»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.
TimToady nr: my $x; BEGIN $x = 42; say $x; 00:32
p6eval rakudo b10e2b, niecza v24-24-gbdc3343: OUTPUT«42␤»
TimToady nr: our $x; BEGIN $x = 42; say $x;
p6eval rakudo b10e2b: OUTPUT«===SORRY!===␤Cannot modify an immutable value␤»
..niecza v24-24-gbdc3343: OUTPUT«42␤»
diakopter r: my $*a = 33; BEGIN { say($*a); $*a = 44; }; say $*a 00:33
p6eval rakudo b10e2b: OUTPUT«===SORRY!===␤Dynamic variable $*a not found␤»
diakopter r: my $*a = 33; BEGIN { $*a = 44; }; say $*a
TimToady try the our in the repl and it coredumps
p6eval rakudo b10e2b: OUTPUT«===SORRY!===␤Dynamic variable $*a not found␤»
diakopter the repl's antimatter containment is very weak 00:34
TimToady sure, that's fine, since there's no dynamic context yet
doesn't blow up till you exit though, oddly 00:35
(if you exit with ^D) 00:36
uvtc TimToady: thanks for trying to help earlier with your examples using .list, .tree, and .lol , though I don't think I understand what you were trying to illustrate. 01:30
TimToady mostly I was just trying to understand how much of .tree was implemented :) 01:33
uvtc What should .tree do?
TimToady r: say (<a b c>, (<d e f>, <g h i>)).tree(*.Array,2).perl
p6eval rakudo b10e2b: OUTPUT«Cannot call 'tree'; none of these signatures match:␤:(Any:U : Mu *%_)␤:(Any:D : Mu *%_)␤:(Any:D : &c, Mu *%_)␤:(Any:D : Cool $count, Mu *%_)␤␤ in method tree at src/gen/CORE.setting:1366␤ in block at /tmp/sAzmCQh4zz:1␤␤»
TimToady it's supposed to transform a structure of parcels into something more solid 01:34
uvtc Parcels.
TimToady r: say <a b c>.WHAT
p6eval rakudo b10e2b: OUTPUT«(Parcel)␤»
TimToady r: say (<a b c>, (<d e f>, <g h i>)).tree(*.Array)[0].WHAT
p6eval rakudo b10e2b: OUTPUT«(Array)␤»
TimToady like that
it's supposed to be able to turn all the leaf nodes into arrays too, but it can't do that yet 01:35
uvtc r: my @a = ('a', 'b', 'c'); say @a.WHAT; say ('a', 'b','c').WHAT;
p6eval rakudo b10e2b: OUTPUT«(Array)␤(Parcel)␤»
TimToady as someone was mentioning, assignment is not binding
r: my @a := ('a', 'b', 'c'); say @a.WHAT; say ('a', 'b','c').WHAT; 01:36
p6eval rakudo b10e2b: OUTPUT«(Parcel)␤(Parcel)␤»
uvtc That sounds deep. To me, assignment *is* binding.
TimToady no, binding is assignment of references rather than values
do you know C
TimToady ? 01:36
uvtc Oh, yes. Right. A container is involved.
geekosaur roughly: in perl 6, assignment assigns into a container bound to a name; binding binds a container to a name
TimToady assignment is *destination = *source, while binding is destination = source 01:37
uvtc So `:=` is the binding operator, correct? It binds the parcel ('a', 'b', 'c') to the name `@a`?
TimToady yes
whereas assignment copies the values into the container
uvtc I used C years ago.
TimToady that is initialized to be an Array because of the @
it's the list assignment to the array that, er, turns it into an Array 01:38
r: my $item = (1,2,3); say $item.WHAT
p6eval rakudo b10e2b: OUTPUT«(Parcel)␤»
uvtc I'd thought "binding" was just a generic term to mean "this name refers to this cubby hole (memory)".
TimToady assignment to a scalar container doesn't do list assignment, so it just copies in the parcel ref
yes, but when you do assignment, you aren't doing that, unless you happen to be assigning a reference 01:39
in particular, list assignment doesn't do that
it copies the values of the list into the (existing) array
uvtc steeples fingers in deep thought... 01:40
TimToady mostly, assignment is there to preserve the sense of value semantics that a Perl 5 programmer wants
binding is what a Python programmer is taught to expect
(when they say =)
if a programmer wants to program pythonically in p6, they just need to use $vars and := consistently 01:42
then no copying of anything except pointers ever happens without an explicit .clone 01:43
basically, = is value semantics rather than pointer semantics, how your C program manages to actually set an int location in memory to 42 01:44
(but as in C, sometimes pointers can be treated as native values too)
that's what happened when I assigned a parcel to $item
it ignored the ref-ness of it and just poked it into the scalar container currently bound to $item 01:45
however, since all scalar containers automatically deref that for you, you can think of it as just putting the object into the variable
(if you admit that an object can be in more than one place at the same time) 01:46
which is how value objects work anyway; 42 is the same value in $foo as in $bar
TimToady it's just that normal objects take their identity from their location rather than their value 01:47
so when you copy around object identities, you're just copying pointers to locations, really
most of the time all this referencing and dereferencing 'cancels out' so you can just ignore it 01:48
uvtc `my $x = ...` and `my @a = ...` make containers and then copy values into them. `=` is "value semantics" (like C and Perl 5 (when not using references in Perl 5)). 01:58
uvtc is still here. Thanks so much for your help, TimToady. 01:59
TimToady also like refs in P5, except in P6 you never have to use \ to make a ref explicitly
since everything makes a ref automatically
r: my $ref = &die; &die.("Something to die for") 02:00
p6eval rakudo b10e2b: OUTPUT«Something to die for␤ in block at /tmp/xvq2Kw6sVs:1␤␤»
TimToady &die creates a ref, and .() automatically derefs it
@foo creats a ref, and .[] derefs it
%foo creates a ref and .{} derefs it 02:01
so mostly we just got rid of the backslashes in P6
assignment stayed much the same, and we added binding so we didn't have to play with typeglobs anymore ever again
[Coke] r: my \atnight = "mostly"; say atnight; 02:02
p6eval rakudo b10e2b: OUTPUT«mostly␤»
TimToady that's a pseudoassignment really, and notices that it's really supposed to bind
uvtc Does the `:=` explicitly cause a ref to be taken of the rhs? 02:03
TimToady nope
it merely treats it differently on the left
the RHS has no clue it's inside a := or a =
it just does its think of always making a ref to something 02:04
*thing
uvtc Sounds like we're saying the same thing: `:=` makes sure a ref is assigned to the LHS. 02:04
TimToady was just quibbling with "cause a ref to be taken", when it just uses the ref that's there, it doesn't cause it, except insofar as it simply evaluates the RHS 02:05
uvtc TimToady: re `my $item = (1,2,3);` above, you wrote, "assignment to a scalar container doesn't do list assignment, so it just copies in the parcel ref". But, why is ref copying happening ("binding"), if the assignment operator is being used? 02:11
TimToady it's copying the ref as a value into the container, it's not replacing the container as := would 02:12
r: my $item = (1,2,3); $item = 42; say $item 02:13
p6eval rakudo b10e2b: OUTPUT«42␤»
TimToady r: my $item := (1,2,3); $item = 42; say $item
p6eval rakudo b10e2b: OUTPUT«Cannot modify an immutable value␤ in method STORE at src/gen/CORE.setting:5186␤ in block at /tmp/eqxv8WxF7A:1␤␤»
uvtc `:=` replaces the container of the object on the RHS with that of the LHS?
TimToady yousaid that backwards
uvtc ! :)
`:=` takes what's on the RHS and puts it into the container on the LHS?
Oh, that's worse. 02:14
`:=` takes what's in the container on the RHS and puts it into the container on the LHS?
