»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or /msg camelia p6: ... | irclog: irc.perl6.org or colabti.org/irclogger/irclogger_log/perl6 | UTF-8 is our friend! 🦋
Set by Zoffix on 25 July 2018.
00:00 Manifest0 left 00:01 Sgeo_ joined 00:04 Sgeo left 00:05 Manifest0 joined 00:09 Manifest0 left 00:13 Altai-man_ left 00:14 Manifest0 joined 00:23 Manifest0 left 00:28 Sgeo__ joined 00:29 Manifest0 joined 00:30 kurahaupo_ joined 00:31 kurahaupo left 00:32 Sgeo_ left 00:37 Manifest0 left 00:42 Manifest0 joined 00:44 ribasushi joined 00:46 Sgeo__ left 00:47 Sgeo__ joined 00:55 espadrine_ left 00:58 TreyHarris left
cpan-p6 New module released to CPAN! Podviewer (0.0.1) by 03UZLUISF 00:59
00:59 Manifest0 left 01:01 kurahaupo_ left 01:02 kurahaupo joined
Kaiepi m: module A { role B { }; class B::C does B { } }; say A::B::C 01:03
camelia (C)
Kaiepi hm 01:04
01:04 Sgeo__ left, Sgeo__ joined, netrino left
Kaiepi m: module A {}; module A { role B { }; class B::C does B { } }; say A::B::C 01:05
camelia 5===SORRY!5=== Error while compiling <tmp>
Redeclaration of symbol 'A'
at <tmp>:1
------> 3module A {}; module A7⏏5 { role B { }; class B::C does B { } };
expecting any of:
generic role
01:05 Manifest0 joined
Kaiepi i have a module that's similar to what i'm trying to do here but it's not working hastebin.com/ivuvizebap.rb 01:05
it claims the symbol Net::Telnet::Subnegotiation::NAWS doesn't exist
why? 01:07
Xliff "Subneotiation::NAWS does Subnegotiation" ... Um ... "Subneotiation" is missing the "g"?
But even so, that's namespace issue. That should still work.
Kaiepi oh whoops lmao 01:08
Xliff :)
Kaiepi ah there we go it works now 01:09
thanks
Xliff LOL! That was it?!
Kaiepi yes
Xliff \o/
Kaiepi i feel like i should rename the serialize method to Blob so i can pass subnegotiation objects to methods that take Blob() 01:10
Xliff Method::Also allows you to do both. 01:11
use Method::Also
method serialize is also<Blob> { ... }
Options are good.
Also. TIMTWODI
Kaiepi oh huh didn't know about also 01:13
btw what does the handles trait mod do?
Xliff Yeah. I'm using it for p6-GtkPlus
handles?
Ah! Delegation. The docs do not do that explanation justice. 01:14
m: class A { method b { say "From B"; }; }; class B { method b { A.new.b }; }; 01:15
camelia ( no output )
Xliff ^^ A pattern like that is MADE for delegation.
Which can be written as:
m: class A { method b { say "From B"; }; }; class B { has A $.a handles<b>; }; B.new.b 01:16
camelia From B
Xliff m: class A { method b { say "From A"; }; }; class B { has A $.a handles<b>; }; B.new.b
camelia From A
Xliff For that arbitrary example, it's hard to see the use, but imagine a 4-D Point class being wrapped by a Box class.
s/Box/MetaPoint/ 01:17
"class MetaPoint { has Point $p handles<x y z Ⓐ>; }". Prevents you from having to write silly wrapper methods for .x, .y, .z and .Ⓐ 01:19
01:19 andrzejku joined
Xliff Push that out to a wrapper class for a large-scale application that handles 4 different kinds of objects for 4 disparate systems. 01:20
01:20 andrzejku_ left
Xliff And then push that further. 01:20
So... long story short: the handles trait can be viewed syntactic sugar for methods that stub in methods from other class attributes. 01:21
But it's more than that.
Oh.. syntax error in my class explanation... 01:22
"class MetaPoint { has Point $!p handles<x y z Ⓐ>; }". Prevents you from having to write silly wrapper methods for .x, .y, .z and .Ⓐ
01:22 Manifest0 left, Matthew[m] left, MitarashiDango[m left, tyil[m]1 left, unclechu left, Seance[m] left, Demos[m] left, AlexDaniel` left, mack[m] left
Xliff THAT allows you to still expose $!p.x, $!p.y, $!p.z, $!p.Ⓐ, without exposing $!p 01:22
01:22 tyil[m] left
Xliff Or writing wrappers. 01:23
01:23 lance_w[m] left, rba[m] left, Garland_g[m] left 01:25 vs^ left, irdr left 01:26 vs^ joined, vs^ left, vs^ joined, irdr joined 01:27 vs^ left, Manifest0 joined, vs^ joined, vs^ left, vs^ joined 01:28 vs^ left, vs^ joined, vs^ left 01:40 Manifest0 left 01:42 pnu__ left, pnu__ joined, b2gills left, afresh1 left 01:43 peteretep left 01:45 peteretep joined, Manifest0 joined 01:47 Sharparam_ joined, gks_ joined 01:48 Kaiepi left, lichtkind left, MilkmanDan left, Geth left, vxe left, synopsebot left, xinming left, huf left, agentzh left, gks left, Sharparam left, KotH left, Grrrr left, stux|RC left, nightfrog left, bobby left, tyil left, hahainternet left, krunen left, silug_ left, Ekho left, kubrat left, jdv79 left, revdiablo left, Woodi left, Sgeo_ joined, afresh1 joined 01:50 zacts joined 01:51 Sgeo__ left 01:56 Woodi joined 01:57 molaf left 02:01 b2gills joined, Manifest0 left 02:05 mack[m] joined 02:09 unclechu joined, Manifest0 joined 02:10 molaf joined, lance_w[m] joined 02:11 Matthew[m] joined 02:12 Seance[m] joined, rba[m] joined 02:13 tyil[m]1 joined 02:14 AlexDaniel` joined, hythm joined 02:15 MitarashiDango[m joined 02:22 Ekho joined 02:23 Manifest0 left 02:24 Sgeo__ joined 02:25 Demos[m] joined, Garland_g[m] joined 02:26 tyil[m] joined 02:27 Sgeo_ left 02:28 hythm left 02:29 zacts left 02:31 Manifest0 joined 02:34 zacts joined 02:44 zacts left 02:45 Kaiepi joined, lichtkind joined, MilkmanDan joined, Geth joined, vxe joined, synopsebot joined, xinming joined, huf joined, agentzh joined, KotH joined, Grrrr joined, stux|RC joined, nightfrog joined, bobby joined, tyil joined, hahainternet joined, krunen joined, silug_ joined, kubrat joined, jdv79 joined, revdiablo joined 03:11 Manifest0 left 03:16 Manifest0 joined, ChoHag joined 03:20 mowcat joined 03:23 Manifest0 left 03:24 zacts joined 03:28 Manifest0 joined
SmokeMachine pmurias: it seems not hard to use Red for it... 03:29
nine: I’ve done something similar on Red... 03:32
03:32 ChoHag left 03:35 Manifest0 left
SmokeMachine .tell nine github.com/FCO/Red/tree/master/lib/Red/AST 03:36
yoleaux SmokeMachine: I'll pass your message to nine.
