»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg p6eval perl6: ... | irclog: irc.perl6.org/ | UTF-8 is our friend!
Set by sorear on 4 February 2011.
azawawi consider running farabi6 inside a directory with fewer files/directories... i.e. not $HOME 00:00
to test open file "autocomplete" feature
sorear: Hi. How safe is {Niecza --safe}? 00:02
00:03 cognominal left 00:04 sftp joined
azawawi remind me again, where is the rakudo perl 6 safe mode code? 00:08
nevermind, found it... github.com/moritz/try.rakudo.org/b.../p6safe.pl 00:09
sorear Niecza --safe is a quick hack and should never be used to protect a valuable system; it operates at far too high a level 00:10
properly sandboxing perl 6 should be done at the VM level or below (ideally seccomp) 00:11
tadzik azawawi: I don't think that's the way to do it
there is SAFE.setting for this these days
github.com/rakudo/rakudo/blob/nom/...FE.setting
which apparently is deprecated in favor of github.com/rakudo/rakudo/blob/nom/...ED.setting 00:12
azawawi :)
i think the wording 'restricted' is better than 'safe'
00:13 cognominal joined
azawawi sorear: thx 00:13
00:15 kshannon joined
azawawi tadzik: any idea how to use it? :) 00:15
00:15 benabik joined
azawawi hmm i will add $FARABI6_INSECURE which is by default disabled... 00:18
00:19 armitage81 joined 00:20 armitage81 left, armitage81 joined 00:21 snearch left
azawawi FARABI6_UNSAFE=1 farabi6 # enable unsafe mode... i.e. run local code 00:30
00:38 kda left 00:53 rindolf left 00:56 daniel-s_ is now known as daniel-s 01:00 tokuhiro_ joined
azawawi hmmm... http::server::simple and firefox = crash 01:02
i was wondering why farabi6 kept crashing randomly
01:04 bowtie left, DarthGandalf left 01:05 Chillance left 01:07 kcwu left 01:09 DarthGandalf joined
diakopter I had a dream that the Rakudo team implored me to "please stop finding/submitting [WEIRD] rakudobugs" 01:30
(it was an ego-boosting dream, I guess) 01:31
r: my $a = .HOW.new(); say $a.^methods 01:40
p6eval rakudo 64208d: OUTPUT«No such method 'gist' for invocant of type 'NQPRoutine'␤ in method gist at src/gen/CORE.setting:4875␤ in sub say at src/gen/CORE.setting:7304␤ in block at /tmp/zGdVxRn9kr:1␤␤»
diakopter boo. I wanna see ClassHOW methods 01:41
r: my $a = .HOW.new(); print $a.^methods
p6eval rakudo 64208d: OUTPUT«Nominal type check failed for parameter '$got'; expected Any but got NQPRoutine instead␤ in method REIFY at src/gen/CORE.setting:6091␤ in method reify at src/gen/CORE.setting:5247␤ in method gimme at src/gen/CORE.setting:5612␤ in method print at src/gen/CORE.se…
01:42 fgomez left, dayangkun joined
diakopter r: my $a = .HOW.new(); .($a) for $a.^methods 01:43
p6eval rakudo 64208d: OUTPUT«too few positional arguments: 1 passed, 2 (or more) expected␤ in any set_invoke_forwarder at src/gen/Metamodel.pm:2207␤ in block at /tmp/WcpjEgyZDO:1␤␤»
diakopter r: my $a = .HOW.new(); .($a, $a) for $a.^methods
p6eval rakudo 64208d: OUTPUT«too many positional arguments: 2 passed, 1 expected␤ in any archetypes at src/gen/Metamodel.pm:2203␤ in block at /tmp/id6bir1sfS:1␤␤»
diakopter set_invoke_forwarder and archetypes seem pretty leaky to me.
phenny: tell masak please see here VVDKVMLRSNCD in the logs
phenny diakopter: I'll pass that on when masak is around.
azawawi feather.perl6.nl:3030/ # Farabi6 with the new Run dialog and Open dialog... feedback appreciated... Firefox crashes HTTP::Server::Simple server... hence please use Chrome or something else... Have fun 01:47
farabi6 server now restarts itself... just in case :) 01:49
azawawi sleep & 01:54
01:54 azawawi left 01:55 fgomez joined 01:59 hypolin joined
skids gist.github.com/4014812 # workaround for crony roles used more than once 01:59
02:03 kcwu joined 02:06 chee joined 02:15 leont left 02:22 whiteknight left 02:28 sizz_ joined, sizz left 02:30 cognominal left 02:32 armitage81 left 02:37 tokuhiro_ left 02:55 araujo left 02:57 araujo joined 03:26 fgomez left, dayangkun left 03:34 dayangkun joined 03:35 dayangkun left 03:36 dayangkun joined 04:11 raiph joined
raiph phenny, tell azawawi Syntax Check (F6) seems to do nothing; Rakudo and Niecza give "ENV{FARABI6_UNSAFE} is not enabled"; If I click in text area when a menu (File, Build...) is dropped, that menu disappears and will no longer drop. (Have to refresh page to clear.) (From Chrome's about: "10.0.648.82 (75062) Built on Joli OS 1.2, running on Joli OS 1.2") 04:33
phenny raiph: I'll pass that on when azawawi is around.
raiph exit 04:34
heh. goodnight #perl6... 04:35
04:35 raiph left 04:44 havenn joined 05:10 plobsing left 05:23 telex left 05:25 telex joined 05:30 telex left 05:37 telex joined 05:52 fgomez joined 05:57 SamuraiJack joined 06:24 araujo left, kaleem joined 06:30 skids left 06:37 araujo joined 06:38 havenn left 06:50 azawawi joined 07:09 havenn joined 07:18 havenn left 07:34 bonsaikitten left 07:35 bonsaikitten joined 07:38 domidumont joined
azawawi hi 07:43
phenny azawawi: 04:33Z <raiph> tell azawawi Syntax Check (F6) seems to do nothing; Rakudo and Niecza give "ENV{FARABI6_UNSAFE} is not enabled"; If I click in text area when a menu (File, Build...) is dropped, that menu disappears and will no longer drop. (Have to refresh page to clear.) (From Chrome's about: "10.0.648.82 (75062) Built on Joli OS 1.2, running on Joli OS 1.2")
07:44 domidumont left
moritz \o 07:45
phenny moritz: 04 Nov 01:38Z <[Coke]> ask moritz you added S03-operators/andthen.t to rakudo's spectest - did you forget to push the test file?
moritz seems like, yes 07:46
azawawi phenny: tell raiph 1. very old chrome version, F6 at the moment does not do anything when there is no error. Try it with some errors in it. You can install Farabi6 to test it on your machine and set FARABI6_UNSAFE=1
phenny azawawi: I'll pass that on when raiph is around.
azawawi moritz: hi
moritz unfortunately I just moved to a different flat, and won't have internet access until 2012-11-12 or so
hi azawawi 07:47
azawawi Niecza wins on ease of installation (provided you install mono first... a lot of time wasted there) 07:51
Rakudo wins on completeness and modules support.. You lose a lot of time building a rakudo star though 07:52
std on the near perfect error message support. make takes a bit of time though. TimToady++
bonsaikitten azawawi: both are trivial to install
azawawi im thinking from a user's point of view 07:53
how much time is spent doing task X before using starting with real programming perl6 task
moritz we really want to do be able to install it with 'apt-get install rakudo' 07:54
and on windows there's an .msi
bonsaikitten "emerge niecza rakudo" <-- problem solved ;)
azawawi installation time... apt-get solves it... emerge = build time 07:55
07:55 hoelzro|away is now known as hoelzro
bonsaikitten azawawi: only if you don't have binpkgs already (and apt-get is really bad with some common issues) 07:55
azawawi bonsaikitten: true but a binary package is really handy... 07:56
bonsaikitten azawawi: takes me 8 minutes to build *and* run tests 07:57
azawawi moritz: how do i get the p6doc index? i need it for help index autocompletion and F1 context help
bonsaikitten azawawi: that's a tolerable tradeoff if I get a recent version out of it
azawawi programmers are lazy... :) 07:58
azawawi is building rakudo on feather3 for farabi6 auto deployment 07:59
moritz azawawi: the best option is probably to hack htmlify to dump the index as .json or so
jnthn morninä
.oO( this is not the keyboard layout you are looking for )
 08:00
moritz \o 08:01
azawawi moritz: github.com/perl6/doc/blob/master/htmlify.pl , right?
moritz right
jnthn moritz: oh noes no internets?!
