»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, or /msg camelia p6: ... | irclog: irc.perl6.org or colabti.org/irclogger/irclogger_logs/perl6 | UTF-8 is our friend!
Set by moritz on 22 December 2015.
00:02 BrassLantern joined 00:13 labster left
Zoffix It's so quiet.... You could download the sound of a pin dropping... 00:18
00:19 kurahaupo left 00:21 addison left
TEttinger sprint! 00:23
dj_goku ........| 00:34
00:35 RabidGravy left 00:56 BrassLantern left 01:07 rgrinberg left, cdg left 01:10 BenGoldberg_ joined 01:14 labster joined
Zoffix If you are "neuron", "ianmcb", and "LLFourn" wish to use their full names instead of nicknames in this month's Rakudo's contributors list, let me know. 01:14
llfourn, ping ^
01:22 silug left 01:31 huggable joined 01:35 molaf left 01:37 silug joined 01:45 ilbot3 left 01:47 molaf joined, ilbot3 joined
TEttinger Zoffix, that's their real name, it's welsh 01:51
it's pronounced by wrapping your tongue around your left eye and sneezing 01:52
01:54 kalkin- left, kalkin-_ joined
jdv79 could you demonstrate? 01:54
TEttinger I haven't made my sacrifice to Wlldrkrrl today 01:57
01:58 MasterDuke joined
Zoffix Ah :) OK then :) 02:06
02:11 Khisanth left 02:14 itaipu joined
dalek sectbot: 6f9e9a6 | (Daniel Green)++ | benchable.pl:
Discard STDOUT and round timing results
 02:19
02:25 itaipu left 02:26 _28_ria left 02:28 Khisanth joined
dalek sectbot: b56e519 | (Daniel Green)++ | benchable.pl:
Add missing initial '|' from error output
 02:35
02:43 Madcap^Jake^ joined, Madcap^Jake^ is now known as MadcapJake 02:44 noganex_ joined 02:47 noganex left
dalek sectbot: f579c64 | (Daniel Green)++ | benchable.pl:
Better associate errors with the commit they came from
 02:49
02:57 Tonik joined
Tonik I write 03:00
my $a = 42;
my $b := $a;
... and I do some awesome stuff ...
but then I want $b to become a happy free variable again, not bound to $a. How do I do that without touching $a?
AlexDaniel great question 03:02
Zoffix Bind it to something else?
geekosaur just tried Scalar.new but it doesn't like that
Zoffix m: my $a = 42; my $b := $a; $b = 72; say $a; my $z = 42; $b := $z; $b = 'meow'; say $a
camelia rakudo-moar e2ec35: OUTPUT«72␤72␤»
AlexDaniel Zoffix: what if you want $b = 42 without binding?
Zoffix TBH, there's likely a much better solution to Tonik's problem. 03:03
Tonik Zoffix, $b wants to be on its own, not bound to anyone :(
Zoffix Tonik, why?
Tonik because if it can't, it will feel something is wrong with Perl 6 03:04
Zoffix :/
Very well then.
konobi yeah, you'd just use a different variable in real life, no?
but it's probably doable 03:05
geekosaur hack alert 03:07
m: my $a = 4; my $b := $a; $b = 6; say $a; $b := $; $b = 2; say $a; say $b;
camelia rakudo-moar e2ec35: OUTPUT«6␤6␤2␤»
Tonik I've been trying to wrap my mind around the idea of binding in Perl 6, and I almost convinced myself it makes sense, until I found out you can rebind what's already been bound, and it all fell apart for me. Seems that every variable is in fact a pointer, but you don't get to enjoy the full power of pointers as in other languages, and these pointers are shoved down your throat whether you want it or not
konobi well, everything's a reference
Tonik except "my $a := 42" which is not, right? 03:08
AlexDaniel wait, um… if you don't want to rebind then just don't? Doctor, it hurts when I do this…
konobi Tonik: github.com/rakudo/rakudo/blob/nom/...ariable.pm 03:10
so gotta be able to do it somehow =0) 03:11
geekosaur points up at hack
an alternative would be for Scalar.new to not produce Nil...
