»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.
00:03 benabik joined 00:09 cognominal left
dalek c: 923f9e1 | (Geoffrey Broadwell)++ | htmlify.pl:
htmlify.pl: Minor code readability improvements, no functional changes
 00:18
c: a7d2b13 | (Geoffrey Broadwell)++ | htmlify.pl:
htmlify.pl: Factor two major Pod file processing blocks out of MAIN; no functional changes
c: 3fb2dfe | (Geoffrey Broadwell)++ | htmlify.pl:
Fix naive refactoring bug in previous commit
00:19 gfldex left, gfldex joined 00:23 benabik left, dayangkun joined 00:29 fgomez left 00:31 _jaldhar left 00:37 fgomez joined 00:38 cognominal joined 00:46 cognominal left
dalek ecs: 65c3a06 | larry++ | S11-modules.pod:
spec both ::() and :file() require forms
 00:47
ecs: 0a50011 | larry++ | S11-modules.pod:
semantic clarifications of new require forms

fixes #23
00:48 cognominal joined 00:49 _jaldhar joined 00:51 immortal left
[Coke] gist.github.com/3628577 00:53
00:54 scott_ joined, scott_ is now known as Guest38681
[Coke] (getting an error trying to compile array.pm) 00:55
00:55 scottdware joined 01:03 sisar joined 01:04 scottdware left 01:08 benabik joined 01:09 spider-mario left 01:12 hypolin joined 01:15 whiteknight left
[Coke] wonders if it's a pod commenting issue. 01:17
01:22 sisar left 01:27 fgomez left 01:29 fgomez joined 01:31 topologist joined, MikeFair_ joined 01:40 topologist left, Cofyc joined 01:42 tokuhiro_ joined
[Coke] I'm seeing this a lot: Error while compiling block : ResizableStringArray: Can't shift from an empty array! 01:42
(trying to use nqp latest) 01:44
gist.github.com/3629000 01:45
MikeFair_ [Coke]: I saw that too 01:46
I fixed that one
It had to do with the implemention of say I believe 01:47
I'm currently dealing with either:
Method 'parse' not found for invocant of class 'NQPMu'
(when I get it to compile and run
or
method 'blocktype' not found for QAST::CompUnit 01:48
or something close to that
But i'm not sure how much of this is NQP and how much of this is Parrot and how much of this is my not unerstanding what _should_ be happening -- so I've tried not to make so much noise about this 01:49
Especially here on #Perl6 :) 01:50
benabik QAST is largely #perl6 territory. 01:51
MikeFair_ [Coke]: One change I had to make to Grammar / Actions / Compiler - that instead of the Compiler instantiation being a statement it's now a class 01:52
so this should be at the top of those .pms : grammar safire::Grammar is HLL::Grammar {
With the corresponding } at the bottom
The can't shift from empty array had something to do with the command line ARGS IIRC 01:53
So I just commented that line out to get passed compilation and would deal with it later
benabik: Yeah, but it seems that what I'm dealing with is that the deafult shell language Parrot comes with doesn't seem to instantiate the language in the same way NQP does. I'm not sure if this a problem or not though 01:55
benabik: So I'm not clear on the separation between the two domains. :)
01:56 dayangkun left
MikeFair_ [Coke]: I also had to add a: use NQPHLL; line to the top of each .pm as well 01:59
[Coke] You mentioned a lot of issues that I already addresses, yah. 02:00
MikeFair_ [Coke]: Grammar for say had to be rewritten from [ <EXPR> ] ** ',' }
to [ <EXPR> ]* %',' }
MikeFair_ nods.
[Coke]: Figured you would have as these are the stupid obvious ones. :)
[Coke] partcl-nqp/nqp2 now builds again, but I've gutted the grammar so that the only thing that parses is 'three' which generates a '3'. 02:01
MikeFair_ I gleaned a lot of this from the examples/json.nqp
[Coke] but at least now I have something that builds and runs, and I can slowly add functionality back in.
02:02 dayangkun joined
MikeFair_ [Coke]: Any ideas what my "method 'parse' not found for 'NQPMu'" message might be about 02:03
[Coke] you need to register your compiler. 02:04
github.com/partcl/partcl-nqp/blob/...tcl.pm#L96 02:05
02:07 sisar joined
MikeFair_ [Coke]: Yeah I saw that function but I'm not sure what file it should go in.. my main is currently in a 'setup.pir' module 02:07
MikeFair_ will poke around partcl-nqp and learn. Thanks! 02:08
[Coke] I moved my main file from a .pir to a .pm
I am sorry for you if that's your best example. ;) 02:09
where can I find a list of nqp opcodes? 02:11
02:12 Vlavv` left
MikeFair_ hehe - there doesn't seem to be much in the way of docs, I've jsut been grepping -ri through the source tree 02:13
[ <EXPR> ]* %',' }
err
grep -ri t/qast/pirt.t
err grep -ri pirop t/qast/pirt.t
That might not be all of them, but it's a start ;) 02:14
02:19 orafu left, orafu joined 02:23 dayangkun left, tokuhiro_ left
pmichaud nqp opcodes: github.com/perl6/nqp/blob/master/d...opcode.txt 02:27
02:29 popl joined
MikeFair_ pmichaud: Even better :) 02:29
pmichaud that list is fairly up-to-date 02:30
sorear o/ pmichaud 02:31
phenny sorear: 02:31Z <zort-> tell sorear 4
pmichaud the complete list would be in src/QAST/Operations.nqp
phenny sorear: 02:31Z <zort-> tell sorear 5
sorear: 02:31Z <zort-> tell sorear 6
sorear: 02:31Z <zort-> tell sorear 7
Further messages sent privately 02:32
pmichaud o/ sorear
sorear who is zort-
02:33 dayangkun joined
benabik Someone who likes to count? 02:34
phenny benabik: 02:34Z <zort-> tell benabik 65
benabik: 02:34Z <zort-> tell benabik 66
benabik: 02:34Z <zort-> tell benabik 67
benabik uhhhh
phenny benabik: 02:34Z <zort-> tell benabik 68
benabik: 02:34Z <zort-> tell benabik 69
benabik: 02:34Z <zort-> tell benabik 70
benabik: 02:34Z <zort-> tell benabik 71
benabik Someone trying to crash phenny? 02:35
phenny benabik: 02:35Z <zort-> tell benabik 72
benabik: 02:35Z <zort-> tell benabik 73
benabik: 02:35Z <zort-> tell benabik 74
sorear hmm 02:36
phenny sorear: 02:36Z <zort-> tell sorear 105
sorear: 02:36Z <zort-> tell sorear 106
sorear: 02:36Z <zort-> tell sorear 107
sorear: 02:36Z <zort-> tell sorear 108
Further messages sent privately
02:38 ChanServ sets mode: +o sorear, phenny was kicked by sorear (phenny))
benabik I wonder where he's sending them... PM to phenny? 02:42
sorear or another channel
phenny is in quite a few places
I sent off a ping to sbp
benabik sorear++ 02:43
02:47 daniel-s__ joined
TimToady wonders if it was a python injection attack of some sort 02:50
02:51 daniel-s_ left
TimToady probably not, since the time changes 02:53
03:34 skids left 03:36 jaldhar_ joined 03:37 _jaldhar left
MikeFair_ pirs pirs everywhere but not an nqp :) 03:41
dalek c: 1f37c81 | (Geoffrey Broadwell)++ | app.pl:
Trivial Mojolicious::Lite app to serve static content from html dir; currently uses redirects instead of content negotiation
 03:47
04:08 dayangkun left, colomon left
MikeFair_ [Coke]: Did you get your error figured out? 04:08
04:09 kaleem joined 04:13 jaldhar_ left 04:16 colomon joined 04:21 jaldhar_ joined 04:22 Cofyc left
__sri japhb: stay away from $ENV{MOJO_APP}! 04:25
just use the app function instead 04:27
MOJO_APP actually doesn't work anymore in newer Mojolicious releases, since putting data structures into %ENV was never a Perl feature and has been broken in 5.17.3 04:29
you may want to visit #mojo :) 04:31
04:35 marmay joined 04:43 colomon left 04:45 colomon joined
geekosaur wat. that should never have worked 04:52
05:07 cognominal left
sorear geekosaur: that's what #p5p said too. sri was not amused 05:11
__sri yea :,( 05:12
it was only a side effect of a mojolicious detection feature, but many people have actually been depending on it 05:19
japhb __sri, that's because old versions of Mojolicious::Lite had no other obvious way to grab the app object. :-/ 05:34
__sri perldoc Mojolicious::Lite
the app function existed since day 1
japhb Ah, my bad. It is documented near the end, but confusing in the first few sections of the doc (where it appears to be only some magic to make app->start work) 05:36
But I understand now, fixing ...
__sri you must be using a very old version 05:37
i don't think there are any examples without app->start anymore
mojolicio.us/perldoc/Mojolicious/Lite#SYNOPSIS # :) 05:38
japhb Yeah, the box I'm working on is very out of date.
No worries.
05:38 popl left
dalek c: ba8b767 | (Geoffrey Broadwell)++ | app.pl:
Replace Mojolicious guts-evilness with normal API/attribute access
 05:40
japhb There you go ^^
__sri \o/
05:41 cognominal joined
japhb __sri, while you're here -- what is the "proper" way to do what I did in that app? (Emulate Apache doing .../ => .../index.html and .../file => .../file.html) 05:41
__sri word of warning, ->root() is gone for like 6 months, we now support multiple ->paths() ;)
japhb Gone as in "old API no longer exists" or gone as in "deprecated and not documented, but still exists for back compat"? 05:42
05:42 benabik left, benabik joined
__sri it was deprecated and then removed 6 months ago :) 05:42
japhb *sigh*
__sri redirect seems fine i guess 05:43
japhb OK, question for the 6'ers: Is there a way to use multi *subs* from a Role? Or does a Role even have a way of feeding multi subs to a composer? 05:44
__sri i'm not quite understanding the problem, but maybe all you want is default values for placeholders
japhb __sri, Well, I guess what I mean is "How would you make that just one route?" 05:45
__sri default values i suppose, the lite tutorial should have all you need
japhb I found the lite tutorial to be a bit ... lite.
Too terse.
Not enough descriptions of edge cases and interactions between things. 05:46
I found myself thinking "Do I really need to go to the full Mojolicious just to do this sanely?"
__sri lite and full are 99% the same, you can use the exact same features 05:47
japhb What I ended up with was my best guess (from the M::Lite doc) of how to capture the "spirit" of a M::Lite app while still handling the task.
__sri anyway, we are quite off topic, you should join #mojo over on irc.perl.org
japhb Fair enough. 05:48
benabik sometimes wonders if #p6 has an on topic.
japhb benabik, I've asked two on-topic questions tonight. They were both Warnocked. ;-/ 05:49
tadzik I think the rule is mostly "talk about whatever you want as long as you don't interrupt any ongoing on-topic discussion"
japhb tadzik, Ah, you're awake, coo. 05:50
er cool.
tadzik my problem with Mojo vs Mojo::Lite was that the documentation for libraries always covers only the Lite part, same with tests
it's sometimes hard to figure out how to do the same thing in full mojolicious
japhb My question above about multi subs from Roles was because I'm hacking at a possible refactorization of Pod::To::* 05:51
__sri tadzik: it gets much easier later on
tadzik __sri: what do you mean by later on?
__sri once you've worked a bit with both, lite and full apps
sorear japhb: if you'd just ask questions I know the answer to...
japhb sorear, *chuckle* 05:52
tadzik I see
__sri tadzik: usually i only hear that complaint from those who skipped the lite tutorial :)
tadzik hehe
__sri you're kinda supposed to learn lite and move on to full 05:53
tadzik which says "this is what lite does under the hood"
05:53 sorear sets mode: -o sorear, cognominal left, cognominal joined
japhb tadzik, Why do you use multis in Pod::To::HTML and not in Pod::To::Text? Was one just a newer design than the other? 05:54
tadzik japhb: Pod::To::HTML is not really written by me :) 05:55
japhb Oh.
japhb fail
tadzik I started it, but flussence++ wrote most of it I think
japhb Ah, OK 05:56
phenny, ask jnthn Is there a way to use multi *subs* from a Role? Or does a Role even have a way of feeding multi subs to a composer? 05:57
Oh right, phenny got kicked.
Is it safe to bring phenny back?
TimToady stop asking hard question! :) 06:00
*tions
__sri free phenny!
benabik Did phenny used to cost something? 06:04
sorear aloha: msg japhb we have two tellbots here 06:05
msg japhb we have two tellbots here
aloha: tell japhb we have two tellbots here 06:06
benabik aloha is somewhat crippled here...
sorear :/
japhb wonders if aloha knows I'm here
Apparently not.
sorear aloha: seen japhb
aloha sorear: japhb was last seen in #perl6 1 seconds ago saying "Apparently not.".
benabik seen japhb
aloha japhb was last seen in #perl6 2 seconds ago saying "Apparently not.".
japhb heh
06:06 lambdabot joined
benabik thinks aloha is just here so we in #parrot can tell if someone is over here. 06:06
sorear ok, that worked at least. 06:07
@tell japhb the old standby.
lambdabot Consider it noted.
japhb @tell jnthn Is there a way to use multi *subs* from a Role? Or does a Role even have a way of feeding multi subs to a composer?
lambdabot Consider it noted.
sorear I'm a little suprised I still have my special-command bit
japhb Hopefully he gets that message ...
lambdabot japhb: You have 1 new message. '/msg lambdabot @messages' to read it.
sorear '/msg lambdabot @join #perl6'
japhb break & 06:08
06:25 aindilis left
moritz \o 06:38
MikeFair_ Hey moritz 06:42
06:42 dayangkun joined
moritz oh hai MikeFair_ 06:43
MikeFair_ moritz: You have any experience with using nqp to implement another language on parrot. For the life of me I can't figure out how to get the thing to actually build (I keep running into various 'I don't know about that' kind of errors 06:44
or perhaps can point to something that does? Some example perhaps that I can copy.
I've built nqp and gotten it installed, and it works, but I'm just not able to use it successfully during the build process of my language 06:45
06:51 xinming_ left
moritz MikeFair_: my only experience with using NQP is rakudo 06:55
MikeFair_ Yeah that's what I figured ppl's experience would be
moritz MikeFair_: so I'd copy the interesting parts of rakudo's src/main.nqp
and then the heads of src/Perl6/{Grammar,Actions}.pm 06:56
MikeFair_ I'm just not sure what the build process is. Actually I just did that! It had the bits in PIR that I'd only been seeing in a .pm file before
The problem I'm encountering now is nqp seems to want Grammar.pbc which doesn't exist because it hasn't been built
:)
MikeFair_ looks at the heads of actions/Grammar to make sure he's not missing anything 06:57
06:58 FROGGS__ joined, FROGGS joined
MikeFair_ Hmm, I've got use NPQHLL at the heads now (stolen from the json.nqp example) 06:58
FROGGS good morning
MikeFair_ hello FROGGS
06:59 fgomez left
MikeFair_ moritz: I clearly don't have the rest of that stuff in top and those methods though 07:00
:)
07:01 xinming joined
sergot hi o/ ! 07:01
07:07 brrt joined, stepnem joined
tadzik hellllo 07:11
frettled hiho!