Hope you survive :) 08:02
azawawi he did it before if i remember correctly :)
08:04 fhelmberger joined
sorear o/ jnthn 08:07
azawawi is there a plan/grant to upgrade feather.perl6.nl?
jnthn hi, sorear 08:08
sorear feather is a commons sort of thing
a large number of people are authorized to update it
so, nobody does
08:10 havenn joined
azawawi sorear: hardware that is... 08:10
moritz: what happened to try.rakudo.org? maintenance problems? 08:13
moritz jnthn: only internets at $work
azawawi: yes; somhow the server hangs up after a few days 08:14
azawawi: and too few people know how to restart it
08:14 havenn left
sorear is farabi a js app? 08:15
moritz: phew, good to know you won't be going completely dark :D 08:16
jnthn Seems it's about time to start teaching here. bbl :)
azawawi takes a look at try.rakudo.org 08:18
we have great perl6 evalbot here but nothing exposed to the outside world 08:19
s/evalbot/eval bots/
sorear azawawi: what is farabi implemented in? 08:20
azawawi sorear: farabi6 => perl6 + js frontend
sorear: farabi (perl 5): perl + mojo + js frontend
sorear azawawi: I wonder how feasible it would be to run niecza on the client using Silverlight and broker communications through js 08:21
azawawi sorear: interesting
sorear: i have done a lot as3 flash programming but silverlight nothing atm 08:22
sorear: s/a lot/a lot of/
08:24 mhasch joined, domidumont joined 08:55 sorenso joined, PZt left 08:58 domidumont left 08:59 domidumont joined 09:05 sorenso left 09:06 sorenso joined 09:10 havenn joined 09:11 Rix left 09:12 Rix joined 09:15 havenn left 09:21 dayangkun left 09:28 flightrecorder joined 09:33 imarcusthis left 09:36 imarcusthis joined 09:48 kresike joined
kresike hello all you happy perl6 people 09:48
09:48 circlepuller left
tadzik hey kresike 09:49
kresike tadzik, o/
09:54 circlepuller joined 09:56 azawawi left 10:02 azawawi joined 10:03 leont joined 10:07 leont left 10:08 hypolin left
masak morning, #perl6 10:11
phenny masak: 01:43Z <diakopter> tell masak please see here VVDKVMLRSNCD in the logs
10:11 havenn joined
masak (it's OK, I read all of the logs) :) 10:11
10:11 fhelmberger left
azawawi star: "Howdy".say; 10:15
p6eval star 2012.10: OUTPUT«(timeout)»
10:16 havenn left
azawawi star: "Test".say; 10:18
p6eval star 2012.10: OUTPUT«Test␤»
10:22 sorenso left 10:30 hoelzro left 10:31 hoelzro joined
masak backlogs 10:31
10:31 cognominal joined
moritz star: "Howdy".say; 10:34
p6eval star 2012.10: OUTPUT«Howdy␤»
masak r: my $a = .HOW.new(); .($a) for $a.^methods
p6eval rakudo 64208d: OUTPUT«PackFile_unpack: Buffer length 0 is shorter than PACKFILE_HEADER_BYTES 18.␤current instr.: '' pc 4075 (src/gen/perl6-moduleloader.pir:1632) (src/Perl6/ModuleLoader.pm:256)␤called from Sub '' pc 4235 (src/gen/perl6-moduleloader.pir:1692) (src/Perl6/ModuleLoader.pm:2…
masak r: my $a = .HOW.new(); .($a, $a) for $a.^methods
p6eval rakudo 64208d: OUTPUT«PackFile_unpack: Buffer length 0 is shorter than PACKFILE_HEADER_BYTES 18.␤current instr.: '' pc 4075 (src/gen/perl6-moduleloader.pir:1632) (src/Perl6/ModuleLoader.pm:256)␤called from Sub '' pc 4235 (src/gen/perl6-moduleloader.pir:1692) (src/Perl6/ModuleLoader.pm:2…
masak <diakopter> set_invoke_forwarder and archetypes seem pretty leaky to me.
no, I think that's what the methods you're calling might be called. 10:35
oh, the above executions failed. wonder why.
10:36 sorenso joined 10:39 PZt joined, azawawi left 10:49 azawawi joined 10:53 am0c joined
azawawi feather.perl6.nl:8080 # F6 syntax check fixed, Run dialog is more cool :) 10:55
masak finds the page en.wikipedia.org/wiki/Egbert_B._Gebstadter hilarious 11:01
11:01 gootle left 11:11 havenn joined 11:17 havenn left 11:19 hoelzro is now known as hoelzro|away, arlinius left 11:36 SmokeMachine left, fgomez left 11:39 SmokeMachine joined 11:48 xdbr joined 11:50 fhelmberger joined 11:52 chee left 11:55 chee joined 11:59 hoelzro|away is now known as hoelzro 12:12 havenn joined 12:17 havenn left 12:22 tokuhiro_ joined 12:34 azawawi left 12:35 bluescreen10 joined, Psyche^ joined 12:38 Patterner left, Psyche^ is now known as Patterner 12:43 brrt joined 12:52 cognominal left 12:53 cognominal joined
felher masak: thanks for the pointer to your discussion with TimToady. It answered exactly my question :) 12:58
masak nice. glad to hear it.
it's part of an intermittent discussion that stretches over several months. 12:59
13:00 ponbiki left 13:01 ponbiki joined, ponbiki is now known as Guest20434
felher masak: btw: I wasn't sure how to translate your 'I don't think it's at all unreasonable'. Does this mean "It's not completely unreasonable" (my first guess) or "It's totally reasonable" ? (not that it would make a difference. I'm just curious ) 13:07
13:07 grondilu joined
felher this certainly has to do with my english rather than with your sentence :) 13:08
felher afk for a while 13:12
masak felher: "I think the reasonableness is very far from the setting 'unreasonable' (i.e. quite reasonable)"
r: macro capture-output($code) { my $result; my $out; quasi { $result = ""; $out = $*OUT; $*OUT = class { method print(*@args) { $result ~= @args.join } }; {{{$code}}}; $*OUT = $out; $result } }; my $s = capture-output say "OH HAI"; print $s
p6eval rakudo 64208d: OUTPUT«OH HAI␤»
masak this is way cool.
moritz masak++ 13:13
13:13 havenn joined
masak I had written this example off as a case of "Better Written As Sub"... but now that I implement it, I see that it's not. just like with the logging example, the 'say "OH HAI"' is bare. it can't be bare with subs. 13:13
now what I *didn't* show is how I had to try the above in 10 different ways and ran into three different macro-related bugs on the way. :P
arnsholt =D 13:14
masak the first one is related to using 'temp' inside of a macro.
moritz it's like November 2009 again, but with macros this time :-)
arnsholt And in theory this shold work as capture-output { ... } right?
masak arnsholt: if you do it as a sub, you'd do it with the curlies, yes. 13:15
moritz masak: it's basically a custom phaser
with macros
r: BEGIN say 42;
p6eval rakudo 64208d: OUTPUT«42␤»
masak r: macro foo { quasi { temp $*OUT } }; foo 13:16
p6eval rakudo 64208d: OUTPUT«Null PMC access in elements()␤ in sub prefix:<temp> at src/gen/CORE.setting:11961␤ in at /tmp/XCSZksD14I:1␤ in block at /tmp/XCSZksD14I:1␤␤»
masak submits rakudobug
arnsholt masak: How about macro with more than a single statement? 13:17
13:17 havenn left
masak arnsholt: we can only pass terms to macros at present. 13:17
arnsholt Right =)
brrt i have some Very Interesting Behavior, although it is correct afaik
masak arnsholt: however, if you put parens or curlies around your statements...