03:40 Manifest0 joined 03:45 Manifest0 left, Sgeo__ left 03:46 Sgeo__ joined 03:49 mowcat left 03:50 Manifest0 joined 03:57 Manifest0 left 03:58 MasterDuke left 04:03 Manifest0 joined, zacts left 04:14 Manifest0 left
Kaiepi should geth post when new issues are opened on github? i think it might be useful 04:18
04:19 Manifest0 joined 04:20 adu left 04:31 skids left
Kaiepi alias prove6="PERL6LIB=lib prove -e perl6 -vr $@" 04:37
ik there's a prove6 package out there but this is useful for people who don't like installing too many packages 04:38
04:47 Manifest0 left 04:51 Manifest0 joined 04:56 Manifest0 left 05:01 Manifest0 joined 05:05 Manifest0 left 05:11 Manifest0 joined 05:17 Manifest0 left 05:19 kurahaupo left 05:22 Manifest0 joined 05:31 Manifest0 left 05:36 Manifest0 joined
ugexe I’d be surprised if most people *werent* using some variant of that 05:38
of course the pure way is prove -v -r -e ‘perl6 -I .’ ... 05:39
05:41 Manifest0 left 05:46 ChoHag joined, Manifest0 joined 05:51 Manifest0 left 05:57 Manifest0 joined 06:02 Manifest0 left, cpan-p6 left 06:03 cpan-p6 joined, cpan-p6 left, cpan-p6 joined 06:07 Manifest0 joined 06:12 Manifest0 left 06:14 jmerelo joined
jmerelo releasable6: status 06:15
releasable6 jmerelo, Next release will happen when it's ready. 1 blocker. 140 out of 317 commits logged (⚠ 2 warnings)
jmerelo, Details: gist.github.com/53983eedc0359f3e5e...f438cb8a44
06:18 Manifest0 joined 06:22 zacts joined 06:27 Manifest0 left 06:33 andrzejku_ joined, andrzejku left, andrzejku_ is now known as andrzejku 06:34 Manifest0 joined 06:38 Manifest0 left 06:44 Manifest0 joined 06:48 zacts left
Kaiepi all my telnet library needs is documentation and examples and it's done! 06:58
well
"done"
done enough to release on cpan
jmerelo Kaiepi: all perl6 needs is documentation and it's done!
Kaiepi: anyway, good luck! 06:59
06:59 MilkmanDan left
Kaiepi also needs the features i'm adding when i start my grant work next week jmerelo :) 07:00
07:01 jaldhar joined, MilkmanDan joined
jmerelo Kaiepi: congrats on your grant work. 07:01
Maybe you can help with a problem I have now.
How do we get the name of the class we're in within a class?
m: class Foo { say .^mro }; 07:02
camelia ((Any) (Mu))
jmerelo m: class Foo { say .^WHICH };
camelia Cannot resolve caller WHICH(Perl6::Metamodel::ClassHOW:D: Any:U); none of these signatures match:
(Mu:U: *%_ --> ValueObjAt:D)
(Mu:D: *%_ --> ObjAt:D)
in block <unit> at <tmp> line 1
jmerelo m: class Foo { say .^WHo };
camelia No such method 'WHo' for invocant of type 'Perl6::Metamodel::ClassHOW'. Did you mean 'WHY'?
in block <unit> at <tmp> line 1
07:02 zacts joined
Kaiepi m: class Foo { say .^name } 07:03
camelia Any
Kaiepi hm
m: class Foo { say ::?CLASS.^name }
camelia Foo
Kaiepi there we go 07:04
jmerelo Kaiepi: thanks!
xinming [499280.758321] perl6[1374]: segfault at 78 ip 00007f88dfb21b10 sp 00007f88c6ffc990 error 4 in libmoar.so[7f88df88e000+642000] │·
[499280.761621] Code: 41 5d 41 5e 41 5f c3 0f 1f 00 48 8b 16 48 63 46 14 48 89 54 24 20 83 f8 fe 74 3f 48 8b 92 80 00 00 00 48 8d 04 40 48 c1 e0 04 <48> 03 42 78 4c 8b 68 08 0f b7 40 14 89 4│·
4 24 1c e9 66 fe ff ff 0f
Kaiepi np
jmerelo I was about to post it in StackOverflow, so I'll do it anyway
xinming Is there a way how I can troubleshoot this kind of bug?
jmerelo Kaiepi: stackoverflow.com/questions/560969...in-a-class
xinming I got segfault on moar
but no clue how I can repeat it, it's in a large program 07:05
timotimo there are some ways. if the program doesn't run terribly long, "perl6-valgrind-m whatever-your-commandline-was" can give a little hint
by default moarvm isn't compiled with debug symbols, though
if you can, re-run moarvm's Configure.pl with the same flags you can see near the beginning of the "Makefile", and add --debug=3 07:06
xinming Ok
timotimo then just "make install -j4" or so, no need to rebuild rakudo or nqp 07:07
xinming Will try, it's not a long run, But takes my cro app running 3 minutes, and then fail 07:09
timotimo oof 07:10
cro is multi-threaded and valgrind only supports a single thread, so it'll be a bit slower still 07:11
you can also try "perl6-gdb-m"
as soon as it crashes, you can "call MVM_dump_backtrace(tc)" to at least see where in the code it happens
xinming timotimo: So, I'll use perl6-gdb-m, right? 07:12
bbl
timotimo probably
valgrind can get a lot more details out of the crash usually
xinming timotimo: You'll need to give me hints on how to use gdb to get the context 07:13
timotimo no prob 07:14
xinming I'm installing gdb
07:14 holyghost joined
timotimo oh, if you're on a mac, you might already have lldb, in that case you can use perl6-lldb-m i think 07:15
xinming And thanks, I'm really glad to get into deeper part of the perl6.
I'm in gentoo
I use gentoo
timotimo OK
07:24 Manifest0 left
holyghost . 07:26
hello 07:27
.tell Xliff you did ,tell instead of .tell yesterday, I sent you some email for the new permanent ip
yoleaux holyghost: I'll pass your message to Xliff.
07:29 Manifest0 joined 07:32 Kaypie joined, Kaiepi left
holyghost .tell Xliff kibology :-) 07:32
yoleaux holyghost: I'll pass your message to Xliff.
07:43 Sgeo joined 07:45 zacts left, espadrine_ joined 07:46 Sgeo__ left 07:56 Manifest0 left 08:00 Manifest0 joined 08:09 Manifest0 left
xinming !paste 08:12
timotimo: paste.scsys.co.uk/584453
This is where the segfault happen 08:13
08:14 Manifest0 joined
xinming in gdb, what I do next? 08:16
Geth doc: 73314def1d | (JJ Merelo)++ | doc/Type/Metamodel/ParametricRoleHOW.pod6
Rephrasing
 08:19
doc: 4311d97033 | (JJ Merelo)++ | 3 files
Rewrites user example compilation test

Now everything is compiled via -c, instead of using EVAL. This is slower, but closes #2764 since apparently the evaluation of so much code was gobbling up memory, and closes #2782 since either the memory itself or the creation of multiple objects was affecting this somehow. It also causes #2789, so some changes are trying to fix that.
synopsebot Link: doc.perl6.org/type/Metamodel::ParametricRoleHOW
doc: 490133e6a0 | (JJ Merelo)++ | doc/Type/Test.pod6
Adds preambles closes #2789
synopsebot Link: doc.perl6.org/type/Test
timotimo xinming: sorry, i was AFK for a little bit 08:21
a "bt full" could be interesting
probably not quite as interesting is "thread apply all bt full"
most threads will be waiting at that point i assume 08:22
08:23 nadim joined
timotimo i can say for sure that the spesh/frame_walker.c shouldn't crash like this :) 08:24
can you "print f", "print f->spesh_cand"?
xinming timotimo: paste.scsys.co.uk/584454 08:25
timotimo ok, cand = 0x0 is an interesting output 08:26
xinming I'm restart the test, so, I can get "print f" 08:27
timotimo oh, right, you may want "--optimize=0" so that "optimized out" goes away
xinming Do you mean in configure? 08:28
a minute
re-compiling
timotimo yeah, in moarvm's configure 08:30
08:31 imcsk8 left
xinming timotimo: paste.scsys.co.uk/584455 08:31
with the --optimize=0
No symbol "f" in current context. 08:32
If I type print f in gdb cli, I got that error.
08:33 imcsk8 joined
timotimo interesting; perhaps you have to first "frame 0" to "go to" the go_to_first_inline frame 08:34
xinming timotimo: paste.scsys.co.uk/584456 08:35
timotimo oh, interesting 08:37
what's your moarvm version?
in my version there's an "if (f->spesh_cand && f->spesh_cand->inlines) {" on line 74
xinming This is MoarVM version 2019.03-110-g081823ffc built with JIT support
timotimo can you give me the output of "list"? that should give the source code from around that place 08:38
xinming if (f->spesh_cand && f->spesh_cand->inlines) {
I also have this.
