»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.
ugexe .tell pmurias re: `zef clone Foo::Bar` question from last week - this mostly exists already using either `zef fetch Foo::Bar` or `zef look Foo::Bar`. If it happens that Foo::Bar resolves to a .tar.gz (like from cpan) and not a git uri it will extract it for you as well 00:51
yoleaux ugexe: I'll pass your message to pmurias.
bstamour /? 01:01
ugexe ?$/ 01:06
gfldex m: my @tests = { say 1; False }, { say 2; True }, { say 3; True }; say [&&] @tests; 01:50
camelia rakudo-moar d2b115: OUTPUT«2␤3␤True␤»
gfldex i would expect the first block to be executed as well 01:51
gfldex and it's return value to be taken into consideration 01:51
BenGoldberg m: my @tests = { say 1; False }, { say 2; True }, { say 3; True }; @tests.say 02:06
camelia rakudo-moar d2b115: OUTPUT«[-> ;; $_? is raw { #`(Block|80637344) ... } -> ;; $_? is raw { #`(Block|80637416) ... } -> ;; $_? is raw { #`(Block|80637488) ... }]␤»
BenGoldberg m: my @tests = { say 1; False }, { say 2; True }, { say 3; True }; so $_ for @tests;
camelia rakudo-moar d2b115: OUTPUT«WARNINGS for <tmp>:␤Useless use of "so " in expression "so $_" in sink context (line 1)␤»
BenGoldberg m: my @tests = { say 1; False }, { say 2; True }, { say 3; True }; $_() for @tests; 02:07
camelia rakudo-moar d2b115: OUTPUT«1␤2␤3␤»
BenGoldberg m: my @tests = { say 1; False }, { say 2; True }, { say 3; True }; say all(@tests);
camelia rakudo-moar d2b115: OUTPUT«all(-> ;; $_? is raw { #`(Block|69009920) ... }, -> ;; $_? is raw { #`(Block|69009992) ... }, -> ;; $_? is raw { #`(Block|69010064) ... })␤»
BenGoldberg m: my @tests = { say 1; False }, { say 2; True }, { say 3; True }; say so all(@tests);
camelia rakudo-moar d2b115: OUTPUT«True␤»
BenGoldberg m: my @tests = { say 1; False }, { say 2; True }, { say 3; True }; say [&&] @tests;
camelia rakudo-moar d2b115: OUTPUT«2␤3␤True␤»
BenGoldberg m: my @tests = { say 1; False }, { say 2; True }, { say 3; True }; say [&&] !@tests; 02:08
camelia rakudo-moar d2b115: OUTPUT«False␤»
BenGoldberg m: my @tests = { say 1; False }, { say 2; True }, { say 3; True }; say [&&] |@tests;
camelia rakudo-moar d2b115: OUTPUT«2␤3␤True␤»
BenGoldberg I don't understand why any of those are getting called.
gfldex m: say Regex ~~ Callable; 02:09
camelia rakudo-moar d2b115: OUTPUT«True␤»
BenGoldberg Is it a bug?
gfldex that could be the reason
BenGoldberg ?
BenGoldberg doesn't get it. 02:09
gfldex my guess is there is a ~~ somewhere and that is getting the Block called 02:12
m: my @tests = { say 1; True }, { say 2; True }, { say 3; True }; multi sub tester(&test){ test() }; multi sub tester([]){ True }; multi sub tester([&test, *@rest]) { test() && tester(@rest) }; say so tester(@tests);
camelia rakudo-moar d2b115: OUTPUT«1␤2␤3␤True␤»
gfldex that works as a short circuiting && 02:13
it's a bit long tho
Nelsen Good evening, mates 02:14
I was wondering if there was a "Codecademy"-like site to learn Perl?
zostay m: say "test.foo".trans('.' => '::'); 02:42
camelia rakudo-moar d2b115: OUTPUT«test:foo␤»
zostay m: say "test.foo".trans(['.'] => ['::']);
camelia rakudo-moar d2b115: OUTPUT«test::foo␤»
zostay that seems like a bug
MasterDuke zostay: i got bit by that too, but it's in the docs 02:44
zostay then the docs are too subtle for me, i don't read that in the docs at all 02:46
MasterDuke well, it's not really spelled out here ( docs.perl6.org/routine/trans ), but it's shown in the examples 02:48
zostay which example? 02:49
gfldex examples ain't no roast 02:50
it's a bug
MasterDuke $str.=trans( [ '<' , '>' , '&' ] => [ '&lt;', '&gt;', '&amp;' ]);
gfldex the examples got it because they are meant to work with rakudo
zostay: please rakudobug
zostay that does not say that $str.=trans('<' => '&lt;', '>' => '&gt;'); shouldn't work...thx, gfldex 02:51
MasterDuke github.com/perl6/roast/blob/master...rans.t#L55 02:52
(not that i like the current behavior, just playing devil's advocate here) 02:53
zostay at the very least, the docs are incomplete... but i think there's an opportunity to call this a bug still ;) 02:53
gfldex it should either work or complain if it got a multi char replacement. Silently dropping stuff is LTA. 02:55
MasterDuke doesn't that match Perl 5's behavior? 02:57
zostay gfldex, that's what i just suggested on the bug i posted... either warn and fix docs or make the behavior consistent (with my personal vote being for the latter) 03:00
i don't think there's an analogous function to .trans in perl5, tr/// does not have comparable calling conventions at all 03:01
.trans() is awesome as a multiple string find and replace at the same time thing
TimToady R3b0zak. 04:34
dur
so much for that password :) 04:35
grondilu lol 04:36
grondilu hey TimToady did you have a look at github.com/perl6/roast/issues/147 ? 04:37
TimToady I don't see how to allow that, offhand, given the list is evaluated before anything is declared 04:41
something fiddly like that probably wants to use constants anyway
grondilu then it wouldn't be a list but a syntactic construct. Like in C I suppose. 04:42
TimToady in p6 it's just a general list, which you could generate from any compile-time values 04:42
grondilu then allow a syntax that differ from the syntax that uses a list. 04:43
TimToady it seems like kind of a rare thing to want
grondilu :/
that may be true. 04:44
TimToady bein's you're the first person to ask for it in 15 years... 04:44
grondilu yet it bothered me badly that as I was looking at a C header I realized I could not make a straightforward translation.
(well, "bothered me badly" that was maybe strong an expression) 04:46
it was frustrating though.