brrt nr: sub foo { return { say "OH HAI"; return 42; }; }; foo().(); 13:18
p6eval niecza v22-16-g4c016f5: OUTPUT«OH HAI␤Unhandled exception: Illegal control operator: return␤ at /tmp/hVv4WvilD8 line 1 (ANON @ 6) ␤ at /tmp/hVv4WvilD8 line 1 (mainline @ 5) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4215 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line …
..rakudo 64208d: OUTPUT«OH HAI␤Attempt to return outside of any Routine␤ in block at src/gen/CORE.setting:466␤ in block at /tmp/0esC6gtKK2:1␤␤»
masak brrt: yes, it's correct. 13:19
brrt: in short, bare blocks don't bind 'return'.
so your inner return ends up binding to 'foo'. 13:20
13:20 GlitchMr joined
masak which has already returned by the time the inner block runs. so the error message is correct. 13:20
moritz and by the time you run the inner block, &foo already has returned
moritz too slow
brrt yes, now, what happens if i call it in another dynamic context?
moritz r: sub foo { return { say "OH HAI"; return 42; }; }; try foo().(); say $/.WHAT 13:21
p6eval rakudo 64208d: OUTPUT«OH HAI␤Any()␤»
moritz r: sub foo { return { say "OH HAI"; return 42; }; }; try foo().(); say $!.WHAT
p6eval rakudo 64208d: OUTPUT«OH HAI␤X::ControlFlow::Return()␤»
13:21 bowtie joined
moritz r: sub foo { return { say "OH HAI"; return 42; }; }; sub bar { try foo().(); say $!.WHAT }; bar() 13:21
p6eval rakudo 64208d: OUTPUT«OH HAI␤X::ControlFlow::Return()␤»
moritz brrt: same thing
masak r: sub foo { bar({ return 42 }); return 3 }; sub bar(&c) { c }; say foo
brrt r: sub foo { return { say "OH HAI"; return 42; } }; sub bar { foo().(); }
p6eval rakudo 64208d: OUTPUT«42␤»
rakudo 64208d: ( no output )
masak I believe the first output is mine.
and it prints what I expected. 13:22
brrt it is expected, but madness
masak that is, you can *return* from a subroutine that's above your currently executing subroutine in the call stack.
it's not madness. it's just regular exceptions.
moritz I think that's a bug.
masak I don't.
moritz return is supposed to be tied to the outer lexical routine 13:23
masak it is.
that's why you see this behavior.
the 'return 42' in &foo binds to &foo.
flussence oh, I did a hackish $*OUT capture thing a long time ago, if you see any useful ideas there feel free to steal 'em - gist.github.com/1606837
moritz masak: oh, I was confused
flussence (I was thinking, maybe Cat should have IO methods on it...)
masak moritz: for a little while there I thought we were disagreeing on something. then normality returned. :P 13:24
brrt masak: returns and exceptions, that weird 13:25
GlitchMr This channel is always active ;-)
brrt exceptions are weird in their own right, but return-as-an-exception
its easy in continuations, i guess
masak brrt: how would you implement it? :)
brrt well, i would create it by making the inner block a lambda with a fixed continuation - that of foo, determined at run time 13:26
so it would do the same thing
masak brrt: or rather, if you accept that exceptions allow you to effectively return from several routines at once, why don't you like it when &return does?
brrt masak, mostly because return means a simpler thing in the c world 13:27
masak "Toto, I don't think we're in C-land anymore"
brrt this being perl6 i should not at all be surprised by finding there is more than one way to do it
masak r: macro foo { quasi { my &bar = { say "OH HAI" }; bar } }; foo 13:36
p6eval rakudo 64208d: OUTPUT«===SORRY!===␤CHECK FAILED:␤Undefined routine '&bar' called (line 1)␤»
masak r: macro foo { quasi { my &bar = { say "OH HAI" }; &bar() } }; foo
p6eval rakudo 64208d: OUTPUT«Cannot assign into a PMCNULL container␤ in at /tmp/GfBF7Zue77:1␤ in block at /tmp/GfBF7Zue77:1␤␤»
masak submits rakudobug
none of this is surprising, of course, since we already know variable declarations don't work properly in macros. 13:37
er, in quasi blocks, I mean.
r: macro foo { quasi { my $bar = "OH HAI"; say $bar } }; foo 13:38
p6eval rakudo 64208d: OUTPUT«Cannot assign into a PMCNULL container␤ in at /tmp/euDdsux_9o:1␤ in block at /tmp/euDdsux_9o:1␤␤»
masak r: macro foo { quasi { my $bar; say $bar } }; foo
p6eval rakudo 64208d: OUTPUT«Mu()␤»
masak sorry; declarations work. assignments somehow don't.
assignments to locally declared variables, that is.
moritz remember --ll-exception when debugging :-) 13:39
masak thanks for le reminder :) I need it. 13:40
r: macro foo { my $bar = "OH HAI"; quasi { say $bar } }; foo
p6eval rakudo 64208d: OUTPUT«OH HAI␤»
masak r: macro foo { my &bar = { say "OH HAI" }; quasi { &bar() } }; foo
p6eval rakudo 64208d: OUTPUT«OH HAI␤»
masak r: macro foo { my &bar = { say "OH HAI" }; quasi { bar } }; foo
p6eval rakudo 64208d: OUTPUT«===SORRY!===␤CHECK FAILED:␤Undefined routine '&bar' called (line 1)␤»
masak huh! 13:41
I wonder if that's a different failure mode, or the same one.
moritz r: macro foo { sub bar { say "OH HAI" }; quasi { &bar() } }; foo 13:42
p6eval rakudo 64208d: OUTPUT«OH HAI␤»
masak r: macro foo { sub bar { say "OH HAI" }; quasi { bar } }; foo 13:43
p6eval rakudo 64208d: OUTPUT«===SORRY!===␤CHECK FAILED:␤Undefined routine '&bar' called (line 1)␤»
masak well, it's consistent, at least.
[Coke] I made very little (but some!) progress on muddle this weekend, taking the little bit of functionality that was there and putting a slightly nicer skin on it 13:45
feather.perl6.nl:2653/
masak nice. [Coke]++
flussence [Coke]++ 13:46
[Coke] Hopefully by the end of the week, it'll be functional enough that we can add a make target to rakudo or niecza to submit reports.
Thanks to me from a year ago, since dropping in twitter bootstrap is a pretty easy upgrade. :)
flussence how fast is it? one of the big problems with smolder was that it collapses if you look at it funny... 13:47
brrt basically, 'take' both yields its value and returns, right? 13:49
so my $x = take 3; sets $x to 3? 13:50
[Coke] It's a mojolicious app - right now I'm running it under morbo, but we could shove it into a more robust container.
brrt r: my $x = take 3; say $x;
p6eval rakudo 64208d: OUTPUT«take without gather␤ in block at src/gen/CORE.setting:429␤ in block at src/gen/CORE.setting:478␤ in block at /tmp/fBpNUYcX69:1␤␤»
flussence [Coke]: should be pretty fast then :)
[Coke] it's also using sqlite3, because that's zero install and I can run it anywhere. We can add another backend if there's a better choice on feather. (or wherever.)
13:53 cognominal left
moritz (starting simple)++ 13:54
13:55 bluescreen10 left
masak r: my @a = gather { say take 3 }; 13:57
p6eval rakudo 64208d: OUTPUT«0␤»
masak brrt: that's interesting.
flussence mojolicious (and other stuff like it) seem perfect for this sort of thing - replacing a flaky old web service with one that actually works. I've done it a few times at $dayjob already.
masak what does the 0 mean?
r: my @a = gather { say take 3; say take 42 };
p6eval rakudo 64208d: OUTPUT«0␤0␤»
masak n: my @a = gather { say take 3; say take 42 };
p6eval niecza v22-16-g4c016f5: OUTPUT«Potential difficulties:␤ @a is declared but not used at /tmp/YeVGB755T9 line 1:␤------> my ⏏@a = gather { say take 3; say take 42 };␤␤3␤42␤»
masak submits rakudobug 13:58
n: my @a = gather { say take 3; say take 42 }; say @a.perl
p6eval niecza v22-16-g4c016f5: OUTPUT«3␤42␤[3, 42].list␤»
flussence take returns 0?
masak n: my @ = gather { say take 3; say take 42 };
p6eval niecza v22-16-g4c016f5: OUTPUT«3␤42␤»
flussence (sometimes?)
masak flussence: in Rakudo, yes. in Niecza it gives back the value.
