»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg camelia perl6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by sorear on 25 June 2013.
lizmat hoelzro, will do a spectest 00:00
ah, roast :-) 00:01
hoelzro lizmat: let me know if you find any failures
if you're running with the master roast, you should only see a few under S26-documentation, but that's to be expected 00:02
lizmat I see none, only one flapper 00:06
dalek kudo/nom: ce0c258 | TimToady++ | src/core/ListIter.pm:
Reify lists with correct count.

The ListIter reify called p6arrfindtypes with $max (which is 100000), but should be using $count, how many we really want. Fixes RT #122604.		 00:07
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...?id=122604
TimToady PerlJam: ^^ 00:08
TimToady doesn't remember his RT login—if he ever had one—so someone else can close or test-needed it. 00:17
00:18 BenGoldberg joined 00:20 danaj joined
danaj TimToady: I don't have a bug ref, I should test again and file if still there with latest github. 00:22
TimToady have a small test case I can try?
danaj TimToady: The issue was: take an example like rosettacode.org/wiki/Count_in_factors#Perl_6 and when run from a file, everything was fine. When running perl6 then typing at the prompt, the first line hangs. 00:23
I had this happen on multiple machines, with Ubuntu perl6 or self-built, with all 3 back ends.
TimToady that is probably because the REPL is trying to evaluate it to print out the value.
and because the ... * is embedded in parens, it doesn't realize it's infinite 00:24
danaj Sounds like a feature request rather than bug then.
TimToady well, we really ought to have a version of eager evalution that "gets tired" quickly :) 00:25
lizmat t/spec/S02-types/array.t fails
TimToady so yeah, it's really kind of a feature request
hoelzro lizmat: should I go ahead and merge, then?
TimToady lizmat: so it does, hmm, could be my fault
lizmat it's the only commit between working and not working :-) 00:26
m: my @a = 1..*; @a[Inf] = "dog" 00:27
camelia rakudo-moar ef54ae: OUTPUT«Cannot index Array with Inf␤ in method assign_pos at src/gen/m-CORE.setting:8996␤ in sub postcircumfix:<[ ]> at src/gen/m-CORE.setting:2511␤ in block <unit> at /tmp/njESZ2fALS:1␤␤»
danaj I hate how I get all excited at conferences and have a big list of things I'm eager to work on, then get home and .... blah.
lizmat TimToady: it tickles something in Test.pm, throws_like 00:29
TimToady that's the error I get locally too
dunno why Test gets a different error
BenGoldberg Here's an evalbot feature request: Filter the errors, so that src/gen/m-CORE.setting gets changed to some http:// thing, so I can click on it and go from IRC to whatever :) 00:30
TimToady but of the two errors, I suspect the one I just fixed is more important :) 00:31
BenGoldberg danaj, to renew your excitement, maybe you should drink more caffiene?
TimToady danaj++ for good intentions anyway :)
lizmat TimToady: the test is checking specific values in the exception object
if I take the extra check out, it passes
TimToady wonders if a count of Inf is coming in rather than Whatever 00:32
BenGoldberg danaj, To make that rosettacode example work from the REPL, try adding a 1; to the end of the first line, right after the *);
lizmat seems like the check on aggregate => 1..* is causing the problem 00:33
TimToady maybe need a specific test for Inf in there somewhere 00:35
00:36 aborazmeh left
TimToady I really think we oughta teach .gist to give up after a while on long lists--that's why it's called "gist", after all 00:36
BenGoldberg m: (1..*).gist.say 00:37
camelia rakudo-moar ce0c25: OUTPUT«1..Inf␤»
TimToady m: constant @primes = grep &is-prime, 2, (3, 5, 7 ... *); say @primes
camelia rakudo-moar ce0c25: OUTPUT«(timeout)»
BenGoldberg m: constant @primes = grep &is-prime, 2, (3, 5, 7 ... *); say @primes.gist 00:38
m: constant @primes = grep &is-prime, 2, (3, 5, 7 ... *); say @primes.gist
TimToady say already does .gist
BenGoldberg Oh, nmind.
TimToady but gist isn't really giving you the gist of it...
camelia rakudo-moar ce0c25: OUTPUT«(timeout)»
btyler .tell colomon I added added 'dump directly into P6 data structures' and made that behavior the default for 'from-json'. I'm debating tossing the heap of code for letting jansson objects kinda-sorta behave like P6 ones (that is, aside from what's needed to convert)
yoleaux btyler: I'll pass your message to colomon.
TimToady except for the "this is really long" part of the information :)
lizmat maybe we should limit gists to 140 characters :-) 00:40
dalek ast: ad7a578 | (Elizabeth Mattijsen)++ | S02-types/array.t:
Fix fragile test
 00:46
flussence a .gist(:$length) would make some sense, doesn't have to be exact either (.gist itself isn't exact after all) - just make it elide stuff more aggressively after that many characters pass. 00:47
00:48 xenoterracide joined
lizmat flussence: "=item length 00:48
This word is banned in Perl 6. You must specify units.
flussence okay, you got me there :) 00:49
awwaiid slots
flussence
.oO( $elasticity? )
lizmat :chars probably 00:50
:tweet maybe ?
:-)
00:58 kurahaupo joined
btyler .tell colomon ok, I suppose the debate about whether to keep the emulation code was premature: on -really- big JSON blobs (~180M) keeping the data in jansson-land is radically faster (14 second parse time vs long enough that I gave up, around the 10 minute mark) 01:02
yoleaux btyler: I'll pass your message to colomon.
btyler but this is a great opportunity to take the profiler for a spin :) 01:03
hoelzro has anyone run the spectest on the JVM recently? 01:05
oh, nvm
I thought it was hanging
is there a way for a role to require a consuming class to provide certain methods? 01:13
as a sort of way to document what interface consumers of that role will provide?
01:15 raiph joined
lizmat in the role: method required-method-name { ... } ? 01:17
01:17 dayangkun joined 01:18 FROGGS_ joined 01:19 Sqirrel left
hoelzro that's kind of what I figured 01:20
01:20 Sqirrel joined 01:21 FROGGS left 01:31 aborazmeh joined
lizmat shutting down, decommute in 30 mins& 01:34
01:34 lizmat left 01:35 kurahaupo left 01:37 aborazmeh left 01:51 ardeshir joined 01:57 mberends left 02:07 chenryn joined 02:09 colomon joined 02:20 kurahaupo joined 02:34 jack_rabbit joined 02:36 BenGoldberg left 02:39 noganex_ joined 02:42 noganex left 02:43 grondilu left 02:51 bjz left 03:08 jack_rabbit left
colomon btyler++ 03:14
yoleaux 00:38Z <btyler> colomon: I added added 'dump directly into P6 data structures' and made that behavior the default for 'from-json'. I'm debating tossing the heap of code for letting jansson objects kinda-sorta behave like P6 ones (that is, aside from what's needed to convert)
01:02Z <btyler> colomon: ok, I suppose the debate about whether to keep the emulation code was premature: on -really- big JSON blobs (~180M) keeping the data in jansson-land is radically faster (14 second parse time vs long enough that I gave up, around the 10 minute mark)
colomon btyler: Today I wandered back to working on the JSON::Tiny version of my code. Given your changes today, though, I'll probably take another stab at jansson very soon. 03:16
btyler: (Basically I'm trying to balance changes to my $work testing scheme (which relies on p6 and JSON) with actually making forward progress in the non-testing areas of $work. :) ) 03:17
03:23 kurahaupo left 03:27 kaare_ joined 03:46 ingy left, ingy joined
dalek ecs: 84550b3 | (Rob Hoelz)++ | S26-documentation.pod:
Document how S26 behaves regarding roles
 03:47
ecs: 404f1ea | (Rob Hoelz)++ | S26-documentation.pod:
Clarify leading block declarative comments
03:59 [Sno] left 04:03 hagiri joined 04:05 rindolf joined 04:14 raiph left 04:17 kaare_ left 04:20 xinming left 04:21 xinming joined 04:29 beastd left 04:30 anaeem1 joined 04:32 anaeem1 left, xenoterracide left 04:38 ardeshir left 04:42 mr-foobar left, chenryn left 04:48 kaare_ joined 04:59 chenryn joined 05:05 anaeem1 joined 05:14 ecocode joined 05:16 hagiri left 05:19 kuroseki is now known as Salai 05:22 chenryn left 05:26 chenryn joined 05:32 gfldex joined 05:47 ecocode_ joined 05:48 ecocode left 05:52 [Sno] joined, kaleem joined 05:58 gfldex left 06:01 SamuraiJack joined
TimToady added a more performant version of rosettacode.org/wiki/Iterated_digit...ing#Perl_6 06:02
takes good advantage of the jit, except that div doesn't seem to jit fully yet 06:03
anyway, does the first million in about 5 seconds, and 100 million to 586 seconds (slight slowdown due to heat throttling, probably) 06:05
which is respectable, if not quite in the same ballpark as D yet 06:06
TimToady used the profiler heavily to get it this fast :)
moritz wonders if the profiler can be (ab)used for test coverage reporting 06:08
TimToady this problem would parallelize to multiple cores pretty easily
well, maybe some of the same apparatus would be useful in that regard 06:09
but probably only for the bits written in Perl 6 or nqp; it wouldn't really tell you much about the C code's coverage underneath 06:10
06:10 anaeem1 left
moritz nor would I be very interested in that 06:10
TimToady and test coverage is only going to tell you about code that is written, not code that should've been written but wasn't :) 06:11
06:11 anaeem1 joined
TimToady so it won't get you spec coverage 06:11
at one point we were referencing tests back to the specs, but that has bitrotted 06:12
moritz I was more thinking about evaluating test coverage of modules
TimToady though I guess there are a lot of valid backlinks from the tests into the specs that show up online 06:14
yes, that might be a good sort of tool to have
though it only gets you so far there too, since they might run code but test the wrong thing, like we were arguably doing before dies_ok turned into throws_like 06:17
have to test both the happy paths and the sad paths, and there's a lot of overlap 06:18
dalek c: bf04937 | moritz++ | lib/Language/traps.pod:
[traps] add a SUBTITLE
 06:26
c: d444665 | moritz++ | lib/Type/IO/Path.pod:
fix pod error
06:39 kurahaupo_ left
dalek c: ce7613f | moritz++ | htmlify.p6:
[htmlify] only process .pod files

in particular, do not process .Foo.pod.swp files, which are not valid UTF-8		 06:41
c: e83e69e | moritz++ | lib/Type/IO/Path.pod:
[IO::Path] remove TODO comment

it was intefering with pod parsing; probably because I messed it up.