grondilu also, it's not as if NativeCall has existed for 15 years, is it? 04:47
ShimmerFairy fwiw I ran into the same issue once when using enums in P6. The use case is typically to define "synonyms" for some reason or another. 06:28
I'd also like to tell enums not to pollute the namespace with unqualified names, while we're at it. 06:29
moritz ShimmerFairy: you might be able to do that with anonymous enums 06:32
moritz constant EnumName = enum <a b c>; 06:33
or something like that
haven't tested it
grondilu is not sure about what difference there is between constant EnumName = enum <a b c> and enum EnumName <a b c> 06:41
grondilu m: constant FooBar = enum <Foo Bar> 06:42
camelia ( no output )
grondilu m: constant FooBar = enum <Foo Bar>; say FooBar.WHAT
camelia rakudo-moar d2b115: OUTPUT«(Map)␤»
grondilu m: enum FooBar <Foo Bar>
camelia ( no output )
grondilu m: enum FooBar <Foo Bar>; say FooBar.WHAT
camelia rakudo-moar d2b115: OUTPUT«(FooBar)␤»
ShimmerFairy moritz: not synonyms to the enum, but synonyms for its values. In C++, something like e.g. enum class Feature { Off, OneChoice, TwoChoice, On = OneChoice }; for a feature that used to be binary. 06:51
shinobicl_ m: sub postcircumfix:<' '>($it, $arg, :$double) { say $double ?? 2 * $it * $arg !! $it * $arg }; 2'3'; 2'3':double 06:54
camelia rakudo-moar d2b115: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Unable to parse expression in postcircumfix:sym<' '>; couldn't find final $stopper ␤at <tmp>:1␤------> 3 ?? 2 * $it * $arg !! $it * $arg }; 2'3'7⏏5; 2'3':double␤»
moritz ShimmerFairy: I wasn't talking about the synonyms aspects, rather about not polluting your namespace with short names 06:56
ShimmerFairy ah 06:57
masak morning, #perl6 07:05
moritz \o masak
masak m: enum Feature <Off OneChoice TwoChoice>; constant On = OneChoice; say On 07:06
camelia rakudo-moar d2b115: OUTPUT«OneChoice␤»
ShimmerFairy m: enum Feature <Off OneChoice TwoChoice>; constant On = OneChoice; say Feature::On # if you, like me, don't much care for the namespace pollution of 'traditional' enums 07:08
camelia rakudo-moar d2b115: OUTPUT«Could not find symbol '&On'␤ in block <unit> at <tmp> line 1␤␤Actually thrown at:␤ in block <unit> at <tmp> line 1␤␤»
masak you defined `On`, not `Feature::On` 07:10
there's nothing magical going on with that `constant` statement. it's just a name binding. 07:11
ShimmerFairy masak: right, I'm just pointing out that I'd _like_ to be defining Feature::On instead :)
masak m: enum Feature (Off => 0, OneChoice => 1, TwoChoice => 2, On => 1); say Feature::On 07:12
camelia rakudo-moar d2b115: OUTPUT«On␤»
ShimmerFairy The issue with that though is that it's not semantically correct, and that it requires remembering that you have to keep the two options in sync. 07:13
masak you *almost* sound like you have a spec patch coming up :) 07:17
moritz and a roast and rakudo patch :-) 07:20
masak and lots of good feedback on the respective PRs \o/ 07:21
ShimmerFairy fwiw, <TimToady> I don't see how to allow that, offhand, given the list is evaluated before anything is declared 07:22
and there's already a roast issue here :P github.com/perl6/roast/issues/147
masak so, you're saying what you want to do is only impossible once? :P
masak fwiw, grondilu's `enum ("foo", bar => foo);` can't work -- "foo" is just a string, and so there is no `foo` identifier declared when using it as a pair value 07:24
guess this is exactly what TimToady is pointing out
ShimmerFairy nah, I personally think that enums don't necessarily need to keep that behavior :) . In fact I think it would be nice if they worked like function signatures: sub foo($a, $b = $a) { ... }
I can see why that example in particular doesn't work, though. 07:25
masak that `("foo", bar => foo)` thing is literally a list expression. much less magical than a parameter list. 07:26
enums currently ride on the fact that it gets lists in various forms.
ShimmerFairy right, but considering it's a 'package'-style thing to declare, maybe they could stand to be more magical?
masak: I ultimately think that Perl 6's enums are poorly designed. Sure they have type safety, and EnumClass($out-of-range) won't produce a valid defined object (like it does in C++), but they don't really feel worthwhile to me otherwise. 07:28
masak that opinion is fine. 07:28
I disagree, though. 07:29
you should have seen back when they *were* badly designed :P
ShimmerFairy masak: In C++ land, I always use 'enum class' over 'enum', where my only sticking point is that doing that kind of conversion I mentioned will live perfectly fine. But otherwise I like them :) 07:30
masak ShimmerFairy: maybe macros will be able to help you in the future. 07:33
I love being able to *say* that with enough confidence nowadays! :D
leont ShimmerFairy: yeah, enum class is quite awesome :-) 07:35
ShimmerFairy masak: maybe so, as long as it's easy to fix broken-to-me language aspects, there's not much reason to complain :)
leont: agreed, the only issues I have are of the kind that you get for a low-level language: "type conversion doesn't check if the int is in range!" "why won't you give me a postfix:<++> that DWIMs!?" 07:36
(and even on the latter one, I personally don't see it as a really big issue)
leont The main thing I miss is methods on enum classes ;-) 07:37
ShimmerFairy Well, introspection is a separate issue :P 07:38
(imagine if C++ had Signature and Namespace types :P)
leont Signatures can be largely done at compile-time, actually 07:39
ShimmerFairy Honestly, I find C++ to be just as interesting as Perl 6, but coming from the opposite direction on the low-level/high-level ladder 07:42
leont The interesting thing about C++ is how much it is still in development 07:45
C++11 like Perl6 took way too long, but it worth it
leont At least with the new versions they're iterative now. C++14 was a nice minor update and C++17 will bring new awesomeness 07:45
ShimmerFairy leont: and C++11 taught them to maintain their current faster release cycle too, so that's good :)
El_Che wut? C++17? What shall I do with my 1996 book? 09:52
DrForr Burn it, of course. Everything magically corresponds to the new spec. 10:01
andreoss m: multi MAIN() { die $?USAGE }; multi MAIN(Str :$option) { say "hi "}; 11:29
camelia rakudo-moar d2b115: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Variable '$?USAGE' is not declared␤at <tmp>:1␤------> 3multi MAIN() { die 7⏏5$?USAGE }; multi MAIN(Str :$option) { sa␤»
andreoss m: multi MAIN() { die &?USAGE }; multi MAIN(Str :$option) { say "hi "};
camelia rakudo-moar d2b115: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Undeclared name:␤ ?USAGE used at line 1␤␤»
andreoss m: multi MAIN() { die $*USAGE }; multi MAIN(Str :$option) { say "hi "}; 11:31
camelia rakudo-moar d2b115: OUTPUT«hi ␤»
andreoss m: multi MAIN() { die $*USAGE }; multi MAIN(Str :$option) { say $option.perl }; 11:32
camelia rakudo-moar d2b115: OUTPUT«Str␤»
andreoss m: multi MAIN() { die $*USAGE }; multi MAIN(Str:D :$option) { say $option };
camelia rakudo-moar d2b115: OUTPUT«Dynamic variable $*USAGE not found␤ in sub MAIN at <tmp> line 1␤ in block <unit> at <tmp> line 1␤␤»
andreoss how do i get this usage message? 11:33
_4d47 mm how can this method gets called ? github.com/rakudo/rakudo/blob/nom/...ay.pm#L419 11:52
TheLemonMan _4d47, Array.new(VV, :shape(NN)) 11:55
_4d47 TheLemonMan: aww thanks, my eyes where reading Mu:D: instead of Mu:D 11:59
MasterDuke andreoss: $?USAGE isn't implemented yet 11:59
and i believe MAIN will just call USAGE() if it can't dispatch. there's an autogenerated one, but if you create your own sub USAGE it will get called instead
RabidGravy boom 12:27
MasterDuke big bada boom? 12:28
RabidGravy it's amazing how shutting down all those spreadsheet instances frees up memory
anyhow reboot
ab6tract clickbaits blogs.perl.org/users/ab5tract/2016/...print.html 12:30
TimToady bl 20 12:44
Xliff \o 12:49
Morning all.
gist.github.com/Xliff/fceb9354193e...68b991d6ec
^ halp!
