»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.
00:01 pyrimidine joined 00:05 pyrimidine left 00:06 Cabanossi left 00:08 Cabanossi joined 00:10 jp_ left 00:14 cdg joined 00:17 cdg_ left 00:19 cdg left 00:27 wamba left 00:32 agentzh joined 00:37 pyrimidine joined 00:44 aborazmeh joined, aborazmeh left, aborazmeh joined 00:46 dugword joined, ocbtec left
Geth oc: 29295940fe | (Will "Coke" Coleda)++ | 2 files
Don't generate .png typegraph files
 00:47
oc: 525e94eec1 | (Will Coleda)++ | 2 files
Merge pull request #1131 from perl6/coke/pngless

Don't generate .png typegraph files
00:51 dugword left, cyphase left 00:56 cyphase joined
Geth oc: antquinonez++ created pull request #1171:
added line breaks to grammar_tutorial		 00:57
01:04 bpmedley joined 01:06 astj joined 01:11 astj left, agentzh left, cdg joined 01:12 LeCamarade joined, LeCamarade left
Geth oc/master: 4 commits pushed by Altai-man++ 01:12
01:13 astj joined, LeCamarade joined 01:15 sena_kun left 01:19 LeCamarade left 01:25 kyan left 01:33 labster left 01:40 agentzh joined
brokenchicken wtf.. travis just emailed me about a failed build... of a commit from JAN 19! 01:40
cale2 So "grep" is perl's version of "filter" 01:41
I was wondering for the longest time why there was no filter
brokenchicken
.oO( must've been delayed by Donald's administration... )
cale2 as someone that doesn't have a background in unix admin stuff, that's something I found confusing 01:43
brokenchicken Filter's a strange name since you're doing the opposite. 01:44
01:47 espadrine_ left
cale2 `'`'`'`'`'` filter onlySingleQuotes ''''' 01:47
Just depends on how you write your filtering function
01:48 BenGoldberg joined
cale2 Like you're filtering out the bad stuff 01:48
where does "grep" even come from?
"get representation"
BenGoldberg global regular expression print 01:49
cale2 Is grep just one of those things that you'd never think to change? Because it's so loved and gives warm fuzzy feelings, etc 01:50
BenGoldberg Perl has always had a grep builtin... 01:51
cale2 Just wondering how many things are in the language because "it's always been that way" 01:52
BenGoldberg You could of course add some sort of alias, like "filter", but most programmers are used to the name grep.
cale2 I guess P6 isn't necessarily about starting completely over. More like revolution rather than evolution 01:53
or the other way around haha
BenGoldberg The *feature* grep is important, the *name* grep slightly less so. Removing that feature entirely would be stupid.
Renameing it, while possible, is kinda pointless.
02:01 mst joined 02:02 mst left, mst joined
masak TimToady: sorry about the miscategorization of your hat. can I call it an "Aussie hat" next time? :) 02:06
02:06 user__ joined
masak TimToady: any thoughts on BEGIN firing per expansion? do you mind the semi-consensus that emerged yesterday (that it does)? 02:06
another thing that pushed me in that direction was: we don't mind at all that END fires per expansion. 02:07
02:09 kalkin- left
brokenchicken What's an expansion? 02:12
masak brokenchicken: it's when code returned from a macro gets put in the code where the macro call was. 02:13
brokenchicken Ah
masak it's an old Lisp term. 02:15
little bit misleading, since the result can be smaller, bigger, the same size or even divergent 02:16
brokenchicken No, makes sense to me :) 02:18
cale2 the new rakudo star release says advanced macros aren't implemented yet. what does that mean exactly?
masak TimToady: also, there are two kinds of consistency at odds here for quasis with BEGIN blocks without unquotes. they could fire ASAP (once) or per expansion.
cale2 simple macros work, but advanced ones dont? 02:19
masak cale2: there's no short answer. but somewhat simply, I wouldn't want to use the macros that are implemented so far.
cale2: if you're *really* interested in Perl 6 macros, then the development of 007 is worth paying attention to.
it has some cutting-edge insights into macros and slangs that sometimes leak over into this channel. 02:20
implementation of all those new things will commence in Rakudo at some point. don't know exactly when.
cale2: "advanced macros" are based off something called the Qtree. it's an AST structure for the program. 02:21
02:21 cdg left
cale2 Is it basically Racket macros in perl6? 02:22
02:28 travis-ci joined
travis-ci Doc build errored. Will Coleda 'Merge pull request #1131 from perl6/coke/pngless 02:28
travis-ci.org/perl6/doc/builds/196794815 github.com/perl6/doc/compare/c17ac...5e94eec1d5
02:28 travis-ci left 02:30 lostinfog left
brokenchicken hmmm... moar configure failed :/ 02:35
02:35 Cabanossi left
brokenchicken don't really say why 02:36
Ahhh git error: fatal: unable to access 'github.com/libuv/libuv.git/': Failed connect to github.com:443; Connection timed out
samcv :\
02:37 LeCamarade joined, Cabanossi joined 02:40 _28_ria left 02:41 hartenfels joined 02:42 _28_ria joined, gregf_ left 02:45 MasterDuke joined 02:46 dugword joined 02:47 travis-ci joined
travis-ci Doc build errored. Altai-man 'Merge pull request #1165 from Altai-man/comments 02:47
travis-ci.org/perl6/doc/builds/196797707 github.com/perl6/doc/compare/525e9...efe96978ce
02:47 travis-ci left 02:54 _28_ria left 02:56 _28_ria joined 02:58 LeCamarade left, LeCamarade joined 03:06 Cabanossi left 03:08 Cabanossi joined 03:16 cibs left 03:18 cibs joined 03:19 agentzh left 03:20 dugword left 03:21 dugword joined 03:29 xtreak joined 03:32 LeCamarade left 03:34 LeCamarade joined 03:35 avuserow left 03:37 avuserow joined 03:39 noganex_ joined 03:43 noganex left 03:48 cdg joined 03:51 cale2 left 03:59 xtreak_ joined 04:01 xtreak left, xtreak joined 04:02 Cabanoss- joined 04:04 xtreak_ left 04:05 solarbunny left 04:06 Cabanossi left, Cabanoss- is now known as Cabanossi 04:12 solarbunny joined 04:18 LeCamarade left 04:20 itcharlie1 joined
brokenchicken POTATO! 04:29
04:32 khw left 04:36 BenGoldberg left 04:48 dugword left 04:53 _28_ria left 04:55 _28_ria joined 04:56 itcharlie1 left 04:57 dugword joined 05:00 Cabanoss- joined 05:02 agentzh joined, cdg left 05:03 Cabanossi left, Cabanoss- is now known as Cabanossi 05:05 Vynce left 05:27 astj left 05:30 labster joined
samcv brokenchicken, what's the best way to create all possible orderings of an array? 05:32
m: my @a = 1,2,3; @a.op # (1,2,3), (1,3,2), (3,2,1), (3,1,2) etc 05:33
camelia rakudo-moar 3e28b1: OUTPUT«No such method 'op' for invocant of type 'Array'␤ in block <unit> at <tmp> line 1␤␤»
masak cale2: re "grep" -- I feel your confusion, FWIW. 05:37
cale2: I had that function/method as "grep" in 007 for a long while too, until I took a fresh look at it and decided that the name doesn't make sense unless you have a very specific (Unix-y) history. 05:38
"filter" ain't perfect, no, but it's probably the least bad among the various contenders.
samcv a Grep is a goblin like creature 05:39
well that's what the name sounds like at least
masak, any clue how to generate all possible orders of a list in p6?
masak my s/grep/filter/ commit at the time explains my reasoning: github.com/masak/007/commit/9097ba...1438bb9b77 05:40
samcv: I believe that's a built-in these days
m: say [1, 2, 3].permutations.perl
camelia rakudo-moar 3e28b1: OUTPUT«((1, 2, 3), (1, 3, 2), (2, 1, 3), (2, 3, 1), (3, 1, 2), (3, 2, 1)).Seq␤»
samcv nice 05:41
masak yeah, we're nice that way
samcv prolly gonna use that for my bitfield packing
masak Perl 6 no longer has the problem that there are too *few* built-ins :P
I think sometimes we're slightly bad at holding back and looking for that wonderful generalization that could reduce the built-ins a tad more 05:42
perhaps that's an inevitable consequence of the community having as much steerage as it does
samcv i think that's the most trivial way to do it. since it's much easier to write a function to compute how items in a bitfield will be padded than to write something to try and find the best possible ordering programmatically
heh
masak also, I happened to see the .kxxv method name the other day, and was reminded how I wanted to institute an "ugliest built-in name ever" prize :P 05:43
05:43 agentzh left
masak (no, I don't have a better suggestion) 05:43
samcv heh
TEttinger yay bitfields 05:45
is this for unicode stuff, samcv? 05:46
samcv yea
05:50 agentzh joined 05:52 Vynce joined 05:55 _28_ria left 05:58 _28_ria joined, CIAvash joined
Geth oc: antquinonez++ created pull request #1172:
edits to exceptions pod		 06:20
06:21 dugword left 06:25 bjz joined 06:29 agentzh left 06:31 agentzh joined 06:32 domidumont joined 06:35 hartenfels1 joined 06:36 bjz left 06:37 hartenfels left, hartenfels1 is now known as hartenfels 06:38 domidumont left, domidumont joined 06:40 domidumont left 06:41 jraglin left 06:44 agentzh left 06:52 RabidGravy joined, xtreak left 06:53 travis-ci joined
travis-ci Doc build passed. Antonio Quinonez 'edits to exceptions pod' 06:53
travis-ci.org/antquinonez/doc/builds/196846373 github.com/antquinonez/doc/commit/bd26cbe7509b
06:53 travis-ci left 06:55 xtreak joined
masak m: macro swap($a, $b) { quasi { ({{{$a}}}, {{{$b}}}) = {{{$b}}}, {{{$a}}} } }; my $w = 5; my $z = 10; swap $w, $z; say [$w, $z].perl 07:00
camelia rakudo-moar 3e28b1: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Use of macros is experimental; please 'use experimental :macros'␤at <tmp>:1␤------> 3macro7⏏5 swap($a, $b) { quasi { ({{{$a}}}, {{{$b␤»
07:00 domidumont joined
masak m:use experimental :macros; macro swap($a, $b) { quasi { ({{{$a}}}, {{{$b}}}) = {{{$b}}}, {{{$a}}} } }; my $w = 5; my $z = 10; swap $w, $z; say [$w, $z].perl 07:00
m: use experimental :macros; macro swap($a, $b) { quasi { ({{{$a}}}, {{{$b}}}) = {{{$b}}}, {{{$a}}} } }; my $w = 5; my $z = 10; swap $w, $z; say [$w, $z].perl 07:01
camelia rakudo-moar 3e28b1: OUTPUT«[10, 5]␤»
masak \o/
07:04 aborazmeh left
masak tried the corresponding macro in 007, and found a bug :P 07:06
07:12 wamba joined 07:15 hchienjo joined 07:16 hchienjo left 07:17 jraglin joined 07:18 vendethiel left 07:20 darutoko joined 07:25 mr-foobar left 07:27 domidumont left 07:31 mr-foobar joined 07:35 domidumont joined 07:36 mr-foobar left 07:37 agentzh joined 07:47 agentzh left 07:48 e1 joined
samcv masak, i want to use a macro to simplify this: 07:57
macro dump($a) { $a.VAR.^name ~ $a.gist }
i try this and i get "Too few positionals passed; expected 3 arguments but got 2"
masak m: my $a = [1, 2, 3]; say :$a.perl 07:58
camelia rakudo-moar 3e28b1: OUTPUT«:a($[1, 2, 3])␤»
masak :)
samcv that doesn't answer my question
masak agreed
samcv tho :\
masak I notice that your macro doesn't have a quasi 07:59
samcv i know .perl, but i am doing some custom dumping, and want to macro it
samcv has no clue what quasi does
masak they are the thing you put your generated code in
otherwise the code just ends up running as part of the macro call
samcv m: macro dump($a) { quasi { $a.VAR.^name ~ $a.gist } }; my $var = 'test'; say dump($var)
camelia rakudo-moar 3e28b1: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Use of macros is experimental; please 'use experimental :macros'␤at <tmp>:1␤------> 3macro7⏏5 dump($a) { quasi { $a.VAR.^name ~ $a.gi␤»
samcv m: use experimental :macros; macro dump($a) { quasi { $a.VAR.^name ~ $a.gist } }; my $var = 'test'; say dump($var) 08:00
camelia rakudo-moar 3e28b1: OUTPUT«ScalarAST.new␤»
samcv :\
what's the closest to c's #define ?