TimToady you know, even without support for native pointers, it wouldn't be all that hard to get the compiler to notice $native op EXPR and get it to desugar to $native = $native op EXPR for us 06:46
$native op= expression, I meant to say 06:47
where that's presumably declared 'my int $native;' or so 06:48
and once you've done that, it's not so hard to desugar ++$native into the same thing
that's probably a nice NQLHF for someone to tackle 06:49
anyway, that would make a lot of our native code a lot prettier 06:53
TimToady -> zzz however... &
moritz tries to profile htmlify in perl6/doc
runtime without profilinging: 6m16s
ecocode_ since we can call perl5 code from perl6 (seen on yapceu talk) ... does it mean we can parallelize perl5 code ? 06:54
moritz ecocode_: do you mean call out to a perl5 interpreter, or the p5 reimplmentation in rakudo? 06:56
TimToady the lightning talk was about embedding a real P5 06:57
and then embedding a real Python inside that...
moritz and it worked? (ish)
TimToady yes
moritz wow
how did that?
*who 06:58
and how did they do it?
TimToady nativecall to p5 embedding interface
and Inline::Python for the other, iirc
ecocode_ yep, as in the lightning talk, but I would see more interesting stuff if we could parallelize p5 in p6 ;)
moritz ecocode_: then it probably comes with all the usual gotchas you get when you embed p5 in a multi-threaded app 06:59
ecocode_ I'm especially thinking about (easy) parallelizing of website scraping
TimToady you could, potentially, start up separate Perl interpreters, one per P6 thread, which was diakopter's plan
but if P5 can call into P6, then you get all of P6's concurrency, as long as it gets marshalled back to a single P5
moritz TimToady: re =op with natives, one thing that makes this non-trivial is that method infixish does not actually see its LHS arg 07:01
TimToady: so it must be done either in the level above, or in a tree transformation afterwards
TimToady has to be done at the reduction
07:01 darutoko joined
moritz TimToady: so method EXPR? 07:02
TimToady STD had ways to hook things to that reduction in EXPR, dunno if rakudo does
07:03 zakharyas joined
ecocode_ wishes HTML::TreeBuilder and HTML::TableExtract get ported to p6 07:03
TimToady could be done in a later pass as well
TimToady isn't zzz'ing very well here... :)
moritz noticed
TimToady is okay, only have to get up to make coffee for the in-laws and bid them adieu, then I could sleep more tomorrow :) 07:04
moritz profiling htmlify.p6 OOMed after 12m25s 07:06
(presumably during profiling report generation; it completed the processing successfully)
07:06 avuserow left
TimToady ah, it was _reducecheck 07:09
used by ??!! and by .=
it would appear that rakudo also supports a reducecheck 07:10
moritz the special cases //=, infix does etc. are all in method EXPR 07:12
07:13 djanatyn1 is now known as djanatyn
TimToady not sure that's the right place to do such transformations, if we want to support an AST that is closer to the original language, as masak++ wants 07:13
07:15 avuserow joined
TimToady though I suppose he might decide he wants a certain amount of desugaring before his AST sees it 07:15
masak: ^^ talking about where to desugar op= wrt your AST notions
as a higher-order function, op= has to know about both lvalues and rvalues, but we don't have native lvalues (yet) so we're wondering where the right place to do the rewrite is, assuming we want to work around the absence of lvalue natives 07:19
if it's going to be handled later longterm, I'm a bit suspicious of handling it before your AST 07:20
though I suppose it merely degenerates to the ordinary op= handling in that event 07:21
but we could have people depending on native-ish macros that assume the transformation has been done for them 07:22
and then we'd be breaking htat
*th
rindolf TimToady: hi, are you feeling better? 07:26
07:30 kurahaupo joined
nine_ Should I use p5_ or perl5_ as prefix for the helper C functions in Inline::Perl5? 07:33
moritz has no idea, but can point nine_ to doc.perl6.org/type/IO%3A%3APath (in a totally unrelated matter) 07:35
nine_ moritz++ cool! 07:36
nwc10 has the #perl6 hive mind yet identified why rel2abs is such a big number in the profile for startup
moritz nine_: not complete yet, but should be enough to get you started 07:37
nine_ moritz: a great start in any case! 07:39
07:41 _thou left
moritz decides that the native =op rewriting rabbit hole is too deep for him atm 07:42
nine_ Btw. the type graphs could be included directly in the page source instead of converted to PNG. One would only use IE < 9 and Android browser < 4 but gain working links and improved rendering. 07:43
s/would only use/would only lose/
sergot o/ 07:45
moritz nine_: you mean as embedded SVG?
if yes, there are two problems with that: 1) I never got it to work relibably and 2) it only works in XHTML, but doc.perl6.org is html5 07:46
nwc10 all spectests pass 07:49
\o/
ship it!
nine_ moritz: SVG is actually part of HTML5: www.w3.org/TR/html5/embedded-content-0.html#svg 07:52
moritz nine_: wow, things have progress in the last few years :-)
07:53 kurahaupo left
nine_ moritz: at $work we are currently replacing rendered pie and bar charts with inline SVG to gain animatability and styleability. It is really a good time to be in web business right now :) 07:54
moritz now I just have to figure out how to get raw HTML through Pod::To::HTML 07:59
08:00 donaldh joined
dalek kudo/nom: cdf80c7 | moritz++ | src/core/Pod.pm:
Add Pod::Raw type

it is not hooked up to the parser, but can be used to communicate raw HTML to Pod::To::HTML		 08:14
08:24 mathw left 08:53 chenryn left 08:57 fhelmberger joined 08:59 chenryn joined 09:02 kjs_ joined 09:14 aindilis joined
nine_ How do I write a destructor in Perl 6? 09:14
moritz you don't :( 09:15
09:16 _thou joined
nine_ moritz: not planned or not yet implemented? 09:16
moritz nine_: not yet implemented 09:17
nwc10 nine_: but also, destructors happen at $whenever in the future, but I assume that you knew that
("proper" GC, not refcounting)
nine_ nwc10: yep. As long as it would run some time before exit it would be fine. 09:20
09:20 _thou left 09:22 kaleem left
mrf is asuming that after release of 6.0.0 perl6 will be installable via apt? 09:23
nwc10 why wait that long? 09:24
This is perl6 version 2013.12 built on parrot 5.9.0 revision 0
That's what current Ubuntu ships
mrf nwc10: is there a private repo that has perl6 on moarvm?
nwc10 that however is about 8 months out of date
mrf: I'm not aware of one, and I'd love someone (else) to make it.
(That isn't a "well volunteered", unless you'd enjoy doing it) 09:25
mrf can't say I have ever looked into making a dist repo sorry
moritz consider taking the bait
nwc10 yay!
09:25 kaleem joined
mrf having a bleed-repo and a "stable-ish" repo would make things easier for early adopters (vs download tar && make && make install etc) 09:26
nwc10 totally agree. I don't have time to help make this happen. 09:27
09:28 mathw joined
mrf thats cool more a suggestion that a demand 09:28
09:37 darutoko left, darutoko joined
nine_ I just realized that when I read Perl 6 documentation, I read it with jnthns voice in my head. 09:38
mrf is anyone able to give me a very top level reason why running perl6 -e 'say "foo"' is noticably slower than perl -E 'say "foo"' 09:39
I assume this is due to VM startup time
moritz mrf: it is also due to perl 6 loading a full object system
nwc10 mrf: compare with perl -MMoose -E 'say "foo"' and it's closer
it's not so much VM startup time, as VM loading all the built in stuff 09:40
(strictly)
but, as an end user, "VM startup" vs "default stuff" isn't a useful distinction
it's like "I don't care that it's TalkTalk's fault. A&A, please make my ADSL work again"
also, making the time smaller is being worked on.
mrf nwc10: that is a fantastically American metaphore :D 09:41
cool. Despite language speed dick waving being rather pointless, comparisons to other language load times do seem to be important to business people 09:42
I see perl6 likely to be competing in the same market as things like Go 09:43
moritz you mean, concurrent network stuff? 09:46
I tend to see Perl 6 more as a general-purpose language, with some focus on text processing (think grammars), and the tools to make big applications possible 09:47
mrf indeed. Go is being suggested to replace the use of Python as well as C++ for both networking and general purpose scripting 09:49
09:50 spider-mario joined, anon joined 09:51 spider-mario left, woolfy joined 09:55 spider-mario joined 10:02 ggoebel11111119 joined 10:06 ggoebel11111118 left
nwc10 mrf: startup time is a very easy thing to measure, and to think is important. So it's useful to at least "not be the slowest" 10:09
mrf nwc10: I think its probably more important for any language a user expects to run on the command line. 10:12
nwc10 yes. and there's quite a lot of Perl used on command lines 10:13
mrf indeed 10:15
to be fair its not a huge slowness but it is definately noticable in very short -e examples 10:16
10:17 Ven joined
anon java is much slower, if you talk about startup time 10:17
mrf but you almost never run java as a command line tool. So it gets hidden in the "This compiled app is slow to start" 10:18
nine_ Well it's about 90 times slower than perl 5 so it's probably not gonna replace perl5 in tight shell script loops... But other than that it's quite ok.
anon aye 10:19
10:20 telex left 10:22 telex joined
masak oh hai #perl6 10:30
Ven o/ 10:36
10:46 spider-mario left 10:50 SamuraiJack left, SamuraiJack_ joined
timotimo o/ 10:57
10:58 BizarreCake joined 10:59 FROGGS_ is now known as FROGGS
FROGGS o/ 10:59
nwc10 aren't we getting a bit unbalanced here? 11:00
timotimo someone's trying to rock the boat
11:00 spider-mario joined
moritz \\\o 11:01
FROGGS \ohh n\oes!
damn
we cannot escape!