What do you mean the output of "list"?
timotimo the command "list" gives you source code from where the program currently is
xinming Do you mean the source code for src/spesh/frame_walker.c 08:39
ok
timotimo more like "what does gdb think is there"
xinming paste.scsys.co.uk/584457 08:40
timotimo huh. so it segfaults on line 75? and f->spesh_cand is null?
xinming I think so 08:42
timotimo wtf how does it get past that check :( 08:43
xinming The problem is, it may take a while to run the problem for this to occur.
timotimo right
08:43 |oLa| left
xinming I think we also need to check f->spesh_cand->jitcode? 08:44
soemthing like if (f->spesh_cand && f->spesh_cand->inlines && f->spesh_cand->jitcode) ....
timotimo that check is on the next line, but if f->spesh_cand is already null, how would it even get past that
unless another thread is messing with the data
xinming I think so.
timotimo MVMFrame shouldn't be accessed from threads the frame doesn't belong to 08:45
xinming we need a lock somewhere :-)
timotimo if you're okay with a bunch more work, you could install rr, which can record the execution of the whole program and lets you step forwards as well as backwards 08:46
xinming If you wish to continue troubleshooting the bug with me, I can install rr
I'm installing rr now 08:48
timotimo worst case you can share the rr recording, but that will leak lots of data from your computer and what the program was doing so that may not be an option actually
xinming timotimo: It's ok, Not much sensitve data at all. 08:50
timotimo anyway, you haven't used rr before, right? 08:51
xinming Never used.
timotimo OK
so with "rr record -n perl6-m blahblah" will do the recording and it'll just segfault like the regular program
xinming I used gdb for a week, and that's really a long time ago.
Ok
timotimo then you can "rr replay" and it'll give you a gdb that's connected to the recording
xinming got it.
Still compiling 08:52
timotimo i'm not sure how you can get the right number to "goto" to, but in theory you don't have to "c" to let it run until it crashes again
or even if goto is the right command
oh, maybe not "goto", but "run" 08:54
so try "run 100000" and see if that's too far or too early
08:56 kurahaupo joined
timotimo and retry with progressively larger numbers until it complains i guess :) 08:59
cpan-p6 New module released to CPAN! Math::Polygons (0.0.4) by 03PSIXSTEVE 09:01
09:06 Manifest0 left, debprav joined
xinming rr: Saving execution to trace directory `/home/xm/.local/share/rr/perl6-gdb-m-0'. 09:08
[FATAL /tmp/portage/dev-util/rr-5.2.0-r1/work/rr-5.2.0/src/PerfCounters.cc:197:get_cpu_microarch()] AMD CPUs not supported.
For Ryzen, see github.com/mozilla/rr/issues/2034.
>_<
I'll think other way to use rr to debug this
I have many machines 09:09
09:15 Manifest0 joined 09:19 Manifest0 left
ufobat_ is this supposted to work? 09:20
timotimo nooooooooo 09:21
ufobat_ m: Mu.can("something"))
camelia 5===SORRY!5=== Error while compiling <tmp>
Unexpected closing bracket
at <tmp>:1
------> 3Mu.can("something")7⏏5)
ufobat_ m: Mu.can("something").say
camelia ()
ufobat_ m: Mu.^can("something").say # thats the only version i would expect to work
camelia ()
09:22 |oLa| joined
timotimo hm, yeah, that's kind of odd that it'd have a can method of its own 09:23
09:27 Manifest0 joined 09:30 pecastro joined
ufobat_ created a issue for that 09:35
09:42 Manifest0 left 09:44 Sgeo left, Sgeo joined 09:45 rindolf joined 09:47 Manifest0 joined 09:51 Manifest0 left 09:58 Manifest0 joined 10:01 andrzejku left 10:03 Sgeo left 10:08 debprav left 10:12 Manifest0 left 10:18 Manifest0 joined 10:20 Sgeo joined 10:24 Sgeo left 10:30 andrzejku joined 10:36 Manifest0 left 10:42 Manifest0 joined 10:46 andrzejku left, Manifest0 left 10:51 Manifest0 joined
nine Yeah, a little sugar makes all the difference: SQL::Generator.new.generate(select(*, from(subquery(select(*, from('customers')), :as<c>)))).say 10:58
SELECT * FROM (SELECT * FROM customers) AS c
With the option to go deeper at every level if needed of course 10:59
11:04 Manifest0 left
timotimo a teaspoon of sugar helps the SQL go down 11:08
11:12 Manifest0 joined 11:18 Manifest0 left 11:19 sena_kun joined 11:23 Manifest0 joined 11:26 pmurias joined
pmurias SmokeMachine: using Red would introduce an OO layer, if I understood what nine wants he just wants to build SQL programatically without extra OO abstraction 11:27
yoleaux 11 May 2019 19:23Z <MasterDuke> pmurias: have you seen news.ycombinator.com/item?id=19875335 ? seems like something you might be interested in
11:28 Manifest0 left
pmurias xinming: importing a perl6 module into nodejs is not implemented yet, it's definitely in reach so I'll look into it when I get back to rakudo.js soonish 11:31
11:32 netrino joined 11:33 Manifest0 joined 11:35 llfourn left
nine Getting there: select(*, from('customers'), where(boolean_test(comparison(|<id = 1>)))) 11:39
pmurias: it's not mutually exclusive. In fact I think the SQL AST will lend itself to integration into ORMs quite nicely.
11:40 Manifest0 left 11:41 llfourn joined
nine I think the mistake all ORMs so far made is to start with objects representing roles and simple CRUD features and add additional SQL features on top. This must inevitably become harder and harder as the implemented features become more complex. I'm for example not aware of an ORM that still shines in the face of common table expressions or window functions. 11:43
SmokeMachine pmurias: yes, but for doing that, Red uses a AST internally... it is planned to write a way to direct manipulate this AST...
nine s/roles/rows/
11:44 Manifest0 joined 11:46 holyghost left, MasterDuke joined, MasterDuke left, MasterDuke joined
SmokeMachine nine: I think Red is different from that... once the idea of it isn’t map SQL to object, but map perl6 to SQL... 11:47
11:56 molaf left
Geth doc: 098f2c456f | (JJ Merelo)++ | doc/Language/variables.pod6
Deindent
 11:56
doc: a7c52cdf76 | (JJ Merelo)++ | doc/Language/variables.pod6
Clarification about @*ARGS closes #2790
synopsebot Link: doc.perl6.org/language/variables
nine Who came up with names like "nonparenthesized value expression primary"? 12:02
12:03 Manifest0 left
lizmat a committee ? 12:04
nine Probably :) Though to be fair, they needed to come up with hunderts of names for the different syntax elements 12:05
12:09 Manifest0 joined
pmurias nine: sql {select * from customers where id = 1} seems like a lot nicer syntax 12:11
nine: sql {select * from customers where id = $id and {{generate-condition()}}}} in a more complex case 12:12
nine pmurias: absolutely! And once I have the SQL::Grammar going, we can have that. The sugar I mentioned is mostly for making writing of tests easier. Though I have to admit, that I'd already like to use it for $work...
pmurias to be honest {{{ }}} is beyond ugly tho 12:13
nine pmurias: do you know this? niner.name/talks/DBIC%20in%20Perl%2...revisited/
pmurias haven't seen that before
nine: that creates a DSL to build SQL::Abstract objects? 12:15
nine It's exploration of different ways to do that. One of them makes sql{"name=$name and active=$active"} actually safe, just like you proposed 12:16
12:23 Manifest0 left 12:26 ChoHag left 12:29 Manifest0 joined 12:35 Manifest0 left
nine Composing queries: my ($query, $subquery) = select(*, from(table('customers'))) xx 2; $query .= where(eq(column('id'), scalar_subquery($subquery))); # SELECT * FROM customers WHERE id = (SELECT * FROM customers) 12:38
pmurias nine: LTA compared to a grammar integrated solution, but seems very workable 12:40
12:40 Manifest0 joined
nine To a _hypothetical_ grammar integrated solution ;) But year, will invest some time in that after getting my exercise 12:41
12:44 jmerelo left
timotimo m: my @foo; with @foo[4] { say "entry 4 is already there!" } else { $_ = 99 }; dd @foo 12:45
camelia Array @foo = [Any, Any, Any, Any, 99]
timotimo i wonder what i did wrong in my local code
m: my @foo; with "hello" { with @foo[4] { say "entry 4 is already there!" } else { $_ = 99 }; dd @foo }
camelia Array @foo = [Any, Any, Any, Any, 99]
timotimo oh, native int array 12:46
m: my uint8 @foo; with "hello" { with @foo[4] { say "entry 4 is already there!" } else { $_ = 99 }; dd @foo }
camelia entry 4 is already there!