13:58 cognominal joined
masak which seems much more useful. 13:58
flussence agreed
tadzik I wonder if we can marry that and emmentaler somehow 13:59
(muddle)
brrt masak: returning the value is basically what is spec'-ed 14:00
masak brrt: any reference to that bit of spec?
(so I can include it in the RT ticket) 14:01
brrt s04-somewhere, will be a few minutes
masak S04:735. 14:02
brrt perlcabal.org/syn/S04.html#The_gath...ent_prefix 14:03
masak The value returned by the C<take> to the C<take>'s own context is that same
returned argument (which is ignored when the C<take> is in sink context).
brrt say is maybe seen as sink context?
flussence r: my @a = gather { say eager take 3 }; 14:05
p6eval rakudo 64208d: OUTPUT«0␤»
14:05 marloshouse_ joined
masak brrt: no. sink context is when you don't use the value from 'take' at all. 14:05
GlitchMr What eager take is supposed to do?
14:06 marloshouse left, marloshouse_ is now known as marloshouse
flussence not sink context 14:06
14:06 SamuraiJack left, PacoAir joined
brrt r: say sink 3; 14:07
p6eval rakudo 64208d: OUTPUT«Nil␤»
[Coke] I wonder if that's actually a Kitchen sink. it would explain a lot. 14:09
moritz maybe it's a kitten sink
brrt that association also came to my mind
perl6, comes with (kitchen) sink 14:10
14:11 bluescreen10 joined, tokuhiro_ left 14:13 havenn joined 14:14 cognominal left
masak std: KitchenSink 14:16
p6eval std 04216b1: OUTPUT«ok 00:00 41m␤»
masak STD comes with a kitchen sink.
std: GardenHose
p6eval std 04216b1: OUTPUT«===SORRY!===␤Undeclared name:␤ 'GardenHose' used at line 1␤Check failed␤FAILED 00:00 41m␤»
14:18 havenn left 14:28 kaleem left 14:32 plobsing joined 14:39 atrodo_ joined 14:40 kaare_ joined 14:42 atrodo left, atrodo_ is now known as atrodo, atrodo left, atrodo joined, mtk joined 14:43 sivoais left 14:46 kaleem joined 14:47 plobsing left 14:50 marloshouse left 14:54 arlinius joined 14:55 marloshouse joined 14:56 spider-mario joined 14:57 benabik left
felher masak: ah, okay, now i get it. Thanks :) 14:58
(About the reasonable-thingy)
masak also, your request for clarification was not at all unreasonable :P 15:01
15:04 telex left
felher masak: :D 15:04
15:06 telex joined 15:11 stopbit joined
felher oha, the capture-output macro is nice :) 15:12
15:14 havenn joined
masak :) 15:14
felher masak: do you have to save $*OUT explicitly? Couldn't one just write "my $*OUT = class { ... };" ? 15:16
flussence $*OUT is an IO object, not a string 15:17
masak felher: that might work too, yes.
15:17 brrt left, brrt joined
masak r: say "before"; { my $*OUT = class { method print(*@) {} }; say "inside" }; say "after" 15:18
p6eval rakudo 64208d: OUTPUT«before␤after␤»
masak yes, that works. but probably not in quasis yet, because there seems to be something wrong with quasi { my ... }
[Coke] wonders why these gists are gists and not blog posts.
felher masak: okay, nice :)
[Coke] (pointing at masak's recent blog post mentioning some gists)
15:18 skids joined 15:19 havenn left
masak [Coke]: because gists feel like they have a smaller "contract" to be kept up-to-date, or to explain everything to an uninitiated audience. 15:19
it's as if gists are mostly for me, whereas blog posts are mostly for other people.
n: my @sorted = 1, 2, 2, 2, 3, 4, 4, 5; my @uniq = gather for @sorted { state $previous = take $_; next if $_ === $previous; $previous = take $_ }; say @uniq 15:21
p6eval niecza v22-16-g4c016f5: OUTPUT«1 2 3 4 5␤»
masak \o/
[Coke] I feel like you're hiding things you should be showing people.
masak n: my @sorted = 1, 2, 2, 2, 3, 4, 4, 5; my @uniq = gather for @sorted { my $previous = START take $_; next if $_ === $previous; $previous = take $_ }; say @uniq
p6eval niecza v22-16-g4c016f5: OUTPUT«1 2 2 2 3 4 4 5␤»
masak [Coke]: well, I am taking pains to link to them from the blog posts. 15:22
brrt masak: does that work with rakudo?
[Coke] gist.github.com/abd360db840a0719da2f in particular.
masak [Coke]: and the gists really are less authoritative.
[Coke]: yes, that's perhaps the best of them. it's also quite recent.
[Coke]: it was written for YAPC::EU, for the proceedings handed out to all participants. 15:23
(which means that at the time, it did not make sense to blog it)
15:26 adu joined
[Coke] heh. I happened to read that gist first (the proceedings), and assumed the others were also in a similar vein. 15:27
masak most of them are just random musings and ruminations. 15:29
almost all are on the theme "wow, macros are like closures, except they're not closures OMG I'm so confused here's what's really going on..." 15:30
I'm still a little nervous to make the edit to Wikipedia mentioned in that gist. 15:34
maybe I should go talk to some Lisp people on IRC first.
15:35 havenn joined 15:37 sivoais joined
masak heh, #lisp channel redirects me to #scheme, because those are the hygienic people :P 15:39
moritz lol 15:40
TimToady
.oO(Uncommon Lisp)
arnsholt masak: Heehee. That's Common Lispers for ya. "Who needs hygiene anyways?!" ^_^ 15:42
masak I haven't been put on fire yet. people are very friendly on both channels. 15:44
hah, LeoNerd at #scheme makes a really good point about macro expansions and deparsing. 15:48
arnsholt Not bad. #lisp can get real snarky real quick
Oh?
masak if you ever try to deparse the expanded code, suddenly you have to handle variable collisions somehow. 15:49
my factoring of macro expansions means I don't have to handle it with gensymming and such. but during deparse, the problem would resurface.
essentially because some blocks have "impossible" OUTERs.
arnsholt Huh
[Coke] would we ever be able to deparse the fact that a macro was there? special flag on the AST? 15:50
or would it all just look like old school deparsed Java?
15:51 brrt left
masak no idea. 15:51
moritz doesn't understand "impossible" OUTERs 15:52
[Coke] (that is, not much like the original code with artificial var names.... oh - "like minified javascript" is probably better. :)
TimToady the main point of an AST is to be A...not necessarily to let you reconstruct the C 15:53
masak moritz: first-class functions allow you to separate a block from its OUTER at *runtime*. macros allow you to do so at *compile time*. which is why "surgery"/recontexting by fiddling with Parrot internals is needed.
moritz masak: erm, what? 15:54
a function never gets separated from its OUTER, afaict
15:54 thou joined
moritz sub a { sub b { } }; # &b's outer is always &a, no matter what, no? 15:55
masak I'm being a bit sloppy and unclear here. sorry.