MikeFair_ hello
moritz MikeFair_: well, you can compile .pir to .pbc with parrot -o file.pbc file.pir
07:14 SamuraiJack joined
FROGGS will it be again a bit faster then? 07:14
07:14 kresike joined
kresike good morning all you happy perl6 people 07:14
FROGGS .pbc is portable, right?
gmorning 07:15
tadzik FROGGS: portablish 07:17
FROGGS k, thanks 07:18
tadzik well, as it's essentialy pir, you probably don't want to ship it 07:19
dalek ast: e40bbd1 | moritz++ | S05-match/capturing-contexts.t:
fudge a changed test that rakudo does not like in its current form
 07:23
FROGGS tadzik: speed matters a lot, and as it seems than when I ship A.pm6, A.pir and A.bpc it will pick the last one, I am fine with that 07:24
but it might should compile the stuff when compiling, if it is able to, otherwise skip
tadzik FROGGS: it may just not work 07:25
moritz FROGGS: the problem with shipping compiled .pir and .pbc files is that they depend on exact revision of rakudo and all dependencies
tadzik see also: panda not working after recompiling rakudo
FROGGS ohh, thats mad
so it will just be helpful for some kind of staticperl6 later on 07:26
or there must be an automatic rebuilt when it detects changes
moritz well, it just means you have to precompile when you install a new rakudo 07:27
and only the hardcore developers do that very often
but we plan to change the whole precompilation thing 07:28
to make it a cache managed by rakudo, not by the user
FROGGS that sounds like somthing I wanna have ;o)
(some day)
moritz and you will 07:29
(some day)
:-)
FROGGS **g
TimToady rosettacode.org/wiki/Roots_of_a_function#Perl_6 # 450 entries now 07:34
MikeFair_ wooot 07:39
MikeFair_ built and installed perl6
Wish it was getting my language working but as all I really need/want right now is mostly Grammar.pm, i'm thinking I might just do it straight in P6 instead of Parrot 07:40
07:41 Cofyc joined
MikeFair_ Goodnight all! 07:54
MikeFair_ sleeps.
08:05 BenBE left
masak 'ning, #perl6 08:16
FROGGS \o masak
masak TimToady++ # 450
08:19 leont joined, Psyche^ joined, Patterner left, Psyche^ is now known as Patterner
masak I like rosettacode.org/wiki/Roots_of_a_function#Perl_6 -- the NEXT happens in a narrower scope than the for loop itself, and somehow that surprises me, even though it shouldn't feel strange at all at this point. 08:19
08:27 sivoais left 08:32 dakkar joined 08:34 Circlepuller_ joined 08:35 Circlepuller left 08:36 fhelmberger joined 08:59 glosoli joined, glosoli left
masak hm. looking at my last attempt at hygienifying D2 at github.com/rakudo/rakudo/commit/1a...584ea291f3 , I can only assume I was horribly confused, or tired, or both. 09:01
masak reverts that commit and makes a new attempt
today is my Perl 6 day for the week, by the way. expect lots of ramblings about macros.
tadzik hooray 09:02
09:04 hoelzro|away is now known as hoelzro
masak jnthn is unfortunately busy today. so if someone else wants to act as vict^Wsounding board for my heinous ideas, that would help a lot. 09:04
masak draws ASCII art to make things clear to himself 09:08
moritz can act as macro victim, yes 09:09
masak \o/
masak was hoping for moritz++
moritz: let me draw this ASCII art for you; then we can talk over that 09:10
r: macro twice($code) { quasi { {{{$code}}}; {{{$code}}} } }; my $counter = 0; twice $counter++; say $counter
p6eval rakudo cc1858: OUTPUT«0␤»
masak :(
r: macro twice($code) { my $counter = 4; quasi { {{{$code}}}; {{{$code}}}; say $counter } }; my $counter = 0; twice $counter++; say $counter 09:11
p6eval rakudo cc1858: OUTPUT«6␤0␤»
masak well, *something* gets incremented twice...
09:11 hoelzro is now known as hoelzro|away
moritz r: macro twice($code) { my $counter = 4; quasi { {{{$code}}}; {{{$code}}}; say $counter } }; my $counter = 0; twice $counter += 2; say $counter 09:15
p6eval rakudo cc1858: OUTPUT«8␤0␤»
moritz r: macro twice($code) { my $counter = 4; quasi { {{{$code}}}; {{{$code}}}; say $counter } }; my $counter = 0; twice say $counter; say $counter
p6eval rakudo cc1858: OUTPUT«4␤4␤4␤0␤»
moritz r: macro twice($code) { my $counter = 4; quasi { {{{$code}}}; {{{$code}}}; say $counter } }; my $counter = 0; twice say ++$counter; say $counter
p6eval rakudo cc1858: OUTPUT«5␤6␤6␤0␤»
masak heh, my ASCII art turned out to be words rather than pictures. but writing this down seems to help. 09:18
moritz: gist.github.com/3634046 09:28
it's two worlds. the parse world and the runtime world. macros and quasis are all about safely handing values over from one world to the other. 09:34
essentially, the parse world deals in QAST trees, and the runtime world deals in Perl6::AST trees. the latter are a kind of monadic wrapper around the former.
the only place where this isn't quite true is at (2), where parsetime packages the QAST arguments into Perl6::AST objects. but that's just a kind of polite hand-over from parse to runtime. 09:35
moritz masak: ah, I see the problem 09:39
masak: and I see why it is related to hygiene
masak well, it's less of a problem and more of an "it goes both ways" thing. 09:40
moritz (the problem in the example with macro twice, I mean)
that is a bug, IMHO
masak there are Perl6::ASTs going into the macro and Perl6::ASTs going out of the macro. both kinds must close over their parsed context. only the latter does, currently. 09:41
<moritz> that is a bug, IMHO 09:43
could you be more specific?
moritz well, the argument to twice is supposed to be a *value*, not a thunk 09:44
correct? 09:45
and you don't evaluate the argument before passing it
hence the variable gets the wrong context
and hence double execution
masak oh, the double execution is correct. 09:46
moritz why?
masak and if you're thinking thunk, then you're too far ahead in the compilation process. 09:47
at this point `$count++` is an AST.
moritz why?
I mean, why should it be an AST?
masak because ASTs are what's passed into a macro as arguments.
macros are AST manipulators.
moritz r: macro a($x) { say $x }; a 2 09:48
p6eval rakudo cc1858: OUTPUT«AST.new()␤===SORRY!===␤too few positional arguments: 2 passed, 3 (or more) expected␤»
moritz what am I doing wrong here?
masak nothing. you just found a bug, I think.
but you also showed that $x is an AST.
moritz no, I didn't
masak r: macro a($x) { say $x; Nil }; a 2
p6eval rakudo cc1858: OUTPUT«AST.new()␤»
moritz I showed that, somewhere in the process, it tries to generate an AST :-) 09:49
r: macro a($x) { say {{{$x}}}; Nil }; a 2
p6eval rakudo cc1858: OUTPUT«Block.new()␤»
masak point of order: AST.new() is what the gist calls Perl6::AST.
moritz hm, so you can't pass a string to a macro?
masak moritz: I'm not implementing string macros. I'm implementing AST macros. 09:50
moritz: you "can't" use {{{}}} outside quasi blocks.
moritz: in the sense that they will just be three block inside each other.
moritz yes, I figured
but I'm stil not convinced about macro arguments being ASTs
masak this half surprises and half amuses me. 09:51
moritz I mean, not automatically
masak because it's totally clear to me that they are.
though I hadn't factored textual macros into it before, I admit. because I'm not implementing those. 09:52
moritz I'm not talking about textual macros either
masak ok. 09:53
moritz wonders if him being the victim does more harm than good 09:54
masak no, this is interesting so far. 09:55
though I'm a little scared of where the discussion is going. I do hope I'm right ;)
I got the feeling above that you didn't expect `$counter++` to run only once. I wonder why you thought that.
moritz because that's what routines do :-) 09:56
masak r: run-twice(&c) { &c(); &c() }; my $counter = 0; run-twice { $counter++ }; say $counter
p6eval rakudo cc1858: OUTPUT«===SORRY!===␤Confused␤at /tmp/FTGEZVVqdh:1␤»
moritz erm
masak r: sub run-twice(&c) { &c(); &c() }; my $counter = 0; run-twice { $counter++ }; say $counter
p6eval rakudo cc1858: OUTPUT«2␤»
masak even in the case where the macro is a sub, it runs twice.
moritz only if you add a block around it
masak the biggest difference is that the above sub makes two *calls*, whereas our example macro makes two *code insertions*. 09:57
moritz: the macro arguments have implicit blocks around them.
moritz masak: that's what I meant with "thunking"
masak moritz: those blocks must be there for there to be somewhere to attach the surrounding context to.
moritz: right, but they're ASTs.
masak .oO( here, I'll be Abbott, and you'll be Costello ) 09:58
moritz ok, I think you convinced me 09:59
masak .oO( who's on thunk? yes. )
moritz so, on with the show
masak \o/
moritz: some of the things I know today about this, I've discovered in earlier excellent conversations with you. 10:00
moritz: but I don't blame you for forgetting and rediscovering details. I keep doing that, too.
macro implementation seems to be a steep, and slippery, learning curve ;) 10:01
moritz: so, ASTs going in and coming out of a macro are really tree representations of code, pretending to be thunks/closures. 10:02
moritz aye
masak the "pretending to be thunks/closures" bit is the important part to get hygiene right.
it currently only works in the outwards direction.
the patch I need to apply today is a fixup that sets the context correctly in the inwards direction.
github.com/rakudo/rakudo/commit/1a...584ea291f3 is a failed attempt. 10:03
I think it mixes up layers and levels that are much more clearly exaplained in gist.github.com/3634046
cognominal TimToady, what is the relationship between the Summer Institute of Linguistic and tagmemics?
masak cognominal: ooh, I was wondering that too, after reading about tagmemics on Wikipedia.
cognominal Did Kenneth Pike was teaching there
Were you one of his people? 10:04
Question for TimToady when he wakes up :)
s/people/pupils/
oops, my english is so broken 10:05
masak .oO( minions )
moritz masak: reading the failed patch I have no idea how it tried to do what it should do
masak moritz: well, it was essentially a blind copy-and-paste of the things that worked for context-anchoring the quasi. 10:06
moritz: I might start by explaining to you how that works, in fact.
moritz yes please
10:06 sivoais joined
moritz it doesn't work that way the because context-anchoring the quasi is simply anchoring to the outer scope of the quasi, which is the state that $*W currently reflects 10:07
but for context-anchoring the unquote you need the context where the AST came from
10:08 hypolin left
masak right. 10:08
moritz so, not the same thing at all
10:08 Guest38681 left
masak I think I need to do something when passing in the arguments, not where I'm doing it there. 10:08
i.e. at (2)
moritz that sounds sane 10:09
masak let's jump straight to the deep-weird magic: github.com/rakudo/rakudo/blob/nom/...ld.pm#L837 10:10
well, github.com/rakudo/rakudo/blob/nom/...ld.pm#L834 , really. 10:11
moritz looks... overwhelming at first, sensible but opaque at second glance 10:12
masak this $*W method, add_quasi_fixups, takes a Perl6::AST and a block (an actual runtime block, I think), and says "hey $block, your outer is now what this Perl6::AST tells you it is!"
in a sense, this is the invasive surgery version for ASTs of what falls out naturally with closures. 10:13
again, the reason this has to be done is that AST fragments being passed back and forth between a macro and its calling context aren't meant to be evaluated in their surrounding context, but in the context where they were first built. 10:14
and before macros, this never happens, because AST fragments don't move around like that in a pre-macros world. 10:15
hm. correction. $block is a QAST::Block. 10:16
10:16 spider-mario joined
moritz right, doesn't need to be an actual Perl 6 level block 10:16
masak jnthn spoke about the need to simultaneously think of code as text, AST, and runtime objects. I feel that now :)
here's one of the three loci where $*W.add_quasi_fixups is called: github.com/rakudo/rakudo/blob/nom/...s.pm#L3440 10:17
10:18 dayangkun left
masak (three loci because there are three ways to call a macro: paren call, listop call, and operator call. the code ought probably be unified into one code path.) 10:18
as you see, that's pretty straightforward as well. the lines above wrap the QAST from the Perl6::AST into a block. the line in question "blesses" the block with the right context. 10:19
masak is a bit influenced by Io right now, and feels that the line between prototypal inheritance and lexical scoping has blurred 10:20
moritz because both walk (outer/parent) chains? 10:21
10:21 sivoais left
masak right. inner scopes essentially perform a sort of differential inheritance on outer scopes. 10:21
now, if instead of blindly copy-pasting code, I would attempt to blindly transfer the ideas across the itemized list of the gist... :) 10:22
at (2), I need to correctly set the $!quasi_context of the Perl6::AST. man, that's a misleading attribute name.
moritz it is, if you reuse it for that purpose 10:23
masak at (3), I need to either make a call to $*W.add_quasi_fixups, or at least do the corresponding thing to what that method does.
the method call is also misnamed. this is not about quasis at all.
moritz I wonder if you can't do the call to add_quasi_fixups in (2) too 10:24
masak I don't think so. I need to carry code around in a Perl6::AST until it's time to unpack it.
and add_quasi_fixups is executed at unpack-time. 10:25
moritz but in principle you already have all the necessary information in step (2), right?
masak so, here's a better way to highlight the symmetry between macro arguments and macro return values: both are packed, and both are unpacked.
so the order is: pack arguments, unpack arguments, pack return value, unpack return value. 10:26
moritz so it's only a question of implementation where you put the 'unpack arguments' step
masak moritz: yes. but there's some cloning of QAST trees going on here, so I don't think this is an over-complicated solution.
moritz ok, I won't press the point 10:27
sjn \o 10:28
moritz: any thoughts on architecture.html? :)
masak moritz: the two {{{$code}}} insertions are a good example. 10:30
moritz sjn: sorry, I totally forgot about that
sjn has at least one typo fix in the pipeline
masak moritz: it turns into two identical but distinct AST fragments.
moritz masak: oh, I didn't realize there were other operations that might not commute with the context fixing
"identical but distinct"? 10:31
did you mean s/identical/equivalent/ ?
10:32 brrt left
masak heh, yes. :) I hesitated for exactly the same reason when I wrote that. 10:33
it's two equivalent subtrees with distinct-enough identities.