TimToady no 02:14
:= merely changes the meaning of the name on the left to refer to whatever reference there was on the right 02:15
there isn't necessarily a container on either side
(except insofar as a name can refer to something else)
uvtc Would you say, "$x, @y, and %z are names which refer to containers"? 02:16
TimToady by default, yes
unless you bind something else to them
uvtc And then what would you call them? 02:17
(instead of names which refer to containers)
TimToady when you say: my @y; it's equivalent to my @y := Array.new;
uvtc Makes sense. 02:17
TimToady the @ merely claims that the objects supports .[] style subscripting
TimToady and @ and % tend to "autovivify" whatever it is you expect to be there 02:18
uvtc Oh, and when you use `:=`, you're saying, "don't create a new object, just binding to <this> one"
TimToady an array or hash in this case
yes
uvtc s/binding/bind/
TimToady name -> container -> value(s) is the default, but you can override it to name -> whateveryoulike 02:19
uvtc Thanks so much for the help, TimToady. I'm going to let this stew simmer just a bit longer. Which Synopsis would you say covers this particular topic? 02:20
TimToady++ # sitting down at the chalkboard after hours 02:21
TimToady um, that topic tends to be distributed over many synopses
anyway, you're welcome--I need to figure out the best way to explain it in a book soon in any case
uvtc Very well. {tips hat} :) 02:22
mjreed perl6: sub fib($n) { state @fib := 0,1,*+*...*; @fib[$n]; }; fib(0); fib(0); 02:33
p6eval niecza v24-24-gbdc3343: OUTPUT«===SORRY!===␤␤Cannot use bind operator with this LHS at /tmp/PUCoYjKDaH line 1:␤------> sub fib($n) { state @fib := 0,1,*+*...*⏏; @fib[$n]; }; fib(0); fib(0);␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza…
..rakudo b10e2b: ( no output )
mjreed sub fib($n) { state @fib := 0,1,*+*...*; @fib[$n]; }; fib(0); say fib(0);
perl6: sub fib($n) { state @fib := 0,1,*+*...*; @fib[$n]; }; fib(0); say fib(0); 02:34
p6eval rakudo b10e2b: OUTPUT«(Any)␤»
..niecza v24-24-gbdc3343: OUTPUT«===SORRY!===␤␤Cannot use bind operator with this LHS at /tmp/AW4jBzc4RL line 1:␤------> sub fib($n) { state @fib := 0,1,*+*...*⏏; @fib[$n]; }; fib(0); say fib(0);␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/ni…
[Coke] aw, panda install NativeCall barfs. 02:43
is NativeCall expected to work on OS X?
mjreed binding `state` variables - valid per the spec? current rakudo allows it but the binding doesn't actually persist.. 02:46
skids r: sub OHAI0 { "OHAI".say; 0 }; sub fib($n) { state @s = (OHAI0(),1,*+* ... *); @s[$n] }; fib(13).say; fib(8).say # is the binding needed? 02:51
p6eval rakudo b10e2b: OUTPUT«OHAI␤233␤21␤»
skids (S04 'state automatically applies "start" semantics to any initializer') 03:03
mjreed hm. thought that assignment would evaluate the list, and therefore was a no-no for infinite ones. 03:08
colomon mjreed: I believe Rakudo is doing some sort of magic to make infinite list assignment work *if* it knows the list is infinite. 03:10
[Coke] r: (say 1,1,*+*...*) 03:11
p6eval rakudo b10e2b: OUTPUT«1 1 2 3 ...␤»
skids What magic is required seems to be specced out in S03 at the end of the sequence operator text. 03:17
r: sub OHAI0 { "OHAI".say; 0 }; sub fib($n) { START { OHAI0(),1,*+* ... * }[$n] }; fib(13).say; fib(8).say; 03:20
p6eval rakudo b10e2b: OUTPUT«OHAI␤233␤21␤»
TimToady perl6: sub fib($n) { constant @fib = 0,1,*+*...*; @fib[$n]; }; fib(0); say fib(0); 03:23
p6eval rakudo b10e2b, niecza v24-24-gbdc3343: OUTPUT«0␤»
TimToady perl6: sub fib($n) { constant @fib = 0,1,*+*...*; @fib[$n]; }; say fib(^10); say fib(^10);
p6eval rakudo b10e2b, niecza v24-24-gbdc3343: OUTPUT«55␤55␤»
skids
.oO(It's all fun and games until someone loses half their RAM to cached fib values)
 03:24
TimToady perl6: sub fib(\n) { constant @fib = 0,1,*+*...*; @fib[n]; }; say fib(^10); say fib(^10);
p6eval rakudo b10e2b, niecza v24-24-gbdc3343: OUTPUT«0 1 1 2 3 5 8 13 21 34␤0 1 1 2 3 5 8 13 21 34␤»
TimToady skids: in which case the compiler can notice it's a constant, and deduce that it could reconstruct the values if it needed to :) 03:25
TimToady can see that it will take some persuading to get people to think of an infinite series as a constant... 03:27
TimToady it's just a very Large constant 03:27
s/series/sequence/ 03:28
skids Not really that difficult once you teach them that code can be a constant. 03:31
diakopter all deterministic pure code is a constant... with a very big range 03:35
japhb timotimo, Some time ago I believe you pasted this link: github.com/timo/iperl6kernel/blob/...l.nqp#L156 ... it occurs to me that lines 153..157 put @ARGS into the wrong order. 03:38
I'm guessing you instead wanted just: @ARGS.unshift('-MIPerl6::ZMQ'); @ARGS.unshift('-Ilib'); 03:39
Though come to think of it, this is NQP, so I guess you do still need the $pname movement. So really, just swapping lines 155 and 156 ought to correct things. 03:41
TimToady rosettacode.org/wiki/File_modificat...ime#Perl_6 04:44
labster TimToady++ Shouldn't that be in a module? 04:50
TimToady well, eventually; at the moment it's in the category of difficult but possible 04:53
TimToady and I'm trying to pass up Mathematica :) 04:58
3 to go...
searcher-perl hi all 05:35
diakopter hi 05:36
searcher-perl I study perl6 again 05:37
But I read parrot has gone
Is that trueX
?
diakopter parrot hasn't gone
rakudo has been transitioning to use multiple backends/VMs for years now 05:38
searcher-perl I know that
diakopter ok :)
searcher-perl When I use perl6 binary, it's too late. So I searched some news for perl6. 05:39
Someone said Parrot will or may stop. 05:40
diakopter well, all the active developers are otherwise occupied, generally
sorear why do you care? what are you trying to do? 05:41
searcher-perl I wonder I can use perl6 for production program 05:42
diakopter depends for what purpose
it's sufficient for some things; not ready for other things 05:43
searcher-perl sorry, I'm not English native speaker. So I can't understand nuiance very well 05:44
diakopter okay. What kind of production program do you want to try? 05:45
searcher-perl I knew about parrat long time ago, so now I'm thinking parrot can substitute Java. 05:47
perl6 was too late so something was wrong. 05:48
geekosaur that's one of the active projects, yes
diakopter the JVM is one the VMs to which rakudo is being ported
geekosaur mostly what was wrong was people got so busy working on features and forgot to tell the world they were still there... 05:49
sorear searcher-perl: "production" is too vague to say anything. What, specifically, do you want to use perl 6 for?
searcher-perl I mean general purpose like perl5. 05:50
perl6 is very late than my expect. 05:51
diakopter your expectation were set incorrectly :)
searcher-perl I studied groovy. It's later than perl5 very much. perl6 is later than groovy. 05:54
So I want to ask something about perl6. It's fit for general purpose language now. 05:56
diakopter searcher-perl: most of the implementations are Turing complete, so... 05:59
searcher-perl visual basic also say so :) 06:00
I think perl6 is enough then study perl6 or use perl. 06:02
of course v5.
TimToady study many languages, including p6 :) 06:03
searcher-perl Thank you. I thought perl6 is complete. But still has long way. 06:04
sorear do we have any regulars from .kr?
searcher-perl I met Seoul.pm members in 2008.