Xliff Oh. Lemme update that a sec. Because it's a weird edge case. 12:51
OK. Updated 12:53
nine Xliff: looks terrible (the error message) 12:55
Xliff: you could try package TheTest { ... }; sub EXPORT(+@a) { ... } 12:56
i.e. make sure TheTest is defined before you refer to it 12:57
ab6tract shouldn't EXPORT be inside the respective package? 12:58
nine It belongs to the compilation unit and has to be in the top level. 12:58
AFAIK
Xliff nine: OK, one sec. 12:59
ab6tract One day I'm going to have to learn me some of the distinctions between a compilation unit vs a file vs a 'unit module' 13:00
Xliff nine: No change.
ab6tract nine: so then if i do a 'unit class', the EXPORT will work because that's the top level of the compunit?
Xliff ab6tract: The original draft of the code had it that way, but that wouldn't parse.
El_Che ab6tract: looks empty 13:01
Xliff ab6tract: Or gave strange errors, depending.
ab6tract El_Che: the blog post?
El_Che ab6tract: yes
ab6tract El_Che: try a refresh please
that was my fault 13:02
El_Che better :)
Xliff nine: Another weird thing... if I drop the signature from EXPORT, it still fails in the same manner.
For all 3 cases.
ab6tract thanks for pointing it out :)
Xliff Another odd thing: If I return an empty map from EXPORT, the error message goes away for all cases. 13:04
Should I RT?
ab6tract Xliff: I am using EXPORT::DEFAULT like this github.com/ab5tract/Terminal-Print...t.pm6#L324
and it seems to be working so far
Xliff ab6tract: Yes, but is that a map definition or a sub? 13:05
Xliff On first glance, looked like a map. I will try. 13:06
ab6tract it is neither, AFAICT :)
it seems to be autotriggered on import
i'm just doing some binding into the OUR:: namespace there 13:07
Xliff SSDD 13:08
(still getting the error)
ab6tract Xliff: have you updated the gist? 13:13
Xliff About to. I think I found a work around. 13:15
Xliff Updated. 13:17
Yup! Workaround applied to original code and is working! \o/ 13:20
ab6tract Xliff++ 13:28
Xliff The more I write in P6, the more I find that the block notation is more reliable than using "unit ..." 13:29
mst hmm 13:30
this sort of sounds like maybe a case for finding all the tests using block notation and making sure there's a unit version
Xliff ¯\_(ツ)_/¯ 13:31
m/...(>.<)...m/
Well, after this I get to move to nqp.... 13:32
Xliff Still worried that's a bit over my head at this point, but hey... 13:33
TimToady sounds to me more like problems with how EXPORT works; 'unit' is conceptually very simple
unit should probably be disallowing those declarations before it, however, since the point of unit is to wrap the entire file as if it were a block declaration 13:37
jnthn Does unit sneakily introduce a lexical scope? 13:39
TimToady probably
jnthn That could be surprising in that a sub EXPORT is meant to be located in UNIT 13:40
TimToady it might be assuming wrongly that it's just about the first thing in the file
jnthn So folks doing `unit class Foo; sub EXPORT() { ... }` could get caught out
TimToady 'tis possible
would have to glare at the code some time after I wake up :) 13:41
RabidGravy heads up to anyone using chrome on FC24 (I think timotimo does,) if it offers an update of it, you are better just uninstalling and re-installing first
Xliff My original version of the module this popped up in had EXPORT located after the "unit" declaration. 13:51
I could never get that to work and the error messages were less than helpful when trying to figure out a solution. 13:52
Should I RT this issue?
Here's is a writeup, with solution... gist.github.com/Xliff/fceb9354193e...68b991d6ec
skids There is RT#127305 13:55
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=127305
skids Lots of metaish modules already forgo the unit form. 13:56
TimToady there is precisely one test of 'sub EXPORT', and it doesn't involve 'unit'
it pulls in a file that reads:
class Foo {}
sub EXPORT(|){ { 'Bar' => Foo } }
which seems...a bit rudimentary 13:57
RabidGravy not exhaustive for sure 14:00
TimToady jnthn: yes, it calls <.newpad>, which it should probably avoid for a 'unit' declaration 14:10
TimToady experiments with the simple fix first to see how much stuff blows up 14:13
TimToady
.oO(phase one bug trials)
 14:15
oh, heh, if I refrain from pushing a lexpad, I should also refrain from popping it... 14:16
timotimo RabidGravy: can you explain that a bit more? the chrome thing? 14:17
also, they no longer call it "Fedora Core" ;)
TimToady Fedora Not-So-Core? 14:18
timotimo fedora shell? 14:19
fedora fruit?
RabidGravy timotimo, the upgrade seems to mess up the chain of links for the "alternatives" resulting in it disappearing completely
timotimo hah, that's fun 14:21
TimToady well, simple fix is not simple, insofar as it blows up tools/build/install-core-dist.pl 14:22
timotimo :|
TimToady maybe a 'make clean' issue
but if there's code that depends on a unit's OUTER meaning UNIT, we have a problem 14:23
we have a problem 14:27
getting a push on an NQPMu, probably something assuming OUTER gets to UNIT somewhere, I'm guessing 14:28
TimToady well, this will take at last another cuppa 14:29
mspo ugh alternatives 14:33
timotimo with alternatives it seems like i'm always getting a 32bit wine binary but a 64bit wineserver binary served, or the other way around 14:35
every single time i want to run a windows program i have to shuffle the two around
it's hilariously bad
but perhaps it's because i get at least one wine update between two events of wanting to run any windows program
andreoss m: my $x = { $_++ }; $_ = 10 ;$x();$x();say $_; 14:40
camelia rakudo-moar 6dab6b: OUTPUT«12␤»
andreoss m: my $x = sub { $_++ }; $_ = 10 ;$x();$x();say $_;
camelia rakudo-moar 6dab6b: OUTPUT«10␤»
timotimo yeah, bare blocks don't get their own $_ 14:41
m: my $x = -> { $_++ }; $_ = 10 ;$x();$x();say $_;
camelia rakudo-moar 6dab6b: OUTPUT«12␤»
timotimo like pointy blocks
TimToady well, fixed the NQPMu.push, but now install-core-dist merely runs forever 14:45
nwc10 lizmat: domm is down to 5 minutes, so would be useful if your workshop could get ready to return 14:47
(he's still speaking - not yet time to invade the room)
lizmat ok, I put out the message 14:48
pmurias ab5tract: using 'no precompilation' seems horrible 14:49
yoleaux 00:51Z <ugexe> pmurias: re: `zef clone Foo::Bar` question from last week - this mostly exists already using either `zef fetch Foo::Bar` or `zef look Foo::Bar`. If it happens that Foo::Bar resolves to a .tar.gz (like from cpan) and not a git uri it will extract it for you as well
pmurias is also slightly upset people are using Terminal:: instead of the traditional Term:: 14:51
TimToady I carped about that at the time, but nothing changed 14:52
timotimo pmurias: i think we "recently" got someone ranting about how Term collides with the term "term"
though i think if you had a term, you'd term it P6::Term or something, that'd be a better term for term
TimToady Term vs term is not even the same case, if people are worried about grammar collisions 14:53
"term" is such a generic term that I don't see any collision with people wanting to use that for a name root 14:54
timotimo what, modules are case sensitive? why can't i install ACME::MEOW :(
TimToady and Term has a long history of meaning "terminal" at the root of the namespace 14:55
next thing you know, we'll all be starting our programs with "IDENTIFICATION DIVISION" or so, just because longer is always clearer... 14:56
timotimo "THIS IS THE PLACE WHERE THE NAME OF THE PROGRAM GOES, BEN. PLEASE PUT THE NAME HERE THERE IS NO SUCH THING AS A PROGRAM NAME DIVISION GEEZ"
TimToady and PERFORM is obviously better than DO, and COMPUTE is obviously better than, well, nothing 14:57
timotimo every mathematical expression needs to start with COMPUTE 14:58
DrForr STORING RESULT INTO...