well, actually, the hack will lose if it's used in a scope that gets re-entered at some point, since $ will use the same Scalar each time whereas $a will be a new one each time, so not 100% the same behavior 03:13
Tonik geekosaur, uh, what is $ ? 03:15
geekosaur anonymous state variable
(like a C `static`)
konobi geekosaur: seems to me like a lot of that sort of state management would be a pretty interesting pattern for IOC 03:16
Tonik I see
geekosaur but since it's anonymous, every use of `$` is distinct; so if you did `$b := $; $c := $;` then each one gets its own `state` variable which will be reused every time the scope is re-entered 03:17
would have to figure out how to bind it to a new Scalar to avoid this (which currently throws) 03:19
m: my $a = 4; my $b := $a; $b = 6; say $a; $b := Scalar.new; $b = 2; say $a; say $b;
camelia rakudo-moar e2ec35: OUTPUT«6␤Cannot call method 'BUILDALL' on a null object␤ in block <unit> at <tmp> line 1␤␤»
geekosaur (I suspect it can be made to work but Scalar.new needs more information, which if I had to guess is an initial content --- but it also does not take positionals. so I'd have to dig into implementation of Scalar to see what to feed it) 03:20
konobi could you apply a role to the instance of the Scalar and override the scope? 03:21
*blink* could do some crazy stuff 03:23
03:24 adu joined, kid51 joined
geekosaur looks at Scalar impl and gets a headache --- need to create a Perl6::Metamodel::ContainerDescriptor, then the Scalar with named parameters descriptor and value 03:25
gfldex konobi: could can't get hold of an instance of scalar. There is no way to tell the compiler you are talking to the container. There is Proxy tho.
adu TimToady!
Zoffix Would be neat if bisectable could bisect hangs. 03:28
Can it take a gist? 03:29
AlexDaniel sure
Tonik if some of you guys have a sensible visual representation of what a scalar is and how binding works, please share it with me 03:30
or draw it on a piece of paper and show it to me
Zoffix ," gist.githubusercontent.com/zoffixz...tfile1.txt
bisect: gist.githubusercontent.com/zoffixz...tfile1.txt
bisectable Zoffix: Successfully fetched the code from the provided URL. 03:31
Zoffix: Exit code is 1 on both starting points, bisecting by using the output
Zoffix: bisect log: gist.github.com/1314ad2dec81edc86f...a94eb64bc8
Zoffix: (2015-12-25) github.com/rakudo/rakudo/commit/07fecb5
Zoffix Tonik, there's Containers page on docs. It may be useful: docs.perl6.org/language/containers
Tonik nono, I want to know how *you* *visualize* it
adu Tonik: "scalar"?
Zoffix I don't use binding. 03:32
Tonik sighs
adu Tonik: my personal definition of scalar is "not deep"
gfldex Zoffix: you do use binding by calling subs and methods
my personal definiton of scalar is "double indirection". Also somebody else wrote that scalars are object the forward pretty much all their methods to another object. 03:33
Tonik gfldex, double indirection sounds about right :( 03:34
geekosaur the "natural" state of a name is to bind to something. if you bind it to a value, the name is readonly. if you bind it to a container, the name is read-write and you can store values in the container and replace them with new values. `=` stores a new value in a container (and throws if it's not bound to a container). `:=` binds to something else, which may be a container from some other place.
gfldex don't look so much at := and ::= to understand binding. Look at signatures and calls. 03:35
also, default values.
Tonik hm. what is ::= ?
gfldex compile time binding 03:36
Tonik oh
geekosaur it makes a reasonable amount of sense to me, but then I program in Haskell so I have a clear distinction between the concepts to start with :)
gfldex m: my Int $i = Failure.new('Dude, I'm empty!'); say $i;
camelia rakudo-moar e2ec35: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Unable to parse expression in argument list; couldn't find final ')' ␤at <tmp>:1␤------> 3my Int $i = Failure.new('Dude, I'7⏏5m empty!'); say $i;␤ expecting any of:␤ infix␤ i…»
gfldex m: my Int $i is default Failure.new('Dude, I'm empty!'); say $i; 03:37
camelia rakudo-moar e2ec35: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Two terms in a row␤at <tmp>:1␤------> 3my Int $i is default7⏏5 Failure.new('Dude, I'm empty!'); say $i␤ expecting any of:␤ constraint␤ infix␤ infix stopper␤ …»
gfldex m: my Int $i = 1is default Failure.new('Dude, I'm empty!'); say $i;
camelia rakudo-moar e2ec35: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Confused␤at <tmp>:1␤------> 3my Int $i = 17⏏5is default Failure.new('Dude, I'm empty!␤»
gfldex m: my Int $i = 1 is default Failure.new('Dude, I'm empty!'); say $i;
camelia rakudo-moar e2ec35: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Two terms in a row␤at <tmp>:1␤------> 3my Int $i = 17⏏5 is default Failure.new('Dude, I'm empty␤ expecting any of:␤ infix␤ infix stopper␤ postfix␤ statement…»
Tonik "::=" not yet implemented. Sorry.