i literally only need verbatim insertion
masak m: use experimental :macros; macro dump($x) { quasi { say VAR({{{$x}}}).name, ": ", {{{$x}}}.perl } }; my $foo = "OH HAI"; dump($foo)
camelia rakudo-moar 3e28b1: OUTPUT«$foo: "OH HAI"␤»
masak there ya go 08:01
samcv so many {{{}}}
08:01 bwisti left
masak feel free to write it in any other way that works 08:01
samcv what works ;P
TimToady masak: the consensus view seems fine to me
samcv ?
masak also, you're welcome ;)
TimToady: cool
geekosaur samcv, quasi "quotes" code for later evaluation (at run time); {{{ }}} is what "unquotes" stuff to be substituted into the quoted code 08:02
masak samcv: you're not the first person to complain about the {{{ }}}
samcv: there's a lighter syntax being thrown around in 007 (though NYI): `unquote(...)`
samcv can we have a 'define'? 08:03
geekosaur in common lisp, quasi is `() and {{ }}} is ,@ or , depending on whether lists are to be "flattened" or not (respectively)
samcv that can do less fancy things?
masak samcv: no :)
samcv :|
masak samcv: at least, I'm not going to implement one
someone else might get a kick out of repeating C's mistakes. not me.
samcv i guess i'll just have to type lots of braces
haha
masak no-one is forcing you to use macros
samcv well i can't reliably do .VAR.^name without them 08:04
if I make a sub even is pure i often get some random thing outputted that's not the variable's name
masak this is true
VAR(...).name is a very macro-ish wish to have 08:05
note that it's .name, too, not .^name as you wrote
TimToady goodness, it's tomorrow already
samcv no it is .name 08:06
m: my $var; say $var.VAR.^name; say $var.VAR.name;
camelia rakudo-moar 3e28b1: OUTPUT«Scalar␤$var␤»
samcv oh. yeah
sorry i got mixed up
i will take my macros where I can get them though ;) 08:07
08:07 khw joined
masak samcv: I wouldn't be so cranky about `#define` if I didn't deeply believe that what Perl 6 is aiming for is strictly better ;) 08:09
samcv agreed lol
i just want less {{}} tbh
geekosaur perl6 specifically went with that because it was ugly (and therefore highly obvious that Something Weird is going on) 08:10
samcv ah
<----begin weirdness---->
geekosaur because $x means the $x at the time the expanded macro is evaluated, but {{{$x}}} is the $x at the time the macro is being expanded
masak "ugly" here can actually be taken to mean "doesn't collide with normal language syntax" 08:11
see strangelyconsistent.org/blog/macros...eholdeeers
08:12 lolo78 joined 08:14 xtreak left 08:16 xtreak joined
samcv masak, are you implying perl 5 uses macros to do sprintf? 08:20
masak no, not implying that 08:22
samcv "Don't believe me when I say that all templating syntax sucks? Fine, I'll bring you examples:" not sure what you mean by this?
geekosaur no, the %-escapes in the format string are a similar templating notion
samcv it's just a bad way to template items into the string?
ah
yeah that's pretty bad
geekosaur and TT as the second example :) 08:23
08:24 TimToady left
samcv wow python's is really weird 08:24
08:25 tojo joined
arnsholt samcv: I know! What weirdness did you trip over today? 08:25
samcv docs.python.org/3/library/string.h...matstrings
08:26 TimToady joined
tojo how can I take class name from variable? something like this: my $obj = $classname.new 08:27
samcv m: my $var = 'te'; say $var.WHAT.name;
camelia rakudo-moar 3e28b1: OUTPUT«No such method 'name' for invocant of type 'Str'␤ in block <unit> at <tmp> line 1␤␤»
samcv m: my $var = 'te'; say $var.WHAT.^name;
camelia rakudo-moar 3e28b1: OUTPUT«Str␤»
samcv like that kind of class name?
masak m: my $var = 'te'; say $var.^name 08:28
camelia rakudo-moar 3e28b1: OUTPUT«Str␤»
masak samcv: no need for .WHAT
samcv oh. ok 08:29
08:29 e1 left
tojo sorry, bad question.. i hold a class name in variable and want to use that to make object from that class name 08:29
for 'Foo', 'Bar' -> $t { my $obj = $t.new; } 08:30
masak m: for Int, Str -> $t { say $t.new } 08:31
camelia rakudo-moar 3e28b1: OUTPUT«0␤␤»
masak m: for Int, Str -> $t { say $t.new.perl }
camelia rakudo-moar 3e28b1: OUTPUT«0␤""␤»
CIAvash tojo: ::($t).new 08:32
samcv is there a page on perl 6 special methods and stuff? 08:33
like .^methods and uh, how to get the possible signatures of a method or sub
tojo CIAvash: yes! that worked, thanks :)
samcv tojo, docs.perl6.org/language/packages#i...A%3A%28%29 08:34
08:35 dakkar joined 08:36 parv joined
tojo samcv: thx! that page didn't come up on my previous searches 08:37
gfldex samcv: .^ is the same as .HOW, so you need to get the *HOW and look that up in the docs 08:41
most of the time you will be looking for docs.perl6.org/type/Metamodel$COLO...ONClassHOW
samcv what am i doing wrong here gist.github.com/b34f9f7a9a5ee6c9b6...56cae47b3d 08:43
stderr: Too few positionals passed; expected 2 arguments but got 1
gfldex samcv: you are trying to call $c in line 9 08:45
samcv well i created the variable on line 7 08:46
gfldex samcv: try this gist.github.com/gfldex/006d385aa45...b5727cda17 08:48
samcv yeah i was about to try that
08:49 KDr2 joined, g4 joined, g4 left, g4 joined
gfldex samcv: actually, your version should have worked to 08:50
08:52 as__ joined
as__ radudo: my $x = 42; say $x[0]; 08:52
rakudo: my $x = 42; say $x[0]; 08:53
camelia rakudo-moar 3e28b1: OUTPUT«42␤»
samcv also what's the shortcut for getting the last element of an array?
as__ Guys is it because there is a default circumfix for []?
gfldex [*-1]
or .tail(1)
samcv [*-1] seems a lot faster 08:54
gfldex as__: yes
method calls are always dynamic and can't be inlined 08:55
moritz they can, heuristically, at run time
gfldex does that depend on traits? 08:56
geekosaur as__, it's because there's code to pretend that a scalar is a single-element list when used in a place that requires a list 08:57
as__ aha ok thanks! 08:58
parv is there any plan for ";" to be the implicit default on perl6 prompt (REPL), as "my $x = 2" gives error but not with ";"?
09:00 KDr2 left
hartenfels parv: `my $x = 2` on the REPL doesn't give me any errors. 09:02
09:02 ocbtec joined 09:04 jonas2 joined
parv after definition+assignment with ";" (my $x = 2;), there is no problem afterwords. with any new variables sans end ";". 09:07
09:08 rindolf joined
parv where can i post screenshot without an account? 09:08
hartenfels parv: you could post the text to pastebin. 09:09
09:10 aindilis` joined 09:12 aindilis left
parv error: pastebin.com/DTWXN2ws 09:15
09:16 mxco86 joined
hartenfels parv: It works fine for me, maybe your Perl 6 is really old? What do you get when you run `perl6 -v`? 09:16
parv i am using rakudo ae9d517
09:16 domidumont left 09:17 domidumont joined
hartenfels I'm not sure then, because it's supposed to work I think. 09:19
parv actual version output: pastebin.com/sEZ1NuW0 09:21
thanks for trying, hartenfels
09:24 rba__ left 09:25 wamba left 09:27 pyrimidi_ joined, pyrimidine left 09:35 tojo left, zakharyas joined 09:42 domidumont left 09:50 xtreak left, user10 joined 09:51 xtreak joined 09:52 user9 left, user10 is now known as user9 09:54 breinbaas joined 09:56 hartenfels left, carl_ joined
carl_ working? 09:56
hello beings
samcv hello 09:57
hi carl_
carl_ quick question... 09:58
want to do something like this: if $a.WHAT eq (Int) {} , but no good.
moritz if $a ~~ Int { }
that's a type check that propery respects subtypes 09:59
carl_ ~~ ? I'll look it up
moritz it's the smart-match operator
10:00 devtom30 joined 10:03 wamba joined
parv is there really no way to compile+install rakudo/star from github to get the latest rakudo/rakudo? 10:06
else, i woud need to gather & install the bits myself to add to plain rakudo install. 10:08
samcv ok this isn't going to work. using permutations. gonna have to actually write something to pack the bits myself :\ oh well
yeah i guess i can see how they can add up... 10:09
and sadly permutations is naive to repeated numbers
m: (1,1).permutations
camelia ( no output )
samcv m: (1,1).permutations.say
camelia rakudo-moar 3e28b1: OUTPUT«((1 1) (1 1))␤»
stmuk_ parv: there is a Task::Star which resembles Rakudo Star although I suspect the module is dsynced
parv stmuk_: thanks, i will check that. 10:10
geekosaur desynched and currently broken
stmuk_ there is a report on twitter that 2017.01 release doesn't work in at least VS15 (fixed after release)
doesn't compile I mean 10:11
10:13 domidumont joined
parv rakudo/star/README is frankly misleading in that "perl Configure.pl" dies noting that I tried to run it from clone of the git repo. (of course 10:13
)
samcv .tell AlexDaniel East_Asian_Width property has just been merged into MVM 10:14
yoleaux samcv: I'll pass your message to AlexDaniel.
stmuk_ parv: I don't see why it's misleading .. it's intended to catch a common mistake .. not handle people being "clever" :) 10:15
10:15 rba_ joined
moritz parv: in the early days of R*, people would regularly clone the star repo and expected it to behave like the tarball 10:16
hence the current behavior
10:20 wamba left
timotimo eco: fig 10:20
ah, it's wig!
.tell cale2 have a look at the wig module :) 10:21
yoleaux timotimo: I'll pass your message to cale2.
10:21 geekosaur left 10:22 geekosaur joined
parv well, i am back to plain rakudo then: github.com/rakudo/star/issues/28 (--xform error on FreeBSD 11) 10:22
stmuk_ you could install gmake from ports
timotimo um 10:23
parrot hasn't been in rakudo star for a few years
10:23 zakharyas left
timotimo also, if it's only to be used by the release creator, what's your problem with it? 10:24
parv timotimo: it's the tar option that causes the similar error (with MoarVM)
timotimo oh? but why do we need that?
just to throw out the version number?
stmuk_ the easiest fix is to look for gnu make and die with an error saying to install it
10:24 wamba joined
timotimo why would gnu make magically give your tar executable a --xform option? 10:25
stmuk_ because it's a gnu make only option
timotimo parv: i think you're missing the point of rakudo star
what. 10:26
parv timotimo: only the bug submitter mentions that "only to be used by the release creater" without any follow up
timotimo rakudo.org/2017/01/30/announce-raku...e-2017-01/
you download this
parv timotimo: i am aware of that release. 10:27
in any case, i am not trying any more to complile myself rokudo star.
timotimo what do you mean by "gather & install the bits"?
you can also just install rakudo manually, and then use the "install all modules" part of rakudo star to get the modules 10:28
parv find zef/panda etc.; then install it locally.
timotimo not necessarily
parv ? 10:29
timotimo oh, rakudo star does use zef for that
my apologies
10:29 wamba left
Ulti whoa tests down to 0.8s 10:31
timotimo stmuk_: i still don't understand why installing gnu make would change how the tar command works. or does gnu make also come with gnu tar?
parv stmuk_: were the gmake comments meant for me? (at least gmake 4.2.1 is installed here, FWIW) 10:32
timotimo yes, they were 10:33
10:34 labster left
stmuk_ it was a typo for gnu tar 10:35
need coffee
parv stmuk_: ah, ok.