array[uint8].new()
timotimo even more interesting though: 12:47
m: class testclass { has uint8 @.foo; method doit { with "hello" { with @.foo[4] { say "entry 4 is already there!" } else { $_ = 99 }; dd @foo } } }; testclass.new.doit
camelia 5===SORRY!5=== Error while compiling <tmp>
Variable '@foo' is not declared. Did you mean '@!foo'?
at <tmp>:1
------> 3 already there!" } else { $_ = 99 }; dd 7⏏5@foo } } }; testclass.new.doit
timotimo m: class testclass { has uint8 @.foo; method doit { with "hello" { with @.foo[4] { say "entry 4 is already there!" } else { $_ = 99 }; dd @.foo } } }; testclass.new.doit
camelia assign requires a concrete object (got a Nil type object instead)
in block at <tmp> line 1
in method doit at <tmp> line 1
in block <unit> at <tmp> line 1
timotimo is my expectation unreasonable? 12:48
this easily golfs to a shorter thing: 12:53
nine SmokeMachine: to be clear: I think it's fantastic what you've done so far with Red!
timotimo m: class testclass { has uint8 @.foo; method doit { my $thing := @.foo[4]; $thing = 9; dd @!foo} }; testclass.new.doit
camelia assign requires a concrete object (got a Nil type object instead)
in method doit at <tmp> line 1
in block <unit> at <tmp> line 1
13:03 Manifest0 left 13:09 Manifest0 joined
Geth doc: 518d24b8fc | (Jonathan Worthington)++ | 2 files
Improve Lock and Lock::Async documentation

  * Better explain what Lock::Async is
  * Provide a more complete and accurate introduction to Lock, moving
   general information of the mechanism out of the description of the
   lock method
  * Improve the lock, unlock, and protect descriptions for both
... (5 more lines)		 13:12
13:19 lucasb joined 13:22 pmurias left 13:38 Manifest0 left 13:43 reach_satori left, reach_satori joined 13:44 Manifest0 joined
Voldenet I wish github.com/perl6/roast/blob/master...truction.t worked, locks would be a lot easier to use 13:47
13:48 Manifest0 left 13:55 Manifest0 joined
timotimo in what way does it not work? 13:56
oh, undefine
that's not a thing any more, is it?
m: my $foo; undefine $foo
camelia Saw 1 occurrence of deprecated code.
================================================================================
undefine seen at:
<tmp>, line 1
Will be removed with release v6.e!
Please use another way: assign a Nil; for Arrays/Hashe…
timotimo oh, it is still a thing, wow
though i don't think DESTROYALL is a thing?
at least i've never heard of that 13:57
or is that just what calls every applicable DESTROY submethod?
the number 100 is not high enough to actually cause a GC run 14:00
but parent and child do run backwards 14:01
is there no ticket for this? 14:02
if not, one should be made. if there is one, it should be put in the test file for the fudge :)
Voldenet: what exact problem do you have? you need the order to be "child first" and get the opposite? 14:03
Voldenet well, there's no destructors :) 14:04
/at all/
14:05 ChoHag joined
Voldenet it's not a big deal though - hardly any language has working destructors 14:07
14:07 ChoHag left
timotimo um 14:07
there is destructors 14:08
m: class bloop { submethod DESTROY { say "CARNAGE!" } }; for ^100_000 { bloop.new }
camelia ( no output )
timotimo m: class bloop { submethod DESTROY { say "CARNAGE!" } }; my $a = bloop.new; for ^100_000 { $a = bloop.new }
camelia ( no output )
timotimo m: class bloop { submethod DESTROY { say "CARNAGE!" } }; my $a = bloop.new; for ^1_000_000 { $a = bloop.new }
camelia ( no output )
timotimo um, what.
Voldenet :P
timotimo works locally 14:09
m: use nqp; class bloop { submethod DESTROY { say "CARNAGE!" } }; my $a = bloop.new; for ^1_000_000 { $a = bloop.new; nqp::force_gc }
oh, one million gc runs is gonna be a bit slow
camelia (timeout)
Voldenet m: sub { class :: { submethod DESTROY { "beep".say }}.new() }(); use nqp; nqp::force_gc 14:10
camelia ( no output )
timotimo m: use nqp; class bloop { submethod DESTROY { say "CARNAGE!" } }; my $a = bloop.new; for ^1000 { $a = bloop.new; nqp::force_gc }
camelia ( no output )
timotimo wtf camelia
14:10 debprav joined
Voldenet doesn't work for me either 14:10
timotimo hm. for 1000 i get only three carnage outputs
Voldenet for 100 i get none 14:11
timotimo i thought perhaps the optimizer threw out the entire bloop creation and destruction
Voldenet for 1000 i get 999 ಠ_ಠ
timotimo but MVM_SPESH_DISABLE=yes doesn't give me carnage
without force_gc it works, too
anyway, destruction isn't guaranteed to happen timely, or at all (if the program exits before the object needs to be destroyed, say) 14:12
14:13 atweiden-air joined
timotimo but destructors definitely do exist 14:13
Voldenet Yeah, but with their behaviour it's kind of java-ish - you do have them
but relying on them for things like locks is a terrible idea
timotimo ah, yeah 14:14
if you're waiting on a lock that's supposed to be destroyed by GC, you won't be generating objects, nor GC pressure
$lock.protect: { blah } is probably good enough for a whole lot of cases though 14:15
14:16 namibj_ is now known as namibj
nine Voldenet: what would destructors give you that you can't do with $lock.protect or LEAVE $lock.unlock? 14:18
14:18 debprav left
atweiden-air i just released an awesome perl 6 arch and void linux installer for btrfs fde if anyone's interested 14:18
github.com/atweiden/archvault 14:19
github.com/atweiden/voidvault
Voldenet nine: with($logger.Scope) { $logger.Log("Something"); } - if destructors worked, $logger could synchronize accesses to a file with this 14:21
but currently, it's just more feasible to queue the log requests and dispatch saving to file in a thread
nine Voldenet: why not just $logger.Scope: { $logger.Log("Something"); }? 14:22
Voldenet it works with this code, but then you're forced to create a lot of lambdas 14:24
14:26 atweiden-air left
nine Which is downright trivial and not different from a with statement 14:27
Voldenet it is going to lead to deeper stack traces 14:30
nine Are you worried about performance or about readability of those stack traces there? 14:31
Voldenet both :)
imagine composing a lot of objects with tracked lifetime
14:32 ChoHag joined
nine For performance it's not that bad. It may as well happen that the block is inlined into a $logger.Scope method that's again inlined into the caller. There'd be no stack frames or code objects anywhere in that case. 14:32
timotimo put an "is-hidden-from-backtrace" somewhere :D
nine Perl 6 actually relies on the runtime to be smart like that. Think for example about negative array indexing. That's a lambda right there: @a[*-1]
And for loops are actually compiled to calls to .map 14:33
timotimo except sometimes we compile them into a while loop with .pull-one on the .iterator 14:34
Voldenet + to use object with its lifetime, you need to know to track its lifetime 14:35
you can't just { my $log = $logger.Scope; $log.Log("…"); } without a care in the world 14:36
timotimo if you "will leave .close" you can though
Voldenet you actually need: { my $log = $logger.Scope; LEAVE $log.dispose; $log.Log("…"); } 14:37
timotimo i guess that's your point
no, that's more than you need
m: { my Str $foo will leave .say; $foo = 99 }; say "yo"
camelia 5===SORRY!5=== Error while compiling <tmp>
Missing block
at <tmp>:1
------> 3{ my Str $foo will leave7⏏5 .say; $foo = 99 }; say "yo"
expecting any of:
block or pointy block
timotimo m: { my Str $foo will leave { .say }; $foo = 99 }; say "yo"