I'm talking about moving the code itself. the OUTER link remains the same, yes. both with closures and macros.
but the code gets pulled from the block denoted by this OUTER. 15:56
moritz normal closures never move the code, right?
just macros
masak the code "moves" in the sense that I can return it from its original context and pass it somewhere else.
but you're right, in some sense it never leaves its original environment. 15:57
only the references to the closure do.
moritz you can move references to code+lexpad
masak right.
with macros, we *copy* the code itself, which is why it's harder to re-establish OUTERs. and it happens on two levers, once with the quasi and once for each unquote. 15:58
levels*
15:58 kaleem left
moritz I'm with you so far 15:58
how does that lead to "impossible" OUTERs?
masak well, suppose you deparse the code after the macro expansion happened. 15:59
there's no explicit mechanism to denote what OUTER a block has. it's just indicated by nesting, right?
moritz is that because you can copy two macros into the same block, with conflicting OUTERs?
ah
masak and yet code expanded by a macro will have an OUTER that points back into the macro body.
that's genuinely impossible. 16:00
in the sense that you can't write code like that directly in Perl 6.
moritz so, impossible in the sense that you can't reproduce the same thing with non-macro Perl 6 code
masak right.
moritz but, not impossible in the sense of "impossible to resolve deterministically" or so
(which I thought first you meant)
ok, got it
it's a bit like writing closure code without closures, where you have to do massive refactorings 16:01
flussence macros are interdimensional wormholes
moritz like carrying objects around that simulate the lexpads
masak no, it's possible to resolve deterministically.
the impossibility resurfaces when you try to deparse the code, though. I hadn't thought of that before.
skids WOuld you be able to key off the fact that you've reached an impossibility to have your decompiler say: oh I guess I should be looking to generate a macro now? 16:02
moritz it's probably not too hard to detect, but hard to turn into a macro (just my shallow guess) 16:03
masak ooh, an interesting objection at #scheme.
moritz: (re massive refactorings and carrying objects around) yes, exactly.
skids: it's certainly possible to detect when a variable doesn't resolve properly. like referring to $var without $var being in scope. 16:04
skids: but though this is a sure-fire way to detect macros *today*, it may not always be. also, it will not detect all macros.
16:06 sivoais left
skids But it will detect macros that are impossible top express in macro-free perl6, yes? In which case a patch macro that just supplies the appropriate OUTER might be generated. 16:06
moritz sounds possible 16:08
though not trivial
skids Well, I wouldn't expect a decompiler to be a trivial endeavor :-)
16:09 benabik joined
masak skids: well, a naive deparser might easily miss the re-bound OUTER links and just generate illegal code. 16:09
16:09 thou left 16:12 domidumont left, sivoais joined
kresike bye folks 16:14
16:14 kresike left, sorenso left
jnthn evening o/ 16:15
moritz \o jnthn, how was the teaching?
16:15 hoelzro is now known as hoelzro|away
jnthn Fine, but my throat hurts after talking lots. 16:15
masak r: macro foo { quasi { return 42 } }; sub bar { foo }; say bar 16:18
p6eval rakudo 64208d: OUTPUT«Attempt to return outside of any Routine␤ in block at src/gen/CORE.setting:466␤ in at /tmp/Pr3vSisyzw:1␤ in sub bar at /tmp/Pr3vSisyzw:1␤ in block at /tmp/Pr3vSisyzw:1␤␤»
masak aww
that should work, methinks.
jnthn I'm less than convinced.
Well
Hm 16:19
masak :)
jnthn I'm not sure of any way to do it that won't be an epic hack :/
masak remember, a quasi is a *deferred* context.
jnthn I remember. I also remember that return is lexical. Lexical things in quasis refer to...the scope the quasi is in.
quasi { $a } # refers to the $a where the quasi is
masak hm, troo. 16:20
so the 'return' here gets bound to the macro?
jnthn And the way we look up the return handler is just as an ordinary lexical.
I think it uses find_lex even, just like normal variables. 16:21
16:22 cognominal joined
moritz what does it find_lex? &ROUTINE? 16:23
jnthn RETURN 16:25
.lex "RETURN", $P107 # from the generated code
masak hygienic macros reading recommendations from #scheme: gist.github.com/4018113 16:26
as predicted, I am being taught a bit of humility by listening to the peeps on that channel. they haven't crushed my macros factoring yet, but their skepticism is... refreshing. 16:27
moritz I kinda hope your factoring survives
masak me too! 16:28
16:29 am0c left 16:33 flightrecorder left
[Coke] news.perlfoundation.org/2012/11/goo...de-in.html - TPF post about this year's google code in. 16:35
go, add sixtasks.
16:43 am0c joined 16:45 benabik left, [particle]1 is now known as [particle] 16:50 benabik joined
masak I hadn't seen this naggum post about hygiene in Lisps before: www.xach.com/naggum/articles/323678...m.net.html 16:54
(got it from the #lisp people, not from the #scheme people) :P
16:56 Chillance joined 16:57 brrt joined 17:05 MayDaniel joined 17:06 adu left 17:07 fgomez joined
arnsholt masak: Ah yes, Naggum is always an entertaining read. I'm not surprised it was the CL camp that gave you the link though 17:12
Naggum is their prophet, after all =)
masak indeed. 17:16
hence the ':P'.
arnsholt Heh. Indeed
cognominal rn: class A { sub hi { say 'hi' }; method hi { hi }; }; class B is A { method hi1 { hi } }; A.new.hi; B.new.hi; B.new.hi1 17:17
p6eval rakudo 64208d: OUTPUT«===SORRY!===␤CHECK FAILED:␤Undefined routine '&hi' called (line 1)␤»
..niecza v22-16-g4c016f5: OUTPUT«===SORRY!===␤␤Undeclared routine:␤ 'hi' used at line 1␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 1437 (die @ 5) ␤ at /home/p6eval/niecza/src/STD.pm6 line 1147 (P6.comp_unit @ 37) ␤ at /home/…
cognominal I am not sure it make any sense to inherit subs. The idea just crosses my mind :) 17:18
*crossed
masak cognominal: subs are not inherited. 17:19
cognominal I see
masak only methods do lookup along the MRO axis.
which is why in B &hi is not even visible.
cognominal indeed 17:20
masak well, that and the fact that subs are lexical by default. :)
rn: class A { our sub hi { say 'hi' } }; class B { our &hi; method hi1 { hi } }; B.new.hi1 17:21
p6eval niecza v22-16-g4c016f5: OUTPUT«Unhandled exception: Unable to resolve method postcircumfix:<( )> in type Any␤ at /tmp/3v0tIoAW6g line 1 (B.hi1 @ 4) ␤ at /tmp/3v0tIoAW6g line 1 (mainline @ 8) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4215 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/…
..rakudo 64208d: OUTPUT«No such method 'Any' for invocant of type 'Parcel'␤ in at src/gen/BOOTSTRAP.pm:845␤ in at src/gen/BOOTSTRAP.pm:839␤ in any at src/gen/BOOTSTRAP.pm:836␤ in method hi1 at /tmp/tFywyq7xl8:1␤ in block at /tmp/tFywyq7xl8:1␤␤»
masak inneresting.
oh, different classes => different packages.
so not too surprising.
cognominal I like stupid questions, that always lead to interesting thoughts 17:22
masak aye.
diakopter r: our macro foo() { }; foo()
p6eval rakudo 64208d: OUTPUT«===SORRY!===␤Error while compiling block : Error while compiling op call: Error while compiling block : Error while compiling op bind: First child of a 'bind' op must be a QAST::Var␤»
cognominal so I am not ashamed to make a fool of myself if I can learn in the process
diakopter masak: let me type that for you: /me submits rakudobug 17:23
cognominal one just need the right place to bounce them around. :)
diakopter r: temp macro foo() { }; foo() 17:24
p6eval rakudo 64208d: OUTPUT«===SORRY!===␤Method 'returns' not found for invocant of class 'NQPMu'␤»
masak diakopter: indeed.
masak submits rakudobug 17:25
diakopter++ diakopter++
cognominal also, with macros, masak has a keen eye for anything related to names in time and in space.
diakopter r: module foo; macro foo() { }; foo; foo(); # foo works; foo() doesn't 17:26
p6eval rakudo 64208d: OUTPUT«invoke() not implemented in class 'foo'␤ in block at /tmp/JEyQXBbqaU:1␤␤»
masak it does give a keener understanding of context and phases, yes.
