6.2.12 released! | paste: sial.org/pbot/perl6 | pugs.blogs.com | pugscode.org | pugs.kwiki.org
Set by audreyt on 26 June 2006.
fglock for/if/sub/method/class/... 00:00
nothingmuch okay
ingy hola
nothingmuch that sounds pretty consistent with what I have
ingy hi audreyt
nothingmuch hi ingy
fglock ... block/expression
ingy hi nuffin 00:01
fglock hi ingy
clkao ingy: you've done the generator
?
nothingmuch ingy: clkao was pointing me at ACME::Goto::Line but he knows NOTHING 00:02
RIGHT CLKAO?!
clkao_ ya
00:02 bsb joined
fglock and these nodes can have other nodes inside (that's it, mostly) - sub declarations are a bit complex because of the attributes, etc 00:02
nothingmuch hola bsb
bsb hey nothingmuch
lumi Hi bsb
bsb hi lumi 00:03
ingy hi fglock
fglock hi bsb 00:04
hi lumi
nothingmuch hi fglock
hi lumi
hi ingy
ingy hi clkao 00:07
fglock so there is nothingmuch in the v6.pm syntax tree 00:09
00:10 frederico joined
nothingmuch fglock: i'm going to dive into v6 00:14
give me 10 mins
but a friend of mine is having a nervous breakdown on MSN, and i'm on the receiving end, so my latency will be bad
ingy haha
fglock ok - maybe a possible path is progressive migration
k
ingy is that friend me?
nothingmuch fglock: yes, i'm for it 00:17
ingy: no
you are just having a code breakdown
audreyt nothingmuch++
00:19 hikozaemon joined
clkao nothingmuch: you can dive to v6 with me as well 00:21
nothingmuch clkao: okay
i'll tak e you up on that once i'm done having my conversation 00:22
00:27 tup joined 00:35 ludan joined, stevan joined
fglock clkao: svk is working now 00:37
ludan hi
stevan good-m??se all
ludan clkao: how are you? do you remember me? we met in Pisa
ingy hi stevan
stevan hullo ingy
stevan is very tired from 12 hours of travel,.. half of which was improvised
clkao_ fglock: you might want to have a mirror for the last 100 revisions only 00:38
clkao ludan!!
ludan :D
clkao: how is going?
clkao ludan: ok, hackathon'ing 00:39
ingy audreyt: ping
ludan clkao: i'm in the netherlands now
clkao oh
ludan are you always in London?
audreyt pong 00:40
ingy will add .xxxc support to M::C on the plane
audreyt: did you get my msg?
ludan i've to go to sleep now...read you tomorrow guys 00:41
bye
00:42 ludan left
fglock ingy: can M::C touch the pmc file when Makefile.PL runs, so that we can distribute precompiled pmc? 00:43
audreyt that's what pmc_support does in M::I
ingy fglock: it does
fglock: see orz.pm 00:44
fglock ok!
ingy catches flight
gaal audreyt: is Noop and Exp or a Stmt? (or both?) 00:48
audreyt Exp
anything that can happen at rhs of assignment
is exp
gaal so ($x, undef, $z) = @blurp has a noop in the tree? 00:49
audreyt it's illegal
($x, *, $z) = @blurp
is it
gaal ok
nothingmuch clkao: ping-ish after zev 00:50
gaal hm. where does old Exp's Prim go now? 00:51
and parse error (NonTerm), can that stay Exp? 00:52
audreyt NonTerm is no longer used. 00:53
Prim... is interesting
thinking
probably still needs to be there 00:54
probably
clkao nothingmuch: after i cool down my laptop and myself 00:56
01:00 ajs_home joined
gaal is label really Indetifier or can it be Maybe Identifier? 01:02
audreyt an empty label? 01:03
Set Identifier ? 01:04
Maybe Identifier?
I think Maybe is it. maybe.
gaal Set makes no sense to me? But the other two both work. since we don't patmatch against it a lot I thought Maybe can make more formal sense 01:05
.oO( Indentifier? Attack of the tab keys )
01:06
audreyt Identifier means we'll use "null" as Nothing
formal is fine. let's do that.
gaal k 01:07
svnbot6 r11080 | fglock++ | px/fglock/v6.pm - fixed t\03-equal.t 01:10
audreyt whew
audreyt is done with Jifty for the day
however it's also rather late now :/
audreyt decides to 1)shower 2)sleep and resume pugshacking tomorrow 01:11
obra sounds sane 01:12
though it's not That late
;)
rw-rw-r-- pugshacking sounds really dirty out of context... 01:16
audreyt true. 01:20
01:21 mdiep__ joined, scook0 joined
audreyt fglock: try your grant again at TPF? 01:21
fglock audreyt: sure 01:22
01:27 weinig is now known as weinig|sleep
nothingmuch clkao: ping 01:48
clkao nothingmuch: pong 01:49
nothingmuch v6
??
