > node.js === no idea about node.js
xq POE is very good :) 03:02
ktown I wrote my first IRC bot using POE in 2003
Zoffix I released probably over 200 POE modules on CPAN
eco: IRC::Client
buggable Zoffix, IRC::Client 'Extendable Internet Relay Chat client': github.com/zoffixznet/perl6-IRC-Client 6 other matching results: modules.perl6.org/s/
Zoffix ^ compare that to PoCo::IRC interface or even that of Bot::BasicBot. :) 03:03
timotimo i'm sure that link at the end was supposed to somehow have the search query in there?
Zoffix Oh right, something's broken
ktown I agree perl6 is better the POE
tiz my point actually. perl6 should also be better for aync servers then node.js 03:05
lernc0de Hey everyone. Was wondering if there were any things that Perl6 objectively does more smoothly than Common Lisp. I've noticed Perl6 seems to have lisp-style macros, and there is something called grammars that seems to be completely new. 03:07
Note: not trying to start some sort of language war - only want your insights
Zoffix ktown: as for python popularity: it's a snowball effect. Who cares if "the dress" is blue or yellow, but enough people talk about it and it snowballs into other people thinking they have to talk about. Python was/is used in education, so you got a populace of people who only ever learn 1 language religiously crusade for it. Then you got the "There should be one and preferably one way to do it" mentally going
for it. If you find programming hard, learning 1 way vs. many ways is easier. Then you got the "readability" con. It's easy to dupe people that just because you got words instead of symbols that that somehow makes a language more readable (it doesn't; you still have to know acceptable parameters, return values, and special cases to correctly understand a program), so that eliminates a crop of languages that
would be an alternate choice. Lastly, you got the timing: while Perl struggled, Python was one of the best alternatives, so programmers flocked to it.
timotimo as far as i've heard, CL already had everything that's called "new and hot" in today's languages when i was born
Zoffix ktown: it might be. Check out Cro mi.cro.services 03:08
timotimo joking aside, i barely know more than surface-level stuff about CL
i'm off for today o/ 03:11
lernc0de I barely know more than barely surface-level stuff about P6. ;)
ktown Zoffix: when I was a kid I used BASIC... then in university Pascal. then the kids used JAVA, now python
Zoffix I did Pascal in college and when I flunk out I kept trying to use Delphi to write programs in., 03:12
ktown not that crappy VBA. C64 BASIC 03:13
ktown plans to implement a C64 irc bot 03:14
C64: 10 PRINT "HI" 03:15
TEttinger CL has a lot of nice things from modern languages but still has a lot of old and cold misfeatures from before there was a Common Lisp, and there were divergent standards
ktown C64: 20 GOTO 10
TEttinger there are still a ton of variants
lernc0de I find CL's syntax and functions difficult to remember (but this could be with every modern language I suppose). Making an array has so many flags for example: (make-array 3 :adjustable t :element-type something :fill-pointer something :initial-content '(1 2 3)) 03:17
Is P6 more modular in this respect? Sorry if this is an annoying beginner question 03:18
ktown Zoffix: my daugher is taking CS 101 in python, this summer. I'm following along and doing the exersises in p6. 03:21
Zoffix cool :)
ktown acmebot.com/py/ 03:23
if she needs help I will have to learn python :(
Zoffix CaR Grant Report for May: blogs.perl.org/users/zoffix_znet/20...-2018.html 03:25
xq nice 03:31
native rational numbers is a great feature that perl 6 and common lisp share 03:32
lernc0de Probably going to purchase Perl 6 Fundamentals. Are there any other books that may be better? To phrase it differently, is there a K&R C for Perl 6? 03:43
b2gills perl6book.com/ 03:44
lernc0de I've seen the chart previously, but couldn't decide between fundamentals and deep dive. 03:46
ktown lernc0de: are you proficient in another lang or is this your first? 03:48
lernc0de I know C quite well, and am familiar with Common Lisp/Python 03:49
ktown so it's just syntax, and the perl6 lang does sooo0o much of the heavy lifting, have fun 03:51
lernc0de I will! :] 03:54
Thank you, have a great night everyone. Till next time
b2gills I would say that I would probably have difficulty choosing between fundamentals and deep dive.
lernc0de Perhaps I will invest in both this summer
ktown cool: POE::Component::WWW::XKCD::AsText 04:06
lookatme The documents is K&R for Perl 6 05:20
tyil b2gills:thanks for the heads up 05:31
yoleaux 00:06Z <b2gills> tyil: If you have two multis that differ in just a named parameter, it should be marked as being required `:$named!` www.reddit.com/r/ProgrammingLangua...s/dz0mmxn/
tyil weekly: www.reddit.com/r/ProgrammingLangua...t/dz1lc4t/ 05:49
notable6 tyil, Noted!