[Coke] r: my $a = macro foo() {}; say $a; 17:27
p6eval rakudo 64208d: OUTPUT«===SORRY!===␤Error while compiling block : Error while compiling op call: Error while compiling block : Error while compiling op p6store: Unknown QAST node type NQPMu␤»
masak diakopter: both should work.
masak submits [Coke]'s rakudobug
[Coke] I have no idea what I expect $a to contain there, btw. :)
diakopter masak: both should invoke macro foo? or one should invoke macro foo and the other invoke class foo (like it's doing)? 17:28
cognominal :)
masak diakopter: both should invoke macro foo, IMO.
r: module foo; sub foo() { say "OH HAI" }; foo; foo()
p6eval rakudo 64208d: OUTPUT«invoke() not implemented in class 'foo'␤ in block at /tmp/n8jWXMP4lX:1␤␤»
masak hm. 17:29
same there; really unrelated to macros, I think.
diakopter oh
masak [Coke]: a Macro (which is a Routine)
masak submits diakopter's rakudobug
good rakudobug weather tonight :)
that's four in five minutes. 17:30
diakopter r: my $*a = macro { } 17:31
p6eval rakudo 64208d: OUTPUT«===SORRY!===␤Error while compiling block : Error while compiling op call: Error while compiling block : Error while compiling op p6store: Unknown QAST node type NQPMu␤»
[Coke] thinks that's the same as mine. 17:32
masak yes; I'll report them together.
[Coke] pugs: say 3;
p6eval pugs: OUTPUT«3␤»
[Coke] Any haskell folks, Pugs is no longer building for me in a fresh checkout on feather.
jnthn module foo; sub foo() { say "OH HAI" }; foo; foo() 17:33
In this one, foo() parses as a coercion.
[Coke] rakudofolks: github.com/coke/perl6-roast-data/b...pass_rates - line 5 vs. line 9 - pretty sure those are from "running via cron" and "running from my shell"; Any suggestions about what to add to my "run rakudo tests" scripts to avoid that are welcome. 17:34
17:34 robinsmidsrod left, robins joined
masak jnthn: oh! 17:34
masak adds that to the appropriate rakudobug
jnthn Anyway, it's right it doesn't call the sub. Just an LTA error I guess...though I ain't immediately sure what to do with it... 17:35
17:41 sizz joined, sizz_ left 17:42 brrt left 17:50 adu joined 17:51 adu left
sorear o/ 17:56
diakopter o 17:57
17:58 SamuraiJack joined
jnthn o/ 17:59
masak \o/
18:01 domidumont joined
[Coke] anyone use "Tapper" ? 18:06
it appears to be in vaguely the same market as smolder.
18:07 arlinius left, domidumont1 joined 18:11 domidumont left 18:13 fhelmberger left 18:18 kaare_ left, domidumont1 left 18:19 domidumont joined, snearch joined 18:20 flightrecorder joined 18:22 cognominal left 18:23 Milbourne joined 18:24 am0c left, kaare_ joined 18:34 robins is now known as robinsmidsrod 18:43 grondilu left 18:45 brrt joined 18:47 colomon left 18:48 xinming_ joined 18:51 xinming left 19:03 domidumont left 19:05 mtk left, mtk joined
masak n: constant @c := 1, -> $n { 2 * (2 * $n - 1) / ($n + 1) * @c[$n - 1] } ... *; say @c[$_] for 0..5 19:08
p6eval niecza v22-16-g4c016f5: OUTPUT«===SORRY!===␤␤Cannot use bind operator with this LHS at /tmp/B5bAObkC37 line 1:␤------> $n - 1) / ($n + 1) * @c[$n - 1] } ... *⏏; say @c[$_] for 0..5␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/li…
masak n: constant @c = 1, -> $n { 2 * (2 * $n - 1) / ($n + 1) * @c[$n - 1] } ... *; say @c[$_] for 0..5 19:09
p6eval niecza v22-16-g4c016f5: OUTPUT«1␤1␤1␤1␤1␤1␤»
masak hm, but that block isn't getting an index, of course...
TimToady which usually means you shouldn't be using ... for that 19:10
masak n: constant @c = 1, { state $i = 0; $i++; 2 * (2 * $i - 1) / ($i + 1) * $_ } ... *; say @c[$_] for 0..5
p6eval niecza v22-16-g4c016f5: OUTPUT«1␤1␤2␤5␤14␤42␤»
masak there we go.
r: constant @c = 1, { state $i = 0; $i++; 2 * (2 * $i - 1) / ($i + 1) * $_ } ... *; say @c[$_] for 0..5
p6eval rakudo 64208d: OUTPUT«1␤1␤2␤5␤14␤42␤»
masak woo 19:11
TimToady nr: constant @c = 1, { (state $i)++; 2 * (2 * $i - 1) / ($i + 1) * $_ } ... *; say @c[$_] for 0..5 19:12
p6eval rakudo 64208d, niecza v22-16-g4c016f5: OUTPUT«1␤1␤2␤5␤14␤42␤»
TimToady nr: constant @c = 1, { state $i++; 2 * (2 * $i - 1) / ($i + 1) * $_ } ... *; say @c[$_] for 0..5
p6eval niecza v22-16-g4c016f5: OUTPUT«1␤1␤2␤5␤14␤42␤»
..rakudo 64208d: OUTPUT«===SORRY!===␤Preceding context expects a term, but found infix + instead␤at /tmp/T18sEnSDMe:1␤»
masak rn: constant @c = 1, { (state $i)++; 2 * (2 * $i - 1) / ($i + 1) * $_ } ... *; sub b($l) { my $r = (^@c[$l]).roll; my $s; for reverse(^$l) Z ^$l -> $l1, $l2 { $s += @c[$l1] * @c[$l2]; return [~] "[", b($l1), "]", b($l2) if $r < $s }; return "" }; say b 5
p6eval niecza v22-16-g4c016f5: OUTPUT«Use of uninitialized value in numeric context␤ at /home/p6eval/niecza/lib/CORE.setting line 1295 (warn @ 5) ␤ at /home/p6eval/niecza/lib/CORE.setting line 294 (Any.Numeric @ 8) ␤ at <unknown> line 0 (ExitRunloop @ 0) ␤ at /tmp/mV4ZDw7cxC line 1 (b @ 15…
..rakudo 64208d: OUTPUT«[[][[]]][]␤»
masak a little update to the "balanced brackets" solution, now with 'constant'. 19:13
std: state $i++
p6eval std 04216b1: OUTPUT«ok 00:00 42m␤»
masak TimToady: nieczabug that 'state $i++' doesn't parse?
TimToady rakudobug
masak oh!
yes.
masak submits rakudobug
r: state $i++ 19:14
p6eval rakudo 64208d: OUTPUT«===SORRY!===␤Confused␤at /tmp/q3jGE_C9uJ:1␤»
jnthn Already TD'd... 19:15
*RT
masak oh!
jnthn And probably exact same issue as the sub foo { }() thing submitted the other day
masak looks for the previous ticket 19:16
jnthn Now if only somebody could find out why...
masak no, I can't find any previously submitted RT ticket for that. 19:17
19:17 Milbourne left
masak rn: my $a++ 19:17
p6eval rakudo 64208d: OUTPUT«===SORRY!===␤Confused␤at /tmp/txWl1n_0np:1␤»
..niecza v22-16-g4c016f5: OUTPUT«Potential difficulties:␤ $a is declared but not used at /tmp/K5JEon7nDZ line 1:␤------> my ⏏$a++␤␤»
masak maybe it's submitted for something like that instead.
jnthn isn't particularly disappointed Rakudo doesn't parse it :P
It visually chunks pretty awfully. 19:18
19:18 grondilu joined
grondilu in S29, S32/Scalars is mentionned but it doesn't exists, does it? 19:18
(see 'defined' and 'undefine') 19:21
19:21 SamuraiJack left
masak jnthn: same awful visual chunking with sub {}() 19:22
jnthn masak: Yeah 19:23
masak grondilu: right. neither does 'undefine' anymore.
grondilu rn: my %h; say "ok" unless %h<foo> :exists; 19:24
p6eval rakudo 64208d: OUTPUT«===SORRY!===␤Confused␤at /tmp/L2WIT6sx75:1␤»
..niecza v22-16-g4c016f5: OUTPUT«ok␤»
grondilu how do I test the existence of a hash entry?