Tonik sighs
gfldex m: my Int $i is default Failure.new('Dude, I'm empty!') = 1; say $i;
camelia rakudo-moar e2ec35: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Two terms in a row␤at <tmp>:1␤------> 3my Int $i is default7⏏5 Failure.new('Dude, I'm empty!') = 1; sa␤ expecting any of:␤ constraint␤ infix␤ infix stopper␤ …»
geekosaur doesn't is default require some kind of parens/braces? (i.e. it is syntactically an adverb)
Zoffix It's a trait 03:39
m: my Int $i is default(Failure.new: 'Dude, Im empty!') = 1;
camelia rakudo-moar e2ec35: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Default value '(HANDLED) Dude, Im empty!␤ in any at src/Perl6/World.nqp line 2199␤ in any trait_mod:sym<is> at gen/moar/m-Perl6-Actions.nqp line 5056␤ in any trait_mod:sym<is> at src/Perl6/Grammar.…»
gfldex m: my Int $i is default(42); $i = 10; $i = Nil; say $i; 03:40
camelia rakudo-moar e2ec35: OUTPUT«42␤»
gfldex m: my $i is default(Failure.new("Dude, I'm empty!")) = 1; say $i; 03:41
camelia rakudo-moar e2ec35: OUTPUT«1␤»
gfldex m: my $i is default(Failure.new("Dude, I'm empty!")); say $i;
camelia rakudo-moar e2ec35: OUTPUT«===SORRY!===␤Dude, I'm empty!␤»
gfldex the type constraint seams to peek into the container and triggers the Failure 03:42
m: my Int $i is default(Nil); 03:44
camelia rakudo-moar e2ec35: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Default value 'Nil' will never bind to a parameter of type Int␤at <tmp>:1␤------> 3my Int $i is default(Nil)7⏏5;␤ expecting any of:␤ constraint␤»
gfldex m: my $i is default(Nil);
camelia ( no output )
gfldex m: my $i is default(Nil); dd $i;
camelia rakudo-moar e2ec35: OUTPUT«Nil $i = Nil␤»
03:44 Zoffix left
gfldex ha! I finally found a way to hold onto Nil. :) 03:44
that leaves the question if that is sensible to allow 03:45
03:45 rgrinberg joined 04:00 kid51 left
dalek sectbot: a8df0e3 | (Aleks-Daniel Jakimenko-Aleksejev)++ | benchable.pl:
Oops, restore server
 04:00
04:02 benchable joined, jsimonet left, jsimonet joined 04:08 khw left 04:12 skids left 04:13 BrassLantern joined
MasterDuke benchable: HEAD~50..HEAD my $a = "a" x 2**16;my $b = $a.chop($_) for ^1000 04:18
04:18 cpage__ joined
benchable MasterDuke: gist.github.com/0d621f847cc2b80db1...e77d20c71b 04:19
04:19 cpage_ left, cpage__ is now known as cpage_
AlexDaniel benchable: 2015.10,2016.02,2016.04,HEAD my $a = "a" x 2**16;my $b = $a.chop($_) for ^5000 04:20
benchable AlexDaniel: |«2015.10»:1.1509␤|«2016.02»:1.0659␤|«2016.04»:1.0590␤|«HEAD»:1.1272
04:23 djbkd left 04:24 djbkd joined
dalek sectbot: f763de7 | (Aleks-Daniel Jakimenko-Aleksejev)++ | / (2 files):
Fix encoding issue

get_output should probably decode automatically, but for now this is enough.		 04:26
04:27 adu left
dalek sectbot: 57bd995 | (Aleks-Daniel Jakimenko-Aleksejev)++ | committable.pl:
Oops, missed “use”
 04:28
04:29 bisectable left, committable left, bisectable joined, committable joined 04:37 Cabanossi left 04:40 Cabanossi joined 04:57 holyghost left 04:58 holyghost joined 05:01 samb1 left 05:12 samb1 joined 05:20 AlexDaniel left 05:29 CIAvash joined 05:31 huggable left, huggable joined 05:35 BrassLantern left 05:39 cpage_ left 05:40 cpage_ joined 05:42 wamba joined 05:48 BenGoldberg_ left 06:01 canopus left 06:05 furelypunctional joined 06:06 canopus joined 06:12 djbkd left 06:16 TreyHarr1 left 06:30 domidumont joined 06:34 domidumont left 06:35 domidumont joined 06:40 domidumont left
masak in a moment of weakness, I expected .prompt to be a method on $*IN, not just a global sub 06:40
now that I think about it, I'm not sure that was reasonable -- prompt() does involve both output and input 06:41
but I will not that I did expect it :)
note* 06:42
no big deal either way -- I'll solve what I wanted to do in a different way ;) 06:44
06:49 maddingue joined 06:58 darutoko joined 07:01 firstdayonthejob joined 07:03 donaldh joined 07:04 canopus left
moritz could $*OUT.prompt($message, :readfrom($*IN)) 07:05
or $*IN.prompt($message, :out($*OUT))
07:09 canopus joined
Ulti lizmat++ another tenth of a second saving on my tests 07:18
07:21 rgrinberg left, domidumont joined 07:32 domidumont left 07:40 RabidGravy joined 07:49 donaldh left 07:51 wamba left 08:00 firstdayonthejob left 08:04 furelypunctional left
Xliff_ Will .defined return true if scalar is set to a type object? 08:07
Always thought it returned false.
