Summer of Perl | 6.2.12 released! | paste: sial.org/pbot/perl6 | pugs.blogs.com | pugscode.org | pugs.kwiki.org | www.treehugger.com/files/th_images/paradigm.jpg
Set by nothingmuch on 12 July 2006.
00:31 nekokak joined
robkinyon nothingmuch: pong (kinda) 00:40
00:49 hikozaemon joined 01:03 tup joined 01:14 weinig|away is now known as weinig 01:29 tup joined 01:46 xern joined 01:47 xern joined 01:48 weinig joined 01:53 cmarcelo left 01:56 bsb joined 02:06 Yappobot is now known as Yappo 02:09 weinig is now known as weinig|sleep 02:11 dolmans joined 02:12 weinig|sleep is now known as weinig 02:20 bsb_ joined 02:25 aufrank joined 02:31 weinig is now known as weinig|sleep
aufrank anyone around? 02:40
audreyt not me 02:41
<- at pdx airport
but I guess I'll be around for an hour or so :)
aufrank do you know if TT or Mr. Bunce put their slides online? 02:42
(hi!)
audreyt TT's slides will appear on perl.com in a few weeks, usually
TimToady probably has a copy of script somewhere in his hard driv, though, so maybe you can ask him to nopaste it :) 02:43
not sure about tim bunce; mail him?
(hi!)
thedamian usually does not post his slides online.
aufrank ok, probably will do that tomorrow
audreyt cool 02:44
aufrank did you read tim bray's entry about perl6 and parrot?
www.tbray.org/ongoing/When/200x/200...-Languages
lambdabot Title: ongoing &#xb7; OSCON&#x2014;Perl &#x26; Python
audreyt no. reading
aufrank I think we've already got the unicode string handling he's so excited about in python 3, eh? 02:45
it's specced that way, at least
audreyt right. 02:46
also docs/notes/unicode_draft.
but not coded that way yet, no. though perl5 can use some retroporting
to get the same sort of sane behaviour, that is 02:47
aufrank so I've been kicking around an idea, and I can't tell if it would be useful or not
audreyt pray tell
aufrank I was reading the atom spec that is written in the compact relaxNG syntax
and the compact syntax is quite, um, compact 02:48
it would fit into a pretty easy grammar, I think
so I started thinking that if you had a grammar that parsed arbitrary compact relaxng schemas, you could use the match object to build perl6 classes out of the schemas 02:49
and then use those classes to build parsers for instances of the schema
which would give you well-formed objects 02:50
which could then be serialized any way you wanted
audreyt have you read emeijer's X/O paper? :)
aufrank similarly, if you parsed schemas into roles, you could then emit well-formed documents for anything that did the role 02:51
no, I did not
audreyt ok. it's a step by step recipe of how to do that :)
aufrank is it linked on your site right now?
02:51 AppleIver joined
audreyt homepages.cwi.nl/~ralf/xo-impedance-mismatch/ 02:51
yes
lambdabot Title: Revealing the X/O impedance mismatch
aufrank excellent
I'll grab a copy
audreyt but yes, having that technology would be great
actually it could be built on Moose 02:52
using PCR and friends
AppleIver hi all
aufrank yeah, I think we've got what we need to do it
audreyt AppleIver: greetings.
aufrank hi AppleIver
audreyt aufrank: yup. just need to implement the encoding
I mean embedding
AppleIver what you talk about?
02:53 robkinyon joined
audreyt AppleIver: making classes out from XML schema 02:53
aufrank audreyt: first step would be a p6 grammar of www.oasis-open.org/committees/relax...21121.html ?
lambdabot Title: RELAX NG Compact Syntax
AppleIver it's great!
audreyt aufrank: EBNF2rules? 02:54
AppleIver audreyt where are u from? 02:55
audreyt aufrank: take search.cpan.org/dist/Parse-EBNF/ and emit PCR? 02:56
lambdabot Title: Cal Henderson / Parse-EBNF - search.cpan.org
audreyt aufrank: I think that's perhaps easiest path
AppleIver: I'm from audreyt.org and pugs.blogs.com
aufrank audreyt: I wasn't aware of that module-- good place to start for sure
audreyt enjoys this "CPAN: The Gathering" game :)
aufrank audreyt: also satisfyingly general 02:57
audreyt: I keep meaning to ask about your history with M:tG
AppleIver c++?
aufrank did/do you play much? 02:58
nothingmuch aufrank: moose
can you post the summary?
aufrank nothingmuch: sure. of what ;)
nothingmuch oop 02:59
audreyt: ping =)
audreyt aufrank: I quit a few years back. I was playing quite competetively -- top 8 in asia pacific championships, a couple showing at worlds -- but that's all before I discovered the far more satisfying cpan: the gathering
nothingmuch competitive sports is not much fun =/
audreyt especially with a bunch of barely showered teenage males! 03:00
clkao (!)
nothingmuch oh no! males!
audreyt nothingmuch: posting
nothingmuch nothingmuch.woobling.org/may.html
use that
it's got better html
audreyt moose
SamB cpan: the gathering?
who publishes that? 03:01
audreyt SamB: cpan.org
nothingmuch wait, don't use that yet
SamB audreyt: how is it played?
nothingmuch now you may use it as it is actually longer than 0 bytes
SamB: it may or may not involve drinking 03:02
SamB it isn't a TCG?
audreyt SamB: in normal constructed play, you build a deck (also known as a stack) using about 60 CPAN modules
the trick is that 90% of CPAN is useless to your tasks
aufrank nothingmuch: you use Quantum::Superposition for the CPAN drinking game?