But I don't know about them. 06:05
TimToady well, you are very welcome here
searcher-perl Thank you. They say about perl5 not so much perl6. I don't know why. 06:06
TimToady perl5 started earlier :) 06:07
but perl6 will end later
searcher-perl Good.
I learned perl in 2002.
TimToady you will like Perl 6 even better 06:08
when we get it to run faster, it will become a lot of people's favorite language
searcher-perl Grammar is good, it's more like functional. 06:09
TimToady more functional, and more OO too
searcher-perl true.
TimToady and quite a bit more powerful
searcher-perl Perl5 has some obfuscating things.
TimToady we've fixed most of those things 06:10
searcher-perl So I consider Perl6.
TimToady nr: say "Welcome searcher-perl!" for ^10;
p6eval rakudo b10e2b, niecza v24-24-gbdc3343: OUTPUT«Welcome searcher-perl!␤Welcome searcher-perl!␤Welcome searcher-perl!␤Welcome searcher-perl!␤Welcome searcher-perl!␤Welcome searcher-perl!␤Welcome searcher-perl!␤Welcome searcher-perl!␤Welcome searcher-perl!␤Welcome searcher-perl!␤»
TimToady see, both rakudo and niecza agree that you are welcome 06:11
searcher-perl Time to learn again perl6. But too late as I thought. Interating.
what is nieczaX
?
TimToady it's an implementation that runs on .NET or mono 06:12
I use it in production :)
(because it's faster than rakudo, so far)
searcher-perl Sorry, Microsoft platform is not my destiny.
TimToady mono is not on Microsoft platform :) 06:13
searcher-perl Is that soX
TimToady well, you can soon run on JVM instead if you like
diakopter it does have decent chunks of MS-PL code
searcher-perl mono is porting of .Net.
diakopter not really porting; more like clean-room reimplementation
TimToady yes, but just because it has the same API does not mean that it belongs to Microsoft 06:14
searcher-perl No licence problem?
TimToady anyway, Perl 6 will run on many platforms in the future
so you will be able to pick
rakudo: < mono .NET JVM Parrot llvm Haskell Go C>.pick.say 06:15
p6eval rakudo b10e2b: OUTPUT«Go␤»
TimToady ooh, it picks Go :)
Timbus rakudo on llvm would be amazing 06:16
you could call it lakudo
searcher-perl :) 06:17
TimToady llakudo, surely 06:20
searcher-perl Is it possible socket, db in perl6?
TimToady yes, though it is not as developed as p5 yet 06:21
searcher-perl How can it be connect to C library like xs in perl5? 06:22
TimToady currently through the NativeCall interface
for an example, see rosettacode.org/wiki/Call_a_functio...ary#Perl_6 06:23
searcher-perl good. 06:24
Difficult. 06:25
TimToady people will write modules that make it easy
searcher-perl Ok. I'll read perl6 book first. Thank you for your kind answers. Bye! 06:32
grondilu yeah! I'm now officially a perl monk: perlmonks.org/?node_id=946047 06:38
FROGGS wants to do Gokudo soon 07:38
mathw: you are the one who wants to port NQP to Go, right? 07:46
diakopter swarley was working on it last week
moritz wouldn't that be Ragodu?
FROGGS ahh, k, thanks
hehe, ragodu .oO( ragodü? ) sounds like a saxonian name to me :P 07:47
FROGGS awww ó.ò 08:34
gccgo test.go
test.go:21:16: Fehler: import file »./src/sixModel« not found
moritz swarley: ^^ seems you forgot to commit a file 08:37
tadzik RaGodost maybe :) 08:43
hello #perl6
moritz \o tadzik 08:44
sorear sounds like a slavik adjective 08:45
sorear glasnost, blizkost 08:45
or was that the point
hello tadzost
FROGGS is that morally reprehensible if I have to think of blitzkrieg now? 08:48
FROGGS .oO( RATZKRATZNFATZ! )
sorear blizkost means 'togetherness, closeness' IIRC 08:49
not related to Blitzkrieg
sorear Blitz referring to speed, or lightning specifically 08:49
Kugelblitz
FROGGS I know, it is a german word :o) 08:50
sorear ... now that I think about it, aren't you kind of a person who speaks both german and polish
why am I telling you about those languages
I must be tired 08:51
FROGGS I dont speak polish, older east-germans usually speak +russian, older west-germans usually +french, junger germans mostly +english and +french 08:53
moritz s/junger/younger/ :-)
FROGGS IMO polish ppl speak german and english too, but germans dont tend to learn polish language 08:54
thanks moritz
sorear FROGGS: I thought you were specically polish, living in germany
FROGGS when creating another sentences in my head I tend to write german words ó.ò
sorear: no
nwc10 straw poll of one day visiting Bratislava - older people we interacted with were more comfortable with German. Younger with English. 08:55
sorear was everyone comfortable with Slovak, though? :) 08:56
FROGGS in school we had russian language for 3 years or so but I only can read the letters now, we had english for seven years, but no polish or other languages
sorear FROGGS: was this before or after reunification 08:57
moritz I think there's a deep racist current against eastern european countries in the west of germany
FROGGS sorear: after 08:58
sorear interesting. did not expect that
is russia considered part of eastern europe, or its own thing
FROGGS I went to school (with age of seven) 1989 08:59
sorear ooh, exciting times
moritz the fall of the Berlin wall is the first political event that I remember 09:00
FROGGS moritz: it is not only in the west of germany, since generations ago the eastern ppl had "contact" with russian soldiers... and these impression stay long
moritz: I dont even remember that 09:01
moritz though at that time I didn't quite understand why my parents were so excited, and how one country can be two countries or not
sorear moritz: where were you living at the time? (west, east exclusive of berlin, east berlin, west berlin, other) 09:03
moritz sorear: west 09:04
sorear: though my faster came from the east, and "lost" his father by the division
sorear erlangen? 09:05
FROGGS I lived in thuringia, pretty in the center of germany, but still "east"
moritz yes
nwc10 sorear: we didn't speak any Slovak, so we couldn't test this :-) 09:06
jnthn morning, #perl6 o/ 09:21
FROGGS hi jnthn
jnthn nwc10: dobre rano # two words you can test with next time ;)
FROGGS guten morgen? 09:22
err, good morning?
jnthn yes :)
FROGGS dobre, means good at least
:o)
jnthn despairs at how quickly language knowledge rust
*rusts 09:23
nwc10 Pivo!
FROGGS nqp: say("[" ~ ("ab123c" ~~ /<-[\w]>+/) ~ "]") 09:24
p6eval nqp: OUTPUT«[]␤»
FROGGS nqp: say("[" ~ ("ab123c" ~~ /<-[abc]>+/) ~ "]")
p6eval nqp: OUTPUT«[123]␤»
FROGGS -.-
moritz \w matches digits too
jnthn nwc10: That's a useful word :)
FROGGS ahh, of course
thanks, that saves me hours >.<
nwc10 jnthn: visit Bratislava. Or visit Vienna, and we can make a day trip there
jnthn should do that sometime :) 09:25
nwc10 currently the car actually has enough fuel to get to Croatia 09:27
FROGGS nwc10: and to get back too?
nwc10 no. 09:27
but in Croatia, they took more credit cards than I was aware existed
FROGGS: actually, might be able to make a round trip to the nearest part of Croatia 09:28
nwc10 but the car thinks its range is about 850km when full 09:28
we've not tried driving slowly to see if we can up that number. :-)
dalek p: 906d7fa | jnthn++ | src/HLL/Compiler.pm:
Remove some dead code.

Couldn't possibly have worked, since it calls a method that doesn't exist.
p: a93281b | jnthn++ | src/HLL/Compiler.pm:
Elimiante Parrot-hash-specific .update.