m: my $a = Str; say $a.defined; 08:08
camelia rakudo-moar e2ec35: OUTPUT«False␤»
Xliff_ And so it doesn't. Hmmm...
m: use NativeCall; class A does Str { }; my $a = nativecast(A, Str); say $a.defined; 08:09
camelia rakudo-moar e2ec35: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Str is not composable, so A cannot compose it␤at <tmp>:1␤»
Xliff_ m: use NativeCall; class A does Int { }; my $a = nativecast(A, Int); say $a.defined;
camelia rakudo-moar e2ec35: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Int is not composable, so A cannot compose it␤at <tmp>:1␤»
Xliff_ Hrm. 08:10
08:19 FROGGS joined
FROGGS o/ 08:22
08:35 rindolf joined 08:39 girafe joined 08:41 cpage_ left
holyghost »ö« 08:53
09:01 Ven joined 09:10 wamba joined
masak »ö« 09:21
09:24 setty1 joined 09:31 huggable left, huggable joined 09:36 Senji left, Senji joined 09:39 TreyHarr1 joined 09:49 wamba left 09:55 donaldh joined, espadrine joined 10:02 rindolf left 10:03 TEttinger left 10:06 rindolf joined 10:16 donaldh left 10:19 labster left
Xliff_ FROGGS: \o 10:53
m: use NativeCast; class A is repr('CStruct') { }; class B does A { }; my $a = A; my $b = nativecast(B, $a); say $b.defined; 10:55
camelia rakudo-moar cd19db: OUTPUT«===SORRY!===␤Could not find NativeCast at line 1 in:␤ /home/camelia/.perl6␤ /home/camelia/rakudo-m-inst-1/share/perl6/site␤ /home/camelia/rakudo-m-inst-1/share/perl6/vendor␤ /home/camelia/rakudo-m-inst-1/share/perl6␤ CompUnit::Re…»
Xliff_ m: use NativeCall; class A is repr('CStruct') { }; class B does A { }; my $a = A; my $b = nativecast(B, $a); say $b.defined;
camelia rakudo-moar cd19db: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤A is not composable, so B cannot compose it␤at <tmp>:1␤»
Xliff_ m: use NativeCall; class A is repr('CStruct') { has int32 $.a }; class B does A { }; my $a = A; my $b = nativecast(B, $a); say $b.defined; 10:56
camelia rakudo-moar cd19db: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤A is not composable, so B cannot compose it␤at <tmp>:1␤»
Xliff_ m: use NativeCall; class A is repr('CStruct') { has int32 $.a }; class B is A { }; my $a = A; my $b = nativecast(B, $a); say $b.defined;
camelia rakudo-moar cd19db: OUTPUT«Internal error: unhandled target type␤ in sub nativecast at /home/camelia/rakudo-m-inst-1/share/perl6/sources/24DD121B5B4774C04A7084827BFAD92199756E03 (NativeCall) line 406␤ in block <unit> at <tmp> line 1␤␤»
jnthn Xliff_: At a guess, you'd need to mark B as "is repr('CStruct')" too for that to work 10:57
Xliff_ m: use NativeCall; class A is repr('CStruct') { has int32 $.a }; class B is A is repr('CStruct') { }; my $a = A; my $b = nativecast(B, $a); say $b.defined;
camelia rakudo-moar cd19db: OUTPUT«False␤»
Xliff_ jnthn: Yep. Thanks.
jnthn Also, you can factor things out in to roles and compose those into classes marked "is repr('CStruct')" if you like
Since composition is flattening, that should Just Work. 10:58
Xliff_ jnthn: Thanks! That's good to know.
