»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.
sena_kun How can I dereference Pointer using NativeCall(for tests)? 00:00
gfldex m: my $s = '42'; put Q:s:b{\$s} 00:01
camelia rakudo-moar d7f3f0: OUTPUT«\42␤»
gfldex shouldn't the \ vanish?
00:01 skids joined 00:02 bpmedley left, regreg joined, hankache joined
gfldex and stop $s from being interpolated on top of that 00:03
RabidGravy profan, I think, without looking at the code, that it always follows redirects, but yeah it probably needs some "redirect-ok" callback. I have a plan for adding some phase callbacks to influence the way the requests and responses are handled
profan RabidGravy: sounds good :), i'll let you know if anything else comes up or things go out of wack 00:05
RabidGravy If you really, really don't want redirects then you can just set max-redirects to 0, of course it will crap out but the exception will have the response in it ;-) 00:07
sortiz Hi, While setting PERL6LIB=/some/path works, if I use the switch "-I /some/path" rakudo ignores me, someone has an idea why? 00:09
00:12 wamba joined
gfldex m: my $s = '42'; put Q:q:s{\$s} 00:12
camelia rakudo-moar d7f3f0: OUTPUT«\$s␤»
00:13 AndyDee left
timotimo hmm. i'd always write it -I/some/path, but that may not be a/the reason 00:13
flussence -I doesn't get passed to subprocesses... and doesn't precomp use those?
Skarsnik sena_kun, depend, etheir nativecast of you type your pointer and use .deref on it, if you just want the address, I think .Int or .gist show you
-I path work
sortiz timotimo: Thanks but with or without space I see the same results: "Could not find Foo" at Module use time :( 00:16
sena_kun Skarsnik, I have a pointer to integer with errcode(from C library) and want to compare it with zero(for example). I'll try .deref then.
00:17 BenGoldberg joined 00:18 lostinfog left, bpmedley joined
AlexDaniel what about “Perl 6 takes off, Perl 5 has enough developers to improve and maintain it for the next 10 years” scenario? Why isn't it an option? 00:20
(www.dagolden.com/index.php/2589/per...-enemies/)
gfldex m: my $s = '42'; put Q:qq{\$s$s {$s}}, ' ', Q:s:a:h:f:c:b{\s$s {$s}}
camelia rakudo-moar d7f3f0: OUTPUT«$s42 {42} \s42 {42}␤»
stmuk__ I can't see any probs with the RC0 R* 00:21
Skarsnik Good night #perl6
AlexDaniel it would be weird to assume that perl 5 developers would just abandon the ship immediately. Sure they may have less interest, but as far as I can see Perl 5 is still going strong and it does not look like it will change any time soon
00:21 Skarsnik left
stmuk__ I don't think anyone can guess how computer langs get adopted or abandoned (if they are ever abandoned which I doubt) 00:24
gfldex .u $
yoleaux U+0024 DOLLAR SIGN [Sc] ($)
gfldex m: my $s = '42'; put Q:s:b{\c[DOLLAR SIGN]s$s$s} 00:26
camelia rakudo-moar d7f3f0: OUTPUT«$s4242␤»
RabidGravy is there any way that one can gain control of the way that multi candidates are selected programmatically, (without rewriting the language)?
AlexDaniel stmuk__: let's use it as a secure random number generator!
00:27 leont joined
timotimo sortiz: did you try that environment variable that turns on module loading debugging? 00:27
sortiz: or try printing the repo chain?
00:27 yurivish left
stmuk__ people still use LISP and Fortran afterall (I've even seen the latter in jobs and, horror of all horrors csh) 00:27
hankache we still have COBOL people at work 00:28
and AS400 machines 00:29
profan * "Run To The Hills" starts playing in the background * 00:30
RabidGravy as/400 is a rather nice platform 00:31
timotimo excuse my ignorance, but what do i turn "int" from a C header into with NativeCall?
stmuk__ I bet people use VMS still even
RabidGravy timotimo, probably int32 00:32
hankache timotimo int32 ?
sena_kun timotimo, int32.
timotimo oh? well, all right
flussence you wouldn't think it on x64, but apparently that's it in all cases... 00:33
sena_kun I have a native function that returns void. Obviously, I don't return anything in perl code also. But I've got "Native call expected return type with CPointer representation, but got a VMArray" message. From where should I start? 00:34
profan flussence: only long does the switcheroo there
RabidGravy if you're unsure it's just compile a little C program that prints sizeof(foo)
sena_kun, is that "void" or "void *" if the former just omit any retuns thing 00:36
timotimo fixes up SDL2::Raw as well as Cairo 00:37
RabidGravy timotimo++
sena_kun RabidGravy, it's plain "void". Signature in perl6 code doesn't have any "returns". Should I made a pastebin page with code to show example? 00:38
RabidGravy yeah I think so, that seems odd
timotimo i'm not sure what exact version i should supply ...
sena_kun RabidGravy, just a second... 00:39
timotimo tell me, what did .first-index turn into again?
flussence .first(:k)
iirc
timotimo thx
flussence m: say 'foo'.first('o', :k)
camelia rakudo-moar d7f3f0: OUTPUT«Nil␤»
flussence m: say 'foo'.first('o', :v)
camelia rakudo-moar d7f3f0: OUTPUT«Nil␤»
flussence m: say 'foo'.first('o')
camelia rakudo-moar d7f3f0: OUTPUT«Nil␤»
sortiz Where in $*REPO.repo-chain list can I look for the -I switch effects ? 00:40
flussence m: say 'foo'.comb.first('o', :k) # I'm dumb...
camelia rakudo-moar d7f3f0: OUTPUT«1␤»
00:40 hankache left
RabidGravy timotimo, most linux make links for 1, 1.0, 1.0.4, 1.0.4.uncletomcobbleyandall 00:40
00:41 virtualsue left, joydo joined
timotimo hm 00:41
flussence
.oO( wonder if it'd be worthwhile to detect list methods called on scalar literals and emit a warning... )
timotimo there's libSDL2-2.0.so.0.2.1 here
how the ...
is that considered normal?
flussence libSDL has a ... fun numbering scheme 00:42
timotimo how does that map to what i supply to nativecall's "native" thread?
stmuk__ graphics and video libraries are always a total PITA
sena_kun RabidGravy, pastebin.com/47tDwiwc - here is the link.
00:42 alpha123 left
flussence ‘SDL2-2.0’ is the name and everything after ‘.so.’ is the version 00:43
fwiw I have a libSDL-1.2.so.0.11.4 here too...
timotimo well, that's libSDL1
i'm building a libSDL2 binding here
flussence same crazy libname scheme though :) 00:44
sortiz With PERL6LIB set, a 'file#/some/path' appears in repo-chain head, but with -I/some/path, nothing changes.
stmuk__ that's probably SDL 1.2.11 (?) 00:45
flussence stmuk__: mine is, yeah
RabidGravy sena_kun, I'd say the Pointer[void] there wants to be a CArray instance with as many elements as the len argument 00:46
flussence sortiz: you might get something useful out of RAKUDO_MODULE_DEBUG=1
stmuk__ I had to build 1.2.15 from source recently and it doesn't even compile without a patch from a mailing list
RabidGravy so CArray rather than Pointer[void]
00:47 joydon left
sena_kun RabidGravy, and I should remove last int then? 00:47
flussence oh hang on, distro says this is 1.2.15-r9
RabidGravy no, you need the last int, but CArrays don't know their own own length 00:48
so "my $a = CArray[uint8].new; $a[$len] = 0;" for example 00:49
stmuk__ distro libraries are often early versions with patches backported from later versions
sena_kun RabidGravy, I'll try it now, thanks.
timotimo okay, the two modules are up to date. except i decided against putting a version into the libsdl2 thingie 00:52
so that i don't go insane
profan hows the state of the JVM backend by the way? 00:56
if anyone knows anything that is :) 00:57
TimToady it was back-burnered a couple of months while we got 6.c out the door, but I expect various champions will step up to bring it up to snuff over the next few months 00:58
profan ah, i figured it hasn't been a focus lately :) appreciate the info TimToady++
RabidGravy TimToady++ expressed it better, I was going to say it needed some love
sena_kun RabidGravy, it fails anyway. I think I should make a short example of my code for someone free to try it. I'm trying to write a bzip2 bindings btw. 00:59
zwu Hi, I didn't find the doc about the type modifier especially for the class method, such as class A { method foo(::A:D:) {}; }
RabidGravy sena_kun, it's a bit late here now but if you stick it on github or something I'll take a look tomorrow or so 01:00
sortiz Folks: "perl6 -I /some/path -e '...'" works, but not in REPL mode! 01:01
zwu how can we put the type modifier to the Role method too? 01:02
sena_kun RabidGravy, it's already 3 a.m. here, so I'm done with hacking too. Okay, I'll make a quick reproducible gist and paste link to you. 01:03
RabidGravy sure 01:04
just .tell yoleaux and it will remind me :)
toodles folks
01:09 RabidGravy left 01:13 jack_rabbit_ joined 01:14 sena_kun left 01:18 cognominal left 01:29 leont left 01:33 mike1 joined 01:36 mike1 left 01:42 xpen joined 01:46 xpen left
zwu question on module and import. I created a module file under directory p/m/testm.pm6, in which module A { sub foo() { say "hello"}, class MyClass {...} }, 01:49
in the test code, I can load through need p::m::testm; and import A;
I can use the class through A::MyClass, and use function foo defined in A, but not A::foo; 01:50
the sub foo is exported through the trait is export for foo 01:51
skids Maybe try "our sub foo()" ?
zwu but I didn't export MyClass in module A, 01:52
skids classes may have different default behavior?
zwu ah, is that our, my ... for the scope related issue?
01:52 wamba left
skids Yes "our" is a scoping thing like "my" 01:53
zwu yes, using our scope for the sub, I can use A::foo even it is not exported. 01:55
skids Yay! I approximately knew what I was talking about :-) 01:56
zwu so what I understand is that the import actually put the symbols into current lexical scope without the prefixes of the package scopes. 01:57
skids Yes import is for moving symbols from one Stash to the current Stash.
01:59 lichtkind left
zwu another question, since class is default in our scope, that means when a module is loaded, all classed defined in that module file are actually accessible through full package name? 02:00
that makes the export for class almost useless, 02:01
and hard to protect the usage for the non exported classes?
02:02 protium joined 02:03 molaf_ joined
zwu and the difference for export sub routing is because the subs are defined in defined my scope instead of our scope, so the is export trait helps to export the symbol, but when you import them, only the symbol name is imported into current lexical scope, 02:04
if there are multiple module export the same names, there are more chances of collisions than just use our scope for the subs. 02:05
02:07 molaf left 02:08 glucosade joined 02:09 nzkindest left 02:10 AlexDaniel left
dalek c: 9da18b2 | skids++ | doc/Type/Stash.pod:
Verbiage tweaks
 02:15
02:15 kid51_ joined
skids zwu: you can "my" a class 02:16
02:16 cognominal joined 02:18 kid51 left 02:19 kid51 joined 02:22 kid51_ left
sortiz ls 02:22
02:24 Ch0c0late left 02:28 Guest19685 joined
Guest19685 Yesss there's a perl6 channel 02:28
02:29 Guest19685 is now known as linuxuser9000
linuxuser9000 What's up perly-perlers 02:29
TimToady not most of the Eurofolk :)
skids o/
linuxuser9000 I'm so sick of people dissing perl. So I decided to go where other perl lovers are 02:34
Have you guys seen the reverse polish example on the perl6 website? I think it's amazing: examples.perl6.org/categories/inter...s/RPN.html
sortiz linuxuser9000: You are welcome!
linuxuser9000 Also, the quicksort implementation on wikipedia is so nice looking: en.wikipedia.org/wiki/Perl_6#Quicksort 02:35
02:36 linuxuser9000 left
zwu I remembered someone has said that use module is not any more just as the combine of need and import; is the use the same thing as the need just for loading? 02:37
02:37 linuxuser9000 joined
linuxuser9000 Ugh, accidentally disconnected myself -_- 02:38
I have a question. I used the socket.io library for node.js, but the library is really two parts, front end and back. Is there a module for perl that gives the features of socket.io? (socket.io lets you recieve and emit events; the events have a string key, and you can send anything with the event: json, functions, whatever. it's sockets) 02:39
zwu How to make the dynamic module loading by require to load the module specified by a file path, i.e, do not need to search the lib path 02:41
02:42 setty1 left
linuxuser9000 This example for sockets looks cool: poe.perl.org/?POE_Cookbook/Chat_Server 02:43
02:47 keix left
timotimo doc.perl6.org/type/IO::Socket::Async - check this for some asynchronous chat server/client stuff 02:50
extremely basic, but a good start point, IMO
making a socket.io compatible module would be implementing whatever protocol they've created and speaking that via TCP or whatever it is that they use 02:51
skids zwu: Loading is still a bit unpolished. For now you are better off using PERL6LIB= than trying to get absolute-path loading to work right. 02:57
(Or "use lib '/path/to'; use Mymod") 02:58
zwu I'm designing a build system for dependency management on C++, trying to using perl 6, I need to load a configuration or even just the perl script it self for each projects. 03:00
Dynamic load a script will make it convenient, instead of having to load it as string and then import? by the way how can we load to a string and then import it? 03:01
03:02 jdrab_ joined
ugexe zwu: assuming you have a interfaced defined so you know what they will provide you can do something like this github.com/ugexe/zef/blob/master/l...m6#L20-L26 03:04
03:04 vendethiel left
ugexe that loads plugin modules and passes them options from a config file 03:04
zwu ugexe: thanks a lot! 03:05
03:05 bakedb_ joined 03:07 livyathan joined 03:10 mr-foobar joined 03:11 geraud left, pierrot left, jdrab left, noganex_ left, Khisanth left, crux left, troydm left, jojotus left, Util left, cbk__ left, jervo left, wtw left, bakedb left, mspo left, larion_ left, dj_goku left, musca left, pjcj left, ranguard left, jast left, bowtie left, apejens left, avar left 03:14 keix joined
zwu I saw the doc S11.pdf that it said Alternately, a filename may be mentioned directly, as require "/home/non/Sense.pm" <common @horse>;. 03:20
that's great, it already support the dynamic import on any file path without searching.
03:21 yqt left 03:22 jervo joined, Khisanth joined, apejens joined, jojotus joined, Util joined, cbk__ joined, musca joined, jast joined, pjcj joined, mspo joined, larion_ joined, dj_goku joined, ranguard joined, literal joined, krakan joined, samb1 joined, jantore_ joined, skaji joined, d^_^b joined, garu joined, cfloare joined, bowtie joined, avar joined, troydm joined 03:24 corbyhaas left 03:25 wtw joined, crux joined, zemmihates left
ugexe true. it will load slower though as it wont be precompiled which may or may not matter to you 03:25
03:26 zemmihates joined, corbyhaas joined 03:27 deepika joined 03:30 noganex joined
zwu oh, seems the require does not work, i changed "need some_module;" to require 'full_path file.pm6' or require some_module:('full_path.pm6'). 03:31
it is best that require can also use a precompiled file like the import in python for the performance issue, also it is good if it support data files too, such as json, image etc resource files. 03:32
ugexe using a path works 03:34
zwu I may have to explicit import the symbols, looks like the error is the compiling; 03:35
ugexe and require will use a precompiled module but if you give it a .pm6 file instead of a name it will use the .pm6 code
zwu can I just use the fullpath except the extension? 03:36
ugexe no 03:37
if your plugins follow normal perl naming convention you can determine the file path from the name, and pass the beginning part of it via -I or PERL6LIB or use lib 03:38
so My::Plugin in /home/projectX/Plugins/lib/My/Plugin.pm6; -> -I/home/projectX/Plugins/lib -e 'require My::Plugin;' 03:39
then it will precompile the code and use that the next time
sortiz Btw, earlier I found that -I don't seems to work in rakudo REPL. 03:40
ugexe nope
sortiz Why?
ugexe in REPL do `use lib 'xxx'` 03:41
id guess it happens too late
i dont think its on purpose, just a known bug
sortiz Yes, that and PERL6LIB works, but take me hours!
zwu it think your plugin is still using the searching path, since I know already the location, it is better to load without searching in case name collisions. 03:43
ugexe but searching can find the precompiled code. and it doesnt really do much searching... it just breaks the name into path parts 03:44
timotimo don't forget that importation (that is, making symbols available to the lexical environment) can only work if you're running the importation at compile time, as lexical symbols are fixed after compile time.
zwu I changed from the need to require the file, the output is "Could not find symbol", while using need is fine.
timotimo yeah, need is compile-time, require is run-time 03:45
but you can pass a list of symbols you expect to get from the module after the require
require Foo::Bar, <&a-sub $some-var>
zwu I tried to use: require 'full_path.pm6' < module_name::Class_name>, or just: require 'full_path.pm6' <Class_name>, 03:47
03:47 bpmedley left
timotimo i don't have much experience with the whole importation business, sadly 03:47
ugexe you can also do my `require whatever; my $class = ::("Class_name").new; $class.some-method` 03:49
s/my//
linuxuser9000 I'm reading the documentation for metacpan.org/pod/DBIx::Class, and it seems neat. I've never used a database abstraction layer; I always use wrappers that let me just send SQL queries to mySQL. I feel like I shouldn't use my time to learn a wrapper, but I have to ask, what are your guys' opinions on DB wrappers like DBIx::Class?
timotimo SQLAlchemy from python is really awesome 03:50
ugexe ORMs are something you love or hate
zwu in a module file: module A { sub foo is default { say "foo" }}; in a script, I use 'require 'file_path.pm6' <&foo>"; foo; " the output is "trying to import symbols &foo from ..., but it doesn't export anything in block <unit> ... 03:51
timotimo you mean "is export", right?
anyway, i've got to go to sleep 03:52
zwu yes, "is export"
timotimo good night everyone!
maybe try "our sub foo is export"?
o/
zwu it is actually our sub foo in my test code.
same error.
sortiz Good night timotimo. 03:53
03:53 regreg left
zwu Good night timotimo 03:53
linuxuser9000 @ugexe: So it isn't "bad practice" if I prefer to stick to modules/packages that simply let me send queries and read results? 03:56
I ask because I'm an amateur. I only have 2 years of experience as a programmer
awwaiid linuxuser9000: the path to ORM usually starts when you are dynamically generating the queries, in my experience 03:57
linuxuser9000: then you might find yourself concatenating strings, and then building tools to do that, and then encapsalating the results into domain objects, and then adding more dynamic to infer relationships, and then you built an ORM 03:58
ugexe im a barely functioning sql person myself. but the series people either like ORMs or they like stored procedures
awwaiid linuxuser9000: er, concatenating strings containing SQL queries, that is
03:59 noganex_ joined 04:00 joydo left 04:02 noganex left 04:03 kid51 left 04:04 deepika left
sortiz zmu: Found that in rakudo much of import is NYI, see: github.com/perl6/roast/blob/master....t#L22-L34 04:08
04:09 Herby_ joined
Herby_ Evening, everyone! 04:09
04:15 sortiz left
zwu Thanks. What is NYI? 04:15
geekosaur "not yet impleented" 04:17
*implemented
zwu ah, thanks!
have these assigned? i'd like to contribute but I just started to explore perl 6. 04:18
good night! 04:29
04:29 zwu left 04:30 BenGoldberg left
Juerd Weird bug: "Cannot find method 'map': no method cache and no .^find_method" 04:35
m: my $year = 2016; .say for (Date.new("$year-01-01") .. Date.new("$year-12-31")).grep(*.day-of-week == 5) . classify(*.month).map(*.value.tail).sort; 04:36
camelia rakudo-moar d7f3f0: OUTPUT«Cannot find method 'map': no method cache and no .^find_method␤ in block <unit> at /tmp/CHFa2xkFdL line 1␤␤»
Juerd Without sort it works:
m: my $year = 2016; .say for (Date.new("$year-01-01") .. Date.new("$year-12-31")).grep(*.day-of-week == 5) . classify(*.month).map(*.value.tail);
camelia rakudo-moar d7f3f0: OUTPUT«(2016-05-27)␤(2016-12-30)␤(2016-09-30)␤(2016-07-29)␤(2016-11-25)␤(2016-08-26)␤(2016-04-29)␤(2016-03-25)␤(2016-06-24)␤(2016-01-29)␤(2016-10-28)␤(2016-02-26)␤»
Juerd Am I doing something weird or is this a bug?
dalek c: f86c26d | skids++ | doc/Language/list.pod:
Second pass at list-centric page
 04:44
Herby_ wish I could help you Juerd but I don't have a strong grasp on the language yet 04:46
Juerd I'm off to bed. Maybe when I wake up someone knows the answer :) 04:48
Good night! 04:49
Herby_ night! 04:51
m: $a = "hello"; my @count = $a ~~ / <[aeiou]> :g/; say @count.elems 04:53
camelia rakudo-moar d7f3f0: OUTPUT«5===SORRY!5=== Error while compiling /tmp/ZzNEiz1nb_␤Variable '$a' is not declared␤at /tmp/ZzNEiz1nb_:1␤------> 3<BOL>7⏏5$a = "hello"; my @count = $a ~~ / <[aeio␤»
Herby_ m: $a = "hello"; my @count = $a ~~ / <[aeiou]>/:g ; say @count.elems
camelia rakudo-moar d7f3f0: OUTPUT«5===SORRY!5=== Error while compiling /tmp/BRyPke4QEa␤Variable '$a' is not declared␤at /tmp/BRyPke4QEa:1␤------> 3<BOL>7⏏5$a = "hello"; my @count = $a ~~ / <[aeio␤»
Herby_ how would I go about counting the number of vowels in a string?