FROGGS I once was able to drive more than 1100km with a renault trafic, but then you can't drive faster than 80-100km/h 09:29
nwc10 anyway, we're out of pelinkovac and pruscht, and low on olive oil and olives, so we have a shopping list 09:30
Ulti just tried building R* on my windows machine, I get a make error just after parrot gets compiled 09:54
Command failed (status 65280): C:\Dwimperl\perl\bin\perl.exe Configure.pl --with-parrot=C:/rakudo-star/install/bin/parrot.exe --make-install 09:55
Ulti I dont have ICU in the build path atm so I need to do that anyway 09:55
moritz Ulti: it would help if you could nopaste output above that error 09:56
Ulti k 09:57
just running that command I get that it doesn't like the --with-parrot flag
nwc10 jnthn: still works on my machine 09:58
Ulti oh man cmd.exe is so lame 09:58
daxim kill it with fire, then replace it 09:59
stackoverflow.com/questions/440269/...ws-console
FROGGS daxim: well, if you want to test things on windows you dont have many choices 10:00
daxim I recently tried a few, and they all suck compared with konsole, which isn't available for windows
FROGGS what about powershell? never used it though
daxim CMIIW, but isn't powershell just a shell? 10:01
BinGOs shell and .NET-ish scripting environment 10:05
Ulti daxim: thanks for the link
sorear reminder that cmd.exe and conhost.exe are different things 10:11
last time I was on windows I used the version of bash.exe that came with msysgit. it's good 10:12
Ulti nopaste.info/9e0bbcac4e.html <-- output from perl Configure.pl --gen-parrot
sorear conhost is icky but it's much harder to replace, it's somewhat more integrated with system
it can probably be done but I haven't tried
moritz Ulti: you got a warning "Warning: Building a shared parrot library may conflict with your previously-installed C:rakudobinlibparrot.dll " 10:13
Ulti: maybe remove that, and try to buld nqp again? 10:14
cd nqp
gmake install
Ulti k
sorear C:rakudobinlibparrot.dll is kind of red flaggy
timotimo somebody used backslashes without escaping them or osmething? 10:26
dalek p: 1c6c851 | jnthn++ | src/ (3 files):
Eliminate use of Parrot dumper.

We didn't use it for --target=ast, but --target=parse relied on it. The output was incredibly verbose, so replace it with something that does not rely on the Parrot dumper and gives more compact, and hopefully more useful, output. Tweaks welcome.
kresike hello all you happy perl6 people 10:27
Ulti hmmm its flaking out on compiling libtommath or at least between there and nqp_group.c
also I think the previous problem was that I had R* installed elsewhere so parrot was in my path I assume 10:28
I should probably head to day jerb now, sorry :( I'll try properly later tonight to build a .msi as my gf is away playing with clay for three hours... 10:30
timotimo r: use Test::More; throws_like q{class Foo { method bar { Quux.parse('OH HAI'); } }; grammar Quux { rule TOP { .* } }; Foo.bar;}, X::Undeclared::Symbols, post_types => "Quux" => *; 10:32
p6eval rakudo b10e2b: OUTPUT«===SORRY!===␤Could not find Test::More in any of: /home/p6eval/nom-inst/lib/parrot/4.10.0-devel/languages/perl6/site/lib, /home/p6eval/nom-inst/lib/parrot/4.10.0-devel/languages/perl6/vendor/lib, /home/p6eval/nom-inst/lib/parrot/4.10.0-devel/languages/perl6/lib, /h…
timotimo oh 10:33
star: use Test::Util; throws_like q{class Foo { method bar { Quux.parse('OH HAI'); } }; grammar Quux { rule TOP { .* } }; Foo.bar;}, X::Undeclared::Symbols, post_types => "Quux" => *;
p6eval star 2013.02: OUTPUT«===SORRY!===␤Could not find Test::Util in any of: /home/p6eval/star/lib/parrot/4.10.0/languages/perl6/site/lib, /home/p6eval/star/lib/parrot/4.10.0/languages/perl6/vendor/lib, /home/p6eval/star/lib/parrot/4.10.0/languages/perl6/lib, /home/p6eval/.perl6/2013.02.1/lib…
FROGGS it is not part of anything ó.ò
timotimo right, this is not going to work
so, uh, the problem i'm getting is that i get a SORRY outputted from that eval and the $! seems to be Nil and i'm confused. 10:34
moritz timotimo: first, put the string declaration outside the throws_like, and print it, and see if it's what you expected 10:35
second, try some parens in post_types => "Quux" => *; 10:36
mberends does p6 LWP::Simple support https? Here on Linux it dies "could not parse headers" in parse_response in Simple.pm. cosimo? anyone? It's a total blocker :( 10:51
cosimo mberends: it doesn't afaik 10:54
mberends cosimo: thanks 10:55
mberends thinks about an ugly workaround shelling wget on Windows
timotimo moritz: haha, get this: i was adding this to the S32-excepitons/misc.t and moving it all the way up makes it work. looks like the evals were all in the same scope and there were name clashes or something 11:00
dalek ast: 7d2d272 | (Timo Paulssen)++ | S32-exceptions/misc.t:
add a test for RT #69760
 11:03
nwc10 jnthn: still "works" on "my" machine
jnthn nwc10: I've no "beef" with that. 11:06
timotimo may i change the gistification or whatever it is in the Whatever object? i get ok 521 - .post_types matches Quux => Whatever<-4272955866496737429> for one of my tests and likely that number is going to change every test run, so it will be in each and every diff in roast-data :( 11:07
FROGGS change it to what? 11:08
jnthn r: say *
p6eval rakudo b10e2b: OUTPUT«*␤»
jnthn r: say *.Str
p6eval rakudo b10e2b: OUTPUT«WhateverCode.new()␤»
jnthn r: say (*).Str
p6eval rakudo b10e2b: OUTPUT«WhateverCode.new()␤»
timotimo oh, weird. apparently it's the fault of dies_like 11:09
jnthn I think it's more likely to be Str rather than gist, fwiw
timotimo doesn't need to be fixed right away anyway 11:12
moritz timotimo: just don't do a string comparison 11:14
the right of the => is used for smart-matching 11:15
so you can write post_types => { .value ~~ Whatever } or so
timotimo not sure what type exactly that is, but i can look. it's the line information from the exception
oh, it's just a list of line numbers 11:16
FROGGS nqp: say("[" ~ ("ab123c" ~~ /<-[a]+[b]>+/) ~ "]") 11:19
p6eval nqp: OUTPUT«[b123c]␤»
FROGGS nqp: say("[" ~ ("ab123c" ~~ /<-[b]>+/) ~ "]")
p6eval nqp: OUTPUT«[a]␤»
FROGGS does that make sense? 11:20
dalek kudo/nom: c2910e2 | jnthn++ | src/Perl6/ (4 files):
Be sure to use line info cache.
p: 6613df0 | jnthn++ | src/ (2 files):
Move compilation unit handling to backend class.

This should also help make the addition of --target=pbc a bit cleaner.
p: 9acc219 | jnthn++ | src/HLL/Compiler.pm:
Eliminate getprop/setprop usage.
timotimo oh, my test match is actually borked 11:21
moritz is this actually something you want to test? 11:22
timotimo # Got: PostDeclaredGrammar 1 / # Expected: PostDeclaredGrammar 1 11:23
great, i love this kind of error :(
moritz :(
timotimo what i would like to test is that "PostDeclaredGrammar" shows up as a key in the %.post_types hash 11:24
currently i'm trying to match against post_types => (PostDeclaredGrammar => 1).hash or without .hash 11:25
moritz pass a closure and do say .perl; in the closure, for debugging 11:26
timotimo oh, that sounds like a good idea 11:27
didn't know i could do that
Method 'perl' not found for invocant of class 'Integer' - er ... ? 11:28
dalek p: 3b3d8bb | jnthn++ | src/QAST/Operations.nqp:
Add nqp::getcodecuid and nqp::forceouterctx.