[Coke] WE HAVE THE BEST PROGRAM NAMES.
timotimo afkbbl
TimToady anyway, I agree with pmurias++ here
DrForr JOY IDENTIFICATION DIVISION 14:58
ugexe just use the sha1 as the module name 14:59
pmurias ugexe: unfortunately the modules end up as sha1 in stacktraces anyways 15:02
ab5tract: the only thing 'no precompilation' in Terminal::Print::Dimensions seems to be doing is wasting CPU time 15:04
MetaZoffix m: class Meow { has $!s = Supplier.new; has $.sup = $!s.Supply; method do-things { $!s.emit: $_ for <a b c>; } }; react { with Meow.new { whenever .sup { $^m.say }; .do-things; }; done } 15:11
camelia rakudo-moar e8d0d0: OUTPUT«a␤b␤c␤»
MetaZoffix Any idea why this isn't printing the stuff with whenevers in my code even tho it works in the example above? I got my react block github.com/zoffixznet/na/blob/mast...nner.p6#L6 and my Meow class that emits messages github.com/zoffixznet/na/blob/mast...er.pm6#L33 15:12
And in the output I do get output from `say "Got stuff: $mes";` that's a line above the emits, but the whenevers don't react and don't print anything 15:13
[Coke] tries to figure out his bailador/start problem now that the latest crisis is over 15:13
MetaZoffix And even if I put $!shell-messenger.emit: "built stuff!!!!!"; above this `given` I get the message received :S weird github.com/zoffixznet/na/blob/mast...er.pm6#L19 15:17
smls m: say (1..*).rotor(1..*)[^5]; # Floyd's triangle 15:18
camelia rakudo-moar e8d0d0: OUTPUT«((1) (2 3) (4 5 6) (7 8 9 10) (11 12 13 14 15))␤»
smls m: say (1..*).rotor(1...*)[^5]; # Floyd's triangle
smls ^^ Why does it work with a Range argument but not a Seq? 15:19
camelia rakudo-moar e8d0d0: OUTPUT«(timeout)»
smls Does it have to do with how parameters are passed to routines, or does `rotor` mishandle it?
MetaZoffix s: (1..*), "rotor", \(1...*) 15:20
SourceBaby MetaZoffix, Something's wrong: ␤ERR: Type check failed in binding to &code; expected Callable but got Method+{<anon|70743856>} (Method+{<anon|7074385...)␤ in sub do-sourcery at /home/zoffix/services/lib/CoreHackers-Sourcery/lib/CoreHackers/Sourcery.pm6 (CoreHackers::Sourcery) line 42␤ in sub sourcery at /home/zoffix/services/lib/CoreHackers-Sourcery/lib/CoreHackers/Sourcery.pm6 (CoreHackers::Sourcery) line 29␤ in block <unit> at -e li
MetaZoffix s: (1..*), "rotor", \(1..*)
SourceBaby MetaZoffix, Something's wrong: ␤ERR: Unhandled exception: While looking for '/home/zoffix/services/sourceable/building-perl6/perl6.moarvm': no such file or directory␤
MetaZoffix great
s: (1..*), "rotor", \(1..*)
SourceBaby MetaZoffix, Something's wrong: ␤ERR: Unhandled exception: Missing or wrong version of dependency 'src/Perl6/Grammar.nqp' (from 'gen/moar/m-main.nqp')␤ at <unknown>:1 (/home/zoffix/services/sourceable/building-perl6/perl6.moarvm:<dependencies+deserialize>)␤
Xliff smls: WAG - Eager vs lazy? 15:21
MetaZoffix m: my $w = 1...*; say $w.list[^5]
there's yo problem
camelia rakudo-moar e8d0d0: OUTPUT«(timeout)»
Xliff m: my $w = 1..%; say $w.list[^5]
camelia rakudo-moar e8d0d0: OUTPUT«(Nil Nil Nil Nil Nil)␤» 15:22
[Coke] m: my $w := 1...*; say $w.list[^5]
camelia rakudo-moar e8d0d0: OUTPUT«(1 2 3 4 5)␤»
Xliff m: my $w = 1..*; say $w.list[^5]
camelia rakudo-moar e8d0d0: OUTPUT«(1 2 3 4 5)␤»
MetaZoffix smls: rotor takes a List:D, so for Seqs it goes through Any, which calls .list which loses laziness
Xliff \o/
My WAG was right!?
MetaZoffix smls: it may or may not be an LTA :) feel free to rakudobug
smls m: say (1..*).^find_method("rotor").signature 15:23
camelia rakudo-moar e8d0d0: OUTPUT«($: |c is raw)␤»
timotimo "WAG"?
Xliff Wild Assed Guess
timotimo oh
smls MetaZoffix: Where does it say it takes a List:D?
Xliff =)
MetaZoffix s: (1..*), "rotor", \(1..*) 15:24
SourceBaby MetaZoffix, Something's wrong: ␤ERR: Type check failed in binding to &code; expected Callable but got Method+{<anon|73357248>} (Method+{<anon|7335724...)␤ in sub do-sourcery at /home/zoffix/services/lib/CoreHackers-Sourcery/lib/CoreHackers/Sourcery.pm6 (CoreHackers::Sourcery) line 42␤ in sub sourcery at /home/zoffix/services/lib/CoreHackers-Sourcery/lib/CoreHackers/Sourcery.pm6 (CoreHackers::Sourcery) line 29␤ in block <unit> at -e li
MetaZoffix I hate that bug
smls: in the source code: github.com/rakudo/rakudo/blob/nom/...t.pm#L1152
smls m: my $w = (1...*); say $w.list[^5] 15:25
MetaZoffix smls: oh, wait, I'm even looking at the wrong thing. It's the divisors that are the problem >_<
camelia rakudo-moar e8d0d0: OUTPUT«(1 2 3 4 5)␤»
smls ^^ MetaZoffix: Your timeout earlier was because missing parens
MetaZoffix right
MetaZoffix m: sub foo ( *@what ) { say @what[^5] }( 1...* ) 15:26
camelia rakudo-moar e8d0d0: OUTPUT«(1 2 3 4 5)␤»
MetaZoffix m: sub foo ( *@what ) { say @what.is-lazy }( 1...* )
:/ 15:27
camelia rakudo-moar e8d0d0: OUTPUT«(timeout)»
MetaZoffix m: sub foo ( *@what ) { say @what.is-lazy }( 1...* )
timotimo potentially only works with +@?