10:37 TEttinger left 10:42 wamba joined, carl_ left
timotimo that makes more sense, then 10:43
10:44 wamba left 10:46 bjz joined 10:51 zakharyas joined, patzim joined
patzim @SmokeMachine: Wrt. implementation of P6W: github.com/zostay/Smack 10:53
10:55 ufobat joined
timotimo a nonbreaking space would be a good idea in the title 11:00
it breaks just between Perl and 6 for me %)
samcv what is the best way to iterate through a list, possibly removing items from the list 11:02
11:02 astj joined
samcv doing for ^@array.elems will stop matching up when i remove elements, but i'm not sure how to remove an item from the array by doing `for @array` 11:02
timotimo i'd just map, or even grep 11:03
jnthn samcv: I'd probably @array .= grep: { ... } and hand back False to remove
arnsholt I think I'd do a grep removing elements, and then do the map
samcv well i need to do things to those elements 11:04
process those items
not _just_ remove them
jnthn You can in a grep?
samcv ok you think that would be suited fine?
seems a little weird to me
arnsholt If the processing is disjoint from the decision to remove or not, I'd filter first and then process
(Incidentally, it might be easier to help with a bit more context about what the problem you're solving is) 11:05
samcv they are not disjointed
11:05 parv left
timotimo if by "process, not just remove" you mean that you need to change individual values, i'd go with map and return Empty when you want to remove elements 11:06
samcv ok i have an Array of numbers, and I need to go through the array of numbers, and remove all the easy numbers and put it into another array
no changing values
11:06 pyrimidi_ left, pyrimidine joined
samcv trying to write something to pack a bitfield 11:06
timotimo m: my @a = ^10; @a .= map({ $_ %% 3 ?? $_ ** 2 !! Empty }); say @a
camelia rakudo-moar 3e28b1: OUTPUT«[0 9 36 81]␤»
11:06 astj left
jnthn Maybe .categorize? 11:07
(Which is for breaking a list into sub-lists matching conditions)
samcv hmm
that may work
11:08 devtom30 left
samcv let me try categorize 11:09
maybe that'll work for finding ones that are divisible by 8 cleanly, but then i will need to go through and find any complements
such as 3 and 7 = 8 for example
such as 3 and 7 = 8 for example 11:11
oops sent that twice
DrForr www.youtube.com/watch?v=vYmhapbBrig
11:12 rba_ left
samcv categorize is pretty cool 11:13
timotimo it's very good
samcv i'm trying to think what to look up for mathematical theory on this sort of thing 11:14
timotimo hm. bin packing?
samcv i mean if i take out all 8's and 16's for example, is there any chance of me losing some packing?
well just this sort of thing 11:15
like fitting certain numbers into another number
most efficiently
also fitting like 2d shapes into the smallest size 11:16
timotimo i would have assumed they fit like 1d shapes
samcv err
yeah
1d shapes but
it would be nice to know for sure that taking out all 8's and 16's will never prohibit one best possible packing 11:17
timotimo what do you fit stuff into?
samcv pretty sure it will? but idk. there must be some way to prove it
8 bits
timotimo ah
then you can transform the problem into another problem
arnsholt Oh. You have a series of field bit-widths, and want an algorithm to find the smallest number of bytes needed to pack them 11:18
samcv pretty sure divisible by 8 things will be fine
yes
more importantly the ordering
timotimo i.e. "fit all those 1, 2, 3, 4, 5, 6, 7 wide things into N 8-wide bins"
samcv of the smallest bytes
timotimo is the same as
"fit all those 1, 2, 3, 4, 5, 6, 7, 8, or 16 wide things into N + 10 8-wide bins"
i.e. just pretend those extra bins were never even there
and just sprinkle your 8- and 16-bit wide things at the end or beginning 11:19
samcv yeah, also want to know if i take out 4+4 and 3+6, will that also definitevely be the best? 11:20
i mean i know by themself, but hmm. i'm thinking i may need to sort odd and even
timotimo you mean 3+5?
samcv yes
timotimo i'd say so
there's this dynamic programming thing
where a perfect sub-solution can be part of a perfect super-solution 11:21
i.e. a solution that has one 8-bit bin filled up with 4 + 4 can be as perfect as that sub-solution for that single 8-bit bin
samcv yeah 11:22
i want to say that is true, but some part of me wants to look at a mathematical proof for it lol 11:23
SmokeMachine patzim: thanks! 11:24
11:26 xtreak left, gregf_ joined
jast if you have M bins of size N each and ignore elements of size i*N (for integer i), that's equivalent to packing the remaining elements into M-i bins 11:27
in other words, these elements don't really contribute anything interesting to the result
Ulti samcv: so long as the number you're removing isnt prime it shouldn't be a big deal?
because there's always some box that fits a different multiple 11:28
samcv many of them will be prime. but that is a good point
so should do all the non-prime ones first
then deal with the other ones last
11:28 ufobat left
Ulti also I wasnt following what you're actually up to other than the abstract thought problem :3 11:28
samcv oh
packing C bitfields in for our unicode database rewrite 11:29
for MVM
11:29 xtreak joined 11:30 rba_ joined
Ulti not sure im fully clear 11:30
is the problem the numbers are sparse?
so like you have lots of integers where you could fit the variety into 5bits but they use integers 8bits wide?
timotimo something like that, yeah 11:31
samcv i have bins of size 8, i need to fit a series of numbers, most of them 1-8 into the minimum nubber of bins
so need to do like 4+4 = 8, 2+2+2+2 = 8, 2+6= 8, but then also deal with all the odd numbers
11:32 rba_ left
moritz greedy approaches are typical 11:32
so you start with placing all the big chunks
11:32 rba_ joined
Ulti so in that example you can think of addition as a hash digest is there a reason something like that wouldnt work 11:33
moritz and when there's space left (for example when packing an 8-bin with 5, 3 remains), you do that recursively with all the chunks that are small enough to fit in the remaining space
samcv yeah
DrForr Keeping in mind that the optimal algorithm for bin packing is NP-hard, so from a practical standpoint heuristics are the only way to go.
samcv NP? 11:34
DrForr Non-Polynomial.
samcv ah
ok so just try and not find optimum and find close to optimum :)
DrForr Another way of saying the number of steps *rapidly* tends to infinity as the number of bins goes up.
arnsholt (Non-deterministic polynomial, actually)
samcv yeah 11:35
arnsholt But in practical terms, non-polynomial, yes =)
DrForr I sit corrected.
Ulti samcv: is this meant to be an on the fly thing or you want to do it once to find a good representation for an encoding?
timotimo only once
samcv yeah it's done only once
timotimo well, whenever the unicode db source data changes
Ulti ahh thats a lot easier so you have all the data
samcv yep
Ulti if its all internal though you dont really need to keep the representation consistent between the db changes?
DrForr This also sounds close to packing a C bit struct :) 11:36
timotimo correct, Ulti 11:37
samcv well it should generate the same thing every time
but it doesn't have to be consistent between the numbers changing
that it packs in
Ulti any reason you dont just enumerate all the things :S
surely thats easiest
samcv just try all combinations?
is that what you mean? tried it. like
over 9k years
Ulti yeah systematically enumerate whatever the space is
samcv was fine when it did like 8 different ones
but then past like 15.... 11:38
waited 20 minutes then killed the program
there's gonna be like 30+
could be 50
Ulti im assuming from recent things this is like all combos of combining emojii or something?
samcv well. like a lot of them are boolean, some are enums
hold on let me find something to show you 11:39
Ulti also has no one come up with an encoding for this already? thats kind of surprising 11:41
samcv Ulti, gist.github.com/samcv/1c2070ff0bb4...eld-h-L183 11:42
an encoding?
so if you know c, the :2 means that takes up only 2 bits width 11:43
and binary properties take up only 1 bit
and all the characters which have identical property values have rows in the bitfield deduplicated
11:44 rba_ left, sufrostico joined 11:45 sena_kun joined
samcv though to be honest what takes the most space in my proof of concept is storing the values of each cp -> bitfield row 11:45
like for every codepoint you have to store which bitfield row it is mapped to, for all the unicode codepoints, which is a huge number 11:46
i have it fixed for the name lookup at least, but need to implement the same for the bitfield rows
11:51 bjz left, bjz_ joined, xtreak left
jast example heuristic: pack all 8/16 bit fields. pack the largest unpacked field that will still fit in the currently open bin; leave bin incomplete if not possible. repeat until all fields are packed. the basic idea of this heuristic is the idea that large fields will not become easier to pack later, when there's less choice of what to pack them together with. it's easy to come up with examples for which it will perform poorly, but it should be okay if you have enough 11:54
11:54 ufobat joined
timotimo add a phase of wiggling stuff about randomly at the end in the hopes of finding something better :P 11:55
"okay, add an hour of random experimentation here, see if we find anything better"
samcv lol 11:57
jast that's actually not that easy, you have to generate mutations that fulfill the constraints
timotimo nah, just check the constraints after generating each mutation 11:58
jast virtually all of them will have to be eliminated, though 11:59
alternatively you could backtrack but chances are that the bad packings don't happen near the end of that process
timotimo *shrug* ;)
jast well I guess you could turn it into a search problem, change the heuristic into a score function and try the "best" packing first 12:01
then apply dynamic pruning rules and other fancy things like that
and down the rabbit hole we go 12:02
DrForr Heuristics FTW. 12:04
jast tree search is about optimizing brute force using heuristics and formal reasoning, it's pretty neat but easy to get wrong 12:05
12:05 rurban joined, rurban left 12:06 astj joined
samcv i gotta go to bed now guys 12:09
thanks for all your help :)
SmokeMachine I've read constraints and heuristic and would propose to use modules.perl6.org/#q=ProblemSolver but that's nothing related... :( 12:12
12:14 parv joined 12:15 gregf_ left
SmokeMachine S/propose/suggest/ 12:16
12:18 parv left
masak TimToady: would you also be fine with BEGIN-in-quasi firing only at expand time, even if it could fire sooner? that is, which consistency do we aim for in that case? 12:20
12:25 xtreak joined, xtreak left 12:27 rba_ joined 12:28 rba_ left 12:32 astj left 12:33 astj joined 12:34 gregf__ joined 12:36 lukaramu joined 12:38 bjz joined 12:39 bjz_ left, andrzejku left 12:44 agentzh joined 12:48 agentzh left 12:51 rba_ joined 13:03 astj left
lizmat clickbaits p6weekly.wordpress.com/2017/01/30/...e-mondays/ 13:15
13:18 cdg joined
brokenchicken clicks 13:19
m: "\c[Grinning cat Face]".say 13:20
camelia rakudo-moar f22170: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Unrecognized character name [Grinning cat Face]␤at <tmp>:1␤------> 3"\c[Grinning cat Face7⏏5]".say␤»
13:20 cdg left, cdg joined
brokenchicken star: class Foo::Bar {}; my Foo::Bar $a .= new; dd $a 13:21
camelia star-m 2016.10: OUTPUT«Foo::Bar $a = Foo::Bar.new␤»
brokenchicken star: class Foo::Bar {}; class { my Foo::Bar $.a .= new }.new
camelia ( no output )
brokenchicken star: class Foo::Bar {}; class { my Foo::Bar $.a .= new }.new.qa
camelia star-m 2016.10: OUTPUT«No such method 'qa' for invocant of type '<anon|62232192>'␤ in block <unit> at <tmp> line 1␤␤»
brokenchicken star: class Foo::Bar {}; class { my Foo::Bar $.a .= new }.new.a
camelia ( no output )
brokenchicken star: class Foo::Bar {}; class { has Foo::Bar $.a .= new }.new.a 13:23
camelia star-m 2016.10: OUTPUT«===SORRY!===␤Could not locate compile-time value for symbol Foo::Bar␤»
brokenchicken ah
13:34 wamba joined 13:35 mr-foobar joined 13:37 wamba left 13:41 ChoHag left 13:47 geekosaur left 13:48 geekosaur joined, khw left 13:49 astj joined 13:51 khw joined, b^_^d left 13:52 d^_^b joined, d^_^b left, d^_^b joined 13:54 astj left 14:00 KDr2 joined 14:16 skids joined
SmokeMachine I remember that! 14:17
brokenchicken Well, it *was* just last week :P
stmuk_ anyone know if the ORA Perl 6 for CS book is available in any sort of ebook draft form? 14:18
perlpilot shop.oreilly.com/product/0636920065883.do talks about "Early Release Ebook" 14:20
14:20 gregf__ left
perlpilot unless you also mean "free" 14:20
stmuk_ yes I noticed that! maybe needs a safari (if it's caled that now) signup 14:21
14:25 mienaikage joined
moritz read the first 120~ pages of an early version as a technical review 14:25
like, before O'Reilly was involved
lizmat stmuk_: I understand that the first 7 chapters will become available as an early version ebook within days 14:27
fwiw, I've proofread the entire book 14:28
brokenchicken Is it good? 14:30
lizmat is slighlt biased
brokenchicken And is it for total programming beginners or all levels? 14:31
moritz it's a very introductory book
brokenchicken Hmm.
stmuk_ its a port of a similarly named python book I think?
moritz it's about leraning to program, and Perl 6 takes a second place after that
stmuk_: yes
stmuk_ actually it does look a bit Noddy 14:32
lizmat chapter names are:
1. The way of the program
perlpilot btw, I'm encouraged by all of the Perl 6 books out there. It gives me hope that many people will use Perl 6.
DrForr This explains a conversation I had a few weeks ago.
lizmat 2.