11:04 _thou joined 11:07 Ven left 11:09 _thou left
sergot .tell hoelzro How do you imagine UA as a middleware? :) 11:11
yoleaux sergot: I'll pass your message to hoelzro.
sergot .seen hoelzro
yoleaux I saw hoelzro 01:20Z in #perl6: <hoelzro> that's kind of what I figured
dalek c: f2d8873 | moritz++ | htmlify.p6:
embed typegraph SVG directly into the HTML

because we can. nine++		 11:12
moritz so much yak shave, so little time.
.tell Mouq FYI, perl6/doc htmlify.p6 now needs latest rakudo + latest Pod::To::HTML 11:13
yoleaux moritz: I'll pass your message to Mouq.
mrf does perl6 do compile time type checking if types are used? 11:14
11:14 akaseki joined
moritz mrf: where it can, yes 11:14
mrf: and where it's implemented, that is 11:15
m: say 'alive'; sub f(Str $x) { }; f 42
camelia rakudo-moar cdf80c: OUTPUT«===SORRY!=== Error while compiling /tmp/jfPuVLG298␤Calling 'f' will never work with argument types (int)␤ Expected: :(Str $x)␤at /tmp/jfPuVLG298:1␤------> say 'alive'; sub f(Str $x) { }; ⏏f 42␤»
moritz that's a compile-time error, the "say 'alive'" is never executed
OTOH:
m: say 'alive'; my Str $x = 42; # only caught at run time 11:16
camelia rakudo-moar cdf80c: OUTPUT«alive␤Type check failed in assignment to '$x'; expected 'Str' but got 'Int'␤ in block <unit> at /tmp/ldqoA6bRHC:1␤␤»
moritz that one would be possible to catch at compile time too
11:16 Salai left
mrf i have method foo (Str $bar){} method bar {$.foo(1)} which doesn't seem to have compile time checks 11:17
moritz mrf: method calls are polymorphic, so it's much harder to check statically
mrf: after all, self could be an instance of a subclass, where foo has a more permissible signature 11:18
mrf moritz: How is that handled in a strictly types language?
moritz mrf: they throw an error anyway 11:19
mrf ahh
moritz mrf: because they check the nominal type, not the runtime (actual) type
sergot ==> Testing OpenSSL 11:20
t/01-basic.t .. 1/9 Internal error: invalid thread ID in GC work pass
ohoh, what is this?
mrf moritz: cheers. That makes sense 11:21
11:22 Ven joined
sergot this was like: i.imgur.com/kzRmuyw.gif 11:23
hoelzro sergot: this is kind of what I had in mind: gist.github.com/hoelzro/671912c49b0efcc8d2aa 11:33
yoleaux 11:11Z <sergot> hoelzro: How do you imagine UA as a middleware? :)
11:33 xinming left 11:34 xinming joined
sergot hoelzro++ 11:35
it looks nice!
11:38 Alula_ left
timotimo stealthily published a p6weekly 11:50
11:51 chenryn left 11:52 Alula_ joined
hoelzro sergot: thanks =) 11:52
11:53 Ven left 11:54 anon_ joined, BizarreCake left 11:56 anon left, anon_ is now known as anon
sergot m: say ::; 12:04
camelia rakudo-moar cdf80c: OUTPUT«PseudoStash.new("\$!", Nil, "\$/", Nil, "\$_", Nil, "GLOBALish", GLOBAL, "EXPORT", EXPORT, "\$?PACKAGE", GLOBAL, "::?PACKAGE", GLOBAL, "\$=pod", Array.new(), "!UNIT_MARKER", !UNIT_MARKER, )␤»
12:07 darutoko- joined 12:08 Alina-malina left 12:10 darutoko left
anon timotimo: the link about "Perl 6 in Context" is wrong? 12:11
12:11 Sqirrel left
timotimo it is? 12:12
one sec
indeed 12:13
updated
anon great 12:14
sergot how to check if one has included a module using use/require? I want to perform an action on whether one wrote 'use MODULE;' or not. 12:16
m: ::IO::Socket::INET 12:17
camelia ( no output )
timotimo sergot: you can have an our sub EXPORT
sergot m: use Pod::To::HTML; ::Pod::To::HTML 12:18
camelia rakudo-moar cdf80c: OUTPUT«===SORRY!===␤Could not find Pod::To::HTML in any of: /home/p6eval/rakudo-inst-2/languages/perl6/lib, /home/p6eval/rakudo-inst-2/languages/perl6␤»
timotimo but it's hard to have a "export was not called" thing going on
as the mainline will be executed before the EXPORT sub ...
sergot timotimo: I want to make IO::Socket::SSL as an optional dependency in HTTP::UserAgent, so it should print "hey, write 'use IO::Socket::SSL' to turn on SSL connections" or something. 12:19
timotimo ah, so only if the user writes "use IO::Socket::SSL; use HTTP::UserAgent" should it even try to load IO::Socket::SSL? 12:20
sergot or, make HTTP::UserAgent include IO::Socket::SSL if it is installed.
timotimo: yes
timotimo "if it is installed" is much easier
sergot And even better in this case I think :)
timotimo m: my $res = try require IO::Socket::SSL <IO::Socket::SSL::TheClass>; say IO::Socket::SSL; say $res;
camelia rakudo-moar cdf80c: OUTPUT«Could not find symbol '&SSL'␤ in method <anon> at src/gen/m-CORE.setting:13173␤ in any find_method_fallback at src/gen/m-Metamodel.nqp:2625␤ in any find_method at src/gen/m-Metamodel.nqp:944␤ in block <unit> at /tmp/W_AYIH8ZhG:1␤␤» 12:21
timotimo m: my $res = try require IO::Socket::SSL <IO::Socket::SSL::TheClass>; say $res;
camelia rakudo-moar cdf80c: OUTPUT«(Any)␤»
12:21 spider-mario left
sergot TheClass? what is this? 12:22
m: use IO::Socket::INET; 12:23
camelia rakudo-moar cdf80c: OUTPUT«===SORRY!===␤Could not find IO::Socket::INET in any of: /home/p6eval/rakudo-inst-2/languages/perl6/lib, /home/p6eval/rakudo-inst-2/languages/perl6␤»
sergot timotimo: it doesn't seem to work. 12:25
12:25 Alina-malina joined
timotimo well, "require" will run-time-load a module (that's why you have to give it a list of things that it'll import; rakudo absolutely has to know the lexical symbols you have available at *compile* time.) 12:26
if it fails, it'll raise a run-time exception, which you can catch
m: say 1 cmp 6 12:28
camelia rakudo-moar cdf80c: OUTPUT«Less␤»
timotimo it's harder to react to the user's code having use'd a "IO::Socket::SSL" before it use'd your module 12:29
not sure how to do it without going into rakudo internals or looking up the source code (if it exists) and grepping for use IO::Socket::SSL 12:30
but you can ask the user to use HTTP::UserAgent <SSL>; or something
sergot Does camelia have access to some p6 modules? 12:31
12:31 kaare_ left
sergot to try it live 12:31
timotimo if you use "star:" instead of "m:" 12:32
but the latest rakudo star is kinda old by now 12:33
star: say $*PERL
camelia star-m 2014.04: OUTPUT«{"compiler" => {"release-number" => "", "build-date" => "2014-05-06T17:09:26Z", "ver" => "2014.04", "name" => "rakudo", "codename" => ""}, "name" => "rakudo"}␤»
..star-p 2014.04: OUTPUT«{"compiler" => {"codename" => "", "ver" => "2014.04", "build-date" => "2014-05-06T17:16:05Z", "name" => "rakudo", "release-number" => ""}, "name" => "rakudo"}␤»
sergot oh, right, it is
anyway
timotimo++
thanks :)
timotimo YW :)
12:34 Sterver joined
Sterver Hi... Where can I find file on github with perl6 example code for learning? Some time ago anybody sent me it, but I lost that link :-P :D 12:36
timotimo it could be either github.com/perl6/perl6-examples (just example code) or learnxinyminutes.com/docs/perl6/ (much more explanation)
you can find more example code on rosettacode.org - each of the pages in that wiki will have the same task solved in a whole bunch of different languages, many of them include Perl 6 examples, too 12:37
Sterver hm, it was like learnxinyminutes.com/docs/perl6/ but it was file on github, and it was more recent
timotimo github.com/Nami-Doc/learnxinyminut...l.markdown - could be this 12:38
sergot: you have passed GSoC with flying colours, right? 12:39
sergot timotimo: I always get (Any) in $res. :(
timotimo: yes :) 12:40
timotimo could be that require doesn't have a return value
in that case you can inspect $! i believe?