 11:30
p: f5648cd | jnthn++ | src/stage0/ (9 files):
Update bootstrap to get added nqp:: ops.
p: 0968572 | jnthn++ | src/ (2 files):
Use new nqp:: ops in HLL::Compiler.

With this, it should now be fairly generic, with all VM-specifics in either the backend object or behind nqp:: ops.
FROGGS n: say("[" ~ ("ab123c" ~~ /<-[a]+[b]>+/) ~ "]") 11:38
p6eval niecza v24-24-gbdc3343: OUTPUT«[b123c]␤»
FROGGS okay, then maybe it should be "everything but a, plus b
" 11:39
jnthn iiuc, when you start with - you're subtracting from everything 11:40
rn: say("[" ~ ("ab123c" ~~ /<-[ab]+[b]>+/) ~ "]")
FROGGS k, well, makes sense somehow
p6eval rakudo b10e2b, niecza v24-24-gbdc3343: OUTPUT«[b123c]␤»
shinobicl rakudo: my @a = ('A' ... 'N'), 'Ñ' ,( 'O' ... 'Z'); for @a {print $_} 12:35
p6eval rakudo c2910e: OUTPUT«ABCDEFGHIJKLMNÑOPQRSTUVWXYZ»
shinobicl rakudo: my @a = <A .. N>, 'Ñ' ,<O .. Z>; for @a {print $_}
p6eval rakudo c2910e: OUTPUT«A..NÑO..Z»
shinobicl rakudo: my @a = <A ... N>, 'Ñ' ,<O ... Z>; for @a {print $_}
p6eval rakudo c2910e: OUTPUT«A...NÑO...Z» 12:36
jnthn Note .. is sufficient in this case 12:37
And you can probably drop the parens if you usethat
shinobicl and how i do it with < and > ?
< and > does interpolate too?
jnthn No 12:38
shinobicl thanks
jnthn << and >> do, but not ranges, afaik
dalek p-jvm-prep: d6a55c4 | jnthn++ | / (9 files):
Implement the various nqp:: IO ops.
 13:24
p-jvm-prep: bc0b098 | jnthn++ | nqp-src/NQPCORE.setting:
Include I/O subs in the setting.
p-jvm-prep: 084eb8b | jnthn++ | nqp-jvm-cc.nqp:
Chase latest HLL backend API.
p-jvm-prep: b52d132 | jnthn++ | t/nqp/63-slurp.t:
Now pass 63-slurp.t.
arnsholt japhb: Looking at the most-wanted, we have basic DB interaction via github.com/perl6/DBIish 13:31
Mostly moritz++'s work I think 13:32
moritz and I mostly build on the work of mberends++ and arnsholt++ 13:32
and jnthn++
arnsholt Oh, and I see that you want unsigned native types. I'm not entirely sure how we'd do that, since both JVM (IIRC) and Parrot don't really have those concepts 13:39
moritz newer java verisions have support for unsigned operations 13:41
arnsholt Ah, spiffy
moritz and if we nicely ask the parrot folks, maybe they add some too
arnsholt Troo, troo 13:42
dalek p-jvm-prep: fb2d288 | jnthn++ | nqp-src/QRegex.nqp:
Uncomment !reduce methods.
 14:45
p-jvm-prep: 37e267c | jnthn++ | / (2 files):
Support long string literals.

Turns out the serialization blob hit the limit.
p-jvm-prep: 6b1059f | jnthn++ | nqp-src/NQPHLL.pm:
Add HLL::Compiler and partial HLL::Backend.

HLL::Compiler has just two lines commented out; HLL::Backend has some methods that need implementing.
p-jvm-prep: dcdf4a1 | jnthn++ | / (2 files):
Start cross-compiling the P6 regex compiler.
[Coke] Is there a sixier way to write this: 14:56
hoelzro haha
"sixier" 14:57
I love it
[Coke] (@a.pairs.sort:{$^a.value <=> $^b.value}).map:{$^a.key} # return positional keys in order by ascending value
colomon +*.value should work with the sort, right? 14:58
moritz you can even do @a.pairs.sort(+*.value).map: *.key 14:58
or @a.pairs.sort(+*.value)>>.key
drKreso Hi I am having problems with Bailador
moritz colomon++ was faster :-)
hi drKreso
[Coke] moritz: is >>. guaranteeing ordering?
drKreso moritz: hi 14:59
moritz [Coke]: return order, yes.
drKreso ./panda/bin/panda install Bailador ==> t/03-response-content.t .. 2/7 Not enough positional parameters passed; got 0 but expected 1
moritz [Coke]: not execution order though
drKreso: looks like github.com/tadzik/Bailador/issues/5
drKreso: I'm not very familiar with Bailador, so I don't know if that's a real problem or not
drKreso can i skip it somehow? 15:00
moritz drKreso: I think you can do a panda --notest install Bailador or so
nwc10 jnthn: the usual. 15:02
drKreso moritz: Thanks. 15:03
colomon wow, Bailador has a lot of dependencies 15:08
hmmm, all tests pass for me.
drKreso Maybe I have older rakudo 15:10
moritz I doubt it; the first report of that issue is with a pretty new rakudo (2013.02.1 release) 15:12
colomon hmmm, that's a bit newer than mine, maybe. update rakudo...
drKreso I have it installed, I can see Bailador in lib folder... but how to run "site" wizard something like "dancer -a MyWeb::App" 15:15
moritz drKreso: there doesn't seem to be one yet 15:16
drKreso Hm ,so how do I use it? 15:17
mberends regarding an improved DBI for Perl 6, I am very impressed with Java Data Objects (JDO). The abstractions are very well thought out and portably cover both relational and non relational stores (eg Google AppEngine). The code is implemented as Aspect Oriented Programming which Perl 6 could also do very nicely. 15:18
arnsholt Cool! Have you mentioned this to tbunce? 15:19
mberends No, haven't contacted him for a few months. He is also interested in making Perl's database object hierarchies analogous to Java's but he was thinking of only JDBC so far. 15:21
arnsholt Yeah, mostly JDBC was my impression as well 15:24
moritz drKreso: you look at how the examples work, and at the documentation, and go from there 15:27
colomon 's Linux box shut down compiling rakudo again. :( 15:28
hoelzro =( 15:29
colomon the new box cannot get here soon enough. 15:30
moritz colomon: happens my laptop too :( 15:31
and ENOFUNDSFORNEWONE 15:32
drKreso moritz: Thanks.
dalek p-jvm-prep: 85bea4b | jnthn++ | / (2 files):
Implement nqp::getcodecuid.
 15:43
p-jvm-prep: 1bd3f65 | jnthn++ | nqp-src/NQPCORE.setting:
Fix thinkos.
p-jvm-prep: d1de3ee | jnthn++ | src/org/perl6/nqp/runtime/ (2 files):
Implement providing argv to the entry point.
p-jvm-prep: e58be0c | jnthn++ | src/org/perl6/nqp/runtime/Ops.java:
Fix index out of range SC bug.
FROGGS home sweet home \o/ 16:04
sergot hi o/ 16:08
FROGGS hi 16:08
nwc10 jnthn: breaks on my machine :-(
pasta.test-smoke.org/451 16:09
looks to be all the same: Exception in thread "main" java.lang.RuntimeE
xception: Wrong number of arguments passed; expected 0..0, but got 1
at org.perl6.nqp.runtime.Ops.checkarity(Ops.java:722)
jnthn nwc10: Does nqptest work, just not test? 16:10
nwc10 looks like it 16:11
pass-so-far
[Coke] masak? 16:13
nwc10 yes, passes all but the test that needs ICU 16:14
dalek p-jvm-prep: be4c3da | jnthn++ | src/org/perl6/nqp/sixmodel/reprs/ContextRefInstance.java:
Implement existskey in ContextRef.