audreyt and some modules don't play well together
nothingmuch search.cpan.org/dist/Perl6-Bible/li...le/S27.pod
lambdabot Title: Synopsis_27 - Perl Culture [DRAFT] - search.cpan.org
03:02 AppleIver left
nothingmuch audreyt: i've never actually played =/ 03:02
03:02 cmarcelo joined
aufrank I built an all ACME deck once 03:02
audreyt so you'll have to pick and choose the one that has synnergy, those that APIs fit together, so it can run smoothly 03:03
also, it's a trading card game, so sometime you will have to trade your modules for other people's
and build new modules in exchange of new features in other modules that use them
also, just like MtG, new generations of modules come at a very fast pace
so you'll have to continuously look at the "recents" catalog and remember new cards, what they are useful for, and their fluctuating popularity 03:04
03:04 robkinyon joined
audreyt most newcomers download deck lists from reputable sites like "cpan advent" or "perl.com" 03:05
aufrank a couple of modules are just totally overpowered and lead to changes in the core mechanics
audreyt but most of the fun comes from unearthing previously-thought-as-uninteresting modules
aufrank like Switch.pm
audreyt and turn them into interesting combinations
yes. we refer to them as "broken" 03:06
aufrank I didn't want to call Switch.pm broken in this channel ;)
SamB they don't allow that kind of thing in Haskell...
audreyt the core team eventually updated the core mechanics for that, making it a new keyword ability
aufrank thought I might call some of your wackier preprocessor antics broken =P
nothingmuch acme++ 03:07
ingy++
audreyt and of course, the Sith Edition Rules are totally simplified and would require some relearning...
nothingmuch we need to play that game at my next birthday
audreyt *Sixth Edition
aufrank Sith Edition
priceless
audreyt incidentally, I was part of the expert team helping drafting MtG's Sixth Edition rules...
and that was the last time the rule changed in a major way 03:08
everything worked flawlessly since then.
nothingmuch you seem to weasel your way into various expert groups fairly consistently
SamB: did you know that unicode is turing complete because of audreyt?
;-)
SamB turing complete is it?
it has symbols for turing machines?
nothingmuch IIRC there was a fixed point combinator somewhere
aufrank I heard audreyt invented an internet that actually is a series of tubes, and not a big truck
nothingmuch or maybe that was a joke 03:09
audreyt nothingmuch: that was a joke
nothingmuch see also ';-)' 4 lines up
SamB where is the internet that is a big truck?
nothingmuch audreyt: thanks for posting summary
audreyt SamB: the .html is ready? 03:10
er sorry.
nothingmuch: the .html is ready?
SamB heh
aufrank ok, I've got to climb into bed and do some reading for $work 03:11
audreyt SamB: I was referring to the IDEOGRAPHIC DESCRIPTION CHARACTER characters, which allows for fractal-building algorithm in unicode
SamB oh?
nothingmuch audreyt: yes, i told you that it wasn't 0 length anymore =)
PAY ATTENTION WOMAN! 03:12
nothingmuch.woobling.org/may.html <-- here
lambdabot Title:
audreyt SamB: i.e. you can say "I have a grapheme here. its upper side is "X", its lower left is "Y", its lower right is yet another composed char, which has the upper side as ..."
nothingmuch hmm, i should add a title
to the template
*shrug*
note amazingly cool rt ticket autolinkfying
and <msgid:...> based links
aufrank audreyt: are you flying back to boston, all the way home, or are you still wandering (while presumably not lost)?
audreyt SamB: however, recursion depth is currently limited. so I joked that we should have a IDEOGRAPHIC DESCRIPTION CHARACTER FIXPOINT
SamB: then it will be turing complete
nothingmuch which refer to named anchors when the thread is in the summary
but point to the archives when it is hidden 03:13
SHAZAAAAM
SamB audreyt: heh
aufrank nothingmuch++ # fancy linkifying
audreyt just like SQL and its "WITH RECURSIVE" selects
nothingmuch aufrank: nothingmuch.woobling.org/Mail-Summary-Tools 03:15
lambdabot Title: Index of /Mail-Summary-Tools
nothingmuch it should be updated with this weekend's hacking by tomorrow
and on the CPAN within a few days
audreyt aufrank: pdx -> sea -> tpe (tthat's .tw) in 48 hrs
nothingmuch although i doubt you actually care ;-)
good night folks
03:15 lollan joined
nothingmuch adhers to the /away message that has been set for the past 4 hours 03:15
audreyt nothingmuch: 'nite. (posted, btw.)
aufrank audreyt: good luck with the traveling! 03:16
nothingmuch pugs.blogs.com/pugs/2006/07/the_per....html#more # EDOUBLEESCAPEDHTML
audreyt # FIXEDTHANKS
nothingmuch <h2> not allowed?
i can emit with alternate headers, like the use.perl html 03:17
audreyt no it's a misclick. nox fixed
*now
nothingmuch audreyt++ 03:18
kudra++
*poof*
audreyt kudra++ nothingmuch++
aufrank ok, gnight all. congrats on what seems to have been a good OSCON!
audreyt nite aufrank :) 03:19
merlyn Yeah, wished I could have been there. But since they haven't invented cloning or teleportation yet...
audreyt prepares for boarding
aufrank merlyn: how'd the cruise go?
audreyt well, cloning is actually available, but probably not the variant of cloning that can help you...
merlyn the cruise was great. very very warm in europe though 03:20
aufrank glad to hear it went well :)
SamB also we don't have the mind-transfer thing
merlyn pompeii was amazing
aufrank I bet
merlyn and finally seeing classic sites like the colliseeum of rome, etc 03:21
cmarcelo audreyt: hi! how's going? when will you be online again (supposing "boarding" => leaving)?