you just have to make sure you don't put { } around the require statement
otherwise the symbols will only be visible in there; i made that mistake before :)
Sterver timotimo, yeah, it is :) thanks
sergot timotimo: $! contains error, thanks 12:42
I thoght I can use :: somehow. 12:43
thought
timotimo you mean to access the "caller's" pseudostash? 12:44
sergot yes
timotimo that's what i was unsure about :\ 12:45
12:45 xfix left, xfix joined, xfix left, xfix joined
nine_ Inline::Perl5 now supports calling Perl 5 functions with multiple Int/Str parameters and multiple Int/Str return values :) We also now have a proper test suite. 12:47
sergot timotimo: I can use $! 12:48
timotimo: you asked about GSoC
nine_ tadzik: you may want to join soon ;) 12:50
12:51 Sterver left 12:52 chenryn joined
PerlJam greetings 12:52
12:52 _thou joined
rurban nine_: Impressive lightning talk about inline::perl5! 12:56
12:56 chenryn left
rurban blitzkost #3 12:56
12:57 _thou left 13:00 anaeem1 left 13:01 xenoterracide joined
jnthn oh my, how much backlog... 13:06
moritz jnthn: could you please test if the makefile-line-length rakudo branch builds on windows?
jnthn: if yes, please merge into nom; it fixes a mac os X build issue ("write error")
nwc10 much backlog, and still no-one seems to have actually used the profiler for great evil, er profit, yet 13:07
moritz nwc10: I disagree
13:07 brrt joined
nwc10 oops sorry. I failed to spot something? 13:07
moritz nwc10: TimToady++ mentioned an optimized version of an RC entry, made with the help of the profiler
nwc10 oh. oops. 13:08
moritz irclog.perlgeek.de/perl6/2014-08-26#i_9243290
13:08 anon_ joined
jnthn moritz: Sure, lemme see how git pull works out... 13:08
anon_ uses profiler to test jnthn++'s gtk exmaple and found it doesn't works well here and 90%+ cost is at 111 lines of NativeCall.pm6
brrt moritz: that is so totally cool and awesome 13:09
TimToady++
jnthn anon_: What, specifically, doesn't work well? :)
colomon nwc10: I dunno about good or evil, but I used the profiler to get a substantial in one of my $work scripts.
13:10 mr-foobar joined, anon left, anon_ is now known as anon, raiph joined
anon jnthn: only "Characters: " text works 13:10
nwc10 colomon: cool
colomon nwc10: it's hard to say how substantial, alas, as the profiler seems to return significantly different overall times on different runs. 13:11
but it seemingly involved hacking about 30% off the overall running time. 13:12
jnthn moritz: Tryig it.
13:12 spider-mario joined
jnthn *Trying 13:12
13:12 spider-mario left
anon also add --profile-compile to CORE setting compiling and it is killed (maybe due to memory) 13:13
nwc10 how much memory do you have?
jnthn Well, I have a 16GB box and that wasn't enough o.O 13:14
nwc10 was assuming to wait a bit to see what gets fixed as a result of smaller things, before trying that
anon 4G
jnthn I need to do something about that...
13:14 spider-mario joined
jnthn The profiler as it stands is "what I had time to do in spare bits of time in the week or so before YAPC::EU". So don't expect it to be perfect. :P 13:14
nwc10 I think the largest thing that I have access to has 96G of RAM 13:15
apart from the main work DB server, which has more, but that is the main work DB server...
jnthn moritz: It seems good to me.
moritz: Manages to build and make test just fine.
13:16 chenryn joined
jnthn nwc10: Yes, maybe don't use that. 13:16
anon there is 128G of RAM server here, but it's a online db server 13:17
nwc10 exactly. I'm waiting for the stuff related to startup of -e0 to be done, and other more basic compiling
arnsholt My research group's compute server has 256 gig, but I'm not sure if my boss would appreciate me seeing how much I can hog at once =)
13:17 chenryn left
brrt 96G :-o 13:17
how even
13:18 xenoterracide left 13:20 guru joined 13:21 guru is now known as Guest65827, Guest65827 is now known as ajr_
dalek kudo/nom: dea7adf | moritz++ | / (5 files):
Read bootstrap sources from file

because make on Mac OS X has a line length limit in commands
  (symptom: "write error")		 13:22
kudo/nom: c693d75 | moritz++ | / (3 files):
Do not pass M_CORE_SOURCES on the command line

it is too long for Mac OS X make
kudo/nom: 3e55a0d | moritz++ | src/core/IO.pm:
[IO::Path] be more consistent with IO::Path:D markers for the invocant
kudo/nom: ba9d01d | moritz++ | / (6 files):
Merge branch 'makefile-line-length' into nom
moritz that should fix the "write error" bug on mac os x for MoarVM builds
s/fix/work around/
colomon moritz++
jnthn oh hurrah! 13:23
moritz++ indeed
moritz tgt++ golfed it down
maybe I should do the same for the JVM build
jnthn Wow
moritz (the problem being that in certain constellations, make doesn't like command lines > 1024 chars on mac os X) 13:24
13:24 jnap joined
masak moritz` 13:24
er, moritz++
13:25 anon is now known as JimmyZ
nine_ Yes, this now works: perl6 -Ilib -e 'use Inline::Perl5; my $i = p5_init_perl(); $i.run("use DBI; 1;"); my @dbh = $i.call("DBI::connect", "DBI", "dbi:Pg:database=timemngt"); my $dbh = @dbh[0]; say $i.call("DBI::db::selectrow_array", $dbh, "select count(*) from products"); $i.DESTROY; 13:25
nwc10 curious. worked on my machine.
clearly different OS X is different 13:26
nine_ I can now use Perl 5's DBI in Perl 6 :)
JimmyZ nine_++ # that's fast!
moritz woah
nine_ Everything available on github.com/niner/Inline-Perl5 13:27
And the weird my $dbh = @dbh[0]; is just because I really suck at Perl 6... 13:29
13:29 bjz joined
colomon nine_: would be good to add that to the p6 module ecosystem.... 13:32
nine_ colomon: yes, would be awesome if someone who actually knew how to write a proper Perl 6 module could have a look at it. These are literally my first steps in Perl 6. 13:33
colomon nine_: if someone else hasn't done it by this evening (US time) I'll take a stab at it. Right now I should be $working. :) 13:35
nine_ colomon: great! :)
colomon is dismayed that searching for info on creating p6 modules turns up a bunch of obsolete information... 13:36
13:36 JimmyZ_ joined 13:37 JimmyZ__ joined 13:39 JimmyZ left, JimmyZ__ is now known as JimmyZ 13:40 agen4 joined
agen4 /msg camelia 13:40
13:41 JimmyZ_ left
hoelzro hmm...S26 doesn't seem to work on jakudo or pakudo 13:42
13:42 kaare_ joined, agen4 left
betterworld m: "" ~ Buf.new 13:43
camelia rakudo-moar cdf80c: OUTPUT«(signal )»
betterworld this hangs rakudo. In an older version, it printed error messages in an endless loop
masak cute graphs: github.com/Dobiasd/programming-lan.../README.md 13:46
FROGGS betterworld: there is already an RT ticket...
betterworld oh, I did not check that, sorry 13:47
13:48 JimmyZ left 13:51 potatogim_ joined 13:52 potatogim_ left, SamuraiJack_ left, SamuraiJack joined 13:54 potatogim_ joined 13:55 potatogim_ is now known as PotatoGim^Home 13:56 PotatoGim^Home is now known as PotatoGim_Mobile 13:57 PotatoGim_Mobile is now known as PotatoGim-Tab
nine_ How can I avoid an "Illegally post-declared type" error? Class A has a method that calls B.new() and B has an attribute of type A. 13:57
13:58 grondilu joined
jnthn Pre-declare one of them 13:58
class B { ... }
13:58 PotatoGim-Tab is now known as PotatoGim_Tab 13:59 PotatoGim_Tab is now known as PotatoGim_Home
nine_ jnthn: ah with "...". Thanks! 13:59
14:00 _thou joined 14:01 PotatoGim_Home is now known as PotatoGim^Home 14:02 mberends joined
mrf in perl5 I often use the construct sub init {$self = $self->new unless ref $self} to initialize objects in class methods. Is there a similar construct in perl6? 14:05
jnthn I guess you could write it as method init() { self // self.new } 14:12
If I remember what the p5 code does correctly, anyway. ;) 14:13
nine_ Now simplified to: use Inline::Perl5; my $i = p5_init_perl(); $i.run("use DBI; 1;"); my $dbh = $i.call("connect", "DBI", "dbi:Pg:database=timemngt"); say $dbh.call("selectrow_array", "select count(*) from products"); $i.DESTROY; 14:14
Now I only need to find out how to overload call and it will be $dbh.selectrow_array("select count(*) from products); :) 14:15
mrf jnthn: the idea being to swap a class str for an object ref 14:16
jnthn String?!
nine_ s/overload call/do something like AUTOLOAD/ 14:17
jnthn thinks he'll leave this one to somebody who knows Perl 5 :)
We almost never think of types by their string name in Pelr 6.