camelia (Str)
Type check failed in assignment to $foo; expected Str but got Int (99)
in block <unit> at <tmp> line 1
nine Voldenet: or: $logger.Scope: { $log.Log("...") }
timotimo m: { my Str $foo will leave { .say }; $foo = 100000 }; say "yo"
camelia (Str)
Type check failed in assignment to $foo; expected Str but got Int (100000)
in block <unit> at <tmp> line 1
timotimo m: { my Str $foo will leave { .say }; $foo = "100000" }; say "yo"
camelia 100000
yo
timotimo no, timo, 100000 isn't a Str either
nine As to performance: the lambda trick actually gives the runtime more opportunity to optimize. The runtime would hardly be able to optimize the Scope object away in the destructor case 14:40
timotimo nine: actually, with our escape analysis we might be able to do that, if the destructor behaves at least :) 14:42
14:44 pmurias joined
Voldenet actually, with destructors you could return a scope with some object 14:45
so object could have opened logger for its lifetime
timotimo if you have that kind of lifetime stuff then you can just do that 14:54
Voldenet Well, currently I'm just passing the scope object with on-end triggers
timotimo with something not not called "Scope"
so one method for getting one with a destructor and one for "will leave" or using the $log.Scope: { blah } thing or whatevs
Voldenet m: class Lifetime { has @!s; method LEAVE(&m) { @!s.push(&m) }; method end { .() for @!s } }; sub req { my $s = Lifetime.new; LEAVE $s.end; $s.LEAVE: { say "bye" } }; req; req
camelia bye
bye
Voldenet it's terrible, but it works
in fact, I attach it to the thread with $*THREAD.id so I don't have to pass it around, but it's still a terrible hack 14:58
I'd gladly see some better method - the one I'm using now looks bad and feels bad to use 15:02
15:04 Manifest0 left
ugexe $*THREAD.id doesn't identify only a specific thread over the lifetime of a program 15:04
Voldenet Whoa, it doesn't? 15:05
ugexe when i needed to do this i ended up doing %cache{$*THREAD} = ... instead of %cache{$*THREAD.id} 15:06
Voldenet thanks ugexe, I'll change it then
15:11 Manifest0 joined 15:19 zakharyas joined 15:20 Manifest0 left 15:25 Manifest0 joined 15:31 Sgeo joined, jmerelo joined, cpan-p6 left 15:32 Manifest0 left, cpan-p6 joined, cpan-p6 left, cpan-p6 joined 15:37 Manifest0 joined 16:06 Manifest0 left 16:11 Manifest0 joined 16:16 nebuchadnezzar joined, vrurg left 16:22 Manifest0 left 16:28 Manifest0 joined 16:34 Manifest0 left 16:38 molaf joined 16:39 Manifest0 joined 16:48 Kaypie is now known as Kaiepi
xinming timotimo: ping 16:51
timotimo: When you are at computer, Please ping me, I got intel computer now, and will try with rr 16:55
though, it's a 3930 machine, which is really slow
timotimo xinming: i'm here now, but i'll be AFK for maybe an hour 16:56
xinming Ok
timotimo i mean, i'm here for a few seconds before being AFK
Voldenet m: use MONKEY-SEE-NO-EVAL; my $fn = EVAL("sub { print 42 }"); $fn();
camelia 425===SORRY!5=== Error while compiling /home/camelia/EVAL_0
Missing block
at /home/camelia/EVAL_0:1
------> 3sub True7⏏5<EOL>
expecting any of:
new name to be defined
16:57 Manifest0 left
xinming timotimo: Ok, I'll setup the test environment. 16:57
Kaiepi +1 jnthn on the Lock/Lock::Async docs rewrite 16:58
16:58 zakharyas left
Kaiepi much better now :) 16:58
Voldenet How to do something similar to: `my $fn = eval("sub { print 42; }"); $fn->();` in perl 5? I'm trying, but I probably don't know how to use eval 17:01
17:01 sena_kun left 17:02 Manifest0 joined
Voldenet m: use MONKEY-SEE-NO-EVAL; my $str = q`sub { print 42; }`; EVAL($str)() 17:03
camelia 42
Voldenet Eh, Nevermind
17:09 Manifest0 left, ayerhart left
xinming Anyone here knows, for rr utility, What if we need network support? For now, I found that when I start the app with rr, It doens't start the cro service 17:12
17:13 Manifest0 joined, ayerhart joined 17:14 ayerhart left, ayerhart joined 17:22 Manifest0 left
Geth doc: af2115cc0e | (Daniel Green)++ | 2 files
Some typo fixes
 17:24
xinming I'm wrong to use perl6-m-dbg with rr
17:26 vrurg joined 17:27 Manifest0 joined
Kaiepi i'm having an issue with blobs 17:29
lizmat weekly: blogs.perl.org/users/laurent_r/2019...dders.html
notable6 lizmat, Noted!
Kaiepi can someone take a look at hastebin.com/azepasogup.rb and tell me what i need to do to get the output i want?
17:31 sena_kun joined
lizmat dd 17:33
Kaiepi: what does `dd $!data` say ?
Kaiepi Buf $!data = Buf.new(195,191,195,186,195,191,195,176) 17:34
it's supposed to be latin1 but it looks like it's utf8 instead
not sure why
xinming timotimo: What do we do if this bug appear on AMD ryzen 2700, But when I try with G3930 from intel, It works fine. 17:36
timotimo: I'm not sure wether it's a cpu problem on SMT in ryzen, or a problem kernel deals with SMT in ryzen 17:37
hmm, I'll recompile the linux kernel without optimization. 17:38
Kaiepi ok i figured out why 17:42
IO::Socket::Async.Supply ignores :enc if :bin is also passed (???) 17:43
ugexe well, the only other sane option there would be to throw an exception 17:44
lizmat Kaiepi: yeah, it does, also with IO::Handle afaik 17:46
Kaiepi i tried without :bin and it's still ignoring the encoding? 17:47
also tried with IO::Socket::Async.connect(:enc<latin1>)
lizmat then maybe we haz a bug ? 17:48
causal glance at the code, it looks like it doesn't even check for a :bin ? 17:49
*casual
17:50 TreyHarris joined
lizmat docs.perl6.org/type/IO::Socket::As...od_connect doesn't even mention an :enc parameter 17:51
fo looks like it is also a doc issue
jmerelo hears doc 17:52
Kaiepi wait hold on i forget to set the encoding of the server to latin1 as well
jmerelo There are a few issues: github.com/rakudo/rakudo/issues/2039 17:53
Anyway, you find anything missing, please raise an issue. That might be part of the 6.d unfinished update, but it might hide something deeper. An issue will help anyway 17:55
Kaiepi yeah, that was the issue
i got it working properly now 17:56
jmerelo Kaiepi: great! 17:57
18:00 adu joined
Xliff m: ('Line' «~» ^6).join("\n").say 18:04
yoleaux 07:27Z <holyghost> Xliff: you did ,tell instead of .tell yesterday, I sent you some email for the new permanent ip
camelia Line0
Line1
Line2
Line3
Line4
Line5
yoleaux 07:32Z <holyghost> Xliff: kibology :-)
Xliff m: sub a ($a) { $a.say }; a(1|3) 18:06
camelia 1
3
Xliff m: sub a ($a) { $a.say }; |a(1|3)
camelia 1
3
Xliff m: sub a ($a) { $a }; a(1|3)
camelia ( no output )
Xliff m: sub a ($a) { $a }; a(1|3).gist.say
camelia any(1, 3)
Xliff m: sub a ($a) { $a }; |a(1|3).say
camelia any(1, 3)
Xliff m: sub a ($a) { $a }; |a(1|3).flat
camelia ( no output )
Xliff m: sub a ($a) { $a }; |a(1|3).flat.say 18:07
camelia any((1), (3))
xinming also occur on G3930, So, I'll use rr to record it
Xliff How do I turn a junction into a list?
m: sub a ($a) { $a }; |a(1|3).List
camelia ( no output )
Xliff m: sub a ($a) { $a }; a(1|3).List.say
camelia any((1), (3))
Xliff m: sub a ($a) { $a }; a(1|3).Array.say
camelia any([1], [3])
jmerelo finds that Perl questions in StackOverflow reached a new minimun last month, since early 2009 data.stackexchange.com/stackoverfl...0977#graph 18:09
Xliff m: sub a ($a) { $a }; .say for a(1|3);
camelia (Mu)
jmerelo m: say (1^3).iterator.pull-one 18:10
camelia (Mu)
jmerelo Xliff: ^^^ Does that help? 18:11
Xliff Nope! ;)
jmerelo Xliff: what do you wanna do?