(i.e. we can cheat and not clone everything if we know we can get away with it)
10:34 leont left
masak moritz: unless it's perfectly clear, this discussion helped *a lot*. thank you. I know what to try now. 10:34
10:34 sivoais joined
moritz you're very welcome 10:35
I learned that thunking with macros is easier than I feared
and a lost of other details
masak heh, "a lost of other details" is how I feel about macros, too :P 10:36
masak .oO( how to get unstuck: (1) make a gist with a concrete example, (2) discuss it with moritz++, (3) ???, (4) profit! )
Juerd
.oO( WD40 )
 10:41
10:43 M_o_C joined 10:48 pat_js joined
moritz lunsj 10:48
tadzik hello Ronja :) 10:49
masak no, that's Norwegian ;)
pat_js hey, how can i make hash slices in perl6 (equvalent to p5's @hash{qw[hello what is up]}
masak r: my %hash = <hello 1 what 2 is 3 up 4 yo 5>; say %hash{<hello what is up>}.perl 10:50
p6eval rakudo cc1858: OUTPUT«("1", "2", "3", "4")␤»
masak that's one way. 10:51
r: my %hash = <hello 1 what 2 is 3 up 4 yo 5>; say %hash<hello what is up>.perl
p6eval rakudo cc1858: OUTPUT«("1", "2", "3", "4")␤»
masak oh, that works too, of course.
pat_js thanks, ok my problem is an other one.
masak so the answer is, "just put several space-separated things in the <>"
pat_js: hehe. X/Y. :P
pat_js yeah thought that, but forgot that p6-objects aren't hashes anymore 10:52
10:53 hoelzro|away is now known as hoelzro
masak they may be, but you're not supposed to care anymore. 10:53
that is, objects are opaque even if their repr is just a normal hash. 10:54
pat_js i wanted to do sth. like DateTime.now<year month day hour minute second>»...
masak r: say .year, .month, .day, .hour, .minute, .second given DateTime.now 10:55
p6eval rakudo cc1858: OUTPUT«201295125520␤»
masak r: (.say for .year, .month, .day, .hour, .minute, .second) given DateTime.now
p6eval rakudo cc1858: OUTPUT«2012␤9␤5␤12␤55␤49␤»
cognominal TimToady bigthink.com/ideas/21744 I suppose the transcription is wrong and you meant Cobol? "the old Cobalt (sic) language sort of had stock phrases that you plugged things into it, that was sort of cargo-cult natural language". Also the assertion should stand for AppleTalk? 10:56
masak r: my $d = DateTime.new; say $d."$_"() for <year month day hour minute second>
p6eval rakudo cc1858: OUTPUT«Int()␤1␤1␤0␤0␤0␤»
masak hrm.
r: my $d = DateTime.new; say $_ for <year month day hour minute second> 10:57
p6eval rakudo cc1858: OUTPUT«year␤month␤day␤hour␤minute␤second␤»
masak why doesn't the `$d."$_"()` variant work? 10:58
10:59 daxim joined
dalek kudo/macros-d2: fa46512 | masak++ | src/ (2 files):
Revert "halfway commit, does not work"

This reverts commit 1ac5431b62bd313bde32417e387881584ea291f3.		 11:00
11:00 M_o_C left
masak oh! 11:00
r: my $d = DateTime.now; say $_ for <year month day hour minute second> 11:01
p6eval rakudo cc1858: OUTPUT«year␤month␤day␤hour␤minute␤second␤»
masak r: my $d = DateTime.now; say $d."$_"() for <year month day hour minute second>
p6eval rakudo cc1858: OUTPUT«2012␤9␤5␤13␤1␤11␤»
masak simple s/now/new/ typo :)
pat_js p6eval: (given DateTime.now {.year,.month,.day,.hour,.minute,.second}).join('-') 11:02
but i could just use DateTime.now.Str instead 11:03
masak r: say Date.new
p6eval rakudo cc1858: OUTPUT«2012-12-24␤»
masak moritz: it's a bit weird that we have this default for Date.new, but not for DateTime.new
r: say DateTime.new.Date 11:04
p6eval rakudo cc1858: OUTPUT«Cannot unbox a type object as a native int␤ in any at src/gen/BOOTSTRAP.pm:97␤ in sub sprintf at src/gen/CORE.setting:2106␤ in method gist at src/gen/CORE.setting:11087␤ in sub say at src/gen/CORE.setting:7140␤ in block at /tmp/Y2qX9eZb0D:1␤␤»
masak oh wow.
we have some ways to go to get proper type conversion.
a DateTime should effortlessly convert to a Date, IMO.
masak submits rakudobug
11:13 Gesh left 11:15 Gesh joined 11:21 JimmyZ joined
moritz r: say DateTime.new 11:28
p6eval rakudo cc1858: OUTPUT«use of uninitialized value of type Int in string context in block at /tmp/7kDir3GHgy:1␤␤DateTime.new(year => , month => 1, day => 1, hour => 0, minute => 0, second => 0/1)␤»
11:32 sisar left
moritz sjn: re architecture.html, looks good 11:34
it still doesn't quite capture how actions and grammar are interleaved, but it's better than before 11:36
11:36 hoelzro is now known as hoelzro|away 11:37 Cofyc left
sjn moritz: I'd love to look into that, if I can have someone's time to probe and ask stupid quesions :) 11:41
moritz "The Parser creates a parse tree out of the Perl 6 source code and then gives control to appropriate action methods that annotate the parse tree" 11:43
Might be better to say "For each syntactic rule, the parser creates a ..." 11:44
11:44 brrt joined
moritz so that it's clear that the scope for that step is rather small 11:44
11:44 jaldhar_ is now known as jaldhar
moritz the next sentence makes it clearer, but I think by that time the reader already has an inaccurate image in his head 11:44
FROGGS r: say DateTime.new( now ) 11:45
p6eval rakudo cc1858: OUTPUT«DateTime.new(year => 2012, month => 9, day => 5, hour => 11, minute => 45, second => 19.640998840332e0)␤»
FROGGS thats what I would expect to get when just calling new
moritz r: say DateTime.now 11:46
p6eval rakudo cc1858: OUTPUT«DateTime.new(year => 2012, month => 9, day => 5, hour => 13, minute => 46, second => 0, timezone => $*TZ)␤»
moritz that's what I would expect you to write if that's what you mean
masak hm, micro-nit: "parser" should not be spelled with a capital "P", unless it's revered or holy in some way.
moritz aye
FROGGS but IMO DateTime.new should be the same like DateTime.now, when passing no args 11:47
11:47 JimmyZ left
masak same argument could be made for Date.new and Date.today, I guess. 11:48
pmichaud good morning, #perl6
FROGGS r: say Date.new
masak pmichaud! \o/
p6eval rakudo cc1858: OUTPUT«2012-12-24␤»
masak so there are three symmetry relations here.
FROGGS r: say Date.today
p6eval rakudo cc1858: OUTPUT«2012-09-05␤»
masak Date.new<->Date.today, DateTime.new<->DateTime.now, and Date.new<->DateTime.new 11:49
FROGGS Date.new gives me december the 24th?
moritz next Christmas eve
FROGGS but why?
masak seems to me we need either back down on Date.new being next Christmas eve (and make it today), or make DateTime.new next Christmas eve, too.
moritz a Christmas egg :-) 11:50
masak FROGGS: because you didn't specify the date.
moritz FROGGS: you asked for a Date, and I gave you one
FROGGS pah
so Int x; might give me 42?
moritz no, there's a very precise spec for that
masak FROGGS: Int has a natural origin. Date doesn't.
moritz nr: my Int $x; say $x.perl 11:51
p6eval rakudo cc1858, niecza v21-1-ga8aa70b: OUTPUT«Int␤»
moritz nr: my Date $x; say $x.perl
p6eval rakudo cc1858: OUTPUT«Date␤»
..niecza v21-1-ga8aa70b: OUTPUT«===SORRY!===␤␤Malformed my at /tmp/UDQQ0IFpmI line 1:␤------> my⏏ Date $x; say $x.perl␤␤Parse failed␤␤»
masak FROGGS: it's a slightly frivolous way to discourage use of Date.new -- we consider Date.today to be clearer.
FROGGS well, somebody could say that beginning of unix epoch might be the natural origin...
(for unix ppl)
masak FROGGS: somebody else could say that the birth of Mohammed might be the natural origin. for Muslim ppl. 11:52
pmichaud the beginning of the Perl epoch might be more natural. :-)
FROGGS hehe, ya
moritz and so we return the old joke about the Perl 6 release date
masak FROGGS: or the formation of the solar system. that's a pretty special date.
moritz or of the Universe
masak r: say Date.new(-13e9, 1, 1)
p6eval rakudo cc1858: OUTPUT«Type check failed in assignment to '$!year'; expected 'Int' but got 'Num'␤ in block at src/gen/CORE.setting:656␤ in method BUILDALL at src/gen/CORE.setting:640␤ in method bless at src/gen/CORE.setting:630␤ in method new at src/gen/CORE.setting:11005␤ in metho… 11:53
FROGGS its just that "time" gives me the current time, I expected DateTime.new to give me the same
masak heh :)
moritz we just don't want to change our code whenever there's a new scientific paper on the age of either
masak r: say Date.new(-13_000_000_000, 1, 1)
p6eval rakudo cc1858: OUTPUT«-115098112-01-01␤»
masak o.O
moritz r: say Date.new(-13_000_000_000, 1, 1).day-of-week
p6eval rakudo cc1858: OUTPUT«7␤»
masak haha
masak .oO( and God rested )
FROGGS *g*
masak I suppose -115098112 above is a bug? 11:54
r: say -13_000_000_000
p6eval rakudo cc1858: OUTPUT«-13000000000␤»
masak submits rakudobug
flussence tries "printf '@%024d' 0 | tai64nlocal" and gets a segfault ...guess 96 bits isn't enough for everyone. :) 11:57
12:02 mtk left 12:04 lizmat joined 12:09 mtk joined 12:14 jaldhar left
masak ah yes. 12:21
add_macro_arguments in Perl6::Actions creates Perl6::AST objects, and gives them $!past attribute values, but not $!quasi_context attribute values.
from a more modern, enlightened viewpoint, that is clearly Wrong. 12:22
12:23 sisar joined
mst POSTing upload for Rakudo-Star-2012.08_001.tar.gz 12:24
now with more installing to the right f*cking place :) 12:25
masak mst++
12:28 spider-mario left
dalek osystem: b2abe53 | (Yecheng Fu)++ | META.list:
Add Redis module.
 12:31
osystem: 66171fe | tadzik++ | META.list:
Merge pull request #11 from Cofyc/redis

Add Redis module.
masak whoa, redis! 12:32
is Yecheng Fu on #perl6?
tadzik we now have 2 redis drivers, but this one is said to work on new Star :)
12:35 dayangkun joined
[Coke] MikeFair_: (error) the partcl-nqp/nqp2 branch is now building, yes. (it only understands one command, "three", which returns a 3. more to come.) 12:36
tadzik :)
moritz but it's a very useful command :-) 12:37
flussence that reminds me, I've got a perfectly good repo sitting all alone on github that I've been meaning to add...
moritz considers uploading Acme::Three 12:38
dalek osystem: 92e7ceb | flussence++ | META.list:
Add XMMS2 module
moritz nr: class My::3ish { } 12:39
p6eval rakudo cc1858: OUTPUT«===SORRY!===␤Name My:: ends with '::' and cannot be used as a package name␤»
..niecza v21-1-ga8aa70b: OUTPUT«===SORRY!===␤␤Illegal explicit declaration of a symbol table at /tmp/TeAuZ9VK8_ line 1:␤------> class My::⏏3ish { }␤␤Unable to parse class definition at /tmp/TeAuZ9VK8_ line 1:␤------> class My::⏏3ish { }…
moritz eval: package My::3ish;
buubot_backup moritz: No output.
moritz p5 allows that
moritz learned that while appling for a job at 123people.com
masak moritz: it was an interview question? 12:40
well, with that company name, they should know :P
moritz masak: no, just normal conversation 12:41
12:41 PacoAir joined
flussence I'm curious, has anything in an Acme:: module ever been promoted into core p5? :) 12:42
12:43 sbp joined 12:44 sbp left
moritz I don't think so 12:46
but last I looked, there were two Acme:: modules included in Debian stable
12:47 Targen left 12:49 snearch joined
moritz chases a change in behavior of a CPAN module 12:52
the release that changed in an incompatible way was issues in 2008.
12:56 ijkkk joined 12:58 guest01 joined 12:59 ijkkk left
[Coke] oh, redis++ . we are now one step closer to letting me use perl6 at work. ;) 12:59
colomon \o/ 13:01
13:01 guest01 left
[Coke] nqp: pir::sleep__vN(1) 13:01
p6eval nqp: ( no output )
[Coke] I am stunned this works, as: github.com/partcl/partcl-nqp/commit/5a368f7cfc was needed to make that file compile. 13:02
the --target=pir worked but trying to "use" it later failed.
colomon is taking his $work p6 involvement to new levels with github.com/colomon/perl6-ISO_10303-21 and wishes he had time to do a blog post on it.
masak is making progress 13:04
FROGGS progress++
13:06 kaleem left
masak 'maximum recursion depth exceeded' -- dang. seems I jinxed the progress. 13:06
sjn so...
Anyone want to give me a short and sensible (beginner-level) definition of what a "parse tree" is? :)
arnsholt It's a tree describing a parse O:) 13:07
sjn hehe
arnsholt But more seriously
PerlJam sjn: did you ever have t o do those sentence diagrams in English class where they said "this is the noun, this the verb, etc." ?
arnsholt Given a rule A -> B C D, the root node will be labelled A and have three children labelled B, C and D 13:08
sjn a parsemonius definition if there ever was one :)
PerlJam: nah
arnsholt Essentially, it describes the decisions the parser made to create the parse
masak sjn: imagine the string to be parsed divided into "regions" of what goes with what. like "5 + 5 * 10" gets divided like this: "(5 + (5 * 10))". the tree is just a representation of that definition. 13:09
s/definition/division/
so, from a procedural point of view, the parse tree tells you which pieces of the text are related. 13:10
sjn vaguely remembers polish notation being a good way of illustrating the tree stucture in code. E.g. (+ 5 (* 5 10)) 13:11
masak an AST for the above would be something like :add[ :const(5), :multiply[ :const(5), :const(10) ]]
sjn lisp++ # teachability
masak sjn: aye, what you just wrote and what I just wrote are isomorphic.
frettled RPN++ # practical for calculations when you've learned it 13:12
But now sjn++ got me wondering what reverse Hungarian notation would look like.
colomon sumPI 13:13
13:13 bluescreen10 joined
colomon .... actually, I kind of prefer that to normal Hungarian notation. 13:13
sjn hungarian notation, eh :-6 13:14
sjn can't imagine reversing it would ever improve it :) 13:15
masak frettled: reverse hungarian notation must be what some BASIC dialects use for typing: A$ for strings, etc.
moritz double-reversed hungarian notation \o/
frettled ho ho ho
moritz: what about out-of-order Hungarian notation? 13:16
13:16 PacoAir left
moritz if it's out of order, I don't care for it. I have enough trouble keeping up with all the working notations :-) 13:17
frettled Massively parallel notation? 13:18
masak moritz: how do I refer to an nqp class such as QAST::Block from the setting? is there prior art?
13:18 PacoAir joined, PacoAir left
moritz masak: erm, you (almost certainly) don't. What are you trying to do? 13:18
13:19 PacoAir joined
frettled sjn: on a more serious note: is that explanation going to be made for Norwegians, native English-speakers, non-native English-speakers, or some other language group? And is that going to be a written explanation, an illustrated one, or a verbal explanation? 13:19
verbal? yikes, Norwenglish FTW.
masak in the method evaluate_unquotes in core/AST.pm, I need to wrap the QAST contained in $!past in a QAST::Block and a QAST::Stmts before I pass it on to $!past.evaluate_unquotes 13:20
moritz: this is the "unpack arguments" step I mentioned above.
i.e. as part of the unpacking, I need to wrap the QAST tree in a QAST::Block, and twiddle the context of that QAST::Block. 13:21
moritz masak: make the wrapping a method in whatever object is in $!past ?
sergot I need some help. :)
moritz (maybe QAST::Node)
masak moritz: there are arbitrary QAST objects in $!past.
moritz: that sounds like a good idea.
frettled sjn: I think Wikipedia's explanation is fairly okay, FWIW: en.wikipedia.org/wiki/Parse_tree
13:22 pat_js left
moritz sergot: then shout HALP! HALP! and run around in circles 13:22
or ask your question here :-)
masak moritz: there's also a call to $*W.add_quasi_fixups in there. that's the twiddling part.
moritz: calling $*W from core feels even more wrong.
moritz: but I might simply inline that logic.
sergot I'm working on an article titled "Who may I speak with?" - where I'll write some information about people connected to Perl 6 and theirs responsibilities. 13:23
e.g.
