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.
00:09 xinming_ is now known as xinming
meppl gute nacht 00:16
00:29 dolmans joined 00:44 xinming joined 00:50 fglock joined 00:53 loumz joined 00:57 kanru joined 01:13 dbrock joined 01:18 byte joined 01:25 robkinyon_ joined
robkinyon_ obra: ping 01:25
01:44 lollan joined
xinming ?eval sub hello { "hello".say; }; hello(); 01:47
evalbot_11936 OUTPUT[hello ] Bool::True
01:50 nirgle joined 01:57 weinig is now known as weinig|sleep 01:59 nirgle left 02:08 knewt_ joined 02:28 knewt joined 02:35 knewt joined 02:45 knewt joined 02:59 knewt joined 03:09 knewt joined 03:17 knewt joined 03:24 knewt joined 03:26 MacVince joined 03:32 knewt joined
azuroth what IM client do ion users use? 03:36
Khisanth why would the wm determine what IM client someone would use? 03:50
azuroth I just find that some programs don't quite fit in with ion 03:51
MacVince how about centericq? 03:54
Khisanth install bitlbee and then just use your IRC client :)
since you are already on IRC, finding an IRC client that fits is assumed to be a problem that has already been solved! 03:55
azuroth it works, but it doesn't quite fit :-p
Khisanth well go find an IRC client, at least you can reduce two problems into one 03:57
MacVince azuroth: there are plenty of choices: xchat, irssi, Emacs, etc.
03:57 rashakil joined
azuroth hmm. maybe I'll look at centericq 03:57
04:03 knewt joined 04:05 azuroth_centeric joined
azuroth_centeric test 04:14
gaal morning! 04:30
04:30 MacVince left
azuroth_centeric good morning 04:31
04:33 azuroth_centeric is now known as azuroth
azuroth NICK azuroth 04:33
04:45 azuroth_centeric joined 04:46 azuroth_centeric joined
azuroth_centeric hmm. I can't seem to figure out how to close windows in centericq... 04:52
Khisanth bad interface? 04:54
azuroth_centeric yeah, it may seem so 04:56
I don't think I would mind it if I could close windows though 04:57
Khisanth and why is ion causing so much trouble? :)
azuroth_centeric Ion rocks, just gaim doesn't ;-p 04:58
05:07 amnesiac joined 05:18 lambdabot joined 05:19 lollan joined 05:33 kanru joined
svnbot6 r11937 | audreyt++ | * third-party/: Update fps to the released version of 0.7. 05:57
r11937 | audreyt++ | (you may need to nuke dist/build/ and rebuild after this change.)
r11938 | audreyt++ | * HsSyck: Chase up fps-0.7 changes. 06:06
r11939 | audreyt++ | * de-bump version for now 06:19
06:32 shachaf joined
gaal I keep forgetting how to do this. I want to 'svk desc' r11939, but not my local 11939. How? 06:37
clkao uhm, you can't easily
gaal This is a mirror, and alarmingly 'svk desc 11938 svn.openfoundry.org/pugs' doesn't do what I want 06:38
that's it, I'm moving to MS VSS
:-p
clkao svk desc `svk log -qr1766@ | grep ^r | cut -f 1 -d ' '` 06:39
does desc 11938@ //depotpath work?
gaal sec, need to pulll... what does @ mean? 06:41
clkao @ = the number is remote revision 06:44
gaal "Can't find local revision for 11938 on /." 06:45
clkao ya. so use the line above 06:46
gaal work, thanks. going to my aliases.. 06:48
06:53 lambdabot joined 06:54 nothingmuch joined 06:59 dolmans joined
audreyt gaal: ping 07:09
gaal: open types without preprocessing: 07:10
07:10 Twiztid joined
audreyt homepages.cwi.nl/~ralf/gpce06/paper.pdf 07:10
gaal: seems to be a much better solution to the scalar_fetch problm 07:15
07:20 lollan joined 07:23 lambdabot joined 07:25 lambdabot joined 07:45 rindolf joined
rindolf Hi all. 07:45
Where's Ovid tips for a TPF grant?
Google is no help. 07:46
07:47 lambdabot joined
rindolf Does use.perl.org has a search facility? 07:48
svnbot6 r11940 | audreyt++ | * newtypes are now DrIFTed correctly.
