Zoffix This returns 0 for me :/ I wonder if that's correct: dd run(:out, $*EXECUTABLE, "-e", "exit 255").exitcode 00:03
This (no :out) returns 25. dd run($*EXECUTABLE, "-e", "exit 25").exitcode
ugexe run/shell are full of weird things like that 01:04
you would have to call .out.close first to get what you want as it current is 01:06
when using both :out and :err I encounter situations where the order that :out and :err are read in matters 01:07
and its not always the same
jdv79 sounds sad 01:28
jdv79 sheds a tear 01:29
AlexDaniel rt.perl.org/Ticket/Display.html?id=125757 01:59
Zoffix: ↑ this, I think
Zoffix yeah, seems to be 02:01
m: BEGIN { UNIT; Nil } 02:38
camelia ( no output )
Zoffix m: UNIT.WHAT
camelia ( no output )
Zoffix m: UNIT.WHAT.say
camelia rakudo-moar d789da: OUTPUT«(UNIT)␤»
Zoffix Interestingly, that BEGIN stuff dies when run on command line 02:42
(there's a ticket for it rt.perl.org/Ticket/Display.html?id...et-history ) 02:43
m: say NaN cmp -Inf 03:02
camelia rakudo-moar d789da: OUTPUT«More␤»
Zoffix m: say NaN cmp Inf
camelia rakudo-moar d789da: OUTPUT«More␤»
Zoffix m: say minmax NaN, Inf, -Inf;
camelia rakudo-moar d789da: OUTPUT«-Inf..NaN␤»
Zoffix m: say max NaN, Inf, -Inf;
camelia rakudo-moar d789da: OUTPUT«NaN␤»
Zoffix m: say min NaN, Inf, -Inf;
camelia rakudo-moar d789da: OUTPUT«-Inf␤»
Zoffix What does "# NaN always sorts as if "NaN" instead." mean? 03:16
geekosaur hm. I would understand that as meaning it uses the string instead of the putative numeric value (because the numeric value is not actually sortable) 03:19
but...
m: say minmax "NaN", Inf, -Inf;
camelia rakudo-moar d789da: OUTPUT«-Inf..Inf␤»
geekosaur m: say minmax NaN, Inf, -Inf; 03:20
camelia rakudo-moar d789da: OUTPUT«-Inf..NaN␤»
geekosaur (given that IEEE floating point says the result of a comparison involving a NaN is NaN, arguably minmax should produce either NaN or NaN..NaN) 03:21
...or a range involving NaN should produce Failure 03:22
Zoffix m: say NaN..NaN
camelia rakudo-moar d789da: OUTPUT«NaN..NaN␤»
Zoffix m: say eager NaN..NaN
geekosaur um...
camelia rakudo-moar d789da: OUTPUT«(timeout)» 03:23
geekosaur NaN is non-comparable, so ranges involving NaN really should be illegal as they have no meaning
Zoffix watches the worms escape from the now-open can 03:24
(started with this ticket :) rt.perl.org/Ticket/Display.html?id=127506 )
Marked as [@LARRY]. My vote is to return NaN if any of the values is a NaN: rt.perl.org/Ticket/Display.html?id...et-history 03:39
m: .say for NaN..NaN 03:44
camelia rakudo-moar d789da: OUTPUT«(timeout)NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤N…» 03:45
Zoffix
.oO( ... batman... )
geekosaur that follows from NaN being non-comparable; it will never reach the endpoint because the values can never compare equal
which is why it should be a Failure
guh. if I try to comment with anonymous login it won't let me. if I log in with bitcard it says I don't have display permission 03:49
Zoffix Another @LARRY for the NaN seq/ranges (there was actually a ticket for that since pre-Christmas): rt.perl.org/Ticket/Display.html?id=125337 03:50
Zoffix goes to bed
geekosaur right, logging in removed my access completely. thanks perl.org :/
dalek ast: 5cef331 | usev6++ | S05-metasyntax/charset.t:
Skip test for ignoremark on JVM (ordbaseat NYI)
 04:10
psch j: A: for ^1 { B: do for ^1 { say "in B"; last A } }; 08:46
camelia rakudo-jvm cd19db: OUTPUT«in B␤»
psch ...does that mean non-phasery sinking is wonky..?