Geth doc: mryan++ created pull request #2027:
Make the availability of p6doc clearer		 06:01
Geth doc: 94754ee350 | (Martin Ryan)++ (committed by Moritz Lenz) | doc/Language/faq.pod6
Make the availability of p6doc clearer
 06:08
synopsebot Link: doc.perl6.org/language/faq
Geth doc: 48a2d0e84f | (JJ Merelo)++ | doc/Language/variables.pod6
Change the default number of threads

Following @jnthn advice. Closes #1065, maybe for good this time.		 06:24
synopsebot Link: doc.perl6.org/language/variables
jmerelo moritz: we seem to have a proxy error in whateverable.6lang.org travis-ci.org/perl6/doc/jobs/379566309 06:39
moritz jmerelo: yes, just saw that. I don't think I can do anything about that though 06:40
who maintains that? AlexDaniel` maybe?
jmerelo moritz: who's in charge of those servers? AlexDaniel` ?
He
We are using that for retrieving perl6 binary for the tests. Maybe there's another high-availability server for that? 06:41
AlexDaniel` I'll check in 15 mins 06:45
moritz ++AlexDaniel` 06:48
jmerelo AlexDaniel`: thanks 06:49
tyil weekly: www.reddit.com/r/perl6/comments/8j...sparrowdo/ 06:52
notable6 tyil, Noted!
AlexDaniel jmerelo: I have no idea what was wrong 06:52
jmerelo: a bunch of folks here have access to that server btw 06:53
jmerelo: timotimo MasterDuke Zoffix
jmerelo: but you should have access too 06:54
jmerelo: which of the keys should I use? github.com/JJ.keys
jmerelo AlexDaniel: let me see 06:55
AlexDaniel: the first one should work 06:56
AlexDaniel: would there be any way to upload the latest binary to some high-availability server? That's the only thing the tests need...
jmerelo AlexDaniel`: maybe bintray? blog.bintray.com/2013/05/30/google...per-place/ 06:59
I think El_Che uses it for his packages...
AlexDaniel` jmerelo: try `ssh [email@hidden.address] 07:12
jmerelo AlexDaniel`: got it, thanks. 07:13
AlexDaniel` jmerelo: as for uploading the binary somewhere, I can do it
jmerelo AlexDaniel`: I guess it's a matter of rebooting when it fails, right?
AlexDaniel` jmerelo: but at the same time it's the last commit on master, so people shouldn't be expecting it to always work anyway 07:14
jmerelo AlexDaniel`: that would be great. bintray seems to have an API so you can do it automatically...
AlexDaniel` like, there are some rare moments when it doesn't even compile…
jmerelo AlexDaniel`: right, but at least we will avoid the problems of it being unavailable...
AlexDaniel` jmerelo: you can restart bots with `sake kill:botname`
like `sake kill:shareable` 07:15
jmerelo AlexDaniel`: Great, thanks!
AlexDaniel` another option is to run `htop` or the like and just do it there, works too 07:16
systemd will restart bots if they're dead for any reason
jmerelo AlexDaniel`: OK
AlexDaniel` jmerelo: filed an issue here: github.com/perl6/whateverable/issues/313 07:24
jmerelo Just created the Whateverable tag in StackOverflow stackoverflow.com/questions/tagged/whateverable 08:36
It might come in handy for the next squashaton
squashable6: status
squashable6 jmerelo, Can't parse the wiki page 08:37
jmerelo Er
Geth doc: 544bd80924 | (JJ Merelo)++ | doc/Language/contexts.pod6
First version of the contexts document

This would close #732.
Problem is it is part of a larger document, #1225, which is by no means complete. Also was recently added to #114 as part of the language docs.		 08:54
synopsebot Link: doc.perl6.org/language/contexts
doc: 08e976bb54 | (JJ Merelo)++ | doc/Language/contexts.pod6
Adds numeric context

Refs #1225 and #124
doc: ce6422af84 | (JJ Merelo)++ | 2 files
Adds string context

Refs #114 and #1225. There are still many contexts to go, but these are the three most important.