m: my $a = "hello"; my @count = $a ~~ / <[aeiou]>/:g ; say @count.elems
camelia rakudo-moar d7f3f0: OUTPUT«5===SORRY!5=== Error while compiling /tmp/oScpY8_t5N␤You can't adverb ␤at /tmp/oScpY8_t5N:1␤------> 3hello"; my @count = $a ~~ / <[aeiou]>/:g7⏏5 ; say @count.elems␤ expecting any of:␤ pair value␤»
skids m: my $a = "hello"; my @count = $a ~~ m:g/ <[aeiou]>/ ; say @count.elems 04:55
camelia rakudo-moar d7f3f0: OUTPUT«2␤»
Herby_ m: my $a = "hello"; my @count = $a ~~ m:g / <[aeiou]> / ; say @count.elems
camelia rakudo-moar d7f3f0: OUTPUT«2␤»
ugexe also $string.comb(/a|e|i|o|u/).elems
Herby_ m: my $string = "hello world"; $string.comb(/a|e|i|o|u/).elems.say; 04:56
camelia rakudo-moar d7f3f0: OUTPUT«3␤»
Herby_ sweet, thanks
m: my $a = "hello 123"; my @count = $a ~~ m:g / <[aeiou]> <[1..4]> / ; say @count.elems 04:57
camelia rakudo-moar d7f3f0: OUTPUT«0␤»
Herby_ m: my $a = "hello 123"; my @count = $a ~~ m:g / <[aeiou]> | <[1..4]> / ; say @count.elems
camelia rakudo-moar d7f3f0: OUTPUT«5␤»
Juerd Hey, "E" and "é" are vowels too! 04:58
04:58 zemmihates left
Herby_ :) 04:58
m: my $a = "HELLO"; my @count = $a ~~ m:g / <[aeiou]> | <[1..4]> / ; say @count.elems 04:59
camelia rakudo-moar d7f3f0: OUTPUT«0␤»
Herby_ m: my $a = "HELLO"; my @count = $a ~~ m:g:i / <[aeiou]> | <[1..4]> / ; say @count.elems
camelia rakudo-moar d7f3f0: OUTPUT«2␤»
04:59 bpmedley joined
ugexe m: my $string = "hEllo"; say $string.comb(/[:i a|e|i|o|u]/).elems 05:00
camelia rakudo-moar d7f3f0: OUTPUT«2␤»
05:00 bpmedley left
Herby_ m: my $string = "count words in string"; $string.comb(/ ' ' /).elems.say; 05:01
camelia rakudo-moar d7f3f0: OUTPUT«3␤»
05:01 bpmedley joined
ugexe m: my $string = "one two three"; say $string.words.elems 05:01
camelia rakudo-moar d7f3f0: OUTPUT«3␤»
Juerd m: "hElló".comb(/:i :m [a|e|i|o|u]/).elems.say 05:02
camelia rakudo-moar d7f3f0: OUTPUT«2␤»
Juerd m: "hElló".comb(/:i :m <[aeiou]>/).elems.say # bug
camelia rakudo-moar d7f3f0: OUTPUT«===SORRY!===␤chr codepoint cannot be negative␤»
Juerd (rakudobugged)
Herby_ ugexe, that is pretty slick 05:03
Juerd Okay, going to sleep for real now :)
Herby_ night!
05:04 pjscott joined
Herby_ m: "here is another string \n test".words.elems.say 05:04
camelia rakudo-moar d7f3f0: OUTPUT«5␤»
ugexe m: "here is another string \n test".lines.elems.say 05:05
camelia rakudo-moar d7f3f0: OUTPUT«2␤»
Herby_ m: "here is a test \n".lines.elems.say 05:06
camelia rakudo-moar d7f3f0: OUTPUT«1␤»
05:09 pjscott left 05:13 joydo joined, yurivish_ left 05:16 pjscott joined 05:19 pierrot joined 05:24 jack_rabbit_ is now known as jack_rabbit 05:26 khw left 05:27 pjscott left 05:28 shadowpaste left 05:30 aindilis left, aindilis joined 05:33 pjscott joined 05:34 Cabanossi left 05:35 Herby_ left 05:37 Cabanossi joined 05:40 pjscott left 05:43 agentzh joined, agentzh left, agentzh joined 05:44 agentzh left, agentzh joined, agentzh left, agentzh joined, bjz joined
awwaiid Now I have this working: use Inline::Ruby::Sweet; use csv:from<Ruby>; say ~CSV.parse("demo.csv")[1][1] #=> Andy 05:45
05:45 bjz left, agentzh left, agentzh joined, agentzh left, agentzh joined, agentzh left 05:46 agentzh joined, agentzh left, agentzh joined
skids awwaiid++ 05:49
05:49 psy_ left, yurivish joined
awwaiid it looks a bit funny though if you get a ruby-JSON object and then do like $data[0]["children"][1]["depth"] :) 05:50
[] gets passed into ruby as method invocation, so that is like $data.send("[]", 0).send("[]","children")... 05:51
I wish I could specify the branch in ecosystem -- I want to make the "develop" branch the default on my libraries, and "master" to be what is released. But I think panda pulls from the default branch 05:57
ugexe panda pulls from whatever source-url links to 05:58
awwaiid I tried embedding a branch in there a while back and it didn't work. Yeah... I should try again
or maybe I should just leave master default and give instructions to would-be contributors to PR against something else. hmmm. 05:59
so many choices!
time for bed :)
06:16 jameslen_ joined 06:18 adhoc left 06:20 adhoc joined 06:35 Zoffix joined, pjscott joined 06:40 CIAvash joined
skids m: my int $x; $x = Nil; 06:52
camelia rakudo-moar d7f3f0: OUTPUT«Cannot unbox a type object␤ in block <unit> at /tmp/kuKiLJejXe line 1␤␤»
skids wonders why it was decided that Nil does not assign the default value of the native type. 06:53
06:56 pjscott left 07:00 keix left
DrForr .tell jnthn Inconistency that probably is out of scope for Grammar::{Debugger,Tracer} - once the TOP rule has matched, if there is more text to be matched by the parse('foo') method it's not reported, so Grammar::Tracer could conceivably report a MATCH when the parse() still fails. 07:02
yoleaux DrForr: I'll pass your message to jnthn.
07:10 Tonik joined 07:14 zengargoyle left 07:18 abraxxa joined
Quom__ skids: I doubt decided — native types are known to be troublesome… :9 07:19
07:19 ChoHag left 07:22 abraxxa left 07:26 uruwi left 07:27 uruwi joined 07:35 jeek left 07:36 abraxxa joined 07:41 jeek joined 07:47 joydo left 07:50 leat left 07:51 schnuppi joined 07:53 leat joined 07:55 daotoad joined
daotoad Anyone have any thoughts on free and/or cheap Perl6 web app hosting? 07:57
I looked at heroku, but the buildpack is horribly out of date.github.com/pnu/heroku-buildpack-rakudo 07:58
DrForr Linode or AWS and roll your own?
There's a Docker buid.
*build
08:01 kaare_ joined 08:04 abraxxa left 08:15 daotoad left, daotoad joined, bpetering joined
bpetering .tell Juerd I'm going through /^F/, skip to /^G/ please :) 08:17
yoleaux bpetering: I'll pass your message to Juerd.
08:17 wamba joined 08:20 rindolf joined 08:23 darutoko joined 08:24 CIAvash left 08:25 CIAvash joined 08:42 bpetering left 08:47 spider-mario joined 08:50 obra left
dalek c: 2b3c920 | skids++ | doc/Type/Nil.pod:
Bring page on Nil into sync with post-GLR reality.
 08:50
DrForr Is it just me, or is s/// not documented in the regex page? Or anywhere really, for that matter? 08:54
(ignoring the synopses/apocalypses, because someone coming in from outside won't know about those...) 08:55
skids Well, it is in 5to6 pages. But... well vounteered! :-) 08:56
skids goes to sleep off his cold. 08:57
08:57 skids left 09:09 citizenX joined 09:10 llfourn joined 09:12 lnrdo joined 09:13 hankache joined
hankache good morning everyone 09:17
09:19 sno left 09:20 sno joined 09:25 firstdayonthejob left 09:30 Ch0c0late joined
masak good antenoon, #perl6 09:30
hankache hi masak 09:33
09:34 vendethiel joined
DrForr m: my $x="foo\nbar"; $x~~s:g/\n/\c[2424]/; say $x 09:35
camelia rakudo-moar d7f3f0: OUTPUT«fooॸbar␤»
DrForr Oh, wait, what's 2424 in hex... 09:36
Ah, that's it. 09:37
09:42 daotoad left 09:43 virtualsue joined
DrForr Well, feel free to mock #127220 roundly as you see fit in that case. RT doesn't allow me to see the ticket so I can't comment on it. 09:43
pnu daotoad: you can "echo 2015.12.152.gd.7.f.3.f.0.b >.rakudo-version" (and commit that to git) to get the latest version. I have daily automated builds but the "latest" version is updated only when spectests pass.. so latest "latest" is 2015.11.100.g.1.ae.6805 at the moment. 09:47
[Tux] csv-ip5xs 50000 18.278 18.166 09:48
test 50000 24.503 24.390
test-t 50000 12.645 12.533
csv-parser 50000 50.714 50.602
09:48 _nadim left 09:49 _nadim joined
pnu daotoad: However, I'm updating and integrating the perl6 support to my heroku-buildpack-perl. Just having some trouble with relocatable bins, to get it cleanly done. After that is done, a separate rakudo buidpack is not needed. 09:51
09:54 keix joined
hankache how can i make tests stop if one of them fail ? 09:54
i have 3 tests and if number 2 fails, i do not want to run number 3
09:55 keix left
pnu daotoad: err.. "latest" is updated when spectest *and* task::star installation pass (github.com/pnu/heroku-buildpack-ra...ld.sh#L77) 09:55
nine .tell awwaiid nice blog post! Actually...I do recommend to people to use Perl 5's Inline::Python even for critical production code. I can recommend this in good concience because that's what we do.
yoleaux nine: I'll pass your message to awwaiid.
09:56 wamba left 09:57 daotoad joined, vendethiel left 09:58 keix joined 10:03 TEttinger left 10:04 spider-mario left 10:05 vendethiel joined 10:06 linuxuser9000 left 10:08 RabidGravy joined 10:09 abcxyz joined
RabidGravy morning 10:11
yoleaux 01:13Z <sena_kun> RabidGravy: gist.github.com/Altai-man/d2fb78521fe2eb62ae73 from sena_kun.
10:13 abcxyz left 10:14 daotoad left, FROGGS_ joined 10:17 daotoad joined 10:18 leat left, citizenX left, FROGGS left, leat joined 10:31 fdsfdsfds joined
fdsfdsfds hi, how can I log in Perl6? By log I mean writing stuff to a file for debugging 10:32
not authenticating
I haven't found any module
Quom__ … modules.perl6.org/#q=log 10:33
Though I haven't used any of these myself 10:34
masak .oO( lmmpotfy )
fdsfdsfds thanks 10:36
10:36 daotoad left
Quom__ np :) 10:37
10:37 fdsfdsfds left
hankache how can i stop testing when one of the tests fail? 10:37
10:43 fdsfdsfds joined
fdsfdsfds could you tell me where I can find an instruction of how to submit my own perl 6 module to modules.perl6.org/ ? 10:43
RabidGravy hankache, there's a die-on-fail in Test but it doesn't seem to be exporter 10:44
DrForr fdsfdsfds: There's a link on the front page, see 'this guide'. 10:45
RabidGravy fdsfdsfds, see doc.perl6.org/language/modules
fdsfdsfds tnx
10:45 fdsfdsfds left
RabidGravy .tell sena_kun I added a working example to your gist, it doesn't fail but I don't have the bzip2 dev packages to check it is doing the right thing 10:47
yoleaux RabidGravy: I'll pass your message to sena_kun.
dalek c: a17ff0d | (Siavash Askari Nasr)++ | doc/Language/unicode_entry.pod:
Mention TeX input method for entering unicode characters in Emacs
 10:49
c: 176bf8a | (Siavash Askari Nasr)++ | doc/Language/unicode_entry.pod:
Merge pull request #333 from CIAvash/unicode_entry_emacs_tex_input_method

Mention TeX input method for entering unicode characters in Emacs
10:55 citizenX joined 10:56 glucosade left, spider-mario joined 11:00 citizenX left 11:01 domidumont joined
dalek ast: c435912 | lizmat++ | S02-types/int-uint.t:
Fix test, fix indenting, Juerd++
 11:01
11:03 ely-se joined 11:04 sena_kun joined 11:06 domidumont left, domidumont joined
bartolin_ hi, #perl6 -- and belated congrats for the 6.c release! 11:07
bartolin_ was away for a while
nine Hi bartolin_! 11:08
moritz welcome back bartolin_
bartolin_ \o/ 11:09
RabidGravy rarr!
bartolin_ it looks like the recent nqp bump broke the build for nqp-j: gist.github.com/usev6/057d64b0fc9b3c830c2f 11:10
the culprint seems to be nqp commit 4b1236e535
lizmat that's probably because JVM doesn't have readlinechompfh yet ? 11:13
funrep just read perl6intro, is there any tutorial that follows you how to build a program of sort? 11:14
ely-se is the GLR still ongoing?
11:15 frankjh joined
RabidGravy ely-se, I don't think so 11:16
nine ely-se: the GLR has been pretty much done since September
ely-se nice :)
is there a change list of it some where?
llfourn ely-se: check out perl6advent.wordpress.com/2015/12/...-refactor/ 11:22
ely-se thanks!
DrForr funrep: There's a 'perl 6 in X days' guide out there, I don't remember the exact URL. 11:26
funrep DrForr: thanks! i found this slide: jnthn.net/papers/2015-spw-perl6-course.pdf seems pretty good i gonna go through it 11:27
p6: say 2; 11:32
camelia rakudo-moar d7f3f0: OUTPUT«2␤»
11:36 AlexDaniel joined
nine panda now actually checks a dist's Perl 6 version requirement and errors out if it cannot be met. This should make it much easier to diagnore installation issues caused by the user having a too old Rakudo. 11:38
bartolin_ lizmat: there is a readlinechompfh for JVM. jnthn++ introduced it with 05d7b20f, but the commit message states that it's "cheating". maybe it gets caught now :-/ 11:39
lizmat bartolin_: that may be the case :-(
11:39 ely-se left
nine bartolin_: maybe it isn't readlinechompfh but stat_time and lstat_time? 11:40
The patch did not change any JVM related files
bartolin_ nine: I was able to build rakudo-j with 4b1236e535 reverted (actually I added an '#?if jvm' case) 11:41
but that doesn't look like a nice solution ...
nine Ah, ok
11:41 mr-foobar left
CIAvash Is there a reason '≠', '≤' and '≥' are not used for '!=', '<=' and '>='? 11:44
funrep what is an IntStr?
lizmat CIAvash: you would have to ask TimToady that 11:45
nine A data type that can act like an Int and a Str. It's purpose is for example for command line arguments "foo.pl 3" will pass a "3".IntStr to the MAIN function which can then decide if the argument should be an Int or a Str 11:46
lizmat personally, I would be in favour of adding them, simply because things like × and / also got added
funrep okey make sense
another question, i assume arrays are heterogeneous, is that correct? 11:47
nine yes
lizmat native arrays aren't ?
funrep alright
nine untyped arrays anyway
RabidGravy m: my $a = IntStr.new(10, "ten"); say +$a; say ~$a;
camelia rakudo-moar d7f3f0: OUTPUT«10␤ten␤»
funrep so i can optionally make them homogenous?
same for lists or are lists same as arrays? maybe i should read more before asking qusetions :P 11:48
lizmat funrep: sorry, misundersttood earlier
technically, all arrays are homogenous
and by default Any
funrep aah
okey
lizmat since Any accepts just about anything, it looks like they're heterogenous 11:49
nine Like Object[] in Java
funrep (i noticed perl added (any) objects to my list when i added something out of range)
lizmat but e.g., they will not accept Mu or Junctions, as they don't inherit from Any
hankache can we catch the result of a test without printing it to the terminal ?
funrep array*
lizmat hankache: you can change $*OUT/$*ERR before the test ? 11:50
funrep okey, thanks for the replies!
CIAvash lizmat: ok. Whenever I ask TimToady something, he disappears :) maybe I should .tell him.
lizmat CIAvash: that might help 11:51
hankache lizmat can you give me an example?
i want to print to text file
lizmat in any case, I don't think we'll be changing much before the next release
AlexDaniel CIAvash: haha! Ten minutes ago I was thinking about asking the same question 11:52
RabidGravy hankache, there are "our" scoped subs to control the output: Test::output(), Test::failure_output() etc, you can redirect the output, and all the test functions return a boolean to indicate whther they passed or not
AlexDaniel CIAvash: but hey, I've mentioned tha a couple of times already
CIAvash: so generally the question is “how much unicode stuff do we really want in Perl 6?” 11:53
CIAvash: for example, you can use ∞ instead of Inf now, but it was added just a month ago or so 11:54
RabidGravy as long is there an alternative I'm cool with it, some of these characters are hard to type
AlexDaniel CIAvash: and × ÷ operators are rather new
11:54 pjscott joined
CIAvash AlexDaniel: oh! :) So I shouldn't ask him again? 11:55
AlexDaniel CIAvash: why not? 11:56
dalek c: 244bd51 | lizmat++ | doc/Language/unicode_texas.pod:
Document − and … and texas - and ...
c: 2ab3ad6 | lizmat++ | doc/Language/unicode_texas.pod:
Add ÷ , AlexDaniel++ for making me remember
timotimo Juerd: i think the exception you got when you put .sort in your code comes from inside the sort code, which uses a map call to turn the list of indices it sorted back into elements from the original list; perhaps somehow the sort sub returned NQPMu instead of throwing an error in some way or other 11:57
11:57 Skarsnik joined
CIAvash AlexDaniel: If it was already disccused 11:57
Skarsnik hi
AlexDaniel CIAvash: ah, also a fairly new addition is stuff like ½ and superscripts
CIAvash: there was no final decision whether we should add ≤ ≥ ≠ or not 11:58
RabidGravy It surprises me how many of these things are available via "Compose Key"
AlexDaniel .u / 11:59
yoleaux U+002F SOLIDUS [Po] (/)
AlexDaniel .u /
yoleaux U+002F SOLIDUS [Po] (/)
CIAvash AlexDaniel: Yeah, I know about the characters you mentioned. That's why I thought why aren't these characters available. 12:00
DrForr There are Unicode lookalikes up above 0x7f :)
AlexDaniel .u // 12:02
yoleaux U+002F SOLIDUS [Po] (/)
AlexDaniel lizmat: any idea why / is listed?
CIAvash .tell TimToady Is there a reason '≠', '≤' and '≥' are not used for '!=', '<=' and
yoleaux CIAvash: I'll pass your message to TimToady.
CIAvash '>='?
lizmat AlexDaniel: no idea 12:03
AlexDaniel lizmat: perhaps somebody was going to add ÷ but wrote a regular slash instead
CIAvash I was surprised to see / too. 12:04
12:04 Ch0c0late left
AlexDaniel CIAvash: there are also things like ≔ ⩵ ⩶ 12:06
CIAvash: though if the first two are more or less OK, the third one is just nonsense when you think about monospace fonts 12:07
CIAvash: also, if we add these ops then there is a high chance that some people (like me) will start using those extensively. Imagine people on the internet complaining that the code is not just “write-only”, but now you can't even type it. 12:09
Like, I have no problems entering common unicode chars, why would I want to use <= >= != then? 12:10
RabidGravy well, for most people it would be more key strokes 12:12
12:12 kid51 joined
RabidGravy ≤ is one more stroke than <= for me 12:12
AlexDaniel sure, but for me it would be less keystrokes than ascii equivalents… 12:13
CIAvash AlexDaniel: Yeah, I think someone mentioned that before.
RabidGravy and that's if I can remember how to type it
huf some people were burned by mojibake and broken encoding in their youth so much that they'll never accept non-ascii operators :)
CIAvash Well, it depends. You can define your own keys too, I think 12:14
AlexDaniel CIAvash: one other thing to note is that « » … and maybe even ｢｣ already exist on many keyboard layouts 12:16
(I'm not sure about ｢｣ though, it seems like these are not the ones that are used in Japan) 12:17
RabidGravy but yeah, the point about it becoming thought of as an un-typable langiage is something to bear in mind
hankache we need a Perl 6 keyboard
RabidGravy but it comes down to self control
AlexDaniel CIAvash: superscripts are also typable most of the time 12:18
12:18 _nadim left
DrForr votes for foot pedals. 12:18
RabidGravy hankache, look how well that went with APL
AlexDaniel hankache: keyboard layout you mean?