Right now I'm trying to figure out why certain routines seem to be returning type objects even though they've gone through a .defined check. 10:59
So far every use case I've tried through camelia seems to work fine, but in my code it doesn't. Must be missing something.
jnthn Odd... Guess it's a case of trying to boil it down to the smallest possible test case... 11:04
11:04 avenj left 11:08 TreyHarr1 left 11:19 kid51 joined 11:20 girafe left 11:22 girafe joined
timotimo also, just to make sure it's not something wonky in spesh or something, try MVM_SPESH_DISABLE=yesplease 11:23
11:40 pmurias joined
pmurias .tell konobi using v8-profiler and loading the profiles in chrome turned out the thing I needed for profiling 11:42
yoleaux pmurias: I'll pass your message to konobi.
pmurias do we have an API for profiling things?
that is for turning profiling on from Perl6/NQP code? 11:43
11:43 cpage_ joined
timotimo we do for moarvm 11:46
but it's just two ops
11:48 girafe2 joined, TreyHarr1 joined 11:50 girafe left
masak moritz: I'd argue that .prompt is "more" a method on $*IN than on $*OUT, since $message is optional but the .get is not ;) 11:55
pmurias timotimo: how are they used? 12:09
jnthn pmurias: I think the most natural way would be to nqp::getcomp('perl6').eval('some code', :profile(1)) or some such 12:10
pmurias: (Which already should work, or something close to it) 12:11
See run_profiled in the MoarVM Backend.nqp in the NQP repo for how it's done
In fact, you can probably use nqp::getcomp('perl6').backend.run_profiled($code_object)
12:12 kaare_ left
jnthn (Note, method names from memory, may not 100% match reality ;)) 12:12
12:27 amoe_ joined 12:37 pmqs_ joined 12:41 pmqs_ left 12:46 kid51 left 12:52 pmurias left 12:56 BenGoldberg_ joined 13:01 avenj joined, avenj left, avenj joined
dogbert17 it's very quiet around here, is everyone out tapping their beer supplies 13:01
m: say 'a b' ~~ /a b/ 13:02
camelia rakudo-moar 88c35e: OUTPUT«Potential difficulties:␤ Space is not significant here; please use quotes or :s (:sigspace) modifier (or, to suppress this warning, omit the space, or otherwise change the spacing)␤ at <tmp>:1␤ ------> 3say 'a b' ~~ /a7⏏5 b/␤Nil␤»
dogbert17 do we really need to list this as a trap when we get this warning? docs.perl6.org/language/traps#Regexes 13:03
13:05 sammers left
MasterDuke IMHO, yes. Perl 5 regexes are almost muscle memory now, but i have to stop and think about Perl 6 syntax. anything to warn people ahead of time is good 13:10
psch the warning somewhat goes against /x as default 13:12
dogbert17 MasterDuke: perl6 does warn you, but the text implies that is does not, perhaps it needs to be rephrased a bit 13:13
psch and perl5 doesn't warn on 'say "foo" =~ / f o o /x'
(yes, i had to check :P )
which is to say i'd rather like the warning to go and only keep it around in language/traps 13:15
m: $ = quietly / a b / # especially as it's not quietly-able 13:16
camelia rakudo-moar 88c35e: OUTPUT«Potential difficulties:␤ Space is not significant here; please use quotes or :s (:sigspace) modifier (or, to suppress this warning, omit the space, or otherwise change the spacing)␤ at <tmp>:1␤ ------> 3$ = quietly / a7⏏5 b / # especial…»
dogbert17 that's one way I guess. to me it seems that the warning was added after the text was written
13:16 sammers joined 13:18 pmurias joined
pmurias are the situations where you want use /a b/ instead of /ab/ common? 13:20
13:21 holyghost left
psch matching any continuous over an (arbitrary) length of 40 chars in a regex is where it applies 13:21
i don't know how common that is, but i also don't think that "how common" matters for my argument
13:22 holyghost joined
psch but i also don't really feel strongly about it 13:22
it just seems weird to me that we "force" /x on people and complain when they use it to its full extend
jnthn m: no worries; / a b /
camelia rakudo-moar 88c35e: OUTPUT«Potential difficulties:␤ Space is not significant here; please use quotes or :s (:sigspace) modifier (or, to suppress this warning, omit the space, or otherwise change the spacing)␤ at <tmp>:1␤ ------> 3no worries; / a7⏏5 b /␤»
jnthn Hm, I'd have expected that to suppress it
13:28 kurahaupo joined 13:30 holyghost left, CIAvash left, holyghost joined 13:31 huggable left, huggable joined
pmurias psch: "how common" matters a lot for warnings, what the warning is IMHO intented to catch is people assuming :sigspace is in effect when it's not 13:35
psch pmurias: as i said, i don't think that matters for my argument 13:43
13:43 skids joined
psch pmurias: i mean, the intend behind making all regexen assume /x semantic is to allow arbitrary whitespace 13:43
pmurias: and then we complain about arbitrary whitespace, which just seems weird 13:44
but, as also said, i don't feel strongly about it
13:48 cpage joined 13:49 cpage__ joined
pmurias psch: warnings are for things that are allowed but confusing/seem to be a mistake, for things that are not allowed we have errors 13:49
13:50 cpage_ left, cpage__ is now known as cpage_
psch m: /a b+/ # i want to seperate, purposefully, for clarity 13:51
camelia rakudo-moar 88c35e: OUTPUT«Potential difficulties:␤ Space is not significant here; please use quotes or :s (:sigspace) modifier (or, to suppress this warning, omit the space, or otherwise change the spacing)␤ at <tmp>:1␤ ------> 3/a7⏏5 b+/ # i want to seperate, p…»
psch because ab+ is easily misunderstood as "ab"+
having to silence the warning by rewriting that just makes it worse
granted, if we could silence it that particular case stops being valid 13:52
but we can't
m: say "foo" or "bar" # why aren't we warning here?