(I mean, until the :exists adverb is implemented on rakudo?) 19:25
jnthn %foo.exists('bar')
grondilu oh yeah. ok 19:26
jnthn My last efforts to look into operator adverbs led to headaches and a busted operator expression parser... :/
masak rn: my %h; say "ok" unless %h.postcircumfix<( )>('foo', :exists)
p6eval niecza v22-16-g4c016f5: OUTPUT«Unhandled exception: Unable to resolve method postcircumfix in type Hash␤ at /tmp/NwdlGnPlct line 1 (mainline @ 4) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4215 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4216 (module-CORE @ 579) ␤ …
..rakudo 64208d: OUTPUT«No such method 'postcircumfix' for invocant of type 'Hash'␤ in block at /tmp/L0vpRT6hMB:1␤␤»
19:26 leont joined
masak rn: my %h; say "ok" unless %h.postcircumfix<{ }>('foo', :exists) 19:27
p6eval niecza v22-16-g4c016f5: OUTPUT«Unhandled exception: Unable to resolve method postcircumfix in type Hash␤ at /tmp/arCFrHNqzI line 1 (mainline @ 4) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4215 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4216 (module-CORE @ 579) ␤ …
..rakudo 64208d: OUTPUT«No such method 'postcircumfix' for invocant of type 'Hash'␤ in block at /tmp/hTdh3q9sf0:1␤␤»
jnthn masak: You missed the :
masak oh right.
rn: my %h; say "ok" unless %h.postcircumfix:<{ }>('foo', :exists)
p6eval rakudo 64208d, niecza v22-16-g4c016f5: OUTPUT«ok␤»
masak \o/
19:27 cognominal joined, Chillance left 19:28 Chillance joined 19:29 azawawi joined
jnthn OK, I don't understand how STD's parsing of my $a++ works out. 19:29
azawawi hi
jnthn Unless there's a trait or post_constraint, the last thing that matches is a <.ws>
grondilu rn: my %foo = 'bar' => 1; undefine %foo<bar>; say "ok" unless %foo.exists: 'bar';
p6eval rakudo 64208d, niecza v22-16-g4c016f5: ( no output ) 19:30
jnthn Which will match between $a and ++
dalek ecs: 30974c6 | masak++ | S29-functions.pod:
[S29] updated references to S32 synopsis
grondilu how do I delete a hash entry (not just set the entry to 'Any')?
jnthn And then POST does:
<!{ @*MEMOS[$¢.pos]<ws> }>
grondilu: .delete('foo')
masak grondilu++ # finding broken links
rn: my %h = foo => 1; %h.postcircumfix<{ }>('foo', :delete); say %h.perl 19:31
jnthn As in, we didn't just match ws.
p6eval niecza v22-16-g4c016f5: OUTPUT«Unhandled exception: Unable to resolve method postcircumfix in type Hash␤ at /tmp/qisI11z9zk line 1 (mainline @ 4) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4215 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4216 (module-CORE @ 579) ␤ …
..rakudo 64208d: OUTPUT«No such method 'postcircumfix' for invocant of type 'Hash'␤ in block at /tmp/ZhniHBeFdc:1␤␤»
jnthn So I'm kinda at a loss as to why it works in STD :)
masak rn: my %h = foo => 1; %h.postcircumfix:<{ }>('foo', :delete); say %h.perl
p6eval niecza v22-16-g4c016f5: OUTPUT«{}.hash␤»
..rakudo 64208d: OUTPUT«("foo" => 1).hash␤»
masak submits rakudobug
dalek rl6-roast-data: 302162b | coke++ | / (3 files):
today (automated commit)
grondilu rn: my %foo = 'bar' => 1; %foo.delete: 'bar'; say "ok" unless %foo.exists: 'bar';
jnthn masak: todo, not bug...
p6eval rakudo 64208d, niecza v22-16-g4c016f5: OUTPUT«ok␤»
jnthn masak: And I'd again be surprised if the adverb stuff ain't already TODO ticketed. 19:32
TimToady 'my foo' is just a term, and POST will match after that without intervening ws
masak jnthn: oh! :exist worked above for the wrong reasons...
:exists
jnthn TimToady: But variable_declarator calls <.ws> explicitly.
19:33 hash_table joined
TimToady well, if there's no actual whitespace there, it can still be a postfix 19:33
jnthn TimToady: But ws matches in this case, no? 19:34
TimToady doesn't matter
POST checks for the absence of ws explicitly
line 3302 19:35
grondilu also, honnestly I kind of liked P5's syntax: "... if exists $foo{bar}"
jnthn Yes, that's what I'm looking at
brrt is the CONTROL phaser implemented anywhere?
19:35 chee left
brrt and if so, hiow 19:35
jnthn Rakudo does <!MARKED('ws')>
In the same place 19:36
Which should be equivalent.
oh.
if ($¢.pos == $startpos) {
@*MEMOS[$¢.pos]<ws>:delete;
}
Is in ws. Does that mean "don't mark it unless we actually had some characters?"
19:37 snearch left
brrt rn: sub foo { say "hi"; return 42; CONTROL { say "bye!"; } } 19:37
p6eval rakudo 64208d: ( no output )
..niecza v22-16-g4c016f5: OUTPUT«Potential difficulties:␤ &foo is declared but not used at /tmp/GMj3js4ZOx line 1:␤------> sub foo ⏏{ say "hi"; return 42; CONTROL { say "by␤␤»
brrt rn: sub foo { say "hi"; return 42; CONTROL { say "bye!"; } }; foo; 19:38
p6eval rakudo 64208d, niecza v22-16-g4c016f5: OUTPUT«hi␤»
brrt is CONTROL not supposed to be called?
19:38 Guest52580 joined
TimToady jnthn: yes, we memoize the space at the end to point to the beginning, but only if it's different 19:39
masak grondilu: to me that syntax looks odd. it looks like we're doing something with the value $foo{bar}. but we're not. we're asking the hash whether that entry is even in the hash. 19:41
grondilu masak: but we do something similar with // 19:42
masak we do?
grondilu r: my %foo; say %foo<bar> // "there is no bar in foo"
p6eval rakudo 64208d: OUTPUT«there is no bar in foo␤»
masak sure, but that's a straight fetch from the hash. 19:43
not an existence check.
meaning that if someone stores an undefined value, you won't be able to tell that apart from no entry at all.
grondilu is skeptical
r: my %foo = 'bar' => undef; say %foo<bar> // "there is no bar in foo" 19:44
p6eval rakudo 64208d: OUTPUT«===SORRY!===␤Unsupported use of undef as a value; in Perl 6 please use something more specific:␤ Mu (the "most undefined" type object),␤ an undefined type object such as Int,␤ Nil as an empty list,␤ !*.defined as a matcher or method,␤ Any:U as a type constraint␤ o…
grondilu r: my %foo = 'bar' => Nil; say %foo<bar> // "there is no bar in foo"
p6eval rakudo 64208d: OUTPUT«there is no bar in foo␤»
19:44 leont left
grondilu indeed 19:44
masak I'm happy empirical evidence still cures misinformed skepticism :) 19:45
grondilu so // has not the exact same semantic as in P5, has it?
doy that's how // works in p5 too 19:46
grondilu oh, I thought // tested existence.