I'll also revise #1225 to close it when enough meaningful contexts done. In general, most of them will be like ... (6 more lines)
AlexDaniel` jmerelo: ouch :) 09:08
Geth whateverable: 031cf5de52 | (Aleks-Daniel Jakimenko-Aleksejev)++ | bin/Squashable.p6
Change heuristic to support md cell alignment
 09:10
AlexDaniel squashable6: status 09:11
squashable6 AlexDaniel, Next SQUASHathon in 16 days and ≈0 hours (2018-06-02 UTC-12⌁UTC+14). See github.com/rakudo/rakudo/wiki/Mont...Squash-Day
AlexDaniel jmerelo: fixed, thanks
Geth doc: a993e37fab | (JJ Merelo)++ | doc/Language/contexts.pod6
Adds string contextualizer ~ refs #1225
 09:35
synopsebot Link: doc.perl6.org/language/contexts
thundergnat .tell Zoffix There are a total of 8 tests that specifically look for exactly 6 digits of precision for stringified representations that can't be exactly represented. EG. (2/3).Fatrat.Str is checked to be specifically 0.666667 10:01
yoleaux 01:18Z <Zoffix> thundergnat: in FatRat.Str: "speed increase, 16 digits would work fine; but it isn't spec"... Are you sure those tests were actually spec (as in 6.c-errata branch tests) and not just proptests (new tests in master branch)?
thundergnat: I'll pass your message to Zoffix.
thundergnat tell Zoffix As I recall there was 1 in S32-num/fatrat.t and 7 in S32-num/stringify.t. The tests exist in the errata branch too. 10:05
.tell Zoffix As I recall there was 1 in S32-num/fatrat.t and 7 in S32-num/stringify.t. The tests exist in the errata branch too.
yoleaux thundergnat: I'll pass your message to Zoffix.
tbrowder_ .tell Zoffix my Rakudo PR #1826 (which fixes your GH #1821) is ready for merging 10:53
yoleaux tbrowder_: I'll pass your message to Zoffix.
tbrowder_ .ask jnthn it’s may, how does one sign up for commaide? 13:11
yoleaux tbrowder_: I'll pass your message to jnthn.
jkramer m: my $b = bag 'a'..'c'; say $b<x>; say $b<a b x>; say $b<a b x>:p 13:32
camelia 0
(1 1 0)
(a => 1 b => 1)
jkramer Is this correct? Shouldn't the last say include a "x => 0"?
Geth doc: bd5617199a | (Will "Coke" Coleda)++ | doc/Language/contexts.pod6
whitespace
 13:33
synopsebot Link: doc.perl6.org/language/contexts
doc: 9abdb9550e | (Will "Coke" Coleda)++ | doc/Language/5to6-nutshell.pod6
fix typo
synopsebot Link: doc.perl6.org/language/5to6-nutshell
lucasb m: multi f(NaN) {}; multi f(NumStr) {}; multi f(ProtocolType) {}; multi f(UInt) {}; f 14:24
camelia ===SORRY!===
Circularity detected in multi sub types for &f
lucasb ^^ I wonder what's the interaction between these values?
mcmillhj jkramer: according to the docs on the subscript :p adverb, it skips non-existent elements: docs.perl6.org/language/subscripts..._%3Ap-%3Ap 15:03
jkramer: :!p can be used to non silently skip them 15:04
m: my $b = bag 'a' .. 'c'; say $b<a b x>:!p;
camelia (a => 1 b => 1 x => 0)
jkramer mcmillhj: Sweet, thank you! 15:10
hythm_ :m grammar G {token a {'a'}; token b {<?after 'a'> 'b'}; token TOP {<a><b>};}; G.parse('ab').say; # works 15:37
timotimo needs to be m: not :m
hythm_ m: grammar G {token a {'a'}; token b {<?after 'a'> 'b'}; token TOP {<a><b>};}; G.parse('ab').say; # works 15:38
camelia ｢ab｣
a => ｢a｣
b => ｢b｣
hythm_ m: grammar G {token a {'a'}; token b {<?after <a>> 'b'}; token TOP {<a><b>};}; G.parse('ab').say; # works
camelia ｢ab｣
a => ｢a｣
b => ｢b｣