Variables, expressions and statements
3. Functions and subroutines 14:33
4. Conditionals and recursion
stmuk_ python one is interactivepython.org/courselib/sta...index.html
lizmat 5. Fruitful subroutines 14:34
6. Iteration
perlpilot "fruitful subroutines"? That's an interesting title.
lizmat 7. Strings
perlpilot seems a bit out of place with the other, more boring titles
moritz vegetableful subroutines would be a bit weird 14:35
lizmat m: sub a { say "Hello World" } # not fruitful
camelia ( no output )
brokenchicken perlpilot: yeah, especially written by people who aren't regulars in this channel :)
lizmat m: sub a { "Hello World" } # fruitful
camelia ( no output )
brokenchicken I also notice this channel's average usercount has grown by ~15 people since December 14:36
lizmat regulars here would be people who know how to program :-)
moritz functions that return a value, rather than having a side effect, I guess
stmuk_ www.openbookproject.net/thinkcs/pyt.../ch05.html 14:37
yes
14:39 jonas2 left, ufobat left 14:41 mienaikage left, cale2 joined, user9 left
stmuk_ there are a bunch of "learn $foo the hard way" books as well 14:41
14:42 user9 joined 14:45 gregf_ joined
SmokeMachine lizmat: I'm not sure... I'm connected here all the time... 14:46
lizmat :-) 14:47
14:51 Dandel_ joined
Dandel_ skids, you around? 14:53
14:54 mienaikage joined
skids o/ 14:54
brokenchicken \o 14:55
Dandel_ I was wanting to ask ya a few things about an old project that you where working on a few years back.
cale2 Think Perl 6 book says it delves into declarative programming and grammars even
yoleaux 10:21Z <timotimo> cale2: have a look at the wig module :)
cale2 wig module? huh? 14:57
brokenchicken buggable: eco wig
buggable brokenchicken, wig 'where is grep: Make grep prettier, put a wig on it!': github.com/0racle/p6-wig
lizmat cale2: chap 14: regexes and grammars 14:59
cale2: chap 15: functional programming 15:00
skids Dandel_: which one? 15:01
Dandel_ GGI.
skids Ahah. Yeah that was some time ago :-)
Dandel_ all I'm doing is trying to get in contact with the contributors so i can map things into git. 15:02
skids Sure, what do you need?
15:03 zakharyas left
Dandel_ Just the current contact information you want to use for git. 15:03
brokenchicken What's the name of <>? As in $blah<>? Also, is its purpose just to decontarize or is that a sideeffect?
15:03 zakharyas joined
skids skids is my git uid as well. 15:03
Dandel_ I figured.
skids Or ... oh you mean email? 15:04
Dandel_ ya
I was wanting to get current email addresses... I plan on updating the tree once I get the map to also mention which old devs do not wish to be contacted again :) 15:05
15:08 zakharyas left
DrForr On boston-pm there's a posting about rakudo.org/2017/01/30/announce-raku...e-2017-01/ being blocked by WBRS with 'Reason: BLOCK-MALWARE', 'Threat Type: othermalware' 15:08
brokenchicken :( 15:09
15:09 pyrimidine left 15:10 pyrimidine joined
brokenchicken Um. WBRS being what? 15:10
DrForr WBRS seems to be Cisco firewall filtering, not sure what else. Where should I forward the message?
brokenchicken WBRS 100.1 FM in google has "this site may be hacked" message under it
DrForr "Web Reputation Filters" appears in the message, there's really nothing else useful that I see in the message. 15:11
It's not appropriate for rakudo-bugs, where should it get sent?
mspo wordpress can get quietly hacked 15:12
brokenchicken DrForr: well, does it even need to be sent?
DrForr (I suspect there's no real solution because it's bayesian.)
brokenchicken DrForr: wtf is WBRS?
mspo rakudo.org/how-to-get-rakudo/ doesn't have checksums and doesn't use https 15:13
DrForr brokenchicken: The message doesn't mention what the filtering system is. I believe it's Cisco blocking, but I don't know. 15:14
brokenchicken Yeah, it's pretty frustrating to have such a simple problem unsolved for months.
DrForr: but what's WBRS?
15:14 pyrimidine left
mspo www.cisco.com/c/en/us/support/docs/...sa-00.html 15:14
15:15 rurban joined
DrForr brokenchicken: You can repeat your question as often as you like. The message does *not* state what it is, all I can do is guess that it's a Cisco filter. 15:15
mspo could be more related to the hosting IP and not specifically rakudo.org
DrForr All I'm asking is "Is there a security@ or admin@ email address that I should forward this to?"
mspo DrForr: is www.pmichaud.com/ also blocked? 15:17
15:17 wamba joined
mspo or www.pmwiki.org/ (wikis are also often quietly hacked) 15:17
DrForr mspo: brokenchicken: I did not make this discovery, I have no more information than what I passed along. This is a report to boston.pm that I'm mentioning, I have no more information than that. 15:18
brokenchicken DrForr: but what's WBRS?
:)
mspo DrForr: I would send it to pmichaud guy 15:19
brokenchicken DrForr: sorry, I thought WBRS was some organization at first :)
mspo and also consider ssl and checksums on rakudo :)
brokenchicken DrForr: and mention that we need Wordpress upgrade and HTTPs
He's a bunch of bot messages.
mspo or maybe some kind of perl6-based blog
DrForr I know Patrick, I don't have his email to hand. What is it?
15:20 cyphase left
brokenchicken Rakudo releases are signed, though I see not R*.. 15:20
15:21 wamba left 15:23 wamba joined
DrForr (privately if you like) It's not a download of a specific file, but it could very well simply be blocking the site because of a download link that's not over https, yes. 15:25
15:25 as__ left
brokenchicken [rakudo@power ~]$ grep wp-login access-logs/rakudoperl.org | wc -l 15:25
15:25 cyphase joined
brokenchicken 384 15:25
[rakudo@power ~]$ grep downloads access-logs/rakudoperl.org | wc -l
214
Wonder if our WP install is up to date... I know php is waaaay out of date :( 15:26
moritz it's not 15:27
brokenchicken yeah :(
15:27 CIAvash left
brokenchicken And more bots try the wp-login page than people downloading releases 15:28
DrForr I've dealt with that on tau-station before, WP seems to be fairly stable but I'd still make sure to do a full backup of DB and web directory, because it may not come back. And breaking templates is *always* a concern.
brokenchicken wow "Akismet has protected your site from 872,376 spam comments already." :)
15:31 dugword joined
DrForr Also, I don't see pmichaud's email address in the IRC logs, anyone willing to send it my way privately? 15:33
brokenchicken There's one listed on github profile: github.com/pmichaud 15:36
15:40 AlexDaniel joined
DrForr sorry if I was a little churlish. News from home is not fun. 15:42
AlexDaniel heh… somebody just wrote “poophole optimizer”
yoleaux 10:14Z <samcv> AlexDaniel: East_Asian_Width property has just been merged into MVM
mspo I use http simple auth on wp-login to stop a *ton* of bot traffic 15:44
but it didn't stop a sql injection (or something, i'm not sure really) from hitting one of my sites *badly* over this weekend
15:45 wamba left
cale2 I like this: github.com/0racle/p6-wig 15:45
It feels like it belongs in the language with the "where" guards on parameters and subsets as well 15:46
"where" in subsets and params is basically the same thing as grep anyway. `.where` shouldn't even be a method. It should just be re-used as a keyword for most types 15:47
brokenchicken 0.o
AlexDaniel samcv: great! Amazing! Now I have to find why I needed that… 15:49
15:49 dugword left 15:50 g4 left
[Coke] u { / ROMAN NUMERAL /} 15:50
AlexDaniel u: { / ROMAN NUMERAL / } 15:51
hm, maybe not that
u: { .uniname ~~ / ROMAN NUMERAL / }
unicodable6 AlexDaniel, gist.github.com/047dfa7d80ea019d36...e4a870ef23
AlexDaniel, Oops, something went wrong!
AlexDaniel hey why so slow
[Coke] u: { .uniname ~~ / 'ROMAN NUMERAL' /} 15:52
unicodable6 [Coke], U+2160 ROMAN NUMERAL ONE [Nl] (Ⅰ)
[Coke], U+2161 ROMAN NUMERAL TWO [Nl] (Ⅱ)
[Coke], gist.github.com/5476fccd59fe9f5214...62c7801d28
brokenchicken prints a billion of warnings, I'm guessing.
Oh, just one. 15:53
AlexDaniel hm, but still it timed out for some reason
ah, right 15:54
u: ROMAN NUMERAL
unicodable6 AlexDaniel, U+2160 ROMAN NUMERAL ONE [Nl] (Ⅰ)
AlexDaniel, U+2161 ROMAN NUMERAL TWO [Nl] (Ⅱ)
AlexDaniel, gist.github.com/df0db785c43d5745c1...8e7b4891fb
AlexDaniel no need for a code block
the one without a code block is optimized to do the whole thing in about two seconds 15:55
the other one may be pretty slow, depends on what you write
Geth oc: df9d1155b8 | (Zoffix Znet)++ | doc/Language/operators.pod6
Formatting fix: Markdown -> POD
 15:56
15:58 geekosaur left, matt__ joined 16:00 geekosaur joined
perigrin w/ 48 16:01
brokenchicken m: say quietly ($ = 42) ~~ S/4/2/; say $/; say $_
camelia rakudo-moar f94cb2: OUTPUT«Potential difficulties:␤ Smartmatch with S/// is not useful. You can use given instead: S/// given $foo␤ at <tmp>:1␤ ------> 3say quietly ($ = 42) ~~ 7⏏5S/4/2/; say $/; say $_␤False␤22␤(Any)␤»
brokenchicken You can't even quiet this warning? Laaame
perlpilot It's compile-time. quietly only works at run-time. 16:03
AlexDaniel I'm not even sure why does it have to run at all
16:04 wamba joined, KDr2 left
AlexDaniel I mean, what's the point of proceeding when there's that? 16:04
brokenchicken AlexDaniel: are you talking about my warning? 16:05
AlexDaniel yea
brokenchicken AlexDaniel: then I don't follow. The warning is a lie.
You can get the result from $/
AlexDaniel but why not use given instead? 16:06
brokenchicken doesn't understand the question 16:07
perlpilot doesn't understand either
AlexDaniel m: say S/4/2/ given ($ = 42); say $/; say $_ 16:08
camelia rakudo-moar f94cb2: OUTPUT«22␤22␤(Any)␤»
AlexDaniel and look, you're not getting a useless False
the error message says that smartmatching with S/// is not useful. What's useful about getting False here? 16:09
Geth oc: 08cfdd8e8a | (Zoffix Znet)++ | doc/Language/operators.pod6
Minor clarifications for S/// / s///

  - Mention the operators set $/
  - Reword the warning explanation: saying there's no way to get the result is a lie, as it can be obtained from $/.
   I'm not even sure why the warning exists at all. Feels the original design that set result into $/ did not
   prescribe a warning for S///
brokenchicken AlexDaniel: but it's not useless. And your given reverses the arguments, making it read backwards and you did it just to avoid the warning that lies :) 16:10
AlexDaniel /o\
the result of ~~ is useless
brokenchicken No, it tells you whether any changes were done
16:10 pyrimidine joined
brokenchicken But you don't need to use it 16:10
AlexDaniel brokenchicken: so what's False in this case?