masak no, infix:<//> falls back to the rhs if the lhs is undefined. nothing more. 19:47
doy eval: my %foo = (bar => undef); say $foo{bar} // "there is no bar in foo"
buubot_backup doy: ERROR: syntax error at (eval 20) line 1, near "$foo{bar"
doy eval: use 5.010; my %foo = (bar => undef); say $foo{bar} // "there is no bar in foo" 19:48
buubot_backup doy: there is no bar in foo 1
grondilu needs to read perlop again
doy where is that trailing 1 coming from 19:49
huf return value of say 19:50
the perl5 evalbot evals your code in scalar context and also prints the return value 19:51
doy ah, okay
huf if it's a ref, it gets pretty printed, otherwise just shat out
put a '' as the last expression if you want to suppress it ;)
masak or just don't 'say' the final value ;) 19:53
eval: use 5.010; my %foo = (bar => undef); $foo{bar} // "there is no bar in foo"
buubot_backup masak: there is no bar in foo
huf well yes, but maybe you have a valid reason for saying so much 19:54
19:54 brrt left 19:55 MikeyG joined 20:03 grondilu left 20:05 bruges left
jnthn TimToady: What happens if you have a "my @foo[10];" and assign something like 1..* into it? 20:06
20:07 bruges joined 20:08 bluescreen10 left
geekosaur I would expect the whatever to appropriately limit itself... isn't that the point? 20:08
jnthn geekosaur: Note that 1..* isn't forming a closure; it's a distinct case from @foo[1..*]. 20:09
r: say (1..*).max 20:10
p6eval rakudo 64208d: OUTPUT«Inf␤»
jnthn 1..* is just sugar for 1..Inf
20:12 gaussblurinc joined
TimToady nr: my ($a,$b,$c) = 1..*; say "$a $b $c" 20:18
p6eval rakudo 64208d, niecza v22-16-g4c016f5: OUTPUT«1 2 3␤»
TimToady jnthn: it's just like that
nr: my @a; @a[0..9] = 1..*; say @a; 20:19
p6eval rakudo 64208d, niecza v22-16-g4c016f5: OUTPUT«1 2 3 4 5 6 7 8 9 10␤»
TimToady or that
jnthn my @a[10] = 1..*; # is this done eagerly? 20:20
(the @a[0..9] = 1..* case would be...)
masak I'm fine with it being done eagerly.
TimToady I think it must be eager when there are known targets on the left
jnthn OK, makes sense, thanks.
What of push, shift, etc on a fixed size array? 20:21
TimToady my ($a,$b,@c) = 1..* can treat only the final assignment as lazy
error
jnthn OK, presumably a splice that would result in a change is the same. 20:22
TimToady despises the use of pop on things that can't be popped :)
[Coke] r: my $a = 1..*; say $a.pop; 20:29
p6eval rakudo 64208d: OUTPUT«No such method 'pop' for invocant of type 'Range'␤ in block at /tmp/8cOHCm35jB:1␤␤»
[Coke] r: my $a = 1..*; say $a[*];
p6eval rakudo 64208d: OUTPUT«(timeout)»
20:51 kaare_ left 21:05 kurahaupo joined, GlitchMr left 21:09 azawawi left 21:13 colomon joined
dalek osystem: 0b0c582 | (Timothy Totten)++ | META.list:
Added PSpec
 21:15
sorear is Timothy Totten a #perl6-er? 21:18
21:19 leont joined
sorear o/ colomon 21:19
colomon \o
21:23 leont left 21:26 leont joined 21:28 MayDaniel left
[Coke] sorear: github.com/supernovus 21:29
sorear I see 21:31
dalek kudo/nom: 0dd4514 | jnthn++ | src/Perl6/ (2 files):
Implement INIT as r-value.
21:33 MayDaniel joined
dalek ast: 1d5f961 | jnthn++ | S04-phasers/rvalue.t:
Fix and unfudge phaser r-value tests.

Correct a mis-paren'd eval call, and correct a test that had wrong expectations about containers.		 21:33
21:45 havenn left 21:46 gaussblurinc left, havenn joined 21:50 havenn left 21:54 hash_table left
dalek kudo/nom: e472c9b | jnthn++ | src/Perl6/Actions.pm:
Fix our ($x, $y).
 21:54
kudo/nom: ef70aad | jnthn++ | src/core/control.pm:
Fix return value of take/take-rw.
21:56 arlinius joined 21:57 domidumont joined 22:03 skids left
dalek kudo/nom: 0e41354 | jnthn++ | src/Perl6/World.pm:
Only SET_FILE_LINE on exceptions that can.
 22:07
22:07 benabik left, cognominal left 22:09 MayDaniel left 22:14 domidumont left
masak jnthn: I see you rejected rt.perl.org/rt3/Ticket/Display.html?id=115586 rather than merge it with its earlier duplicate. 22:18
in the long run, merging tickets will simplify searching for old bugs. 22:19
jnthn Not filing dupes within a couple of days of each other would save me time in the short term :P 22:20
Feel free to twiddle it as you prefer.
sorear masak \o/
masak will do. 22:21
jnthn r: say '' ~~ / $d /; 22:22
p6eval rakudo 0e4135: OUTPUT«===SORRY!===␤Variable $d is not declared␤at /tmp/l4Zsow9e9k:1␤»
masak and yes, sorry about filing the same thing twice within such a short time span. there's been a flurry of tickets, and I wanted to make sure I submitted all of them...
merged. 22:23
diakopter \o/
[Coke] always ok to just note it here and some bugadmin will take care of it. 22:26
(mainly masak. :)
diakopter jnthn: sorry masak and I have feeble memories about this :D 22:27
masak generally I catch overlaps, but my overlap detector must have underlapped this time. 22:28
jnthn r: my ${a} = 5; 22:29
p6eval rakudo 0e4135: OUTPUT«===SORRY!===␤Unsupported use of ${a}; in Perl 6 please use $a␤at /tmp/9PM7O_R0P8:1␤»
22:29 spider-mario left
diakopter this cannot be. 22:31
22:31 sizz left
diakopter er, ww 22:31
22:33 cognominal joined 22:35 Guest52580 left, flightrecorder left
jnthn std: sub f { f(|$) }; say "alive" 22:35
p6eval std 04216b1: OUTPUT«===SORRY!===␤Unsupported use of $) variable; in Perl 6 please use $*EGID at /tmp/kH8aP4cv2r line 1:␤------> sub f { f(|$)⏏ }; say "alive"␤Parse failed␤FAILED 00:00 43m␤»
22:35 chee joined, chee is now known as Guest64859 22:36 Guest64859 left, Guest64859 joined 22:37 Guest64859 is now known as chee
jnthn Is rt.perl.org/rt3/Ticket/Display.html?id=102276 rejectable? 22:39
sorear yes and no 22:40
.duckmap still exists, I think
but hypers are specced to do something different now
jnthn r: $_ = 'foo'; s:g/ /a/; 22:41
p6eval rakudo 0e4135: ( no output )
jnthn r: $_ = 'foo'; s:g//a/;
p6eval rakudo 0e4135: OUTPUT«===SORRY!===␤Unrecognized regex metacharacter / (must be quoted to match literally) at line 2, near "a/;"␤»
jnthn std: $_ = 'foo'; s:g/ /a/;
p6eval std 04216b1: OUTPUT«===SORRY!===␤Null pattern not allowed at /tmp/5RJPtyw4vg line 1:␤------> $_ = 'foo'; s:g/ ⏏/a/;␤Parse failed␤FAILED 00:00 43m␤»
jnthn std: $_ = 'foo'; s:g//a/;
p6eval std 04216b1: OUTPUT«===SORRY!===␤Null pattern not allowed at /tmp/4_SLCIBEbD line 1:␤------> $_ = 'foo'; s:g/⏏/a/;␤ expecting colon pair (restricted)␤Parse failed␤FAILED 00:00 43m␤»
22:41 flightrecorder joined
jnthn Well, that's a bunch of tickets moved to testneeded status... 22:49
masak jnthn++ 22:50
jnthn Various of them quoting related.
22:52 leont left 22:53 sftp left
jnthn I wonder if we can use Google Code In to reduce the size of the testneeded list :P 22:53
(If so, need to select ones that aren't really hard to test well...)
dalek kudo/nom: 0a320a4 | jnthn++ | docs/ChangeLog:
A few ChangeLog additions.
 22:56
23:00 Entonian joined 23:01 Entonian left
jnthn 'night, #perl6 23:06
masak 'night, #perl6 23:11
23:11 stopbit left
tadzik good knight 23:11
23:17 PacoAir left 23:18 mtk left 23:21 fgomez left 23:22 benabik joined 23:23 mtk joined 23:36 skids joined 23:39 Guest20434 is now known as ponbiki 23:43 jeffreykegler joined 23:45 jeffreykegler left 23:47 sftp joined 23:59 fgomez 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!