07:49 weinig|sleep is now known as weinig 07:55 iblechbot joined
rindolf audreyt: here? 07:58
Hmmm... use.perl.org/~Ovid/journal/30143 07:59
lambdabot Title: Journal of Ovid (2709)
svnbot6 r11941 | audreyt++ | * do not revisit third-party unneccessarily. 08:10
r11942 | audreyt++ | * util/drift.pl - handle "newtype" instances; carefully 08:19
r11942 | audreyt++ | sidestep the nonsupported multi-parameter data types
r11942 | audreyt++ | like "EvalT m a".
r11943 | audreyt++ | * regen instances. 08:22
r11944 | audreyt++ | * DriFT.YAML and Perl6Class: add instancesfor native types such as Word 08:25
08:25 Jedai joined
svnbot6 r11945 | audreyt++ | * CapInternals.hs: gaal's CITest and CMTest now compiles again. 08:37
r11946 | audreyt++ | * util/drift.pl: handle consecutive data declarations 08:40
r11946 | audreyt++ | without blank lines in between.
StudiesInWords is ghc available as a subversion archive? 08:49
audreyt it's available via darcs 08:50
gaal audreyt++ # wow!
audreyt darcs get darcs.haskell.org/ghc
gaal I'm very much at $work now, just popping in to see the fireworks :)
lambdabot Title: Index of /ghc
audreyt gaal: ok. my immediate next step:
Pugs.Val.Native Pugs.Val.Pure Pugs.Val.Mut Pugs.Val.IO 08:51
StudiesInWords audreyt++
gaal Oh? for smaller files?
or shorter names? :)
audreyt both :)
actually just smaller files
gaal how do you hope to avoid circularity?
audreyt and also encapsulation
I don't see an obvious case of circularity besides the main Val type itself 08:52
gaal also, did you notice the comment about the sadness of
VNative $ NInt $ IFinite 42
in haskell that's not expensive in anything by fingertyping 08:53
but in other implementations that might even be slow, no?
audreyt that can easily be solved
gaal not that I worry about tHAT :)
audreyt by instance Num Native
then you just type 42.
gaal hee :)
audreyt: circularity e.g. in objects
audreyt other implementations may surely encode things differently
below the NInt level 08:54
gaal ACK.
audreyt also, NativeInt doesn't have IFinite.
it's just haskell Int
PureInt has IFinite
(already fixed)
gaal that's worth documenting then, which parts of the AST are "official"
duh, right! thanks :)
audreyt anything in Pugs.Val is; anything not mentioned there is not, how about that :) 08:55
(for Val constructors)
gaal ;)
audreyt anyway, let me finish the refactoring first... this is falling into palce nicely
re the two main knot types
Val and Native (aka ValId)
err ObjId 08:56
(very confusing; maybe we just rename it to ValId to agree with TimToady )
gaal nothingmuch and I think you too mentioned it was worth having another level of parser produce, closer to a parse tree. but this is the nice level in which regular macros operate.
audreyt I'm ready to use .hs-boot for the two knot types
since we dropped 6.4.0 support a while ago
that means .hs-boot will actually work
SPJ says 6.6 "by ICFP" 08:57
gaal but but! objects are not generally values!
audreyt which is september
I agree. so ObjId still :)
I think timtoady means ".id as a native value"
shortening to .valid
also, I think I'm making "buf" immutable 08:58
since that's what .valid will return
("Buf" will stay mutable)
gaal ah. how can we guarantee native ids for arb objects?
audreyt: there are two Bufs (well, many more) in the spec...
brb
rindolf Hi gaal, audreyt
audreyt well, maybe we can only guarantee pure ids 08:59
instead of native ids
in which case CapInternals has it wrong currently
I kinda like pure ids. going with that then
rihi. 09:00
rindolf: hi.
TimToady: VB9 has this nice solution to the timely destruction problem:
using open('/etc/passwd') -> $fh { ... };
where "using" is just like "given" except it calls DESTROY.
wolverian hm, isn't that what C# does too? 09:01
audreyt yup
rindolf Hi wolverian
wolverian hello, rindolf
audreyt we currently has no solution to that
so it might be worth stealing...
(well, our solution is spelled as LEAVE { $fh.close })
reading SPJ's commit logs is quite inspiring :) 09:03
gaal audreyt: where is the association of "using" and "close"?
in ::IO ? 09:04
IO::DESTROY ?
audreyt "... implement some property ... strangely, I can't quite see how to prove it! ... So I've added a test instead ..."
gaal: yeah
gaal lol
09:04 revision17_ joined
gaal audreyt: we need to remember sometime that there's buf32, buf64, and stuff (s02 IIRC) with interesting coersions between them... but not yet 09:05
also, int1 which is -1 or 0 :-)
but generally, type_n types 09:06
audreyt I think it's enumerated 09:07
it's okay to fail to handle int3
so that's fine
we just add more Native ytypes
variants are cheap :)
gaal agrees ObjId should be pure. and wonders how come it wasn't that laready
ooh open types, saving the paper for tonight 09:08
cheaps except that sometime we'll have to patmatch this stuff :)
s/s//
audreyt we patmatch in exactly one place 09:09
which is class instances :)
and we drift that whenever possible
gaal which reminds me we'll need to generate ids for types too (Boxable)...