brokenchicken $x ~~ s/a/b/; say $x 16:11
$x ~~ S/a/b/; say $/
I don't see why you think one version is useless and needs to be rewritten with `given` while the other is fine
16:11 bwisti joined
brokenchicken AlexDaniel: return = "is result after changes the same as the original string" 16:11
It basically ends up being Str ~~ Str 16:13
AlexDaniel which is the opposite of what regexes usually do
but at least I see your point now
brokenchicken hehe
But my jorney to this point started with this: 16:14
AlexDaniel m: given $ = 42 { S/4/2/ }; say $/; say $_
camelia rakudo-moar f94cb2: OUTPUT«22␤(Any)␤»
brokenchicken m: say <a1 a2 a3>.map({S/huh//});
camelia rakudo-moar f94cb2: OUTPUT«(a1 a2 a3)␤»
brokenchicken m: say eager <a1 a2 a3>.map({S/huh//});
camelia rakudo-moar f94cb2: OUTPUT«(a3 a3 a3)␤»
brokenchicken Note, this is NOT a bug :)
Or
m: my $fun = <a1 a2 a3>.map({S/huh//}); say $fun[^1]; say $fun[^2]; say $fun[^3]; 16:15
camelia rakudo-moar f94cb2: OUTPUT«(a1)␤(a2 a2)␤(a3 a3 a3)␤»
brokenchicken also not a bug :)
AlexDaniel why isn't it a bug by the way
ah
I see…
brokenchicken And pondering wtf S/// returns $/ anyway. And I came to conclusion original design intended for $/ to be obtained the result from, and no warning :) 16:16
Hm. OTOH it still doesn't need to return it
16:16 ChoHag joined
brokenchicken And I was gonna say worth seeing if spectests explode, but now I remember we got like just a couple of tests for S/// 16:18
.seen FROGGS
yoleaux I saw FROGGS 17 Jan 2017 21:11Z in #perl6: <FROGGS> dataf3l: to this: hg.dyncall.org/pub/dyncall/dyncall/...allf.c#l91
AlexDaniel c: 2015.07 say ($ = 42) ~~ S/4/2/; say $/; say $_ 16:19
committable6 AlexDaniel, ¦«2015.07»: False␤22␤(Any)
AlexDaniel ok let's see what was the motivation behind it exactly
bisect: old=2015.07 say ($ = 42) ~~ S/4/2/; say $/; say $_
bisectable6 AlexDaniel, Bisecting by output (old=2015.07 new=f94cb21) because on both starting points the exit code is 0
AlexDaniel, bisect log: gist.github.com/3477b17991231dcdfe...afd0dad75e
AlexDaniel, (2015-10-13) github.com/rakudo/rakudo/commit/f5...427990bada
AlexDaniel ok that commit message isn't super helpful 16:20
brokenchicken hehe
Well, kinda is. It proves my point the warning wasn't a part of original design :)
perlpilot yeah, you were being far too optimistic
AlexDaniel now I remember, the message was a bit different initially 16:21
and then was changed because of my complaints I think
bisect: old=f54ff833 say ($ = 42) ~~ S/4/2/; say $/; say $_
bisectable6 AlexDaniel, Bisecting by output (old=f54ff83 new=f94cb21) because on both starting points the exit code is 0
AlexDaniel, bisect log: gist.github.com/c5f49e9a3da6e94fda...823990b2a7
AlexDaniel, (2015-12-04) github.com/rakudo/rakudo/commit/58...a3ed6723f3
brokenchicken bisect: old=2014.01 new=2015.12 S/4/2/.VAR.name.say given "x"
bisectable6 brokenchicken, Bisecting by exit code (old=2014.01 new=2015.12). Old exit code: 1
AlexDaniel no… no… it only goes to 2015.07 16:22
brokenchicken oh, oops :)
AlexDaniel it will give you the best answer it can
perlpilot Getting rid of that worry will make compilation a tiny bit faster ;) 16:23
brokenchicken looks for the ABORT button
abraxxa great, libgd2 2.2.1 on Ubuntu doesn't install gdlib-config any more which is required by GD
AlexDaniel brokenchicken: it's fine, it should make it :)
ilmari abraxxa: I'd expect that to be in libgd-dev 16:24
abraxxa ilmari: it was
16:24 koki1 joined
abraxxa ilmari: search for gdlib-config: changelogs.ubuntu.com/changelogs/po.../changelog 16:24
2.2.1-1: Don't install obsolete gdlib-config
ilmari abraxxa: oh, when you said libgd2 I thought you meant libgd2, not some vaguely related other package 16:25
AlexDaniel
.oO( come on!! )
bisectable6 brokenchicken, bisect log: gist.github.com/7d3988cf257d6eed63...9434dfb296
brokenchicken, There are 209 candidates for the first “new” revision. See log for more details
abraxxa what vaguely related other package?
brokenchicken wooo
AlexDaniel \o/
bisect: old=587f700e9 say ($ = 42) ~~ S/4/2/; say $/; say $_ 16:26
bisectable6 AlexDaniel, Bisecting by output (old=587f700 new=f94cb21) because on both starting points the exit code is 0
AlexDaniel, bisect log: gist.github.com/5f1c53cd786e200556...0458940705
AlexDaniel, (2016-01-05) github.com/rakudo/rakudo/commit/ac...ac9584db28
ilmari abraxxa: libgd-dev
AlexDaniel haha… “Zoffix++, AlexDaniel++”
brokenchicken AlexDaniel: so I can still run my last query if I change 2015.12 to 2015.07?
AlexDaniel brokenchicken: you mean 2014.01 to 2015.07? 16:27
brokenchicken mhm
abraxxa ilmari: moved to #perl-help as I've initially mixed up the channels
AlexDaniel commit: all S/4/2/.VAR.name.say given "x"
what are you looking for exactly
committable6 AlexDaniel, gist.github.com/7280de5b08abec1b01...a5d603000e 16:28
brokenchicken To see whether when S/// was implemented it was returning $/
Or more precisely, explanation for why it returns $/ :)
AlexDaniel brokenchicken: looking at committable output, it was implemented in 2015.06 and that's the initial behavior 16:29
brokenchicken bisect: old=2015.05 new=2015.06 S/4/2/.VAR.name.say given "x"
bisectable6 brokenchicken, Bisecting by exit code (old=2015.05 new=2015.06). Old exit code: 1
AlexDaniel that will give you the same 209 candidates, as it has nothing in between
bisectable6 brokenchicken, bisect log: gist.github.com/1eb6ff11bc6aba4b51...91b61d3e5e
brokenchicken, There are 209 candidates for the first “new” revision. See log for more details
brokenchicken AlexDaniel: ah, so pre-Christmas only release commits have builds? 16:30
AlexDaniel brokenchicken: 2015.07~HEAD and all tags are built
so you can use 2014.02 but not 2014.02^ :)
brokenchicken Well, there's always New and Improved™ GitHub blame :) 16:31
16:31 wamba left, wamba joined
AlexDaniel was it changed? 16:31
“view blame prior to this change” 16:32
finally
brokenchicken Yeah, recently. They made it much easier to view previous commits
perlpilot brokenchicken: S/// returns $/ probably because s/// returns $/. The real problem is ... why isn't $/ a Match with S/// and why does it contain the entire string intead of just the matching part.
16:33 dugword joined
brokenchicken perlpilot: because S/// returns the resultant string. 16:33
perlpilot Sure, but that has nothing to do with $/ 16:34
16:34 zakharyas joined
brokenchicken I see what you're saying. 16:34
perlpilot m: my $foo = "blah"; $foo ~~ s/a/u/; say $/; 16:35
camelia rakudo-moar f8b346: OUTPUT«｢a｣␤»
perlpilot I'd expect the S/// variant to behave identically
(I mean with respect to $/)
brokenchicken m: my $foo = "blaaaaaah"; $foo ~~ s:g/a/u/; say $/;
camelia rakudo-moar f8b346: OUTPUT«(｢a｣ ｢a｣ ｢a｣ ｢a｣ ｢a｣ ｢a｣)␤»
brokenchicken Agreed.
AlexDaniel yea 16:36
brokenchicken And that'll fix the .map({S///}) thing too
brokenchicken will fix it after lunch
And the warning on S/// will make sense :D 16:40
on ~~ S///
perlpilot brokenchicken++
brokenchicken perlpilot++ 16:42
16:46 agentzh joined 16:48 domidumont left 16:49 pyrimidine left, pyrimidine joined 16:50 agentzh left, dugword left 16:51 lukaramu left 16:54 pyrimidine left 16:56 cdg left 16:58 Sound joined, zakharyas left 17:01 zakharyas joined 17:02 noro joined
TimToady masak: I thought that was the consensus I was already agreeing to 17:03
17:07 wamba left, ggoebel left 17:10 girafe joined 17:13 wizard_ joined
perlpilot brokenchicken: Here's a rough hack at fixing S/// gist.github.com/perlpilot/bb541544...64a6eb1b27 17:19
17:19 matt__ left
perlpilot brokenchicken: I'd commit it, but I don't have the time to really test it out. 17:20
17:20 matt___ joined
brokenchicken awww, you took all the fun out of it :) I was *just* booting up my VM :) 17:20
17:20 ggoebel joined
brokenchicken Never heard of QAST::Node tho, so heh :) 17:21
What's it do? The my $Sname := QAST::Node.unique("__Sname"); part? 17:22
perlpilot makes a unique identifier.
each call you'll get something like __Sname_1 __Sname_2 __Sname_3 etc.
anyway, like I said, it's a rough hack. You'll do better I'm sure. 17:23
:)
brokenchicken Now you put me under pressure to do better! :) 17:33
17:35 noro left 17:36 rba_ left
Geth oc: 0abaf2508f | (Christopher Bottoms)++ | doc/Language/5to6-perlfunc.pod6
Mention "chomp in-place"

Also touched up definition to ensure it is clear that the target itself isn't returned.		 17:36
perlpilot brokenchicken: I figure you've been touching the code lately and I can't remember the last time I did anything more than update a string (though git tells me it's been 17:40
a while)
brokenchicken Yeah, but I barely know what I'm doing :P 17:41
perlpilot fair enough :) 17:42
17:43 zakharyas left 17:45 domidumont joined 17:46 wamba joined 17:50 labster joined 17:55 dakkar left
Geth oc/master: 4 commits pushed by Altai-man++ 17:59
sena_kun Hmm, shouldn't it show the original contributor, not the person who did a merging? 18:00
18:00 labster left
brokenchicken What do WANTED/wanted subs do in Perl6/Actions? 18:00
TimToady tell things they're not in sink context 18:01
brokenchicken Thanks. 18:02
TimToady it's basically an interleaved top-down 2nd pass
interleaved with the 1st bottom-up pass, that is 18:03
(to the extent that the 1st pass is bottom-up, which it is from the standpoint of Actions, if not always from the parser's point of view)
brokenchicken sena_kun: it reports the pusher, which is accurate. On 1-2 commit merges, it'd report individual authors. 18:08
sena_kun brokenchicken, oh, okay. 18:09
brokenchicken sena_kun: this is the whole webhook: gist.github.com/zoffixznet/4e9304c...c38b821219 The pusher is given as a separate hash, and I'm unsure how reliable it'd be to pick through individual comits to find authors
sena_kun: hmm... it could @all-authors.unique in the message maybe? Abridging them if they're too long 18:10
This is the line in code github.com/perl6/geth/blob/master/...ub.pm6#L92 and that object is populated here: github.com/perl6/geth/blob/master/...6#L96-L114 18:11
Eh, looks simple enough :) I'll do it 18:12
sena_kun brokenchicken, it is rare(afaik) to have a crowd of people in one PR. I just thought that the actual contributor did more than pusher in this case.
brokenchicken Agreed :)
sena_kun brokenchicken, ah, I wanted to try it.
After your proper explanation. 18:13
18:15 Geth left, Geth joined, ChanServ sets mode: +v Geth 18:16 eroux_ joined, Geth left 18:17 Geth joined, ChanServ sets mode: +v Geth
brokenchicken
.oO( need to add tests to it... )
 18:17
Geth oc/master: 4 commits pushed by (Antonio Quinonez)++, (Antonio Quinonez)++, (Antonio Quinonez)++, Altai-man++
brokenchicken oops, forgot unique :P
18:17 Geth left 18:18 matt___ left, Geth joined, ChanServ sets mode: +v Geth 18:19 Geth left 18:20 Geth joined, ChanServ sets mode: +v Geth, eroux_ left
Geth oc/master: 4 commits pushed by (Antonio Quinonez)++, Altai-man++ 18:20
brokenchicken yey
sena_kun It was quick! brokenchicken++
18:21 pyrimidine joined, agentzh joined, eroux_ joined 18:24 eroux_ left
brokenchicken My OCD senses are disturbed by that extra space before Geth's stuff :} 18:26
18:26 _eroux_ joined
brokenchicken AlexDaniel: how about sticking some fancy pants Unicode char there instead? :) 18:26
18:26 lolo78 left
AlexDaniel u: zero width space 18:26
unicodable6 AlexDaniel, U+200B ZERO WIDTH SPACE [Cf] (​)
AlexDaniel, U+FEFF ZERO WIDTH NO-BREAK SPACE [Cf] ()
brokenchicken heh 18:27
u: box
unicodable6 brokenchicken, U+2327 X IN A RECTANGLE BOX [So] (⌧)
brokenchicken, U+237D SHOULDERED OPEN BOX [So] (⍽)
brokenchicken, gist.github.com/aa021254726f35a59e...520b7b328d
AlexDaniel u: bullet
unicodable6 AlexDaniel, U+2022 BULLET [Po] (•)
AlexDaniel, U+2023 TRIANGULAR BULLET [Po] (‣)
AlexDaniel, gist.github.com/63fd72e1328055bd1b...d05811ba8c
brokenchicken geth: f3eec4957b | (Zoffix Znet)++ | 2 files
Report all authors when we don't print full commits
 18:29
AlexDaniel maybe ¦ ?
brokenchicken u: ¦
unicodable6 brokenchicken, U+00A6 BROKEN BAR [So] (¦)
AlexDaniel that's even latin-1 18:30
so should render like everywhere
18:30 telex left
brokenchicken OK 18:31
18:32 telex joined, Geth left 18:33 cdg joined, Geth joined, ChanServ sets mode: +v Geth 18:38 patzim left 18:39 FROGGS joined, pyrimidine left 18:40 pyrimidine joined
brokenchicken What does QAST::Op.new( :op('p6scalarfromdesc'), QAST::Op.new( :op('null') ) ) do? That op is mentioned without any description in rakudo/docs 18:42
Makes a scalar from descriptor? What's a descriptor? 18:43
18:43 labster joined
El_Che brokenchicken: on a perl5 channel? you're in a good mood :) 18:43
brokenchicken brokenchicken: hm? 18:44
What?
El_Che brokenchicken: didn't you post a "screw you guys, I'm going home" some time ago?
(it good to see you there, it's not a critique) 18:45
brokenchicken El_Che: where?