*Perl
nine_ mrf: I'd guess that in Perl 6 classes are full blown objects instead of just names.
geekosaur perl5's oo is an out and out hack. yes, strings. 14:18
there's a reason they adopted an old prototype of p6's mop as a real OO interface :) 14:19
jnthn chuckles
my SomeClassName $x .= new; # is a common Perl 6 idiom
Not sure if it applies in this case
Please could somebody on non-Windows do a perl6-m --profile-compile -e "say 1" and upload the result for me somewhere? 14:22
nwc10 optimised build?
jnthn Preferably. 14:23
nwc10 JIT or not?
jnthn Since that's what I've got on Windows.
JIT.
I've just compared rel2abs on Windows and posix and the Windows one is probably rather more costly.
nwc10 I hadn't said (here) that I tried it on OS X yesterday (or maybe Sunday) with the debugging build from Sunday or Saturday, and it was 16% 14:24
but I have no idea how representative that is of optimised builds
mrf nine_: I seem to get issues if I call an accessor method on an unititialised object in the sub init. e.g sub init {$.foo} which is why I ask 14:26
PerlJam mrf: care to show your whole code? 14:27
nine_ How would a CANDO method declaration really look like? If I do it like shown in S10, I only get an "Invalid typename 'Item'"
nwc10 OK, 6 minutes later, spectest pass 14:28
mrf PerlJam: Any specific pastebin?
PerlJam mrf: any is fine with me.
nwc10 pypy - can you beat that? :-)
rubinius - how low can you go? 14:29
PerlJam nwc10: none of those have spectests though ;)
nwc10 good point
nine_ Is CANDO even implemented? acking rakudo's source yields nothing at all 14:31
PerlJam nine_: I doubt it.
nwc10 jnthn: ccl4.org/~nick/profile-1409063326.04419.html
nine_ PerlJam: damn...so no way to make progress on that front for now.
nwc10 "my" machine is x86_64 CentOS, and has loadsaRAM 14:32
rel2abs at 13.46%
about to go to www.meetup.com/Vienna-Perl-Mongers/...201277862/ so won't be around to re-run stuff this evening
jnthn nwc10: Thanks! 14:33
nwc10 no problem
jnthn waits for it to download :)
14:33 PotatoGim^Laptop joined
nwc10 did you want it compressed? 14:34
jnthn It's fine; this wifi ain't the fastest
Though I just noticed there seems to be a network cable on the desk here too 14:35
14:35 PotatoGim^Home left
dalek kudo/nom: cfdf260 | jonathan++ | src/core/IO.pm:
Don't try to rel2abs open '-'.

Seems this isn't the source of the costly rel2abs calls, however, so only a minor win from this.		 14:35
kudo/nom: 8af523b | jonathan++ | src/core/Supply.pm:
Don't derive grammar in CORE.setting.

Fixes an issue reported by donaldh++ where we were declaring a new op inside of the setting. This caused a new grammar to be derived, which in turn ended up including fresh NFAs and an entire copy of the source to end up the file too! Fixing this shaves ~6MB off CORE.setting size, about 14MB off base memory, and decreases startup time too.
14:36 PotatoGim^Laptop is now known as PotatoGim^Home
mrf PerlJam: pastebin.com/XEsxLfVG 14:37
japhb Whee, may other wins be as easy as that one!
14:37 PotatoGim^Home left
mrf PerlJam: That code snippet should show what I am attempting to do 14:37
14:37 PotatoGim^Home joined 14:39 PotatoGim^Home left, PotatoGim^Home joined
masak nine_: try 'git grep' instead of 'ack'. it's faster, and it plays really well with git repositories. 14:40
colomon ponders writing a grammar to parse the output of git log --stat 14:43
nwc10 jnthn: new profile at profile-1409064143.23213.html
er
ccl4.org/~nick/profile-1409064143.23213.html
but you might want ccl4.org/~nick/profile-1409064143.23213.html.gz if your bandwidth is bad
203K rather than 4.1M 14:44
colomon nwc10: when I go to ccl4.org/~nick/profile-1409064143.23213.html… oh! It just takes a while to load!
nwc10 Oh yes. 14:45
sadly that friend's toybox is no longer in the office of an ISP
colomon didn't post a profile that way the other day because he thought it wasn't working.
nwc10 using all their unsold fibre
which was fun
Casey was downloading a Perl 5 RC
the bottleneck was in the wifi in starbucks
14:45 anaeem1_ joined
nwc10 on the other side of the pond. 14:45
14:46 rindolf left
TimToady too much pond scum, maybe 14:46
nwc10 no, the transatlantic link was fine
it was faster than the starbucks wifi
and I don't think that the starbucks wifi was particularly slow.
jnthn nwc10: urgh, yes, it's bad enough the gz is taking a while o.O 14:47
TimToady we still multiple terahertz of dark fiber running within a block or two of our house and across the Pacific
*accidentally 14:48
colomon jnthn: suggestion: could we include the full command line used in the profiler output?
colomon finds it too easy to get confused about what he was actually profiling. 14:49
nine_ Is there any way at all to have a class fake the existence of methods at runtime?
TimToady too
[Coke] (RT #122604) added "testneeded" 14:50
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...?id=122604
Timbus mrf, if you want to check if self is defined in your method
dalek kudo-star-daily: b21121a | coke++ | log/ (14 files):
today (automated commit)
Timbus .. do exactly that
if self { } 14:51
14:52 molaf joined
TimToady there's a specced FALLBACK method, not yet implemented 14:54
mrf Timbus: if I call Foo::Bar.init whould self be defined?
TimToady it would presumably call the existing .^add_fallback metamethod to install itself
14:54 akaseki left
TimToady nine_: ^^ 14:54
nine_ TimToady: ah too bad. So I guess I won't finish Perl 5 Perl 6 interop in a week after all. 14:55
Timbus mrf: no
mrf Timbus: I am more attempting to accertain if self is a class or object ref which was a fairly common check in perl5
14:56 xinming left
Timbus when undefined, i believe self is a type object 14:56
14:56 spider-mario left
jnthn Time for some sleep... & 14:57
mrf which would explain the error I get when calling $.args when self is undefined 14:58
jnthn nwc10: Got the .gz; will study it tomorrow. :)
mrf Timbus: can I do self .= new ? to allow $.args later in that sub. 14:59
Timbus .. woa that seems like a bad idea 15:00
15:00 spider-mario joined
mrf maybe trying to force bad perl6 paradigms on perl6 15:00
15:00 anaeem1_ left
Timbus if you just want a class variable that isn't tied to an instance of the class, use 'my' instead of 'has' 15:01
mrf Timbus: I want a method where the invocant could be either a class name or an object ref. init and run are good examples of this. These usually handle the duality by normalising to having an instantiated object 15:03
nine_ mrf: how about using a multi method for that? 15:04
Timbus ^ beat me to it
mrf Multi method?
nine_ mrf: class Foo { multi method init (Str) { self.new().init() } multi method init (Foo) { ... } } or something like that
Timbus well. id probably use the :D check on the invocant, probably 15:05
nine_ is still an absolute Perl 6 beginner
mrf nine_: you seem to be doing better than me :D
Timbus m: class Wot { multi method init(Wot:U: $args) { say 'type class' }; multi method init(Wot:D: $args) { say 'actual instance' } }; Wot.init(1); Wot.new.init(1); 15:08
camelia rakudo-moar ba9d01: OUTPUT«type class␤actual instance␤»
15:09 akaseki joined 15:10 xinming joined
psch m: class Foo { has @.args is rw; method init(Foo $self is rw: @args) { $self //= Foo.new; $self.args = @args } }; my $f = Foo; $f.init([1,2,3]).say; $g = Foo.new; $g.init([4,5,6]).say 15:10
camelia rakudo-moar ba9d01: OUTPUT«===SORRY!=== Error while compiling /tmp/YFIE4kpOAC␤Variable '$g' is not declared␤at /tmp/YFIE4kpOAC:1␤------> }; my $f = Foo; $f.init([1,2,3]).say; $g⏏ = Foo.new; $g.init([4,5,6]).say␤ expecting any of:␤ …»
psch my my 15:11
m: class Foo { has @.args is rw; method init(Foo $self is rw: @args) { $self //= Foo.new; $self.args = @args } }; my $f = Foo; $f.init([1,2,3]).say; my $g = Foo.new; $g.init([4,5,6]).say
camelia rakudo-moar ba9d01: OUTPUT«1 2 3␤4 5 6␤»
psch that's slightly horrible though, in my opinion
Timbus i think i did that in some code somewhere and i hate it
15:13 MilkmanDan left, MilkmanDan joined 15:14 chenryn joined 15:15 akaseki is now known as Salai, mrf left
btyler colomon: ok, no worries if JSON::Tiny's a better fit. I'd love to get your changes/added tests as a pullreq if you ever have a moment, though 15:17
colomon oh, absolutely! … when I get a second.
btyler ultimately healthier for the ecosystem if JSON::Tiny is sufficient for most needs, I think
colomon dunno if JSON::Tiny's a better fit, it's just something I've got working NOW. :) 15:18
15:18 treehug88 joined
colomon btyler: doing a crazy balancing act here, if you know what I mean. 15:19
btyler colomon: totally hear you on the value of 'it works right now' :) 15:20
15:20 ajr_ left
colomon needs to improve his code, his tests, his testing infrastructure, and the speed of all three things, all at the same time. 15:21
15:21 mrf joined 15:22 guru joined 15:23 guru is now known as Guest86251, Guest86251 is now known as ajr_
colomon ponders the difference between v and n 15:23
TimToady well, something is usually better than nothing, if only as a bad example :) 15:24
15:26 mrf left, chenryn left, Alina-malina left 15:30 chenryn joined
colomon panda install Grammar::Debugger 15:31
===SORRY!===
Could not find Shell::Command
:\
15:31 mrf joined 15:32 brrt left
TimToady colomon: I suspect the only difference is that \n will match CRLF 15:33
colomon ah, as one thing?