I really need to get back to $work, which is a shame because this stuff is so interesting :) 09:10
pasteling "rindolf" at 88.155.102.228 pasted "TPF Grant Proposal for audreyt and gaal" (59 lines, 2K) at sial.org/pbot/18805 09:11
nothingmuch why "valid" and not "val_id" ? 09:19
if ( $form.valid ) { ## aacckkkkkkkkkkk }
audreyt nothingmuch: builtin methods in p6 has no underscores I elieve that 'what TimToady is after
nothingmuch =( 09:20
audreyt $form.is_valid surely?
nothingmuch yes
but...
suckage
audreyt I share that sentiment
better suggestions?
nothingmuch valueid
audreyt .objid
nothingmuch "valid" is just asking for trouble
audreyt .identity
nothingmuch something involving "flat" maybe
audreyt Dog.valid is false; Dog.new.valid is true 09:21
that may or may not help :)
nothingmuch heh
mot really
but thanks for trying
audreyt I think we are in agreement.
but I can't find a convincingly enough word :)
09:21 larsen joined
nothingmuch i think we need a secret agent in @Larry 09:21
we should drug chromatic 09:22
and hyponotize him
audreyt heh
I'm a synopsis committer, but otoh I stay away from ergonomics, especially when English words is involved...
nothingmuch yes
i belive that is the right course
hmmm 09:23
has the no underscores thing been justified?
i mean, it's been working out great everywhere else
and in some places you just can't have c-ish naming conventions
e.g. MOP introspection
that's just eeeeeeeeevvvvvvil to the user base 09:24
audreyt I can see maybe that .defined should return a value.
it was my counterproposal
but that means we .defined in bool context must be fast 09:25
svnbot6 r11947 | audreyt++ | * Pugs.Val: New module space to host the "Val" type and all
r11947 | audreyt++ | types it uses; this will be the "knot" of mutually-recursive
r11947 | audreyt++ | modules where most containers must refer to the Val type
r11947 | audreyt++ | whilst being a Val subtype themselves.
r11947 | audreyt++ | There beryl, pearl, and opal pale,
r11947 | audreyt++ | And metal wrought like fishes' mail,
r11947 | audreyt++ | Buckler and corslet, axe and sword,
r11947 | audreyt++ | And shining spears were laid in hoard...
nothingmuch it can be context sensitive
but
urff
let's explain what it does
and see if we find some pretty fnords
the purpose, as I reunderstand it
09:25 weinig joined
nothingmuch is to have something slightly akin to STORABLE_* hooks 09:25
audreyt ok. it returns an immutable object that uniquely identify this object among instances of the same class 09:26
that's it
nothingmuch that is, the bare essence of the data the object is representing
audreyt @foo.vlid may just be its memory address.
lambdabot Unknown command, try @list
audreyt .valid
nothingmuch so it's just for comparison?
audreyt and indexing 09:27
nothingmuch hmm
okay
i thought it was going to give us a nice convention
audreyt it's what .object_id in other languages return
rindolf audreyt: can you please review sial.org/pbot/18805 ?
lambdabot Title: Paste #18805 from "rindolf" at 88.155.102.228
nothingmuch but maybe that's for $obj.does("ReversibleValId")
audreyt nothingmuch: reversible? 09:28
nothingmuch my $clone = Class.invalid( $obj.valid ); 09:29
that is
audreyt rindolf: 1. the url you gave on "Project Details" does not describe your project
nothingmuch perhaps it should be more
audreyt rindolf: 2. saying "may be somewhat slow" without benchmark numbers is probably not helpful
nothingmuch but
orthogonal
to digress a bit further: if an obj does the MeaningfulValId role maybe is_deeply can use that instead, and Storable can be assisted, and etc etc 09:30
anyway
audreyt rindolf: 3. need to explain a bit more of how HTML-Widgets-NavMenu helps other people; who is already using it; how it integrates with other web frameworks; etc
rindolf: sorry, I'm in the middle of a hacking run, that's all I can think of right now
nothingmuch the fnords we found: immutable, uniquely identifies, instances, essence, data
represents
audreyt that suggests .uniqid 09:31
or .uuid
j/k
nothingmuch i thgink i like "represents" a bit, maybe
$obj.urepr 09:32
$obj.urp
for "unique representation"
=/
audreyt goes back to hacking
let's call it getHashCode ;) 09:33
nothingmuch heh
"urp"
hmm
i think i prefer urp to valid
it's much more nonsensical 09:34
*shrug*
rindolf audreyt: sure, no problem, let me remedy it.