hythm_ m: grammar G {token a {'az'}; token b {<?after 'az'> 'b'}; token TOP {<a><b>};}; G.parse('ab').say; # works
camelia Nil
hythm_ m: grammar G {token a {'az'}; token b {<?after 'az'> 'b'}; token TOP {<a><b>};}; G.parse('ab').say; 15:40
camelia Nil
hythm_ m: grammar G {token a {'az'}; token b {<?after 'az'> 'b'}; token TOP {<a><b>};}; G.parse('azb').say; # works 15:41
camelia ｢azb｣
a => ｢az｣
b => ｢b｣
hythm_ m: grammar G {token a {'a'}; token b {<?after <a>> 'b'}; token TOP {<a><b>};}; G.parse('ab').say; #Does not work works 15:42
camelia ｢ab｣
a => ｢a｣
b => ｢b｣
hythm_ Sorry for confusion, but looks like last one did not work on my box... it worked here though 15:44
timotimo committable6: releases grammar G {token a {'a'}; token b {<?after <a>> 'b'}; token TOP {<a><b>};}; G.parse('ab').say 15:47
committable6 timotimo, ¦releases (29 commits): «｢ab｣␤ a => ｢a｣␤ b => ｢b｣␤»
hythm_ m: grammar G {token a {'az'}; token b {<?after <a>> 'b'}; token TOP {<a><b>};}; G.parse('azb').say; # not working
camelia Nil
hythm_ Why this does not work^. If i replaced <a> with 'az' ..inside <?after...> ,, it works fine 15:49
timotimo hm, something wrong with the flipping algorithm, i wonder 15:52
mcmillhj m: my $b = bag 'a' .. 'c'; say $b<a b x>:!p; 15:56
camelia (a => 1 b => 1 x => 0)
mcmillhj oops, mt
tobs m: say $[1,2,3,4][*..*-3] 16:33
camelia Cannot convert -Inf to Int:
in block <unit> at <tmp> line 1
lucasb c: 2018.04.1 ()[*..5] 16:40
committable6 lucasb, ¦2018.04.1: «Cannot convert -Inf to Int: ␤ in block <unit> at /tmp/CV8y4wQJiA line 1␤␤ «exit code = 1»»
lucasb strange, locally I get "Unhandled exception: No exception handler located for warn" 16:41
tobs I get an additional "Use of uninitialized value $!reason of type Any in string context." but my rakudo is old 16:42
timotimo a bare * has a special meaning to the .. operator 16:59
m: say $[1, 2, 3, 4][^(*-2)]
camelia (1 2)
timotimo m: say $[1, 2, 3, 4][{ $_ .. $_ - 2 }]
camelia ()
timotimo m: say $[1, 2, 3, 4][{ $_ .. ($_ - 2) }] 17:00
camelia ()
timotimo oh?
m: say $[1, 2, 3, 4][{ say $_; $_ .. ($_ - 2) }]
camelia 4
()
timotimo m: say $[1, 2, 3, 4][{ say $_; say ($_ .. ($_ - 2)) }]
camelia 4
4..2
2
timotimo ah, of course
tobs m: say $[1, 2, 3, 4][{ $_ ... ($_ - 2) }]
camelia ((Any) 4 3)
timotimo ah, is that what you wanted? 17:00
tobs Almost. I wanted the last $k elements in reverse order, without an (Any) 17:01
I'm using using .reverse now :-)
*just using
timotimo my suggestion would probably be:
m: say $[1, 2, 3, 4].tail(4).reverse 17:02
camelia (4 3 2 1)
timotimo m: say $[1, 2, 3, 4].tail(2).reverse
camelia (4 3)
tobs Ah, yes. That reads better
jmerelo O/ 18:15
samcv i can't seem to install IRC::Client because of a failure in IO::Socket::Async::SSL (this is on the latest star) 18:58
lizmat bummer, also not with --force ? 18:59
samcv yeah well then the irc server doesn't work :)
i get Cannot locate symbol 'sk_num' in native library 'libssl.so' 19:00
samcv this seems similar? github.com/wkhtmltopdf/wkhtmltopdf/issues/3001 it says i need to get the legacy libssl (i'm on debian sid) or something 19:00
timotimo sounds like that's an unsupported version?
samcv what? sid? or openssl 1.1h? 19:01
seems like functions changed maybe
El_Che samvc: $ docker run -ti --entrypoint="" rakudo/ubuntu-amd64-18.04 bash -c "apt-get update && apt-get install -y libssl1.0 && zef install IRC::Client" 19:08
...