camelia rakudo-moar e8d0d0: OUTPUT«(timeout)»
smls That one seems related to RT #128201
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=128201
MetaZoffix m: sub foo ( +@what ) { say @what.is-lazy }( 1...* ) 15:28
camelia rakudo-moar e8d0d0: OUTPUT«True␤»
MetaZoffix bisect: m: sub foo ( *@what ) { say @what.is-lazy }( 1...* )
oh shit.. that was probably a bad idea
bisect: m: Promise.in(2).then: {exit}; sub foo ( *@what ) { say @what.is-lazy }( 1...* )
smls: seems to be yeah. 'cause .rotor is checking .is-lazy: github.com/rakudo/rakudo/blob/nom/...t.pm#L1154 15:29
MetaZoffix &
MetaZoffix I got distracted and totally forgot that I originally came here with my emit issue :) Any idea why it's not emitting? 15:33
timotimo what isn't emitting? 15:36
MetaZoffix timotimo: my code, about two screenfuls above. irclog.perlgeek.de/perl6/2016-09-02#i_13139346 15:37
timotimo thank you
MetaZoffix Even though small sample works... even though doing $!shell-messenger.emit: $mes; in another method works. Even though putting "say $mess" in the original method with .emit also prints stuff :S
timotimo very distracted 15:41
MetaZoffix figured it out by writing it like this: gist.github.com/zoffixznet/ee5e2a1...60f79b5831 15:48
awwaiid hoelzro: Add yourself to the STL unsession I made for Rakudo Perl 6, github.com/strangeloop/StrangeLoop...Unsessions
MetaZoffix Seems it's getting its head screwed when the method calls are inside the react{}
awwaiid and anyone else going to StrangeLoop as well :)
MetaZoffix Maybe I should try a Channel instead of Supplies in here
MetaZoffix Thanks for the help! 15:56
ab6tract pmurias: then someone fixed a bug in precompilation, because previous editions of this library would get a single value for terminal width and height 16:16
pmurias: WRT to 'term' vs 'terminal' -- sorry, I cannot disagree more
single value for each of width and height, which will be wrong if the terminal is resized after precompilation 16:17
if you find that behavior to be in error, please submit an issue
I think it would be a crying shame if P6 ended up with P5's namespacing by fiat
ab6tract after some initial testing, I am seeing that removing no precompilation does indeed have the undesired effect of pinning the grid size of the first invocation 16:25
pmurias: if you find it behaving otherwise, I would be interested to know about it
[Coke] .seen supernovus 16:27
yoleaux I saw supernovus 22 Apr 2016 23:27Z in #perl6: <supernovus> Well, I'm going to have to run. Have a great day/night everyone. I hope to fix up some of my long neglected libraries at some point when I'm not completely overloaded with work! :-)
ab6tract further, it seems to settle the size at precompilation time, which means it isn't even accurate to the size you had when you first ran it 16:28
[Coke] .seen ufobat
yoleaux I saw ufobat 27 Aug 2016 06:53Z in #perl6: <ufobat> yay :D
hoelzro awwaiid: done! 16:41
tailgate Is Pod::To::HTML a library that comes with rakudo-star? I tried deleting all copies of it on my computer, but it's still loading it 16:46
how do I find where it's getting the file from?
ab6tract .tell pmurias I've tried to make the the 'no precompilation' hit as minimal as possible. are you using a terminal client which automatically updates COLUMNS and ROWS in the environment for you as you resize? 16:50
yoleaux ab6tract: I'll pass your message to pmurias.
timotimo rakudo has Pod::To::Text, nothing else in Pod::
bioduds hey guys 17:03
im not able to rua a one liner 17:04
only a 2 liner
this is the best I can do so far
67.205.136.118/install.sh
timotimo how does it not work? 17:05
i mean, what's wrong?
timotimo hm. you set PREFIX, is that enough? i thought we have to --prefix in the configure line, too 17:08
and if you put stuff into /usr/local, it's not going to be able to install anything into the prefix
because you're asking the user to run it as "sh" not as "sudo sh" 17:09
and also:
compiling and such shouldn't be done as root
so maybe it'd be better to extract tha building into three steps so we can "sudo make install" only
bioduds --PREFIX ? 17:10
how so?
timotimo no, --prefix
timotimo i've never tried to use the prefix env var 17:10
also, i think if you don't "export" it, it won't land in the environment anyway 17:11
bioduds i believe that lies the magic of making all work, but I really dont quite understand what it does
bioduds what im testing now is to run the following line: curl 67.205.136.118/install.sh | sh; . ~/rakudo/setpath.sh 17:12
as it would actually be 2 lines in one
that would be the 'ONE' liner
timotimo ok. at what point does it break?
bioduds it does not break
it installs fine
you just need to . ~/rakudo/setpath.sh manually after 17:13
timotimo oh
bioduds which Im hoping to have already on with ;. ~/rakudo/setpath.sh
testing now, if works, should be it 17:14
bioduds WORKED 17:20
ONE LINER IS HERE
its actually a 2-liner
so
right now you may do: curl 67.205.136.118/install.sh | sh ;. ~/rakudo/setpath.sh 17:21
and it will install perfectly
mspo that looks totally safe 17:24
sudo it?
bioduds no need
simply run it
mspo was joke :) 17:25
bioduds oh
:)
now, there need to be SSL on to have it on install.perl6.org
making the 'one liner': curl install.perl6.org | sh;. ~/rakudo/setpath.sh
which will work like a charm 17:26
anyone would care to test it on a mac?
sjohnson let's encrypt!
bioduds i believe it will work on mac too
mspo The certificate is only valid for the following names: design.perl6.org, doc.perl6.org, docs.perl6.org, modules.perl6.org, perl6.org, tablets.perl6.org, www.p6c.org, www.perl6.org 17:27
timotimo we've already got SSL for the rest of perl6.org, no?
bioduds im gonna start testing on a few other linux distros as well
mspo new firefox doesn't have a way to actually view the cert 17:28
that's really annoying
sjohnson it does, you just gotta dig deeper.
mspo I just get the www page from install
timotimo protect the innocent users who don't know what ssl is
sjohnson design.perl6.org, doc.perl6.org, docs.perl6.org, modules.perl6.org, perl6.org, tablets.perl6.org, www.p6c.org, www.perl6.org 17:29
looks like my list matches mspo.
bioduds im going to put up a test list and once its tested on all ill send you guys the table 17:30
japhb ab5tract: Thank you for writing Terminal::Print. I've been putting off doing a couple projects (both work and personal) because I didn't feel like a giant yak shave. You've just removed several layers of yak hair. :-) 17:37
timotimo do we have a certificate behind where we download the rakudo star tarball from? 17:38
bioduds no 17:47
there should be
rakudo.org/downloads/star/rakudo-s....04.tar.gz instead of rakudo.org/downloads/star/rakudo-st....04.tar.gz 17:48
mspo The certificate is only valid for the following names: host.pmichaud.com, www.host.pmichaud.com 17:51
timotimo damn 18:03
bioduds this whole ssl mob is gruesome 18:04
mspo where is perl6.org hosted? 18:05
bioduds tests on ubuntu 14.04 LTS 64 bits with 2 cpus and 2gb ram ran ok 18:05
mspo p6c
why isn't rakudo just uploading to there?
bioduds testing now with 14.04 1 cpu 1 gb ram
digital ocean vps's
gonna try them all
writing a sheet with results too 18:06
wow 18:07
1gb failed
this install script is greedy man
DrForr Perl6 ain't Lua. 18:08
mspo yeah 18:09
bioduds testing CentOS 7.2 x64 2gb ram 2 cpus dig. ocean
timotimo yeah, the core setting is kind of big 18:10
kyclark Given a list of strings, how would I efficiently find all those starting with a given list of prefixes? 18:12
my @list = <foo bar baz quux>
my @want = <fo ba>
I feel something like " @list ~~ any(@want)"
Actually, I just need to do one string at a time, e.g., "foo ~~ any(@want)" 18:13
kyclark @want.map({ 'foo' ~~ /^$_/ }) 18:13
TestNinja m: my @list = <foo bar baz quux>; my @want = <fo ba>; say @want.grep: *.starts-with: @list.any 18:15
camelia rakudo-moar e8d0d0: OUTPUT«()␤»
TestNinja orly
m: my @list = <foo bar baz quux>; my @want = <fo ba>; say @want.grep: /^@list/
camelia rakudo-moar e8d0d0: OUTPUT«()␤»
TestNinja oh right :)
m: my @list = <foo bar baz quux>; my @want = <fo ba>; say @list.grep: /^@want/ 18:16
camelia rakudo-moar e8d0d0: OUTPUT«(foo bar baz)␤»
TestNinja m: my @list = <foo bar baz quux>; my @want = <fo ba>; say @list.grep: *.starts-with: @want.any
camelia rakudo-moar e8d0d0: OUTPUT«(foo bar baz)␤»
TestNinja \o/
kyclark Let me soak that in for a bit. 18:17
dj_goku Is anyone else having issues building 2016.08.1? 18:26
TestNinja dj_goku: what sort of issues? 18:27
kyclark TestNinja, that's some cool stuff. Thanks.