09:46 asz joined 09:54 lambdabot joined 09:57 ludan joined
gaal $vampire.vlad.impale.impale.impale 09:58
nothingmuch gaal++ 09:59
gaal .uid could work... 10:03
or .oid for that matter
or go haskellish and use a ScarySymbol 10:04
.id#
10:04 lambdabot joined
gaal .id$ etc. (find a ligis that makes sense) 10:05
is # cool in identifiers? or is $oops#I_spaced_out a var with a comment in it? 10:06
s/in/after
talking lexically here, of course with symbol table elkage you can have a var with a '#' in it. 10:07
10:18 kane-xs joined 10:24 prefiks joined 10:25 weinig joined
gaal audreyt: shouldn't Pugs.Val be called Pugs.AST.Val (etc.)? 10:28
audreyt well it is used in runtime as well 10:30
as well as serialization
I think toplevel may be warranted
10:48 shachaf joined 10:49 azuroth_centeric joined
svnbot6 r11948 | audreyt++ | * beginning of CapInternals->Val refactoring. 10:49
clkao audreyt: flickr.com/photos/clkao/207896989/ # very cool shop name 10:50
lambdabot Title: IMG_1812 on Flickr - Photo Sharing!
audreyt gaal: my idea is that Val will become abstract data type outside Pugs.Val
gaal: no casing of Vl anymore 10:51
_everything_ done in term of typeclasses
restrict the casing inside Pugs.Val.*; for the evaluator, you can ask for coercion; call methods; promote native values; call "smart" onstructors
but no casing and no explicit tycon calls 10:52
this is to avoid the massive expression-problem thing we are tackling now
clkao: nice! 10:53
nothingmuch's InterfaceIdea also dictates that we model each builtin Role as a typeclass
nothingmuch yes yes yes yes yes
audreyt and we make Val an instance of that interface 10:54
nothingmuch u/w 20
audreyt and let it throw exceptions for variants that does not satisfy them
as well as allow for testing of interface compatibleness
which is simple by collecting instance evidence at Val level
and nothingmuch won't have a reason to accuse me for nonextensible Haskell internals anymore ;)
nothingmuch ? 10:55
audreyt a visitor will take the form of (CanDoSomething a => a -> SomeResult) -> Val -> SomeResult
nothingmuch: you opined a while ago that the adhoc nature of the current runtime by pushing more and more operations into AST.Internals and Eval is harmful 10:56
audreyt goes back hacking 10:57
nothingmuch my mind is a blank 10:58
i think it's a good thing, right?
;-)
(what you're saying)
audreyt it's a good thing :)
11:10 jferrero joined
gaal whee 11:16
audreyt TimToady: is it Bool::True or bool::True 11:17
synopses disagree with itself
I think "bool" can be retired and made synonym of "bit" if really needed
and True would mean "Bool::True"
$x = True; # assume Bool semantics; same as "$x = 3" assumes Int not int 11:18
nothingmuch ++
11:19 elmex joined 11:20 dolmans joined 11:23 lichtkind joined
lichtkind g morning 11:23
11:27 Spania32 joined
audreyt lichtkind: hi 11:27
lichtkind audreyt : :) still no official perl6 tut ? i just yesterday began to continue my work :) 11:28
audreyt lichtkind: no, doc/Perl6/Tutorial/ is still unpopulated :) 11:30
lichtkind audreyt even if you dont speak german would you like to look on my aproach on wiki.perl-community.de/bin/view/Wis...alKapitel3 11:31
lambdabot Title: PerlCommunityWiki ōæ½xBB Wissensbasis ōæ½xBB Perl6TutorialKapitel3
lichtkind whu ?
and BTW audreyt i didnt listen you and still develop my perl editor in perl :) 11:33
but to the tutorial my approach was a bit theory first, but skipable
audreyt looks quite sane
lichtkind then vars then operators than control then subs then oop ..roles
audreyt how's the perl editor doing?
lichtkind quit well last week we released 0.3.3 which behaves a lot more than professional user app :) 11:34
my basic idea for this tutorial was to have one place for all operator related stuff and one place vor variables and so on 11:35
begin with simpel stuff when gets to nifty you can pass on to control structures 11:36
audreyt url to the editor? :)
*nod*
lichtkind but you still know where to find what
first chapter ist a summary what hapened and where to find what and second chapter is a bit theorie goal, design pronciple and so on 11:37
editor : proton-ce.sourceforge.net/site/en/h...news.shtml
sorry but i maintain only win port but alias and me are close before CPAN release 11:38
last major steps are currently taken
so your fine with the concept of the tut? 11:39
audreyt yup
from the bit of German language still left in me 11:40
it looks like pretty concisely written
lichtkind thanks the best are the 4 appendix 11:41
audreyt so will PCE contain more perl-specific features? 11:42
basing on Scintilla seems like a smart choice.