===> Installing: IRC::Client:ver<3.007006>
seems to work on Ubuntu if you install libssl1.0
(it's not a Star image, though)
Zoffix samcv: FWIW, installs fine on openssl-1.1.1-pre6 19:32
on 2018.04-20-g7847768 built on MoarVM version 2018.04-34-g25f165a 19:33
lucasb gist.github.com/lucasbuchala/70473...7dd46192ff 20:01
^^ can anyone confirm this behavior? 20:02
is it a bug?
Zoffix There's a bug that has same error R#131574 20:03
There's a bug that has same error RT#131574
synopsebot RT#131574 [new]: rt.perl.org/Ticket/Display.html?id=131574 [BUG] Incorrect circularity detection with infix:<+>
Zoffix Note the last comment on it tho: "Did anyone actually take all the candidates and manually draw out the DAG? :) 15:17 I can't imagine the cycle detection is a false positive."
lucasb hmm, you already stumbled upon it 20:06
samcv ZofBot: are you sure you don't have libssl1.0-dev? 20:18
or maybe the nondev. anyway i installed it and now it works
El_Che samcv: the way most nativecall modules find their libraries is flaky 20:20
samcv well it finds it 20:21
well. idk.
El_Che it find 1.0 not 1.1
what installed by default on recent distros 20:22
and the user has no clue he needs to install an older version
or, like on a SO question for the latest debian and dbiish: you need to create a link from the libmariadbclient lib to one with the name lf libmysql 20:23
samcv my post on the new perl 6 hashing changes and explaining the security issue it solves is now live: cry.nu/perl6/secure-hashing-for-moarvm/ 20:36
El_Che \o/ 20:37
moritz samcv++ 20:45
samcv i'm going to expand the user facing changes section to have more recommendations of do's and don'ts 21:40
samcv anybody want to check out this section and tell me if it can be improved? cry.nu/perl6/secure-hashing-for-mo...ng-changes 21:48
lizmat samcv: @stuff.push: $i; s/$i/$key/ ? 21:51
samcv ah yes 21:52
lizmat++ and let me know if there's more i should touch on
lizmat "identify items that need fixes" s/fixes/fixing/ ? 21:53
samcv had been way too long since i made a blog post 21:54
samcv also hoping the pictures were helpful. i just used draw.io to make it 21:55
comborico1611 samcv: Very good! 22:15
samcv thanks! 22:22
lucasb samcv: "Assuming s/everthing/everything/ is fine ..." 22:34
samcv lucasb: thanks! 22:36
timotimo does siphash make sure the resulting value is never 0 so we can use 0 as a value for "hash wasn't computed yet"? 22:42
buggable New CPAN upload: PDF-Class-0.2.0.tar.gz by WARRINGD modules.perl6.org/dist/PDF::Class:cpan:WARRINGD
samcv timotimo: uh. i don't know. i mean do we do that now? 22:43
timotimo i don't think we do at the moment
that was also mentioned in the c3 talk 22:44
samcv i don't see us doing that?
timotimo if we sacrifice one other value that we spit out if the result would have been 0, we can prevent hashing the same string object over and over again
samcv also eventually we may want to rekey our hashes if there are too many collisions, perl 5 does thta at least
though. i think they might use a less secure hashing by default, not sure. though when i add siphash i should be moving us to 64bit hashes 22:45
which will make collisions even rarer
timotimo right
does that also make all our hashes bigger? or just our strings for the cached hash value?
samcv timotimo: ah we do do that 22:46
timotimo how would re-keying work?
samcv though it should only happen 1/2147483648 chance 22:47
timotimo have a "salt" for every hash?
right, siphash makes it hard to get that exact value
samcv if our bucket is still full after doubling the buckets then we rekey
timotimo even harder than getting two strings with just the same hash value
samcv timotimo: let me see if i can make a hash that has 0. one sec 22:48
i've already separated out mvm's hash function and made code to brute force codes. so let me see 22:49
timotimo did you vectorize it yet? ;) 22:50
samcv haha 22:52
samcv timotimo: yes it can be one 22:57
(our current hash function, not sure about siphash but probably yes
timotimo sorry, can be one? 22:58
you mean can't be zero?
samcv no it can be zero
timotimo OK 22:59
samcv i mean we could add one to it?
timotimo then we can hit 0 again if we overflow? :D 23:00
n00b_ m: say -15 ** 0.5; say (-15, -16, -17) >>**>> 0.5; 23:02
camelia -3.872983346207417
(NaN NaN NaN)
n00b_ why doesn't the hyper work?