 16:21
p-jvm-prep: 3651628 | jnthn++ | src/org/perl6/nqp/sixmodel/KnowHOWMethods.java:
Ensure KnowHOWs get a .WHO set up.
p-jvm-prep: 923bfe8 | jnthn++ | / (4 files):
Make nqp::list and nqp::hash consider HLL config.
p-jvm-prep: 0d59dca | jnthn++ | nqp-src/NQPCORE.setting:
Make sure .push etc. always work on an @foo.
p-jvm-prep: 76a5199 | jnthn++ | nqp-src/QRegex.nqp:
Sync with nqp QRegex (for %!marks).
jnthn nwc10: None of those fix the busted make test...I'll get there in a bit. 16:26
colomon huh. updated my rakudo, and now I do get the failure in Bailador. Must be a very recent change. 17:17
has there been a recent change to Match.list? 17:22
japhb arnsholt: I did link DBIish in most-wanted, as the WIP for the DBI-like interface style. 17:30
*did already
moritz the most recent change in src/core/Match.pm was 2012-10-10
japhb arnsholt, mberends, is Tim Bunce still interested in working on DB modules for Perl 6? I thought he'd gone off to other pastures .... 17:32
mberends japhb: I think he was disappointed by the orders-of-magnitude difference in efficiency between p6 implementations and p5 (he wrote metacpan.org/module/Devel::NYTProf) and seems to want to provide overall direction to volunteers who will the heavy lifting. 17:40
*who will do
japhb Ah, but he is still interested in having at least that much connection to it still ... good. 17:41
mberends Tim Bunce sent a very upbeat message to the Perl Reunification Summit (mdk.per.ly/2012/08/20/prs2012-perl5...on-summit/ and blogs.perl.org/users/liz/2012/08/as...-may.html) because he regretted being unable to attend. 17:45
PerlJam you could always email him 17:51
or talk to him in real time on #p5p
mberends oh. silly /me 17:52
PerlJam (well, he's not active on #p5p just now, but he was about 19 hours ago) 17:56
diakopter is that magnet or freenode 17:57
PerlJam magnet 17:58
diakopter or twitter hashtag >.<
moritz hopes that some engineer at twitter changes all HASH tags to their sha1 hash on April 1st 18:00
jnthn Talking of hash... 18:01
jnthn just discovered an accidental reliance on hash ordering... :/
Hopefully easy to fix.
In the NQP MOP of all places...
diakopter mop it up... 18:02
arnsholt japhb: Yeah, I noticed afterwards. I was looking at library wrappers first, and then modules =) 18:07
mberends Talking of hash... jnthn++ is already handling all this well, it's a good explanation of the pitfalls of hashes in language implementation: blog.headius.com/2012/09/avoiding-h...-ruby.html (posted by scrottie in magnet:perl11 irclog.perlgeek.de/perl11/2012-11-08#i_6137030) 18:14
moritz mberends: the interesting thing about your link is that it talks (among other things) about invalidating method caches of subclasses, which Rakudo doesn't do yet 18:25
and which is why this example here fails: 18:26
r: use MONKEY_TYPING; augment class Any { method flurb() { } }; "foo".flurb()
p6eval rakudo c2910e: OUTPUT«No such method 'flurb' for invocant of type 'Str'␤ in block at /tmp/2jILNgbo21:1␤␤»
moritz r: say Str ~~ Any
p6eval rakudo c2910e: OUTPUT«True␤»
arnsholt timotimo: Possibly relevant to your interests: github.com/arnsholt/Net-ZMQ/commit...801ef220eb 18:43
arnsholt timotimo++ # figuring out how to do basic testing 18:44
dalek p: 361feca | jnthn++ | src/how/NQP (3 files):
Attribute ordering matters; don't use a hash.
 18:46
p-jvm-prep: 4b71902 | jnthn++ | nqp-src/QRegex.nqp:
ParseShared is no longer lexical.
 18:48
p-jvm-prep: 2fe1938 | jnthn++ | src/org/perl6/nqp/runtime/Ops.java:
Basic support for iterating scopes.

Implement nqp::getmessage.
jnthn :/ 18:48
arnsholt Heh. I suspect dalek only remembers one floodlimit parameter, or something like that 18:49
japhb jnthn, looking at that nqp commit, it appears correct behaving code was actually shorter and more efficient too. Pretty much all win. :-) 18:50
jnthn indeed. 18:51
arnsholt jnthn: Nothing actually using VMArray in Parrot NQP/Rakudo yet, right? 19:10
jnthn arnsholt: Correct, they are but stubs. 19:11
arnsholt Cool. I'll look into extending it so that Buf can be VMArray REPRed then 19:12
arnsholt I suspect it'll involve either a new HOW or modifying ClassHOW, that make sense? 19:15
jnthn I've somewhat expect we'll end up with some kinda primitve array meta-object at some point. 19:16
*expected
arnsholt Right. I'll give that a whack as well, then 19:17
nwc10 jnthn: nqptest passes (except for the ICU one) 19:23
p-jvm-prep: 2219d82 | jnthn++ | src/org/perl6/nqp/ (2 files):
Fix various buffering issues, so we get output.

Need to revisit this area later.
p-jvm-prep: 49b84c9 | jnthn++ | nqp-src/QRegex.nqp:
Add parse tree dumper, from NQP repo.
p-jvm-prep: 7775882 | jnthn++ | nqp-src/NQPHLL.pm:
Use nqp::getmessage rather than stringification.
p-jvm-prep: 458254f | jnthn++ | nqp-src/QASTNodes.nqp:
Missing null check.
alester masak: I still haven't forgotten you. 19:31
phenny alester: 25 Feb 01:36Z <[Coke]> tell alester to check out www.warrenellis.com/?p=14655
alester This work project is just killing me.
We roll tonight.
alester phenny: Tell Coke thanks for the ack-ack link 19:31
diakopter probahly phenny wants [Coke] 19:32
alester I usually do in the afternoon, too.
phenny: Tell [Coke] thanks for the ack-ack link 19:33
japhb Heh, I bet phenny is case-sensitive WRT commands -- and it wants lowercase. :-) 19:34
diakopter phenny: tell alester see 19:36
phenny diakopter: I'll pass that on when alester is around.
nwc10 jnthn: nqptest still passes (except for the ICU one) 19:37
alester I give up w/phenny. 19:37
phenny alester: 19:36Z <diakopter> tell alester see
uvtc What exactly is meant by "container" in Perl 6? From previous discusssions, I see that $s, @a, and %h may or may not refer to containers... Are scalars, arrays, and hashes containers? 20:10
I see that there's a "Containers" doc, perlcabal.org/syn/S32/Containers.html , but 20:12
no intro there.
PerlJam uvtc: those are all containers to me. 20:13
mostly you talk about containers when you're concerned about operations that happen on the container vs. operations that happen on the value that's in the container 20:14
uvtc r: my @a; say @a.WHAT; my $x; say $x.WHAT; #<--- Scalar?
p6eval rakudo c2910e: OUTPUT«(Array)␤(Any)␤»
uvtc Is a Parcel a container?
moritz uvtc: Scalar is a container type. You basically never see it
uvtc I think Perl 6 may use the term "container" in a slightly different sense than, say, Java (where a container is a class that like the ones in java.util). 20:17
rindolf Hi all. 20:18
Is TimToady here?
moritz rindolf: this is not we-contact-TimToady-here channel. If you want to do that, please take it private message 20:19
PerlJam rindolf: depends on what you mean by "here" :-)
timotimo for 1..* -> $i { 1; } # making this not use up a crazy amount of space is how hard? could rakudo see that the early parts of the list are not used any more at all?
rindolf moritz: OK, sorry.
moritz timotimo: in sink context? or other context?
timotimo perl6 -e that code gives me steadily growing memory consumption 20:20
rindolf PerlJam: www.youtube.com/watch?v=9m-kbBamg_U - that's what I mean - excerpt from Sesame Street.
timotimo i believe this is in sink context then?
skids Well, test with "sink for" 20:20
moritz timotimo: doesn't seem to grow all that much for me 20:21
in relation to typical Rakudo memory consuption, that is 20:22
timotimo oh?