DrForr: any cheap one that you could recommend?
hankache yup
huf always have ascii pairs of unicode things and ship a tool that converts source files?
and then nobody has to ever care. 12:19
AlexDaniel huf: hm, I've heard that idea a couple of times
DrForr Just adopt trigraphs and be done with it :)
timotimo uargh
huf DrForr: at this point, ascii has become "trigraphs"
the poor dear
DrForr I was thinking of the equivalents for {} and [] :) 12:20
AlexDaniel CIAvash: ah also, while you are at it, perhaps you can find some good symbol for whatever star
dalek c: 3f2cbc0 | lizmat++ | doc/Language/unicode_texas.pod:
Rework it a bit
 12:21
AlexDaniel CIAvash: ⍰ is very close, but I'm sure that there are better symbols
funrep what capitalization convention is most common in perl6 code?
timotimo usually, we have ClassName, method-name, sub-name, CONSTANT_NAME, INTERNAL-NAME, right? 12:22
12:22 _nadim joined
huf _texas_? 12:22
timotimo yeah
everything's bigger in texas
» is >>, for example
huf oh god... :D 12:23
AlexDaniel though not always: × and *
timotimo m: say "×".encode.perl
camelia rakudo-moar d7f3f0: OUTPUT«utf8.new(195, 151)␤»
timotimo m: say "*".encode.perl
camelia rakudo-moar d7f3f0: OUTPUT«utf8.new(42)␤»
AlexDaniel oh, in that sense…
timotimo i disagree :P
AlexDaniel good point
indeed, everything is bigger… 12:24
timotimo i don't think that's not meant to mean that, though
AlexDaniel funrep: one interesting difference between perl6 and other languages is that we have kebab-case
12:24 leont joined
AlexDaniel m: say 5 12:25
camelia rakudo-moar d7f3f0: OUTPUT«5␤»
AlexDaniel oh wow, did not expect that to work
lizmat: consider mentioning non-break space and perhaps other whitespace characters? 12:26
lizmat AlexDaniel: I'll take PR's :-)
AlexDaniel lizmat: well, in that case I can just commit it myself :D
lizmat please do! :-) 12:27
nine panda will now fall back to curl and then to wget if downloading meta data failed. This should fix the proxy issues.
.tell [Coke] panda will now fall back to curl and then to wget if downloading meta data failed. This should fix the proxy issues. 12:28
yoleaux nine: I'll pass your message to [Coke].
dalek c: 0b3deca | lizmat++ | doc/Language/glossary.pod:
Add some more lemma's
timotimo nice
12:29 sena_san joined, sena_san left
lizmat timotimo: wouldn't mind seeing someone give a better description of QAST :-) 12:29
nine .tell todd_ panda will now fall back to curl and then to wget if downloading meta data failed. This should fix the proxy issues.
yoleaux nine: I'll pass your message to todd_.
timotimo oh
lizmat bootstraps panda 12:30
timotimo =head1 QAST
Successor to L<#QAST>.
we're stuck in an infinite vortex of improvement
lizmat oops 12:31
timotimo: perhaps you can improve ?
12:31 sena_kun left
timotimo it'll be hard to find just the right amount of text to put there 12:31
12:31 _nadim left
timotimo it shouldn't be much 12:31
Skarsnik m: module A { our $var := "hello"; }; use A;
camelia rakudo-moar d7f3f0: OUTPUT«===SORRY!===␤Could not find A in:␤ /home/camelia/rakudo-m-inst-1/share/perl6/site␤ /home/camelia/rakudo-m-inst-1/share/perl6/vendor␤ /home/camelia/rakudo-m-inst-1/share/perl6␤ CompUnit::Repository::AbsolutePath<140519385066968>␤ …»
12:31 _nadim joined
lizmat m: module A { our $var := "hello"; }; need A; 12:32
camelia rakudo-moar d7f3f0: OUTPUT«===SORRY!===␤Could not find A in:␤ /home/camelia/rakudo-m-inst-1/share/perl6/site␤ /home/camelia/rakudo-m-inst-1/share/perl6/vendor␤ /home/camelia/rakudo-m-inst-1/share/perl6␤ CompUnit::Repository::AbsolutePath<140014219795600>␤ …»
Skarsnik m: module A { our $var := "hello"; }; import A;
camelia ( no output )
CIAvash lizmat: So it's ok to push directly instead of submitting pull requests?
lizmat Skarsnik++
RabidGravy frankly I'm waiting for someone to do "package EXPORTHOW { package DECLARE { our constant कामधेनु = Metamodel::ClassHOW }};" ... "कामधेनु Foo { }; say Foo.new" ;-)
lizmat CIAvash: if you have a commit bit, yes
Skarsnik m: module A { our $var is export := "hello"; }; import A;
camelia ( no output )
Skarsnik ho it work now?
to export binded var
lizmat m: module A { our $var is export := "hello"; say $var }; import A; say $var # looks like a bug 12:33
camelia rakudo-moar d7f3f0: OUTPUT«hello␤(Any)␤»
lizmat the export succeeded, cause it has $var in the outer scope now 12:34
m: module A { our $var is export = "hello"; say $var }; import A; say $var # seems binding it creates the problem
camelia rakudo-moar d7f3f0: OUTPUT«hello␤hello␤»
AlexDaniel .u say 5
yoleaux U+0035 DIGIT FIVE [Nd] (5)
U+0061 LATIN SMALL LETTER A [Ll] (a)
U+0073 LATIN SMALL LETTER S [Ll] (s)
AlexDaniel m: say 5
camelia rakudo-moar d7f3f0: OUTPUT«5␤»
12:37 espadrine joined
AlexDaniel m: print ‘say’; print join ‘’, (0..0x1FFFF ==> grep { .uniprop ~~ m/‘Zs’/ })».chr; say 42 12:37
camelia rakudo-moar d7f3f0: OUTPUT«say                　42␤» 12:38
AlexDaniel m: say                　42
camelia rakudo-moar d7f3f0: OUTPUT«42␤»
dalek c: 378c3e9 | (Aleks-Daniel Jakimenko-Aleksejev)++ | doc/Language/unicode_texas.pod:
Mention other whitespace characters (Zs property)
 12:41
12:42 kid51 left
AlexDaniel I'm not sure if there are any other things besides Zs though 12:42
lizmat AlexDaniel++ 12:43
12:43 kid51 joined
AlexDaniel oh 12:44
m: print ‘say’; print join ‘’, (0..0x1FFFF ==> grep { .uniprop ~~ m/‘Zl’/ })».chr; say 42
m: print ‘say’; print join ‘’, (0..0x1FFFF ==> grep { .uniprop ~~ m/‘Zp’/ })».chr; say 42
camelia rakudo-moar d7f3f0: OUTPUT«say
42␤»
rakudo-moar d7f3f0: OUTPUT«say
42␤»
12:45 pjscott left 12:46 pjscott joined 12:47 brrt joined
AlexDaniel m: say
5 12:48
camelia rakudo-moar d7f3f0: OUTPUT«5␤»
AlexDaniel ha
m: say
5 12:49
camelia rakudo-moar d7f3f0: OUTPUT«5␤»
AlexDaniel ok, I'm wrong
dalek c: 49a5110 | (Aleks-Daniel Jakimenko-Aleksejev)++ | doc/Language/unicode_texas.pod:
Zl and Zp are also valid whitespace characters
 12:50
kudo/nom: c0bb692 | lizmat++ | src/core/io_operators.pm:
Remove non-functioning, untested pipe(), cygx++

Part of PR #646, the other part I'm not happy merging at this point
Skarsnik m: my %h; %h<b><id> = 2; %h<a><id> = 1; say %h.values.sort(*<id>); 12:57
camelia rakudo-moar d7f3f0: OUTPUT«(id => 1 id => 2)␤»
Skarsnik m: my %h; %h<b><id> = '_2'; %h<a><id> = '_1'; say %h.values.sort(*<id>);
camelia rakudo-moar d7f3f0: OUTPUT«(id => _1 id => _2)␤»
Skarsnik hm that does not work for me. is that because it can't sort directly _xx ? 12:58
dalek kudo/nom: 7253482 | lizmat++ | src/core/IO.pm:
Remove some superfluous returns
 13:06
13:06 shadowpaste joined
lizmat Skarsnik: not sure what you're getting at 13:07
Skarsnik I want to sort on the value of the id, but I get it work with {$^a<id> > $^b<id>} x) 13:09
lizmat and it doesn't with *<id> ? 13:13
13:13 lnrdo left
lizmat or { $_<id> } ? 13:13
or { $^a<id> } ?
I guess the former can't work because it doesn't indicate the signature
if the signature's arity is 1 (which just using $^a will do), then sort should do TRT 13:14
AlexDaniel TRT?
lizmat The Right Thing 13:15
nine lizmat: have you given further thought to the language_versions branch?
lizmat nine: I have, and I think it's a bad idea
13:15 hankache left
nine :( 13:16
lizmat I don't think we can make it acceptable to module developers or users to tell them that if a feature isn't tested in roast
nine So we need another idea or some really good arguments.
lizmat it doesn't exist
if it works, it works
nine But...we wouldn't have to?
lizmat yes, we would, e.g. in the case of Date.new taking 3 positionals 13:17
it wasn't tested in 3.c
it was added during development of 2016.01
use v6.c will *not* tell you whether it will work
nine I'd say exactly the opposite: we leave the existing code alone as much as possible (except for real bug fixes or performance improvements) by moving new additions to src/core.d/
Developers who want to use the additions of core.d before it's released will have to use v6.d.PREVIEW with the same caveats as pre 6.c nom: you may use this, but it may go away before the release. 13:19
Skarsnik curse you union. you make my life a nightmare!
lizmat nine: and what is the position of nom in this ?
nine Same as it has always been. We do development on it. It's just that changes to the API go into a different directory and have to be written using augment and supersede and friends. 13:21
13:21 virtualsue left
nine Changes that don't touch semantics, like plain performance improvements or fixes to bugs go to core code same as ever (except if they break the 6.c tests in roast) 13:21
13:22 pjscott left
Skarsnik you realise you will need version on the doc too :( 13:22
when 6.c is not even fully documented x)
lizmat nine: I like the idea, but practically, it's going to be a nightmare 13:23
13:23 virtualsue joined
nine Skarsnik: yes, of course. But the only alternative is to freeze all APIs and not have any future language versions at all :) 13:23
lizmat my idea would be:
nine lizmat: backwards compatibility always is. The question is: how can we make the nightmare most bearable.
lizmat 6.c / 2015.12 is frozen in a branch
any work done post-2015.12 (in nom) is part of 6.d.PREVIEW 13:24
any bugfixes / speed improvements are cherry picked in the 6.c branch
doc is frozen in 6.c / 2015.12 branch 13:25
13:25 pjscott joined
Skarsnik I don't want to sound annoying, but for me: Something not documented is not in 6.c :) 13:25
lizmat any additions / changes are for 6.d.PREVIEW
AlexDaniel Skarsnik: not tested you mean?
Skarsnik No, documented
lizmat Skarsnik: but there's the source
AlexDaniel documented as in docs.perl6.org? 13:26
lizmat people *will* use it, especially if it's so P5 centric that nobody thought it worthwhile to document for P6
13:28 brrt left
lizmat but since our spectesting is still only covering a relatively small percentage of the code base 13:28
nine lizmat: if we're using git branches for 6.c/6.d distinction, wouldn't that mean that I'd have to install different rakudo versions if some of my programs are written in 6.c and some are in 6.d? 13:29
lizmat I think we're doing users / module developers a disservice by saying that 2016.01 is 6.c, like 2015.12 is 6.c
nine lizmat: ^^^ that's the reason why I won't add any API method to a rakudo that claims to implement 6.c. Because people will use them regardless of what roast says. And other people will run the code on older versions and it will not work. 13:30
lizmat well, generally, I would think that a feature to be removed in 6.d, should be DEPRECATED at least one language level bump
13:30 CIAvash left
nine lizmat: deprecation is certainly ok. It can throw a deprecation warning in 6.d and vanish in 6.e. The language_version architecture makes this quite easy to do. 13:31
lizmat well, then I guess most of the commits since 2015.12 need to be moved somewhere else then
nine Yes, that's why I'd really like to see a solution soon. Because the work will become more and more :)
But actually I think most of the commits are fine. All the performance improvements and commits like removing a &pipe that didn't work anyway. 13:32
13:33 pjscott left
lizmat well, a large part of the Date/DateTime changes would need to be looked at, as a lot of them depend on the new positional candidates 13:33
13:33 pjscott joined
Skarsnik is there lot of tests that fail in roast with the current nom? 13:34
lizmat we should only have some flappers 13:35
Skarsnik Because I am pretty sure 2015.12 is not that used because of all the panda/precomp issue
nine lizmat: I think for DateTime, we could move the positional constructor to a private method and use that everywhere and add the positional constructor which forwards to the private method in 6.d
13:35 sftf joined
lizmat nine: true, but that would basically make any maintanence on Date/DateTime work in two branches 13:36
and we're going to have more and more of these cases
nine not in two branches but in two files. On the other hand it will be very, very obvious which language version changed what. 13:37
lizmat and we're going to start mistakes with them
*having
13:37 pjscott left
nine Yes, we will make mistakes. Just as with every kind of developing. I don't see a solution that completely avoids that :/ 13:39
orbus not to toss a bomb and then leave (because I have to take off) but it seems like there's some disagreement over whether the language version covers only syntax, or also covers the API for some core set of classes
yoleaux 6 Jan 2016 09:10Z <nine> orbus: The NativeCall test failures are because of issues with passing int8 and uint8 to C and back and because of nativesizeof(SomeStructOrClass) doesn't seem to work correctly. Neither have a high probability of affecting Inline::Perl5, but there may be other issues.
6 Jan 2016 09:28Z <FROGGS> orbus: can I get detailed output of the failing tests? like if you'd run perl6 t/04-nativecall/02-simple-args.t etc
lizmat orbus: no, there's no disagreement there
nine It's always language syntax + core classes, isn't it? :) 13:40
lizmat yep
13:40 pjscott joined
orbus okay 13:40
lizmat It's more about what 6.c means, and how we can make sure that users / module developers can be sure that "use v6.c" is a *FIXED* set of capabilities
rather than a slightly dynamic one depending on which compiler release you're using 13:41
orbus I see
I guess I'm primarily looking at java as a model
lizmat without having to resort to saying "well, it wasn't tested, so booboo for using it anyway"
orbus java has a similar situation where there are multiple implementations and distinct language versions 13:42
13:43 pjscott left
orbus they seem to pretty much freeze the apis for a given version - if they want to add more parameters or more methods or whatever, it waits for a new version 13:43
of course, I have no clue how the code base is managed :p
Skarsnik hm, nativesizeof on a struct work correctly 13:44
orbus .tell FROGGS I'll work on getting you test output when I have some free time - hopefully later today 13:45
yoleaux orbus: I'll pass your message to FROGGS.
orbus I will say, I know I personally would rather rely on what's documented than just some method I found in the source 13:47
lots of software products have undocumented apis 13:48
doesn't mean they are safe to use
13:48 AlexDaniel left
lizmat orbus: sure, but we want to prevent situations like with Coro in Perl 5 13:49
whenever we get a successful module relying on some internal API, you basically cannot break it anyway 13:50
nine lizmat: ensuring that 6.c is a fixed set of capabilities should be a strong point of the language_versions architecture. In reviews we just have to make sure that changes to src/core/ leave semantics alone. Sounds much easier than reviewing code riddled with if $*PERL.version branching (though we will never be able to avoid those completely) 13:51
orbus maybe anything undocumented for a given version should be somehow flagged as such? 13:54
like throw warnings?
nine git branches won't let you support different language versions with the same binary. Let alone mixing different language versions in the same program.
orbus nine: agree 13:55
probably if your aim is to support v6.c, v6.d, etc. in the same binary you're going to have to go back and test later versions of the runtime against each of the old spec tests 13:56
make sure v6.c passes, then v6.d, then v6.e 13:57
etc
[Tux] if I get a Supply, can I ask it its Supplier? 13:58
orbus just looking at the docs - I don't think so 13:59
14:00 lnrdo joined
funrep so i just learned that you can have code inside strings, this begs the question, doesnt this makes it easy to exploit perl6 programs? 14:01
Skarsnik probably
orbus has to afk for a bit
funrep or well, probably there are several types of strings
llfourn funrep: no not at all
funrep: can you suggest an attack vector? 14:02
Skarsnik my $foo = '{say hello}'; "$foo";
m: my $foo = '{say hello}'; "$foo";
camelia ( no output )
14:02 jshy joined, zwu joined
llfourn m: my $foo = '{say hello}'; say "$foo"; 14:02
camelia rakudo-moar 725348: OUTPUT«{say hello}␤»
llfourn that aint gonna work :P 14:03
funrep input in a website, you write "{ rm_rf("../" }" or something
14:03 pmqs_ is now known as pmqs
llfourn funrep: you would need still need to EVAL It 14:03
funrep i started learning this thing yesterday so probably just premature assumption that's not real problems :P
llfourn: isnt application strict though? so if you put that string as an argument it will be evaluated before the function call? 14:04
Skarsnik can we have an operator for defined in 6.d? x) 14:05
funrep well, gotta run!
llfourn funrep: I'm not sure what you mean. "{ }" is just syntactic sugar at runtime it's the same as concatination
Skarsnik I am bored of writting if stuff.defined and stuff < 2;
Zoffix Skarsnik, where's stuff coming from? You may benefit from the :D smiley 14:08
lizmat funrep: only if you use EVAL, which you probably shouldn't
Zoffix (stuff//0) < 2
lizmat funrep: as interpolation works at compile time
Skarsnik Zoffix, I have a ton of code like that, where stuff is various stuff, like if @sorted[$i + 1].defined and @sorted[$i + 1]<obj> ~~ CUnion 14:09
lizmat cycling&
14:10 RabidGravy left
Skarsnik I would like a stuff ? cmpoperator value 14:10
Zoffix +1
llfourn Skarsnik: do you use 'with'? 14:11
with checks for definedness
with @sorted[$u + 1] { say "win" when CUnion } 14:12
Skarsnik that aweful
llfourn I like it :)
14:13 Zoffix left
llfourn with @sorted[$i + 1] { say "win" when .<obj> ~~ CUnion } #or rahter 14:14
14:14 Zoffix joined
Zoffix Seems huggable is leaking. Dies after ~1-2 days. And this time I got a notification from linode about excess disc activity, which I'm guessing is swap 14:15
:( 14:16
Skarsnik NC lead to leak I think
14:16 huggable joined
Zoffix Aw :( I'm using it (indirectly, via DBIish) github.com/zoffixznet/perl6-IRC-Cl...actoid.pm6 14:17
14:17 wamba joined
Skarsnik In dbiish tests we have a weird issue where regular Str receive the explicitlymanaged role of NC and they don't get GC 14:17
Zoffix OTOH, the leak is probably somewhere in the Grammar/Actions github.com/zoffixznet/perl6-IRC-Cl...Parser.pm6 14:18
Because I don't see the bot used much. It's really it just parsing traffic on #perl6
14:18 huggable left
Zoffix Gonna disable the debugger plugin (that uses Data::Dump) to eliminate one variable 14:19
Skarsnik Yes, but if every regular Str get not GC because of a NC call, you are in trouble x)
14:19 huggable joined
Skarsnik Display the ^name of your variable 14:19
Zoffix Skarsnik, there shouldn't be an NC call without DB look up 14:20
I meant the metaphorical variable not literal :)
huggable, IO::Handle
huggable Zoffix, class IO::Handle [Opened file or stream]: doc.perl6.org/type/IO::Handle
Zoffix k, will see if it's still alive come Monday
14:26 hankache joined 14:29 virtualsue left 14:30 DukeOfPerl joined
DukeOfPerl Can Perl6 source code be in plain codee, or must it be Unicode? 14:30
Skarsnik plan codee? 14:31
*plain
DukeOfPerl Ordinary ASCII text as plain code, not Unicode.
DrForr What's the difference? ASCII is the first 0x7f codepoints of UTF-8.
DukeOfPerl But in Unicode one may express logical operators, not usually necessary. 14:32
14:32 _nadim left
Zoffix DukeOfPerl, yes, there are Texas Variants of all Unicode operators 14:32
[Tux] Any other (useful) in or out I could implement for «Text::CSV.csv (in => $in, out => $out)» ? github.com/Tux/CSV/blob/master/lib...CSV.pod#in
DrForr You can express logical operators in ASCII too.
Zoffix DukeOfPerl, doc.perl6.org/language/unicode_texas
DukeOfPerl And is Perl6 only available for Apple macintosh now in January 2016?
DrForr No, it works fine on Windows and Linux as well. Where did you find this information? 14:33
DukeOfPerl When will Perl5 be available for MSOFT Windows or for Linux?
Zoffix DukeOfPerl, Perl 5?
hankache lurkes
DrForr I think you mean Perl 6, and it's already been available for quite a while. 14:34
Zoffix DukeOfPerl, it already is available. For Perl 6, you can use the method described here: perl6.org/downloads/ or you can wait a few weeks and get the Rakudo Star distribution
DukeOfPerl en.wikibooks.org/wiki/Category:Per...rogramming has pretty good info, but old.