this is crazy 09:17
from the looks of it the 'do' example above doesn't actually ever hit a handler that handles EX_CAT_LABELED
right, so part of the bug is definitely that we're doing (candidate_category & category != 0), instead of == category 09:43
which means we currently on jvm hit any "close enough" handler 09:44
...or implementation details are so different between the backends that looking at moar is actually harmful, of course :| 09:48
m: say 3403.base(2).comb.map({ (state $x = -1)++; $_ > 0 ?? 2 ** ($_ * $x) !! 0}) # and how are this meaningful categories for a handler..? 10:00
camelia rakudo-moar d789da: OUTPUT«(1 2 0 8 0 32 0 0 256 0 1024 2048)␤»
psch 'cause afaict we don't have handlers with id 2, 1024, or 2048
or rather categories with those values 10:01
and the combination also seems weird 10:02
CAT_CATCH, unknown, CAT_REDO (without LAST, although we always install them together..?), and CAT_PROCEED and two more unknowns... 10:03
err, s/LAST/NEXT/
oh, except that's not a handlerInfo, carry on >_> 10:06
stmuk I've put the preXmas stars under rakudo.org/downloads/star/archive/ 10:18
Zoffix m: A: for ^1 { B: do for ^1 { say "in B"; last A } }; 10:27
camelia rakudo-moar d789da: OUTPUT«in B␤»
Zoffix doesn't see anything wrong with JVM version. 10:28
psch r: A: for ^1 { B: for ^1 { say "in B"; last A } } 10:33
camelia rakudo-moar d789da: OUTPUT«in B␤»
( no output )
psch r: A: for ^1 { B: for ^1 { last A } } 10:34
camelia rakudo-jvm cd19db: OUTPUT«===SORRY!===␤labeled last without loop construct␤»
( no output )
Zoffix ah
psch curiously 10:35
r: A: for ^1 { FIRST {}; B: for ^1 { last A } }
camelia ( no output )
psch r: A: for ^1 { FIRST {}; B: for ^1 { say "in B"; last A } }
camelia rakudo-jvm cd19db, rakudo-moar d789da: OUTPUT«in B␤»
psch i've tried looking at &sequential-map and what it changes, but nothing that seems different changes the non-phasery version into a working one 10:55
and, well, my stumbling around in ExceptionHandling also seems rather fruitless
tbrowder nqp-m: my @arr := <0 1 2 3>; for @arr -> $a, $b {say("a = $a, b = $b")} 11:08
camelia nqp-moarvm: OUTPUT«a = 0, b = 1␤a = 2, b = 3␤»
stmuk rakudo.org/2016/07/22/announce-raku...e-2016-07/ 11:11
[Coke] reminder: anyone who has weird permission issues on perl.org, please open a ticket with the admins. 12:04
email address at the bottom of the page on the site.
er, on rt.perl.org, specifically
unmatched} Who's got the keys for @rakudoperl twitter account? Looks like it's not been updated for ages: twitter.com/rakudoperl 12:18
RabidGravy hey I didn't even know it existed 12:21
moritz unmatched}: try asking pmichaud++ 12:31
dalek ar: 7f1d752 | (Zoffix Znet)++ | tools/star/release-guide.pod:
Include asking users to advertise the release
 12:38
ar: e74732c | lizmat++ | tools/star/release-guide.pod:
Merge pull request #73 from zoffixznet/patch-1

Include asking users to advertise the release
unmatched} .ask pmichaud is there a possibility of adding several trusted users to twitter.com/rakudoperl Twitter account, so we could post announcements such us weeklies and R* releases? 12:53
yoleaux2 unmatched}: I'll pass your message to pmichaud.