aufrank enjoyed the new article, btw
audreyt cmarcelo: hi! going quite well. will be online again in... I think 48 hours
will pop in from airport(s) if possible 03:22
but not guaranteed
merlyn I'm flying for 16 hours today too.
west though. so it'll be a Very Long Monday
cmarcelo audreyt: in what timezone will you be?
audreyt will stay on land for an entire month (or maybe more) after this
merlyn depart rome 11am, arrive portland 8pm
audreyt cmarcelo: CST, that's GMT+800
merlyn 9 timezones in th emiddle 03:23
audreyt but I'll still be part of the Pacific Standard Tribe
at least for a week or so, I think
cmarcelo ok... re soc: i'll notify haskell-cafe (when my gmail.com connection come back) about progress and will move towards MArray, DiffArray and family.. 03:24
audreyt *nod* does ByteString all work with .Map? 03:26
cmarcelo audreyt: as far as i tested yes. =) it's the best choice for both map types.. tested with checkdup example and with the Atom/Intern experiment... 03:28
audreyt cmarcelo: also, re "reevaluate benefits"
03:29 hikozaem_ joined
audreyt Bulat's comment about Arrays is mostly because Perl's idea of "Arrays" and his idea about arrays are not neccessarily the same thing :) 03:29
for fixed-size arrays with no "holes" between elements, surely Judy cannot be any faster
wolverian heh, I see pugs.blogs.com uses the new MS fonts.. they're nice 03:30
audreyt the use case for pugs's "arrays" is @x with @x[0] set and @x[1000000] set but nothing in between; in that case regular arrays are going to consume a lot of memory
03:30 agentzh joined 03:31 agentzh joined
cmarcelo audreyt: hmmm 03:31
audreyt and though I think bulat's library does array resizing, it does so by a O(n) memcpy into another memchunk
and so is the least efficient possible way to do that
so I'd say it's different meaning we attach to the term "array" 03:32
bottom line is that it's good to offer choices as long as the interface are similar enough.
cmarcelo audreyt: is it indeed very common in perl this use case you described?
audreyt skip arrays? 03:33
cmarcelo case: @x[0] and @x[10000000] set and nothing in between... 03:34
03:34 hikozae__ joined
cmarcelo don't know what is "skip arrays".. just learn last week what's exact is list context =o) [by reading Learning Perl hehe] 03:35
audreyt s/skip/sparse/
cmarcelo oh
well, skip arrays, are they really *that* common? 03:36
audreyt sorry, "sparse array" is the correct term 03:37
cmarcelo I see that Judy could benefit this cases, but I feel that they're a few specialized ones.. 03:38
audreyt *nod* honestly, I do not know how common it will be for perl6, or if it will be the default. 03:39
I know, however, that sparse arrays has come in p6l a few times and it'd be nice to have it as the default container type for arrays
and also that pugs currently use a very slow sparse array as default contianer for arrays 03:40
and also that parrot currently uses a faster sparse array as default container for arrays
cmarcelo the default is a sparse one?
audreyt in both pugs and parrot currently, yes
cmarcelo well, at least we'll catch up with parrot =).. 03:42
audreyt yeah. the current pugs one is horrible :) 03:43
it doesn't segfault immediately like p5 when you do $_[999999999] = 1
but the lazy population is still linear
and so could still take a long time
and JudyL will definitely help.
this is currently O(1) in parrot: 03:45
.loadlib 'perl_group'
.sub main $P0 = new .PerlArray $P0[999999999] = 1
.end
er
cmarcelo but, as I see, it's enough to use already existing Map2 (aka JudyL) for this then? (ie. no need new api)
audreyt .loadlib 'perl_group'
.sub main
$P0 = new .PerlArray
$P0[999999999] = 1
.end
(though with another "9" it would overflow)
yes, I think it would be sufficient without new api 03:46
I wonder if Map2 is better named IntMap...
03:48 hikozaemon joined
cmarcelo probably yes.. my name choices were all very odd.. is the Array stuff "near" that IHash and friends? 03:48
audreyt it's src/Pugs/Types/Array.hs 03:49
look for
instance ArrayClass IArray where
and note the hideously typedef'ed IArray 03:50
type IArray = TVar (IntMap (IVar VScalar))
cmarcelo there's a on-going change on AST, right? this Pugs/Types are kind of independent of this change, right? 03:51
audreyt not entirely independent
but the Pugs/Types/*.hs
should be reusable
that is, the instance definitions
will be reudsable 03:52
so if you make IArray retypedefed into Judy and the instance sitll work
it'll just need minimal retooling (probably automatable) to be used by new CapInternals AST 03:53
(hopefully I'm making some sort of sense :))
03:54 hikozaem_ joined
cmarcelo yes.. at least until /CapInternals/ =o) 03:54
Cap = Captures ?