PerlJam sergot++
sergot Juerd, moritz are responsible for feathers.
masak sorear: "feather", not "feathers" :)
Juerd Responsible? :|
moritz feathers
PerlJam masak: there is more than one :)
moritz there are 3.5 of them 13:24
masak fairenuf.
Juerd moritz: .5?
masak sorear: oops, sorry, mistab. :/
moritz Juerd: I count the host for the VMs as .5
sergot "You my talk to them to ... get an account on feather.perl6.nl"
Juerd moritz: That one isn't feather specific anymore.
moritz Juerd: oh, so 3 it is
Juerd moritz: Feather[123] now run on a shared virtual hosting platform
masak sergot: also, the title sounds slightly off to my (non-native) ears. maybe "Who can I talk to?" 13:25
sirrobert hi =)
tadzik or just "who is who"
masak sergot: "Who may I speak with?" sounds like we mostly don't allow people to speak to us :)
sergot masak: Yes, that sounds better. :)
moritz shouldn't that be "Whom can I talk to?" :-)
\o sirrobert
sergot o/ sirrobert 13:26
masak moritz: yes, in a world where people still care about the word "whom" :P
PerlJam How about just "Help!" :)
masak "Halp!"
sergot nevertheless that's not the point. :D
masak "People who know stuff"
PerlJam People. Stuff. Perl 6.
sergot You know better than me who is responsible for. :) 13:27
Juerd What's this in /etc/sudoers?
# for some community games and stuff
ALL ALL=(nobody) ALL
daxim Butterfly. Onion. Beer.
masak sergot: you came with an idea for a useful document and expected us *not* to bikeshed over the name? :P
daxim: tits. onion. trash.
sirrobert heh
masak daxim: FIVE!
PerlJam more like punning than bikeshedding.
But there's a fine line.
13:28 marmay left
Juerd I think the fine line should be green 13:28
PerlJam blue!
Juerd Compromise on teal? 13:29
PerlJam :)
flussence oh wow, rakudo just compiled without me having to hack the makefile!
Juerd Blog it! 13:30
sergot masak: of course You can do this. :P
masak: I just wanted to start writing. :)
arnsholt flussence: Which platform have you been having trouble on?
flussence x86-32 gentoo with 1GB RAM 13:31
PerlJam "And then there's sergot ... he just wanted to start writing and now look what's happened!"
sirrobert it's extremely awesome that eval closes over sub refs
++whoever-did-that
karma whoever-did-that
aloha whoever-did-that has karma of 0.
sjn frettled: the explanation is for $anyone who's interested in rakudo/docs/architecture.html :)
moritz sirrobert: we only give out pre-increment karma for stuff people have promised to do 13:32
flussence (not really 1GB either, since chrome's eating 150MB of that)
sirrobert moritz: ahh =)
moritz sirrobert: and postfix ++ for things that have actually been done
masak sirrobert: (re eval closing over sub refs) we do like lexical scoping in here.
sirrobert heh
flussence who do I postfix:<++> for having a working perl6 again? :D
13:32 sisar left
moritz flussence: jnthn++ found a fix for some GC trouble in NQP 13:32
flussence jnthn++
arnsholt Right. 1 gig of RAM, I can see how that might be tricky =) 13:33
sirrobert fluessence: did it compile with <1GB?
arnsholt 32-bit has smaller pointers, so that'll help a bit compared to 64-bit, but I can imagine it'd be a bit cramped for physical memory 13:34
flussence sirrobert: it used to, but it was broken for a few months after parrot 4.$someversion 13:35
13:35 GlitchMr joined
sjn wonders if the glossary definition for NQP could be improved: "NQP: Not Quite Perl 6, a small subset of Perl 6 that is used for tree transformations in compilers." 13:35
flussence it just worked now
sirrobert fluessence: nice! That's been a little hangup for a co-worker of mine 13:36
mst sjn: you there
sirrobert was compiling on my computer with a matching disk image and porting it over
sjn Where can one read up on what NQP can do?
mst: I'm here! \o/
mst sjn: cpan/cpanm trout.me.uk/perl/Rakudo-Star-2012.08_001.tar.gz please
moritz in the source
mst sjn: then try 'eval $(perl -MRakudo::Star::Paths); perl6'
sjn: and see if it works
sjn clever :) 13:37
sjn checks
13:39 Chillance joined
sjn moritz: hm. nothing in prose? :-/ 13:39
sergot moritz: who else is responsible for feather.perl6.nl 13:42
mst sjn: I'm hoping 002 can lose the _ :) 13:43
masak moritz: ugh. adding that method to QAST::Node feels Wrong for two reasons. (1) the classes QAST::Block and QAST::Stmts need to be stubbed in QAST::Node's file, which is at least a code smell.
PerlJam sergot: define "responsible"
masak (2) there are ties to Perl6::World's .add_fixup_task, which I don't seem to be able to get rid of.
PerlJam sergot: in some sense everyone who has an account on feather is responsible for it. 13:44
sergot PerlJam: I mean who can create an account there. :)
arnsholt sjn: I'm not aware of any significant docs for NQP. When I've been dabbling it's been mostly finding similar stuff in the NQP or Rakudo sources
sjn hm
ok
PerlJam Didn't pmichaud start an NQP book at one point? 13:45
moritz PerlJam: yes, but last I looked it was empty
masak: point taken
mst PerlJam: could I put upon you to test the dist as well please?
sergot The same with perl6 organization on github - who can give an access?
13:46 fhelmberger left
PerlJam mst: sure 13:46
13:46 stopbit joined
moritz masak: other code from the setting that calls back into the compiler goes through pir::compreg__Ps('perl6') (which gets the Perl6::Compiler object). Not sure if that's of any help 13:46
mst PerlJam: preferably into a perlbrewed perl since that's what went wrong for leont yesterday
PerlJam: I'm pretty sure local::lib installs are fine
moritz sergot: (re perl6 organization) iirc it's TimToady, pmichaud, jnthn, sorear and me 13:47
and it used to work via hugme, but I think it bitrotted
masak moritz: thanks. I might have a look at how those other bits do it, and if I can do the same. 13:49
moritz: I'm very much inclined to make something work now, and put the bits in more appropriate places later.
moritz masak: fwiw I'm aware of two such bits, &eval (in src/core/control.pm) and the backtrace printer (src/core/Exception.pm) 13:50
masak oh, ok.
moritz another evil idea
write a sub that does in the actions
sergot moritz++ thanks
moritz and assign it to a contextual variable
and invoke that from within the setting
masak could work. 13:54
moritz++
kind of like a dynamically accessed closure. 13:55
13:55 Chillance left, skids joined
moritz don't tell anybody it was my idea :-) 13:56
sjn quick! kill the irc logger! :) 13:57
13:58 FROGGS__ left, FROGGS left
sergot Who is the head rakudo's developer? :) 13:59
13:59 gongyiliao left, FROGGS__ joined
sirrobert moritz? 14:00
I think
[Coke] I think it's a tie between jnthn & pmichaud.
sirrobert oh =)
[Coke] moritz++ is up there.
sergot So, those 3? :)
these(?)
sirrobert those
14:01 tokuhiro_ joined
moritz sergot: pmichaud is 14:01
14:02 benabik left
[Coke] huh. I just got an offer for 100$ of free advertising from google. Wonder if I can use it for perl6. 14:02
mst sjn: still installing?
sergot moritz: only he?
kresike bye folks
14:02 Chillance joined, kresike left, nickA joined
moritz sergot: well, he's the project leader 14:02
and pmichaud and jnthn implement most of the features 14:03
I make many commits too, but mostly small (or mechanic) stuff
sjn mst: yeah, my computer is slow :-\
mst sjn: 's cool, was just checking
sjn: I'll have to bounce shortly but /msg me the outcome :)
masak sergot: pmichaud is the project leader. during the later parts of the nom refactor, and the time following it, jnthn was doing most of the commits on Rakudo. 14:04
sjn running tests now
sergot What is masak "responsible" for? :)
14:04 dayangkun left
moritz breaking stuff 14:05
and submitting reports
and macros
mst sjn: ooh, then it's almost done :)
nickA hi perl6.what's the diffrence between #my ($a,$b) = ('a','b')>>.IO>>.open;
my ($a,$b) = ('a','b')>>.IO.open;
mst sergot: we prefer to say "blamed" ;)
sergot hah, so s/responsible/blamed/ :)
moritz nickA: the latter calls .open on a list of IO objects
masak sergot: I'm more irresponsible than responsible, actually. 14:06
moritz nickA: and fails, because there's no such method open
sergot And for YAPSI, right? :)
nickA moritz:why it fail?my ($a,$b) = ('a','b')>>.IO.open;
masak sergot: yeah. I should probably get my vast development team to start releasing Yapsi regularly again. 14:07
sergot masak++ : very good idea. :)
sirrobert nickA: ('a','b')>>.IO creates a list of IO objects after 'a' and 'b'
GlitchMr nickA: It means my ($a, $b) = (('a', 'b')>>.IO).open
moritz nickA: because there's no method 'open' in class List
[Coke] nqp: say(nqp::new('OS').'cwd'();
p6eval nqp: OUTPUT«Confused at line 2, near "say(nqp::n"␤current instr.: 'panic' pc 19998 (src/stage2/gen/NQPHLL.pir:7314) (src/stage2/gen/NQPHLL.pm:325)␤»
sirrobert nickA: .open is not a list method
GlitchMr So it .opens list, not IO
[Coke] nqp: say(nqp::new('OS').'cwd'()); 14:08
p6eval nqp: OUTPUT«Error while compiling block : Error while compiling op call: Error while compiling op callmethod: Error while compiling op new: No registered operation handler for 'new'␤current instr.: '' pc 42779 (src/stage2/QAST.pir:14984) (src/stage2/QAST.nqp:2244)␤»
moritz nqp: say(pir::new__Ps('OS').cwd())
p6eval nqp: OUTPUT«Class 'OS' not found␤current instr.: '' pc 49 ((file unknown):161309279) (/tmp/19wb6fbgKr:7)␤»
[Coke] nqp: say(nqp::new('OS').'cwd'()); INIT { nqp::loadlib('OS'); }
p6eval nqp: OUTPUT«Error while compiling block : Error while compiling op loadlib: No registered operation handler for 'loadlib'␤current instr.: '' pc 42779 (src/stage2/QAST.pir:14984) (src/stage2/QAST.nqp:2244)␤»
moritz nqp: say(pir::new__Ps('os').cwd())
p6eval nqp: OUTPUT«Class 'os' not found␤current instr.: '' pc 49 ((file unknown):167322207) (/tmp/blU32xIoi_:8)␤»
moritz nqp: nqp::loadlib('os'); say(pir::new__Ps('OS').cwd())
p6eval nqp: OUTPUT«Error while compiling block : Error while compiling op loadlib: No registered operation handler for 'loadlib'␤current instr.: '' pc 42779 (src/stage2/QAST.pir:14984) (src/stage2/QAST.nqp:2244)␤»
moritz nqp: pir::loadlib__vs('os'); say(pir::new__Ps('OS').cwd()) 14:09
p6eval nqp: OUTPUT«error:imcc:The opcode 'loadlib_sc' (loadlib<1>) was not found. Check the type and number of the arguments␤ in file '(file unknown)' line 167980855␤»
GlitchMr ('a', 'b')>>.IO.open working as ('a', 'b')>>.IO>>.open would be nice, but it's probably too late for that
mst sjn: tests finished?
[Coke] moritz: fun, innit! ;)
moritz nqp: pir::loadlib__Ps('os'); say(pir::new__Ps('OS').cwd())
p6eval nqp: OUTPUT«/home/p6eval␤»
moritz [Coke]: there ya go
[Coke]: I guess there won't be nqp:: opcodes for that, because it's very parrot-specific
sjn mst: yeah, installing now 14:10
nickA any GlitchMr:why you say that?
mst sjn: \o/
PerlJam mst: Successfully installed Rakudo-Star-2012.08_001
nickA GlitchMr:why you say that?
sergot Is there any person who is blamed for official Perl 6 doc? TimToady? :) 14:11
PerlJam mst: that's on my perlbrewed perl 5.16.0
mst PerlJam: right. does 'eval $(perl -MRakudo::Star::Paths); perl6' work?
14:11 oops left, FROGGS joined
moritz sergot: I'm irresponsible for doc.perl6.org, kinda 14:11
PerlJam mst: yep
sjn mst: it works! :)
GlitchMr Rakudo::Star on CPAN?
mst \o/
GlitchMr: that's what I've been working on :)
PerlJam mst++ quite awesome. 14:12
GlitchMr Well, I find it nice... but Rakudo::Star doesn't have anything with Perl 5, does it?
PerlJam GlitchMr: yet
frettled mst++ very nice
mst GlitchMr: I'm going to be writing an perl.org/Object::Remote backend for rakudo 14:13
GlitchMr: so you can do Some::Perl6::Class->new::on(v6, @constructor_args);
GlitchMr: I figured that that would be more fun if people could get all their deps via CPAN, including rakudo
GlitchMr oh, ok
This would be fun
masak "Namespaces are obsolete". a programming language design opinion. weblogs.asp.net/bleroy/archive/2012...olete.aspx -- what do y'all think? 14:14
sjn mst: perl6-debug works too
mst sjn: I've not done anything particularly clever; if it works at all, I'd hope it works entirely :)
sjn: however, please od play in case I'm wrong :)
[Coke] nqp: pir::loadlib__Ps('os'); say (nqp::new('OS').cwd()) 14:15
p6eval nqp: OUTPUT«Confused at line 2, near "say (nqp::"␤current instr.: 'panic' pc 19998 (src/stage2/gen/NQPHLL.pir:7314) (src/stage2/gen/NQPHLL.pm:325)␤»
sjn wonders where the Perl6 modules are installed
mst sjn: it's all under the Rakudo/Star/Install/ directory
GlitchMr Cannot you do stuff like import __future__ as __past__ in Python too?
[Coke] ah well.
moritz [Coke]: space after 'say' 14:17
nqp: pir::loadlib__Ps('os'); say(nqp::new('OS').cwd())
p6eval nqp: OUTPUT«Error while compiling block : Error while compiling op call: Error while compiling op callmethod: Error while compiling op new: No registered operation handler for 'new'␤current instr.: '' pc 42779 (src/stage2/QAST.pir:14984) (src/stage2/QAST.nqp:2244)␤»
14:17 mucker joined
sirrobert masak: I like the picture. I *really* like Node's package management as well. I'm not sure what the impact is on namespacing in general, but I do like the pattern a lot. 14:18
moritz masak: Perl 6 goes a bit into the "deprecate namspaces" direction by doing much more lexically than Perl 5 14:19
sergot Who is mostly irresponsible for p6eval? 14:20
moritz sergot: sorear and me (mostly me)
masak moritz: right. seems to me the post is advocating replacing namespaces by... lexical scoping. 14:21
nickA so ?any other code for this my ($a,$b) = ('a','b')>>.IO>>.open; if I want to write it one line? 14:22
mst you still need something approximating it though 14:23
[Coke] moritz: that's one 3 line file reclaimed. thanks. ;)
mst "I want to import X from Y" requires some way of finding Y
sirrobert r: :test
p6eval rakudo cc1858: ( no output )
sirrobert r: say :test.WHAT
p6eval rakudo cc1858: OUTPUT«Pair()␤»
sirrobert what's the intended purpose of that?
moritz [Coke]: you're welcome
masak mst: troo.