Xliff I think I found the answer here: perlgeek.de/blog-en/perl-5-to-6/08...tions.html
See "Word of Warning" at the end.
MasterDuke m: sub a($b) { take $b }; my @c = gather a(1|2); dd @c
camelia Array @c = [1, 2]
Xliff "If you try to extact items from a junction, you are doing it wrong" 18:12
Which is sad. Because that's what I was trying to do.
jmerelo Ah, OK, right.
Xliff: you were doing it wrong!
Xliff I was trying to do:
m: ($v1, $v2) = $obj.method(1|3)
camelia 5===SORRY!5=== Error while compiling <tmp>
Variable '$v1' is not declared
at <tmp>:1
------> 3(7⏏5$v1, $v2) = $obj.method(1|3)
Xliff "($v1, $v2) = $obj.method(1|3)"
jmerelo Xliff: you kinda wanted to do autothreading in a method by using Junctions 18:13
Xliff Just so I wouldn't have to do "($v1, $v2) = ($obj.method(1), $obj.method(3))
jmerelo Xliff: In parallel! In constant time!
Xliff: you can still do >>
Xliff ?? 18:14
Example, pls?
jmerelo Xliff: let me see
m: class Foo { has $.bar; method zape { $!bar*2 }; my $foo = Foo.new(bar=> 3); my $bar = Foo.new(bar => 5 ); say ($foo,$bar)».zape 18:15
camelia 5===SORRY!5=== Error while compiling <tmp>
Missing block
at <tmp>:1
------> 3oo.new(bar => 5 ); say ($foo,$bar)».zape7⏏5<EOL>
expecting any of:
postfix
statement end
statement modifier
sta…
jmerelo m: class Foo { has $.bar; method zape { $!bar*2 }; my $foo = Foo.new(bar=> 3)}; my $bar = Foo.new(bar => 5 ); say ($foo,$bar)».zape 18:16
camelia 5===SORRY!5=== Error while compiling <tmp>
Variable '$foo' is not declared. Did you mean 'Foo'?
at <tmp>:1
------> 033)}; my $bar = Foo.new(bar => 5 ); say (7⏏5$foo,$bar)».zape
jmerelo m: class Foo { has $.bar; method zape { $!bar*2 }}; my $foo = Foo.new(bar=> 3); my $bar = Foo.new(bar => 5 ); say ($foo,$bar)».zape
camelia (6 10)
jmerelo Xliff: ^^^ that
Xliff Hmm...
No. That's the wrong way! :) 18:17
jmerelo OK, because you want the arguments, right?
Xliff "($v1, $v2) = ( $obj.method(1), $obj.method(3) )"
jmerelo right
Let me see
Xliff $obj doesn't vary.
jmerelo m: class Foo { has $.bar; method zape( $zipi) { $!bar*$zipi }}; my $foo = Foo.new(bar=> 3); say $foo.zape».()» (3,6) 18:19
camelia 5===SORRY!5=== Error while compiling <tmp>
Missing postfix
at <tmp>:1
------> 3o = Foo.new(bar=> 3); say $foo.zape».()»7⏏5 (3,6)
expecting any of:
argument list
postfix
18:20 pmurias left
jmerelo m: class Foo { has $.bar; method zape( $zipi) { $!bar*$zipi }}; my $foo = Foo.new(bar=> 3); say $foo.zape().» (3,6) 18:21
camelia 5===SORRY!5=== Error while compiling <tmp>
Missing postfix
at <tmp>:1
------> 3oo = Foo.new(bar=> 3); say $foo.zape().»7⏏5 (3,6)
expecting any of:
argument list
postfix
jmerelo m: class Foo { has $.bar; method zape( $zipi) { $!bar*$zipi }}; my $foo = Foo.new(bar=> 3); say $foo.zape.» (3,6)
camelia 5===SORRY!5=== Error while compiling <tmp>
Missing postfix
at <tmp>:1
------> 3$foo = Foo.new(bar=> 3); say $foo.zape.»7⏏5 (3,6)
expecting any of:
method arguments
postfix
Kaiepi Net::Telnet is almost complete! i just need to write more documentation and some example code and release 0.0.1 will be ready 18:22
Xliff m: class A { method($a) { $a² }; }; my $obj = A.new; ($v1, $v2) « a() « (1, 3)
camelia 5===SORRY!5=== Error while compiling <tmp>
Variable '$a' is not declared
at <tmp>:1
------> 3class A { method(7⏏5$a) { $a² }; }; my $obj = A.new; ($v1, $
Kaiepi well
"complete"
18:22 pmurias joined
Kaiepi it will never be complete with how many options telnet has 18:22
Xliff m: class A { method a ($a) { $a² }; }; my $obj = A.new; ($v1, $v2) « a() « (1, 3)
camelia 5===SORRY!5=== Error while compiling <tmp>
Variable '$v1' is not declared
at <tmp>:1
------> 3od a ($a) { $a² }; }; my $obj = A.new; (7⏏5$v1, $v2) « a() « (1, 3)
Xliff m: class A { method a ($a) { $a² }; }; my $obj = A.new; my ($v1, $v2) « a() « (1, 3)
camelia 5===SORRY!5=== Error while compiling <tmp>
Missing infix inside hyper
at <tmp>:1
------> 3² }; }; my $obj = A.new; my ($v1, $v2) «7⏏5 a() « (1, 3)
expecting any of:
infix
infix stopper
jmerelo Xliff: I'm pretty sure there's some trickery to do that...
Xliff m: class A { method a ($a) { $a² }; }; my $obj = A.new; my ($v1, $v2) « $a.a() « (1, 3) 18:23
camelia 5===SORRY!5=== Error while compiling <tmp>
Missing infix inside hyper
at <tmp>:1
------> 3² }; }; my $obj = A.new; my ($v1, $v2) «7⏏5 $a.a() « (1, 3)
expecting any of:
infix
infix stopper
jmerelo m: class Foo { has $.bar; method zape( $zipi) { $!bar*$zipi }}; my $foo = Foo.new(bar=> 3); say $foo.zape.(3)
camelia Too few positionals passed; expected 2 arguments but got 1
in method zape at <tmp> line 1
in block <unit> at <tmp> line 1
jmerelo m: class Foo { has $.bar; method zape( $zipi) { $!bar*$zipi }}; my $foo = Foo.new(bar=> 3); say $foo.zape.($foo,3)
camelia Too few positionals passed; expected 2 arguments but got 1
in method zape at <tmp> line 1
in block <unit> at <tmp> line 1
Xliff m: class A { method a ($a) { $a² }; }; my $obj = A.new; my ($v1, $v2) = $a.a .« (1, 3) 18:24
camelia 5===SORRY!5=== Error while compiling <tmp>
Variable '$a' is not declared
at <tmp>:1
------> 3}; }; my $obj = A.new; my ($v1, $v2) = 7⏏5$a.a .« (1, 3)
Xliff m: class A { method a ($a) { $a² }; }; my $obj = A.new; my ($v1, $v2) = $obj.a .« (1, 3)
camelia 5===SORRY!5=== Error while compiling <tmp>
Malformed postfix call (only alphabetic methods may be detached)
at <tmp>:1
------> 3 $obj = A.new; my ($v1, $v2) = $obj.a .7⏏5« (1, 3)
Xliff m: class A { method a ($a) { $a² }; }; my $obj = A.new; my ($v1, $v2) = $obj.a .» (1, 3)
camelia 5===SORRY!5=== Error while compiling <tmp>
Malformed postfix call (only alphabetic methods may be detached)
at <tmp>:1
------> 3 $obj = A.new; my ($v1, $v2) = $obj.a .7⏏5» (1, 3)
Xliff m: class A { method a ($a) { $a² }; }; my $obj = A.new; my ($v1, $v2) = $obj.a().» (1, 3) 18:25
camelia 5===SORRY!5=== Error while compiling <tmp>
Missing postfix
at <tmp>:1
------> 3obj = A.new; my ($v1, $v2) = $obj.a().»7⏏5 (1, 3)
expecting any of:
argument list
postfix
Xliff Ah well. Thanks for trying, jmerelo
jmerelo Xliff: Only alphabetic methods might be detached?