03:54 nirgle joined
audreyt yeah 03:55
nirgle woves
nirgle waves
audreyt new-ast-based-on-the-capture/signature-calling-convention
nirgle: *particles*
nirgle dual
cmarcelo which Syn describes that?
audreyt S06 03:56
03:58 hikoza___ joined
svnbot6 r11830 | audreyt++ | * Perl6::FAQ::Capture - remove prefix * and change it to prefix [,] 04:00
cmarcelo hmm. for doing pop and unshift would be necessary call (sort . keys) to discover what element should be popped/unshifted.. because judyL doesnt guarantee any order at all.. i wonder if it's too much overhead? 04:01
audreyt judyl is ordered no?
only JudyHS is unordered
judyl is sorted
cmarcelo just a sec 04:02
audreyt in particular, judyLPrev etc guarantees sorted order 04:03
would be somewhat pointless if we had to call sort :)
nirgle waves
04:03 nirgle left
cmarcelo indeed =o) 04:04
04:04 buubot joined
cmarcelo (finally google.* just came back to work in my ISP)... 04:05
audreyt now boarding for real :) 04:06
cmarcelo: sorry for having to disappear again
*wave*
cmarcelo ok. =)
spinclad re big truck internet: 'do not underestimate the bandwidth of a pickup truck tearing down the interstate packed with exabyte tapes' -- ancient usenet proverb 04:10
merlyn it wasn't a pickup truck
it was a station wagon
and it was pre exabyte 04:11
spinclad 1/2 inch tapes?
merlyn 9-track
spinclad it's been filtered through my meme-regenerator 04:12
merlyn do not underestimate the bandwidth of a station wagon loaded with 9-track tapes
it was a UUnet guy, if I recall
spinclad think i remember john gilmore's sig as my source for it 04:13
and that sounds very like it
dolmans how to declare a function parameter which is also a function but with a specific signature? 04:26
gaal hullo. 04:29
nothingmuch: pong 04:30
audreyt dolmans: sub moose (Code:(Specific --> Signature) $x) {...} 04:31
(not yet implemented) 04:32
&x also works in place of $x
dolmans: write a test?
wolverian (and with &x you can call it as x in the body of the sub, iirc) 04:33
audreyt (do you have a commit bit yet? if not, what's your email addr so I can send you one?)
yup, what wolverian said.
dolmans audreyt: oh, i missed a little while, trying to understand what you says now:-P 04:35
gaal audreyt: oh, is [,] final?
04:36 kanru joined
audreyt gaal: as final as anything else... 04:37
dolmans basically, i want to port some haskell functions to perl6, but without a type system, i found it is difficult. what do you guys think?
audreyt dolmans: I think you are totally correct :)
please write tests
including type variables 04:38
dolmans what tests, i just missed
audreyt sub flip (Code:(::a, ::b --> ::c), ::b, ::a --> ::c) {...}
sorry, boarding for real. other lambdacamels here will walk you through :-) 04:39
*wave*
dolmans ah yeah, i'd like to
gaal audreyt: I thought it was dismissed because it didn't make sense with slurpies and namedds.
have a nice flight!
dolmans bonne vayage audreyt
so, does perl6 will have the type system like haskell, or just pugs or not either? 04:48
gaal dolmans: it's not fully specced yet, but the intent is to have some typing inferencing, yes. 04:59
there's considerable desire to bridge between fully static typing and dynamic freedom. 05:01
dolmans gaal, got it. 05:02
introducing static typing into perl6 probably means some very big change perhaps end with uncontrollers, i thought. :-) 05:03
perhaps this is the right time for language design talent. 05:04
gaal I'm not familiar with the term "uncontroller"... 05:05
dolmans i mean ending with a unsteady status 05:06
05:08 amnesiac joined
gaal gotta go off to $work, see 'ya & 05:08
05:21 xern joined 05:44 MacVince joined 05:59 bsb_ left
ingy seen audreyt 05:59
jabbot ingy: audreyt was seen 1 hours 20 minutes 38 seconds ago
ingy audreyt: ping 06:00
audreyt pong 06:06
hi from sea/tac airport
ingy audreyt: hi 06:07
audreyt dolmans: may I have your email addr to send you a commit bit?
ingy: hi
ingy audreyt: I need you to tel me the various ways you use YAML
for my grant submission 06:08
audreyt I use it for ~/.svk/config
for Jifty::Plugin::REST
for handy replacement for Data::Dumper::Dumper
for shipping pugs AST from Haskell to Perl6 to Perl5 and back
gaal Perl 6 bytecode
audreyt for lots of config files 06:09
(pugs smoker info, etc)
gaal handy pugs smoke data intermediate format
audreyt gaal: how's your $job and schedule about 5 days from now? 06:10
once I'm over jetlag, I'd like to make CapInternals reality
gaal whee!
probably I have a few days available 06:11
you're talking saturday? I have saturday :)
but not sunday
svnbot6 r11831 | cmarcelo++ | * Judy.Map2: toRevList. And TODO updates.
audreyt ok.
tentatively bookmarked 06:12
hm. I have 4 hours in the airport
audreyt ponders reading the 8th working draft of XHTML2 as bedtime story
gaal there are some issues with object dumping but unfortunately I can't really stay on the channel now 06:13
but I could have used your help.
basically it's a problem of redispating stuff across two typeclasses
06:13 xern joined
audreyt is there a writeup or backlog somewhere? 06:13
gaal no, it was with lumi in the mini tlv hackathon last week 06:14
but the summary is: there's a typeclass of Perl6Class, with method asPerl6Object
a typical instance method is asPerl6Object (MyCons aa ab ac) = "new(" ++ (intersperse ", " [plShow aa, plShow ab, plShow ac]) ++ ")" 06:16
class PLit a where plShow :: a -> String
ingy audreyt: thanks
gaal for example it makes the string "wouldn't" into qn'would'~"'"~qn't' 06:17
(a better implementation would be welcome, but the point is that there needs to be AN implementation)
audreyt ok
gaal the problem is how to testify that aa, ab, and ac all "do" PLit
audreyt so you want to constraint PLit over all MyCons children 06:18
without saying that in MyCons itself
gaal precisely. and hertofore all we needed when drifting was one 'derives' pseudocomment
audreyt (which would be cumbersome but possible)
gaal precisely :)
audreyt well, the canonical solution is moose
er
I mean overlapping instance 06:19
gaal moose is a good canonical solution
audreyt instance PLit a where plShow = error "hey look moose"
gaal how so?
audreyt and then
instance (Show a) => PLit a where plShow = ...some real show...
that's it I think
the "PLit a" satisfies typechecker 06:20
at runtime it gets dispatched to the more specific instance(s)
gaal ah, excellent, so there*is* a way to bless all types into a class in haskell! cool.