DrForr perl6.org has the most recent info. 14:35
14:35 gregf left
DukeOfPerl And at Perl6.org they speak of Rakudo, which means -- what? Japanese? 14:35
Zoffix DukeOfPerl, I can imagine. There've been many recent changes, prior to the first stable release. Check out perl6intro.com/ and doc.perl6.org/
DukeOfPerl, Rakudo is the compiler of the Perl 6 language. It's Japaneese for The Way of The Camel or something It's in the Wikipedia 14:36
14:36 virtualsue joined
DukeOfPerl Has anybuddy yet offered, or pointed to, a Perl6 "killer app"? 14:36
DrForr It's only been released for a few weeks. Give it at least a month for someone to write its killer app. 14:37
llfourn DukeOfPerl: there are many killer Perl6 feautures but no one has built anything in Perl 6 that is a "killer app"....yet :0
14:37 _nadim joined
DukeOfPerl Why does Larry Wall propject that Perl6 will replace Perl5 after FORTY years? 14:37
Zoffix DukeOfPerl, if "killer apps" were something that easy to come up with... :)
DukeOfPerl, because Perl 6 is written for the next 100 years... Perl 5 isn't
DrForr Ask Larry, he'll be around in a little while. 14:38
DukeOfPerl Anybody contemplating a Perl6 "killer app"? [I am]
Zoffix DukeOfPerl, sweet. Start coding.
DukeOfPerl Wait a miunute-- Larry Wall often comes hither???
Zoffix DukeOfPerl, I suspect the Perl 6 Killer App will be an AI :)
DukeOfPerl, yeah, TimToady is Larry Wall.
DrForr Thought is cheap, bytes are worthwhile. 14:39
Zoffix (as in, he's here now under that nick :))
DukeOfPerl ai.neocities.org/AiSteps.html is my Perl5 AI app. (So who am I really? -)
14:39 FROGGS_ left
llfourn DukeOfPerl: interesting 14:40
Zoffix Yeah, that page lost me at "PERL"
DukeOfPerl An did Larry Wall reall attend Seattle Pacific College (now U) in Seattle?
DrForr *plonk* 14:41
DukeOfPerl Pardon me for aggrandizing "Duke of Perl" (in envy of Duke of URL on another site.
14:42 Ch0c0late joined
Zoffix shakes head and resumes hacking 14:42
DukeOfPerl Are you "friendly" people "plonkinh" me? How rood.
OK I gotta go, amy I yet lurk a while? Toady, are you really Dr. Wall?
14:44 ChoHag joined 14:45 uruwi left
Zoffix huggable, texas :is: All of Perl 6's fancy Unicode operators have traditional ASCII symbol alternatives (aka Texas Variants): See doc.perl6.org/language/unicode_texas 14:47
huggable Zoffix, Added texas as All of Perl 6's fancy Unicode operators have traditional ASCII symbol alternatives (aka Texas Variants): See doc.perl6.org/language/unicode_texas
Skarsnik Am I the only one messing enum syntax everytime? I always write enum Foo = () 14:48
colomon doesn’t remember p6 enum syntax at all... 14:49
14:49 Zoffix left
Skarsnik it's enum Foo (); 14:50
llfourn enum Foo = () # woudn't this mean you want to set a symbol of type enum to an empty list? 14:51
14:52 xpen joined
llfourn though I suppose constant works the same way as you expect so I can see why 14:52
14:52 lucasb joined
DukeOfPerl If someone "plonks" me, am I then disconnected. I hope not. 14:57
DrForr Nope.
lucasb about enum syntax, I wonder why enum E 'a','b','c' doesn't work the same as enum E <a b c>
DukeOfPerl In comp.lang.perl.misc, people were rude to me talking about AI. 14:58
If the Usenet Perl people continue to be rude about Perl AI, I will post in other groups.
DrForr Sorry to hear it. 14:59
DukeOfPerl All right, before I vape (go), I have one big major Perl6 question...
14:59 rindolf left
DukeOfPerl In Perl5 for Windows, I am unable to deal directly with each keyboard stroke. CanPerl6??? 15:00
15:00 rindolf joined
llfourn DukeOfPerl: I doubt p6 is any better in terms of hardware interactions than p5 15:00
DrForr What does that mean?
DukeOfPerl I code AI in Forth (English; German) and JavaScript (English; Russian)... 15:01
In Forth and in JavaScript, the AI Mind processes each character AS IT COMEs In. Perl6?
DrForr Sure, just use the appropriate ReadKey library - modules.perl.org 15:02
DukeOfPerl Thankew, Dr Forr :-) 15:03
Skarsnik hm, xml does not allow <mytag/> ?
lucasb the xml specification? yes, I think it allows it 15:04
hankache DukeOfPerl an AI using Perl 6 would be a hit. If you have some spare time to work on it or blog about it, that would be great. 15:05
DukeOfPerl Right now I have 24/7 spare time for AI in Perl, Forth and JavaScript. 15:06
15:06 virtualsue left
DukeOfPerl I wish I had a 64-bit Linux machine running Perl6 for AI, but I am technologically primitive. 15:06
DrForr Cool, go for it!
DukeOfPerl Which Linux distribution is best for 64-bit Perl6?
hankache DukeOfPerl all of them work. 15:07
DukeOfPerl Also which 64-bit CPU should be on the Perl6 Linux machine?
DrForr They should all work, 'best' is your personal preference.
15:08 lichtkind joined
DrForr As is CPU. It's running on a VM, so it really doesn't matter. 15:08
DukeOfPerl Right now for Perl5 (Windows) all I have is an Acer Aspire One netbook (this machine).
VM (virtual machin e) is the Rakudo thingie, right?
DrForr Should work fine. I'm running rakudobrew on an Ubuntu VM inside a Windows machine. 15:09
15:09 firstdayonthejob joined, virtualsue joined
DukeOfPerl What sort of In ternet connection is needed? Ordinary Wi-Fi okay? 15:09
If I shop around here in Seattle WA USA, maybe I can buy an used Linux 64-bit machine. 15:10
hankache DukeOfPerl ordinary wi-fi is ok
DukeOfPerl hankache, thankew I am sure. But Wi-Fi has never let me do "telnet". 15:11
DrForr If you can download from github, your connection is good enough.
Wifi has nothing to do with that, telnet's just a protocol.
DukeOfPerl Unfortunately, my Radio Shack Acer netbook from 2009 can not get into GitHub :-(
llfourn can you install git? 15:12
DukeOfPerl friedo.com/blog/2016/01/exploring-p...nd-running has helped me a lot. 15:13
Ilfourn: I dunno if I can install git. I code AI, but I am only a Latin-and-Greek major :-( 15:14
llfourn DukeOfPerl: well I'd say that installing git might be necessary right now to get perl 6 up and running
lucs Can I get more elegantly from ｢for %h.keys.sort -> $k { my $elem = %h.{$k} ⋯｣ to ｢for ⋯ -> $elem { ⋯｣? 15:15
DukeOfPerl Thanks, Ilfoourn, I will look into what "installing git" entails and requires.
llfourn I think a windows installer thing is in the works but I don't think it's been released for v6.0c
15:15 iH2O joined
DukeOfPerl When I shop for the 64-bit Linux machine, I will try to accommodate GitHub necessities. 15:16
llfourn DukeOfPerl: there are no github necessities. It's a website.
DukeOfPerl And another thing. Will CPAN be accepting new modules expressly for Perl6?
15:17 xpen left
llfourn no. modules.perl6.org/ is our CPAN for now 15:17
lucs (That is, without requiring the otherwise unused $k variable.)
DukeOfPerl Ilforun, cool!
15:17 Zoffix joined
Zoffix nemo, lol, your instructions completely disabled my XCompose now :) even stuff that did work no longer does 15:18
15:18 iH2O left
Zoffix It looks like it's doing something (sequence characters aren't typed), but end result is no char 15:18
lucs Zoffix: Oh, what are you doing with XCompose? (I've been tweaking mine a lot lately)
Zoffix lucs, I wanted to get a wider range of sequences: doc.perl6.org/language/unicode_entry#XCompose 15:19
lucasb lucs: maybe use: for %h.sort.values -> $v { ... }
15:20 pjscott joined
lucs lucasb: Oh, hmm... Trying. 15:20
15:21 st_iron joined
st_iron good afternoon 15:21
Zoffix \o 15:22
llfourn gday, st_iron
st_iron o/
lucs lucasb: Appears not to work.
lucasb idk, then :(
llfourn m: {three => "four",one => "two", a => "z"}.sort.values.note 15:23
camelia rakudo-moar 725348: OUTPUT«(a => z one => two three => four)␤»
lucs lucasb: Um, I may have done something wrong though...
llfourn lucs: which bit doesn't work?
lucs Right :)
llfourn ah you want just the .value? 15:24
lucs Yes.
llfourn m: {three => "four",one => "two", a => "z"}.sort.map(*.value) #maybe
camelia ( no output )
lucasb ahh, sorry, my mistake. %h.sort returns a list of pairs, I think
llfourn m: {three => "four",one => "two", a => "z"}.sort.map(*.value).note #maybe
camelia rakudo-moar 725348: OUTPUT«(z two four)␤»
lucs Yeah, good ol' map. 15:25
15:25 CIAvash joined
lucs lucasb, llfourn: Thanks 15:26
llfourn nps :)
lucs Zoffix: So, what's your XCompose problem? 15:27
Zoffix lucs, <Zoffix> It looks like it's doing something (sequence characters aren't typed), but end result is no char
lucs Zoffix: Let me rephrase: What did you do, what did you expect, bla bla bla :)
15:28 pjscott left
Zoffix lucs, irclog.perlgeek.de/perl6/2016-01-08#i_11854039 15:28
Skarsnik hm I want that A is the same type that B I need to write constant A := B? 15:29
I want the same thing that typedef in C
lucasb if you don't mind that A.new.WHAT is (B)... 15:31
Skarsnik hm 15:32
15:32 Peter_R joined
Juerd Skarsnik: subset A of B; 15:33
yoleaux 08:17Z <bpetering> Juerd: I'm going through /^F/, skip to /^G/ please :)
15:33 khw joined
Juerd bpmedley: That's great! I got stuck in /^C/ after [^30] 15:33
bpmedley: Tab completian fail, never mind that 15:34
15:34 telex left
Skarsnik Juerd, not sure if subset will work well with nc type 15:34
Juerd What is nc type? 15:35
Skarsnik NativeCall
*types
Juerd Ah
Skarsnik I litteraly want to transform a 'typedef struct A B' to the perl6 eqv
15:36 telex joined
Zoffix m: class C { method gist {rand}}; say C.new; 15:39
camelia rakudo-moar 725348: OUTPUT«This type cannot unbox to a native string␤ in block <unit> at /tmp/r1Zmdma0Mx line 1␤␤»
Zoffix Huh? I thought say() called .gist
15:39 st_iron left
Zoffix m: class C { method gist {"42"}}; say C.new 15:43
camelia rakudo-moar 725348: OUTPUT«42␤»
lucasb m: class C { method gist { Inf } }; say C
camelia rakudo-moar 725348: OUTPUT«This type cannot unbox to a native string␤ in block <unit> at /tmp/rO5YHBBgHs line 1␤␤» 15:44
Skarsnik Not bad: to rewrite my gumbo binding : echo "unit module Gumbo::Binding; " > Binding.pm6 && perl6 -I lib bin/gptrixie --all /usr/local/include/gumbo.h | grep -v GUMBO_TAG >> Binding.pm6
15:45 zmyrgel left
geekosaur m: class C { method gist {~rand}}; say C.new; 15:46
camelia rakudo-moar 725348: OUTPUT«0.913867389771498␤»
lucasb hm, right, .perl, .Str, and .gist should always return strings 15:47
but what if the user made a mistake, can't the compiler give a hand there? :)
...and stringify what was supposed to be a string 15:48
Zoffix That's why my thought would be
I don't see why I have to manually cast to .Str all the time
m: class C { method Str {"42"}}; my @what = C.new, C.new, C.new; @what ».= gist; say @what 15:49
camelia rakudo-moar 725348: OUTPUT«5===SORRY!5=== Error while compiling /tmp/RBulbfJlyc␤Missing « or »␤at /tmp/RBulbfJlyc:1␤------> 3y @what = C.new, C.new, C.new; @what ».=7⏏5 gist; say @what␤ expecting any of:␤ infix␤ infix stopper␤»
Skarsnik m: class C { method gist returns Str { rand } }; say C
camelia rakudo-moar 725348: OUTPUT«Type check failed for return value; expected Str but got Num␤ in method gist at /tmp/IRZ0gs7ykI line 1␤ in block <unit> at /tmp/IRZ0gs7ykI line 1␤␤»
Skarsnik lol really
15:50 havenwood left 15:51 virtualsue left
DukeOfPerl Says good-bye to go work on the AI Mind Perl6 "Killer App". 15:52
15:52 DukeOfPerl left
Zoffix Bastard stole my idea! :D 15:53
arnsholt Skarsnik: A typedef doesn't need much translation when you're wrapping a library, IMO. Just pick one of the names for your class and use that
DrForr The joke will be on us when it achieves Singularity and takes over the world. 15:54
Zoffix :) 15:55
hankache ;)
llfourn he was a perl 5 AI right? That's what I thought he said.
Skarsnik lol
15:57 virtualsue joined
Zoffix m: say $PROCESS::OUT.Str 15:58
camelia rakudo-moar 725348: OUTPUT«IO::Special.new(what => "<STDOUT>")␤» 15:59
15:59 uruwi joined
Zoffix (not on docs: /types) 15:59
autarch is there an IMPORT or import sub called when a module is 'use'd?
Zoffix autarch, doc.perl6.org/language/modules#Expo..._Importing 16:00
doc.perl6.org/language/modules#EXPORT on that page in particular
autarch hmm, but there's no sub that's designed for _accepting_ options and not exporting stuff 16:01
llfourn autarch: yes EXPORT
autarch I'm trying to figure out how to do things we do in Perl 5 by writing an import() that accepts arbitrary parameters and does stuff
llfourn autarch: yeah just do stuff and return an empty {}
Zoffix autarch, sub EXPORT (*@args, *%args) { return {}; } 16:02
autarch well, what if I want to export stuff too?
can I "callsame" to get the defaults to happen?
llfourn autarch: just return stuff in the { }, or do you mean you want is export to work
autarch yes, I want it to work
llfourn yeah you will have to make is export(:MANDATORY) or make your users use MyModule 'some-arg', :DEFAULT; 16:03
autarch sure, that's fine, but how do I invoke the built-in exporter?
llfourn autarch: you don't need to 16:04
it's always invoked
&EXPORT doesn't override anything
autarch ah, cool
llfourn the default exporter is called EXPORTALL or something I think but I've never dealt with it
16:05 _nadim left
llfourn autarch: out of interest what mischief are you planning on getting up to in sub EXPORT {} 16:06
16:07 hankache left
autarch llfourn: I'm working on Test::Stream and I'm building a module that acts like Test.pm6 - by default, I want it to emit certain required test events (Suite::Start & Suite::End) - however, you may want to use these subs in another module that manages suite start & end, so I need a way to disable that 16:07
and since this is not an object, there's really no other API to use 16:08
actually, hmm, I take that back, I suppose I could just have a sub you could call to disable those events
... which I now realize is probably a better way to do this for now
llfourn sounds reasonable to me 16:10
mst autarch: please look at the split front-end/back-end architecture of Test2 16:11
16:11 RabidGravy joined
mst autarch: we put quite a bit of time into figuring that out, and it'd be sad for you to remake the same mistakes in p6 land we just spent time unamking in p5 :) 16:11
autarch mst: what specifically? I think what I have so far is pretty similar to Test2 16:12
mst autarch: ok, but remember that 'the thing that acts like Test.pm6' should then be independet, in a separate namespace, and merely a client of the streaming code
autarch yes, that's exactly how it's designed
it's only named Test::Stream::Subs for now because it's in the same git repo
mst that wasn't clear, especially given you're still using Test::Stream as a name, which also got discarded during this process
autarch it would be very easy to move it to its own distro and make it require the Test::Stream core
well, there's no reason to call this stuff Test2 for perl 6 16:13
timotimo T3st, please
mst ok, so, most of the hassle we've had comes from "it's only named X for now" followed by "wait now people are relying on part of the interface that makes no sense"
could you not give its its own name within your own lib/ tree? otherwise I worry about this going hilariously wrong by accident 16:14
funrep llfourn: aah it's syntax, now i get it thanks!
autarch what should it be called then? I don't particularly care about the name 16:15
llfourn funrep: no probs!
Skarsnik I like T3est
mst autarch: how about a working name of 'Test::Predicates' with an assumption somebody will find you a better one?
autarch sure, that works
mst also note we now have a #perl6-toolchain channel which I'm trying to use to co-ord CompUnitRepo and stuff above and belove it 16:16
autarch I'll join that
mst that's not a "we should relocate this conversation" statement, just a "you're one of the people I'd like to know the channel exists"
16:21 zwu left
autarch what's the operator to iterate over >1 list at a time? 16:22
hoelzro autarch: Z? 16:23
for @first Z @second -> $first-element, $second-element { ... }
autarch yes, that's what I wanted
16:24 pochi_ is now known as pochi
krunen [Coke]: (re moving Rakudo release tags) I thought I only said it was talked about at some point, but confused discussion happened. Sorry. 16:24
RabidGravy Ok darlings, is it actually possible to create a useable role on the fly? 16:26
I thought something like
m: my $a = Metamodel::ParametricRoleHOW.new_type(name => "Bnana"); $a.^compose; my $b = 1; $b does $a
camelia rakudo-moar 725348: OUTPUT«Could not instantiate role 'Bnana':␤Cannot invoke this object␤ in any at gen/moar/m-Metamodel.nqp line 2423␤ in any specialize at gen/moar/m-Metamodel.nqp line 2410␤ in any compose at gen/moar/m-Metamodel.nqp line 2979␤ in any generate_mixin…»
RabidGravy but clearly it doesn't work how I anticipated 16:27
16:27 virtualsue left
moritz RabidGravy: first of all you should bind to $a, not assign to it; otherwise you might end up composing Scalar 16:28
RabidGravy: the MOP is fiddly in this way
RabidGravy actually I think I knew that bit but forgot 16:29
Zoffix :o my 50-line Perl 6 code worked correctly the first time I ran it.
I guess this is a point where I can start calling myself a Perl 6 programmer ^_^
RabidGravy Zoffix, it's a false dawn ;-) 16:31
16:31 virtualsue joined
Zoffix looks at en.wikipedia.org/wiki/False_dawn and doesn't get the reference 16:31
MadcapJake anyone know if I could (when installing rakudo) specify the eventual path of the files but install all files to a temporary location and move them manually? Right now I'm trying `--sysroot=/usr --prefix=<temp-dir>`, would that work? 16:34
hoelzro MadcapJake: I think DESTDIR for make install works
Zoffix ^ this is the second time someone asks that in 2 days :) We should have a solution
hoelzro yes, it does
Zoffix hoelzro, but does that satisfy the "move them manually" portion?
MadcapJake really, so I can specify a DESTDIR on the make install? 16:35
and it won't affect how the paths are written to the scripts?
hoelzro --prefix tells Perl 6 where it will live someday
nine MadcapJake: the openSUSE packages are built just like that
MadcapJake nine, thanks, that'll be really helpful
hoelzro DESTDIR tells make install where to move the files
we have arch PKGBUILDS in perl6/os-build that should be pretty easy to fllow
nine MadcapJake: build.opensuse.org/package/view_fi...c?expand=1 16:36
MadcapJake hoelzro, those don't seem to do this though, i've been looking at the arch packages for a few days now
16:36 zwu joined
MadcapJake nine, thanks! 16:36
hoelzro MadcapJake: do what, --prefix? 16:37
nine MadcapJake: and yes the rm -rf commands are really a good idea. The precomp files generated during make install are useless and contain the build directory.
hoelzro you should probably trust nine's answer more; my arch pkgbuilds are probably rather outdated
MadcapJake hoelzro, oh i see that arch uses destdir actually, gah! I'm just not very knowledgeable with make/make install
16:38 salv0 left
MadcapJake hoelzro, nine: it looks like both of your packages separate moar and nqp out into separate packages, is that a good idea? 16:38
hoelzro MadcapJake: for arch, it's necessary 16:39
although...you *could* probably get away without it
it would just be hacky
MadcapJake yeah that's the predicament i'm in. I've been trying really hard to just generate it all in one package, but i think it will be cleaner to separate them 16:40
hoelzro I recommend giving in and just doing separate packages
MadcapJake but if i can use destdir, i might be able to get away with it B)
hoelzro but that's from my experience
16:41 pjscott joined
hoelzro MadcapJake: the problem is if you install Moar or NQP to DESTDIR, and you try to invoke them to build Rakudo, they'll be looking in PREFIX for their runtime files 16:41
I think Moar can handle it
I don't think NQP can
but I may be wrong
MadcapJake hoelzro, that's exactly the problem i've been having actually!
nqp fails to find moar
hoelzro I just tried once, it didn't work, and moved on =)
MadcapJake ok so it's settled, just make it three packages and be done with it! thanks so much hoelzro & nine! 16:42
nine MadcapJake: thanks for packaging rakudo :) 16:43
hoelzro indeed!