15:35 chenryn left, Alina-malina joined 15:36 chenryn joined 15:37 JimmyZ joined 15:38 zakharyas left 15:41 chenryn left
TimToady m: say so "\xd\xa" ~~ /^ \n $/ 15:41
camelia rakudo-moar ba9d01: OUTPUT«True␤»
TimToady m: say so "\xd\xa" ~~ /^ \v $/
camelia rakudo-moar ba9d01: OUTPUT«False␤»
15:43 chenryn joined 15:53 firefish5000 joined
JimmyZ jnthn++ # CORE.setting.moarvm size is 6M smaller 15:53
15:54 xinming_ joined 15:55 xinming left, raiph left 15:58 MilkmanDan left, MilkmanDan joined
firefish5000 Good day. How do you ignore-case for a token?I have tried <token tkn :i {...}> and <token :i tkn {...}>. 15:59
TimToady try putting the :i inside the brackets 16:00
firefish5000 Ahh. much better. Thanks TimToaday!
PerlJam firefish5000: What kind of grammar are you writing? 16:01
16:03 chenryn left
firefish5000 PerlJam: Im just playing around right now. This one parses time strings like '3d1h4s' 16:04
PerlJam neat.
play on! :)
16:05 chenryn joined 16:07 dayangkun left 16:08 MilkmanDan left 16:09 kaleem left, MilkmanDan joined 16:13 jack_rabbit joined, JimmyZ left 16:16 [Sno] left
colomon just made his grammar explode 16:17
16:21 Rotwang joined 16:22 donaldh left
nine_ What's the closest thing to Perl 5's undef there is in Perl 6? 16:28
moritz nine_: Any
nine_: are you marshalling things between p5 and p6? 16:29
nine_ moritz: yes
moritz nine_: in the p5 -> p6 direction, I'd translate undef to Any
nine_: and in the other direction, any value that's not .defined to undef 16:30
firefish5000 Can we add to the result of rules? such as <token { (\S+) \s+ (\S+) {return "$0 And $1"} }> without exiting the grammar? (bad practice?) 16:32
nine_ moritz: thanks! Now this works: say $dbh.call("selectrow_hashref", "select * from products").perl; 16:35
moritz m: '23 + 42' ~~ /(\d+) \D+ (\d+) { make $1 + $2 }/; say $/.ast # for firefish5000
camelia rakudo-moar 8af523: OUTPUT«use of uninitialized value of type Any in numeric context in block <unit> at /tmp/loqK19mpvH:1␤␤42␤»
moritz m: '23 + 42' ~~ /(\d+) \D+ (\d+) { make $0 + $1 }/; say $/.ast # for firefish5000 16:36
camelia rakudo-moar 8af523: OUTPUT«65␤»
moritz firefish5000: the idiomatic thing is to use make(), which stores the result in $/.ast
m: '23 + 42' ~~ /(\d+) \D+ (\d+) { make $0 + $1 }/; say $/.made
camelia rakudo-moar 8af523: OUTPUT«65␤»
16:36 Util left
moritz in .made; I think .ast is old/deprecated 16:36
firefish5000 moritz: Thanks, that did it! 16:39
TimToady in some respects undef is more like Nil
but Nil is still misbehaving in list contexts
(I think)
m: my @a = 1, Nil, 3; say @a.elems 16:40
camelia rakudo-moar 8af523: OUTPUT«2␤»
TimToady yeah
we need to figure out where we're currently depending on the () semantics 16:41
and use () instead
16:44 Util joined
nine_ Now workign: say $dbh.call("selectall_arrayref", "select * from products where id < ?", Any, 10).perl; 16:44
In other words: only passing arrays and hashes from Perl 6 to Perl 5 missing for full DBI access :)
xfix r: say [+] grep * %% 2, (1, 1, *+* ...^ * > 4_000_000);
nine_ Have to leave now though...more on this tomorrow
camelia rakudo-jvm 8af523: OUTPUT«(timeout)» 16:45
..rakudo-{parrot,moar} 8af523: OUTPUT«4613732␤»
16:47 jnap left 16:48 PerlJam left, tadzik left
colomon jnthn: if the Grammar trace ends with * MATCH, doesn't that mean the entire rule matched? 16:48
16:48 PerlJam joined 16:51 jnap joined
colomon oh, does parse actually fail if the entire string is not parsed now? 16:51
TimToady yes, use subparse if you want to stop in the middle 16:52
xfix r: say [+] grep *%%2,(1,1,*+*...^*>4e6) 16:53
camelia rakudo-{parrot,jvm,moar} 8af523: OUTPUT«4613732␤»
colomon TimToady: I didn't want to stop in the middle, just the optional last section was failing the match. :)
16:53 ajr_ left
colomon failing to match, I should say. 16:53
16:54 guru joined, guru is now known as ajr_ 16:56 akaseki joined, Util left
colomon whomever-fixed-parse++ 16:56
16:57 tadzik joined 16:58 Salai left 17:02 tadzik left 17:03 tadzik joined 17:04 Util joined, telex left 17:06 telex joined
colomon jnthn++ # Grammar::Tracer still rocks 17:07
17:08 jack_rabbit left 17:09 akaseki left, Util left, Util joined 17:11 akaseki joined 17:12 ecocode_ left 17:14 tadzik left 17:15 PerlJam left, tadzik joined 17:17 PerlJam joined, fhelmberger left 17:18 chenryn left 17:21 kaleem joined, ajr_ left 17:22 guru joined, guru is now known as Guest88584, Guest88584 is now known as ajr_ 17:23 tgt joined
timotimo o/ 17:24
moritz \o 17:28
colomon |o| 17:33
TimToady
.oO(touchdown?)
colomon
.oO(TIE fighter)
 17:34
17:34 firefish5000 left
moritz in Perl 6, that would be PROXY fighter, no? :-) 17:35
colomon what's the p6 equivalent of "read files from the command line or standard in, or whatever"? $*IN appears to just be stdin
tadzik is it just me or feather is having problems?
moritz colomon: lines()
colomon: the handle is called $*ARGFILES or so
colomon moritz: what if you want to read everything into a single string?
17:36 ajr_ left
moritz m: say $*ARGFILES.slurp.chars 17:36
camelia rakudo-moar 8af523: OUTPUT«1134␤»
colomon moritz++ # $*ARGFILES is exactly it
moritz m: say slurp().chars
camelia rakudo-moar 8af523: OUTPUT«1134␤»
moritz colomon: slurp() should have the same default
colomon oh!
colomon feels like an idiot now
timotimo .o( DWIM: Dude! What Incredible Magic! ) 17:37
FROGGS colomon: ask me... I was chasing a problem for two or three days just because I am stupid
colomon FROGGS: we should form a club
17:37 jnap left
FROGGS and I mean >10 hours a day 17:38
yeah :/
colomon BTW, being able to add { say $<commit-name> } in the middle of the grammar is pretty darned awesome 17:39
FROGGS my default is to add things like: <?{ say "$?FILE:$?LINE &?ROUTINE.name()" }> 17:41
like in v5's grammar
because grammar tracer can't help me there
17:45 BizarreCake joined
colomon FROGGS: I'm using it as a progress track, how far it's gotten in the file, and where it stops. 17:45
FROGGS I've currently a problem that v5 hangs somewhere in the grammar after parsing a statement 17:46
17:47 rindolf joined
colomon BTW, my grammar parses a 2.8M git log --stat file with 5734 commits in 11.3 seconds on my MacBook Pro. (Before any profiling, mind you.) 17:51
seems like that's much better performance than I've seen in the past...
17:51 Ven joined
FROGGS 11.3s sounds not too bad, aye 17:51
17:53 [Sno] joined
colomon I thought it was going to be too slow to parse the entire log like that, but 12 seconds is fine for my purposes. 17:54
17:58 fling left 18:00 kaleem left
Ven timotimo: the learnxinyminutes file is not on my github account, it's on the original one :) 18:03
(wrt irclog.perlgeek.de/perl6/2014-08-26#i_9244515 )
dalek ecs: 96719fa | (Stéphane Payrard)++ | S99-glossary.pod:
fix formatting errors
18:04 fling joined
dalek rl6-roast-data: 1fef0fa | coke++ | / (6 files):
today (automated commit)
 18:07
18:08 molaf_ joined
timotimo Ven: sorry :| 18:09
sterver isn't around any more, though :(
Ven yes, which is why I answered to you :) 18:10
Well, I already manually merged learnperl6iny PR manually, I could do it again anyway
18:11 molaf left
Ven disagrees a bit with litchkind++'s slides on the brain part :) 18:12
(emotions are in front, graphics is behind) 18:13
TimToady well, emotions are kind of in the middle :)
[Coke] disk access on feather seems really slow.
Ven TimToady: the frontal lobe is responsible for reasoning and emotions 18:14
TimToady mostly for controlling emotions :)
Ven well, it's the most developed part
TimToady [Coke]: maybe feather is having negative emotions deep down inside :) 18:17
[Coke] htop wasn't reporting anything big on cpu or memory, but launching any new command is verrry slow. 18:18
r: say "what?"
camelia rakudo-{parrot,jvm,moar} 8af523: OUTPUT«what?␤»
[Coke] huh: load average: 18.42, 13.12, 9.23
Ven that's a big word. 18:19
[Coke] (that's with 5% of cpu)
TimToady how many processes in D state?