TestNinja Any time. 18:27
dj_goku TestNinja: I can't scroll back give me a few minutes to try and rebuild it.
dj_goku currently trying 2016.07.1 18:27
hmm might be a memory issue: make: *** [CORE.setting.moarvm] Error 137 18:29
Command failed (status 512): make
TestNinja You need around 2GB for RAM+swap
dj_goku hmm 18:30
bioduds yep, around 2gb ram 18:31
from what my tests are showing
dj_goku well that stinks lol 18:32
TestNinja Just resize to get more swap :)
I built on boxes with 1GB RAM and large spawp 18:33
dj_goku this is a production box. :D
TestNinja ooopppss.... 18:34
murdered committable6 with committable6: 2016.04 m: say "meow"; <a b c d b>.rotor: 1 => -2;
Xliff You bastid! 18:35
moritz TestNinja: it's only murder if the dead one is of the same species as you. So you're a bot? :-)
TestNinja I guess it's not comming back :/
Xliff Sorry. South Park ref. I couldn't help it.
dj_goku After a few things finish running I'll cycle a service that is eating some memory
TestNinja moritz: you never know ;)
Xliff "<TestNinja> m: my @list = <foo bar baz quux>; my @want = <fo ba>; say @list.grep: *.starts-with: @want.any" 18:38
Syntax Yum!
bioduds not sure I can resize swap on Digital Ocean VPSs 18:38
mspo do you have shell access? 18:39
bioduds also, main reason for this one-liner p6 installer is to help out for beginners. so, we dont want to tell beginners they should actually go and change swap 18:40
yes, shell access
mst right, 1Gb isn't enough to build rakudo
bioduds out of the box, 2gb is needed
mspo bioduds: just make a swapfile
bioduds in docs, there could be a full explanation on how to change swap in order to make it install on 1gb ram 18:41
mspo dd if=/dev/zero of=/var/db/swap0 bs=1M count=1024 ; mkswap /var/db/swap0; swapon /var/db/swap0
bioduds I'll use this mspo after I test all linux distros i can, thanks :D 18:42
currently on centos 7.2 2b ram 64 bits
it is installing 18:43
all need to have build-essential set before
sudo apt install build-essential on ubutnu
sudo yum groupinstall 'Development Tools' on centos
mst I think ~development also works
bioduds im not sure how to check this in script 18:44
if I manage to have it tested
I can set a line to install it if it is not there I guess
avuserow bioduds: better is to check if the tool exists, maybe something like `command -v make >/dev/null || (echo no make && exit 1)` 18:46
bioduds ok, ill try that
cause if i can do this, the one-liner will work out of the box
since perl5 ships with main linux distros 18:47
i believe it ships with mac as well
bioduds min 1gb swap would be what? 1024? 18:51
bioduds there could be a min reqs telling 2gb ram or 1gb min. 1024Mb swap 18:51
also, 512 RAM and swap 2048 would work? 18:52
[Coke] rakudo isn't on perl6.org because the Rakudo Perl 6 Compiler is a different thing from the Perl 6 Language Specification. 18:54
and this underscores that difference.
bioduds centos installed great 18:56
testing 1GB RAM now 18:57
bioduds avuserow 18:58
bioduds how should i do this test in the shell script? 18:58
[Coke] I doubt there is a portable way for you to do that. 18:59
or at least, an easy way.
bioduds i wont work on that i guess 19:00
maybe after testing
TestNinja m: %{Any} = a => 'B', c => 'b'; 19:03
camelia rakudo-moar e8d0d0: OUTPUT«Odd number of elements found where hash initializer expected:␤Only saw: -> ;; $_? is raw { #`(Block|74651200) ... }␤ in block <unit> at <tmp> line 1␤␤»
TestNinja What's it on about?
bioduds also, installing it after checking that it does not exist would be platform dependent
TestNinja m: my %h{Any}; %h = a => 'B', c => 'b'; dd %h
camelia rakudo-moar e8d0d0: OUTPUT«Hash[Any,Any] %h = (my Any %{Any} = :a("B"), :c("b"))␤»
TestNinja m: my Any %{Any} = :a("B"), :c("b") 19:04
camelia ( no output )
TestNinja mkay
m: %{Any} = a => "B", c => "b"
camelia rakudo-moar e8d0d0: OUTPUT«Odd number of elements found where hash initializer expected:␤Only saw: -> ;; $_? is raw { #`(Block|77295888) ... }␤ in block <unit> at <tmp> line 1␤␤»
TestNinja m: my %{Any} = a => "B", c => "b"
camelia ( no output )
bioduds testing on 1gb ram centos
TestNinja m: %{Any}
camelia rakudo-moar e8d0d0: OUTPUT«Odd number of elements found where hash initializer expected:␤Only saw: -> ;; $_? is raw { #`(Block|76513384) ... }␤ in block <unit> at <tmp> line 1␤␤»
TestNinja Ah, treats it as a coercer 19:05
bioduds i should nevertheless ensure it is runnable 19:07
as avuserow pointed
also, check if min perl5 is there 19:08
do you guys know what is min. perl5 version needed for perl6 to be installed?
MasterDuke Configure.pl has 'use 5.010;' 19:09
timotimo bioduds: did you keep the PREFIX= part of the script? 19:12
TestNinja bioduds: 5.10.1 19:13
(5.10.0 won't work)
bioduds fedora does not come with perl pre-installed 19:15
timotimo yes I did
right now im testing if i can detect make
also will test for perl
so 5.10.1 and above min 19:16
ok
timotimo cool 19:18
leont 5.8.8 may be a more convenient dependency (RHEL5), but OTOH that target group is unlikely to be interested in p6 19:22
mst I would be very surprised if tehre was somebody with 5.8.8 who both wanted p6 and wasn't capable of building a newer perl5 19:23
bioduds there are these legacy users but as you said, it would surprise to see if they wanted to "upgrade" to p6 19:24
good, i could test now 19:26
with
command -v make >/dev/null 2>&1 || { echo >&2 "I require make but it's not installed. Aborting."; exit 1; }
leont True, I doubt it's much of a practical concern 19:27
bioduds i guess i can do the same for perl
good, worked 19:29
now I can prevent installation if no make and no perl is available and warn installers to install it 19:30
67.205.136.118/install.sh
[Coke] If i have a promise with a status of Broken, how can I find out what broke it?
arglebargle. "cause"
TestNinja .result 19:32
Well, the .result would have an Exception if it got broken by exception. Unsure about promises broken manually 19:33
timotimo m: say Promise.^attributes
camelia rakudo-moar 685cc5: OUTPUT«(Mu $!scheduler Mu $!status Mu $!result int $!vow_taken Mu $!lock Mu $!cond Positional @!thens Mu $!dynamic_context)␤»
timotimo m: say Promise.^methods(:local) 19:34
camelia rakudo-moar 685cc5: OUTPUT«(BUILD vow keep break result cause then start in at anyof allof Supply Str Numeric Bool scheduler status)␤»
timotimo ah, yes, cause is probably it
bioduds cool, installing on Fedora 24 x64 2gb ram
TestNinja :o
"If the promise was broken, returns the result (or exception). Otherwise, throws an exception of type X::Promise::CauseOnlyValidOnBroken." 19:35
bioduds timotimo did you say I should remove PREFIX="/usr/local" ?