moritz but still grows 20:23
timotimo i like those bugs that just end in coke asking "can this be closed" and followed by silence for 1 to 2 years
moritz timotimo: it might not be rocket science to improve it.
timotimo: the list iteration code in sink context isn't particularly optimized, iirc 20:24
uvtc How do I create a list in Perl 6? I see that `<foo bar baz>` gives me a Parcel. If I assign it to an array, as in `my @a = <foo bar baz>`, I get an array. So far, I haven't see a "List" obect, afaik.
timotimo would it be acceptable to special-case ranges?
moritz uvtc: try <a b c>.list
timotimo r: <foo bar baz>.list.perl.say
p6eval rakudo c2910e: OUTPUT«("foo", "bar", "baz").list␤»
uvtc r: say <a b c>.list.WHAT 20:25
p6eval rakudo c2910e: OUTPUT«(List)␤»
moritz timotimo: it would, as long as it's not done inside any tight loop
uvtc Thanks. Though, what is the purpose of .list? Where would I need (or, where would I end up with) a list?
moritz timotimo: src/core/MapIter.pm, method reify($n = 1, :$sink)
timotimo: that's what might need optimizing 20:26
timotimo what do you mean "not done inside any tight loop"?
moritz uvtc: it's used for lazy, sequential containers where the items aren't necessarily assignables
PerlJam uvtc: .list is useful when you only want the positional bits of a parcel :) 20:27
moritz timotimo: if you want to special-case ranges, don't do the check if something is a range inside the loop that does the actual iteration
uvtc PerlJam: So, when you have a Parcel in-hand and want to turn it into a list?
moritz another use casaes for .list: 20:28
timotimo oh my. suddenly: PIR code
moritz r: my $a = (1, 2, 3); my $c = 0; $c++ for $a; say $c
p6eval rakudo c2910e: OUTPUT«1␤»
uvtc I see at doc.perl6.org , about Array it says, "An Array is a List which forces all its elements to be scalar containers, which means you can assign to array elements." but am trying to make sense of that.
moritz r: my $a = (1, 2, 3); my $c = 0; $c++ for $a.list; say $c
p6eval rakudo c2910e: OUTPUT«3␤»
moritz r: my @a := <a b c>; @a[0] = 3; # you can't assign here, because @a[0] isn't a container 20:29
p6eval rakudo c2910e: OUTPUT«Cannot modify an immutable value␤ in block at /tmp/fMGh8Vmfaa:1␤␤»
moritz r: my @a := <a b c>.list; @a[0] = 3;
p6eval rakudo c2910e: OUTPUT«Cannot modify an immutable value␤ in block at /tmp/OLsOP8M3ov:1␤␤»
moritz but
my $scalar = 0; my @a := ($scalar, 1, 2); @a[0] = 42; # should work, because the first element is a Scalar container 20:30
r: my $scalar = 0; my @a := ($scalar, 1, 2); @a[0] = 42; # should work, because the first element is a Scalar container
p6eval rakudo c2910e: ( no output )
uvtc Why would I want to do: `my @a := (1, 2, 3)`? It looks like I'm creating an array, but then assigning a Parcel to it. ... Yet it seems to work like an array, regardless. 20:32
r: my @a = <foo bar baz>; say @a.WHAT; # Parcel?
p6eval rakudo c2910e: OUTPUT«(Array)␤»
moritz uvtc: := doesn't assign
PerlJam uvtc: := is binding
moritz assignment to an array empties it, and then re-fills again 20:33
uvtc moritz: Well, `:=` assigns a reference, correct? Which is referred to as "binding", correct?
moritz binding to an array variable replaces the array with whatever is on the right
uvtc: I tend to avoid that phrasing, because it's quite ambiguous, IMHO
uvtc moritz: FWICT, $s, @a, and %h are all references, but they're automatically dereferenced when you use them, so folks tend not to even notice. Is that right? 20:35
moritz uvtc: what do you mean by "are all references"? 20:36
uvtc They are names that refer to objects. That is, even though folks tend to say, "$x is 5", what's really going on under the hood is that "$x refers to an Int object who's value is 5, but it's automatically dereferenced so you won't even notice". Right? 20:37
uvtc Reference == "pointer" 20:38
moritz there's no difference between "5" and "an Int object who's value is 5"
uvtc (Though I seem to recall, in C++ there's a difference between references and pointers, in Perl 6 they seem to be general descriptive terms which are synonyms.)
moritz let my try it this way 20:39
my $x = 42; # $x is a variable, which is bound to a Scalar, which contains the 42
dalek p-jvm-prep: 9a5e5a5 | jnthn++ | nqp-src/QASTNodes.nqp:
Do some null checks correctly.
p-jvm-prep: 9fb6e7e | jnthn++ | / (2 files):
Some small tweaks towards getting backend working.
uvtc "bound to" == "refers to" == "points to" .Correct? 20:40
moritz so you can assign to $x, $x = 23 20:40
which in turn calls the STORE method on the scalar
yes
moritz but you can also say 20:40
my $x := 42;
moritz in which case the variable $x contains the Int directly, without a Container 20:41
which means you cannot assing to it, because Int has no .STORE method
PerlJam reads an interview of moritz
moritz feels famous, in this small echo chamber 20:43
PerlJam oh. Hmm. this is a redux of some earlier interview.
uvtc is currently interviewing moritz , live!
PerlJam uvtc++
moritz PerlJam: yes, there's a link to the original one at the bottom 20:44
jnthn waits for the cross-compiler to cross-compile the compiler...
PerlJam moritz: I want to know more about this "People of Perl 6 club" ;)
nwc10 jnthn: it's getting that close? 20:45
timotimo jnthn: that sounds like we are getting dangerously close to a fully cross-compiled nqp on jvm :)
uvtc moritz: I think I'm having some basic misunderstanding of terminology here. I think of "$x" as a name (an identifier). It may refer to something, or may not. But you're calling it a *variable*. And theres also a container involved sometimes as well. Hm.
s/theres/there's/ 20:46
moritz uvtc: $x is a variable, which means it's an entry in some lexpad. It's name is '$x'
uvtc: the entry in the lexpad is basically a pointer to an object 20:47
uvtc There's that word again: "lexpad". I don't know what that is. It haunts me. :)
moritz uvtc: and usually that object is a container, for $x is Scalar
uvtc dreams (neigh, a nightmare) that he's in a cornfield, late at night, searching for the lexpad where the helipod can land so he can get out... 20:48
moritz uvtc: the compiler keeps track of all the known variables for each scope. The runtime representation of this tracking is a lexpad
uvtc (ouch. s/neigh/nay)
PerlJam uvtc: you can think of a lexpad as a per-scope symbol table.
moritz long name "lexical pad"
PerlJam (sorta)
moritz note that there two sorts of tracking of variables 20:49
the one is inside the compiler, so that we know if a variable is declared or not, and if yes, where it was declared, what's its type etc.
moritz and then there's a runtime structure, which contains pointers to the values stored in the variables. That runtime structure is commonly called "lexical pad" or "lexpad" 20:50
and the reason you can't optimize out the lexpads (in the general case) is closures
moritz my $closure = do { my $x = 42; sub () { say 42 } }; $closure() 20:51
nr: my $closure = do { my $x = 42; sub () { say 42 } }; $closure()
PerlJam moritz: careful ... you'll have covered an entire CS curriculum if you keep going :
p6eval niecza v24-24-gbdc3343: OUTPUT«Potential difficulties:␤ $x is declared but not used at /tmp/wtA5wBMaNF line 1:␤------> my $closure = do { my ⏏$x = 42; sub () { say 42 } }; $closure()␤␤42␤»
..rakudo c2910e: OUTPUT«42␤»
PerlJam er :)
moritz nr: my $closure = do { my $x = 42; sub () { say $x } }; $closure()
p6eval rakudo c2910e, niecza v24-24-gbdc3343: OUTPUT«42␤» 20:52
uvtc (Curious about the name "pad". "pad" ~~ "paddock"? Or maybe like a pad of paper on which to write notes about lexicals.)