MadcapJake++
16:44 Ch0c0late left
MadcapJake no prob! this OS is really neat and need my perl 6 if I'm gonna stick with it :P 16:45
Skarsnik what system? 16:46
MadcapJake solus-project.com it's a really nice custom gnome-shell and a custom-built OS underneath 16:47
I'm not particular knowledgeable of build systems but I think theirs is pretty innovative (along the lines of Nix package manager, I believe) wiki.solus-project.com/EvoBuild 16:48
RabidGravy but linux of some sort under the hood 16:49
MadcapJake yep
RabidGravy I was getting excited, I guess at some point somebody will get it to work on something really exotic 16:50
MadcapJake yeah custom-built everything beyond linux kernel :P
Zoffix I see I'm not the only one posting P6 results on this :) twitter.com/OvidPerl/status/685770487989932032 16:52
<3 Perl 6
Skarsnik m: say "Story" ~~ /^(.+?)"+"?/ 16:53
camelia rakudo-moar 725348: OUTPUT«｢S｣␤ 0 => ｢S｣␤»
Zoffix m: say "Story" ~~ /^(.+?)"+"? $/ 16:54
camelia rakudo-moar 725348: OUTPUT«｢Story｣␤ 0 => ｢Story｣␤»
16:54 psy_ joined
Skarsnik m: say "Story+{Role}" ~~ /^(.+?)"+"?/ 16:54
camelia rakudo-moar 725348: OUTPUT«5===SORRY!5=== Error while compiling /tmp/QsOUQdhBfw␤Undeclared name:␤ Role used at line 1␤␤»
Skarsnik m: say 'Story+{Role}' ~~ /^(.+?)"+"?/
camelia rakudo-moar 725348: OUTPUT«｢S｣␤ 0 => ｢S｣␤»
16:54 skids joined
Zoffix m: say 'Story+{Role}' ~~ /^(<-[+]>+)"+"?/ 16:55
camelia rakudo-moar 725348: OUTPUT«｢Story+｣␤ 0 => ｢Story｣␤»
Zoffix m: say 'Story+{Role}' ~~ /^(<-[+]>+?)"+"?/
camelia rakudo-moar 725348: OUTPUT«｢S｣␤ 0 => ｢S｣␤»
Zoffix :(
Whatcha tryina match?
autarch ok, so Z is not what I wanted - how can I do something like this "for @a, @b -> $a, $b { ... }" where $a & $b get one element at a time from their respective lists? 16:56
Skarsnik Story
Zoffix Why is that not Z?
Skarsnik Well I want to capture Story
autarch Z produces a set of 2 element lists
hoelzro ah, my bad
Zoffix m: say 'Story+{Role}' ~~ /^(\w+)/
camelia rakudo-moar 725348: OUTPUT«｢Story｣␤ 0 => ｢Story｣␤»
hoelzro for @first Z @second -> ($first-element, $second-element) { ... }
autarch oh, I just need to destructure
hoelzro autarch: ↑
Skarsnik \w will probably fail with what perl6 accept as class name 16:57
hoelzro I always forget what needs destructuring and what doesn't
Z does, .kv doesn't
Skarsnik m: class My_class-a {};
camelia ( no output )
Zoffix Skarsnik, that's why I asked what you're trying to match :)
m: say "Story+{Role}" ~~ /^(.+) '+'? '{'/ 16:58
camelia rakudo-moar 725348: OUTPUT«5===SORRY!5=== Error while compiling /tmp/a7hz3Om0Yw␤Undeclared name:␤ Role used at line 1␤␤»
Zoffix m: say 'Story+{Role}' ~~ /^(.+) '+'? '{'/
camelia rakudo-moar 725348: OUTPUT«｢Story+{｣␤ 0 => ｢Story+｣␤»
moritz autarch: for @a Z @b -> ($a, $b) { ... } works fine and is idiomatic
Zoffix use Acme::GiveUp;
autarch moritz: yep, already figured that out - I got hung up on the destructuring at first
moritz autarch: the alternative is to use explicit flattening: for flat @a Z @b -> $a, $b { ... } 16:59
autarch ah
hoelzro ahhh
lucasb or more uglier: @a Z[&slip] @b 17:00
RabidGravy boom
Skarsnik m: say 'Story+{Role}' ~~ /^[!"+"]+/
camelia rakudo-moar 725348: OUTPUT«5===SORRY!5===␤Unrecognized regex metacharacter ! (must be quoted to match literally)␤at /tmp/__ofolUpaN:1␤------> 3say 'Story+{Role}' ~~ /^[7⏏5!"+"]+/␤Unrecognized regex metacharacter " (must be quoted to match literally)␤at /tmp/__ofo…»
Skarsnik m: say 'Story+{Role}' ~~ /^<!"+">+/
camelia rakudo-moar 725348: OUTPUT«5===SORRY!5===␤Unrecognized regex metacharacter < (must be quoted to match literally)␤at /tmp/yhfZ85WfeE:1␤------> 3say 'Story+{Role}' ~~ /^<!7⏏5"+">+/␤Unable to parse regex; couldn't find final '/'␤at /tmp/yhfZ85WfeE:1␤------> 3say …»
timotimo m: my Int @a[5] = 0, 1, 2, 3, 4; @a >>does>> <Hello How Are You>; .say for @a; say @a.gist; say @a.perl;
camelia rakudo-moar 725348: OUTPUT«Cannot mix in non-composable type Str into object of type Int␤ in block <unit> at /tmp/sSonG0rdkh line 1␤␤»
timotimo aaw
Skarsnik I think grouping is weird x) 17:01
m: say 'Story+{Role}' ~~ /^<!+>+/
camelia rakudo-moar 725348: OUTPUT«5===SORRY!5===␤Unrecognized regex metacharacter < (must be quoted to match literally)␤at /tmp/pu3uvlatMb:1␤------> 3say 'Story+{Role}' ~~ /^<!+7⏏5>+/␤Unable to parse regex; couldn't find final '/'␤at /tmp/pu3uvlatMb:1␤------> 3say 'S…»
RabidGravy okay, so I still couldn't work out any incantation to completely make a role from scratch, but this seems to be enough
17:01 zwu left
RabidGravy m: my $a = role {}; $a.^set_name("nana"); $a.^add_method("foo", method {"rarr"}); my $b = 1; $b does $a; say $b.WHAT; say $a.foo 17:01
camelia rakudo-moar 725348: OUTPUT«(Int+{nana})␤rarr␤»
Skarsnik m: say 'Story+{Role}' ~~ /^<[-+]>+/ 17:02
camelia rakudo-moar 725348: OUTPUT«Nil␤»
Skarsnik m: say 'Story+{Role}' ~~ /^(<[-+]>+)/
camelia rakudo-moar 725348: OUTPUT«Nil␤»
Skarsnik m: say 'Story+{Role}' ~~ /^<-[+]>+/ 17:04
camelia rakudo-moar 725348: OUTPUT«｢Story｣␤»
hoelzro there's no way (currently) to get EVAL to behave as if --doc were set, right? like a EVAL $code, :doc ? 17:05
autarch m: say False ~~ False
camelia rakudo-moar 725348: OUTPUT«False␤»
autarch is that really intentional?
Skarsnik lol what
m: Bool::False ~~ Bool::False;
camelia ( no output )
Skarsnik m: say Bool::False ~~ Bool::False;
camelia rakudo-moar 725348: OUTPUT«False␤»
autarch that just seems really really wrong 17:06
leont Yeah, Bools smartmatch that way
It surprised me too
Skarsnik m: my Bool $bool1; my $a = False; if ($a ~~ $bool1) {say 'hello'}; 17:07
camelia rakudo-moar 725348: OUTPUT«hello␤»
autarch that's also confusing
Skarsnik m: my Bool $bool1 = False; my $a = False; if ($a ~~ $bool1) {say 'hello'}; 17:08
camelia ( no output )
nine Why? False is a Bool so it smartmatches
Skarsnik look like that could lead to weird bug
leont The default value for a Bool isn't false, it's a Bool type-object 17:09
nine While nothing smartmatches False and everything smartmatches True. It's...something to get used to but also the only way we could have defaults.
moritz Skarsnik: the alternative is also prone to errors
autarch right, I don't see why a Bool type object matches False
leont It's the other way around
False matches a Bool type object
autarch leont: oh, that does make sense
moritz Skarsnik: having ~~ False always be False and ~~ True always be True enables us to write stuff like ~~ *.defined
autarch so if I'm _expecting_ False how do I check that I got False?
leont === False 17:10
I also have had to use that
moritz autarch: !$thing or === False or eqv False
Juerd Why is it important to you that it's False instead of some other false value?
leont I needed it when dealing with an Enum that could be false, that I needed to handle differently 17:11
autarch Juerd: because I'm comparing object attributes that I expect to get, and the object contains a Bool-typed attribute 17:13
Juerd autarch: If you know it's going to be a Bool, why not just use !$foo ? 17:14
Skarsnik It's annoying to remember that Array attribute are Positionnal and not Array :(
autarch Juerd: I have a hash of key/value pairs where keys are attributes to check and values are ... anything
and I'd rather not special case lots of different types - using === seems to work well 17:15
Juerd Well, === is all about special casing.
mst === will test reference eqaulity and eqv contents?
moritz mst: yes
17:15 leont left
mst egads, a use for my time in clisp land 17:16
Juerd What do you mean by "reference equality"? =:= tests identity equality.
Skarsnik m: my Str @a; say @a.^type.of; say @a.^type.of.defined;
camelia rakudo-moar 725348: OUTPUT«Method 'type' not found for invocant of class 'Perl6::Metamodel::ClassHOW+{<anon>}+{<anon>}'␤ in block <unit> at /tmp/ZEkkvEl8tm line 1␤␤»
moritz Juerd: =:= tests container equality
Juerd moritz: Is this a container versus value thing?
moritz m: my ($a, $b); say $a =:= $b; say $a =:= $a;
camelia rakudo-moar 725348: OUTPUT«False␤True␤»
moritz m: my ($a, $b); say $a =:= $b; say $a =:= $a; say $a === $b; 17:17
camelia rakudo-moar 725348: OUTPUT«False␤True␤True␤»
moritz Juerd: yes
Juerd Thanks. TIL :)
Skarsnik m: my Str @a; say @a.WHAT.of; say @a.WHAT.of.defined;
camelia rakudo-moar 725348: OUTPUT«(Str)␤False␤»
Skarsnik damn how I know the Array is 'typed' ?
Juerd Any array is typed. The default type is Any.
17:18 pjscott left
moritz m: @a; say @a.of.perl 17:18
camelia rakudo-moar 725348: OUTPUT«5===SORRY!5=== Error while compiling /tmp/jIAPcwgECS␤Variable '@a' is not declared␤at /tmp/jIAPcwgECS:1␤------> 3<BOL>7⏏5@a; say @a.of.perl␤»
moritz m: my @a; say @a.of.perl
camelia rakudo-moar 725348: OUTPUT«Mu␤»
moritz Juerd: Mu, not Any
Juerd To know that the array is typed to something other than Any, @a.WHAT.of !=== Any, I'd guess.
Huh?
Skarsnik m: my Str @a; say @a.WHAT.of; say @a.WHAT.of !=== Any;
camelia rakudo-moar 725348: OUTPUT«(Str)␤True␤»
Juerd Why's the default type for array values Mu, but the default type for scalar values Any/
mst but given there's always a type of some sort, one would imagine that testing it that way is odd anyway
moritz Juerd: otherwise you couldn't store a junction in an array
Juerd m: my $a; say $a.WHAT
camelia rakudo-moar 725348: OUTPUT«(Any)␤»
Skarsnik m: my Str @a; say @a.WHAT.of; say @a.WHAT.of !=== Mu; 17:19
camelia rakudo-moar 725348: OUTPUT«(Str)␤Cannot call infix:<===>(Str, Mu); none of these signatures match:␤ ($?)␤ ($a, $b)␤ (Int:D \a, Int:D \b)␤ (int $a, int $b)␤ (Num:D \a, Num:D \b)␤ (Num $ where { ... }, Num $ where { ... })␤ (num $a, num $b --> Bool…»
Juerd m: my @a; say $a.of.WHAT
camelia rakudo-moar 725348: OUTPUT«5===SORRY!5=== Error while compiling /tmp/w6lqe7SMhz␤Variable '$a' is not declared. Did you mean '@a'?␤at /tmp/w6lqe7SMhz:1␤------> 3my @a; say 7⏏5$a.of.WHAT␤»
Juerd er :)
m: my @a; say $a.WHAT.of
camelia rakudo-moar 725348: OUTPUT«5===SORRY!5=== Error while compiling /tmp/Mnq9CwfEOq␤Variable '$a' is not declared. Did you mean '@a'?␤at /tmp/Mnq9CwfEOq:1␤------> 3my @a; say 7⏏5$a.WHAT.of␤»
Juerd I see what I did wrong but I won't spam the channel more than this :)
Skarsnik m: my @a; say @a.WHAT.of; say @a.WHAT.of !=== Mu;
camelia rakudo-moar 725348: OUTPUT«Cannot look up attributes in a type object␤ in block <unit> at /tmp/ZUPn_Z5qjB line 1␤␤»
Skarsnik $a = @a
m: my @a; say @a.WHAT.of; say @a.WHAT.of;
camelia rakudo-moar 725348: OUTPUT«Cannot look up attributes in a type object␤ in block <unit> at /tmp/GeqUJxKqJQ line 1␤␤»
Skarsnik that... 17:20
timotimo why would you inspect the .of of the WHAT?
m: my @a; say @a.of.perl
camelia rakudo-moar 725348: OUTPUT«Mu␤»
Skarsnik I want to know the type of an Array
timotimo don't forget the .WHAT gives you the type object
Skarsnik well the of
dalek osystem: 04668e6 | hoelzro++ | META.list:
Add POD::EOD to ecosystem
Skarsnik m: my Str @b; my @a; say @a.WHAT.can('of'), @b.WHAT.can('of'); 17:21
camelia rakudo-moar 725348: OUTPUT«(of of)(of of of)␤»
Skarsnik m: my Str @b; my @a; say @a.WHAT.^can('of'), @b.WHAT.^can('of');
camelia rakudo-moar 725348: OUTPUT«(of of)(of of of)␤»
Skarsnik gah
moritz Skarsnik: not that you can use @a.of directly; no need to go through @a.WHAT.of
Skarsnik: s/not/note/
timotimo going through .WHAT.of is wrong, too. 17:22
not just unnecessary
Skarsnik why it's wrong?
moritz timotimo: I don't think so
Juerd m: my $a; say $a.WHAT; $a = 1|2; say $a.WHAT;
camelia rakudo-moar 725348: OUTPUT«(Any)␤(Junction)␤»
moritz m: my Str @a; say @a.^name
camelia rakudo-moar 725348: OUTPUT«Array[Str]␤»
timotimo oh, the .of is a property of the type, not the instance?
Juerd moritz: Why can this Any hold a junction?
hoelzro DOC INIT { use Pod::EOD; } # if I'm not using --doc, should a file with this in it fail to compile if Pod::EOD is not installed? that seems LTA
Skarsnik Yes, of in on the type
moritz Juerd: $a is typed Mu but defaults to Any
hoelzro m: DOC INIT { use Pod::EOD }
camelia rakudo-moar 725348: OUTPUT«===SORRY!===␤Could not find Pod::EOD in:␤ /home/camelia/rakudo-m-inst-1/share/perl6/site␤ /home/camelia/rakudo-m-inst-1/share/perl6/vendor␤ /home/camelia/rakudo-m-inst-1/share/perl6␤ CompUnit::Repository::AbsolutePath<140234873832976>…»
timotimo m: my $a; say $a.VAR.of
camelia rakudo-moar 725348: OUTPUT«(Mu)␤»
Juerd I see 17:23
Thanks
17:23 dustinm`_ left
moritz just like 'my @a' defaults to Array and is typed Positional 17:23
Juerd What's the benefit of defaulting to Any instead of Mu?
Skarsnik why array attribute are Positional?
moritz Juerd: Any is the default constraint type for parameters, so you'd get dispatch failures for uninitialized variables 17:24
skids @ sigil's constraint is it must contain a Positional.
Skarsnik It really weird when doing stuff with A.^attributes
moritz Juerd: like my $x; $x++; # would fail if $x defaulted to Mu
orelse $++ on a junction of variables wouldn't work
Juerd moritz: Ah. Thanks.
moritz it tooks us some iterations to find that out the hard way :-) 17:25
Skarsnik How I know a class is not a Core class? 17:26
Yes I have weird need x)
timotimo you can go through CORE::.keys (except be careful of StopIteration being in there)
moritz m: say CORE::<Int>.exists
camelia rakudo-moar 725348: OUTPUT«Method 'exists' not found for invocant of class 'Int'␤ in block <unit> at /tmp/w_irkTOQ0G line 1␤␤»
timotimo oh. or that
moritz m: say CORE::<Int>:exists
camelia rakudo-moar 725348: OUTPUT«True␤»
llfourn m: class Foo {}; CORE::<Foo>:exists.say
camelia rakudo-moar 725348: OUTPUT«False␤»
moritz it's a hash; you don't have to iterate it to check of existence :-) 17:27
timotimo well .....
m: constant Foo = Int; say CORE::{Foo}:exists
camelia rakudo-moar 725348: OUTPUT«Use of uninitialized value of type Int in string context␤Any of .^name, .perl, .gist, or .say can stringify undefined things, if needed. in block <unit> at /tmp/fGzaxdFDgl line 1␤False␤»
timotimo m: constant Foo = Int; say CORE::{Foo.name}:exists
camelia rakudo-moar 725348: OUTPUT«Method 'name' not found for invocant of class 'Int'␤ in block <unit> at /tmp/20QrUEWwqW line 1␤␤»
timotimo m: constant Foo = Int; say CORE::{Foo.^name}:exists
camelia rakudo-moar 725348: OUTPUT«True␤»
moritz could still be redeclared in an inner scope
timotimo ah
moritz m: { my class Date { }; say CORE::{ Date.^name}:exists }
camelia rakudo-moar 725348: OUTPUT«True␤»
timotimo right, that's more problematic. so you'll need to check identity, too 17:28
moritz m: { my class Date { }; say CORE::{ Date.^name}:exists && CORE::{ Date.^name } =:= Date }
camelia rakudo-moar 725348: OUTPUT«False␤»
17:28 virtualsue left
timotimo good 17:28
moritz m: { say CORE::{ Date.^name}:exists && CORE::{ Date.^name } =:= Date }
camelia rakudo-moar 725348: OUTPUT«True␤»
moritz and yes, that's actually abusing =:=
llfourn moritz: shouldn't you check ===
Zoffix m: my \x = 2; say x x x x x # :D
camelia rakudo-moar 725348: OUTPUT«2222␤»
moritz but if you know you have no containers, =:= works, and doesn't require extra care with Mu 17:29
llfourn moritz: ah yeh that makes sense
moritz (which is why the setting code uses =:= to test for StopIteration)
hoelzro I suppose DOC INIT doesn't defer use statements in its block?
llfourn what if you do have containers and have Mu?
hoelzro I can get around it with DOC INIT { require Pod::EOD }, but I'm wondering if use in DOC blocks should be skipped unless %?DOC
er, $?DOC 17:30
moritz llfourn: I don't know; which is why I'd prefer it if === didn't autothread, but I was overruled in that decision
TimToady I see why the warning for ~~ True and ~~ False went away; testing a patch
yoleaux 12:02Z <CIAvash> TimToady: Is there a reason '≠', '≤' and '≥' are not used for '!=', '<=' and
llfourn moritz: ah. I sorta ran into that problem the other day... I think I worked around it somehow 17:31
17:31 dustinm` joined, skids left
Zoffix +1 for '≠', '≤' and '≥' 17:31
I'd use them way more often than ÷ ^_^ 17:32
mst I'm going to end up writing an asciimogrifier for code where people do that
Zoffix mst, why?
mst, get on with the times, old man :P
mst i.e. something that pops them out into the ascii form for me to edit, and then back into whatever they were so the patch is clean
lets me continue to be a grumpy minimalist in the corner without stopping other people :P 17:33
TimToady well, depends on what you're minimalizing there
Zoffix Sounds like an easy thing to do with perl -pi -e''.
moritz mst++
Zoffix Surely ⊈ is much clearer than !(<=)
moritz Zoffix: such things are always easy, until you hit the corner cases 17:34
like, not wanting to do that inside literals
mst TimToady: Bill Joy's visual interface for ex is not what you'd call designed for unicode
and I'm running a private fork of the real thing ;)
17:34 uruwi left
TimToady either you have to track where the non-ascii was, or you have to translate everything back to non-ascii even if it wasn't before 17:34
17:34 citizenX joined
mst well, if somebody's code isn't consistently one or the other I can send them two patches to pick between first 17:35
Zoffix I guess this is the sort of thing good for the /bikeshed repo? I want '≠', '≤' and '≥' and some sort of star char for the [Hyper]Whatever star
17:35 virtualsue joined
TimToady well, a real unicode-aware editor has got to do better on proportional spacing 17:35
Skarsnik if you are curious I am doing that gist.github.com/Skarsnik/ea16c3d9b305bed0f536 with these weird MOP stuff x)
17:36 citizenX left
TimToady merely having single and double-wide characters is not sufficient 17:36
Zoffix Skarsnik, an ORM?