Xliff: maybe try StackOverflow?
Xliff: github.com/rakudo/rakudo/issues/2092 18:26
timotimo are you looking for <== instead of hyperops? 18:27
Xliff Maybe.
timotimo not sure it works with method calls though
Xliff timotimo: "($v1, $v2) = ( $obj.method(1), $obj.method(3) )"
Is there a shortcut so that I can just specify (1, 3) in some way?
MasterDuke .map over them?
jmerelo m: 42 .&abs 18:28
camelia 5===SORRY!5=== Error while compiling <tmp>
Malformed postfix call (only alphabetic methods may be detached)
at <tmp>:1
------> 0342 .7⏏5&abs
timotimo m: class Bloop { method a($a) { $a² } }; my ($v1, $v2) = (1, 3)».&[Bloop.a($_)}
camelia 5===SORRY!5=== Error while compiling <tmp>
Missing infix inside []
at <tmp>:1
------> 3a) { $a² } }; my ($v1, $v2) = (1, 3)».&[7⏏5Bloop.a($_)}
expecting any of:
dotty method or postfix
infix
infix …
timotimo m: class Bloop { method a($a) { $a² } }; my ($v1, $v2) = (1, 3)».&[Bloop.a($_)]
camelia 5===SORRY!5=== Error while compiling <tmp>
Missing infix inside []
at <tmp>:1
------> 3a) { $a² } }; my ($v1, $v2) = (1, 3)».&[7⏏5Bloop.a($_)]
expecting any of:
dotty method or postfix
infix
infix …
jnthn ($v1, $v2) = ($obj.method($_) for 1, 3)
timotimo m: class Bloop { method a($a) { $a² } }; my ($v1, $v2) = (1, 3)».&[{Bloop.a($_)}]
camelia 5===SORRY!5=== Error while compiling <tmp>
Missing infix inside []
at <tmp>:1
------> 3a) { $a² } }; my ($v1, $v2) = (1, 3)».&[7⏏5{Bloop.a($_)}]
expecting any of:
dotty method or postfix
infix
infi…
timotimo yeah the postfix for loop is surely the easiest
Xliff MasterDuke++ # /o\.... yeah.. that works. 18:29
jmerelo Xliff: but I thought you didn't like map for some reason. If you use map, you can also use the (badly named) feed operator ==> (I like rocket) 18:30
MasterDuke m: sub a($b) { take $b }; my $c = a(1); dd $c # is there some reason take without gather couldn't just act like return? 18:31
camelia take without gather
in sub a at <tmp> line 1
in block <unit> at <tmp> line 1
jmerelo m: class Bloop { method a($a) { $a² } }; say (1,3) ==> map: Bloop.a( * ) 18:32
camelia 5===SORRY!5=== Error while compiling <tmp>
Confused
at <tmp>:1
------> 3thod a($a) { $a² } }; say (1,3) ==> map:7⏏5 Bloop.a( * )
expecting any of:
colon pair
jmerelo m: class Bloop { method a($a) { $a² } }; say (1,3) ==> map( { Bloop.a( $_ ) } )
camelia (1 3)
Xliff m: class Bloop { method a($a) { $a² } }; my ($a, $b) <== map: Bloop.a( * ) <== (1, 3); ($a, $b).say 18:33
camelia 5===SORRY!5=== Error while compiling <tmp>
Only routine calls or variables that can '.push' may appear on either side of feed operators.
at <tmp>:1
------> 3class Bloop { method a($a) { $a² } }; 7⏏5my ($a, $b) <== map: Bloop.a( * ) <==…
Xliff m: class Bloop { method a($a) { $a² } }; my ($a, $b) = (1, 3).map: Bloop.a( * )
camelia Cannot resolve caller Numeric(Whatever:D: ); none of these signatures match:
(Mu:U \v: *%_)
in method a at <tmp> line 1
in block <unit> at <tmp> line 1
ugexe feed operator is going to be slower for most use cases.
Xliff m: class Bloop { method a($a) { $a² } }; my ($a, $b) = (1, 3).map( Bloop.a( * ) )
camelia Cannot resolve caller Numeric(Whatever:D: ); none of these signatures match:
(Mu:U \v: *%_)
in method a at <tmp> line 1
in block <unit> at <tmp> line 1
Xliff m: class Bloop { method a($a) { $a² } }; my ($a, $b) = (1, 3).map({ Bloop.a( $_ ) }) 18:34
camelia ( no output )
ugexe significantly slower is many cases even
Xliff m: class Bloop { method a($a) { $a² } }; my ($a, $b) = (1, 3).map({ Bloop.a( $_ ) }); ($a, $b).say
camelia (1 9)
moritz sometimes I just love Python's easy way to get a bound method. With python, you could just map over Bloop.a 18:36
18:38 netrino_ joined, netrino left 18:40 domidumont joined
nadim evening, what's the difference between "class.new() but some_role " and "class.new() but some_role.new()"? 18:53
In my dumper I see that the first object also does the role but the second one doesn't 18:54
ugexe .new on a role creates a class 18:57
nadim so what does "class.new but created_class" do? 19:02
m: class C {} ; role R { has $.x} ; dd C.new() but R ; dd C.new but R.new/) 19:04
camelia 5===SORRY!5=== Error while compiling <tmp>
Missing required term after infix
at <tmp>:1
------> 3; dd C.new() but R ; dd C.new but R.new/7⏏5)
expecting any of:
prefix
term
nadim m: class C {} ; role R { has $.x} ; dd C.new() but R ; dd C.new but R.new()
camelia C+{R}.new(x => Any)
C+{<anon|1>}.new
nadim it's 'annon' that's making me curious 19:05
ugexe github.com/rakudo/rakudo/blob/b2e7...rs.pm6#L65
i dunno what its doing
seems like it shold throw X::Mixin::NotComposable but isnt 19:06
nadim shall I report this somewhere? 19:08
19:09 jmerelo left
MasterDuke m: my $a = 2 but Str; say $a.indent(3) 19:10
camelia Cannot mix in non-composable type Str into object of type Int
in block <unit> at <tmp> line 1
MasterDuke m: my $a = 2 but Str.new; say $a.indent(3)
camelia No such method 'indent' for invocant of type 'Int+{<anon|1>}'. Did you mean any of these?