nine This sounds like an intriguing idea: github.com/niner/Inline-Perl5/issues/69 13:05
Any thoughts on it?
psch the twigil feels somewhat misplaced 13:07
lizmat nine: looks to me something that a sufficiently smart AT-KEY should be able to handle lazily ?
psch aside from that it looks perfectly fine
perlpilot nine: make it so :) 13:08
psch as in, i'm not sure how you get %*PERL5 in time during module import
well, unless simply exporting dynamics works...
nine psch: yes, that's the unknown part
psch m: module Foo { my %*BAR is export = baz => "quux" }; import Foo; say %*BAR.perl 13:09
camelia rakudo-moar d789da: OUTPUT«===SORRY!=== Error while compiling <tmp>␤Can't apply trait 'is export' on a my scoped variable. Only our scoped variables are supported.␤at <tmp>:1␤------> module Foo { my %*BAR is export⏏ = baz => "quux" }; import Foo; say %*BA␤ …»
psch m: module Foo { our %*BAR is export = baz => "quux" }; import Foo; say %*BAR.perl
camelia rakudo-moar d789da: OUTPUT«{}␤»
nine module Foo { our %BAR is export = baz => "quux" }; import Foo; say %BAR.perl 13:10
m: module Foo { our %BAR is export = baz => "quux" }; import Foo; say %BAR.perl
camelia rakudo-moar d789da: OUTPUT«{:baz("quux")}␤»
nine But that's just not as pretty
lizmat m: module Foo { our %BAR is export is dynamic = baz => "quux" }; import Foo; say %BAR.perl # still dynamic
camelia rakudo-moar d789da: OUTPUT«{:baz("quux")}␤»
psch m: module Foo { my %BAR = baz => "quux"; sub term:<%*BAR> is export { %BAR }; }; import Foo; say %*BAR.perl 13:12
camelia rakudo-moar d789da: OUTPUT«{:baz("quux")}␤»
psch ...
nine Well, first things first. I can just start my adding a method to Inline::Python for accessing globals
psch: that's....wow
psch gets out
nine I don't know what to say :)
lizmat actually, why is it that we cannot export a lexically scoped variable ?
dalek kudo/nom: b89bf9a | lizmat++ | src/core/Hash.pm:
Make Hash.DELETE-KEY about 2x as fast

  - rewrite using nqp ops		 13:14
ar: 378fdbc | (Zoffix Znet)++ | tools/star/release-guide.pod:
Include Hacker News as potential places to advertise releases at
 13:17
ar: 0a51fe4 | (Steve Mynott)++ | tools/star/release-guide.pod:
Merge pull request #74 from zoffixznet/patch-1

Include Hacker News as potential places to advertise releases at
nine Oh, in Perl 5 the name of the variable is actually @ not $@ 13:32
stmuk . o ( should we publish on 4chan if we include hackernews ) 13:40
jnthn goodness that gradle container build spews some output 13:41
uh, ww
nine m: module Foo { our $s = 1; our %h := class Foo is Associative { method AT-KEY($name) { "foo" } }.new; }; $Foo::s.say; %Foo::h<bar>.say 14:07
camelia rakudo-moar b89bf9: OUTPUT«1␤(Any)␤»
nine What am I doing wrong there?