17:36 uruwi joined
mst still remembers the days when I could say "I haven't fallen into every classsic trap, I've written a web framework and a templating system but not an ORM yet" 17:36
Zoffix We do need an ORM in P6 tho :P 17:37
Skarsnik yes, but I did not find something that did not ask me specify lot of stuff. It's to cover small case where Class = Table in DB
TimToady m: say m: say 𒐫
camelia rakudo-moar 725348: OUTPUT«5===SORRY!5=== Error while compiling /tmp/LyXmcb8Mgp␤Colons may not be used to delimit quoting constructs␤at /tmp/LyXmcb8Mgp:1␤------> 3say m:7⏏5 say 𒐫␤ expecting any of:␤ colon pair (restricted)␤»
Zoffix I'm only 1 out of three for those :)
TimToady m: say 𒐫
camelia rakudo-moar 725348: OUTPUT«9␤»
Zoffix Skarsnik++ 17:38
.u 𒐫
yoleaux U+1242B CUNEIFORM NUMERIC SIGN NINE SHAR2 [Nl] (𒐫)
Skarsnik and I use WHY for fun ^^
I could even say 'id' that look like a good primary key candidate xD 17:39
17:39 ghost_ joined
Zoffix And I also want subtest 'test description', { code-of-the-subtest }; 17:40
17:41 sena_kun joined
Zoffix starts github.com/perl6/zoffixs-wishes-and-demands 17:41
Skarsnik Oh yeah
DrForr While you're at it I'd like to add $num-tests-to-run to that list.
TimToady maybe the killer app of P6 is to ignore the browsing guts of a browser and use the rendering engine to write a proportional terminal emulator
Zoffix I can't believe I'm in the minority on the subtest() stuff :/
Skarsnik I don't use subtests since you can't name it :)
DrForr Or at least nested 'plan $x' inside subtests, which might FAIK work, haven't tried it.
subtest sub { }, 'name goes here'; 17:42
Zoffix DrForr, yeah, plan inside does work 17:43
lucasb anyone knows if that 'my Int $x = CORE' bug is rakudobugged? I tried searching zefram's bugs, but didn't found it
Zoffix But that's not really now it looks like... It's subtest { TOOONS OF CODE HERE }, 'name goes here'; and it's pretty easy to add a multi for subtest 'name goes here', { TOOONS OF CODE HERE };
mst which, frankly, is the only place a plan is a good idea
top level plans are just merge error generators 17:44
Zoffix s/now/how/;
llfourn m: .say for '𒐬'..'𒐫';
camelia ( no output )
llfourn :(
mst Zoffix: of course, you start discussing this just after leont quits
17:44 vendethiel left
Zoffix m: .say for 𒐬..𒐫; 17:44
camelia rakudo-moar 725348: OUTPUT«1␤2␤3␤4␤5␤6␤7␤8␤9␤»
Zoffix mst, well, I brought it up ages ago, but no one shared my opinion :)
llfourn Zoffix: you can't print the chars through :( 17:45
Zoffix Ah
mst I have to say 'subtest $name, $code;' seems far far nicer
and I'm not sure why it's t'other way round
Zoffix mst, for consistency with all the other &test, $expected, $optional-description
DrForr Because the name is the last argument in the other tests, I'd guess. 17:46
mst right, but a subtest isn't the same thing
sena_kun Another question about NativeCall is here... Is there any way I can print CArray insides? I loss data in between of two loc and don't know how to debug where exactly.
yoleaux 10:47Z <RabidGravy> sena_kun: I added a working example to your gist, it doesn't fail but I don't have the bzip2 dev packages to check it is doing the right thing
mst grumbles
Zoffix nods
DrForr And I'd agree, it should be at the front for readability's sake.
mst a foolish consistency is the hobgoblin of little minds
(for once, the -real- quote applies, not the bastardised version ;)
Zoffix :) 17:47
DrForr Egads, the perl6 tutorial is 1st thing in the morning, 1st day of OSCON.
RabidGravy mst, it could even be a multi so both the "description first" and "description last " be happy
Zoffix Is that bad?
mst RabidGravy: which is -probably- the right thing to do, yeah
DrForr I am *so* not a morning person.
lucasb TimToady: do you think enum E 'a','b','c' should work? or is the parentheses there really required? 17:48
RabidGravy hey I've giveen some great talks with the worst hangovers ;-) 17:49
DrForr This will be 9 hours of jet lag *plus* hangover. 17:50
Zoffix just found 3 Unicode ops without a valid Texas variant :)
17:51 CIAvash left
RabidGravy more booze to balance it out and it will be great 17:51
mst RabidGravy: yeah, and IIRC you also dragged me out of LPW to the pub at lunch for a couple pints because I needed that to de-hangover enough to give mine ;)
Skarsnik sena_kun, I had issue with acceding data with a Carray that was pointing to data on a Str I give the the lib, and I had to explictly manage the Str for no reason
Zoffix m: say 5**+; say 5**-; say 5**⁻
camelia rakudo-moar 725348: OUTPUT«5===SORRY!5=== Error while compiling /tmp/vBOSRI6q5R␤Prefix + requires an argument, but no valid term found␤at /tmp/vBOSRI6q5R:1␤------> 3say 5**+7⏏5; say 5**-; say 5**⁻␤ expecting any of:␤ prefix␤»
RabidGravy ;-)
Zoffix They're listed in doc.perl6.org/language/unicode_texas tho
mst DrForr: drink breakfast. you'll be fine.
17:52 zakharyas joined
DrForr Well, there'll be a bottle of Chartreuse Noix in my bag on the way over, that'll handle the rest :) 17:52
17:52 CIAvash joined
RabidGravy you can't have chartreuse unless BooK is involved ;-) 17:53
dalek c: 7920764 | (Zoffix Znet)++ | doc/Language/unicode_texas.pod:
Remove Texas superscript +/- as those do not work

And they don"t really need to
kudo/nom: 4f2bfcf | TimToady++ | src/Perl6/Actions.nqp:
re-enable warning on ~~ True and ~~ False

The warning was accidentally bypassed by code that makes it possible to chain ~~ on rhs literals.		 17:54
CIAvash m: say 2 ** -3 17:55
camelia rakudo-moar 725348: OUTPUT«0.125␤»
CIAvash m: say 2⁻³
camelia rakudo-moar 725348: OUTPUT«0.125␤»
TimToady I assume the doc problem was the Texas part, not the superscript part
sena_kun Skarsnik, let it be explicit, but how can I do that? I need to convert carray which I passed into function to blob, but blob is empty after "my $blob = Blob.new($buff);". I wonder is it I write something wrong with initialization or it's some type problem. 17:56
s/write/wrote/
Skarsnik $buff is a Buf? 17:57
TimToady lucasb: it's possible we could relax on the parens at some point, but I'm in no mood to redesign the language just yet :)
timotimo sena_kun: the reason is $buff.elems
there's not yet a super clever way to get the first N elements of a CArray into a blob or buf ...
RabidGravy sena_kun, check out my NativeHelpers::Array, it has most of those things
moritz I kinda like the requirements for parens (or a quoting construct), as to limit the impact of two terms in a row
timotimo oh, neat 17:58
sena_kun timotimo, I'll try now, okay.
lucasb TimToady: ok, thanks
sena_kun RabidGravy, thanks for reference. I'll look at it.
Zoffix TimToady, yeah, 5⁻ was listed equivalent of 5**-, and 5**- doesn't actually work (and don't need to) 17:59
RabidGravy most are so simple you can just copy them
TimToady as for ≠ and friends, sure, we could add them specifically, but I'm more interested in fixing the bug that prevents them from being user-defined correctly as chaining operators
sena_kun Skarsnik, $buff is a CArray, actually(name is strongly temportary).
18:01 gist_nb joined 18:02 vendethiel joined
RabidGravy sena_kun, the only way of dealing with CArray is to copy the elements one by one, they don't know how long they are just the same as actual C arrays 18:05
so a C function that populates an array must always tell you how big 18:06
or more typically you give an array and say how big
mst my $buff is Strong; 18:07
18:11 sftf left
lucasb ah, I found the ticket I was looking: rt.perl.org/Public/Bug/Display.html?id=126891 18:11
I was going to add more snippets there, but I think there's no need. the ticket already cover all cases
Zoffix What's the way to check the message from `fail` in the test?
m: sub foo { fail "error"; }; use Test; throws-like { foo }, X::AdHoc.new: :payload<error>
camelia rakudo-moar 4f2bfc: OUTPUT« 1..2␤ ok 1 - code dies␤ not ok 2 - right exception type (X::AdHoc)␤ ␤# Failed test 'right exception type (X::AdHoc)'␤# at /home/camelia/rakudo-m-inst-2/share/perl6/site/precomp/EFD1604865713AD5BB2DC596308FC4E2FDE573CD.1452363029.0371…»
lucasb m: my Int $x; $x = module {}; $x = package {}; $x = role {}; say 'ok' # this snippet
camelia rakudo-moar 4f2bfc: OUTPUT«ok␤»
Skarsnik it ignore the type?! 18:13
lucasb maybe because a Failure is not an Exception? 18:14
Zoffix The expected/got match in the full output: gist.github.com/zoffixznet/313f5520cdd6440a68bf 18:15
It works if I just use `X::AdHoc`, but I need to test the actual message is correct 18:16
moritz m: m: sub foo { fail "error"; }; use Test; throws-like { foo }, Exception, message => /error/; 18:17
camelia rakudo-moar 4f2bfc: OUTPUT« 1..3␤ ok 1 - code dies␤ ok 2 - right exception type (Exception)␤ ok 3 - .message matches /error/␤ok 1 - did we throws-like Exception?␤»
18:17 zakharyas left
Zoffix Thanks 18:18
moritz Zoffix: throws-like uses try, which fatalizes fail()s, so you get an exception back, not a Failure
and the exception has .message, not .payload
TimToady .tell gfldex could you please file a bug for Q:s:b{\$x}?
yoleaux TimToady: I'll pass your message to gfldex.
lucasb timotimo: btw, I was reading the backlog, just want to note that when iterating CORE::.keys, there's no need to worry about IterationEnd, because the hash keys are strings, not the actual objects 18:19
18:20 frankjh left
lucasb you actually wrote StopIteration... I was going to make a joke about this is not Python, but nevermind :) 18:20
gfldex TimToady: i shall file as you wish
yoleaux 18:18Z <TimToady> gfldex: could you please file a bug for Q:s:b{\$x}?
18:21 _nadim joined
timotimo lucasb: oh, durr :) 18:21
18:21 gist_nb left
dalek c: ad791da | (Zoffix Znet)++ | doc/Language/testing.pod:
Add example of testing a fail()
 18:22
c: 542ca73 | (Zoffix Znet)++ | doc/Language/testing.pod:
Strip trailing whitespace
gfldex TimToady: filed as wished under #127226 18:25
18:27 pi4 left
lucasb what should that Q:s:b{\$x} return? "42" or "$x" ? 18:31
gfldex 42
it's The Answer after all 18:32
lucasb but "\$x" returns '$x'
gfldex sorry, i'm confused
lucasb what you get evaluated first, the scalar interpolation or the backslash escaping the dollar sign
?
m: my $x = 42; say "\$x" 18:33
camelia rakudo-moar 4f2bfc: OUTPUT«$x␤»
lucasb *what will get eval first
gfldex \$ should ignore the $ and remove the \
18:33 sena_kun left
TimToady I think :s should turn on backslashing for $ 18:33
gfldex return ENOSLEEP 18:34
TimToady and for \ 18:35
much like q turns on backslashing for \ and the delimiter
lucasb ok, makes sense. so Q:s/\$x/ should return '$x' ? 18:36
TimToady nod
lucasb and Q:b/\$x/ currently returns '\$x', but I think it should return only '$x'. iow, consume that backslash 18:39
18:39 _nadim left
lucasb m: say Q:b/\z/ 18:40
camelia rakudo-moar 4f2bfc: OUTPUT«\z␤»
lucasb perl -wE 'say "\z"' #=> returns 'z' and warning: Unrecognized escape \z passed through
TimToady that's because it's in double quotes 18:44
perl -wE "say '\z'" # returns '\z' 18:45
I'd argue this is more like '' than ""
but only weakly, since I can see an argument for the other side as well 18:47
one could argue that :b is the most important component of :qq 18:49
and that, even though we're building stuff up from Q, :b is what trips it over into :qq semantics
unfortunately, that would really be a breaking change for 6.c, so we'd have to think about it 18:51
Floyd-ATC Trying to IO::Socket::INET.write() a HTTP response header, I use .encode('ascii') on the Str to make it a Blob. According to tcpdump it looks right but for some reason both curl and Firefox ignores the \r\n\r\n and treats the content as part of the header. I'm so confused...
TimToady arguably, the number of programs depending on that "feature" is pretty close to 0 currently, and I doubt roast tests for it 18:53
so it's likely fair game to be classified as a bug
Zoffix Floyd-ATC, what about \n\n? 18:54
CIAvash TimToady: can you please take a look at this question I asked a few weeks ago? irclog.perlgeek.de/perl6/2015-12-23#i_11762824
18:55 llfourn left
Floyd-ATC Zoffix: Same result, the header and content doesn't get separated. I'm fairly sure I'm messing up the type/encoding in some way but I just can't see it 18:56
Zoffix no idea
18:57 nzkindest joined
Skarsnik Floyd-ATC, what is your full output? and did you try using the http::message/http::response class in h:u:a? 18:57
19:00 lnrdo left, patrickz joined
Floyd-ATC I don't want to dump several lines here but the headers themselves are perfectly valid - I'm basically just trying to mimic a response to wrap my head around p6's types 19:01
tried encoding as 'ISO-8859-1' now, same result
looks perfect in tcpdump but it just doesn't work
pnu "home" repo got missing from the default $*REPO.repo-chain in recent rakudos.. I wonder if I got the fix right? github.com/rakudo/rakudo/pull/678 19:02
lizmat Zoffix: re 7920764 , strangely enough, I lifted the superscript +- from the code, so if they don't work, it's a bug :-) 19:05
Zoffix lizmat, they do work, but not their texas variants 19:06
lizmat aaahhhhh... ok
Floyd-ATC I've been programming in various languages since the early 90s but this... this is so difficult to work with I really can't see myself writing anything in it that's meant for production :-( 19:07
19:07 vendethiel left
Zoffix Floyd-ATC, do you have a code example to look at? 19:07
Floyd-ATC Hang on
Zoffix If it looks perfect in tcpdump, I can't imagine why the browsers would be ignoring it :P 19:08
19:08 lnrdo joined
Floyd-ATC pastebin.com/hsGecyGK 19:10
There HAS to be a difference there somewhere, I just can't see it
Zoffix Floyd-ATC, HTTP1/1 is wrong 19:12
Funny, Palemoon captures it as "HTTP/0.9 200 OK"
Yup, changing lines 38 and 39 to HTTP/1.1 showed something in the browser 19:13
Skarsnik And I don't know what is confusing, Str are unicode, if you want unencoded stuff you need Buf x) 19:14
19:14 glucosade joined
Floyd-ATC OK, I'm off to hang myself in the basement now :-) 19:14
Skarsnik or Blobl
Zoffix :D
19:15 nzkindest left, _Vi joined
TimToady CIAvash: I don't see anything wrong with your code, offhand, so please file a bug report if it leaks memory 19:16
Floyd-ATC Well, it starts being confusing when you're only 90% sure what you're doing and it doesn't work. And Perl is traditionally very good at DWIM 19:17
Zoffix Yeah :)
TimToady m: say False ~~ False
camelia rakudo-moar 4f2bfc: OUTPUT«Potential difficulties:␤ Smartmatch against False always fails; if you mean to test the topic for truthiness, use :!so or *.not or !* instead␤ at /tmp/u2LwNJJQ1M:1␤ ------> 3say False ~~ 7⏏5False␤False␤»
TimToady that one works again :) 19:18
Zoffix m: say True ~~ True
camelia rakudo-moar 4f2bfc: OUTPUT«Potential difficulties:␤ Smartmatch against True always matches; if you mean to test the topic for truthiness, use :so or *.so or ?* instead␤ at /tmp/xvj3rNd1FI:1␤ ------> 3say True ~~ 7⏏5True␤True␤»
Floyd-ATC Zoffix: Thanks for lending a fresh pair of eyes though, this was driving me crazy
Zoffix neat
Floyd-ATC, no problem :)
Floyd-ATC stupid typo
CIAvash TimToady: ok, thanks 19:19
lucasb the warning suggests ":so". any example where this adverb can be used? 19:20
TimToady m: say True ~~ :so
camelia rakudo-moar 4f2bfc: OUTPUT«True␤»
lucasb ah, like that
TimToady m: say "anything" ~~ :so
camelia rakudo-moar 4f2bfc: OUTPUT«True␤»
lucasb TimToady: thanks!
TimToady m: say Cool ~~ :!so 19:21
camelia rakudo-moar 4f2bfc: OUTPUT«True␤»
Juerd not so Cool after all?
m: say not so Cool
camelia rakudo-moar 4f2bfc: OUTPUT«True␤»
TimToady it's not so true, anyway
Zoffix m: say Perl or not Cool 19:22
camelia rakudo-moar 4f2bfc: OUTPUT«(Perl)␤»
lucasb wrong precedence :P 19:23
Juerd Perl is a type?
What does it do?
Zoffix huggable, Perl
huggable Zoffix, nothing found
Zoffix shrugs 19:24
awwaiid huggable?
yoleaux 09:55Z <nine> awwaiid: nice blog post! Actually...I do recommend to people to use Perl 5's Inline::Python even for critical production code. I can recommend this in good concience because that's what we do.
Zoffix huggable, help 19:25
huggable Zoffix, I'm a factoid bot. See me my source at github.com/zoffixznet/perl6-IRC-Cl...in-Factoid
dalek kudo/nom: a540b79 | (Panu Ervamaa)++ | src/core/CompUnit/RepositoryRegistry.pm:
Use p6box_s for version when setting home repo
kudo/nom: 20d5f0e | lizmat++ | src/core/CompUnit/RepositoryRegistry.pm:
Merge pull request #678 from pnu/nom

Use p6box_s for version when setting home repo
awwaiid Zoffix: fancy! 19:29
19:29 Ch0c0late joined 19:30 cooper_ is now known as cooper
lucasb m: 'oh hai' ~~ :say # now I understood the adverb with ~~ . It's just method invocation 19:30
camelia rakudo-moar 4f2bfc: OUTPUT«oh hai␤»
lucasb the warning for False suggest ':!so', but it could suggest ':not' 19:31
awwaiid lucasb: that's interesting. So you're saying that is equivalent to 'oh hai'.say ?
TimToady it could, but there's something to be said for teaching the :!foo form too 19:32
lucasb awwaiid: iiuc, yes, that's equivalent
19:32 lnrdo left 19:33 kjk joined
awwaiid m: say ('oh hai' ~~ :perl) ; say 'oh hai'.perl 19:33
camelia rakudo-moar 20d5f0: OUTPUT«True␤"oh hai"␤»
lucasb TimToady: hm, ok. point taken :)
Skarsnik hm I can't use a Type as a key in a hash?
awwaiid m: say (5 ~~ :is-prime) 19:34
camelia rakudo-moar 20d5f0: OUTPUT«True␤»
19:34 sena_kun joined
dalek kudo/nom: 8c97759 | lizmat++ | src/core/ (3 files):
Remove the Date/DateTime.new positional candidates

To make it more like 6.c , even though there were no tests in roast to test the positional interface, it was felt it would be better to make sure the interface does not change at all.		 19:34
19:34 lnrdo joined
awwaiid lucasb: I think it calls the method AND forces .Bool? 19:34
Zoffix m: my %foo = Int => 42; say %foo<Int> 19:35
camelia rakudo-moar 20d5f0: OUTPUT«42␤»
Zoffix Skarsnik, you can if you make it a string first :P
awwaiid still -- that's pretty powerful and convenient
lizmat marks 8c97759 as a commit to be reverted for 6.d 19:36
Skarsnik make me think, is there a way to get the basename of a class?
TimToady m: my %foo := :{ (Int) => 42 }; say %foo{Int}
camelia rakudo-moar 20d5f0: OUTPUT«42␤»
Zoffix :o
Skarsnik like without the namespace or its roles
TimToady m: my %foo{Any}; %foo{Int} = 42; say %foo{Int} 19:38
camelia rakudo-moar 20d5f0: OUTPUT«42␤»
awwaiid Skarsnik: what's an example of the input and output you desire?
Skarsnik m: class A {}; sub foo($a) {my %h{$a} = "foo"}; foo(A);
camelia rakudo-moar 20d5f0: OUTPUT«5===SORRY!5=== Error while compiling /tmp/VdVYllw4Gp␤Invalid hash shape; type expected␤at /tmp/VdVYllw4Gp:1␤------> 3class A {}; sub foo($a) {my %h{$a}7⏏5 = "foo"}; foo(A);␤ expecting any of:␤ constraint␤»
Skarsnik m: class A {}; sub foo($a) {my %h; %h{$a} = "foo"}; foo(A); 19:39
camelia rakudo-moar 20d5f0: OUTPUT«Use of uninitialized value $a of type A in string context␤Any of .^name, .perl, .gist, or .say can stringify undefined things, if needed. in sub foo at /tmp/i0H9g1FAGD line 1␤»
TimToady use %h{Any} if you want an object key
awwaiid m: class A {}; sub foo($a) {my %h<Any>; %h{$a} = "foo"}; foo(A);
TimToady as I already demoed above
camelia rakudo-moar 20d5f0: OUTPUT«5===SORRY!5=== Error while compiling /tmp/Tu_RUrYeye␤Shaped variable declarations not yet implemented. Sorry. ␤at /tmp/Tu_RUrYeye:1␤------> 3class A {}; sub foo($a) {my %h<Any>7⏏5; %h{$a} = "foo"}; foo(A);␤»
Skarsnik m: class A {}; sub foo($a) {my %h{Any}; %h{$a} = "foo"}; foo(A); 19:40
camelia ( no output )
awwaiid m: class A {}; sub foo($a) {my %h{Any}; %h{$a} = "foo"}; foo(A);
camelia ( no output )
awwaiid oops
Zoffix lizmat, what should be done with rt.perl.org/Ticket/Display.html?id=127186 ?