meh
i'm not that restless, i just can't type
fglock uh? 01:50
01:51 amnesiac joined
nothingmuch fglock: hu 01:51
gaal audreyt: we forgot Param traits (is rw etc.). data ParamTrait = TraitRW | TraitCopy | {- other common ones -} | TraitSimple Identifier | TraitArg Identifier [Value] 01:56
?
nothingmuch looks at clkao in an insulted manner
nothingmuch begs for attention
gaal looks at nothingmuch in a mooselike manner 01:57
audreyt gaal: sounds good.
gaal elkcellent. 01:59
bsb Is sprintf still around in p6? or has it been rethunk somehow? 02:02
gaal hmm. likewise code traits, but should those be on a SimpleCode or on named code nodes? 02:03
hey bsb. I think it's still in but you can also do $smth.as()
bsb I've been having to look at Python, and they have % as a sprintf like operator 02:05
audreyt bsb: it's .as()
but sprintf is still there
bsb and they have named params to formatting
"%(name)02s" or something 02:06
thought there might be a capture <-> sprintf synergy
which Syn has .as()?
.. s02 02:07
gaal do subroutine PRE and POST blocks receive arguments? 02:27
(I'm presuming they do not) 02:28
TreyHarris gaal: they had better, if they're going to fulfill the DBC stuff that I thought was their raison d'etre. 02:30
if you're talking about can they see the subroutine's arguments
gaal hm. so implicitly same signaure as the do block? 02:31
(the "do" block is the main sub part) 02:32
TreyHarris yes, i would assume so. in DBC, PRE is supposed to be used to set preconditions that must be met for the sub to work correctly, and POST expresses postconditions that must be met for the sub to successfully complete. they may both be keyed to the arguments, but they shouldn't get anything apart from what the do block gets
(the idea being that in DBC you don't have to read the subroutine to ensure you're meeting the API, just look at the pres and posts. it's also assumed that in production, you may turn pres and posts off for performance, they're not supposed to have side-effects.) 02:33
gaal ...which leads me to wonder whether they are not represented internally as traits at all but rather as fields in a Code data type 02:34
TreyHarris: yes, you're making sense to me
TreyHarris gaal: is that a polite way of saying, "i've got it already, now shut up"? :-)
gaal well, if I were *polite* I'd say thanks 02:35
...thanks!
TreyHarris hee
the only weird thing about DBC preconditions and postconditions is that both are stacked from most derived to least derived method, but pre's are OR'd together, and posts are AND'd together. the idea being that if the precondition of Dog::praise is "have food ready", the precondition of Labrador::eat must allow you to obtain simply by having food ready, but it might also let you "have a hand ready to pet with" 02:38
in other words, in pure DBC, while you may decide whether a subclassed method calls its supermethod or not, you cannot choose to avoid calling your superclassed pre- and postconditions. dunno if Larry wants to relax that requirement though. 02:40
gaal should precendence and associativity be represented just as Yet Another Typed Trait on code? I'm not sure if traits should be [CodeTrait] or a map; some seem to accept multiple values, some none.
TreyHarris gaal: i'm not sure i understand the implications of either choice. 02:41
gaal the first gives flexibility and possibly poor performance and somewhat clumsy code. 02:42
02:42 reZo joined
gaal background, we're refactoring the pugs AST. 02:42
my question was primarily aimed at audreyt but apparently she's sleeping already 02:43
(though I haven't seen evidence to that effect) 02:44
TreyHarris she just sent mail to p6-l 02:45
so jump up and down and wave, she might see you :-) 02:46
gaal Ah, so there is evidence she's asleep. There you go :) 02:47
TreyHarris ah, coding is her waking activity and emailing is her sleeping one?
audreyt gaal: prec is best stored as a rational number 02:48
fixity is enumerated
gaal not exactly... but we're used to seeing her after she sleeps.
hi audreyt
audreyt hi, and btw I'm sleeping :) 02:49
miyagawa lol
audreyt gaal: Code should admit fixity and prec and explicit fields
gaal audreyt: but how is the traitbad stored?
audreyt the idea is all typed info we get on compile time
we use distinct fields
to make casing easier
and all the usertags we just lump together in a set or map 02:50
gaal s/d/g/
audreyt I'd say just a Map really
Map Identifier Val
Map Ident Val
gaal (renamed to Ident)
hah
nothingmuch .'ןמ13 02:51
audreyt dreams that she is "audreyt" on freenode and IRC'ing on #perl6
gaal nothingmuch: מוה
audreyt kinda like Zhuang-Zhou and the butterfly
nothingmuch מוה
obra audreyt: can you also dream that p6 is done 02:52
?