travis-ci Rakudo build errored. Elizabeth Mattijsen 'Make Hash.DELETE-KEY about 2x as fast 14:10
travis-ci.org/rakudo/rakudo/builds/146641338 github.com/rakudo/rakudo/compare/d...9bf9af76db
jnthn m: module Foo { our $s = 1; our %h := class Foo is Associative { method AT-KEY($name) { "foo" } }.new; }; say %Foo::h.WHAT
camelia rakudo-moar b89bf9: OUTPUT«(Hash)␤»
jnthn Curious... 14:11
lizmat m: module Foo { our $s = 1; our %h := class Foo is Associative { method AT-KEY($name) { "foo" } }.new; }; $Foo::s.say; %Foo::h.AT-KEY("bar").say 14:13
camelia rakudo-moar b89bf9: OUTPUT«1␤(Any)␤»
lizmat m: module Foo { our $s = 1; our %h := class Foo is Associative { method AT-KEY($name) { "foo" } }.new; }; $Foo::s.say; %Foo::h.Foo::AT-KEY("bar").say 14:14
camelia rakudo-moar b89bf9: OUTPUT«1␤Cannot dispatch to method AT-KEY on Foo because it is not inherited or done by Hash␤ in block <unit> at <tmp> line 1␤␤»
lizmat m: module Foo { our $s = 1; our %h := class Foo { method AT-KEY($name) { "foo" } }.new; }; $Foo::s.say; %Foo::h.Foo::AT-KEY("bar").say
camelia rakudo-moar b89bf9: OUTPUT«Type check failed in binding; expected Associative but got Foo::Foo (Foo::Foo.new)␤ in block <unit> at <tmp> line 1␤␤»
lizmat m: module Foo { our $s = 1; our %h := class Foo is Hash is Associative { method AT-KEY($name) { "foo" } }.new; }; $Foo::s.say; %Foo::h.Foo::AT-KEY("bar").say 14:15
camelia rakudo-moar b89bf9: OUTPUT«1␤Cannot dispatch to method AT-KEY on Foo because it is not inherited or done by Hash␤ in block <unit> at <tmp> line 1␤␤»
lizmat m: module Foo { our $s = 1; our %h := class Foo is Hash is Associative { method AT-KEY($name) { "foo" } }.new; }; $Foo::s.say; Foo.new.AT-KEY("bar").say
camelia rakudo-moar b89bf9: OUTPUT«1␤You cannot create an instance of this type (Foo)␤ in block <unit> at <tmp> line 1␤␤»
lizmat m: module Foo { our $s = 1; our %h := class Foo is Associative { method AT-KEY($name) { "foo" } }.new; }; $Foo::s.say; Foo.new.AT-KEY("bar").say 14:16
camelia rakudo-moar b89bf9: OUTPUT«1␤You cannot create an instance of this type (Foo)␤ in block <unit> at <tmp> line 1␤␤»
lizmat m: module Foo { our $s = 1; class Foo is Associative { method AT-KEY($name) { "foo" } }; our %h := Foo.new; }; $Foo::s.say; Foo.new.AT-KEY("bar").say 14:17
camelia rakudo-moar b89bf9: OUTPUT«1␤You cannot create an instance of this type (Foo)␤ in block <unit> at <tmp> line 1␤␤»
lizmat hmmmm
afk& 14:20
nine m: module Bar { our $s = 1; class Foo is Associative { method AT-KEY($name) { "foo" } }; our %h := Foo.new; }; $Foo::s.say; Foo.new.AT-KEY("bar").say 14:23
camelia rakudo-moar b89bf9: OUTPUT«===SORRY!=== Error while compiling <tmp>␤Undeclared name:␤ Foo used at line 1␤␤»
nine m: module Bar { our $s = 1; class Foo is Associative { method AT-KEY($name) { "foo" } }; our %h := Foo.new; }; $Foo::s.say; Bar::Foo.new.AT-KEY("bar").say
camelia rakudo-moar b89bf9: OUTPUT«(Any)␤foo␤»
nine m: module Bar { our $s = 1; class Foo is Associative { method AT-KEY($name) { "foo" } }; our %h := Foo.new; }; $Foo::s.say; Bar::Foo.new<bar>.say 14:24
camelia rakudo-moar b89bf9: OUTPUT«(Any)␤foo␤»
gfldex please review github.com/perl6/doc/commit/fbb443ed73 as it sounds utterly wrong 14:26
[TuxCM] This is Rakudo version 2016.07.1-29-gb89bf9a built on MoarVM version 2016.07-4-g236058a
test 15.154
test-t 8.061
csv-parser 16.160
lizmat nine: shouldn't that be "does Associative" ? 14:37
really afk&
[Coke] gfldex: "its default value", not "it's"
gfldex that's not what i meant with "sounds utterly wrong" 14:38
[Coke] I know
I think the focus should be on "Nil is a special sentinel value" rather than "Nil breaks things", but I don't think your statements there are incorrect. 14:39
also the section immediately before that already says that if you assign Nil you get the default back. 14:40
jnthn Note that Nil is just a more general kind of Failure 14:42
m: say Failure ~~ Nil
camelia rakudo-moar b89bf9: OUTPUT«True␤»
jnthn Nil is the absence of a value. Failure is the absence of a value together with a reason for said absence. 14:43
[Coke] I would put any thoughts about Nil/Failure into a separate section, rather than putting them in every spot where you could return a Nil.