It's about positionals and Date[Time]
Skarsnik Zoffix, is that valid sqlite: author_id Int REFERENCES Author(id) or I need the foreign key keyword? 19:42
lizmat Zoffix: did'n't moritz fix that already in c19142cf26eaeff ?
kjk hello~ how do you refer to a method of an object without calling it? "asdf".say
Zoffix lizmat, right, but it's marked as "tests needed". Are they still needed if the positionals are not longer an option? 19:43
lizmat not for 6.c, so let's wait until we have 6.d testing set up
Zoffix kjk, what do you mean "refer to"?
k
kjk Zoffix: I want to hold on to it, with the object bound to it, and call it later 19:44
Zoffix kjk, .^can I believe
m: class Foo { method bar { say "barrrs!" }; }; my $sub = Foo.new.^can: 'bar'; $sub()
camelia rakudo-moar 20d5f0: OUTPUT«Invocant requires a type object of type List, but an object instance was passed. Did you forget a 'multi'?␤ in block <unit> at /tmp/xo6duPpKX2 line 1␤␤»
Zoffix wat
CIAvash For those interested, I blogged about entering unicode characters in Emacs(the rfc1345 part can be used for Vim too) blog.ciavash.name/2016/01/09/enteri...emacs.html 19:45
Skarsnik m: class Foo { method bar { say "barrrs!" }; }; my $sub = Foo.new.^can('bar'); $sub()
camelia rakudo-moar 20d5f0: OUTPUT«Invocant requires a type object of type List, but an object instance was passed. Did you forget a 'multi'?␤ in block <unit> at /tmp/yV4xhRSsmr line 1␤␤»
Skarsnik and can call it I think 19:46
lucasb you have to ^can('method')[0], because it's a list
Zoffix m: class Foo { method bar { say "barrrs!" }; }; my $sub = Foo.new.can('bar'); $sub.perl.say
camelia rakudo-moar 20d5f0: OUTPUT«$(method bar (Foo $: *%_) { #`(Method|79663304) ... },)␤»
Zoffix Ah
m: class Foo { method bar { say "barrrs!" }; }; my $sub = Foo.new.can('bar'); $sub[0]()
camelia rakudo-moar 20d5f0: OUTPUT«Too few positionals passed; expected 1 argument but got 0␤ in method bar at /tmp/RwVlfT7MBN line 1␤ in block <unit> at /tmp/RwVlfT7MBN line 1␤␤»
Zoffix m: class Foo { method bar { say "barrrs!" }; }; my $sub = Foo.new.^can('bar'); $sub[0]()
camelia rakudo-moar 20d5f0: OUTPUT«Too few positionals passed; expected 1 argument but got 0␤ in method bar at /tmp/Z_bFYjzdg2 line 1␤ in block <unit> at /tmp/Z_bFYjzdg2 line 1␤␤»
Juerd CIAvash: When typing the unicode character is more work than typing the texas variant, I'm not inclined to use the unicode variant :)
Zoffix kjk, well you get the idea ^_^ 19:47
TimToady Juerd: well, do you care about the readers of your code?
Juerd s/more/much more/
TimToady most code is read many more times than it is written
Juerd TimToady: That's another thing, I'm afraid that other people will have worse unicode/utf8 setups than I do :)
kjk i see.., though it's not bound to the object it seems. but i can work with that.., thanks all! 19:48
TimToady by that argument no progress will ever be made
lucasb My terminal is not even configure for unicode! :) That's why at least errors/warnings could avoid unicode chars
Juerd TimToady: This is one of those changes where I'm not convinced that it actually is progress, but I am happy that Perl 6 supports this so that people who care can at least try :)
Zoffix lucasb, heh, what's next? Avoid images on the docs site, because you're using a 56K modem? :P 19:49
TimToady 56K? wow, that's fast!
Juerd I might use → though if it were supported. Because it's so pretty compared to -> ;)
lucasb well, I'm browsing with lynx, so I don't see images :)
Zoffix :o
Juerd lucasb: For real?
lucasb just kidding :)
Juerd I was hoping that :) 19:50
sena_kun I'm annoynig today, probably, but what can be wrong here - pastebin.com/braJQ3M9 ? It's already three hours I'm trying to find out why value just vanishes.
Zoffix Juerd, the other day there was a person in here who uses w3m for everything
Juerd I've used w3m a lot and I still use it from time to time, but I mostly ditched it as sites began to become unusable without things such as Javascript. 19:51
19:51 llfourn joined
Juerd sena_kun: Where's the rest of the code? Like copy-carray-to-buf... And what's the issue? 19:53
TimToady sena_kun: in the absence of any information about what you're calling, I'd guess it's pre-GLR or pre-Christmas code 19:55
19:56 llfourn left
RabidGravy sena_kun, I'm surprised this doesnt segfault, you use "my $temp = CArray[uint32].new; my $len = BZ2_bzRead($bzerror, $bz_read, $temp, $size);" without extending the $temp to the required length 19:56
dalek osystem: 9ae6497 | (Zoffix Znet)++ | META.list:
Add IO::MiddleMan to the Ecosystem

IO::MiddleMan - hijack, capture, or mute writes to an IO::Handle - github.com/zoffixznet/perl6-IO-MiddleMan
RabidGravy $temp[$len] = 0; should do the trick 19:57
TimToady sena_kun: and if there's additional info in the backlog, I haven't got to it yet
sena_kun Juerd, copy-carray-to-buf is a part of NativeHelpers::Array module. Issue is: I'm trying to write bindings to bzip2. I can write things, but reading don't work as it should. C function push result of reading to CArray(with correct values in length) and... And result is zeroed.
lizmat hmmm... apparently Date.new *did* have a positional candidate, DateTime.new didn't
TimToady so obviously other people are tracking this better than me, so I'll shut up now :)
lizmat star-m: say DateTime.new(1999,12,12,0,0,0) 19:58
camelia star-m 2015.09: OUTPUT«Default constructor for 'DateTime' only takes named arguments␤ in block <unit> at /tmp/rAVgnOdcUR:1␤␤»
19:58 jshy left
lizmat star-m: say Date.new(1999,12,12) 19:58
camelia star-m 2015.09: OUTPUT«1999-12-12␤»
lizmat continues fixing
CurtisOvidPoe p6: say DateTime.now.perl 19:59
camelia rakudo-moar 8c9775: OUTPUT«DateTime.new(:year(2016),:month(1),:day(9),:hour(20),:minute(59),:second(28.8908677101135e0),:timezone(3600))␤»
19:59 patrickz left
CurtisOvidPoe Hmm, on my box that’s this: DateTime.new(2016,1,9,20,59,6.68587803840637,3600) 19:59
19:59 patrickz joined
Zoffix CurtisOvidPoe, lizmat just pushed a commit that changed that 20:00
s/just/25m ago/;
CurtisOvidPoe Zoffix: thanks! Missed that bit. It was part of why I changed my slides in my FOSDEM talk, so I was curious to see it coming up here.
lizmat CurtisOvidPoe: to preserve 6.c.ness
CurtisOvidPoe I’m too 6.c for my code …
Juerd CurtisOvidPoe: Upcoming fosdem talk? Looking forward to it! 20:01
CurtisOvidPoe Jeurd: it’s a major rewrite of last year’s FOSDEM talk, aimed at those who’ve never heard of Perl before. Works much better.
“Juerd”, even. (Who’s Jeurd?) 20:02
Juerd I don't know who Jeurd is :)
CurtisOvidPoe I hear he’s a jerk. 20:03
20:04 sena_kun left
Juerd How come you talk with proper quotes on IRC? 20:04
And an actual …, even. 20:05
TimToady why not?
Juerd I'm not complaining :)
TimToady ASCII is so 1970s 20:06
Juerd I'd like to know why, or actually, what's making it so.
TimToady a number of folks here are using clients that tend toward smartquotes
that's why Perl 6 handles them now 20:07
Juerd Let's see if irssi can do this
Zoffix P6's module production grew by 300% on www.modulecounts.com/ :)
What's curious is P5's doubled too.
20:07 sena_kun joined
arnsholt (If it's irssi, it's mostly a question of your terminal emulator, font and screen setup [if any]) 20:08
Zoffix (as compared to, like, last couple weeks of last year)
CurtisOvidPoe I’m using Colloquy on a Mac and I’ve tried to disable smart quotes, but no love. On the plus side, I can type code directly into my Perl 6 PowerPoint presentations without toggling smart quotes all the time :)
Juerd Apparently irssi doesn't have this feature...
20:09 lnrdo left
Skarsnik Zoffix, where do you see the grow? x) 20:10
stmuk__ I hadn't noticed fosdem.org/2016/schedule/track/perl/ is up
dalek kudo/nom: dde4cef | lizmat++ | src/core/Date (2 files):
Turns out Date.new *did* have positionals in 6.c
El_Che stmuk__: the last speaker confirmed today 20:11
so the list is definitive
Zoffix Skarsnik, www.modulecounts.com/ The last few weeks of 2016, the 'Avg Growth' for P6 was 1/day, dropping to 0/day sometimes, and for P5 it as 4/day. Now it's 3/day for P6 and 8/day for P5
20:15 patrickz left, hankache joined 20:16 lnrdo joined
hankache good evening #perl6 20:16
20:16 patrickz joined
Zoffix \o 20:17
20:18 marmay joined 20:26 virtualsue left 20:27 vendethiel joined, darutoko left 20:28 _nadim joined 20:31 domidumont left
lizmat TimToady: any thoughts an allowing ≤ and friends ? 20:32
(in 6.d. obviously :-)
TimToady irclog.perlgeek.de/perl6/2016-01-09#i_11859391 20:34
20:35 _nadim left
lizmat TimToady++ 20:35
20:37 frankjh joined 20:41 RUR joined
RUR hi 20:42
i am trying to call some native library function which takes pointer to filehandler 20:43
i am getting 20:44
Native call expected return type with CPointer representation, but got a P6opaque
sena_kun RUR, did you meant "FILE*" pointer?
RUR yes
20:45 CIAvash left
autarch RUR: a gist would probably help 20:45
sena_kun Paste somewhere C function signature and your perl6 function signature. Probably you want just "Pointer".
arnsholt Sounds like you're trying to pass a Perl 6 file handle 20:46
That's not gonna work
We don't have a good story for passing Perl 6 file handles off to native functions 20:47
sena_kun arnsholt, we always can bind "fopen" and "close" from default C library at the place and use it.
RUR gist.github.com/anonymous/9a1e1917c7eb737c0310 20:48
arnsholt Sure, but then you have to do ALL of your IO with libc functions
Which is probably going to suck
20:48 vendethiel left
RUR yes i am passing perl filehandle 20:48
arnsholt Yeah, that's not supported
You'll have to wrap fopen as well 20:49
RUR hmm yes i was already thinking about opening it C library, but i though there can be easier way
arnsholt 'Fraid not
sena_kun RUR, and it seems you're writing a bz2 library bindings, aren't you?
RUR just trying
:)
arnsholt In your gist, you can't pass $error there either
RabidGravy or use native-descriptor on the perl filhandle, and use fdopen with result 20:50
sena_kun RUR, it's fun, because I'm just trying too currently.
arnsholt RabidGravy: Someone implemented that? Cool!
RUR: Apparently I'm way out of date!
Try what RabidGravy said
Skarsnik it's like 2 weeks old?
RabidGravy yeah, I put a ticket jnthn fixed it :)
yeah RUR, apparently you are doing the same as sena_kun have a chat so we don't have to explain the same stuff again ;-) 20:52
20:52 llfourn joined
RUR yup :), ok thanks for help 20:52
sena_kun RUR, just write to me directly then(through I'm not confident too) if any. 20:53
Skarsnik and is the great and powerful trixie!
*use
RabidGravy arnsholt, you can even do ioctl, syswrite and everything with it :)
arnsholt Oh, wait. It returns the int fd, or a FILE*? 20:54
RabidGravy fd
20:54 pi4 joined
RabidGravy more useful for most things 20:54
and if you must you can bind fdopen to get a FIL* 20:55
arnsholt Then it won't work directly for RUR. But there's fdopen that converts an fd to a FILE*
RabidGravy yeah
arnsholt Yup! =)
20:57 llfourn left
lizmat not to mean to rain on the parade, but I think we cannot allow the native descriptor in 6.c 20:57
RabidGravy why? 20:58
lizmat because the library that is now is being developed, would *NOT* work on all compiler releases of 6.c
RabidGravy it was on the xmas rt list and got implemented before xmas
lizmat did it ?
RabidGravy yes
lizmat ok, then I will most gladly shut up :-) 20:59
nine lizmat: since jnthn went on holiday right at release time, we can be quite sure that everything he implemented is in 6.c :)
RabidGravy like 15th dec or something
masak lizmat++ # staying vigilant 21:00
21:00 frankjh left
RabidGravy yeah, I wrote code that used it on 16th december 21:01
so definitely before then :)
lizmat 344113cecbf5ffafe84da to be precise
Implment fileno_fh op. 21:02
Provides access to the native descriptor/handle underlying some OS
handle. Platform specific.
21:02 TEttinger joined
lizmat *phew* 21:02
RabidGravy groovy
nine Oh wow, while I've been lazily watching Raiders of the Lost Ark, lizmat++ has already prepared DateTime for 6.d :) 21:04
lizmat: I guess with that the 6.c spec tests for DateTime pass again? 21:05
lizmat yeah, working on IO::ArgFiles c.s
no, because the *test* was faulty
and I didn't undo the fix
dalek kudo/nom: a28270f | lizmat++ | / (8 files):
Restore old IO::ArgFiles behaviour

This removes IO::CatPath and IO::CatHandle from the core. Hope to see it back some time in 6.d.		 21:07
21:07 wamba left
nine lizmat: as soon as you say yes to language_versions, I'll start bringing all the goodness back in 6.d :) Or as soon as we agree on another solution. 21:08
TimToady this seems the best way for now
TimToady would really like to get jnthn++ back online before commiting to anything more solid 21:09
lizmat well, an additional fear I have, is that we will have a 6.c that is being optimized, and a 6.d. that will not get optimized until somebody trues to merge changes from the nom branch
and then will see massive merge issues
TimToady Fear is the mind-killer, I fear... :) 21:10
autarch I hope that whatever branch plans you have include regular (daily, ideally) merges 21:11
anything else is asking for trouble
RabidGravy indeed
also it makes cherry picking less onerous 21:12
nine lizmat: with jnthn's plan, most of 6.c's code will still be used in 6.d. In fact almost all, so why would we miss those optimizations?
And why are we talking about branches?
autarch because if 6.d is going to contain breaking changes then you need a branch for it
nine autarch: no we won't. That's the whole point of jnthn's suggestion. Please have a look at docs/language_versions.md 21:13
dalek ast: ece2218 | lizmat++ | S02-types/WHICH.t:
Bring test back into 6.c. state
masak if the plan means we can avoid having branches, then I like the plan :)
TimToady well, if the blan to avoid branches makes merging/cherrypicking harder, then I'm not so sure 21:14
*plan
nine github.com/rakudo/rakudo/blob/nom/...ersions.md
Why would it?
lizmat nine: how can we "augment" additional files in moar_core_sources ?
masak why would it make merging/cherry picking harder?
nine It's completely ortogonal to git and branches
lizmat without a branch ?
masak the idea is to have a single code base, which is beneficial in various ways. merging/cherry picking is not affected, except perhaps for the better. 21:15
RabidGravy I'm just confused, ignore anything I might say on the matter
nine "Additions to the setting will be made in a "nested" setting. That is, a "use v6.d" will load CORE.d.setting or so. It can use augment and supercede in order to effect changes." 21:16
lizmat well, I'm not seeing it either, RabidGravy
nine: how can we "augment" new files like IO::CatHandle in moar_core_sources ?
Skarsnik I though supercede was not implemented?
autarch nine: a master & nom branch are branches
nine lizmat: IO::CatHandle will always be contained in the build in the 6.d setting that will only be loaded by source files containing a use v6.d; or use v6.d.PREVIEW; 21:17
lizmat nine: there's a file called moar_core_sources that contains a list of files to *build* the setting
if we don't have branches, there's only 1 file of it 21:18
nine We have one perl6 binary containing code for all Perl6 versions. The setting that's built from the src/core/ directory contains everything that is 6.c. There are additional settings in directories src/core.d/ src/core.e/ and so on that contain (much smaller) nested settings that get loaded on demand.
lizmat how can that be shared between 6.c and 6.d ?
ok, so we just add files
fine by me
nine yes
TimToady as near as I can tell, jnthn's proposal is only that the *tests* go off in their own directory, and everything else stays branches 21:19
lizmat so every compiler will contain all code of all language versions ?
nine I've started doing that in the language_versions branch. I can do more changes (like bringing back IO::CatFiles) if if helps demonstrating how this is supposed to work
lizmat: yes, yes, yes :)
lizmat tries to let that sink in 21:20
TimToady we would like to be able to emulate older versions
RabidGravy I would hope only current and future versions, otherwise it will be immense in 10 years time :)
nine Ok, I'll just do further work in the language_versions branch tomorrow. I think it will be quite clear then :)
lizmat so if I understand the language_versions branch: it is a prototype of the future nom? 21:21
nine yes
TimToady and nom is currently the "maint" branch of 6.c?
lizmat still not seeing the "change setting" happening
21:22 hankache left
lizmat I mean, how can you now change to the RESTRICTED setting when already running, without shelling out ? 21:22
nine lizmat: github.com/rakudo/rakudo/blob/lang....nqp#L1514
21:23 mohij joined
lizmat tries to wrap her mind around that 21:23
nine That loads the CORE.d.setting.moarvm which is compiled against the CORE.setting. Those are just nested lexical scopes, so we have { CORE.setting { CORE.d setting { user code } }
That's why all 6.c code is still there and still in use except for the parts that are "overwritten" by the 6.d setting
21:23 patrickz left
RabidGravy falls back in his confidence that everyome is clever enough to make it work out 21:24
TimToady had a lot of trouble trying to get -n and -p to be implemented with nested settings, hope you have better luck :)
21:25 virtualsue joined
lizmat well, I guess I'm with RabidGravy 21:25
RabidGravy :) 21:26
TimToady well, our standard here is only "different mistakes" :)
Juerd m: use 🐵-SEE-NO-EVAL; 21:27
camelia rakudo-moar a28270: OUTPUT«5===SORRY!5=== Error while compiling /tmp/OIWJhJTI33␤Bogus term␤at /tmp/OIWJhJTI33:1␤------> 3use7⏏5 🐵-SEE-NO-EVAL;␤ expecting any of:␤ argument list␤ infix␤ infix stopper␤ postfix␤ prefi…»
Juerd Aww :)
lizmat .u 🐵
yoleaux U+1F435 MONKEY FACE [So] (🐵)
RabidGravy I'm with Juerd :)
TimToady .u 🙈
yoleaux U+1F648 SEE-NO-EVIL MONKEY [So] (🙈)
TimToady they misspelled it 21:28
lizmat fwiw, I hope that MONKEY-SEE-NO-EVAL will not become one of the category "you think that's cute now"
like "tuits", I think it's going to be very hard to explain to non-native english speakers :-(
Juerd rotflol, there actually is a SEE-NO-EVIL emoji. 21:29
Wow.
TimToady if it becomes horribly un-cute later, it will only add to the stigma of using EVAL 21:30
lizmat m: "say 42".EVAL 21:31
camelia rakudo-moar a28270: OUTPUT«42␤»
lizmat m: EVAL "say 42"
camelia rakudo-moar a28270: OUTPUT«42␤»
TimToady no problem till you interpolate
lizmat m: my $a = 42; "say $a".EVAL 21:32
camelia rakudo-moar a28270: OUTPUT«42␤»
lizmat is that a bug ?
TimToady hmm
I'd say so
lizmat looks at fixing
21:33 llfourn joined
TimToady m: say $a = 42; EVAL "say $a" 21:33
camelia rakudo-moar a28270: OUTPUT«5===SORRY!5=== Error while compiling /tmp/xg4_abXWRL␤Variable '$a' is not declared␤at /tmp/xg4_abXWRL:1␤------> 3say 7⏏5$a = 42; EVAL "say $a"␤»
TimToady m: my $a = 42; EVAL "say $a"
camelia rakudo-moar a28270: OUTPUT«5===SORRY!5=== Error while compiling /tmp/bjADu7u0QY␤EVAL is a very dangerous function!!! (use MONKEY-SEE-NO-EVAL to override,␤but only if you're VERY sure your data contains no injection attacks)␤at /tmp/bjADu7u0QY:1␤------> 3my $a = 42; EVAL…»
TimToady m: use MONKEY; my $a = 42; EVAL "say $a"
camelia rakudo-moar a28270: OUTPUT«42␤»
21:34 wamba joined, glucosade left
dalee so if i have a CPAN module that is already on github, and i want to publish a perl 6 port, anyone have any idea on what an appropriate naming scheme should be? 21:35
My-Module-Perl6?