18:19 kjs_ left, ecocode_ joined
[Coke] 6 18:20
TimToady could well be disk then 18:21
tadzik what's up with feather :(
[Coke] kjournald, flush-254:0, syslogd, sshguard, ./run ... now 5. root & syslog
weechat-curses from tjs was in there for a bit 18:22
TimToady disk might be overloaded or failing
18:22 ardeshir joined
[Coke] urk. now like 20, including a ton of apache2 processes. 18:22
TimToady does it have a log of failures/resets somewhere
[Coke] .seen juerd?
yoleaux I haven't seen juerd? around.
[Coke] .seen juerd
yoleaux I saw Juerd 16 Aug 2014 14:35Z in #perl6: <Juerd> Wrong way around perhaps, but with an output mechanism that doesn't do partial lines, it may make sense
tadzik me weechat has been causing trouble before, I hope it's not at it again
TimToady might just be robots banging on it
or it might just have been promoted to its level of incompetence 18:24
[Coke] moritz++ os x fixed. 18:26
18:26 Ven left, Ven joined
FROGGS [Coke]: what is fixed? 18:27
[Coke] pinged Randal to let him know.
the make write error.
FROGGS ohh!
moritz++ # \o/
colomon just did $match<date><year month day>.join("-"), which he thinks is quite elegant. 18:30
tadzik < 18:31
<3
18:32 Akagi201 left 18:33 Akagi201 joined
[Coke] install rakudo-moar. install panda. install rakudo-moar. sadface. 18:34
should I have to reinstall everything after updating rakudo? 18:35
timotimo yes, sadly 18:36
all the precompiled modules will have references and dependencies on specific compunits and serialization IDs
making a "migration strategy" for that is probably a very Hard Problem. 18:37
PerlJam rakudobrew++
18:37 Akagi201 left
colomon [Coke]: panda rebootstrap 18:38
FROGGS [Coke]: just chdir panda && perl6 rebootstrap.pl && coffee
(or coke)
colomon but yeah, rakudobrew++
18:41 SamuraiJack left
colomon in which case a full update is normally just rakudobrew build moar 18:42
and panda rebootstrap happens automatically.
timotimo cool
colomon tadzik++
tadzik :) 18:43
18:44 Ven left 18:45 kjs_ joined
colomon m: say <a b c>.first-index('b') + 1 18:47
camelia rakudo-moar 8af523: OUTPUT«2␤»
colomon m: say <a b c>.first-index(* eq 'b') + 1 18:48
camelia rakudo-moar 8af523: OUTPUT«2␤»
TimToady the nqp version of rosettacode.org/wiki/Iterated_digit...ing#Perl_6 runs in 40% of the time of the best jittable pure Perl6 version
colomon huh, that's getting me No such method 'item' for invocant of type 'NQPMu'
locally
18:48 Ven joined
colomon oh 18:49
FROGGS colomon: can you pass --ll-exception to it?
colomon FROGGS: actual error was in a different line
FROGGS I'd like to know where in Perl6::Actions this happens
colomon m: sprintf ("%02d", 2) 18:50
camelia rakudo-moar 8af523: OUTPUT«No such method 'item' for invocant of type 'NQPMu'␤ in method message at src/gen/m-CORE.setting:12859␤␤»
colomon that's the code
FROGGS ohh
colomon (extra space is the problem)
FROGGS that explodes while constructing the exception me thinks
Ven "@list = 1, 2, 3 same as $list = (1, 2, 3)" wut?? 18:51
@fib and $fib, actually. from lichtkind++'s slides
18:51 gfldex joined
Ven "for forces list context" er, that seems kinda wrong 18:52
FROGGS colomon: that might fix it: 18:53
-"Type check failed in $.operation; expected '{$.expected.^name}' but got '{$.got.^name}'";
+"Type check failed in $!operation; expected '{$!expected.HOW.name($!expected)}' but got '{$!got.HOW.name($!got)}'";
[Coke] $ perl6 rebootstrap.pl
use of uninitialized value of type Any in string context in block <unit> at rebootstrap.pl:27
FROGGS in my class X::TypeCheck is Exceptio
colomon [Coke]: hmmm… either you've got an out-of-date panda, panda is broken, or the module being rebootstrapped is broken. 18:54
[Coke]: it does work in general, honest
[Coke] it worked, but I got 3 or 4 warnings. 18:55
if I want to test a local change to panda, how do I invoke it? just ./bin/panda in the sourcecode dir? 18:59
tadzik perl6 -Ilib may be helpful
FROGGS or rebootstrap it 19:01
this will also use and install itself
[Coke] my $list = from-json slurp $!projectsfile;
^^ $!projectsfile is undefined when I run "panda"
complains when I have http_proxy set. 19:02
I have an non empty projects.json file installed. when I try to run panda again, the from-json slurp dies on it. 19:12
m: say from-json " "
camelia rakudo-moar 8af523: OUTPUT«No such method 'ast' for invocant of type 'Any'␤ in sub from-json at src/gen/m-CORE.setting:21321␤ in block <unit> at /tmp/wqeccb94jU:1␤␤»
[Coke] ^^
19:13 PerlJam left
colomon hmmm, anyone have an elegant way to ensure that the last character of a string is a period? 19:13
19:14 PerlJam joined
[Coke] also, the local panda should probably not be pointing there. 19:14
(to the install dir)
19:16 Sqirrel joined 19:18 Util left, kjs_ left
TimToady colomon: $string ~= '.'; # :P 19:18
flussence
.oO( subset ProperlyPunctuatedStr of Str where /\.$/ ... )
 19:19
19:20 tadzik left
TimToady hmm, p5 does an optimization where it knows it can start scanning that one from the end of the string 19:21
I'll bet we don't do that one yet
19:22 darutoko- left
TimToady colomon: by "ensure", do you mean test, or do you want add one if there isn't one? 19:22
m: my $_ = 'foo'; s/ <-[.]> <( /./; .say 19:23
camelia rakudo-moar 8af523: OUTPUT«Potential difficulties:␤ Redeclaration of symbol $_␤ at /tmp/Uf9HpOWlul:1␤ ------> my $_ ⏏= 'foo'; s/ <-[.]> <( /./; .say␤f.oo␤»
TimToady m: $_ = 'foo'; s/ <-[.]> <( $ /./; .say
camelia rakudo-moar 8af523: OUTPUT«foo.␤»
TimToady like that?
m: $_ = 'foo'; s/ <-[.]> $ <( /./; .say 19:24
camelia rakudo-moar 8af523: OUTPUT«foo.␤»
TimToady slightly faster, probably
m: $_ = 'foo'; s:p(.chars-1)/ <-[.]> $ <( /./; .say
camelia rakudo-moar 8af523: OUTPUT«foo.␤»
TimToady faster yet
Ven m: my @a = ^10; say @a[1].WHAT; say @a[1,].WHAT
camelia rakudo-moar 8af523: OUTPUT«(Int)␤(Parcel)␤»
colomon TimToady: I want to add one if there isn't one. 19:25
Ven I must admit that, *sometimes*, sigil variance has advantages
I'd get bit by it so many times tho...
colomon TimToady++
TimToady m: say (1).WHAT; say (1,).WHAT; say (1;).WHAT 19:27
camelia rakudo-moar 8af523: OUTPUT«(Int)␤(Parcel)␤(Int)␤»
TimToady m: say (1).WHAT; say (1,).WHAT; say (1;2).WHAT
camelia rakudo-moar 8af523: OUTPUT«(Int)␤(Parcel)␤(LoL)␤»
TimToady looks like a buglet there, I'd think (1;) should also LoL
19:28 donaldh joined
TimToady gee, who wrote that code? <whistles> 19:28
Ven :P 19:31
19:32 tadzik joined 19:35 daxim_ left 19:36 Util joined
TimToady problem is it's just testing # of statements, and seeing 1 19:38
19:39 guru joined, guru is now known as Guest78247, Guest78247 is now known as ajr_
TimToady can think of several ways to fix it, so probably better nap on it 19:39
19:42 Util left 19:43 tadzik left, donaldh left, spider-mario left 19:44 Util joined, tadzik joined 19:45 spider-mario joined, spider-mario left 19:46 jack_rabbit joined 19:48 spider-mario joined
dalek kudo/nom: d8048c9 | coke++ | docs/ChangeLog:
keep ChangeLog up to date
 19:55
kudo/nom: 399faac | coke++ | docs/ChangeLog:
remove trailing whitespace
19:57 kjs_ joined 20:04 Akagi201 joined 20:09 Akagi201 left 20:26 Rotwang left
Ven .seen raiph 20:35
yoleaux I saw raiph 15 Aug 2014 18:33Z in #perl6: <raiph> cjqwesc
20:36 itz_ joined 20:38 BizarreCake left
tgt r: (*..100)[131072] 20:39
camelia ( no output )
20:39 itz left
tgt Huh. That breaks here with "P6opaque: no such attribute '$!storage'" 20:39
20:41 kjs_ left
tgt r: (*..100)[1250000] 20:42
camelia ( no output )
moritz tgt: fixed in latest rakudo
tgt How latest? I cloned a few minutes ago. 20:43
moritz tgt: in commit ce0c258640222761ab8176aff92cf9bbe5b3665f by TimToady++
tgt Ah, sorry, nvm. Was using the wrong perl6. 20:44
20:44 itz joined, itz_ left
tgt Did whoever it was that looked into fixing smartmatch with ranges sometime in July get anywhere? (Might've been FROGGS?) 20:45
FROGGS tgt: I cannot remember fixing something along these lines
tgt I'll see if I can find it in the archive. It was about smartmatching against a stringy range. 20:46
FROGGS ahh
m: say "b" ~ "a" .. "z" 20:47
camelia rakudo-moar 8af523: OUTPUT«"ba".."z"␤»
FROGGS m: say "b" ~~ "a" .. "z"
camelia rakudo-moar 8af523: OUTPUT«True␤»
FROGGS m: say "a" ~~ "b" .. "z"
camelia rakudo-moar 8af523: OUTPUT«False␤»
FROGGS okay, not trivial to guess :o)
tgt m: say "33" ~~ 1..5
camelia rakudo-moar 8af523: OUTPUT«True␤»
FROGGS ahh, that one
no, I was unable to fix it
tgt m: say "bb" ~~ "a".."aaa" 20:48
camelia rakudo-moar 8af523: OUTPUT«False␤»
tgt Ah, okay. 20:49
With that last one, I disagree with the spec.