I was wondering about that.
audreyt :)
though I wonder if it makes more sense to just use "show".
plShow :: Show a => a -> String
gaal another problem is really technical
audreyt then you don't have to do that anymore 06:21
gaal yes, in most cases it is show
audreyt what cases isn't Show?
gaal the 2nd problem is that there is a circularity beween the actual file and the instances file, if I want to define PLit instances locally
I want to avoid havingDrIFT.Perl6Class need to know about AST internals
sorry, in all cases it is Show 06:22
audreyt erll then you don't need the PLit class
gaal in most cases plShow = show
how come? lumi mentioned something like open datatypes and maybe open types, but that's not in ghc is it? 06:23
06:23 tup joined
audreyt it's available as a preprocessor. 06:23
from andres and ralf
gaal We don't want to do that do we :) 06:24
audreyt # lambda-the-ultimate.org/node/1453
lambdabot Title: Open data types and open functions | Lambda the Ultimate
audreyt we don't unless we run out of options :)
in any case
gaal okay, so I don't yet understand the comment about not needing a type PLit
s/type/class/
audreyt so basically the types that does Show is exactly the same as types that does PLit 06:25
but PLit may override plShow for some of the types
correct?
ingy audreyt: rafb.net/paste/results/S3EVM640.html
buubot The paste S3EVM640 has been copied to erxz.com/pb/2417
ingy bbiab
gaal I guess... but you can't add a method to a class, isn't that it?
audreyt right... so you do need class PLit, sorry 06:26
the canonical solution is something like
instance Show a => PLit a where plShow = show
instance PLit SomeSpecificType where plShow = ...somespecificcode...
and enable overlapping instances
that's it
gaal okay, let's see if that works. 06:27
wait no, it doesn't -- I already have that!
let's dig up a sample error 06:28
audreyt (and let me know what commands I can run to dup it :)) 06:29
gaal audreyt: wget perlcabal.org/~gaal/tmp/plit.patch ; apply it ; util/drift.pl src/Pugs/AST/CapInternals.hs ; ghc -package-conf third-party/installed/packages.conf -isrc -imisc/pX/gaal -O0 --make misc/pX/gaal/CITest.hs src/pcre/pcre.o -o atest 06:32
dolmans audreyt: i already signed in pugs
gaal the first issue you'll see is actually (I think) not directly related, it's a circularity thing between the main module and the Instances.hs
also note the bogus types Bogus and Eval a 06:33
which were meant to be placeholders for future stuff but seem to uncover polymorphism trouble 06:34
dolmans and trying to write some test code, but first i need to coordinate myself to pugs style or any open source project way.
06:34 bsb joined
gaal dolmans: take a look around the t/ directory 06:34
the README file there is useful.
dolmans gall: yes that what exactly what i do now 06:35
*that is
sorry, gaal indeed.
audreyt dolmans: also, add yourself to AUTHORS :)
gaal dolmans: and don't worry about committing not 100% perfect stuff, we have source control :)
dolmans audreyt: i'd like to do that!
gaal: that what i worried 06:36
*that is
gaal audreyt: also, there's something I don't understand about cabal. I wanted to register a newish fps so I can run lambdabot and GOA (xerox++). but when I have a systemwide fps 0.7, pugs won't build. 06:38
audreyt gaal: duped
gaal I thought cabal was meant to solve that?
audreyt gaal: it was. and we need to tell cabal to use installed/package.conf and shadow systemwide one
06:39 cookys joined 06:40 tup joined
gaal I must return to some C++ debugging, sorry, will check back occasionally 06:40
audreyt ok. so it's data Perl6Class a => Eval a = Eval a 06:42
that was causing all this
why do we need the constraint here?
if it's relaxed all works
gaal really?
audreyt yeah.
data Eval a = Eval a 06:43
then it compiles
gaal I added it because I was geting an error about 'can't deduce context Perl6Class'...
audreyt ok. the typesystem-sudoku seems to have been solved 06:46
I'll post a new patch
gaal &audrey .= wrap { &?ROUTINE++; call } 06:50
huh that doesn't work. 06:52
guess I'll have to keep doing it manually then: audreyt++ # sudoku solver with class 06:53
06:54 nekokak joined 06:57 shachaf joined
audreyt gaal: perlcabal.org/~audreyt/tmp/plit-compiles.patch 07:03
note that this is purely sudoku-solving
and I have no idea what you are doing :)
(still uploading)
gaal just putting the chocolate on the moose 07:04
audreyt mm syntactic chocolate 07:06
ok, wget now
(this is against trunk)
gaal ACK 07:07
audreyt looks at ingy's url now 07:08
07:08 iblechbot joined
audreyt ingy: s/YAML::Syck.pm/YAML::Syck/ 07:09
ingy: not sure if "Syck lacks good unicode support" is at this moment relevant... 07:10
with $ImplicitUnicode=1
miyagawa yeah, YAML::Syck works without any problem
it has "good unicode support" 07:11
audreyt ingy: also, the proposal isn't clear whether you'll do a libyaml binding
or a pure-perl port of pyyaml
or both
ingy ok
audreyt arguably you should say pure-perl, and then list its benefits
and mention that a libyaml binding would be available on machines with compilers
but a pure-perl available-anywhere implementation that is also compat with libyaml 07:12
is clearly desirable
ingy yep
audreyt (though I wonder if we can just embed libyaml and be done with it...)