My-Module-6?
21:36 vendethiel joined
CurtisOvidPoe Anyone know why when I run “panda install Linenoise” on with the latest (several) builds I get “Could not find Shell::Command in …:”? 21:36
TimToady maybe we just need some regular naming convention such that it can be removed if the 6-ish modules is dropped into an implicitly p6 namespace
Skarsnik I use perl6-stuff 21:37
for the github name 21:38
RabidGravy dalee, I have just named the *module* exactly the same, the github repo prexixed with p6-
prefixed
TimToady hmm, when I trying install Linenoise, I get: Could not find Native::Resources::Build
dalee CurtisOvidPoe Linenoise seems to install fine via panda for me 21:39
TimToady *try to
Skarsnik Native::Ressouces?
moritz somthing funny in the META.info maybe?
CurtisOvidPoe This is happening on two of my OS X boxes, only one of which is El Capitan. I can’t install any modules :(
Skarsnik Shell::Command is a panda issue
CurtisOvidPoe This happened in the past, but seemed to have been resolved.
Now it’s back. 21:40
Skarsnik What did you use to install perl6/panda?
CurtisOvidPoe Srarksnik: rakudobrew build moar && rakudobrew build panda
RabidGravy goats blood and black candles
lizmat $ ls -ls lib/.precomp | wc -l 21:41
323
Skarsnik I prefer dark purple candles
lizmat looks like we need some cleaning logic as well soonish
RabidGravy hipster ;-p
TimToady wow, after a bootstrap.pl, I get: resolve stage failed for Linenoise: Dependency Native::Resources is not present in the module ecosystem
RabidGravy eugh
21:41 CurtisOvidPoe left
Skarsnik CurtisOvidPoe, did you rakudebrew nuke before updating? 21:42
TimToady he dropped off
21:42 CurtisOvidPoe joined
Skarsnik CurtisOvidPoe, did you rakudebrew nuke before updating? 21:42
TimToady this is running panda without rakudobrew, btw, just in my git dir
21:43 virtualsue left
CurtisOvidPoe Skarsnik: how do I do that? “rakudobrew nuke” doesn’t appear to be a command. 21:43
21:44 schnuppi left
lizmat good night, #perl6! 21:44
Skarsnik Good night lizmat
RabidGravy ah, the Native::Resources in the Linenoise meta 21:45
maybe a "panda update" may help here
TimToady panda update does nothing 21:46
RabidGravy Hmm Native::Resources is definitely there in the ecosystem, maybe its meta is broken 21:48
no looks fine
TimToady afks for a bit * 21:49
RabidGravy is the thing that builds the modules.perl.org idfferent to the thing that generates the project.json 21:50
21:53 sena_kun left
Zoffix RabidGravy, yes 21:54
Shell::Command should be installed when panda's installed
RabidGravy okay, well the Native::Resources is in the projects.json so i dunno 21:55
Zoffix I get '==> Successfully installed Linenoise' with my This is Rakudo version 2015.12-23-ga7ea8e6 built on MoarVM version 2015.12
dalee, I use perl6-Module-Name 21:56
RabidGravy yeah I think it's a case of "unlucky" 21:57
dalee yeah, i'm going with perl6-XXX 21:58
21:58 vendethiel left
Zoffix tries a fresh brew of perl6 21:59
dalek kudo-star-daily: 4989e4c | coke++ | log/ (5 files):
today (automated commit)
RabidGravy but in the end of the day it's just to disambiguate for github's thing so it's just please yourself
dalee is there a perl6 sort of equivalent to rt.cpan.org, or would i perhaps be better off just using the github issue tracker? 22:02
Zoffix dalee, github for now. 22:03
Well, actually I don't know how rt.cpan.org learns about the module queries. That thing kinda always happened by magic to me. 22:04
RabidGravy I've been trying to push people over to github for the P5 modules too tbh
Juerd As a CPAN author, I wish I could disable the magic RT thing. 22:05
RabidGravy Juerd, yeah
Juerd I always forget to look there.
Sometimes people will email me and yell at me for not fixing a bug for over a year, or longer.
22:05 BenGoldberg joined
Zoffix Ditto 22:05
Juerd And then it's the first time I learn about the report.
RabidGravy oh I've got ten year old rt tickets 22:06
Juerd You win :) 22:07
RabidGravy there's one about Term::ReadKey and SIGWINCH and sig_action that I just hope if I ignore it long enpugh it will go away ;-) 22:08
Zoffix CurtisOvidPoe, does your 'rakudobrew build panda' output contain "Successfully installed Shell::Command"? It's getting installed before `panda` itself.
Skarsnik I discover a ticket from 2003. I would prefer getting a mail lol
Zoffix rakudobrew build moar && rakudobrew build panda; panda install Linenoise; successfully completes for me on a new Debian Wheezy user 22:09
CurtisOvidPoe Zoffix: “rakudobrew build panda” prints “rebootstrap is obsolete. You can upgrade your rakudo while keeping all installed distributions including panda.”
22:09 kaare_ left
CurtisOvidPoe A couple of lines later it reads “Done, built panda for moar-nom" 22:10
RabidGravy rakudobrew self-upgrade
Zoffix CurtisOvidPoe, sounds like you could benefit from rm -fr ~/.rakudobrew; rm -fr ~/.perl6; and then rakudobrew build moar && rakudobrew build panda; panda install Linenoise;
CurtisOvidPoe Zoffix: trying now.
Zoffix With the git clone github.com/tadzik/rakudobrew ~/.rakudobrew; step after the rm -fr, of course :P 22:12
CurtisOvidPoe Yeah, got that bit :)
stmuk__ w00t! more 6guts 22:17
RabidGravy mmmm guts 22:19
dalek osystem: 00db9ed | (Dale Evans)++ | META.list:
add Text::Spintax to ecosystem, see github.com/daleevans/perl6-Text-Spintax
osystem: 5315bb5 | moritz++ | META.list:
Merge pull request #117 from daleevans/patch-1

add Text::Spintax to ecosystem, see github.com/daleevans/perl…
Zoffix \o/ moar modules \o/ 22:21
RabidGravy Rarr!
CurtisOvidPoe, when are we going to move Corinna to Perl 6? 22:22
I abandoned the moosifaction as I got busy
CurtisOvidPoe RabidGravy: when I don’t have one company and trying to start a second? :)
Zoffix I see a whole ton of people use "author" : "Foo Bar", instead of "authors": ["Foo Bar"]. 22:23
dalee is it supposed to be authors instead of author?
Zoffix dalee, as per the spec, yeah
dalee if so, Module::Minter should be updated
22:23 ely-se joined
dalee i will go make a pull request for that 22:24
RabidGravy CurtisOvidPoe, slacker ;-) I'll start next week
Zoffix But I'd argue we should allow a single "author" with a string to mean "authors": ["blah"}
dalee++
22:24 pjscott joined
RabidGravy dalee, tell them to put a Test::META test in, that actually fails for author only now 22:24
Zoffix dalee, "perl" key is also mandatory; so, "perl": "6.c"
nine TimToady: what exactly does "panda update does nothing" mean? It does not generate any output if everything works. Have you tried installing Linenoise again after the panda update? 22:25
22:25 pjscott left
RabidGravy which is mandatory per the spec 22:25
CurtisOvidPoe With “rakudobrew build panda”, it still fails with “perl 6 needed, this is perl 5…” if $PERL5LIB env var is set.
dalee what does 6.c mean?
RabidGravy but nothing but Test::META cares about
Zoffix dalee, minimum Perl 6 version your module works with 22:26
RabidGravy CurtisOvidPoe, yeah
dalee so 6.0 at this point, or does 6.c have some meaning?
CurtisOvidPoe Zoffix: thank you. It looks like Panda works fine now.
Zoffix m: say so v6.0 before v6.c
camelia rakudo-moar a28270: OUTPUT«False␤»
nine CurtisOvidPoe: can you please create a github issue for the PERL5LIB thing?
RabidGravy it's the current version of the spec 22:27
Zoffix dalee, 6.c is the current version of the spec. There's no 6.0
CurtisOvidPoe nine: creating it now.
Zoffix dalee, and 6.0 also is a larger version number
nine CurtisOvidPoe++
Zoffix grumbles about versions and goes back to watching TV 22:28
nine It's "versions" not "version numbers" ;)
dalee so actually the key should be "perl6": "v6.c" if i understand design.perl6.org/S22.html correctly 22:29
RabidGravy yeah, number is confusing
dalee, no "v"
orbus anybody know if cas actually works? 22:30
RabidGravy the spec is ambiguous
orbus in the sense of, is it really atomic?
because it seems like it's not, or I'm doing something wrong
m: my $x=0; await start { for ^1000 {cas($x,{$x+1} ) } } xx 8; put $x
camelia rakudo-moar a28270: OUTPUT«3525␤»
dalee well, that doc says Mandatory. The minimal perl version for which this distribution can be installed. Specified as a version string. So:
"perl" : "v6.0.1"
RabidGravy but it's a string that represents a version, so the v is redundant at best 22:31
CurtisOvidPoe nine: github.com/tadzik/rakudobrew/issues/79
RabidGravy m: say Version.new("v6.c").Str
camelia rakudo-moar a28270: OUTPUT«v.6.c␤»
RabidGravy m: say Version.new("v6.c")
camelia rakudo-moar a28270: OUTPUT«vv.6.c␤»
RabidGravy the v becomes part of the version which is not what is meant 22:32
22:32 setty1 joined
ugexe that s22 link is not up to date 22:33
theres plenty of stuff in there that doesnt even exist anymore
orbus m: my $x=0; my $l=Lock.new; await start { for ^1000 { $l.protect({$x++}); } } xx 8; put $x 22:36
camelia rakudo-moar a28270: OUTPUT«8000␤»
orbus locks work at least
RabidGravy ugexe, yeah
nine Current panda will warn if the "perl" version has a leading "v" and will strip it. 22:37
RabidGravy orbus, if locks didn't work we'd be in deep do do
nine Otherwise too many modules would be uninstallable I guess since panda now checks the Perl version
orbus lol
RabidGravy nine, yeah META6 warns and strips, if the meta6 version > 0 then it's fatal 22:38
orbus but seems like cas is not working right, or I've misunderstood how to use it
hmm
sub cas (\val,&code) { val = code(val) } # naive implementation of cas 22:39
oh
yeah
that doesn't look super atomic
ugexe its not implemented
orbus guess it's a placeholder
so I see :)
ugexe at least not when i asked a few months ago 22:40
orbus well, that's what's in src/core/asyncops.pm
so I think you're right
stub implementation
Quom__ sub cas (\val,&code) { die "Uh, oops." } # naive implementation of cas 22:41
orbus heh
22:42 marmay left
RabidGravy I thought I so it implemented, or do I need to lay off the crack? 22:42
22:43 nzkindest joined
orbus well, unless someone implemented it post 2015.12, that straight assignment is what I found in the source 22:43
which would be equivalent as long as you don't actually try to do anything multithreaded 22:44
RabidGravy but isn't it that with a $lock.protect round it? or am I misunderstanding the purpose? 22:45
orbus well, it doesn't have a lock.protect around it
22:46 pjscott joined
orbus and I think the idea was to implement without locks 22:46
but sure
that'd probably be a closer implementation
The Atomic Compare and Swap (CAS) primitive is directly supported by most modern hardware. It has been shown that it can be used to build a whole range of concurrency control mechanisms (such as mutexes and semaphores). It can also be used to implement lock-free data structures.
that's what the synopsis says 22:47
since it's talking about lock-free data structures, I'm assuming the backend implementation shouldn't use locks either
but I am not a concurrency wizard
RabidGravy yeah, so it would have to be right in the VM
orbus probably
guess it's not yet 22:48
in the meantime, an implementation with locks might get closer to expected behavior though
I think you'd have to maintain a global "cas lock" just for that function though? 22:49
22:49 marmay joined
orbus hmm 22:49
ugexe the older implementation was just sub cas (\val,&code) { code(val) }
orbus I don't think that would do what the synopsis says it should 22:50
i.e., swap val to point to the result of code
RabidGravy { my Slock = Lock.new; sub cas(*@args) { Slock.protect({ ..... })}} 22:51
22:51 RUR left
RabidGravy but yeah, in the VM 22:51
orbus right - I just mean there'd have to be a lock defined in core for cas's exclusive use 22:52
22:52 cdg joined
ugexe a proper compare and swap compares if the value is different, and changes it only if so 22:52
hence the compare
orbus well
the synopsis actually talks about two forms 22:53
sub cas($ref is rw, $expected, $new)
and the form with val and block
but only the second one seems to be implemented at all
o support this, there is another &cas candidate that takes a scalar and a block. It calls the block with the seen initial value. The block returns the new, updated value. If nothing else updated the value in the meantime, the reference will be updated. If the CAS fails because another update got in first, the block will be run again, passing in the latest value. 22:54
I'm actually not sure how that's going to work from an implementation perspective 22:55
it seems like you have to store the value of val, run the block, then compare val to "old val"
RabidGravy yeah, the intel cmpxchg op might be a bit difficult to implement for perl types
orbus the first form seems more like what I think of as "compare and swap" 22:56
anyway, I have a number of questions about what operations are atomic in perl6 anyway 22:57
and not much of that is documented
$x++ pretty clearly isn't
add, then assign I guess 22:58
22:58 marmay left
ely-se I want to implement Inline::COBOL. 22:59
RabidGravy do it
orbus COBOL programmers will thank you
all seven of them
m: my $x=0; await start { for ^1000 { $x++ } } xx 8; put $x
camelia rakudo-moar a28270: OUTPUT«7271␤»
RabidGravy I learned cobol
orbus I have a friend who learned it in college 23:00
and he's not that old
geekosaur started out doing cobol
RabidGravy like 34 years ago but I learnt it
geekosaur it's still used in business programming contexts
orbus I know - and from what I understand qualified programmers are in fairly short supply
if you can do it, pays well 23:01
23:01 mohij left
orbus at least that's what I hear 23:01
Juerd rosettacode.org/wiki/Command-line_a...nts#Perl_6 could be improved with nicer examples. Anyone interested in writing something up?
23:02 rindolf left
RabidGravy I think the last cobol I wrote around 1997, I needed to alter the general ledger interface to housing system I was working with (which was in informix 4gl) 23:04
geekosaur (sadly my cobol experience is not that useful in the current business world, nobody uses RM/COBOL any more and it had significant differences from more standard versions) 23:05
23:05 patrickz joined
RabidGravy yeah this was all ICL mainframe stuff 23:05
geekosaur (it doesn;t even exist any more, Micro Focus bought Ryan MacFarlane and killed it) 23:06
Skarsnik Write a cobol interpreter with nqp?
ely-se no just call gnucobol and link to the resulting shared object 23:07
Juerd And rosettacode is offline *again* 23:08
I wonder how they manage to be down so much
ely-se they recently destroyed the perfectly good web design 23:09
CurtisOvidPoe ely-se: there’s an interesting business model with Inline::COBOL if it’s done right :)
ely-se now it hurts my eyes
Juerd ely-se: That's just the CSS being broken
ely-se mediawiki default theme is great
CurtisOvidPoe Though a grammar and macros would be better.
ely-se CurtisOvidPoe: more so if implemented in Perl 5 23:10
CurtisOvidPoe Actually, there’s a much better one for Perl 6, but I don’t want to spam the channel with it right now :)
23:11 mohij joined, pjscott left
CurtisOvidPoe In particular, COBOL’s native decimal types map well to Perl 6 rationals and would be crucial to making it work for large scale accounting systems whcih want to gradually migrate away. 23:11
23:12 patrickz left
ely-se COBOL picture clauses don't map well to Perl 6 rationals however 23:13
orbus m: sub casx(\val,&code){state $l=Lock.new; $l.protect({val=code(val)})}; my $x=0; await start {for ^1000 {casx($x,{$_+1})}} xx 8; put $x;
camelia rakudo-moar a28270: OUTPUT«8000␤»
orbus that seems to do it
although I'm a little unsure about using state there
not sure about the concurrency implications 23:14
but the problem with that implementation is the lock is held for the duration of block
really just the compare and swap at the end should be locked, in the absence of a system level implementation 23:15
given the description in the synopsis
hmmm
23:15 Tonik left
CurtisOvidPoe ely-se: COBOL’s packed decimals seem great for Rats. So long as the data returrned maps cleany to COBOL memory, things should be good (with the caveat that the devil is in the details). 23:16
23:16 nzkindest left, patrickz joined
CurtisOvidPoe I was working on a new business idea around that, but with one business running and another one in the works, I don’t have much free time :) 23:16
RabidGravy I was actually thinking about making a thing whereby you could define the attributes of a class with cobol PIC traits so you could read write strait from a cobol made storage 23:17
straight
geekosaur need more than PIC there 23:18
23:18 mohij left
ely-se I yet have to find a use-case for COBOL. 23:18
geekosaur USAGE {DISPLAY,COMPUTATIONAL,COMPUTATIONAL-3, probably others in various implementations}
(storage format for those is ASCII, binary, packed decimal respectively)
CurtisOvidPoe RabidGravy: I suspect you can’t write directly to storage because we’re talking about multiple OSs, versions of those OSs, versions of COBOL, different machine instruction sets, and so on.
geekosaur well, EBCDIC
depending on platform 23:19
CurtisOvidPoe In fact, some companies have built in-house COBOL compilers with custom extensions due to the fear of COBOL providers going bankrupt.
RabidGravy I made some stuff in Perl a long time ago to pick the bits out of cobol storage to migrate an ancient planning system
geekosaur not to mention that some constraints are not defined at the level of COBOL when the implementation is from an OS that uses a record management abstraction (AS/400, VMS, etc.)
RabidGravy yeah, I haven't really worked in those kinds of enterprises for 15 years 23:21
23:21 vendethiel joined
RabidGravy AS/400 you can just do an sql query on *any* file 23:21
ok plain test may only have one column but hey 23:22
geekosaur yep. tjey don;t have a filesystem. they have a global database.
23:22 Zoffix left
RabidGravy all good 23:25
and don't get me started on idms 23:29
23:30 mohij joined
RabidGravy I guess that went away with icl maiframes 23:30
and cafs 23:31
Juerd Going through these rosettacode things is really annoying if it's down more than it's up... :)
23:31 mohij left
RabidGravy actually I've used cafs on unix with informix 23:31
23:31 TEttinger left, patrickz left
orbus .seen FROGGS 23:33
yoleaux I saw FROGGS 8 Jan 2016 16:37Z in #perl6: <FROGGS> nine: maybe we can remove it later on...
orbus .tell FROGGS failing tests on arm you asked about paste.scsys.co.uk/504181
yoleaux orbus: I'll pass your message to FROGGS.
RabidGravy orbus, I RT'd that 23:34
masak 'night, #perl6 23:35
Juerd @rosettacode.grep(/^C/) checked.
9 items with a C in rosettacode.org/wiki/Category:Perl_..._attention
23:36 nzkindest joined, pjscott joined
RabidGravy orbus, I took a look at the code and I think it's mostly due to getting the alignment wrong on arm 23:36
unfortuntely the build slowness on the RPi militated against me investigating further 23:39
23:42 vendethiel left, pjscott left 23:43 pjscott joined, pjscott left 23:44 pjscott joined 23:47 yqt joined 23:50 vendethiel joined, wollmers joined
wollmers hi rakudistas 23:50
p6: my uint64 $z = +^0;say sprintf("%064b",$z); 23:51
camelia rakudo-moar a28270: OUTPUT«00000000000000000000000000000000000000000000000000000000000000-1␤»
dalek blets: 75874f6 | (Herbert Breunung)++ | docs/appendix- (2 files):
added numeric type
wollmers p6: my uint32 $z = +^0;say sprintf("%032b",$z); 23:52
camelia rakudo-moar a28270: OUTPUT«11111111111111111111111111111111␤»
23:52 wamba left
wollmers p6: my uint64 $z = +^0;say sprintf("%064b",$z-1); 23:52
camelia rakudo-moar a28270: OUTPUT«0000000000000000000000000000000000000000000000000000000000000-10␤»
wollmers p6: my uint64 $z = 1 +<63;say sprintf("%064b",$z); 23:53
camelia rakudo-moar a28270: OUTPUT«-1000000000000000000000000000000000000000000000000000000000000000␤»
RabidGravy the unsigned part in the perl world is only a decoration right now
wollmers What's the current policy regarding changes of the spec-tests? 23:54
RabidGravy it may have meaning when passed into the native world
wollmers The unsigned seem to work so far. The spec-tests for sprintf() are wrong. 23:55
23:56 pjscott left 23:58 cdg left
wollmers @RabidGravy As far as I tested them, uint64 work with bitops. 23:59

Powered by Cro and the Raku® Programming Language.

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

Thank you!