unmatched} jnthn: so then assigning a Nil to a :D variable should be acceptable? There's a ticket for that: rt.perl.org/Ticket/Display.html?id...et-history 14:45
gfldex jnthn: I do understand what Nil is for. But I don't agree with initialisers that are return values of routines producing undefined values in definedness constrained containers.
m: my Int:D $i = Nil; 14:46
camelia ( no output )
gfldex m: my Int:D $i = Nil; say $i.defined;
camelia rakudo-moar b89bf9: OUTPUT«False␤»
jnthn Umm...
I don't know. You'd have to ask TimToady.
I didn't really want Failure or Nil to be able to sneak through *anywhere*. :P 14:47
I suspect on :D things they should not.
On unconstrained things they certainly can.
unmatched} .ask TimToady would you offer comments on whether my Int:D $i = Nil; should be accepted? There's a ticket for it: rt.perl.org/Ticket/Display.html?id...et-history and discussion: irclog.perlgeek.de/perl6-dev/2016-0...i_12889080
yoleaux2 unmatched}: I'll pass your message to TimToady.
jnthn (e.g. without :D) 14:48
gfldex m: my Str:D $f = Str:D; $f.say 14:49
camelia rakudo-moar b89bf9: OUTPUT«Type check failed in assignment to $f; expected Str:D but got Str:D␤ in block <unit> at <tmp> line 1␤␤»
gfldex jnthn: also ^^^
jnthn Yeah, that one is in the "should not work" direction :)
The error is interesting though correct :) 14:50
In fact that one's clear cut correct
'cus it doesn't involve a Nil or Failure :)
We could maybe when the value is a type object put the words type object on the end of the error 14:51
"expected Str:D but got Str:D type object"
gfldex that would help
jnthn Should be an easy patch if anyone fancies it :) 14:54
[Coke] is bummed that he's writing this proof of concept in perl 5. :| 14:55
jnthn Phew, $dayjob duties done for the week. Rest time. :) 14:58
[Coke] switches to p6, feeling better about life. 15:14
unmatched} :/ 15:18
Ah, I think Linoide killed off the box to do the Xen vuln update 15:19
nine lizmat: yes it should, but doesn't make a difference :/ 15:24
dalek p: a98d5f5 | (Pawel Murias)++ | src/vm/js/ (2 files):
[js] Remove dead code.
 17:30
p: 3eb8bd8 | (Pawel Murias)++ | src/vm/js/ (5 files):
[js] Inject unwrapped methods into objects.
unmatched} eyes mst 17:38
TrollTracker, eh? :)
gfldex unmatched}: could you run `panda installdeps .` on docs please? 17:56
unmatched} done. 18:00
gfldex thanks. i'm watching the buildlog
not sharing the precompiled .pods on docs is intentionally right now. If the build works I will get that fixed next. 18:18
timotimo: could you copy the updated util/update-and-sync to wherever it needs to go please? 18:37
timotimo what repo is that in? docs? 18:38
gfldex yes
timotimo let me have a look
it doesn't need to be moved anywhere 18:40
gfldex does it call make instead of perl htmlify.p6 ? 18:41
timotimo no
it runs ./doc/util/update-and-sync 18:42
gfldex i mean what does update-and-sync call?