El_Che #perl
brokenchicken I'm not there
El_Che ah ok, someone with a similar nick
18:45 pyrimidine left
El_Che not making sense, then 18:45
:)
El_Che ignores El_Che 18:46
18:46 domidumont left
brokenchicken ... and I thought *I* was sleep deprived :P 18:46
El_Che sadly, that part is true
18:47 avuserow left 18:52 avuserow joined 19:01 andrzejku joined, cdg_ joined 19:04 cdg left 19:05 agentzh left 19:07 agentzh joined
brokenchicken Who's Github's ronaldxs on IRC? It's m-nom or something like that 19:09
don't seem to b ehere 19:10
19:11 darutoko left 19:15 Sound left 19:19 zakharyas joined, domidumont joined 19:22 rurban1 joined 19:25 rurban left 19:26 _eroux_ is now known as eroux 19:27 cibs left 19:29 cibs joined, mr_ron joined
mr_ron ronaldxs - mr_ron 19:30
yoleaux 24 Jan 2017 20:47Z <brokenchicken> mr_ron: now that I remember more specific test files, a lot of them depend on the test file located in t/spec. This is used for auxiliary files locations for some test files, like module loading ones, and the ones that include t/spec/packages
brokenchicken Ah, right :) 19:31
But I no longer have a question. Was just wondering what @already_fudged was about, but read the code :)
19:31 pyrimidine joined 19:34 rurban joined
brokenchicken mr_ron: FWIW we also have #perl6-dev 19:36
19:37 wizard_ left 19:38 rurban1 left 19:40 labster left 19:41 TEttinger joined
brokenchicken What's QAST::Op.new( :op('locallifetime'), ? Don't see it in docs 19:43
19:44 zakharyas left, agentzh left 19:45 rurban left
brokenchicken *crickets* :) 19:46
Maybe I should try learning this stuff from the other end... 19:47
MoarVM -> nqp -> Perl 6 19:48
19:49 pyrimidine left 19:50 pyrimidine joined
brokenchicken wonder if you can write programs with just MoarVM, no other stuff 19:50
perigrin if I understood right MoarVM was a very very thin layer beneath npq 19:51
perlpilot You can make fire with two sticks, but why when there's other incendiary devices to help you? :) 19:52
brokenchicken perlpilot: to discover how fire is made :) 19:54
19:54 agentzh joined, pyrimidine left
jdv79 i have this in a role: "has Hash $!state; method !state { return-rw $!state }" 19:55
is there anthing wrong with that?
brokenchicken jdv79: well, why do you think somethiong's wrong with it? 19:56
jdv79 eventually i get this error from a consuming class: P6opaque: no such attribute '$!state' in type Agent::AtomAgent when trying to get a value
because that
brokenchicken jdv79: is it the same for method !state is rw { $!state } ?
jdv79 having trouble golfing it
takes a while to manifest. i'll try that
19:56 cdg joined
brokenchicken 0.o wow 19:56
jdv79 nah. same error. 19:59
19:59 cdg_ left
brokenchicken has no idea 19:59
jdv79: "takes a while to manifest". So it works fine some of the time? Concurrent code? 20:00
jdv79 its concurrent, yes.
well, the role is composed into half a dozen classes and its not always the same ones that bork 20:01
forgot to try with max threads set to 1 y'day 20:02
20:02 cdg_ joined
brokenchicken Does that even work? It'll prolly just deadlock. 20:03
20:04 astj joined
jdv79 its chugging along just fine. 20:04
brokenchicken weird, I had deadlocks in roast when max threads was too low
jdv79 oh, a failed json parse segfaulted. well, that's at least probably not related
brokenchicken m: role Foo { has Hash $!state; method !state { return-rw $!state } }; for ^100000 .hyper: :batch { $ = class :: does Foo { method x { self!state } }.new.x }; say "nada" 20:05
jdv79 I'm running something like two dozen start blocks at once
camelia rakudo-moar 015231: OUTPUT«(signal XCPU)»
jdv79 weird
brokenchicken maybe to do with await
m: role Foo { has Hash $!state; method !state { return-rw $!state } }; for ^10000 .hyper: :batch { $ = class :: does Foo { method x { self!state } }.new.x }; say "nada" 20:06
20:06 cdg left
camelia rakudo-moar 015231: OUTPUT«(timeout)» 20:06
brokenchicken m: for ^20 { await start await start sleep 2 } 20:08
camelia rakudo-moar 015231: OUTPUT«(timeout)»
jdv79 nope, even with 1 thread it can error
cmon:(
brokenchicken m: use v6.d.PREVIEW; for ^20 { await start await start sleep 2 }
hm.. weird. I thought that'd work
camelia rakudo-moar 015231: OUTPUT«(timeout)»
jdv79 needs food
lunch & 20:09
20:09 astj left
brokenchicken m: use v6.d.PREVIEW; for ^5 { await start await start sleep 2 } 20:09
camelia ( no output )
brokenchicken m: use v6.d.PREVIEW; for ^5 { await start await start sleep 2 }; say now - INIT now;
camelia rakudo-moar 015231: OUTPUT«10.01377108␤» 20:10
brokenchicken Oh, I'm dump
m: await do for ^20 { start await start sleep 2 }; say now - INIT now
camelia rakudo-moar 015231: OUTPUT«8.008532624␤»
brokenchicken wonder how it ends up with 8s 20:11
m: await do for ^20 { start await start sleep 1 }; say now - INIT now
camelia rakudo-moar 015231: OUTPUT«3.007262␤»
brokenchicken m: await do for ^20 { start await start sleep 3 }; say now - INIT now
camelia rakudo-moar 015231: OUTPUT«9.00744710␤»
brokenchicken m: use v6.d.PREVIEW; await do for ^20 { start await start sleep 3 }; say now - INIT now
camelia rakudo-moar 015231: OUTPUT«6.009234␤»
20:15 pyrimidine joined
brokenchicken m: use v6.d.PREVIEW; await do for ^100 { start await start sleep .01 }; say now - INIT now 20:16
camelia rakudo-moar 015231: OUTPUT«0.1125727␤»
brokenchicken aha!
m: await do for ^100 { start await start sleep .01 }; say now - INIT now
there's the goodness of 6.d :P
camelia rakudo-moar 015231: OUTPUT«(timeout)»
20:20 domidumont left, ocbtec left 20:21 Tonik joined
RabidGravy what's the scope visibility of " use v6.d.PREVIEW;" ? CompUnit or global? 20:24
brokenchicken compunit 20:25
you can have one file run 6.c and another 6.d.PREVIEW
m: use lib </tmp/foo>; use c; use d; 20:26
camelia rakudo-moar 015231: OUTPUT«SETTING::src/core/asyncops.pm␤SETTING::src/core.d/await.pm␤»
brokenchicken :)
RabidGravy so if I have a module with 6.d.PREVIEW it doesn't matter what the code that uses ir does?
brokenchicken Right
RabidGravy cool that will make testing easier :) 20:27
brokenchicken Not sure how far we can get away with it. E.g. your module returns a 6.c core `Str` object that's then operated by 6.d stuff that expects 6.d Strs with (potentially) different methods and stuff
Geth doc/coke/bughunt: 107764ae79 | (Will "Coke" Coleda)++ | README.md
Update our symptoms
 20:28
brokenchicken m: '/tmp/foo/d.pm6'.IO.spurt: ｢use v6.d.PREVIEW; sub EXPORT { { '&await' => &await } }｣ 20:29
camelia ( no output )
20:29 cibs left
brokenchicken m: use lib </tmp/foo>; use d; await start say "hi" 20:29
camelia rakudo-moar 015231: OUTPUT«hi␤»
brokenchicken m: use lib </tmp/foo>; use d; say &await.file 20:30
camelia rakudo-moar 015231: OUTPUT«SETTING::src/core.d/await.pm␤»
brokenchicken hehe
a 6.d `await` in 6.c land....
20:31 lukaramu joined, cibs joined
brokenchicken ooh 20:32
m: use lib </tmp/foo>; use d; say (start say "hi").^does("Awaitable")
camelia rakudo-moar 015231: OUTPUT«hi␤False␤»
20:33 nowan_ joined 20:35 nowan left
brokenchicken m: '/tmp/foo/d.pm6'.IO.spurt: ｢use v6.d.PREVIEW; sub EXPORT { { '&new-await' => &await.cando(\(class :: does Awaitable {method get-await-handle {}}.new))[0] } }｣ 20:35
camelia ( no output )
brokenchicken m: use lib </tmp/foo>; use d; new-await start say "hi" 20:36
camelia rakudo-moar 015231: OUTPUT«===SORRY!===␤When invoking 2 '', provided outer frame 0x2a117e8 (4 'EXPORT') does not match expected static frame 0x2a118f0 (3 '')␤»
brokenchicken Hehe. NOW it's starting to crack :P
20:38 andrzejku left
brokenchicken Hm. That's weird really. 'Cause in that previous version, with &await' => &await... It's exporting the proto too, eh? 20:40
And since it didn't crap out, I'm guessing the await start called a 6.c candidate
This is all mind-bending a bit :P
And what else I realize. This use v6.*-per-file limitation needs a better eval bot... 20:43
whoa.. it's quittin' time :o
20:48 nowan joined 20:51 nowan_ left 20:54 Tonik left 20:56 cdg joined 20:57 andrzejku joined 20:59 cdg_ left 21:00 agentzh left 21:04 sufrostico left 21:08 eroux is now known as eroux[away], pmurias joined, cdg_ joined 21:11 labster joined 21:12 cdg left 21:13 cdg joined 21:14 pmurias left 21:16 cdg_ left 21:17 cdg left 21:21 eroux[away] left 21:30 jraglin left
Geth doc: aae50f7b9c | (Wenzel P. P. Peppmeyer)++ | doc/Language/variables.pod6
link to identifier
 21:31
doc: 8553738991 | (Wenzel P. P. Peppmeyer)++ | doc/Language/syntax.pod6
doc interpolation of colon pairs in identifiers
samcv morning all o/ 21:32
RabidGravy Ooh the non-blocking await really breaks HTTP::Server::Tiny 21:33
mst non blocking await? 21:34
El_Che mst: github.com/rakudo/rakudo/pull/1004 21:35
mst aha! 21:36
RabidGravy It winds up with a rather scary "Attempt to unlock mutex by thread not holding it" 21:37
Geth doc: 938c8ea728 | (Wenzel P. P. Peppmeyer)++ | 2 files
link to identifier
 21:39
doc: 041c1591e1 | (Wenzel P. P. Peppmeyer)++ | doc/Language/terms.pod6
fix typo
doc: 1be171ca36 | (Wenzel P. P. Peppmeyer)++ | doc/Language/terms.pod6
more linking to identifier
21:41 labster left 21:42 labster joined
cale2 how do I search for the X operator? 21:46
it's not coming up on the docs search
21:46 skids left
RabidGravy docs.perl6.org/language/operators#...perator%29 21:47
"cross"
21:47 skids joined
cale2 RabidGravy: it shows up in search when you type "xx" in the search box 21:47
could be something to look into
21:48 skids left
cale2 wait it was the wrong routine 21:48
So there's no texas variant of the cross operator :/
21:49 espadrine_ joined 21:52 FROGGS left
gfldex cale2: it in the javascript definition, so it must be a javascript problem 21:52
21:54 andrzejku left 21:56 Vynce left
samcv m: 'a'.uniprop('East_Asian_Width').say 21:56
camelia rakudo-moar 015231: OUTPUT«Na␤»
samcv cool
21:58 pmurias joined 22:00 Actualeyes left, curt_ joined 22:01 cdg joined
AlexDaniel samcv: what's “Na”? 22:02
samcv uh 22:03
idk i forget
i have my ucd rewrite resolving all the short names into long names at least, so hopefully it won't always be as unhelpful
curt_ m: try { my $proc = Proc::Async.new('cat', :r, :w); my $done = $proc.start; say "try sending"; await $proc.say("hi there"); say "worked"; $proc.close-stdin; await $done; CATCH { default { say "caught"; } } }; say "done"; 22:04
camelia rakudo-moar 015231: OUTPUT«caught␤done␤»
samcv AlexDaniel, see unicode.org/Public/UNIDATA/Property...liases.txt
it apparently means Narrow
curt_ m: try { my $proc = Proc::Async.new('nothing', :r, :w); my $done = $proc.start; say "try sending"; await $proc.say("hi there"); say "worked"; $proc.close-stdin; await $done; CATCH { default { say "caught"; } } }; say "done";
camelia rakudo-moar 015231: OUTPUT«caught␤done␤»
samcv m: '≔'.uniprop('ea').say
camelia rakudo-moar 015231: OUTPUT«N␤»
samcv N is neutral 22:05
22:07 Vynce joined 22:08 cdg left, cdg joined
cale2 I want to zip two strings of different length to Pairs 22:11
AlexDaniel samcv: by the way, I want to teach unicodable to dump all properties of a character. Where do I get the list of properties?