14:23 jaldhar joined
sirrobert r: say :test 14:24
p6eval rakudo cc1858: OUTPUT«␤»
moritz nickA: my ($a, $b) = <a b>.map: *.IO.open;
mst masak: I quite like doing 'use aliased "Some::Namespace::Containing::AClass";'
masak: and then doing AClass->new(...)
sirrobert err.. that it puts: 'test' => Bool::True
nickA moritz:wonderful!thanks!
14:24 oops joined
mst masak: that way the name is only used once 14:24
moritz sirrobert: note that :test is a named argument 14:25
nr: say (:test); # positional argument
p6eval rakudo cc1858, niecza v21-1-ga8aa70b: OUTPUT«"test" => Bool::True␤»
masak mst: nice! I want metacpan.org/module/aliased for Perl 6 now.
sirrobert moritz: used outside of that context it's the same?
I'm cool with that, just trying to understand =)
nickA bye perl6 ! thanks your help!
14:25 nickA left 14:26 tokuhiro_ left, atrodo joined
sirrobert moritz: ah, I see it better now. thanks 14:26
[Coke] BY THE POWER OF PERLSIX! 14:27
moritz sirrobert: the difference is only in argument lists
sirrobert ok 14:28
moritz masak: I'm pretty sure I could make that happen by patching rakudo. Not sure about a separate module.
14:29 wtw left
masak hm. 14:30
sergot Who is irresp* for perl6 wiki, p6 on rosettacode, perl6 book? :) 14:31
mst sergot: why does this matter?
sergot: generally it being an open source project "whoever's had time and thought it looked like fun to work on" 14:32
[Coke] the "who to talk to" can be summed up with "this irc channel" for now.
TimToady is mostly responsible for the p6 rosettacode stuff 14:33
masak I agree with mst when it comes to wikis. the idea of wikis is that anyone with tuits can be responsible for the content.
moritz sergot: uvtc did much of the wiki stuff. The perl 6 book is dead, I'm afraid
masak who is taking charge at the moment is largely ephemeral and irrelevant. maybe even harmful to note it down.
TimToady but there must generally be a champion until you get critical mass
masak aye. 14:34
sergot I know, but if someone new (begginer) or from "outside" wants to know or do something, he can ask for sth particular person. :)
that was to mst. :)
PerlJam sergot: if he shows up here, he/she can ask anyone and get help 14:35
sergot That's right.
TimToady which is why IRC is nice, because nobody can be on on the time, and some info is better than none
sergot If someone has an idea about panda, he can write straight to tadzik. 14:36
[Coke] and I'd rather folks came here with questions than pinged pmichaud directly.
TimToady there are many answers of the form "I think it works like this, but ask such and so for more"
[Coke] sergot: that's not always a good thing.
TimToady we try to turn the knob to a point where we rarely have to tell someone "I'm too busy" 14:37
[Coke] if you said "see timtoady about the spec" and folks did that directly, how much of his tuits does that eat into?
sergot So there's no need to write this article. :)
masak moritz: g'ah. figuring out where to declare the contextual closure hurts my brane. I think it should be declared at (2).
TimToady there's tight coupling and loose coupling
(with a developer, in the sense you can have a conversation, or just put something into their queue) 14:38
moritz masak: well, it must be in a caller of the routine where the wrapping occurs
masak: like, the location where the macro is called is a pretty safe bet 14:39
masak right. that's (2).
TimToady sergot: if the main point ends up explaining how the culture works rather than just being a Who's Who, it can be useful to name names as an example of how it works currently
masak moritz: I started defining it at (4), which made sense for a while, and then went "hey, waaaaitaminute..." 14:40
TimToady and to the extent that we are not interchangeable parts, each of us has to provide something that nobody else can
masak sergot: frettled once made a list of bots and who were responsible for each. it's in the blogosphere somewhere. 14:41
s/frettled/frettled++/
moritz it moved on the the perl6 wiki
wiki.perl6.org/IRC%20bots 14:42
seems to be a bit incomplete
sergot Great, thanks. :)
masak moritz: just found github.com/rakudo/rakudo/commit/39...7b#L0R3528 -- what would happen if that line wasn't changed. (trying to understand what it does) 14:44
moritz: once I understand what it does, I will probably add it to the two other macro-call code paths.
moritz masak: erm, which line? the hiliting doesn't seem to work here :( 14:46
FROGGS moritz: I guess first diff, fourth from bottom 14:49
3528 that is
masak moritz: my $quasi_ast := $*W.ex-handle($/, { $routine(|@argument_quasi_asts) }); 14:50
moritz: (the id part of the URL, #L0R3528, also gives the line)
moritz masak: it calls the block (i.e. invokes the routine), catches exceptions, adds line/file info, rethrows 14:51
masak aha.
so it's a kind of "add line/file info" exception wrapper? 14:52
moritz yes 14:53
(it adds a bit more info, but that's the essence) 14:54
anyway, that's the actual invocation of the macro body
masak right.
there's three of them in Actions, currently.
g'ah, another 'maximum recursion depth exceeded'! 14:56
[Coke] time to bump up the limit again, or are you going off into the weeds? 14:58
moritz I guess it's the latter
masak [Coke]: no, it's clearly a thinko on my part, considering the stacktrace.
furthermore, I suspect the stacktrace is of trying and failing to serialize an error message that occurs during compilation. because the routines in the stacktrace are stringifying methods, and my added code isn't doing any stringification. 15:00
it's exciting to be so close to having things working, though. 15:02
[Coke] masak++ # hey, is your grant done yet?
[Coke] hides.
moritz masak: --ll-exception 15:03
not sure if it'll help, but it's worth a shot
masak [Coke]: no, but after I get this working, I will consider it 50%, and take the appropriate blogging actions.
moritz: ah, good idea. 15:04
moritz masak: and look at the bottom of the backtrace
15:04 snearch left
moritz sometimes the calling location tells you what's wrong 15:05
FROGGS .oO( how it would look like at the bottom of the backtrace? )
masak moritz: this is during 'make'. 15:06
moritz masak: and?
arnsholt Hack the Makefile, I guess =)
moritz copy the command that fails
add an --ll-exception
masak ah, yes.
TimToady on passing strings to macros, it should be about as common as Action rules in the grammar dealing in strings rather than ASTs; that is, it happens occasionally, but only with when something forces it to be text explicitly rather than AST, which is the default 15:07
in fact, I think that thinking of macros as user-define action rules is probably a good thing 15:08
because that's more or less what they are
(I mean, that's what hygienic macros are) 15:09
which makes me thing that we might find some way of making P6 grammar actions look more like macros, once we figure out what macros look like :) 15:10
*think
I am pretty much convinced that the triple rule must die; they are very cluttery and actively confusing to the eye 15:12
masak hygienic != AST-based 15:13
moritz triple rule?
TimToady {{{$stuff}}}
masak --ll-exception didn't shed any new light :/
moritz masak: then try the bottom-of-the-stacktrace thing
masak but I know which line in src/core/AST.pm causes the infinite recursion: '$past := &*wrap_and_recontext($_);' 15:14
15:14 erkan joined, erkan left, erkan joined
masak I don't see why it should; it looks perfectly innocent to me. 15:14
hey, wait! probably needs a 'my'! :)
arnsholt I hate those bugs 15:15
moritz well, you probably also see what function it calls
arnsholt So hard to find >.<
moritz and one of those (compile-time) callees is what throws the error 15:16
masak that fixed it \o/ 15:17
moritz \o/
masak bah, I keep tripping on the three-code-paths-instead-of-one thinko myself the whole time. making it more urgent to actually unify them. 15:21
moritz The Mystery of the Three Code Paths, by Carl Masak 15:23
or maybe *Mastery
TimToady cognominal: Ken Pike was one of the founders of SIL, so I was indirectly his pupil. I only saw him for a day or so in real life, but he did a monolingual demonstration that was quite impressive
15:24 thou joined
TimToady they put him up with a speaker of some African language he'd never studied, and by the end of the hour was having rudimentary conversations with him 15:24
they both knew English, but neither spoke it 15:25
Pike spoke and took notes on the blackboard in Aztec
well, and IPA :) 15:26
15:26 jaldhar left
moritz wow, that's impressive 15:26
TimToady as for the relationship of SIL with tagmemics, well, Pike invented tagmemics, at least its early forms 15:27
some later philosophical creations went rather beyond his original vision, and I don't know that he entirely approved
he was mostly a practical linguist :)
cognominal :) 15:28
TimToady, thx for the answer
rjbs Wow. 15:30
TimToady but the important concepts in tagmemics are about how people actually think about language (and just as importantly, which bits they don't think about) 15:33
cognominal fishes don't theorize about water 15:34
masak is having much use for --ll-exception as he keeps debugging his still-not-working patch
TimToady when you use a word, that word has a number of different contexts, and people naturally work with all those contexts simultaneously
like multiple dimensions of water
15:36 crab2313 joined
TimToady Pike liked to use a particle/wave/field analogy for some of the most important contexts 15:36
cognominal btw did you see sigils like flexions in flexionnal language (with the proviso that with identifiers, there is no root to be flexed, just lexed maybe :) 15:37
TimToady I have no idea what you're saying. :) 15:38
cognominal anyway, they literally have a role in Perl 6.
BooK speaking of sigils, is that an invented word?
moritz no, sigil exists outside of programming too 15:39
flussence
.oO( but all words are invented! )
geekosaur @wn sigil
lambdabot No match for "sigil".
geekosaur bah
BooK oh right
got some wikipedia article about them 15:40
weird that I never found that before, or rather, that I don't remember finding it
frettled masak: I don't think that was me, that IRC bots thingy, but I could just be suffering from dementia. :)
TimToady sigil is a good example of a word that means one thing in the dictionary but we use it for something else
cognominal ho that's fussional language in english en.wikipedia.org/wiki/Fusional_language 15:41
TimToady we know "dog" is a noun, but we can also use it as a verb or as an adjective
PerlJam stamps a butterfly-shaped piece of wood into hot wax
BooK no only is perl stealing ideas from other languages, but also words from the dictionary!
TimToady we all steal words all the time 15:42
that's how language works, according to tagmemics
cognominal so english is less a fusional language than french because there is less fusion so more ambiguity when a word is taken alone 15:43
That's probably why the english wikipedia is so short compared to the french one. 15:44
PerlJam cognominal: I thought it was more because the french are obsessed with language more than english-speaking peoples 15:45
moritz or maybe because the English wikipedia simply has more contributors
15:46 kaare_ joined
cognominal anyway my analysis of sigil is wrong according to this criteria. Adding a sigil does not modify the "root" form of the identifier, so by this token perl is an analytic language en.wikipedia.org/wiki/Analytic_lang..._languages 15:47
anyway I have strayed far away from tagmemics
TimToady but in Perl 5, adding {} to $foo changes the root form to %foo, by one way of thinking 15:48
15:48 crab2313 left
TimToady but computers force us to be rather analytical at the best of times 15:49
15:49 hoelzro|away is now known as hoelzro
cognominal well, by root, I meant "foo" here, changing @ or % would be a grammatical way to convay the accessing action :) 15:50
TimToady that would be a highlander variable interpetation, which was proposed for Perl 6 but rejected
we include the sigil with the root, and bare 'foo' has no meaning 15:51
we can, however, use bare foo to mean &foo in some situations
or you can look at the other way around, 'foo' is a verb, and & is noun marker 15:52
cognominal btw I am tempted in Perl 6 when argument(s) of a function fun is/are string(s), to write fun< a b > 15:53
probably ambiguous because < is an operator
masak hm. so. it successfully finds and calls the magical dynamic subroutine &*wrap_and_recontext
but it fails inside of it, for unknown reasons.
the error I get is "get_attr_str() not implemented in class 'Undef'"
from a line which appears to be a nqp::getattr call. 15:54
cognominal r: sub fun { … }; fun< a b > 15:55
p6eval rakudo cc1858: OUTPUT«===SORRY!===␤Unable to parse blockoid, couldn't find final '}' at line 2, near "\u2026 }; fun< "␤»
masak oh, btw, this is no longer during compilation. macros-d2 compiles now. this is during the macro call.
cognominal: your client turns '...' into '…'.
cognominal: tell it not to do that :) 15:56
cognominal yup the error I got here was postcircumfix:<{ }> not defined for type Failure
so it parses "well" after all but fails later
TimToady it parses as (fun){'a','b'} 15:57
cognominal ha!
TimToady that would only have worked if we had unified function invocation with hash subscripting
masak or as fun()<a b> 15:58
15:58 brrt left
TimToady and if we had unified those, we probably couldn't have a decent multi-dimensional slice syntax for subscripting 15:59
(not that anyone implements the latter yet) 16:00
16:00 FROGGS left, FROGGS__ left
cognominal but you already carved a syntactical space for it :) 16:00
masak I wanted to unify hash/array lookup with argument passing in the language I'm designing, but I'm having serious second thoughts now. 16:01
cognominal masak, or we can apply the rule to make things easier to user and harder to the implement and treat … like the triple dot :) 16:03
*implementer
masak cognominal: I'd hate introducing a synonym there just because some people have silly IRC clients. 16:04
TimToady one could think of macro args vs return as a kind of CPS in the compiler 16:05
masak TimToady: they sure feel very symmetrical to me.
TimToady: that's been the insight of late -- "hey, there are *two* cases of AST insertions!" 16:06
we don't think of it with HTML templates, because those encompass whole HTML documents most of the time.
16:07 benabik joined, kaare__ joined, kaare_ left
TimToady hmm, if we replace {{{}}} with ¤, should that be called "ox" or "xo", I wonder... 16:08
(those are the chars in the compose key map, for those wondering)
dalek kudo/macros-d2: 7fb2693 | masak++ | src/ (2 files):
partial commit

almost working... getting this error:
  $ ./perl6 --ll-exception -e \
   'macro twice($code) { quasi { {{{$code}}}; {{{$code}}} } };
   my $counter = 0; twice $counter++; say $counter'
get_attr_str() not implemented in class 'Undef'
  [...]
masak I feel stuck. anyone want to build macros-d2 and help me debug?
16:09 hoelzro is now known as hoelzro|away
masak TimToady: 'ox' is easier to pronounce, and more in line with #perl6's fondness for animals. 16:09
TimToady
.oO(rename AST to Ox)
 16:11
16:11 marmay joined
daxim /usr/share/X11/locale/en_US.UTF-8/Compose suggests both the mnemonics <o> <x> and <x> <o> 16:12
TimToady which is why I was asking
or wondering, more precisely :)
masak :)
TimToady it's ^KCu in vim 16:13
16:13 benabik left
TimToady
.oO(macros are wired with copper)
 16:14
cognominal :) 16:15
TimToady and yes, I said "COBOL" 16:17
masak "New technologies come and go, but COBOL is forever." -- abstrusegoose.com/323 16:18
16:22 Vlavv joined
marmay Out of curiosity, is there something similar to .^methods for modules and subs? Something that lists all subs of a module? 16:22
masak r: module Foo { our sub foo { say "OH HAI" } }; say Foo.HOW.^name 16:23
p6eval rakudo cc1858: OUTPUT«Perl6::Metamodel::ModuleHOW␤»
masak r: module Foo { our sub foo { say "OH HAI" } }; say Foo.HOW.^methods 16:24
p6eval rakudo cc1858: OUTPUT«No such method 'gist' for invocant of type 'NQPRoutine'␤ in method gist at src/gen/CORE.setting:4730␤ in sub say at src/gen/CORE.setting:7140␤ in block at /tmp/l8YZYbApAG:1␤␤»
masak heh.
r: module Foo { our sub foo { say "OH HAI" } }; say Foo::
p6eval rakudo cc1858: OUTPUT«("\&foo" => sub foo() { ... }).hash␤»
masak r: module Foo { our sub foo { say "OH HAI" } }; say Foo::<foo> 16:25
p6eval rakudo cc1858: OUTPUT«Any()␤»
masak r: module Foo { our sub foo { say "OH HAI" } }; say (Foo::)<foo>
p6eval rakudo cc1858: OUTPUT«Any()␤»
masak r: module Foo { our sub foo { say "OH HAI" } }; say (Foo::).perl
p6eval rakudo cc1858: OUTPUT«("\&foo" => sub foo() { ... }).hash␤»
masak r: module Foo { our sub foo { say "OH HAI" } }; say (Foo::).keys
p6eval rakudo cc1858: OUTPUT«&foo␤»
masak marmay: like that?
marmay: only our subs, though. 16:26
marmay masak: Yes, thank you!