...but if you would like to work on pureperl yaml.pm that's still a wonderful thing to do
ingy well it doesn't exist yet...
miyagawa $JSON::Syck::ImplicitUnicode = 1; encode('JavaScript-UCS', JSON::Syck::Dump($data)) makes me happy 07:13
audreyt do you believe his SoC will fail? :)
ooh, Encode::JavaScripdt::UCS. 07:14
I wonder if we should fold that back into JSON::syck
nah. different layers.
miyagawa right, different layers 07:16
but it works pretty well with JSON::Syck, either by encoding to JavaScript-UCS before or after Dum 07:17
Dump
audreyt cool :)
bsb ingy: s/bast/best/ 07:19
07:22 dakkar joined 07:23 Southen joined 07:30 crem joined
gaal if I have a directory that already has some svn:ignore props on it, does 'svk propset svn:ignore dir somefile' just add somefile, or overwrite everything? 07:37
svk propset svn:ignore somefile dir, that is. 07:39
audreyt overwrite 07:43
you want propedit
gaal I really don't, I want this in a script 07:45
pugs has like 30 ignorable files and the property interface sucks
I got a lot of rejects in the patch :( 07:46
audreyt :/ 07:49
07:53 drrho joined
dolmans when make install parrot, i got an error `copy docs/ops/bit.pod to /opt/share/doc/parrot/docs/ops/bit.pod: Permission denied', what dose that mean? i sure have that directory's permission. 07:59
audreyt dolmans: no idea... irc.perl.org #parrot may be slightly more helpful on that 08:01
or [email@hidden.address]
dolmans ah, i got it, it says the docs/ops/bit.pod is unreadable! i dunno why that happend. 08:02
08:02 Noriega joined 08:29 elmex joined 08:33 integral_ joined, integral_ is now known as integral 08:42 marmic joined 08:43 buetow joined
nothingmuch gaal: pong-to-pong 08:46
09:04 revision17_ joined
nothingmuch w 24 09:15
09:24 kanru joined 09:32 pdcawley joined
gaal nothingmuch: well, what? :) 09:33
09:34 iblechbot joined
nothingmuch gaal: nothing now =) 09:36
i wanted you to post a summary when audrey was not here
but she has since done so herself
=)
gaal wonders if his blogs.com account is any good after a month passes since creating it 09:39
...and I don't understand why anyone would want to use that service when LJ gives a lot more features for free
nothingmuch file hosting?
*shrug* 09:40
ask audreyt?
09:40 chris2 joined
gaal does typepad give arbitrary file hosting? 09:40
audreyt yes.
gaal how much weight? 09:41
audreyt 1,000 megabytes
gaal nice
audreyt but the reason I use typepad is just that miyagawa and sky and btrott can be trusted to fix problems speedily :) 09:42
gaal lj gives audiopost and photo hosting only, 2gb for paid accounts
audreyt and I don't know the lj people that well
gaal audreyt: brad's fast too :) 09:43
audreyt sure :)
miyagawa thanks audrey. even though none of Artur, Ben and I are actually working on TypePad
gaal (LJ people)++
miyagawa :)
audreyt miyagawa: I know that :) but you dig up src fast when I asked :)
how's vox doing? 09:44
miyagawa it's in a pretty good shape 09:45
audreyt cool :) 09:46
pasteling "cmarcelo" at 200.232.237.151 pasted "SVK/Command/Propadd.pm followed by the needed diff to make it work on trunk.. just a sketch.." (132 lines, 3.5K) at sial.org/pbot/18721 09:52
cmarcelo gaal: if you use svk from trunk, you might want to take a look on this paste above...
dolmans it seems like to build a binary ghc bundled with all kinds of tools suitable for building pugs is a meaningful work. 10:15
gaal cmarcelo: cool! 10:16
dolmans: ? there aren't very many dependencies for pugs...
and non of them AFAIK need a GHC rebuild! 10:17
dolmans gaal: at least hs-plugins which depends a chain of tools.
not ghc itself, really bundled tools. 10:18
meppl guten morgen 10:22
10:30 crem_ joined 10:33 xern joined 10:38 kanru joined 10:50 elmex joined 11:09 pdcawley joined
gaal dolmans: hs-plugins is really, really optional. 11:19
nothingmuch audreyt: ping? 11:27
gaal nothingmuch: I think she's on a long flight 11:32
nothingmuch yeah, i figured 11:33
nothingmuch needed pause-admin-fu
short notice
?eval say "Module::Build--" for 1 ...
gaal what lowercase module are you uploading now? :-P
nothingmuch nothing of the sort
but Module::Bane decided it's a good idea to store a version object in the meta.yml
11:34 evalbot_11829 is now known as evalbot_11831
evalbot_11831 OUTPUT[Module::Build-- Module::Build-- Module::Build-- Module::Build-- Module::Build-- Module::Build-- Module::Build-- Module::Build-- Module::Build-- Module::Build-- Module::Build-- Module::Build-- Module::Build-- Module::Build-- Module::Build-- Module::Build-- Module::Build-- Module::Build-- Module::Build-- Module::Build-- Module::Build-- Module::Build-- Module::Build-- Module::Build-- Module::Build-- Module::Build-- Module::Build-- Module::Build 11:34
nothingmuch so i uploaded a broken release
11:34 hugues23 joined
nothingmuch karma for Module::Build? 11:34
jabbot nothingmuch: for Module::Build? has neutral karma
nothingmuch karma for Module::Build
jabbot nothingmuch: for Module::Build has neutral karma
gaal haha
nothingmuch jabbot: Module::Build--
jabbot nothingmuch: I'm not sure I understand you fully.