n00b_ How do I do sqrt for each scalar in an array (without having an explicit loop)? 23:03
timotimo you can >>.sqrt for the second one
n00b_ nice
I can never figure out when I need the pair >> >> vs when I can get away with one >>
samcv timotimo: overflow? 23:04
timotimo one >> is only for prefix or postfix operators and method calls
the reason why your two ** 0.5 pieces differ is because of the precedence between - and **
you're actually calculating -(15 ** 0.5)
whereas in the second one you're calculating (-15) ** 0.5
you can get complex numbers if you want, but you'll have to start out with a complex number first
m: say <-15+0i> ** 0.5 23:05
camelia 2.3715183290419594e-16+3.872983346207417i
timotimo m: say (<-15+0i> ** 0.5) ** 2
camelia -15+1.83697019872103e-15i
timotimo does that help?
n00b_ It does help, yes. 23:06
At least I know the syntax >>**>> 0.5 is meaningful
brb
timotimo :)
n00b_ For the sake of readability, I'd like to remove the extra parenthesis in ( [Z-] ( (1,2,3) , (4,5,6) ) >>**>> 2 )>>.abs>>.sqrt so that I could say something like .sqrt>> .abs>> [Z-] ( (1,2,3) , (4,5,6) ) >>**>> 2 23:23
Is there some syntax I could make use of to move the sqrt and abs to the left of the expression?
timotimo yup 23:24
m: say .>>sqrt.>>abs given [Z-] ( (1,2,3) , (4,5,6) ) >>**>> 2
camelia 5===SORRY!5=== Error while compiling <tmp>
Malformed postfix call (only alphabetic methods may be detached)
at <tmp>:1
------> 3say .7⏏5>>sqrt.>>abs given [Z-] ( (1,2,3) , (4,5
timotimo m: say $_.>>sqrt.>>abs given [Z-] ( (1,2,3) , (4,5,6) ) >>**>> 2
camelia 5===SORRY!5=== Error while compiling <tmp>
Missing dot on method call
at <tmp>:1
------> 3say $_.>>7⏏5sqrt.>>abs given [Z-] ( (1,2,3) , (4,5,6
expecting any of:
postfix
timotimo m: say $_>>.sqrt>>.abs given [Z-] ( (1,2,3) , (4,5,6) ) >>**>> 2
camelia (NaN NaN NaN)
timotimo oh?
n00b_ ah, nice, I forget that $_ is there, because it's so often elidable
timotimo m: say [Z-] ( (1,2,3) , (4,5,6) ) >>**>> 2 23:25
camelia (-15 -21 -27)
timotimo ah, right, of course that becomes NaN xx 3
n00b_ m: $_>>.abs>>.sqrt given [Z-] ( (1,2,3) , (4,5,6) ) >>**>> 2 23:26
camelia ( no output )
n00b_ m: say $_>>.abs>>.sqrt given [Z-] ( (1,2,3) , (4,5,6) ) >>**>> 2
camelia (3.872983346207417 4.58257569495584 5.196152422706632)
timotimo ah, that's why
n00b_ m: say $_>>.abs>>.sqrt given 2 R>>**>> [Z-] ( (1,2,3) , (4,5,6) ) 23:27
camelia (3 3 3)
n00b_ m: say $_>>.abs>>.sqrt given ( 2 R>>**>> [Z-] ( (1,2,3) , (4,5,6) ) ) 23:27
camelia (3 3 3)
timotimo should also be able to <<R**<< 23:28
though i suppose the R>>**>> version is a little faster
n00b_ I think your version is easier to understand
or rather maps better to the concept I'm thinking of
timotimo there's also feeds, fwiw 23:29
i don't use them often at all, so let's see if i remember them right
oh, no, that would need maps, that'd be dumb
n00b_ m: 2 RZ** (1,2,3) 23:30
camelia Potential difficulties:
Useless use of RZ** in sink context
at <tmp>:1
------> 032 7⏏5RZ** (1,2,3)
n00b_ m: say 2 RZ** (1,2,3)
camelia (1)
n00b_ m: say (1,2,3) Z** 2
camelia (1)
n00b_ hmm
timotimo Z stops at the shorter list
m: say (1,2,3) Z** 2 xx *
camelia (1 4 9)
n00b_ Nice. 23:31
I think I'm happy with $_>>.abs>>.sqrt given ( 2 <<R**<< [Z-] ( (1,2,3) , (4,5,6) ) ) . 23:32
timotimo m: say >>.abs given (1, -2, 3)
camelia 5===SORRY!5=== Error while compiling <tmp>
Missing << or >>
at <tmp>:1
------> 3say >>.7⏏5abs given (1, -2, 3)
timotimo right, needs the $_ to work
n00b_ I'm trying to get closer to the APL way of doing things, an array-at-a-time, without having to tell the compiler that I want a loop, which feels to me like its job
TimToady m: say (1,2,3) X** 2 23:33
camelia (1 4 9)
timotimo oh, of course 23:33
i'll take this as a sign that i should really go to bed soon
n00b_ m: 2 RX** (1,2,3)
camelia Potential difficulties:
Useless use of RX** in sink context
at <tmp>:1
------> 032 7⏏5RX** (1,2,3)
n00b_ m: say 2 RX** (1,2,3)
camelia (1 4 9)
n00b_ wooo!