cale2 m: 'cale'.comb Z 'helloworld'.comb
camelia rakudo-moar 015231: OUTPUT«WARNINGS for <tmp>:␤Useless use of "Z" in expression ".comb Z 'helloworld'.comb" in sink context (line 1)␤»
cale2 m: say 'cale'.comb Z 'helloworld'.comb 22:12
camelia rakudo-moar 015231: OUTPUT«((c h) (a e) (l l) (e l))␤»
samcv AlexDaniel, second column unicode.org/Public/UNIDATA/PropertyAliases.txt
or
cale2 m: say 'cale'.comb Z 'he'.comb
camelia rakudo-moar 015231: OUTPUT«((c h) (a e))␤»
gfldex m: 'cale'.comb «=>» 'helloworld'.comb
camelia ( no output )
AlexDaniel m: say roundrobin 'cale'.comb, 'helloworld'.comb
camelia rakudo-moar 015231: OUTPUT«((c h) (a e) (l l) (e l) (o) (w) (o) (r) (l) (d))␤»
samcv or see here github.com/perl6/roast/issues/195
gfldex m: dd 'cale'.comb «=>» 'helloworld'.comb
camelia rakudo-moar 015231: OUTPUT«(:c("h"), :a("e"), :l("l"), :e("l"), :c("o"), :a("w"), :l("o"), :e("r"), :c("l"), :a("d"))␤»
cale2 AlexDaniel: thanks
samcv so basically all the ones in the 2nd column plus Emoji 22:13
Emoji_Presentation
Emoji_Modifier
Emoji_Modifier_Base
plus Numeric_Value_Numerator and Numeric_Value_Denominator
cale2 gfldex I need it to always favor one side. what the heck is that operator haha
AlexDaniel samcv: but… any way to get it from perl 6?
samcv no way
AlexDaniel :'(
gfldex => is the Pair constructor and «» is the hyper that doesn't care about length of arguments 22:14
22:14 RabidGravy left
cale2 «» <- not to be confused with something else, as I discovered 22:15
timotimo cale2: but the cross operator is already texas?
22:15 rba_ joined
cale2 timotimo, try searching for it in the docs 22:15
samcv AlexDaniel, also there's NFG_QC too
timotimo aye, it's missing
but that has nothing to do with the fact that X is in ASCII
gfldex timotimo: it's not, it's in the javascript index 22:16
cale2 timotimo: type "xx" and you can find something similar
gfldex for whatever reason jquery is acting up
timotimo xx is very unsimilar, actually
gfldex "X " does work
cale2 yeah, but "xx" and you see the single "x"
22:16 avuserow left
timotimo but x and X aren't the same thing 22:16
22:17 Actualeyes joined
AlexDaniel m: say ‘cale’.comb »=>» ‘helloworld’.comb 22:17
camelia rakudo-moar 015231: OUTPUT«(c => h a => e l => l e => l)␤»
cale2 m: say 'cale'.comb Z 'he'.comb # I need the left side to remain the same length, but the right side to cycle forever
camelia rakudo-moar 015231: OUTPUT«((c h) (a e))␤»
AlexDaniel m: say ‘cale’.comb «=>« ‘helloworld’.comb
camelia rakudo-moar 015231: OUTPUT«(c => h a => e l => l e => l c => o a => w l => o e => r c => l a => d)␤»
gfldex the search is case insensitive
22:17 bjz left
gfldex m: say ‘cale’.comb Z[=>] ‘helloworld’.comb 22:17
camelia rakudo-moar 015231: OUTPUT«(c => h a => e l => l e => l)␤»
cale2 AlexDaniel: Sorry, I need the left side to be the one determining the length
AlexDaniel cale2: so use »=>» then 22:18
22:18 avuserow joined
cale2 gfldex: that's the solution 22:18
what is going on with that extra arrow next to Z?
brokenchicken makes a pair 22:19
cale2 m: say ‘cale’.comb Z[=>] ‘he’.comb
AlexDaniel it's a pair operator
camelia rakudo-moar 015231: OUTPUT«(c => h a => e)␤»
gfldex it's the inner operator of the meta Z
cale2 wait, still doesn't work...
gfldex meta Z is not the same thing as Z
AlexDaniel »=>» is what you want
brokenchicken m: say ‘cale’.comb Z=> ‘he’.comb
camelia rakudo-moar 015231: OUTPUT«(c => h a => e)␤»
cale2 m: say ‘cale’.comb »=>» ‘he’.comb
camelia rakudo-moar 015231: OUTPUT«(c => h a => e l => h e => e)␤»
brokenchicken gfldex: what does the [] on it do?
AlexDaniel m: say 5 [+] 2
camelia rakudo-moar 015231: OUTPUT«7␤»
AlexDaniel nothing 22:20
brokenchicken OK
cale2 »=>» that doesn't work because it just repeats the last letter. I need the right list to cycle
gfldex brokenchicken: precedence, so it forces => to be the inner operator of the meta chain
timotimo Z=> and Z[=>] is the same thing, but the latter is more readable
brokenchicken meta chain?
brokenchicken feels outta the loop :o
AlexDaniel m: say ‘cale’.comb »=>» ‘XY’.comb
camelia rakudo-moar 015231: OUTPUT«(c => X a => Y l => X e => Y)␤»
AlexDaniel cale2: are you reading it correctly? 22:21
cale2 Yeah, that worked. I read it wrong haha
m: say ‘cale’.comb»=>»‘XY’.comb 22:22
camelia rakudo-moar 015231: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Malformed postfix␤at <tmp>:1␤------> 3say ‘cale’.comb»7⏏5=>»‘XY’.comb␤ expecting any of:␤ method arguments␤ postfix␤»
brokenchicken .tell RabidGravy be sure to report your breakages so we iron everything out in v6.d.PREVIEW and not leave it until v6.d ;)
yoleaux brokenchicken: I'll pass your message to RabidGravy.
cale2 m: say ‘cale’.comb»=>» ‘XY’.comb
camelia rakudo-moar 015231: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Malformed postfix␤at <tmp>:1␤------> 3say ‘cale’.comb»7⏏5=>» ‘XY’.comb␤ expecting any of:␤ method arguments␤ postfix␤»
AlexDaniel cale2: nope, you need whitespace there
cale2 I thought hyper operator was supposed to butt up against
list>>.map(stuff)
AlexDaniel that's not that 22:23
timotimo that's hyper methodcall
method calls have only allowed you to put a space in front since half a year or maybe a whole year
22:23 rurban joined
cale2 m: say ‘cale’.comb >>=>>> ‘XY’.comb 22:23
camelia rakudo-moar 015231: OUTPUT«(c => X a => Y l => X e => Y)␤»
cale2 m: say ‘cale’.comb >> => >> ‘XY’.comb
camelia rakudo-moar 015231: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Unsupported use of >> to do right shift; in Perl 6 please use +> or ~>␤at <tmp>:1␤------> 3say ‘cale’.comb >>7⏏5 => >> ‘XY’.comb␤»
timotimo for that case you can use [ ] around => to make it more readable
cale2 m: say ‘cale’.comb >>[=>]>> ‘XY’.comb
camelia rakudo-moar 015231: OUTPUT«(c => X a => Y l => X e => Y)␤»
cale2 nice
y'all are getting very haskelly with these operators 22:24
gfldex brokenchicken: see design.perl6.org/S03.html#Nesting_o...aoperators
AlexDaniel doesn't see how adding another two non-letter characters is going to help with readability, but I'll note that…
gfldex ENODOC on nesting metaoperators btw 22:25
@a >>[>]>> $b # oh yeah
cale2 m: say ‘cale’.comb >>[=>]>> ‘QRSTUVQXYZ’.comb
camelia rakudo-moar 015231: OUTPUT«(c => Q a => R l => S e => T)␤»
22:25 Vynce1 joined
cale2 perfect 22:25
brokenchicken m: my @a <a b c>; my @b = 1, 2, 3; say @a [X+]= @b 22:26
camelia rakudo-moar 015231: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Missing required term after infix␤at <tmp>:1␤------> 3my @a <a b c>7⏏5; my @b = 1, 2, 3; say @a [X+]= @b␤ expecting any of:␤ prefix␤ term␤»
brokenchicken m: my @a = <a b c>; my @b = 1, 2, 3; say @a [X+]= @b
camelia rakudo-moar 015231: OUTPUT«Cannot convert string to number: base-10 number must begin with valid digits or '.' in '3⏏5a' (indicated by ⏏)␤ in block <unit> at <tmp> line 1␤␤Actually thrown at:␤ in block <unit> at <tmp> line 1␤␤»
cale2 I'm sure there is an intuitive explanation for the hyper operator. It's just not clear on first glance.
22:26 Vynce left
brokenchicken m: my @a = 10, 20, 30; my @b = 1, 2, 3; say @a [X+]= @b 22:27
camelia rakudo-moar 015231: OUTPUT«[11 12 13 21 22 23 31 32 33]␤»
22:27 sena_kun left
brokenchicken m: my @a = 10, 20, 30; my @b = 1, 2, 3; say @a X[+=] @b 22:27
camelia rakudo-moar 015231: OUTPUT«(11 13 16 21 23 26 31 33 36)␤»
cale2 Why not instead make a "cycle" method that repeats a list infinitely. Then you don't need "hyper" at all.
brokenchicken cale2: see rotor
cale2 brokenchicken: I did. I thought that was for chunking 22:28
lizmat m: dd ((1,2,3) xx *)[^10]
camelia rakudo-moar 015231: OUTPUT«((1, 2, 3), (1, 2, 3), (1, 2, 3), (1, 2, 3), (1, 2, 3), (1, 2, 3), (1, 2, 3), (1, 2, 3), (1, 2, 3), (1, 2, 3))␤»
brokenchicken lizmat++
lizmat m: dd (|(1,2,3) xx *)[^10]
camelia rakudo-moar 015231: OUTPUT«(1, 2, 3, 1, 2, 3, 1, 2, 3, 1)␤»
brokenchicken gfldex: what about "meta Z is not the same thing as Z". I thought Z was just the meta Z with infix:<,> op
samcv do Augmenting built-in classes make executed code slower? 22:29
lizmat brokenchicken: internally it works a bit more optimized, but otherwise Z is the same as Z,
afaik
samcv or just compiling? also curious about custom operators
22:29 agentzh joined
samcv aside from compilation speed, is there any speed reduction from using them? 22:29
lizmat samcv: creating a custom operator means cloning the grammar 22:30
gfldex i believe to remember that meta Z behaves differently with LoLs but that may just have been a bug
lizmat which is still a lot of work at compile time
samcv what about run-time?
gfldex the inliner should inline them 22:31
timotimo hyper operators will attempt to mirror the incoming structure
lizmat I recall a check in the setting that makes sure we didn't clone the grammar during setting building
brokenchicken cale2: you're right; the `xx` is the right tool
timotimo (which is why they also work on hashes)
brokenchicken Also I notice a bug with .rotor:
mc: m: dd (1,2,3).rotor(3 => -3)[^10]
committable6 brokenchicken, ¦«2015.12»: ((1, 2, 3), (1, 2, 3), (1, 2, 3), (1, 2, 3), (1, 2, 3), (1, 2, 3), (1, 2, 3), (1, 2, 3), (1, 2, 3), (1, 2, 3))
brokenchicken m: m: dd (1,2,3).rotor(3 => -3)[^10]
camelia rakudo-moar 015231: OUTPUT«Rotorizing gap is out of range. Is: -3, should be in -2..^Inf; ␤Ensure a negative gap is not larger than the length of the sublist␤ in block <unit> at <tmp> line 1␤␤»
lizmat because of its runtime ramifications, so I think there are some runtime ramifications as well
brokenchicken m: m: dd (1,2,3).rotor(3 => -3, :partial)[^10] 22:32
camelia rakudo-moar 015231: OUTPUT«Rotorizing gap is out of range. Is: -3, should be in -2..^Inf; ␤Ensure a negative gap is not larger than the length of the sublist␤ in block <unit> at <tmp> line 1␤␤»
timotimo once the grammar is cloned, it's no performance issue
brokenchicken m: m: dd (1,2,3).rotor(3 => -2, :partial)[^10]
camelia rakudo-moar 015231: OUTPUT«((1, 2, 3), (2, 3), (3,), Nil, Nil, Nil, Nil, Nil, Nil, Nil)␤»
jnthn Only slightly longer startup time because of deserializing the extra stuff
And even then you probably won't hit it unless you EVAL
samcv so this github.com/samcv/UCD/blob/master/l...rators.pm6 should not be acause of slowness?
jnthn 'cus we lazily deserialize
samcv unless EVAL?