nothingmuch ugh!
jabbot: IT FUCKING SUCKS OKAY?!
jabbot nothingmuch: Tell me more about that.
nothingmuch head-on-keyboard
gaal ?eval "orz"
evalbot_11831 "orz"
nothingmuch oh dear god 11:35
even when the M::B changelog claims to have fixed it it's still broken
nothingmuch reverts to an older version
gaal This is MyModule, version HASH(0x814ec28)
nothingmuch helgafell.rhi.hi.is/pic26777.jpg 11:37
gaal: yep
11:39 agentzh joined
agentzh an S05 related question: is "m:/.../" valid? 11:42
?eval 'abc' ~~ m:/\w+/
evalbot_11831 Error: unexpected end of input expecting "\\", "$/", "$!" or ":"
agentzh hmm...seems not. maybe it's just another synopsis nit. :) 11:43
?eval my @a = 1..3; @a[0]++; @a.say 11:44
evalbot_11831 OUTPUT[223 ] Bool::True
agentzh ?eval my @a = 1..3; @a>>++; @a.say
evalbot_11831 Error: Can't modify constant item: VInt 1
agentzh anyone can tell me what's happening here? 11:45
?eval my @a = 1..3; @a>>.say; @a.say 11:47
evalbot_11831 Error: unexpected ">" expecting "::", term postfix, comment, operator, postfix conditional, postfix loop, postfix iteration, ";" or end of input
agentzh :(
11:49 agentzh left 12:13 Aankhen`` joined 12:33 Limbic_Region joined 12:49 mjl69 joined 12:51 pdcawley joined
gaal agentzh: I think that's just not implemented yet, but should be made to work. 12:54
12:57 pdcawley_ joined 12:58 Kattana joined 13:01 mauke_ joined 13:06 esammer is now known as esammer_home 13:16 mauke_ is now known as mauke, bpphillips joined 13:26 [particle] joined, mako132 joined
svnbot6 r11832 | fglock++ | pX/Pugs-Compiler-Rule - refactored positional captures in Regex runtime 13:35
r11831 | cmarcelo++ | * Judy.Map2: toRevList. And TODO updates.
13:38 frederico joined 13:46 nothingmuch_ joined 13:50 fglock joined 13:54 cmarcelo left 13:59 nothingmuch joined 14:00 miyagawa joined 14:17 hugues23 joined 14:31 hcarty joined 14:34 vel joined 14:43 renormalist joined 14:51 offby1 joined, offby1 left 14:59 amnesiac joined 15:04 cjeris joined 15:13 cdpruden joined 15:14 bpphillips joined 15:16 foo\ joined 15:24 prefiks joined 15:26 xinming__ joined 15:27 Aankh|Clone joined 15:33 justatheory joined 15:38 MacVince joined 15:39 aufrank joined 15:52 rodi joined 15:53 xinming joined 16:00 xinming joined 16:03 crem joined 16:04 buetow joined 16:08 bpphillip1 joined
gaal mooses 16:10
16:10 hexmode joined
aufrank moose! 16:16
16:17 ludan joined
svnbot6 r11833 | gaal++ | * Runtime dumping of Haskell data as Perl 6 objects: asPerl6Object 16:33
r11833 | gaal++ | method works for many of the simpler types in CapInternals.
r11833 | gaal++ | TODO: hashes, buffers, non-pure types.
r11833 | gaal++ | lumi++ nothingmuch++ audreyt++
r11833 | gaal++ | * update demos for this in misc/pX/gaal/ (see the source or ask)
16:33 xinming joined
svnbot6 r11834 | gaal++ | some ignore props 16:36
16:36 lisppaste3 joined 16:40 Aankhen`` joined
nothingmuch was doing the teddybear role for that commit =) 16:40
btw, how's your new debugging moose acclimating?
ingy hi 16:41
nothingmuch hi ingy
(how|what) was the movie?
gaal nothingmuch: the moose is excellent! doesn't have a name yet 16:42
nothingmuch how about ג'×Øבוהקון
gaal I perfer something I'm more likely to remember :-) 16:43
16:45 MacVince left
svnbot6 r11835 | gaal++ | * add more ignore props, (uncluttered 'svk st')++ 16:48
gaal should .installed-pkg-config, .setup-config, fps.cabal and dist in and around third-party/ be ignored too? I think so but am not sure. 16:50
svnbot6 r11836 | gaal++ | correct ignore prop 16:54
17:08 Aankh|Clone joined
nothingmuch i75.photobucket.com/albums/i309/Cus...secopy.jpg 17:15
i75.photobucket.com/albums/i309/Cus...secopy.jpg 17:16
oops
17:27 fglock joined 17:30 justatheory joined 17:31 penk joined 17:38 pen1 joined 17:47 shachaf joined 17:54 lollan joined 17:57 Aankh|Clone is now known as Aankhen`` 18:01 frederico joined 18:07 daxim joined 18:09 robkinyon joined
pmurias fglock: hi 18:13
18:19 weinig|sleep is now known as weinig
pmurias sorry for not getting round to writing the string parsing part 18:24
i was on a alogrithmic camp and it was too tiring to write much perl 18:29
also should the variable interpolation et al. be done after string concatation works? 18:31
did you look at ext/Parse-Rule?