index
invert

in block <unit> at <tmp> line 1
ugexe probably. im failing to think of a reason the current behavior could be useful
nadim I had no other to try than to see if it would generate an exception that I catch when dumping it but it simply worked! 19:12
19:15 domidumont left 19:16 antoniogamiz joined
antoniogamiz mm should the result of use cmp always return true? O.o 19:16
m: if (1 cmp 2) {say 2;}
camelia 2
antoniogamiz if (2 cmp 1) {say 2;}
evalable6 2
antoniogamiz ah if they are equal it returns false, sorry 19:17
MasterDuke cmp returns an Order 19:18
docs.perl6.org/type/Order
antoniogamiz ty 19:19
MasterDuke np
antoniogamiz mmm methods throwing exceptions should be documented? 19:28
I mean: say (1..2) cmp 3 # OUTPUT: «Less␤»
ups, I meant github.com/rakudo/rakudo/blob/d904...e.pm6#L594 19:29
19:39 vrurg left 19:44 molaf left
Geth doc: 0899a013fa | Antonio++ | doc/Type/Range.pod6
document Range infixs #2739
 19:53
doc: 2729ed3a47 | Antonio++ | doc/Type/Range.pod6
document the rest of the methods, closes #2739
synopsebot Link: doc.perl6.org/type/Range
19:53 cpan-p6 left 19:54 cpan-p6 joined, cpan-p6 left, cpan-p6 joined
Kaiepi i never thought i'd ever need to use a supplier that emits more supplies 19:56
but i found myself doing just that
i'm adding support for receiving multiple binary transmissions over a single connection to Net::Telnet 19:57
originally i had it so whenever binary data was received, it'd emit chunks received to a supplier, but that's not good enough for if multiple binary transmissions occur 19:58
so now when a binary transmission starts i emit a supply to a supplier and emit chunks received to that supply's supplier 19:59
20:03 Xliff left 20:08 ufobat__ joined 20:12 ufobat_ left 20:13 Manifest0 left, antoniogamiz left 20:16 jaldhar left, jaldhar joined 20:19 Manifest0 joined 20:34 zacts joined 20:35 molaf joined 20:44 Manifest0 left 20:51 Manifest0 joined 20:52 zacts left 20:58 Manifest0 left 21:03 Manifest0 joined, pmurias left
cpan-p6 New module released to CPAN! FindBin::Dirs (0.1.0) by 03LEMBARK 21:04
21:10 pmurias joined 21:15 Manifest0 left 21:21 Manifest0 joined 21:25 zacts joined 21:31 Manifest0 left 21:38 Manifest0 joined 21:49 zacts left 21:52 pmurias left 21:53 mowcat joined, mowcat left 21:54 mowcat joined 21:59 Manifest0 left 22:01 adu left 22:02 zacts joined, zacts left 22:04 Manifest0 joined 22:09 Manifest0 left 22:15 Manifest0 joined 22:20 Manifest0 left 22:22 pecastro left 22:26 Manifest0 joined 22:31 Cabanossi left 22:33 Cabanossi joined, rindolf left 22:42 netrino_ left, mowcat left 22:52 mowcat joined, Manifest0 left 22:57 mowcat left 22:58 Manifest0 joined, Yary joined
Geth ecosystem: da2dee9246 | (Ben Davies)++ (committed using GitHub Web editor) | META.list
Add Net::Telnet to ecosystem
 22:59
Kaiepi \o/
Yary m: sub MAIN(Int(IntStr) $value where $value ∈ (1,2,4,8,16)) {say "\$value.perl = ",$value.perl } 23:00
camelia Usage:
<tmp> <value>
23:01 MasterDuke left 23:02 nadim left
Yary m: perl6 -e 'BEGIN @*ARGS=IntStr.new(4,"4"),;sub MAIN(Int(IntStr) $value where $value ∈ (1,2,4,8,16)) {say "\$value.perl = ",$value.perl }' 23:05
camelia 5===SORRY!5=== Error while compiling <tmp>
Two terms in a row
at <tmp>:1
------> 3perl6 -e7⏏5 'BEGIN @*ARGS=IntStr.new(4,"4"),;sub MA
expecting any of:
infix
infix stopper
postfix
statement…
23:05 mowcat joined
Yary m: BEGIN @*ARGS=IntStr.new(4,"4"),;sub MAIN(Int(IntStr) $value where $value ∈ (1,2,4,8,16)) {say "\$value.perl = ",$value.perl } 23:05
camelia Usage:
<tmp> <value>
cpan-p6 New module released to CPAN! Net::Telnet (0.0.1) by 03KAIEPI 23:06
Yary m: BEGIN @*ARGS=<4>;sub MAIN(Int(IntStr) $value where $value ∈ <1 2 4 8 16>) {say "\$value.perl = ",$value.perl } 23:09
camelia $value.perl = IntStr.new(4, "4")
23:09 veesh left, zacts joined
Kaiepi m: Begin @*ARGS=<1,2,3,4>; sub MAIN(Int @foo) { .say for @foo } 23:11
camelia 5===SORRY!5=== Error while compiling <tmp>
Undeclared name:
Begin used at line 1
Kaiepi m: BEGIN @*ARGS=<1,2,3,4>; sub MAIN(Int @foo) { .say for @foo }
camelia Usage:
<tmp> <foo>
Yary - looks like recent Rakudo has better MAIN error than 6.c, I had seen messages like "Type check failed in binding to parameter '<anon>'; expected Any but got Mu (Mu) in block <unit> at main.p6 line 1" which is now showing as the better "Usage:␤ <tmp> <value>␤" 23:12
Kaiepi m: BEGIN @*ARGS=<1 2 3 4>; sub MAIN(*@foo) { .say for @foo }
camelia 1
2
3
4
Yary m: BEGIN @*ARGS=<1 2 3 4>; sub MAIN(*@foo) { .perl.say for @foo } 23:13
camelia IntStr.new(1, "1")
IntStr.new(2, "2")
IntStr.new(3, "3")
IntStr.new(4, "4")
Kaiepi m: BEGIN @*ARGS=<--a=1 --a=2 --a=3>; sub MAIN(Int :@a) { .say for @a }
camelia Usage:
<tmp> [-a=<Any> ...]
Kaiepi m: BEGIN @*ARGS=<-a=1 -a=2 -a=3>; sub MAIN(Int :@a) { .say for @a }
camelia Usage:
<tmp> [-a=<Any> ...]
Yary m: BEGIN @*ARGS=<-a=1 -a=2 -a=3>; sub MAIN(:@a) { .say for @a } 23:14
camelia 1
2
3
Yary m: BEGIN @*ARGS=<-a=1 -a=2 -a=3>; sub MAIN(:@a) { .perl.say for @a }
camelia IntStr.new(1, "1")
IntStr.new(2, "2")
IntStr.new(3, "3")
Yary m: BEGIN @*ARGS=<--a=1 --a=2 --a=3>; sub MAIN(Array[Int] :@a) { .perl.say for @a } 23:17
camelia Usage:
<tmp> [-a=<Any> ...]
Yary m: BEGIN @*ARGS=<--a=1 --a=2 --a=3>; sub MAIN(Array[IntStr] :@a) { .perl.say for @a }
camelia Usage:
<tmp> [-a=<Any> ...]
23:20 veesh joined
Yary m: BEGIN @*ARGS=<--a=1 --a=2 --a=3>; sub MAIN(:@a where {.all ~~ Int}) { .perl.say for @a } 23:23
camelia IntStr.new(1, "1")
IntStr.new(2, "2")
IntStr.new(3, "3")
Yary m: BEGIN @*ARGS=<--a=1 --a=2 --a=foo>; sub MAIN(:@a where {.all ~~ Int}) { .perl.say for @a }
camelia Usage:
<tmp> [-a=<Positional where { ... }> ...]
23:27 reach_satori_ joined, reach_satori left
timotimo Yary: Int(IntStr) will not coerce because IntStr is already an Int; coercions will only fire if the type doesn't match, like with Array and Hash for example 23:28
23:30 zacts left
Yary Thanks for the coercion explanation, I'll follow up in the email thread I started on it. This MAIN issue that Kaiepi hinted at is prompting me to file an issue on constraining its slurpy args 23:32
23:39 sena_kun left 23:45 zacts joined 23:50 Manifest0 left
Yary m: BEGIN @*ARGS=<4>;sub MAIN(Int $value is rw where ($value += 0) ∈ (1,2,4,8,16)) {say "\$value.perl = ",$value.perl } 23:51
camelia Usage:
<tmp> <value>
23:52 MasterDuke joined, MasterDuke left, MasterDuke joined
Yary m: BEGIN @*ARGS=<4>;sub MAIN(Int $value is rw where Int($value) ∈ (1,2,4,8,16)) {say "\$value.perl = ",$value.perl } 23:52
camelia Usage:
<tmp> <value>
Yary m: BEGIN @*ARGS=<4>;sub MAIN(Int $value where Int($value) ∈ (1,2,4,8,16)) {say "\$value.perl = ",$value.perl }
camelia $value.perl = IntStr.new(4, "4")
Yary Hmm, the "is rw" version doesn't match, but removing "is rw" allows matching. Why would that be? 23:53
23:54 zacts left, espadrine_ left

Powered by Cro and the Raku® Programming Language.

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

Thank you!