that's slick.
TimToady actually...
TimToady m: say 2 RX** 1,2,3 23:33
camelia (1 4 9)
TimToady X and Z are listops, so you don't really need the parens on the lists 23:34
(unlike hypers)
n00b_ Thanks, but in my case it'll be [Z-] ((1,2,3) , (4,5,6))
or rather @a and @b because it's a sub
TimToady so you don't need the outer parens, prolly
n00b_ m: say 2 RX** [Z-] (1,2,3) , (4,5,6) 23:35
camelia (9 9 9)
n00b_ hahah
love it
n00b_ I'm a little sad about the $_ needed in $_>>.abs>>.sqrt ... but c'est le vie 23:35
timotimo if you don't want the $_, you can also .map(*.abs.sqrt) 23:36
TimToady m: say .».abs for -1, -2, -3
camelia 5===SORRY!5=== Error while compiling <tmp>
Malformed postfix call (only alphabetic methods may be detached)
at <tmp>:1
------> 3say .7⏏5».abs for -1, -2, -3
tobs m: *».abs».sqrt with 1,-2,3
camelia ( no output )
tobs m: *».abs».sqrt.say with 1,-2,3
camelia (1 1.4142135623730951 1.7320508075688772)
n00b_ Oh! I like that!
Whatever stars are one of my favorite parts of P6 23:37
tobs I stumbled upon it earlier today as well :-)
TimToady m: say .self».abs for -1, -2, -3 23:38
camelia (1)
(2)
(3)
TimToady look no $_ ;)
timotimo then you'll also not need the »
n00b_ m: sub cartesian_distance(@a, @b) { *>>.abs>>.sqrt with 2 RX** [Z-] @a, @b }; say cartesian_distance( (1,2,3) , (4,5,6) ) 23:39
camelia (3 3 3)
n00b_ nice
very clean
TimToady m: say .self».abs for (-1, -2, -3), (-4,-5,-6)
camelia (1 2 3)
(4 5 6)
n00b_ duh, forgot the su
m: sub cartesian_distance(@a, @b) { *>>.abs>>.sqrt with [+] 2 RX** [Z-] @a, @b }; say cartesian_distance( (1,2,3) , (4,5,6) )
camelia (5.196152422706632)
n00b_ Look ma, no loops. 23:40
tobs Perl 6 has all the loops you'll ever need built in 23:41
TimToady
.oO(∞ loops should be enough for anyone...)
 23:42
n00b_ Alright, now that we have cartesian_distance, here's the $64K question: given MxN matrix of M points in N-dimensional space, how do I calculate the MxM distance matrix from each point to every other point, sans loops?
timotimo this is the point where i'd probably go for at least one explicit loop, weak little me 23:43
n00b_ I know. I ask because i'm trying to solve this little twitter puzzle: twitter.com/Hillelogram/status/996...3456390144 23:44
You can ignore most of the code there. We've written +/&.:*:@:- as cartesian_distance here.