lichtkind appendix A ist complete index with all perl6 commands ops and all used terms that need to be axplained
appendix b list for short looking 11:43
appendix c best of feature tour , an apetizer
appendix d greatest changes for quick orientation of good perl5 programmer 11:44
audreyt of course alias is payed by tpf to plug in PPI and things like Perl::Tidy and so on are made to be plugged in 11:45
but we had to make some internal changes first
and its now called Kepher not PCE, that will come in news next week :) 11:46
audreyt what does Kepher mean? 11:47
lichtkind its old egypt term that stands for creation out of it self (writing editor with editor) and spiritual progress
audreyt "bootstrapping" 11:48
lichtkind not exactly
acording the legend it is the beatle that rols every night the sun to its place where it can go up next morning it takes care of creation while it cant run by itself 11:49
audreyt oh! ok.
lichtkind so i stole enough attention its geat to see you still wrestling with pugs :) 11:51
11:51 CodeBook joined 11:53 ruz joined
audreyt :) 11:53
jferrero lichtkind, en.wikipedia.org/wiki/Kephir 11:56
CodeBook do the channels #perl6 and #perl have a direct relation?
lichtkind jferrero kefir is very healthy :) the only problem i had was that the transliteration "xeper " was used by satanists 11:57
Jedai CodeBook : #perl6 is for the development of pugs and modules related to perl6 for perl5, #perl is the channel where you can find help for your perl questions. 11:58
CodeBook Jedai, i see
Jedai, do both channels have the same channel management? 11:59
jferrero lichtkind, Only I have to translate the english.conf file to spanish?
lichtkind yes 12:01
CodeBook i tried to "/cs access #perl list" to be able to talk to an OP there about the reasons for a provider ban. cause i never been there so far. it might be offtopic here. but is there a chance to talk to a person who has access on #perl?
but the access list is restricted 12:02
lichtkind jferrero but still i need to add 2 lines of perl code and something in configs to be able to switch to spanish via language menu in the app :) 12:05
but that easy for me done in 2 minutes
jferrero In coding = line, I can set the coding to 8859-15? 12:07
audreyt TimToady: how lazy is a capture? 12:09
my $x = \(exit(1), exit(2));
it should exit even if $x is not usef, right?
jferrero lichtkind, In coding = line, I can set the coding to 8859-15? 12:10
azuroth_centeric what sort of laptops does everyone use? 12:11
12:12 elmex joined
StudiesInWords #perlhelp is the channel for help with perl questions. #perl is the social channel for the perl community. (imho) 12:15
azuroth_centeric aww, okay 12:16
gaal audreyt: surely! my $x = \(1+2)_ # you would expect this to be reduced too
likewise for a function call, IMHO 12:17
if you want Sclar::Defer or lazy, I can set you up with the person who wrote that...
lichtkind jferrero no its a planned feature, now its treated like normal perl strings 12:18
audreyt gaal: well then Cap can't contain Exp can it. 12:19
it has to contain Val.
for the PureCap.
gaal huh. 12:24
btw didn't you rename it to Capt? or was it Capt you'd renamed it from? :)
audreyt you renamed it from Capt to Cap iirc. 12:25
it's Cap at this moment
gaal NiCap
lazy{} and defer{} in pugs, how would they be implemented? cps voodoo? 12:27
audreyt VThunk 12:28
which would be in IO 12:29
12:32 macroron joined
audreyt Pugs.Val.Foo would be the API of the Foo builtin class, I think 12:35
I've pondered Pugs.Class.Foo and Pugs.Data.Foo instead
or even Pugs.Pure.Code 12:37
but Pugs.Val.Code seems to work well, so I'll put it there for now 12:38
gaal sorry for the lag
Pugs.Val.Code for the Code ASTish type you mean? or the Perl6-visible Code type? or are you intentionally coupling them? 12:39
audreyt the perl6 visible Code type
it would define the Haskell-land PureCode type 12:40
as well as its instances for the various interfaces
aka roles
which we may call Pugs.Interface.*
or Pugs.Role.*
or some such
the idea is to decouple the constructors from the implementations 12:41
for ValPure, a PureFoo is always just Purefoo
for ValPure, a PureFoo is always just PureFoo
the "data PureFoo ..." should not occur in the same module
anyway, you'll see pretty soon :) 12:42
gaal :)
audreyt meanwhile read that ralf paper :)
gaal $work :(
audreyt oy :/ 12:43
gaal i only show up here when my stomach won't let me do any more c++
language of whipped backs instead of whipuppededness 12:44
svnbot6 r11949 | audreyt++ | * Pugs.Val - sketch of the public-facing Val interface.