16:26 Vlavv_ joined 16:29 Vlavv left
dalek kudo/macros-d2: 8c85e7a | masak++ | src/Perl6/ (2 files):
fix embarrassing typo

Also comment out some code. Yes, I expect to rebase these commits later.		 16:30
masak this gets me a little further. 16:31
16:31 stopbit left
masak now the only problem is that the custome op perl6_get_outer_ctx gets a String where it expected a Parrot Sub. 16:31
custom*
which should be a relatively easy problem to track down.
yay! :) 16:32
16:32 stopbit joined
tadzik \o/ 16:32
masak my problems get smaller and smaller :)
TimToady wow, that's that only problem? cool!
everything else should be a piece of cake
pity I'm allergic to cake...
masak I would have been more afraid of that custom op, if I hadn't been sitting at the keyboard as it was created ;) 16:33
I would've said "if I hadn't authored it", but that feels a bit too strong.
GlitchMr perl6: IDENTIFICATION DIVISION. 16:34
masak for those interested, that custom op is essentially the unholy rite you have to perform in order to make AST behave as closures.
p6eval rakudo cc1858: OUTPUT«===SORRY!===␤Confused␤at /tmp/6Rcp4ggmEh:1␤»
..niecza v21-1-ga8aa70b: OUTPUT«===SORRY!===␤␤Confused at /tmp/7Ed_fvdhLd line 1:␤------> IDENTIFICATION DIVISION⏏.␤␤Undeclared names:␤ 'DIVISION' used at line 1␤ 'IDENTIFICATION' used at line 1␤␤Parse failed␤␤»
TimToady what about the rest of us? :P 16:35
masak for the rest of you, it's absolutely no cause for concern.
TimToady oh oh... 16:36
masak tries to treat the type error as a retrograde analysis problem 16:39
16:44 benabik joined, phenny joined, benabik left 16:45 leont joined 16:48 benabik joined
sorear .u 44 16:49
phenny U+4400 CJK UNIFIED IDEOGRAPH-4400 (䐀)
masak hm. that seemed to lead me somewhere. unexpected.
sorear phenny: tell sorear hi
phenny You can tell yourself that.
masak phenny++
sorear in query with sbp now. once phenny is operational, I'll have lambdabot leave 16:50
masak \o/ 16:51
sorear good * #perl6
masak sorear! \o/
16:51 FROGGS__ joined
masak heh. trying to print the offending value causes multi dispatch of 'print' to fail ;) 16:52
sorear hii 16:53
phenny sorear: 16:53Z <sbp> tell sorear this is a test. please let me know if it succeeded!
16:56 sbp joined
sorear um, wtf 16:58
I kicked phenny last night
phenny is in the channel now
I don't have a phenny join or nick-change-to-phenny in my log
sbp I made her join a few minutes ago, just after you asked
phenny!
phenny sbp!
sbp and she's operational 16:59
benabik Dr. Scott!
sbp benabik!
tadzik phenny!
phenny tadzik!
sbp .u circ kat tu
phenny U+32E1 CIRCLED KATAKANA TU (㋡)
sorear oh
.u circ kat tu
phenny U+32E1 CIRCLED KATAKANA TU (㋡)
sbp apologies to all here for the disruption last night
masak goes for a brief run
sorear sorry for dragging you into this just now, sbp
benabik sbp: What happened?
sorear the most recent problem: 17:00
phenny was spamming me in pm, so I did /ignore phenny
benabik More of the zort- stuff, or something new?
sbp benabik: there was spam from some chap who was passing on messages using the wildcard ("*") syntax in the tell feature. I made the wildcards more restrictive as a result. it affected at least one other channel too, annoyingly
just the zort- stuff
sorear @part #perl6
17:00 lambdabot left
benabik wildcards in tell? Eep. 17:00
tadzik phenny: tell bena* foo 17:01
phenny tadzik: I'll pass that on when bena* is around.
17:01 dakkar left
sbp * now matches 0123456789_-[]` — prepended and appended. circumfix! 17:02
17:02 FROGGS joined
benabik I guess that's useful for foo{,_,-work,-away,etc} 17:02
sbp yeah
benabik notes that bena* is around. :-/
sbp "bik" aren't in 0123456789_-[]` :-) 17:03
tadzik :)
benabik Ahhhh.
sbp anyway, enjoy and apologies again. I do plan to reside here again when I actually learn #perl6 and start writing some software in it—otherwise I'll just run my mouth off unnecessarily 17:04
sbp waves
17:04 sbp left 17:05 gongyiliao joined, marmay left 17:09 azawawi joined
azawawi hi #perl6 17:09
Just wanted to showcase feather.perl6.nl:4040/ # Frabi web-based Perl editor
Hopefully we can use it also for Perl 6 via Perlito 6 17:10
sirrobert azawawi: looking 17:11
azawawi brb 17:12
sirrobert nice =)
FROGGS is it possible to share code like a gist? 17:18
or better like a no-paste 17:19
17:20 SamuraiJack left, SamuraiJack joined
benabik FROGGS: Like with gist.github.com or nopaste.snit.ch/ ? 17:24
r: gist.github.com/1524863 17:25
p6eval rakudo cc1858: OUTPUT«test␤t e s t␤»
benabik FROGGS: And p6eval will even take gists and run them.
sorear benabik: I think the context here is generating gists *directly from azawawi 17:27
benabik: I think the context here is generating gists *directly from azawawi's p6 code editor*
17:27 nodmonkey joined
benabik sorear: Oh. Sorry. Totally spaced. 17:27
FROGGS: nvm then
17:29 MayDaniel joined
FROGGS benabik: I just wanted to know if I would be able to give someone a link, he/she clicks on it and get to your page, with my code sample present 17:29
like I can give somebody a no-paste link
benabik FROGGS: Yes. I somehow missed the bit about the editor. 17:30
FROGGS right, azawawi's page... but you are right, you can share gists and run that here too 17:31
GlitchMr Just wondering, does Perl 6 have sets like Python? 17:33
Python has sets which use: {1, 2, 3} syntax
FROGGS do you mean something like an enum? 17:35
GlitchMr No, I don't mean dictionary
I mean sets
>>> {1, 2, 3} & {1, 2, 4} 17:36
set([1, 2])
sorear n: say set(1,2,3) (&) set(1,2,4) 17:38
p6eval niecza v21-1-ga8aa70b: OUTPUT«set(1, 2)␤»
GlitchMr Nice 17:40
(&) looks ugly, but I guess it's good enough
& is junction
sirrobert n: say set(1,2,3) & set(1,2,4) 17:41
p6eval niecza v21-1-ga8aa70b: OUTPUT«all(set("1", "2", "3"), set("1", "2", "4"))␤»
sirrobert ah
sorear .u intersec t
phenny sorear: Sorry, no results for 'intersec t'.
sorear .u intersect
phenny U+2229 INTERSECTION (∩)
sorear n: say set(1,2,3) ∩ set(1,2,4)
p6eval niecza v21-1-ga8aa70b: OUTPUT«set(1, 2)␤»
sirrobert nice
17:44 benabik_ joined, benabik left, benabik_ is now known as benabik
GlitchMr .u negation 17:44
phenny U+2AEE DOES NOT DIVIDE WITH REVERSED NEGATION SLASH (⫮)
daxim .u ¬ 17:45
phenny U+00AC NOT SIGN (¬)
sirrobert what does the $bar here mean: token foo ($bar) { ... }
GlitchMr n: say set(1, 2, 3) ∖ set(1, 2, 4) 17:46
p6eval niecza v21-1-ga8aa70b: OUTPUT«===SORRY!===␤␤Confused at /tmp/dUhmqKuMWV line 1:␤------> say set(1, 2, 3) ⏏∖ set(1, 2, 4)␤␤Parse failed␤␤»
GlitchMr :(
If it supports ∩, why not ∖?
sirrobert what is \?
GlitchMr oh wait, that's ASCII
lol
.u ∖
phenny U+2216 SET MINUS (∖)
GlitchMr or not
This looks confusing in console 17:47
sirrobert n: say set(1, 2, 3) ∖ set(1, 2, 4)
GlitchMr But it's still set minus
p6eval niecza v21-1-ga8aa70b: OUTPUT«===SORRY!===␤␤Confused at /tmp/DDQc707DoY line 1:␤------> say set(1, 2, 3) ⏏∖ set(1, 2, 4)␤␤Parse failed␤␤»
sirrobert n: say set(1, 2, 3) \∖ set(1, 2, 4)
p6eval niecza v21-1-ga8aa70b: OUTPUT«===SORRY!===␤␤Confused at /tmp/vrhp3nhTNt line 1:␤------> say set(1, 2, 3) ⏏\∖ set(1, 2, 4)␤␤Parse failed␤␤»
sirrobert huh
sorear (-) is allowed 17:50
GlitchMr But is there any reason why ∖ isn't?
17:50 benabik left
sorear colomon didn't add an infix:<∖> definition to the setting, most direct reason 17:51
TimToady didn't know there was a character for that
sirrobert neither did I =)
sorear but if you ask me set(1,2,3) ∖ 2 looks too much like an unspace 17:52
TimToady well, in context it might not, if you're doing a lot of set stuff 17:54
[Coke] I would say too confusing for core, but easily addable if you're using sets all the time and speak that dialect.
GlitchMr So, we have ∪, but we don't have ∖ 17:55
This sort of makes sense
TimToady n: my &infix:<∖> := &infix:<(-)>; say set(1, 2, 3) ∖ set(1, 2, 4)
p6eval niecza v21-1-ga8aa70b: OUTPUT«set(3)␤»
daxim where's the documentation for set? 17:56
GlitchMr Also
There is × operator
TimToady S32-setting-library/Containers 17:57
GlitchMr set(1, 2) × set(3, 4) eqv set((1, 3), (1, 4), (2, 3), (2, 4))
sirrobert What is ~$ ?
in a grammar's action class
GlitchMr But I guess that Perl 6 has perfectly fine X operator
TimToady also bits of S02 and S03 17:58
but you'd have to make it back into a set
sirrobert or: what does "text => ~$<text>" mean?
TimToady stringify the node and make a pair of it 17:59
sirrobert ohh
arnsholt Makes a pair, .key is "text" and .value is whatever the <text> subrule matched
17:59 araujo left
TimToady stringified 17:59
sirrobert I was confused by ~$
makes sense now
[Coke] pmichaud: my latest nqp issue is errors like "Error while compiling block : Error while compiling op bind: Error while compiling block gets: Error while compiling op lexotic: Error while compiling op bind: Cannot infer type from ''
" - they don't tell me where in the code the problem is. 18:00
sirrobert Oh, $<text> is a shorthand for $/<text>
thanks
GlitchMr It saves one character, but it's always one character
TimToady less visual clutter, is the main thing
[Coke] (and the backtrace is basically "QAST errored out")
GlitchMr When you make lots of operation on grammar, this is helpful
I guess 18:01
18:01 Chillance left 18:02 Chillance joined, nodmonkey left
[Coke] AHA. new NQP is much finickier about pir opcode modifiers. pir::readline__sp dies horribly, but pir::readline__SP works. 18:02
FROGGS TimToady: an example using SDL: rosettacode.org/wiki/Mouse_position#Perl 18:05
masak does :$<text> expand to text => $<text> ? :)
18:08 araujo joined
sorear n: "foo" =~ /$<text> = [o*]/; :$<text>.perl.say 18:09
p6eval niecza v21-1-ga8aa70b: OUTPUT«===SORRY!===␤␤Unsupported use of =~ to do pattern matching; in Perl 6 please use ~~ at /tmp/Pt6EF_L9Yt line 1:␤------> "foo" =~⏏ /$<text> = [o*]/; :$<text>.perl.say␤␤Parse failed␤␤»
sorear n: "foo" ~~ /$<text> = [o*]/; :$<text>.perl.say
p6eval niecza v21-1-ga8aa70b: OUTPUT«"text" => #<match from(0) to(0) text() pos([].list) named({}.hash)>␤»
sorear masak: looks like so in niecza :D
TimToady but doesn't do the stringify 18:10
GlitchMr perl6: grammar cake { token TOP { <text> }; token text { a } }; cake.parse('a'); (:$<cake>).perl.say
p6eval niecza v21-1-ga8aa70b: OUTPUT«"cake" => Any␤»
..rakudo cc1858: OUTPUT«===SORRY!===␤Confused␤at /tmp/5nS1Joi8d0:1␤»
GlitchMr Doesn't look like it
It's more like cake => Any 18:11
perl6: grammar cake { token TOP { <text> }; token text { a } }; cake.parse('a'); (:$<cake>)['cake'].say
p6eval rakudo cc1858: OUTPUT«===SORRY!===␤Confused␤at /tmp/v9jTHTrRsU:1␤»
..niecza v21-1-ga8aa70b: OUTPUT«Unhandled exception: Cannot parse number: cake␤ at /home/p6eval/niecza/lib/CORE.setting line 1435 (die @ 5) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3525 (ANON @ 11) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3527 (NumSyntax.str2num @ 5) ␤ at …
GlitchMr perl6: grammar cake { token TOP { <text> }; token text { a } }; cake.parse('a'); (:$<cake>)<cake>.say
p6eval rakudo cc1858: OUTPUT«===SORRY!===␤Confused␤at /tmp/UI2ISXyM5g:1␤»
..niecza v21-1-ga8aa70b: OUTPUT«Unhandled exception: Cannot use hash access on an object of type Pair␤ at <unknown> line 0 (ExitRunloop @ 0) ␤ at /home/p6eval/niecza/lib/CORE.setting line 352 (Any.at_key @ 9) ␤ at <unknown> line 0 (ExitRunloop @ 0) ␤ at /tmp/d7CbhxAn7V line 1 (mainlin…
GlitchMr ok...
sorear n: grammar cake { token TOP { <text> }; token text { a } ; cake.parse('a'); (:$<cake>).value.say 18:12
p6eval niecza v21-1-ga8aa70b: OUTPUT«===SORRY!===␤␤Unable to parse block at /tmp/wS2kGc4nCT line 1:␤------> grammar cake {⏏ token TOP { <text> }; token text { a } ␤Couldn't find final '}'; gave up at /tmp/wS2kGc4nCT line 1 (EOF):␤------> ; cake.parse(…
sorear n: grammar cake { token TOP { <text> }; token text { a } }; cake.parse('a'); (:$<cake>).value.say
p6eval niecza v21-1-ga8aa70b: OUTPUT«Any()␤»
sorear n: grammar cake { token TOP { <text> }; token text { a } }; cake.parse('a'); say $/ 18:13
p6eval niecza v21-1-ga8aa70b: OUTPUT«#<match from(0) to(1) text(a) pos([].list) named({"text" => #<match from(0) to(1) text(a) pos([].list) named({}.hash)>}.hash)>␤»
sorear n: grammar cake { token TOP { <text> }; token text { a } }; cake.parse('a'); (:$<text>).value.say
p6eval niecza v21-1-ga8aa70b: OUTPUT«#<match from(0) to(1) text(a) pos([].list) named({}.hash)>␤»
sorear there's no "cake" capture anywhere.