moritz uvtc: pad of paper
PerlJam uvtc: like a pad of paper
uvtc: "scratch pad"
uvtc Ok.
moritz so, $closure must know somethough about $x
which the compiler achives by sticking a pointer the lexpad and to the actual code object into a common structure, and calls it $closure 20:53
uvtc moritz: Right. The sub was declared when $x was in scope, so it knows about it. A closure.
uvtc is rereading what's been written above... 20:54
PerlJam: Yes, I don't have a CS education. Did physics instead.
moritz uvtc: me too 20:55
uvtc And now, here I am, up the creek without a lexpaddle.
nwc10 jnthn: nqptest still passes (except for the ICU one) 20:55
moritz so, if you do 20:56
my $x = 42;
moritz then the lexpad entry for $x is a pointer to a Scalar object 20:56
uvtc yup 20:56
moritz and that Scalar object has another pointer, which points to an Int 20:57
so, that allows you to do
sub f($a is rw) { $a = 18 }; f($x)
so that passes the Scalar object (which we usually just call "the container") to function f 20:58
dalek p-jvm-prep: fd60679 | jnthn++ | src/org/perl6/nqp/ (4 files):
Primitives for in-process JAST => JVM bytecode.
p-jvm-prep: 9819d20 | jnthn++ | / (2 files):
Flesh out backend compilation stages.
moritz which means that if f() modifies $a, then $x also changes
note that we don't only just need that for 'rw' arguments 20:59
but also for things like
my @a; @a[0] = 42;
uvtc Ok. So, when you call `f($x)`, it's changing the Scalar to point to a different Int.
moritz uvtc: that's what the assignment inside of f does, not the call itself (just want to make sure we don't conflate stuff here)
uvtc yup 21:00
moritz ok, back to my @a; @a[0] = 42;
that only works because indexing into an array returns a Scalar
otoh if you do
r: (0, 1, 2)[0] = 42
p6eval rakudo c2910e: OUTPUT«Cannot modify an immutable value␤ in block at /tmp/ZJz_IK2Vrc:1␤␤»
moritz here the [0] indexes into a Parcel 21:01
moritz and indexing into the Parcel just returns whatever is there 21:01
and that's not a Scalar, it's an Int
you can't assing to it
uvtc Gotcha.
moritz but if it happens to be a Scalar, it works again
uvtc Ok.
moritz r: my $s = 42; ($s, 1, 2)[0] = 23; say $s 21:02
p6eval rakudo c2910e: OUTPUT«23␤»
moritz so, a List is like a Parcel in that regard
the difference is that a List has more magic that makes it lazy
which is why the return value from a .map is a List
uvtc So this comes back to the difference between List and Array, where the docs say, "An Array is a List which forces all its elements to be scalar containers, which means you can assign to array elements." 21:03
moritz r: say (0..*).map(*+1).WHAT
p6eval rakudo c2910e: OUTPUT«(List)␤»
moritz uvtc: right
uvtc Wait. Why would `map` return a List? I may want to save what it returns and assign to its elements later on.
raiph heh. if i were to claim that P6 is parsed by a single P6 regex, what would that regex be? 21:04
uvtc Oh, I see. `my @a = map ...` would give me an array in @a.
moritz uvtc: correct
uvtc: what doesn't work is (0..*).map(...)[2] = 42
PerlJam moritz: I know others (and probably you too) have written about Lists and Parcels and such in the past, but you're doing an excellent job explaining it right now. Maybe you should consider another post on the subject. Just take the last several minutes of IRC and munge it into an article.
uvtc Thanks, moritz . Going to go back and read again what you've told me (above). 21:05
moritz PerlJam: I thought the same, basically :-)
PerlJam moritz++
skids definitely one of the clearer and less cluttered explanations which I have seen. moritz++
uvtc PerlJam: Good idea. I'm trying to understand it so I can put an article together, but I've only got this so far: github.com/uvtc/compact-perl6-tut/...binding.md and it's not yet correct. 21:06
PerlJam raiph: complicated. :)
nwc10 jnthn: nqptest still passes (except for the ICU one)
raiph PerlJam: heh. I'm thinking rule comp_unit
moritz TOP 21:07
:-)
dalek kudo/nom: f4e96b0 | moritz++ | README:
README: update download location for star packages
timotimo rt.perl.org/rt3/Ticket/Display.html?id=76928 - is this closable? it looks that way to me. 21:08
moritz takes a look 21:10
yes, closable with tests
timotimo i will think about adding a test for that tomorrow. need to be up early and still got stuff to pack :) 21:12
moritz note that you'll want to use .HOW.gist instead of say .HOW 21:13
and now it's bed time for me
jnthn 'night, moritz 21:14
timotimo sure
good night, happy perl (not only) 6 people! 21:15
jnthn Must doesn't work yet, but: 22:31
$ java -cp 3rdparty/bcel/bcel-5.2.jar;.;bin NQPJVM -e "say('omgz nqp running on jvm')"
omgz nqp running on jvm
colomon \o/
"Must"? 22:32
diakopter o_O 22:33
"Much?"
"Much"?
jnthn *most
Or much
I dunno how much yet or how many root causes :)
diakopter what's bin 22:34
oh I see
dalek p-jvm-prep: a63c38b | jnthn++ | nqp-src/ (3 files):
Use high-level join and fix some op usages.
 22:36
p-jvm-prep: b93079a | jnthn++ | src/org/perl6/nqp/runtime/Ops.java:
Disable colliding fate sorting for now.

Seems it's not quite right...and we get righter results without it.
p-jvm-prep: 266da05 | jnthn++ | / (2 files):
Start cross-compiling NQP itself.

We now survive "say('hello world')" running entirely on the JVM, with NQP's grammar/actions parsing and building the AST, the QAST -> JVM bytecode happening in-process and then evaluating the result. Much to bug hunt, but very simple things work. Note, -e and file input work but the REPL does not yet.
p-jvm-prep: ce56d86 | jnthn++ | src/org/perl6/nqp/runtime/Ops.java:
Fixed fate sort, hopefully.
 22:42
pmurias jnthn: re new backend API will that enable -Cjvm and -Cjs options? 22:47
jnthn pmurias: Not directly I think, but it's a move towards providing that if we want to try and support cross-compilation generally. 22:49
pmurias jnthn: the way I support having common sc ids in nqp-js it that rather than having a stable ids option after parsing I convert the same ast to both pbc and js 22:50
census Hi! nonperl5 q. Does anybody know fortran? 23:11
I would like to translate some fortran code between fortran and perl.
But I don't know a lick of fortran, unfortunately
scsys.co.uk:8002/232718?ln=on&t...Format+it! 23:12
thank you! :)
sorear i know some of it
mostly from trying to read ADVENT
census i see 23:15
census sorear: what do you think of that code i posted? 23:16
diakopter in Soviet Russia, formulas translate you 23:20
uvtc census: maybe try the #fortran channel 23:21
census uvtc: yes i have tried :)
it is a small channel.
diakopter: in soviet union, you drink vodka. and when you are done, you drink some more vodka
ggoebel jnthn++: self-hosting nqp on jvm 23:23
jnthn It'll be better when I figure out why \h+ matches 123...
tadzik oh oh 23:33
ooossom!
pmurias jnthn++ # self hosting nqp on the jvm 23:38
dalek p-jvm-prep: 289ad59 | jnthn++ | lib/JAST/Nodes.nqp:
Avoid nqp::escape for SVal encoding.

It does all sorts of things that need all sorts of untangling...		 23:55
p-jvm-prep: e556c3d | jnthn++ | nqp-src/NQPHLL.pm:
Fix an atkey that shoulda been an atpos.

Powered by Cro and the Raku® Programming Language.

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

Thank you!