18:34 pen1 is now known as penk 18:39 vel6608 joined 18:57 hcarty left
ingy hi audreyt 18:58
er, nm (unless...)
nothingmuch: ping
nothingmuch ingy: pong 18:59
ingy want to review my YAML proposal?
nothingmuch sure
ingy rafb.net/paste/results/QFC4Iw52.txt
about to send it off...
nothingmuch Synopsis: - no quote op? 19:01
|- ?
or >- ?
also, benefits should probably be a list ;-)
ingy oops 19:03
nothingmuch ingy: a little more tech info, methinks
but then again, i don't relaly know
show what YAML.pm doesn't have that PyYAML will give it
ingy ok, thanks 19:04
nothingmuch ingy: i'll be happy to help you with that grant 19:07
that is, without $$$ or anything
just as a volunteer
i want a YAML.pm that i can't complain about ;-)
19:10 Southen_ joined 19:15 crem joined 19:16 weinig joined 19:18 bernhard joined 19:26 vel joined 19:29 prefiks joined 19:39 bpphillips joined 19:40 Southen joined 19:55 bpphillip1 joined 19:57 bpalmer joined
fglock pmurias: hi/pong 19:58
19:59 larsen joined 20:03 HEx joined 20:11 Southen_ joined 20:17 salty-horse joined 20:32 Southen joined
pmurias fglock: pong 20:35
20:53 bpphillip1 left 21:01 crem_ joined 21:15 iblechbot joined 21:21 crem_ is now known as crem
pmurias sleep& 21:25
21:28 SageLT joined, Southen_ joined 21:37 beppu joined 21:42 crem_ joined 21:48 bpalmer left 21:49 cjeris left
xinming 'window stick off' 21:54
21:56 Southen joined 21:58 rnorwood joined 22:02 prefiks left 22:07 crem_ is now known as crem, crem is now known as crem_ 22:08 crem_ is now known as crem
svnbot6 r11837 | Sage++ | Changed the extractType function to really extract on just type, not matchWithoutEnc. 22:20
r11838 | Sage++ | Refined and cleaned the code for the Oo file handle translations .close and .readline. Both should be more correctly applied, with better junk handling, and the code is now much more readable. No claims that they're absolutely complete, but they're better. 22:29
r11838 | Sage++ | Modifiers to regexs are now better handled with cleaner code. Handeled modifiers include /g, /i, and /e (for s/). They can also be handeled in combination (such as /ige). The cide now correctly catches all of these modifiers on all regexs.
r11838 | Sage++ | The regex in split() is now translated, using the same functionality as normal regex translations.
r11838 | Sage++ | The handeling of foreach loops is now improved, both in results and in clarity of code. Foreach loops can vary quite a lot, so I'm sure it can still be broken, but it's now better and easier to fix, when it is broken.
r11838 | Sage++ | Some other minor cleanups to the code have also been made.
r11839 | fglock++ | pX/Pugs-Compiler-Rule - passes 217/241 22:44
clkao almost there! 22:45
fglock clkao: almost :) 22:46
clkao fglock: we need to have runtime detection of function being written in perl6 for using the data::bind callconv. type-based mmd requires this. 22:47
fglock I think PCR now has the internal information to implement the remaining semantics 22:48
clkao: how about calling perl5 subs (p5 cpan modules)
clkao: runtime detection - do you mean in eval() ? 22:49
clkao no, knowing if you should call func with ("1234") or ([\1234], {}) 22:50
fglock clkao: like we do with objects vs. scalar 22:51
it checks if blessed()
we need to check if bind() 22:52
Emitter::Perl5 #586 22:54
does Data::Bind has a function that tells if a sub is bind?
clkao ya. you can see if there's bind info in the gv.. another thing is things like %hash.keys.sort 22:56
fglock I think it should behave like (keys %hash).as(Array).sort 22:58
%hash is unboxed, but keys returns a boxed array 22:59
it would work with unboxed array too 23:00
but then you would have to test for 'objectness' in every single method call 23:01
so maybe 'keys %hash' returns unboxed array, and '%hash.keys' returns boxed 23:02
so that you can chain method calls
back in ~2h & 23:04
23:04 fglock left 23:06 AppleIver joined, AppleIver left 23:17 nothingmuch_ joined
ludan how can i get out from a pugs shell? 23:29
TreyHarris ludan: control-d 23:30
(i.e., send an eof)
ludan thanks
salty-horse ludan, or :q
TreyHarris on windows, i guess that would be control-z
ludan are there some tutorials to understand the main P6 features implemented in pugs ? 23:31
TreyHarris ludan: theoretically, all perl 6 features are implemented in pugs. in reality, that isn't true, but we pretend that it is :-)
so you should look at the synopses and try things
ludan yep, i'm lurking on this channel by a long time :P
TreyHarris do you know where the synopses are?
dev.perl.org/perl6/doc/synopsis.html 23:32
lambdabot Title: Synopses - perl6
ludan thanks
TreyHarris some very advanced and esoteric features are implemented, and some very basic features are not, so you generally should just try and see
23:33 weinig is now known as weinig|bbl
ludan ok 23:33
TreyHarris: which is the major issue to release a stable and complete pugs? 23:34
TreyHarris ludan: heh. for it to be done :-) 23:36
it's a very iterative process, AFAIK, nobody's thinking in terms of project plans and goalposts and gating items
ludan you mean because it's to difficult to implement?
TreyHarris that's as good a non-answer as any.... nothing's finalized. i suspect at some point, somebody will notice that it's useful and stable enough that they'll suggest doing a distribution. but that's many months off, i think 23:38
TreyHarris will be back in a few & 23:39
obra 1: the language spec needs to be finished.
ludan it should be enough :) 23:41
daxim oh fark, I missed qixi 23:46
23:52 weinig|bbl is now known as weinig

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!