The point of the challenge is to write the equivalent to J's / operator
n00b_ which is a generalized outer product operator: x +/ y is the sum table, x */ y is the product table, x ^/ y is the exp table 23:45
it's super useful in everyday programming, as odd as that seems, given outer products are big & expensive, but it's used all the time
I'd like an operator that does that in P6
(the "1 is also relevant, but leave that for now) 23:46
timotimo rosettacode.org/wiki/Kronecker_product#Perl_6 - is this relevant? 23:47
it has an explicit loop, though
n00b_ it's thematically related, but that seems like it does a mmult for each scalar in A with the entire matrix B 23:48
TimToady if you're generalizing to more dimensions, hypers are likely to be better behaved than listops, especially if you use the dwimmy ones «op» which will do the autorepeating at any level that needs it 23:49
n00b_ My gut was we could press X into service
but if it stops at the shorter list, that's problematic
timotimo only Z does that
TimToady but X only extends a scalar to a list at the top level
timotimo X goes through every item on both sides
true that
TimToady «op» doesn't stop at the shorter one 23:50
n00b_ (1,2,3) op/ (4,5) is (1 op>> (4, 5)), (2 op>> (4,5)), (3 op>> (4,5))
TimToady hmm 23:51
n00b_ m: say ((1, 2, 3), (4, 5, 6)) << <<+>> >> ((7,8,9), (10,11,12))
camelia 5===SORRY!5=== Error while compiling <tmp>
Unsupported use of << to do left shift; in Perl 6 please use +< or ~<
at <tmp>:1
------> 3say ((1, 2, 3), (4, 5, 6)) <<7⏏5 <<+>> >> ((7,8,9), (10,11,12))
n00b_ m: say ((1, 2, 3), (4, 5, 6)) <<<<+>>>> ((7,8,9), (10,11,12)) 23:52
camelia 5===SORRY!5=== Error while compiling <tmp>
Missing << or >>
at <tmp>:1
------> 3say ((1, 2, 3), (4, 5, 6)) <<<<+>>>7⏏5> ((7,8,9), (10,11,12))
expecting any of:
infix
infix stopper
TimToady «op» is more like APL, which would autorepeat 4,5,4,5,4,5...
n00b_ Yes, that's what I want.
timotimo hyperops descend into substructures, too
n00b_ m: say ((1, 2, 3), (4, 5, 6)) <<+>> ((7,8,9), (10,11,12))
camelia ((8 10 12) (14 16 18))
n00b_ oh!
timotimo m: my %left = :1hey, :2bye, :3foo; my %right = :9boo, :8bye, :4foo; say %left «~» %right 23:53
camelia {bye => 28, foo => 34}
timotimo oh, i thought if you point the pointies at the hashes missing keys would be treated as "no operand from here" but still show up in the result? 23:54
m: my %left = :1hey, :2bye, :3foo; my %right = :9boo, :8bye, :4foo; say %left »~« %right
camelia {boo => 9, bye => 28, foo => 34, hey => 1}
n00b_ m: sub postfix:<c_d> { *>>.abs>>.sqrt with [+] 2 RX** [Z-] @^a, @^b }; say (1, 2, 3) c_d (4, 5, 6) 23:55
camelia 5===SORRY!5=== Error while compiling <tmp>
Two terms in a row
at <tmp>:1
------> 3] 2 RX** [Z-] @^a, @^b }; say (1, 2, 3)7⏏5 c_d (4, 5, 6)
expecting any of:
infix
infix stopper
postfix
sta…
TimToady s/post/in/ ??? 23:56
Herby_ o/
n00b_ m: sub postfix:<c_d>(@a, @b) { *>>.abs>>.sqrt with [+] 2 RX** [Z-] @a, @b }; say (1, 2, 3) c_d (4, 5, 6)
camelia 5===SORRY!5=== Error while compiling <tmp>
Two terms in a row
at <tmp>:1
------> 3[+] 2 RX** [Z-] @a, @b }; say (1, 2, 3)7⏏5 c_d (4, 5, 6)
expecting any of:
infix
infix stopper
postfix
sta…
TimToady that looks like an infix to me
n00b_ haha, yes
m: sub infix:<c_d>(@a, @b) { *>>.abs>>.sqrt with [+] 2 RX** [Z-] @a, @b }; say (1, 2, 3) c_d (4, 5, 6)
camelia (5.196152422706632)
n00b_ m: sub infix:<c_d>(@a, @b) { *>>.abs>>.sqrt with [+] 2 RX** [Z-] @a, @b }; say ((1, 2, 3), (4, 5, 6)) <<c_d>> ((7,8,9), (10,11,12)) 23:56
camelia Type check failed in binding to parameter '@a'; expected Positional but got Int (1)
in sub infix:<c_d> at <tmp> line 1
in block <unit> at <tmp> line 1
 23:57
n00b_ hmm
Do hypers necessarily descend all the way to scalars? 23:58
TimToady how is it supposed to intuit what level to stop at, if not at the bottom? 23:59
n00b_ the type signature was my thought
the bottom-most thing that matches the type sig
TimToady we know with unaries to check for 'is nodal', but binaries ignore that

Powered by Cro and the Raku® Programming Language.

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

Thank you!