camelia rakudo-moar 88c35e: OUTPUT«foo␤»
psch stops now
i'm getting way to invested :)
masak because for all the compiler knows, `say` might return something falsy, which makes the rhs of the `or` run 13:56
14:02 TreyHarr1 left 14:05 TreyHarr1 joined, kaare_ joined 14:08 canopus left
dogbert17 m: sub postfix:<!>(Int $x where { $x >= 0 }) { [*] 1..$x }; say 5! # oddball question, but why does this take almost one sec to parse 14:12
camelia rakudo-moar cd4265: OUTPUT«120␤»
14:13 rgrinberg joined
moritz dogbert17: adding new operators is fairly expensive, because rakudo has to re-calculate the finite statement machines for the regex engine 14:15
dogbert17: but it surprises me that it takes almost a second
Xliff_ timotimo, what is spesh? 14:16
14:16 Xliff_ is now known as Xlif, Xlif is now known as Xliff
moritz Xliff: just-in-time specialization of code 14:16
things like eliminating unecessary type checks, containers etc. 14:17
timotimo yup
Xliff kk
dogbert17 moritz: on my machine, (old i5) it takes ~0.9 sec
moritz 0.5s here 14:18
dogbert17 moritz: do you have an i7?
moritz dogbert17: I think so
14:19 canopus joined
dogbert17 moritz: compare with this, takes 0.13 sec for me: ulti minfix:<+>(Int $x, "same") { 2 * $x }; 14:19
moritz if I change it to a regular sub, it takes 0.17s instead
dogbert17 aha, interesting
moritz infix:<+> already exists, so no parser changes required there 14:20
dogbert17 ah, didn't think of that
Xliff If I have a base class that has attributes, and a descendent class that wants to set them, how would I do that? 14:26
Example: class A { has $.a }; class B is A { setA { # Want to set class A's $!a here } } 14:27
I think I've asked this before but...
Right now both are repr('CStruct') so I can use nativecast() to access the super, but I don't think that's a wise long-term solution.
timotimo did someone recommend mixing in a role yet? :) 14:28
gfldex m: class B {...}; class A { has $!foo = 42; trusts B; method say { put $!foo } }; class B is A { method m(){ $A::foo = "answer"; } }; B.new.m.say 14:29
camelia rakudo-moar cd4265: OUTPUT«answer␤»
gfldex Xliff: ^^^ (there may be dragons) 14:30
Xliff gfldex: Thanks. Will try. Would prefer to see an example without "has $!att" 14:31
And that doesn't need trust.
But then again, beggars can't be choosers.
gfldex m: class A { has $.foo = 42; method say { put $!foo } }; class B is A { method m(){ $A::foo = "answer"; } }; B.new.m.say
camelia rakudo-moar cd4265: OUTPUT«answer␤»
gfldex there you go
Xliff LOL
timotimo does the attribute really have to be private?
Xliff My attributes are not private. 14:32
gfldex :->
timotimo in that case, why don't you just $.foo = 10?
Xliff Rakudo doesn't let me.
Says attribute is not found.
timotimo please share the code where that errors?
14:33 khw joined
Xliff Let me see if I can create a use case, timotimo 14:33
timotimo thanks 14:34
because what you're describing sounds a bit like a bug
14:36 ggoebel joined
Xliff Attribute $!intSubset not declared in class XML::LibXML::Document 14:38
But $!intSubset is declared in xmlDoc, and class XML::LibXML::Document is xmldoc. 14:39
Er...
class XML::LibXML::Document is xmlDoc.
github.com/Xliff/p6-XML-LibXML/blo...cts.pm#L95 <- xmlDoc def 14:40
github.com/Xliff/p6-XML-LibXML/blo...ent.pm#L18 <- XML::LibXML::Document def 14:41
github.com/Xliff/p6-XML-LibXML/blo...nt.pm#L527 <- Line causing error. 14:43
timotimo, sorry to be throwing the kitchen sink at you, but every time I've tried to pair this down to something small, it doesn't reproduce.