22:33 jp_ joined
lizmat brokenchicken: that last one feels wrong ? 22:33
ah, no, actually, that's correct
brokenchicken lizmat: yeah, seems an off-by-one
jnthn samcv: Those operator defs should only cuase compile-time slowdown; for runtime they'll dispatch just like normal operators
cale2 m: say |'brokenchicken'.comb xx 20
camelia rakudo-moar 015231: OUTPUT«(b r o k e n c h i c k e n b r o k e n c h i c k e n b r o k e n c h i c k e n b r o k e n c h i c k e n b r o k e n c h i c k e n b r o k e n c h i c k e n b r o k e n c h i c k e n b r o k e n c h i ...)␤»
cale2 m: say |'brokenchicken'.comb xx * 22:34
camelia rakudo-moar 015231: OUTPUT«(...)␤»
samcv kk thanks jnthn
gfldex cale2: please note that `say` is a wee bit magic
jnthn samcv: About EVAL - I just mean that if you precompile a module with custom operators used in it, then it will include the language tweaks as part of its precompilation.
lizmat brokenchicken: off by one how ?
brokenchicken lizmat: Actually, not even an off-by-one, but it's checking the negative gap against the positive step instead of whether or not it went past the start of the list
mc: dd (^10).rotor(5, 4 => -5, :partial)[^10] 22:35
committable6 brokenchicken, ¦«2015.12»: ((0, 1, 2, 3, 4), (5, 6, 7, 8), (4, 5, 6, 7, 8), (9,), Nil, Nil, Nil, Nil, Nil, Nil)
jnthn samcv: But since we lazily load stuff then you'll probably only pay for that if you actually need the language
brokenchicken lizmat: ^ like there it isn't even an infinite list
jnthn samcv: EVAL being something that needs it, because an EVAL is done in terms of the enclosing language.
brokenchicken mc: dd (^10).rotor(2, 1 => -2, :partial)[^10]
committable6 brokenchicken, ¦«2015.12»: ((0, 1), (2,), (1, 2), (3,), (2, 3), (4,), (3, 4), (5,), (4, 5), (6,))
brokenchicken m: dd (^10).rotor(2, 1 => -2, :partial)[^10]
camelia rakudo-moar 015231: OUTPUT«Rotorizing gap is out of range. Is: -2, should be in -0..^Inf; ␤Ensure a negative gap is not larger than the length of the sublist␤ in block <unit> at <tmp> line 1␤␤»
22:35 bjz joined
brokenchicken Sad that ain't any test for that behaviour :( 22:36
lizmat yeah, otherwise I would have noticed during the refactor
cale2 What is more intuitive between these two: $secret.comb Z |$keyword.comb xx * OR $secret.comb >>[=>]>> $keyword.comb 22:37
22:37 Vynce joined
brokenchicken cale2: the first one 22:37
cale2: but I don't really program Perl 6...
22:37 Actualeyes left
cale2 brokenchicken: you say that often, but you clearly do haha 22:38
gfldex hyperops are candidates for autothreading, so asking for intuition is wrong
timotimo right, hyper operators are hyper, while Z and X metaops are lazy 22:39
brokenchicken Hm. GitHub added repository "topics" for ease of discovaribility
22:39 Vynce1 left
timotimo oh? interesting 22:39
22:39 rindolf left
jp_ Hi, I have a question. I don't understand why I have to add {;} in a regex to get the "correct" behavour. You can see my test case at : gist.github.com/anonymous/fca80723...c674c50469 22:39
gfldex github.com/blog/2309-introducing-topics 22:40
jp_ If you have any idea ...
timotimo if that's the case, you might want to use || instead of | in some places 22:41
brokenchicken m: say 84+131+84+63+153+4
camelia rakudo-moar 015231: OUTPUT«519␤»
timotimo with {;} (even though {} should also work) you can force the declarative prefix to end earlier than it otherwise would
jp_ good to know...
brokenchicken cale2: well, in the past ~9 months I wrote 519 lines of Perl 6 code. All for a single project :) github.com/perl6/geth Hardly what'd I'd call "programming" 22:42
timotimo the | alternation will first go into the branch that has the longest match in its declarative prefix
jp_ ok
brokenchicken What's a declarative prefix?
timotimo everything from the beginning to the first non-declarative piece 22:43
brokenchicken OK :)
brokenchicken has no idea which pieces are declarative :o
timotimo || for example is non-declarative
it's imperative, because it has an "order of execution" to it
{ say "hi" } is imperative, because nobody knows what a piece of code is able to do 22:44
brokenchicken some comp-sci stuff.
timotimo backreferences are also not declarative
22:44 Actualeyes joined
brokenchicken OK. I think I get it. Thanks. 22:44
[ 'foo' | 'bar' ] is declarative. [ $meow ] ain't. [ .+ ] is 22:45
timotimo $meow could be declarative
but for that we'd have to recompile the NFA every time we do a match
brokenchicken Ah
timotimo (in other words every time $meow could have changed)
mr_ron Warning on mixed | and || and related documentation needs github.com/perl6/doc/issues/1141 22:47
brokenchicken What's QAST::Op.new( :op('locallifetime'), ? It's not documented.
22:48 bjz left
brokenchicken It's here: github.com/rakudo/rakudo/blob/nom/....nqp#L7845 22:48
Geth perl6.org: hankache++ created pull request #71:
Link to "Think Perl 6" page on oreilly.com		 22:49
timotimo it's for defining when local variables can be considered dead
Geth perl6.org: dcfc23ce69 | (Naoum Hankache)++ | source/resources/index.html
Link to "Think Perl 6" page on oreilly.com
perl6.org: fcf6628f0f | lizmat++ | source/resources/index.html
Merge pull request #71 from hankache/patch-1

Link to "Think Perl 6" page on oreilly.com
TimToady brokenchicken: for instance, I think the temporary setting of $_ on the right side of ~~ uses that 22:52
Geth perl6.org: 8afc235eb2 | (Zoffix Znet)++ | source/resources/index.html
s{}{https://}

Where sites offer it		 22:53
timotimo ^- good call
Geth perl6.org: ca5dfe73fa | (Zoffix Znet)++ | source/resources/index.html
s{}{https://}
 22:54
brokenchicken Thanks.
22:57 cdg left 22:58 lukaramu left
jp_ Hi again, I have another testcase with the same grammar parsing another text where this is when I activate the Grammar::Tracer that I get the result I want. You can find my testcase at gist.github.com/anonymous/27c7c108...c491a5064. 23:04
23:04 rurban left
jp_ ( Note this time this is when I drop the {;} that I get the "good" result ... 23:05
23:06 kyan joined
lizmat brokenchicken: re rotor allowing negative gap beyond the step forwar 23:07
d: that would imply that all values of the source iterator should be cached :-(
brokenchicken yup, which is why I didn't say it has to be fixed :P~ 23:08
23:08 RandalSchwartz joined
brokenchicken The merlyn has graced us with his presence! 23:08
RandalSchwartz "rakudobrew build moar" doesn't seem to have updated anything in "bin/*" since January 3rd.
is there something I should change?
23:09 Actualeyes left
brokenchicken RandalSchwartz: maybe run rakudobrew rehash? 23:09
I dunno. It is meant to update bin/*?
RandalSchwartz that's what was building my copy of "perl6"
brokenchicken The rehash thing updates its shims
RandalSchwartz in "bin/*"
brokenchicken Hm. You can run perl6 -v to see which version you got 23:10
m: say $*PERL.compiler.version
camelia rakudo-moar 015231: OUTPUT«v2017.01.127.g.0152316␤»
brokenchicken should be something along those lines ^ with 127 being the number of commits since 2017.01 compiler release
RandalSchwartz This is Rakudo version 2017.01-127-g0152316f5 built on MoarVM version 2017.01-25-g70d4bd53
brokenchicken Yup, You got HEAD
RandalSchwartz Ahh. So, somehow, that's newer than the bin. :)
I was worried since the bin wasn't being updated 23:11
thanks
brokenchicken Looks like ~/.rakudobrew/bin/perl6 is a Perl 5 script that just figures out The Right Thing to become and execs into it
RandalSchwartz ok 23:12
brokenchicken lizmat: I mean, I guess it's useful to have X ways to use rotor than X-1 ways, but there weren't any tests and I'm guessing .rotor would need to be completely changed after your rewrite and there weren't any roast tests 23:14
lizmat yeah, it would have to change rather significantly 23:15
but nothing unoverseeable
it's just that none of the examples in the doc use this feature, and there are no tests for it
brokenchicken yeah
lizmat so I'm going to leave this up to TimToady to decide upon, as he wrote the original List.rotor handler 23:16
if my memory servers me correctly
23:17 RandalSchwartz left
lizmat *serves 23:17
23:19 pyrimidine left 23:20 pyrimidine joined 23:21 wamba left, Actualeyes joined
lizmat brokenchicken: perhaps the solution is to reinstate a separate candidate for List.rotor 23:22
brokenchicken ¯\_(ツ)_/¯ up to you and TimToady :) 23:23
23:24 pyrimidine left 23:26 lep-delete left 23:28 lep-delete joined
lizmat hmmm.. appears List.rotor was a co-production of nine, brokenchicken and me 23:30
brokenchicken: in any case, it appears from the error message "Ensure a negative gap is not larger than the length of the sublist" 23:31
that you should not be able to do 2 => -3
brokenchicken Sounds good :)
23:35 girafe left 23:41 as__ joined
as__ How can I redefine the $?TABSTOP in my code 23:41
?
23:42 Actualeyes left
brokenchicken Never saw those redefined, since they're compile-time constants. 23:42
as__: how come you want to?
[Coke] m: $?TABSTOP=4; # nope 23:44
camelia rakudo-moar b9d927: OUTPUT«Cannot assign to an immutable value␤ in block <unit> at <tmp> line 1␤␤»
timotimo m: my constant $?TABSTOP = 8; 23:45
camelia rakudo-moar b9d927: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Constants with a '?' twigil not yet implemented. Sorry. ␤at <tmp>:1␤------> 3my constant $?TABSTOP = 87⏏5;␤»
timotimo seems to be allowed only in the core setting
as__ so kinda not possible 23:46
[Coke] m: my constant $?HI = 3; $?HI=4;
camelia rakudo-moar b9d927: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Constants with a '?' twigil not yet implemented. Sorry. ␤at <tmp>:1␤------> 3my constant $?HI = 37⏏5; $?HI=4;␤»
as__ it may be useful for heredocs with tabs instead of spaces
timotimo i'm not sure if heredocs and tabs mix well already 23:47
as__ more precisely, with mixture of tabls and spaces
timotimo i.e. whether "\t \t \t" is the same as "\t\t\t"
yeah
as__ a tab is assumed as 8 spaces
brokenchicken m: "\tx\n y".indent(-2).uninames.say 23:48
camelia rakudo-moar b9d927: OUTPUT«(SPACE SPACE SPACE SPACE SPACE SPACE LATIN SMALL LETTER X <control-000A> LATIN SMALL LETTER Y)␤»
lizmat I seem to recall that the use of $?TABSTOP was for Str.indent and Str.de-indent 23:49
oops, scratch de-indent
I'm not sure why Str.indent uses $?STABSTOP instead of a $*TABSTOP that gets its initial value from $?TABSTOP 23:50
as__ perl6advent.wordpress.com/2013/12/...docs-docs/ reads "(Tabs are considered to be 8 spaces long, unless you change $?TABSTOP. "
brokenchicken hehehe STABSTOP :P 23:51
lizmat well, and there you go, you can't
because before Christmas we weren't sure you *could* change something like $?FOO in your own code
brokenchicken And it's pretty easy to workaround it
as__ like how?
lizmat so we made it illegal until we figure it out
brokenchicken m: "\tx\n y".subst(:g, /^(\s+)/, { .[0].subst: "\t", " " x 4 }).indent(-2).uninames.say 23:52
camelia rakudo-moar b9d927: OUTPUT«(SPACE SPACE LATIN SMALL LETTER X <control-000A> LATIN SMALL LETTER Y)␤»
brokenchicken Sorry, ^^ not ^
m: "\tx\n y".subst(:g, /^^(\s+)/, { .[0].subst: "\t", " " x 4 }).indent(-2).uninames.say
camelia rakudo-moar b9d927: OUTPUT«(SPACE SPACE LATIN SMALL LETTER X <control-000A> LATIN SMALL LETTER Y)␤»
as__ Yes but not in heredocs before the closing label
or how do you call it?
brokenchicken as__: personally, I wouldn't be mixing tabs and spaces in my code so would never encounter that issue with heredocs 23:54
as__ theory is more important for me than practice at the moment :) just figured out that you cannot mix it and wanted to know how to deal with that 23:55
brokenchicken as__: we have an acronym for that: DITHWIDT—Doctor, It Hurts When I Do This
You'd have to indent your heredoc with, say, a tab on one line, then not-8 spaces on another to ever encounter this. 23:56
23:59 AlexDaniel left

Powered by Cro and the Raku® Programming Language.

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

Thank you!