timotimo make html || sync-build-log 18:43
./util/sync
gfldex the build log seams to disagree docs.perl6.org/build-log/build-201...0+0000.log 18:44
timotimo i don't get it. 19:03
gfldex timotimo: now make is called and it fails. Looks like panda installed pod2onepage into some /bin but that /bin is not in PATH 19:10
timotimo urgh 19:11
but it should accept your changes via git directly
gfldex timotimo: it should
but i don't know where panda puts stuff on the host
`find / -name 'pod2onepage'` 19:12
it should be `export PATH="$PATH:/home/dex/.perl6/bin/"` 19:13
o.0
export PATH="$PATH:$HOME/.perl6/bin/"
also any changed do /util/update-and-sync must show up in the build that follows the git pull that pulls the changed file 19:17
timotimo ah 19:19
right, that makes sense
we could put a check for that into the update-and-sync, though
git show --shortstat | grep update-and-sync && exec $0 or whatever 19:20
gfldex and to get that we would need 2 commits :->
we need to have that PATH set correctly in any case and that should solve the problem 19:21
timotimo it should have your home's bin in its path? 19:25
because ~/.perl6/bin doesn't exist in the doc.perl6.org user 19:26
gfldex in that case run `find / -name 'pod2onepage'`
that will tell you where panda put the file
timotimo /home/rakudobrew/rakudobrew/moar-2016.06/install/share/perl6/site/bin/pod2onepage
we need to source the global rakudobrew in the user's .profile or something 19:27
i have to destroy some dinner now :)
gfldex that means if we update star on docs, the PATH will be wrong again 19:28
timotimo ... i just said we need to source the rakudobrew; that would make PATH always right 19:29
gfldex m: sub f(){} say "{f(}"; say 'oi!'; 20:23
camelia rakudo-moar b89bf9: OUTPUT«===SORRY!=== Error while compiling <tmp>␤Strange text after block (missing semicolon or comma?)␤at <tmp>:1␤------> sub f(){}⏏ say "{f(}"; say 'oi!';␤ expecting any of:␤ infix␤ infix stopper␤ statement en…»
gfldex timotimo: travis wanted that path fixed too what seams to work 20:28
timotimo aaarrrgghhhh 20:34
yeah.
but that can be changed in the travis.yml only 20:35
or ... just add it to the path, whatever.
gfldex timotimo: could gist what `cat /home/rakudobrew/.rakudobrew-bash` says? 20:38
timotimo PATH=$PATH:/home/rakudobrew/rakudobrew/bin 20:41
timotimo takes a little walk 20:44
(yes, pokemon go. so sue me.)
gfldex timotimo: the workaround worked, the build is building 21:24
dalek kudo/nom: 7257b3e | lizmat++ | src/core/Hash.pm:
Make Hash[Any,Any].AT-KEY about 20% faster

  - rewritten using nqp ops
  - smarter handling of empty typed hashes		 21:57
kudo/nom: a37cd22 | lizmat++ | src/core/Bag.pm:
Make repeated calls on Bag.WHICH 5x as fast

  - by using nqp::attrinited rather than //=		 22:27
lizmat good night, #perl6-dev! 22:29
timotimo gfldex: i'm glad 22:41
travis-ci Rakudo build passed. Elizabeth Mattijsen 'Make Hash[Any,Any].AT-KEY about 20% faster 22:53
travis-ci.org/rakudo/rakudo/builds/146766567 github.com/rakudo/rakudo/compare/b...57b3e22d17
Rakudo build passed. Elizabeth Mattijsen 'Make repeated calls on Bag.WHICH 5x as fast 23:37
travis-ci.org/rakudo/rakudo/builds/146771969 github.com/rakudo/rakudo/compare/7...7cd221210b

Powered by Cro and the Raku® Programming Language.

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

Thank you!