14:43 kid51 joined
Xliff One moar time... 14:43
14:43 kid51 left, kid51 joined
timotimo oh 14:43
it's a regular class and the derived class is a cstruct? 14:44
Xliff Both classes are CStructs
timotimo i don't see the CStruct for xmlDoc
Xliff One is barebones for some nativecall routines other is full object rep.
timotimo when you have the superclass p6opaque and the subclass cstruct, i don't know how that behaves 14:45
Xliff timotimo: that is part of the predef... github.com/Xliff/p6-XML-LibXML/blo...cts.pm#L24
timotimo oh, ok 14:46
14:46 mtj_ left
timotimo i haven't used derived classes with cstruct yet, but i thought it'd Just Work™ 14:46
Xliff m class A { has $.a }; class B is A { setA($a) { $!a = $a } }; my $b = B.new; $b.setA(42); say $b.a; 14:47
m: class A { has $.a }; class B is A { setA($a) { $!a = $a } }; my $b = B.new; $b.setA(42); say $b.a;
camelia rakudo-moar cd4265: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Variable '$a' is not declared␤at <tmp>:1␤------> 3lass A { has $.a }; class B is A { setA(7⏏5$a) { $!a = $a } }; my $b = B.new; $b.se␤»
Xliff m: class A { has $.a }; class B is A { method setA($a) { $!a = $a } }; my $b = B.new; $b.setA(42); say $b.a;
camelia rakudo-moar cd4265: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Attribute $!a not declared in class B␤at <tmp>:1␤------> 3 B is A { method setA($a) { $!a = $a } }7⏏5; my $b = B.new; $b.setA(42); say $b.a;␤ expecting any of:␤ horizontal whitespace…»
Xliff ^^^
m: class A { has $.a }; class B is A { method setA($a) { $A::a = $a } }; my $b = B.new; $b.setA(42); say $b.a; 14:48
camelia rakudo-moar cd4265: OUTPUT«(Any)␤»
Xliff O_o
What did I miss?
14:48 ggoebel_ joined 14:49 mtj_ joined
Xliff m: class A { has $.a }; class B is A { method setA($a) { $A!a = $a } }; my $b = B.new; $b.setA(42); say $b.a; 14:50
camelia rakudo-moar cd4265: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Variable '$A' is not declared. Did you mean '$a'?␤at <tmp>:1␤------> 3$.a }; class B is A { method setA($a) { 7⏏5$A!a = $a } }; my $b = B.new; $b.setA(42␤»
Xliff m: class A { has $.a }; class B is A { method setA($a) { $A::a = $a; say $a; } }; my $b = B.new; $b.setA(42); say $b.a;
camelia rakudo-moar cd4265: OUTPUT«42␤(Any)␤»
Xliff m: class A { has $.a }; class B is A { method setA($a) { $A::a = $a; say $a; say $.a; } }; my $b = B.new; $b.setA(42); say $b.a;
camelia rakudo-moar cd4265: OUTPUT«42␤(Any)␤(Any)␤»
Xliff m: class A { has $.foo = 42; method say { put $!foo } }; class B is A { method m(){ $A::foo = "answer"; } }; my $b = B.new; $b.m; say $b.foo 14:51
camelia rakudo-moar cd4265: OUTPUT«42␤»
Xliff Ah. And thus the dragon breathes fire.
14:52 ggoebel left
Xliff The assignment to $A::foo is not setting the $.foo attribute. 14:52
14:53 ggoebel_ is now known as ggoebel 14:54 Ven left 14:55 Ven joined 14:56 holyghost left
gfldex m: class A { has $.a }; class B is A { method setA($a) { $A::a := $a; say $a; say $.a; } }; my $b = B.new; $b.setA(42); say $b.a; 14:56
camelia rakudo-moar cd4265: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Cannot use bind operator with this left-hand side␤at <tmp>:1␤------> 3s B is A { method setA($a) { $A::a := $a7⏏5; say $a; say $.a; } }; my $b = B.new; $␤»
gfldex $A::a is the accessor method, not the container 14:57
timotimo sorry, i'm quite distracted 14:58
why would you have to $A::a? the method is already public, so it's derived and can be called on self 14:59
15:00 ggoebel left
gfldex m: class A is rw { has $.a = 42 }; class B is A { method setA($a) { $.a = $a; } }; my $b = B.new; $b.setA(42); say $b.a; 15:00
camelia rakudo-moar cd4265: OUTPUT«42␤»
gfldex Xliff: also ^^^
15:00 ggoebel joined
gfldex m: class A { has $.a = is raw 42 }; class B is A { method setA($a) { $.a = $a; } }; my $b = B.new; $b.setA(42); say $b.a; 15:00
camelia rakudo-moar cd4265: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Undeclared routines:␤ is used at line 1␤ raw used at line 1␤␤»
gfldex m: class A { has $.a = is rw 42 }; class B is A { method setA($a) { $.a = $a; } }; my $b = B.new; $b.setA(42); say $b.a;
camelia rakudo-moar cd4265: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Undeclared routines:␤ is used at line 1␤ rw used at line 1␤␤»
gfldex m: class A { has $.a = is rw; }; class B is A { method setA($a) { $.a = $a; } }; my $b = B.new; $b.setA(42); say $b.a; 15:01
camelia rakudo-moar cd4265: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Undeclared routines:␤ is used at line 1␤ rw used at line 1␤␤»
gfldex m: class A { has $.a is rw; }; class B is A { method setA($a) { $.a = $a; } }; my $b = B.new; $b.setA(42); say $b.a;
camelia rakudo-moar cd4265: OUTPUT«42␤»
gfldex ENEEDMORETEA
Xliff gfldex: Thanks! One of those might work. 15:02
timotimo: I was trying to set the attribute. Not access a method.