It says "X.min le $_ le X.max (mod ^'s)", which isn't at all what I'd usually expect. 20:50
moritz most other defnitions rely on some way to generate all possible strings between the end points 20:51
which is kinda tricky in Unicode land
PerlJam moritz: you are a master of understatement :)
tgt Yeah, I was assuming that was the reason. Just seems like it'd not be what you'd want most of the time.
moritz the assumption that 'a'..'aaa' only generates ASCII characters is very much culturally laden 20:52
loden
20:52 kurahaupo joined
moritz whatever 20:52
tgt True. 20:54
Do Perl 6's ranges work outside of ASCII?
timotimo "ranges"? 20:55
.u one
yoleaux U+0011 DEVICE CONTROL ONE [Cc] (␑)
U+001F INFORMATION SEPARATOR ONE [Cc] (␟)
U+0031 DIGIT ONE [Nd] (1)
moritz m: say ('Ⅰ'..'Ⅴ')).list
camelia rakudo-moar 8af523: OUTPUT«===SORRY!=== Error while compiling /tmp/RDusUd6t8U␤Unexpected closing bracket␤at /tmp/RDusUd6t8U:1␤------> say ('Ⅰ'..'Ⅴ')⏏).list␤»
moritz m: say ('Ⅰ'..'Ⅴ').list
camelia rakudo-moar 8af523: OUTPUT«Ⅰ Ⅱ Ⅲ Ⅳ Ⅴ␤»
moritz looks like a "yes" to me :-)
tgt :)
PerlJam I would think they'd *have* to work "outside of ASCII" for character classes to be useful. 20:56
(or it would certainly be LTA for ranges to work differently depending on whether or not you were in a character class)
20:56 kaare_ left
tgt I'll just need to adjust my expectations when smartmatching against stringy ranges. 20:57
m: my $r = "a"..*; say $r.min le "b" le $r.max; 20:59
camelia rakudo-moar 8af523: OUTPUT«False␤»
20:59 kjs_ joined
tgt m: my $r = "a"..*; $r.max.WHAT.say; # Ah 21:01
camelia rakudo-moar 8af523: OUTPUT«(Num)␤»
tgt That's not quite right.
kurahaupo tgt: should it even *have* a max?
FROGGS m: my $r = "a"..*; say $r.max # is meant to be Inf 21:05
camelia rakudo-moar 8af523: OUTPUT«Inf␤»
kurahaupo Inf kinda makes sense, but it's not very stringy
tgt Shouldn't it be a stringy equivalent of Inf?
21:05 Akagi201 joined
FROGGS tgt how? 21:05
gnight #perl6 21:06
kurahaupo a string consisting of an Inf number of \uffffffff chats?
tgt I don't know. My first example above seems surprising. (my $r = "a"..*; say $r.min le "b" le $r.max;)
kurahaupo chars*
moritz m: say 'b' ~~ 'a'..*
camelia rakudo-moar 8af523: OUTPUT«True␤»
tgt hmm. 21:07
kurahaupo Perhap Inf should be in its own class separate from Num?
[Coke] kurahaupo: that way lies madness. 21:08
because then you have come up with a class that is basically a singleton that is also every other numeric (and string) type.
m: say Nil.WHAT 21:09
camelia rakudo-moar 8af523: OUTPUT«Nil␤»
[Coke] Seems like we want something more Nilistic.
kurahaupo true
tgt moritz: I think that's because Range::ACCEPTS uses cmp?
21:10 Akagi201 left
moritz m: say 'z' ~~ 'a'..* 21:10
camelia rakudo-moar 8af523: OUTPUT«True␤»
21:10 Ven left
moritz m: say 'Z' ~~ 'a'..* 21:10
camelia rakudo-moar 8af523: OUTPUT«False␤»
moritz m: say 'B' ~~ 'a'..*
camelia rakudo-moar 8af523: OUTPUT«False␤»
kurahaupo what about a singleton StrInf? 21:11
moritz there's another option
one could store 'a'..* as 'a'..* 21:12
that is, the actual whatever in the right endpoint
and then special-case that in Range.ACCEPTS
and in the generator
tgt I didn't realise * could be stored in things, I thought it was just a thing that had magic done to it when it was compiled. 21:13
PerlJam [Coke]: I dunno ... we have a bottom type (Nil), maybe we need a "top" type too that is a chameleon and becomes whatever we need it to be :)
moritz m: my $x = *; say $x.^name 21:14
camelia rakudo-moar 8af523: OUTPUT«Whatever␤»
21:19 kjs_ left 21:21 Ven joined 21:28 donaldh joined
nine_ Ok, passing arrays and hashes from Perl 6 to Perl 5 is also implemented now. This means for example being able to use full DBI and probably even a Perl 5 DBIx::Class schemafrom Perl 6. 21:32
moritz nine_++
21:33 woolfy left 21:34 woolfy1 joined
Ven nine_++ :) 21:35
donaldh exciting # nine_++
moritz I think I'll have trouble writing Rust code; they use the presence or absence of a final semicolon to indicate if stuff is an expression or a statement 21:37
let x = if 1 == 1 { 42 } else { 23 }
works as expected
but if you write 42; it assumes an empty expression (?) after the 42
21:38 treehug88 left
moritz so x is () after that 21:38
that's quite a difference to perl's "final ; is optional" 21:39
Ven always puts the final ; due to some habit ... :P
21:39 ecocode_ left 21:40 ecocode_ joined
moritz Ven: I tend to do that too 21:40
Ven m: class A { sub foo { self.new(); } }; A.new.foo
camelia rakudo-moar 399faa: OUTPUT«===SORRY!=== Error while compiling /tmp/cc1HKNFX3M␤'self' used where no object is available␤at /tmp/cc1HKNFX3M:1␤------> class A { sub foo { ⏏self.new(); } }; A.new.foo␤ expecting any of:␤ statemen…»
Ven m: class A { sub foo { $self.new(); } }; A.new.foo # err?
camelia rakudo-moar 399faa: OUTPUT«===SORRY!=== Error while compiling /tmp/8HHfOaJNuM␤Variable '$self' is not declared␤at /tmp/8HHfOaJNuM:1␤------> class A { sub foo { $self.new()⏏; } }; A.new.foo # err?␤ expecting any of:␤ method a…»
moritz Ven: because if you add a stament after the current one, it's easy to forgetto check for the absence of a semicolon on the previous line
Ven same problem as the "pending comma" one :) 21:41
oh, sub.
m: class A { method foo { self.new( } }; A.new.foo
camelia rakudo-moar 399faa: OUTPUT«===SORRY!=== Error while compiling /tmp/yojXDoL87Y␤Unable to parse expression in argument list; couldn't find final ')' ␤at /tmp/yojXDoL87Y:1␤------> class A { method foo { self.new( ⏏} }; A.new.foo␤ expect…»
21:43 treehug88 joined 21:44 ecocode_ left, Ven left
tgt Thanks for your help earlier moritz. (Having a go at storing * at the endpoints as you suggested.) 21:48
21:52 ardeshir left
[Coke] Maybe we'll have a good answer to InfStr when we have RatStr settled. 21:54
tgt m: use Test; my $a = *; my $b = *; is $a, $b; 21:58
camelia rakudo-moar 399faa: OUTPUT«not ok 1 - ␤␤# Failed test at /tmp/BGzq9vsvzc line 1␤# got: 'Whatever<140583114232904>'␤# expected: 'Whatever<140583114232992>'␤»
tgt Why is that the case?
Ah, is means $a eq $b. 22:00
22:04 kjs_ joined 22:06 Akagi201 joined 22:10 Akagi201 left 22:13 tgt left, kurahaupo left 22:16 treehug88 left, woolfy1 left 22:18 woolfy joined 22:22 woosley left 22:38 ilogger2_ joined 22:39 Shozan joined 22:41 ggherdov joined 22:42 mtj left, mtj_- left, oetiker left 22:43 flussence joined, isacloud____ joined 22:44 thilp_ joined, mtj joined, raydiak joined, oetiker joined 22:45 sivoais left, sivoais joined 22:48 MilkmanDan joined 22:49 simcop2387 joined, ilbot3 joined 22:50 PotatoGim^Home joined 22:55 ggherdov left, ggherdov joined, isacloud____ left, isacloud____ joined 22:58 xragnar_ joined, xragnar_ is now known as xragnar 23:00 colomon joined 23:02 bowtie_ joined, telex joined 23:03 aborazmeh joined 23:12 kurahaupo joined 23:15 colomon left, kurahaupo left 23:16 colomon joined 23:17 aborazmeh left 23:23 araujo joined 23:25 vike joined 23:34 PotatoGim^Home left 23:35 ardeshir joined 23:40 ardeshir left 23:54 PZt joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!