or leave it there?
timotimo remove
because i don't think it actually has any effect the way it's there 19:36
bioduds ok
timotimo because you said "curl the_script | sh" works, but it can't work if the prefix is really /usr/local
bioduds removed, going to test in a sec
TestNinja m: say (1.1, 3.1).all ~~ any(1.1, 2.1, 3.1)
camelia rakudo-moar 685cc5: OUTPUT«False␤»
TestNinja I excpected that to be True. How come it's false? Is it comparing two Junctions to each other instead of junctioning the values? 19:37
bioduds testing
[Coke] I have a promise $p that is broken. if i then try to print $p.cause.backtrace, I get Backtrace.new 19:38
bioduds fedora is failing 19:38
Can't exec "/home/super/rakudo/rakudo-star-2016.04/install/bin/moar": No such file or directory at tools/lib/NQP/Configure.pm line 450. ===SORRY!=== No suitable MoarVM (moar executable) found using the --prefix (You can get a MoarVM built automatically with --gen-moar.) Command failed (status 512): /usr/bin/perl Configure.pl --prefix=/home/super/rakudo/rakudo-star-2016.04/install --backends=moar --make-install make: *** No targets s
what is this? 19:39
TestNinja bioduds: it can't find moar
m: say .elems == .grep: any(1.1, 2.1, 3.1) given (1.1, 3.1)
camelia rakudo-moar 685cc5: OUTPUT«True␤»
TestNinja This'll do, for me, I guess
timotimo right, the configure.pl is supposed to alreay install a moar when it goes on to build nqp
[Coke] looks like you didn't pass in --gen-moar
bioduds i did 19:40
bioduds perl Configure.pl --backend=moar --gen-moar 19:40
is there in the script
moritz then it looks like compiling or installing moar failed 19:42
bioduds perhaps there is no tar
MasterDuke --backends
bioduds fedora comes all naked it seems
moritz --gen-moar implies backend moar
bioduds no, there is tar 19:43
well, I dont understand then
No suitable MoarVM (moar executable) found using the --prefix
[Coke] did you look at the output of the failed build? or are you hiding it?
bioduds sorry guys, i dont understand what this means
[Coke] if you have it captured, nopaste it somewhere. 19:44
probably that moar didn't build. you have to find the error message on the moar build.
bioduds let me try
[Coke] nopaste the output of your failed attempt to build (put it on gist.github.com or something)
aaaarigh. give my bailador script a POST. fails in a promise, can't get the backtrace of the promise. again, fails, again, now fails different way.... again, now it works. wtf. 19:45
(fresh copy? that still fails) 19:46
bioduds gonna leave fedora for later
heres what i got so far docs.google.com/spreadsheets/d/1dO...sp=sharing 19:47
avuserow bioduds: the snippet I gave you was shell code and seemed to work for me. Not a shell expert though. 19:48
bioduds yes avuserow it was almost correct
i just changed a bit
is working with: docs.google.com/spreadsheets/d/1dO...sp=sharing
opd
ops sorry 19:49
is working with: command -v make >/dev/null 2>&1 || { echo >&2 "make command is required to install Perl6. Please install it and try again."; exit 1; }
also I did the same to test for perl
command -v perl >/dev/null 2>&1 || { echo >&2 "perl is required to install Perl6 (Min. version 5.10.1). Please install it and try again."; exit 1; }
this works
and quits when none is available
[Coke] AIIIGH 19:50
[Coke] (that was for my async bailador stuff, nothing here. ) 19:54
[Coke] docs.perl6.org/routine/run should note if run is synchronous or not. 20:08
leont Yeah, that could be made more obvious, even if synchronous is the default in the rest of the world 20:11
[Coke] m: my $p = start { die 'oh bother' }; sleep 5; say ~$p.cause; 20:14
m: my $p = start { die 'oh bother' }; sleep 5; say $p.cause;
camelia rakudo-moar 685cc5: OUTPUT«oh bother␤»
rakudo-moar 685cc5: OUTPUT«oh bother␤ in block at <tmp> line 1␤␤»
[Coke] there's a surprising Str/gist difference. :| 20:15
geekosaur makes sense to me, although something like $p.cause.payload or etc. might make more sense/be clearer 20:17
[Coke] should at least be documented on docs.perl6.org/type/Exception 20:19
El_Che travis rules: you can test perl6 docker images with it! 20:22
timotimo nice 20:23
El_Che travis-ci.org/nxadm/rakudo-pkg 20:24
El_Che bioduds: I am trying to build native packages. I may have a go at creating fedora packages if I find the time. But you could have a go at the Centos7 pkg: github.com/nxadm/rakudo-pkg (the compile commands are in the docker dir, pkg_rakudo.sh for compiling, the Dockerfile for the prerequisites) 21:08
s/go/look 21:09
grondilu m: say (my %)<foo> // "oops"; 21:15
camelia rakudo-moar 685cc5: OUTPUT«oops␤»
grondilu oh my bad 21:16
grondilu m: my %h; (%h<foo> && die "already defined") = "bar"; say %h<foo> 21:17
camelia rakudo-moar 685cc5: OUTPUT«bar␤»
grondilu nice
grondilu wonders if that's possible in P5 21:23
nope
m: (my $ and die) = "foo" 21:24
camelia ( no output )
harmil_wk I'm really getting into the weeds with this timeout issue. It shouldn't be taking me a week to figure out how to time out a test... 21:39
ugexe m: await Promise.anyof(start { sleep 100 }, start { sleep 3 }); 21:41
camelia ( no output ) 21:42
harmil_wk ugexe: that can't interrupt a long-running call, though, right? 21:42
The sleep 100 just ends up running until you exit, doesn't it?
harmil_wk I have 177 tests to run, and I don't want that many threads generating massive integers until I exit. 21:43
Java has a Thread.interrupt which would be ideal. Then I'd just Promise.anyof($timer, $test) and then $test.interrrupt if the timer hit. 21:44
m: await Promise.anyof(start { loop { sleep 1; say 1} }, start { sleep 1 }); say "Okay, your test is done"; sleep 2; 21:47
camelia rakudo-moar 685cc5: OUTPUT«1␤Okay, your test is done␤1␤1␤»
bioduds guys, what should be the place to ask for help if a user can't get perl6 to be installed? 21:52
I need to put that on the script if it for some unknown reason installation fails 21:53
harmil_wk If I'm reading what you wrote correct, bioduds, you're asking what you should tell people to do if they can't install Perl 6 as a prerequisite for running your script? 21:55
ugexe m: my $s = Supplier.new; my $sup = $s.Supply; await Promise.anyof(start { $sup.tap: { die(); }; for (^100) { say $_; sleep 1 } }, start { sleep 1 }); $s.emit(0); sleep 2; # some sort of kill switch i guess... this blows up the whole thing though
camelia rakudo-moar 685cc5: OUTPUT«0␤1␤Died␤ in block <unit> at <tmp> line 1␤␤»
bioduds yes, if they cant install with the script, where should they go for help 21:56
harmil_wk ugexe: oh, interesting... that's much better than the horrific signal handling I was trying to get working...