GlitchMr lolright 18:14
18:19 jevin left 18:23 mucker left, jevin joined 18:29 jevin left 18:32 jevin joined 18:47 FROGGS__ left 18:49 FROGGS left 18:53 FROGGS__ joined
azawawi FROGGS__: the idea is to build a fully featured editor 18:55
FROGGS__: perl mojo backend 18:56
FROGGS__: js frontend
sorear good luck making a syntax-aware editor for perl 6 18:57
sirrobert azawawi: nice =)
azawawi sorear: hi... long time no see :)
18:58 daxim left
sorear azawawi: hello. i've been here the whole time, I think... 18:58
even when I was 1.5e7 m from home, I was here. 18:59
azawawi sorear: i know. Sadly i havent been here lately :(
19:05 FROGGS joined 19:08 azawawi left 19:16 am0c joined 19:17 benabik joined 19:29 benabik left 19:30 benabik joined
lizmat rn: infix:<+>(1, 2) 19:39
p6eval rakudo cc1858, niecza v21-1-ga8aa70b: ( no output )
19:39 benabik left
lizmat infix:<+>(1, 2).say 19:39
colomon rn: infix:<+>(1, 2).say
lizmat rn: infix:<+>(1, 2).say
p6eval rakudo cc1858, niecza v21-1-ga8aa70b: OUTPUT«3␤»
colomon ;)
lizmat thanks :-) 19:40
sorear o/ colomon, lizmat
colomon \o
tadzik oh hello
lizmat hi sorear
hi tadzik
sorear o/ tadzik
lizmat was just trying some example code
rn: circumfix:«( )»('a', 'b', 'c').say 19:41
p6eval rakudo cc1858: OUTPUT«===SORRY!===␤CHECK FAILED:␤Undefined routine '&circumfix:<( )>' called (line 1)␤»
..niecza v21-1-ga8aa70b: OUTPUT«===SORRY!===␤␤Undeclared routine:␤ 'circumfix:«( )»' used at line 1␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 1435 (die @ 5) ␤ at /home/p6eval/niecza/src/STD.pm6 line 1147 (P6.comp_unit @ 37) …
lizmat rn: circumfix:«( )»('a', 'b', 'c')
p6eval rakudo cc1858: OUTPUT«===SORRY!===␤CHECK FAILED:␤Undefined routine '&circumfix:<( )>' called (line 1)␤»
..niecza v21-1-ga8aa70b: OUTPUT«===SORRY!===␤␤Undeclared routine:␤ 'circumfix:«( )»' used at line 1␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 1435 (die @ 5) ␤ at /home/p6eval/niecza/src/STD.pm6 line 1147 (P6.comp_unit @ 37) …
lizmat rn: circumfix:<( )>('a', 'b', 'c') 19:42
p6eval niecza v21-1-ga8aa70b: OUTPUT«===SORRY!===␤␤Undeclared routine:␤ 'circumfix:<( )>' used at line 1␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 1435 (die @ 5) ␤ at /home/p6eval/niecza/src/STD.pm6 line 1147 (P6.comp_unit @ 37) …
..rakudo cc1858: OUTPUT«===SORRY!===␤CHECK FAILED:␤Undefined routine '&circumfix:<( )>' called (line 1)␤»
19:44 erkan left, snearch joined 19:45 erkan joined, erkan left, erkan joined 19:47 amkrankruleuen joined
lizmat rn: circumfix:« ( ) »('a', 'b', 'c') 19:47
p6eval niecza v21-1-ga8aa70b: OUTPUT«===SORRY!===␤␤Undeclared routine:␤ 'circumfix:« ( ) »' used at line 1␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 1435 (die @ 5) ␤ at /home/p6eval/niecza/src/STD.pm6 line 1147 (P6.comp_unit @ 37…
..rakudo cc1858: OUTPUT«===SORRY!===␤CHECK FAILED:␤Undefined routine '&circumfix:<( )>' called (line 1)␤»
lizmat rn: circumfix:«( )»('a', 'b', 'c')
p6eval rakudo cc1858: OUTPUT«===SORRY!===␤CHECK FAILED:␤Undefined routine '&circumfix:<( )>' called (line 1)␤»
..niecza v21-1-ga8aa70b: OUTPUT«===SORRY!===␤␤Undeclared routine:␤ 'circumfix:«( )»' used at line 1␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 1435 (die @ 5) ␤ at /home/p6eval/niecza/src/STD.pm6 line 1147 (P6.comp_unit @ 37) …
19:49 immortal joined, immortal left, immortal joined 19:50 erkan left
GlitchMr perl6: 26, 32, 35, 48, 12, 11 ==> grep * %% 2 ==> sort ==> my @array; print @array.join(' ') 19:50
p6eval niecza v21-1-ga8aa70b: OUTPUT«===SORRY!===␤␤Preceding context expects a term, but found infix ==> instead at /tmp/5_XEOQMICF line 1:␤------> 35, 48, 12, 11 ==> grep * %% 2 ==> sort ⏏==> my @array; print @array.join(' ')␤␤Parse failed␤␤»…
..rakudo cc1858: OUTPUT«===SORRY!===␤Preceding context expects a term, but found infix ==> instead␤at /tmp/_mv0HVP0hC:1␤»
GlitchMr What's wrong with sort ==>?
sorear you need sort() in current rakudo 19:51
masak and current niecza, it seems.
sorear after sort, the parser is looking for a term, and rakudo doesn't yet handle ==> in term context
masak: niecza doesn't support ==> at all
tadzik probably also in current pugs :)
diakopter phenny: ask TimToady have you seen rosettacode.org/wiki/Category:Draft...ming_Tasks
sorear pugs: 1 ==> sort
p6eval pugs: OUTPUT«*** ␤ Unexpected ">"␤ at /tmp/BL2C6vTadE line 1, column 5␤»
phenny diakopter: I'll pass that on when TimToady is around.
lizmat can someone explain me why "circumfix:«( )»('a', 'b', 'c')" complains about circumfix not being defined? 19:52
it is an example from doc.perl6.org
masak lizmat: looks kosher to me.
diakopter rn: circumfix:«( )»('a', 'b', 'c')
p6eval rakudo cc1858: OUTPUT«===SORRY!===␤CHECK FAILED:␤Undefined routine '&circumfix:<( )>' called (line 1)␤» 19:53
..niecza v21-1-ga8aa70b: OUTPUT«===SORRY!===␤␤Undeclared routine:␤ 'circumfix:«( )»' used at line 1␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 1435 (die @ 5) ␤ at /home/p6eval/niecza/src/STD.pm6 line 1147 (P6.comp_unit @ 37) …
tadzik hrm
rn: sub circumfix:«( )»('a', 'b', 'c')
p6eval rakudo cc1858: OUTPUT«===SORRY!===␤Missing block␤at /tmp/fz1oyu9jrs:1␤»
..niecza v21-1-ga8aa70b: OUTPUT«===SORRY!===␤␤Malformed block at /tmp/2Ft5ZTGWBA line 1 (EOF):␤------> sub circumfix:«( )»('a', 'b', 'c')⏏<EOL>␤␤Parse failed␤␤»
tadzik rn: sub circumfix:«( )»('a', 'b', 'c') {}
p6eval niecza v21-1-ga8aa70b: OUTPUT«Potential difficulties:␤ &circumfix:<( )> is declared but not used at /tmp/749M5vAuEd line 1:␤------> sub circumfix:«( )»⏏('a', 'b', 'c') {}␤␤»
..rakudo cc1858: ( no output )
tadzik rakudo's fine with this one
rn: sub circumfix:«( )»('a', 'b', 'c') { say (a) }; (5)
p6eval rakudo cc1858: OUTPUT«===SORRY!===␤CHECK FAILED:␤Undefined routine '&a' called (line 1)␤»
..niecza v21-1-ga8aa70b: OUTPUT«===SORRY!===␤␤Undeclared routine:␤ 'a' used at line 1␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 1435 (die @ 5) ␤ at /home/p6eval/niecza/src/STD.pm6 line 1147 (P6.comp_unit @ 37) ␤ at /home/p6…
diakopter a 19:54
tadzik r: sub circumfix:«( )»('a', 'b', 'c') { say "asd!" }; (5)
p6eval rakudo cc1858: ( no output )
tadzik the rest is left as an exercise for the reader :P
sorear lizmat: in niecza, circumfix:<( )> is treated specially by the parser and does not exist as a sub/macro
19:57 fibo joined
lizmat r: infix:<+> { say "foo" }; 1 + 2 19:57
p6eval rakudo cc1858: OUTPUT«Cannot call 'Numeric'; none of these signatures match:␤:(Mu:U \v, Mu *%_)␤␤ in method Numeric at src/gen/CORE.setting:706␤ in sub infix:<+> at src/gen/CORE.setting:2505␤ in block at /tmp/Oh4DDFjU3u:1␤␤»
lizmat n: infix:<+> { say "foo" }; 1 + 2
p6eval niecza v21-1-ga8aa70b: OUTPUT«Unhandled exception: Cannot call infix:<+>; none of these signatures match:␤ Any, Any␤ ␤ at /tmp/83PTpRtofc line 1 (mainline @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4138 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4139 (mo…
moritz if you don't tive it a signature, it doesn't take any arguments
nr: sub infix:<+>($a, $b) { say 'foo' }; 1 + 2 19:58
lizmat rn: infix:<+> (1,2) { say "foo" }; 1 + 2
p6eval rakudo cc1858: OUTPUT«foo␤»
..niecza v21-1-ga8aa70b: OUTPUT«Potential difficulties:␤ $a is declared but not used at /tmp/FjhCPipb7i line 1:␤------> sub infix:<+>(⏏$a, $b) { say 'foo' }; 1 + 2␤ $b is declared but not used at /tmp/FjhCPipb7i line 1:␤------> sub infix:<+>($a, ⏏$b) { …
rakudo cc1858: OUTPUT«===SORRY!===␤Confused␤at /tmp/YVJjCUBuVt:1␤»
..niecza v21-1-ga8aa70b: OUTPUT«===SORRY!===␤␤Unexpected block in infix position (two terms in a row, or previous statement missing semicolon?) at /tmp/qmrYbhlFV4 line 1:␤------> infix:<+> (1,2) ⏏{ say "foo" }; 1 + 2␤␤Parse failed␤␤»
sorear more to the point, "sub" is missing
moritz nr: sub infix:<+>($, $) { say 'foo' }; 1 + 2
lizmat aha
p6eval rakudo cc1858, niecza v21-1-ga8aa70b: OUTPUT«foo␤»
lizmat I'm trying to make sense of the examples on doc.perl6.org/language/operators
they are not as enlightening as they should, I think 19:59
20:00 amkrankruleuen left
moritz lizmat: you have to understand that they are calls to operators, not definitions 20:01
lizmat well, I was just being stupid and copy/pasting what looks like examples :-)
sorear which examples?
moritz r: say infix:<+>(1, 2) 20:02
p6eval rakudo cc1858: OUTPUT«3␤»
moritz lizmat: anyway, if you have ideas on how to improve them, please tell
(or patch; I'm happy to give anybody commit access)
lizmat I will push some patches later ok? 20:03
moritz sure
lizmat I already have a commit bit :-)
moritz great
lizmat did feather just die ? 20:04
moritz I know I suck at writing examples
no
20:04 GlitchMr left
lizmat indeed, it just hung for ~ minute for me 20:04
sorear feather does that a lot 20:05
although Juerd says it's actually the router messing up and feather itself is unaffected during the downtime
lizmat so anybody against upgrading feather to some more RAM and some more CPU ?
sorear iirc
20:06 am0c left
Juerd sorear: That was ages ago. 20:06
moritz fwiw feather1 is a horrible mixture of debian unstable and ... unspeakable things 20:07
tadzik sergot: mind updating the modules page? Cronjobs seem not to be running
:)
Juerd sorear: The router has no such issues anymore. If feather still hangs, it's something else.
lizmat I'll take that silence as a "no" then
20:07 sivoais left
tadzik some say, that silence means agreement :) 20:07
Juerd lizmat: Why would anyone be against upgrading?
lizmat indeed
:-)
moritz the proper thing to do would be to set up a new machine, migrate all the services and users, and kill the old one
but currently I don't have time for such fun
sorear Juerd: The people who are against upgrading are the people who 1. would have to pay for the new hardware 2. would have to spend time installing/migrating/fixing breakage 20:08
Juerd moritz: Why the "proper" thing? Sure, feather has accumulated a lot of cruft and dust over the years, but it's mostly manageable because the changes were made in /usr/local for the largest part.
lizmat point 1 is taken care of, I think 20:09
moritz Juerd: running a debian/unstable as a publicly available server on the internet should only be done if security updates are regularly installed 20:10
Juerd: much more regular than what I do now
Juerd: so it would be better to run something less volatile, but debian doesn't support downgrades
tadzik debian testing is alright, I think 20:11
moritz yes, I would be fine with that
but feather is mixed sid/experimental
Juerd moritz: I strongly disagree for feather1 20:13
moritz: For things that are important and that shouldn't crash or be broken into, there's feather2
moritz: Remote exploits in packaged software are incredibly rare, and local users can['t]? be trusted anyway, regardless of the availability of local exploits. 20:14
My password for feather1 has been "insecure", without the quotes, for over a year. That says a lot about how I feel about feather1.
On a machine that is used for development of bleeding edge stuff, you need a bleeding edge platform, IMO. 20:15
PerlJam logs in as Juerd and prepares to wreak havoc
Juerd Even if it does actually bleed from time to time ;D
moritz Juerd: I'm fine with that. If I felt more responsible for it, I'd do things differently though 20:16
Juerd Consider that it has been up and running without major security incidents since 2005. 20:17
diakopter security by obscurity
Juerd Upgrades have caused real trouble about three times. That would have happened with stable Debian too.
moritz diakopter: security by being unimportant enough 20:18
Juerd diakopter: I'd say feather1 has no real security at all.
moritz diakopter: same as p6eval
tadzik in stable debian updates will happen about 3 times :P
Juerd diakopter: If something bad happens, we find who did it, bash the bastard, and restore a backup :)
diakopter :P 'twas a joke guys
Juerd This has happened once, and it wasn't security related. It was an honest mistake.
tadzik in our small community we could as well just raid somebody's house if one did a security breakage :P 20:19
Juerd The other abuse of feather was someone hosting a movie in their ~public_html
That caused a looooot of bandwidth usage.
diakopter ick
20:20 Targen joined
lizmat assumes Juerd watches bandwidth usage ? 20:21
PerlJam he does *now* :) 20:22
sorear Juerd: What was the first thing? 20:23
20:24 MikeFair_ left
Juerd lizmat: Nah, our ISP does that, and warns us :) 20:24
lizmat cool
Juerd lizmat: We do have bandwidth counters but we don't actively monitor that.