gaal that was fast
audreyt *nod* 12:45
12:45 jferrero joined 13:01 jferrero left, jferrero joined 13:03 kanru joined
audreyt bbl... 13:03
svnbot6 r11950 | audreyt++ | * Val: Add UProto for prototype objects 13:04
13:20 mauke_ joined 13:31 jkva joined 13:35 mauke_ is now known as mauke 13:38 ludan joined 13:48 Aankhen`` joined, penk joined 14:28 equinox5 joined
equinox5 hi 14:28
what is the state of Perl 6?
lichtkind whooo big question 14:30
webmind easy question, big answer :)
equinox5 come on
is there going to be a version for production use? 14:31
what backends are working best?
lichtkind of course :)
webmind I'm pretty sure about that there will be
lichtkind the language still evolving a bit but great peaces are finished
pugs is running today
but production .. not this year 14:32
equinox5 I have seen there are many backends.
which are in the best state?
lichtkind perl 5.10 will bring some cool perl6 features in "production " 14:33
equinox5 Such as?
lichtkind err operator or case statements
14:33 xdg joined
lichtkind say 14:34
and moose gives you perl6 OOP today i
pugs is cool but changing too fast for any sane production 14:35
and aprrot still incomplete
equinox5 ok
what about ponie?
lichtkind is sleeping currently 14:38
but translator seeme to awakenin
14:38 nitro44x joined
equinox5 what translator? perl 5 to 6? 14:38
lichtkind yea 14:39
equinox5 what is it called?
officially?
lichtkind its not even born but larry seems to work on it 14:40
audreyt it's Sage's work
based on larry's
misc/pX/Common/P5_to_P6_Translation/
in the pugs tree
lichtkind audreyt how does this work with the commiter bits? 14:41
lichtkind has still no for pugs
xdg lichtkind: wait about 20 secs
audreyt oh, easy
let me know of your email addr :)
lichtkind [email@hidden.address] 14:42
equinox5 I could also make a translator.
audreyt mm, fellow SMAC lover
equinox5 but I cannot see the point.
audreyt welcome aboard!
add yourself to AUTHORS :)
xdg lichtkind: see... wasn't that about 20 seconds?
:)
14:42 foo\ joined
lichtkind :) 14:42
audreyt simply commit as "lichtkind"; you already got an openfoundry account 14:43
14:43 nitro44x left
lichtkind and the pwd? 14:43
audreyt uhm, I suppose you picked a pwd when you signed up
there's a mail-passwd link
on rt.openfoundry.org/ somewhere 14:44
if that fails, I can give you another commit bit on another mail addr, or reset your password manually, etc :)
equinox5 what about the backends? 14:49
lichtkind no thanks audrey im in :)
who runs open foundry?
audreyt some .tw government branch 14:52
I did most of the initial implementaiton; gugod made the UI bearable; but the maintainence team has taken over quite a while ago 14:53
lichtkind cool 14:54
i like open foundry but there are much more folks on sourceforge 15:01
15:05 elmex joined
equinox5 what is your part in this project? 15:05
gaal with EU::MM, how do I override settings when I do 'perl Makefile.PL'? 15:06
eg. perl Makefile.PL cc=/some/alt/cc 15:07
audreyt try CC=
gaal looks like it worked. Thanks :) 15:09
15:17 buetow joined
audreyt gaal: I nuked all our constructors. whee 15:22
svnbot6 r11951 | audreyt++ | * "Pure", "Mut" and "Ext" (nee IO) are now open data types.
lichtkind equinox5 you mean me? 15:23
i just looking and asking questions? may be i manage to write an tutorial :) 15:25
15:26 crem joined
gaal audreyt++ audreyt++ audreyt++ # Mistress * 15:32
Ext as in external?
Worldly Types
nativeId (NBit True) = NInt (-1) 15:35
bogons detected!
audreyt yeah. I'm wondering if it actually pays
for .valid to be true whenever .defined is true 15:36
ext as in external.
gaal dynCompare x y = case cast y of -- interesting 15:37
ah, I understand
cool :)
audreyt :) 15:38
gaal you just made everything comparable
in three lines
Haskell**
15:43 cognominal joined
audreyt ok, that's it for tonight. be back afte a short sleep... 15:52
svnbot6 r11952 | audreyt++ | * valId: for Pure types, force the special pureId method; for Mut and Ext
r11952 | audreyt++ | we simply use memory address.
equinox5 ok, lichkind, 15:54
bye
15:56 ludan joined
gaal night A 16:13
16:18 Bit-Man joined 16:27 chris2 joined 16:40 ludan joined
gaal whoa whoa the open types trick is cool 16:48
big aha on page three :) 16:49
it's also fortunate that their example just happens to be an AST :-)
TimToady audreyt: actually, our current huffmanly solution is: $fh will leave { .close } 17:00
jmf is there any chance of the hyperoperator changing back to something simpler? 17:01
TimToady audreyt: yes, the normal boolean type name is Bool. bool would be a synonym for bit, which is already a synonym for uint1
audreyt: a sub call is never lazy by default, even in a capture. It will prefer to assume list context and run immediately. 17:02
jmf: not in standard Perl 6, but it would be pretty easy to tweak with a grammar mod. 17:03
some mathematicians would rather assume hyper and mark the ones that force scalar. 17:04
jmf i see: and is it supposed to be that >><<< is valid but >>!<<< is not
TimToady the latter is not yet implemented, as far as I know 17:05
but if I were using them, I'd certainly use the french quotes rather than the texas quotes.