bioduds should I tell them to go to chat? or faq? 21:57
perhaps docs.perl6.org/language/faq.html
harmil_wk bioduds: does your script try to install Perl6? I'm just trying to avoid the situation where your script has a problem and we direct people to a general Perl 6 resource...
bioduds yes it tries to install perl6
script is here: 67.205.136.118/install.sh 21:58
it uses main perl6 installation proceedures
if it gets approved, should go run in install.perl6.org 21:59
harmil_wk I would say that the first line of defense should be the script maintainer.
bioduds as top p6 guys here are looking on how would the best way to do this should be 21:59
bioduds I agree. 22:00
So, you think it'd be best if I say, please contact script developer - my email and also look at faq
stmuk bioduds: why doesn't it use the latest Rakudo Star?
bioduds this should be ok?
it is using latest
no?
stmuk rakudo-star-2016.07.tar.gz was the last 22:01
harmil_wk So something like, "if the script fails to install, see <location for script FAQ/bug reporting>, but if you have issues with the installed version of Perl 6 after installation, see <various Perl 6 references and FAQs>"
stmuk it should work if you just change it
bioduds let me change it to the latest then
thanks
dha We didn't have a release last month? 22:02
harmil_wk bioduds: note that you can get the latest release tag here: github.com/rakudo/star/releases 22:03
dha: star releases more slowly now that we're in non-zero-index-Greek-letters 22:04
dha Ah. 22:05
stmuk dha: nope I'm aiming for 3 monthly releases
harmil_wk Though I guess it's not Greek, since "c" isn't Greek... I'll get back to you when I understand the release naming :)
stmuk: is that "3-month"ly releases? 22:06
or 3 per month?
stmuk 1 per quarter (3 months)
harmil_wk that's what I thought, just clarifying 22:06
emdashcomma c is lunate sigma :) 22:07
bioduds yep, changed
going to test with latest now
stmuk bioduds: " 22:13
oops
"You need gcc compiler installed in order to proceed" is only on linux (Mac will need clang installed) 22:14
danlei perl6 repl exits for me on some errors (e.g. when entering 'sub }'). this didn't happen with ubuntu's rakudo package, but with a fresh build (rakudobrew build moar). bug? if not, can it be disabled? 22:15
bioduds clang? 22:16
let me put that on the warnings then 22:17
danlei (2016.08-1-71-g685cc5a/MoarVM 2016.08) 22:19
jnthn danlei: Hm, curious, "sub }" and enter doesn't do it on what I've got locally, which is 2016.08.1-22-g6e76f7e 22:21
danlei: So sounds like a recent regression...please file a bug report.
bioduds i believe my fedora perl installation is the problem
in fedora install failure 22:22
i did sudo yum install perl
perhaps i shouldn't sudo i guess
jnthn (That describe is a tad misleading fwiw, 'cus I'm in a branch...)
bioduds so far ubuntu and centos installations were successful
stmuk bioduds: clang is an alternative compiler to gcc now used on Mac and FreeBSD and confusingly accessible as "gcc" 22:24
I'll try it on a Mac
bioduds great
i cant put up a mac vm, its too complicated
hackintosh
bioduds line now is: curl 67.205.136.118/install.sh | sh;. ~/rakudo/setpath.sh 22:25
later hopefully will become: curl install.perl6.org | sh;. ~/rakudo/setpath.sh
danlei jnthn: ok, thanks. I'll report it.
bioduds great 22:26
if you can, send me results to [email@hidden.address]
bioduds im curious to see if it will work on mac 22:26
El_Che bioduds: you should make the install dir overwritable by a ENV setting (as rakudobrew and perlbrew). Personally I hate stuff in my $HOME (they lib under ~/lib) 22:26
jnthn danlei: Thanks! 22:27
stmuk bioduds: that script should probably exit with a warning if the directory created by tar already exists 22:28
bioduds well, I was rm -rf ~/rakudo but i believe nine told me not to
stmuk sorry I mean the ~/rakudo directory
bioduds it should, I believe, clean up before install yes 22:29
I do rm -rf if it fails though
stmuk yeah I would favour testing for the existance of ~/rakudo, you might delete existing work under ~/rakudo
bioduds sure, cause otherwise it would fail
im putting that in
stmuk someone might be using that with programs existing there and have forgotten
El_Che with rm it's better to err on the safe side 22:30
bioduds correct
I'll think about it
gotta go now, be back soon!
thanks all guys
stmuk bioduds: you have an old 2016.04 reference there 22:31
dalek rl6-most-wanted: cd3e174 | (Tom Browder)++ | most-wanted/modules.md:
update

add a new WIP module: Net::IP delete an unneeded module: Tes::Number::Delta (all function now included in core module Test)		 22:33
AlexDaniel m: my $c = Channel.new; $c.send(42); $c.close; say WHAT $c.list 22:41
camelia rakudo-moar fbeadb: OUTPUT«(List)␤»
AlexDaniel m: my $c = Channel.new; start { $c.send(42); sleep 3; $c.close; }; say $c.list; say now - BEGIN now 22:44
camelia rakudo-moar fbeadb: OUTPUT«(42)␤3.0148542␤»
AlexDaniel m: my $c = Channel.new; start { $c.send(42); sleep 3; $c.close; }; say $c.list.WHAT; say now - BEGIN now
camelia rakudo-moar fbeadb: OUTPUT«(List)␤0.0146833␤»
AlexDaniel I don't understand… How can a List behave like that? 22:44
masak TimToady: questions from charging into the unknown: do you think `is parsed` should be made to work for non-macro subs? do you think `is parsed` has a place on (say) `constant` declarations and/or routine parameters? if possible, please answer modulo any reservations you may have to the viability of `is parsed`. ;) 22:45
masak (to my surprise, `is parsed` is turning out to be solving a lot of interesting problems in 007, at least on paper) 22:48
jnthn masak: Do you handle the parse regex having access to the tokens of the grammar, ooc? 22:49
perlawhirl hi perlers 22:53
perlawhirl ab5tract: Need to update you META6 for Terminal::Print to add OO::Monitors as a dependency 22:54
.seen ab5tract
yoleaux I saw ab5tract 13 Aug 2016 16:17Z in #perl6: <ab5tract> Gar, how did those RHS parens slip in.. Nevermind, my reasoning was based on a grave misunderstanding
TimToady is still deeply suspicious of 'is parsed', but has to run off to play the drums now 22:55
perlawhirl .tell ab5tract You need to update you META6 for Terminal::Print to add OO::Monitors as a dependency 22:56
yoleaux perlawhirl: I'll pass your message to ab5tract.
bioduds stmuk, no, i changed already 22:59
to 2016.07
AlexDaniel oh by the way 23:02
commit: pre-glr,HEAD my @a = <a b c>; .say for 42, @a
committable6 AlexDaniel, ¦«pre-glr»: 42␤a␤b␤c␤¦«HEAD»: 42␤[a b c]
masak jnthn: it's not implemented yet, but (far as I can see) in 007, having access to the tokens of the grammar is higher on the list than having access to lexically defined regexes.
AlexDaniel commit: 2015.07 say ‘Hello from the past!’
committable6 AlexDaniel, ¦«2015.07»: Hello from the past!
masak jnthn: an `is parsed` regex will operate with access to the grammar corresponding to the language currently in effect. 23:03
or, perhaps better to say, the parser currently in effect. 23:04
heh, I asked TimToady to answer modulo reservations, and all I got was the reservations :P 23:05
japhb Isn't that precisely the answer that modulo *can't* give? 23:55

Powered by Cro and the Raku® Programming Language.

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

Thank you!