Most of our contracts are "fair use" based anyway. 20:25
sorear: The first feather abuse recorded was that pirated movie in someone's ~/public_html
sorear I have to admit I've used feather as a largish file host before (0.1 - 10MB files and not for anything I expect to be popular)
Juerd sorear: And the second one was -probably- a password that had leaked; someone used the same password everywhere. 20:26
sorear: At least that's not a 800 MB file that's wildly popular.
sorear a password that had leaked? you mean like how juerd@feather's password is now on perlgeek.de? 20:28
Juerd No
My password has been "insecure", but I changed it to something else a few years ago.
20:29 MikeFair_ joined
masak is generally paranoid enough not to leak old passwords either :) 20:29
Juerd And the other password had probably leaked from somewhere else than feather itself, or IRC. Possibly a key logger, or a website that had its user database opened by accident :)
masak: I'm quite sure that I haven't used this particular password anywhere else :P 20:30
sorear "Anyone caught spamming, hosting pirated bits or kiddypr0n, etcetera, " # now I wonder if the other two ever happened.
Juerd sorear: Not that I know of.
The etcetera hasn't happened either, by the way, even though the pirated bits were clearly there. 20:31
afk 20:34
20:35 MayDaniel left 20:44 immortal left 20:45 erkan joined, erkan left, erkan joined 20:46 sivoais joined
sirrobert in a grammar's actions, does TOP always get called last? 20:46
I'm guessing that's the only time it can be known whether the whole matches... am I missing something in my reasoning? 20:47
arnsholt As long as you don't recurse back into TOP, that should hold I think
sirrobert ahh, good point. ok, I'm pretty safe on that, I think 20:48
is there a way to auto-create an action for each grammar element? 20:49
like ... call a generic sub/method with the symbol name
or something
maybe a "call this method if there's no explicitly matching action" 20:50
sorear you would do that using the fallbacks mechanism. 20:51
sirrobert What's the fallbacks mechanism? where can I read about it? 20:52
(esp. syntax)
sergot tadzik: There's a little bug in mkjson.p6. I was working on it today, ll be repaired this evening. :)
tadzik okay, awesome :)
I'll have fresh smoketest results soonish
TimToady diakopter: I usually get at that view rosettacode.org/wiki/Reports:Tasks_..._in_Perl_6
phenny TimToady: 19:51Z <diakopter> ask TimToady have you seen rosettacode.org/wiki/Category:Draft...ming_Tasks
sorear n: grammar G { token TOP { x } }; class A { method FALLBACK($name,$/) { say $name } }; G.parse('x', :actions(A)) # here's how it works in niecza
p6eval niecza v21-1-ga8aa70b: OUTPUT«TOP␤»
sirrobert sorear: thanks
sorear rakudo has an incompatible (IMO better) system using the .^add_fallback meta method 20:53
but I don't know offhand how to use it
sirrobert sorear: ok, thanks
diakopter ha; no one's attempted rosettacode.org/wiki/Unicode_polynomial_equation
TimToady read the talk page for why
lizmat will continue some more spelunking in the docs tomorrow 20:55
TimToady meeting & 20:59
21:03 skids left 21:08 lizmat left 21:11 kaare__ left 21:16 Targen left 21:17 atrodo left, FROGGS left, FROGGS__ left 21:18 Targen joined
masak jnthn++ # figured out what's wrong with my macros-d2 code, and hinted at what to do instead 21:20
has this one ever gone on record as an autopun? "A Freudian slip is meaning one thing but saying your mother." 21:22
good night, #perl6 21:25
diakopter o/
sorear oo/ 21:30
masak: it hasn't, but it absolutely is one.
\oo/ # two headed sorear
21:33 SamuraiJack left, Circlepuller joined 21:34 Circlepuller_ left
PerlJam sorear: two heads is great, but don't you also need more hands to type code? 21:41
diakopter Abby & Brittney type just fine sharing two hands
*Brittany 21:42
21:42 cognominal left 21:44 cognominal joined
sergot tadzik: done. :) 22:00
New feature added to modules.perl6.org ! 22:01
feather.perl6.nl/~sergot/modules/
feather.perl6.nl/~sergot/modules/mo...ashes.html
\o/ 22:02
TODO: parsing markdown.
22:05 whiteknight joined, snearch left
[Coke] nqp: say(Q:PIR{ %r = 3}); 22:06
p6eval nqp: OUTPUT«Null PMC access in set_integer_native()␤current instr.: '' pc 49 ((file unknown):162810055) (/tmp/Sb0knkHtxE:10)␤»
[Coke] nqp: say(Q:PIR{ %r = box 3});
p6eval nqp: OUTPUT«3␤»
22:08 sHACHAF is now known as Shachaf
[Coke] nqp: nqp::existkey(); say("alive"); 22:09
p6eval nqp: OUTPUT«Error while compiling block : Error while compiling op existkey: No registered operation handler for 'existkey'␤current instr.: '' pc 42779 (src/stage2/QAST.pir:14984) (src/stage2/QAST.nqp:2244)␤»
[Coke] phenny: tell pmichaud that "existkey" is metioned in the docs but doesn't exist. 22:10
phenny [Coke]: I'll pass that on when pmichaud is around.
22:14 cognominal_ joined 22:17 cognominal left, fgomez joined
[Coke] nqp: pir::print__vPS($*OUT,"3") 22:24
p6eval nqp: OUTPUT«get_pointer_keyed_int() not implemented in class 'NQPMu'␤current instr.: '' pc 92 ((file unknown):169239368) (/tmp/FM3XAjhBXd:1)␤»
[Coke] nqp: pir::print__VPS($*OUT,"3")
p6eval nqp: OUTPUT«get_pointer_keyed_int() not implemented in class 'NQPMu'␤current instr.: '' pc 92 ((file unknown):158208840) (/tmp/wkDLfjpdjE:1)␤»
[Coke] HALP. :)
nqp: print("HI"); 22:27
p6eval nqp: OUTPUT«HI»
[Coke] "close enough".
nqp: $*OUT.print("HI")
p6eval nqp: OUTPUT«Method 'print' not found for invocant of class 'NQPMu'␤current instr.: '' pc 96 ((file unknown):149734207) (/tmp/UCa1O__yMH:1)␤»
sorear I think there is no $*OUT in NQP
nqp: print $*OUT 22:28
p6eval nqp: OUTPUT«Confused at line 2, near "print $*OU"␤current instr.: 'panic' pc 19998 (src/stage2/gen/NQPHLL.pir:7314) (src/stage2/gen/NQPHLL.pm:325)␤»
sorear nqp: print($*OUT)
p6eval nqp: ( no output )
sorear nqp: print($*OUT.WHAT)
p6eval nqp: ( no output )
sorear nqp: print($*OUT === NQPMu)
p6eval nqp: OUTPUT«Confused at line 2, near "print($*OU"␤current instr.: 'panic' pc 19998 (src/stage2/gen/NQPHLL.pir:7314) (src/stage2/gen/NQPHLL.pm:325)␤»
sorear nqp: print($*OUT =:= NQPMu)
p6eval nqp: OUTPUT«1»
sorear nqp: print($*OUT =:= NQPInt)
p6eval nqp: OUTPUT«1»
sorear nqp: print($*OUT =:= moof)
p6eval nqp: OUTPUT«1»
sorear nqp: print($*OUT =:= 3)
p6eval nqp: OUTPUT«0»
sorear $*OUT compares equal to any undefined name, hmm. 22:29
(I thought that NQP would throw on undefined names nowadays ??)
22:30 Shachaf is now known as shachaf
[Coke] I'd like a way to call pir's print opcode from nqp. 22:30
22:33 stopbit left
sorear nqp: pir::print__vS("foo") 22:35
p6eval nqp: OUTPUT«foo»
sorear nqp: pir::print__vPS(pir::getstdout__P(), "foo")
p6eval nqp: OUTPUT«foo»
22:35 skids joined
sorear like so? 22:35
22:38 bluescreen10 left 22:39 stepnem left
[Coke] wtf. 22:40
sorear ?
nqp: my $err := pir::getstderr__P(); $err.print("Foo\n") 22:42
p6eval nqp: OUTPUT«Foo␤»
sorear nqp: pir::getinterp__P().stdout_handle().print("foo\n") 22:43
p6eval nqp: OUTPUT«foo␤»
[Coke] nqp: gist.github.com/3646684
p6eval nqp: ( no output )
[Coke] so if you uncomment out the last pir::print there, horrible failure.
nqp: gist.github.com/3646696 22:44
p6eval nqp: OUTPUT«Error while compiling block : Error while compiling op bind: Error while compiling block puts: Error while compiling op if: Cannot infer type from ''␤current instr.: '' pc 42779 (src/stage2/QAST.pir:14984) (src/stage2/QAST.nqp:2244)␤»
sorear nqp: 1 if 1;
p6eval nqp: ( no output )
sorear nqp: pir::print__vPS(pir::getstdout__P(), "foo\n");
p6eval nqp: OUTPUT«foo␤»
sorear nqp: pir::print__vPS(pir::getstdout__P(), "foo\n") if 1;
p6eval nqp: OUTPUT«Error while compiling block : Error while compiling op if: Cannot infer type from ''␤current instr.: '' pc 42779 (src/stage2/QAST.pir:14984) (src/stage2/QAST.nqp:2244)␤»
sorear ah 22:45
the problem is the if.
22:45 fibo left
sorear because it means $nl ?? pir::stuff !! NQPMu 22:45
[Coke] that looks like reasonable code, though, yes?
sorear the left side is void, though
diakopter r: 1 if 1 if 1 { }
p6eval rakudo cc1858: ( no output )
sorear the right side returns a PMC
it cannot infer the type. 22:46
diakopter sorear: ^^ what does that mean, do you know?
sorear nqp: pir::print__vPS(pir::getstdout__P(), "foo\n"), NQPMu if 1;
p6eval nqp: OUTPUT«Error while compiling block : Error while compiling op if: Error while compiling op list: Cannot infer type from ''␤current instr.: '' pc 42779 (src/stage2/QAST.pir:14984) (src/stage2/QAST.nqp:2244)␤»
[Coke] you get the same error if you do if 1 { ... }
sorear what if you add an else?
nqp: pir::print__vPS(pir::getstdout__P(), "foo\n"), NQPMu;
p6eval nqp: OUTPUT«Error while compiling block : Error while compiling op list: Cannot infer type from ''␤current instr.: '' pc 42779 (src/stage2/QAST.pir:14984) (src/stage2/QAST.nqp:2244)␤»
diakopter r: 1 if 1 if 1 { say 4 }
p6eval rakudo cc1858: OUTPUT«4␤»
sorear nqp: (pir::print__vPS(pir::getstdout__P(), "foo\n"); 1) 22:47
p6eval nqp: OUTPUT«Confused at line 2, near "(pir::prin"␤current instr.: 'panic' pc 19998 (src/stage2/gen/NQPHLL.pir:7314) (src/stage2/gen/NQPHLL.pm:325)␤»
sorear nqp: do { pir::print__vPS(pir::getstdout__P(), "foo\n"); 1 }
p6eval nqp: OUTPUT«Confused at line 2, near "do { pir::"␤current instr.: 'panic' pc 19998 (src/stage2/gen/NQPHLL.pir:7314) (src/stage2/gen/NQPHLL.pm:325)␤»
sorear nqp: if 1 { pir::print__vPS(pir::getstdout__P(), "foo\n"); 1 }
p6eval nqp: OUTPUT«foo␤»
sorear nqp: if 1 { pir::print__vPS(pir::getstdout__P(), "foo\n") }
p6eval nqp: OUTPUT«Error while compiling block : Error while compiling op if: Cannot infer type from ''␤current instr.: '' pc 42779 (src/stage2/QAST.pir:14984) (src/stage2/QAST.nqp:2244)␤»
sorear [Coke]: workaround: add block with a ; 1
std: 1 if 1 if 1 { say 4 }
p6eval std b87ea13: OUTPUT«===SORRY!===␤Missing semicolon at /tmp/GKaFRsgI0G line 1:␤------> 1 if 1⏏ if 1 { say 4 }␤Parse failed␤FAILED 00:00 43m␤»
[Coke] lame but effective. danke.
sorear diakopter: It means that rakudo's parser is too lax :D
diakopter yeah; I'm just wondering how it's parsing it 22:48
22:48 Targen left
sorear r: say (1 if 1 if 1 { } ) 22:49
p6eval rakudo cc1858: OUTPUT«Nil␤»
sorear r: say ((1 if 1) if 1 { } )
p6eval rakudo cc1858: OUTPUT«===SORRY!===␤Unable to parse postcircumfix:sym<( )>, couldn't find final ')' at line 2, near "if 1 { } )"␤»
sorear r: say (1 if (1 if 1 { }))
p6eval rakudo cc1858: OUTPUT«===SORRY!===␤Unable to parse postcircumfix:sym<( )>, couldn't find final ')' at line 2, near "if 1 { }))"␤»
tadzik sergot: woow, that's super awesome
[Coke] wonders, if I have nqp-rx code that calls PAST::Compiler.compile, if it's sufficient to s/P/Q/ 22:53
tadzik feather.perl6.nl/~sergot/modules/mo...tines.html \o/ 22:55
ossum
diakopter wonders when VMs will be relatively as lightweight as threads 22:56
22:57 benabik joined
diakopter hardware/OS VMs, I mean 22:57
22:58 xinming left
tadzik when computers become so fast that it won't matter :) 22:58
good knight #perl6
22:59 xinming joined
diakopter sorear: I've tried tons of parentheses pairings without finding on that works 23:00
diakopter gives up and hopes masakbot will submit a rakudobug if it's not already
sorear diakopter: i suppose you are already familiar with VM/CMS? 23:02
geekosaur heh 23:03
leont Don't say such scary things in the channel!
TimToady then there's MVS and VMS... 23:05
tadzik and CVS! \o/
geekosaur bah
tadzik yeah, I should be sleeping
geekosaur VM/SP was doing al this virtualization stuff years ago and way better than the Intel-based stuff ever did. and still does, Ibelieve the same technology is still in the z/OS virtualization products 23:06
23:08 PacoAir left
Tene diakopter: you can get pretty close with linux namespaces. 23:10
23:10 Celelibi left 23:12 Celelibi joined
sorear [Coke]: I'm curious what partcl-nqp's current state of working is 23:12
[Coke] nqp2 branch. the grammar still only recognizes "three". I'm patching up a bunch of code that compiled in nqp-rx but doesn't in nqp. but I'm not actually calling any of it yet. 23:16
23:17 ashleyde1 is now known as ashleydev 23:20 leont left 23:26 Targen joined
felher 'night, #perl6 23:33
colomon o/ 23:39
23:41 thundergnat joined
thundergnat Good * #perl6. 23:46
colomon \o 23:47
23:48 erkan left
thundergnat I recently updated one of my perl6 modules that has been failing for a while so that it no longer ( theoretically ) does. 23:56
github.com/thundergnat/Sort-Naturally
But panda still won't install it, even though it seems to work ok for me when installed manually.
panda fails during the pir generation step (e.g. perl6 --target=pir --output=foo.pir foo.pm6 ) 23:57
===SORRY!=== Serialization Error: unknown static lexical info type for 'infix:sym<ncmp>'
I know I should probably ask this when tadzik / jnthn is around, but my schedule doesn't don't overlap theirs very well.

Powered by Cro and the Raku® Programming Language.

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

Thank you!