just for visual clarity
jmf ahh ok thanks 17:06
17:07 buetow joined
TimToady Ā»<Ā« and Ā»!<Ā« are somewhat easier to look at, if a little harder to type. 17:08
but you read it a lot more times than you type it.
jmf yes
well not in the interpreter
but i see your point
SamB TimToady: I think that must depend on fonts... 17:09
it doesn't look very nice here at all
the Ā»Ā« are itsy-bitsy
and the centerlines don't match up either 17:10
pango needs work :-(
jmf TimToady: what would be the best way to start working on perl6 at this time? 17:14
TimToady depends on your interests
some folks like the low-level work, and so tend to head toward parrot. 17:15
others prefer to be closer to the abstractions, and tend to hang out in pugs-space.
right now, most of the test development is done under pugs 17:16
jmf well working on pugs then?
TimToady "pugs" really means two different things right now.
it started out being the Haskell compiler/interpreter for Perl 6.
and if you run "pugs" that's what you still get. 17:17
but the pugs repository has become home for many other related projects that may or may not involve Haskell.
the whole p6-on-p5 thing is pretty much devoid of Haskell these days.
but that's because Audrey is so good at handing out commit bits. 17:18
speaking of which, do you have one?
jmf no
TimToady gimme an email address and I can send you one. You can use "at" and such to disguise from bots.
jmf jmflizz _AT_ rice _._ edu 17:19
TimToady okay, had to remember where the page was, but an invite should be on your way. 17:23
jmf ok, thanks
TimToady customarily the first commit you do is add your name to AUTHORS 17:24
jmf so will p6-on-p5 and the haskell interpreter both have a place in the future then?
17:24 justatheory joined
TimToady everything that's useful will have a place in the future. :) 17:24
jmf of course :) 17:25
TimToady to be more specific, I don't see either of those withering away any time soon.
p6-on-p5 is probably the most expeditious way to bootstrap the community into using P6.
the Haskell implementation may well remain the implementation of choice for defining the exact semantics desired, and hence regressing the regression tests themselves. 17:26
the parrot implementation will likely be the fastest
jmf and how about p6-on-p6? 17:27
TimToady As soon as any other implementation can run p6 sufficiently well, we can do that too. there's already efforts in that direction. 17:28
I think most of the folks here would rather be writing a Perl compiler in P6 than in P5...
jmf right 17:29
TimToady or in PIR, or in Haskell...
well, I'm gonna need to bail now. have the appropriate amount of fun. :) 17:30
jmf alright, thanks for your help
TimToady bbl & 17:32
17:37 equinox5 joined
equinox5 hi 17:37
what is PIR RUN? 17:38
17:38 jferrero joined
equinox5 Does that mean that Perl 6 is converted into Perl 5 and run using perl.exe? 17:39
Khisanth so how many MB is pugs' AUTHORS up to? :) 17:40
17:42 miyagawa joined 17:48 robkinyon_ is now known as robkkinyon 17:49 robkkinyon is now known as robkinyon
robkinyon obra: ping 17:49
18:12 diotalevi joined 18:18 Limbic_Region joined 18:20 Aankhen`` joined 18:23 StudiesInWords joined, amnesiac joined 18:34 ruz joined, jmf left 18:44 marmic joined 19:00 justatheory joined 19:07 cm joined 19:08 Aankh|Clone joined 19:09 jmf joined
jmf does anyone know if there is a problem with the openfoundry.org registration page? 19:09
19:13 weinig joined
gaal haven't heard of that 19:13
jmf can?t open the page ?rt.openfoundry.org/Foundry/Home/Gue...ter.html?. The error was: ?lost network connection? (NSURLErrorDomain:-1005) 19:14
lambdabot Title: Member Login -- OSSF
gaal worksforme (though slowly) 19:15
jmf well i get that error after I click "I agree" 19:16
gaal oh
breaksforme too then. 19:17
jmf damn
gaal .tw people are probably asleep now...
got something to commit? 19:18
jmf i see, will this probably get fixed on its own in the next few days then or is there someone i should contact?
well no not yet
gaal I think both gugod and audreyt might be able to do something about it, or at least prod someone else to 19:19
I don't know who's the proper person to contact directly though.