15:04 kid51 left, adu joined
gfldex Xliff: you may be aiming at your feed. Esp. if you intend to use threads to make your life harder. 15:04
threads are clearly the work of the devil 15:05
pmurias jnthn: re run_profiled it fits almost perfectly for the kind of profiling support I want to expose on the js backend 15:07
timotimo but why don't you set the attribute via the method? 15:08
Xliff timotimo: ??? 15:09
15:09 Ven left
Xliff timotimo: Via what method? 15:09
timotimo well, make the attribute rw and set the value via the accessor?
the method that has the same name as the attribute
Xliff Oh. Could do that, but sometimes we get "Cannot set an immutable object" so I'm forced to drop back to nqp and bindattr. 15:10
And in those situations I have to nativecast back to the superclass.
And I'm worried that in some situations, when doing that I am confusing something in the model because test that should pass... aren't. 15:11
timotimo just "is rw" :) 15:12
Xliff Even with "is rw" I get that "Cannot set an immutable object" error.
Especially when dealing with structs that originate in libxml2.
dogbert17 m: gist.github.com/dogbert17/dcf078ff...5632b4ad63
15:12 kid51 joined, camelia joined
Xliff ok 9 - DTD's system ID matches $htmlSystem 15:12
Cannot modify an immutable xmlDtdPtr
timotimo wtf?
15:12 Ven joined
Xliff ^^ There's one. 15:12
dogbert17 m: gist.github.com/dogbert17/dcf078ff...5632b4ad63 15:13
timotimo is that attribute rw?
it doesn't seem like you're getting a container there
Xliff I changed xmlDoc to "is rw"
dogbert17 hmm, shouldn't that work, evaluating a gist that is
15:13 ChanServ sets mode: +v camelia
camelia rakudo-moar cd4265: OUTPUT«Attempt to return outside of any Routine␤ in block <unit> at <tmp> line 8␤␤» 15:14
Xliff I can try changing that attribute to "is rw" as well.
dogbert17 why does the example fail with: Attempt to return outside of any Routine␤ in block <unit> at <tmp> line 8
timotimo is rw on the class should do that for you, but i'm not sure how stubbed classes and traits like that interact, really
Xliff Well I have done "is rw" on the class and the trait. 15:15
timotimo OK 15:16
Xliff FROGGS initially wrote most of this and was using nqp::bindattr to get around it.
15:16 donaldh joined, dvinciguerra_ joined
gfldex dogbert17: a pointy block is a Block, what is lower in the chain then A Routine. 15:16
Xliff timotimo: Still getting same error.
gfldex dogbert17: IIRC nexsame and/or callwith fire control exceptions 15:17
dogbert17 gfldex: dunno what it should look like but note that I took this from docs.perl6.org/language/functions
gfldex dogbert17: `return` basicylly _is_ a control exception, that's where the LTA error message comes from
timotimo are you sure nextsame and callwith work via control exceptoins? 15:18
... toins?
dogbert17 gfldex; so does that mean that the example is incorrect?
Xliff timotimo: github.com/Xliff/p6-XML-LibXML.git. Checkout 07dtd branch. Then run "perl6 -Ilib t/07dtd" to see what I am talking about.
Gotta split for a bit.
gfldex dogbert17: since we don't got automatic testing on our examples and some of them are quite old, there will be broken examples 15:19
i know because I fixed a few already
timotimo i'm not sure i actually can figure that out now
dogbert17 gfldex: I've fixed a few too but in this case I don't know how it should look ;( 15:20
in order to do what the author intended
gfldex m: gist.github.com/gfldex/ecde10b59aa...dc4e74ddf0 15:22