jmf alright ill wait it out for now then, thank 19:20
gaal the repo does work though, so you can check out even if you don't have a user.
jmf yeah ive done that already
gaal (it's svn; it nees a user only when you commit)
jmf just wanted to make sure its not something on my end 19:21
19:23 Aankh|Clone is now known as Aankhen`` 19:38 niner joined 20:09 cm joined 20:15 jmf left 20:34 Fawaka joined 20:53 bernhard joined 20:54 weinig joined 21:03 weinig joined 21:07 larsen joined
lichtkind hej im checking out pugs for first time of my liiiiiiffeee 21:17
hehe :)
azuroth_centeric are you _drunk_? :D 21:18
lichtkind better ..happy :)
azuroth_centeric I have work in an hour. :-( 21:20
lichtkind thats no excuse not to be happy :) 21:24
azuroth_centeric I'll try my best, then 21:27
lichtkind love yourself, your great :)
i have problems to check out pugs
server closes connection 21:28
azuroth_centeric like, it gets half-way through and times out? 21:29
lichtkind yes
it stops always on pugs/"svn/vcc/default 21:33
azuroth_centeric yarrg, just a sec 21:36
21:36 lollan joined 21:37 azuroth_centeric joined 21:40 lollan_ joined 21:41 lollan_ left, dduncan joined
lichtkind azuroth_centeric++ 22:03
shii same problem 22:05
22:07 fglock joined
lichtkind please help me i cant commit pugs repository 22:08
22:09 jmf joined, TreyHarris joined
xinming lichtkind: what's the message show? 22:26
lichtkind same as reported
connection closed by server 22:27
xinming lichtkind: I think It might because the server is busy. 22:28
lichtkind: what version control system you use? 22:31
BTW
lichtkind tortoise 22:34
but its not that important the important dir i have now down :) 22:35
23:03 jferrero joined
lichtkind wb 23:03
xinming you was right now i could load all stuff
23:12 reiner_ joined
svnbot6 r11953 | fglock++ | pX/Pugs-Compiler-* - fixed aliasing in Token; '$foo if $bar' works again 23:12
clkao fglock: is it all passing now!
whoot
fglock: did you see my question the other day about removing an infix or stuff. if we leave the scope
fglock clkao: I'm running 08-test.t - but it takes a lot of time here (too little ram) 23:13
clkao syntax error at /home/clkao/work/pugs/misc/pX/Common/Pugs-Compiler-Rule/blib/lib/Pugs/Grammar/Rule/Rule.pm line 17, near "rule ws :"
make clean did the trick
fglock: Use of uninitialized value in subroutine entry at (eval 345) line 441. 23:16
Error in perl 6 parser: Can't use string ("") as a subroutine ref while "strict refs" in use at (eval 345) line 441.
23:17 bcorn_ joined
fglock clkao: we may need to localize the category tables, at both compile time and runtime 23:20
clkao: I had to interrupt 08-test.t, it swaps too much 23:21
clkao fglock: it simply fails early for me
fglock clkao: you may uncomment Pugs::Compiler::Perl6 line 55, to show the current compiling line number 23:23
clkao it's not even there
fglock Perl6Prelude compiles, I'll retry 08-test.t 23:27
clkao fglock: it works 23:38
fglock: it seems to use significant more cpu. ram usage is okay though 23:39
fglock it may be backtracking somewhere 23:41
clkao it all seems beautiful up to now
do you wnt ot merge back? 23:42
-rule is with lots of conflicts because there were lots of manual merging
-perl6 is sane, i can just merge up in a jiffy
fglock clkao: would you help? you have more experience with merging then I 23:47
robkinyon obra: ping 23:48
clkao fglock: i can help with -perl6. can't with -rule, i took a look the other day 23:49
fglock clkao: maybe we could just copy -Rule over
clkao no
i think i have changes in perl5/-rule
maybe manually merge things on perl5/-rule down to px? 23:50
fglock oh, there is scw work
clkao should i just do the perl6 and leave the rest for you? i neeed to go to bed
robkinyon seen obra 23:51
jabbot robkinyon: obra was seen 2 days 5 hours 3 minutes 2 seconds ago
robkinyon anyone here using a MacBook and that DesktopManager replacement?
fglock clkao: how about merge perl6 down to px, so that it doesn't break the perl5/ version; i'll merge -Rule
robkinyon stevan said that obra pointed him to a good replacement (a fork that was better on Intel macs)
fglock clkao: would do the -Perl6 merge? 23:53
svnbot6 r11954 | clkao++ | copy test.pm properly.
obra virtue
fglock I'll work on -Rule then
clkao done 23:54
fglock clkao: thanks! 23:56
svnbot6 r11955 | clkao++ | Remove rule related tests for now, we shuold be using the perl6 version of p6regex.t 23:59

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!