02:52 robkinyon joined
gaal is getting kinda sleepy too... more hacking tomorrow I guess 02:53
gaal dreams of lucid dreams
obra tomorrow, we'll hack from $office 02:54
03:00 lisppaste3 joined
audreyt obra: p6 will be done when I wake up from it :) 03:00
svnbot6 r11081 | cmarcelo++ | * HsJudy minor tweaks: pjerr don't need to be IO type, new 03:19
r11081 | cmarcelo++ | test program.
obra audreyt: :) 03:21
audreyt gaal: can I safely turn off my computer? 03:22
clkao_ audreyt: ;) 03:25
gaal audreyt: I'll log out 03:32
done
audreyt :) & 03:36
svnbot6 r11082 | fglock++ | px/fglock/v6 - passes 15/22 test files 03:59
clkao_ fglock++ 04:02
04:26 fglock left
nothingmuch kolibrie: i screwed up and only remember the picture after we turned in our keyes 04:36
nothingmuch--
clkao_ please do not karma yourself 04:50
04:50 xern joined 04:52 shachaf joined 05:41 nothingmuch joined 05:43 nothingmuch joined 05:48 DebolazX joined 06:11 Aankhen`` joined
svnbot6 r11084 | cmarcelo++ | r11098@roma: cmarcelo | 2006-07-03 03:05:46 -0300 06:11
r11084 | cmarcelo++ | * HsJudy: build elems, keys, toList and map on top of a single
r11084 | cmarcelo++ | map_ function which properly iterates in the Judy type.
06:12 iblechbot joined 06:15 cmarcelo joined
cmarcelo how do you people push changes from a local branch to a mirror in svk? I did with "svk push" and was expecting $EDITOR open up but it didnt.. =P 06:20
06:30 Entonian joined
bsb cmarcelo: there is a way, I forget it. see: svk help sm 06:32
clkao svk sm -fl 06:33
cmarcelo tks 06:34
ingy hola 07:03
audreyt: I put a new Module::Compile on CPAN 07:16
audreyt: and a new Compile::Generators
audreyt: the M::C handles non-module files
audreyt: so adjust v6.pm accordingly 07:17
audreyt: I can try to adjust it, but pugs was failing to build last time I tried
07:28 elmex joined 07:41 shachaf joined 07:45 cmarcelo left 07:54 kanru joined
ingy give up on making Pugs::Compile::Perl6 pass tests 07:57
08:04 drrho joined 08:42 mdiep___ joined 08:44 shachaf joined 08:51 jiing joined 09:09 kanru joined 09:24 shachaf_ joined 09:26 buetow joined 09:42 froh-doh joined
meppl guten morgen 09:48
10:15 Aankh|Clone joined 10:30 rashakil joined 11:07 mj41_ joined 11:12 Debolaz joined 11:17 shachaf joined 11:29 chris2 joined 11:37 szbalint joined
audreyt ingy: you don't have to build pugs 11:45
ingy: just adjust in perl5/Pugs-Compiler-Perl6
nothingmuch morning 11:46
11:47 Bleadof joined
audreyt my room's AC drops to 70F automagically at night 11:48
so I was frozen awake again (for 6 days consecutively)
trying to get some more sleep now
nothingmuch audreyt: i woke up from cold + nightmare ( 11:51
Data::Bind::Compile(
audreyt v6 already do D::B::C
wouldn't hurt to abstract it out a bit into its own layer though
nothingmuch if you have nothing to cook yousrelf with my sleeping bag made sleeping very hard for me in the last week
audreyt that way Perl6::Subs can use it
nothingmuch i guess Per6::Subs should M::C and use D::B 11:52
11:52 shachaf joined
audreyt chip gave me P6::Subs comaint 11:52
so I'm thinking about releasing it as a Module::Compile subclass
nothingmuch okay =)
(re sleeping bag: it was too warm) 11:53
audreyt too /warm/? 11:54
nothingmuch yes
that was my problem despite crazy AC
audreyt is green with envy and purple with frozeness
nothingmuch you can prolly use it as a blanket 11:55
i can bring it up for you
audreyt that sounds good
&
nothingmuch waits for everyone to wake up 11:57
12:05 mauke_ joined 12:06 shachaf_ joined 12:14 Limbic_Region joined 12:20 mauke_ is now known as mauke
gaal meows 12:31
12:37 weinig|sleep is now known as weinig
visq gaal: hello ! you et.al. started refactoring the Haskell implementation ? 12:43
12:48 vytautas joined
gaal visq: yes, but the first commit won't be for a while 12:48
visq I browsed the Haskell source for a while (ghci rocks) 12:49
gaal the refactoring target is not everything in the implementation -- just the AST, but that does imply some futher changes.
12:49 nova-alpha joined
gaal yes it does 12:49
visq The implementation of Numeric.hs seems to need some changes, though 12:50
gaal also ctags
12:50 nova-alpha is now known as alpha
visq Integer.isa(Rat).isa(Num) for example; especially coercion in op2Numeric 12:50
12:50 alpha is now known as nova-alpha
visq s/Num/Double/ ! 12:50
gaal commits welcome if you can figure out a better scheme :) 12:53
gaal goes to look for coffee
12:54 froh-doh joined
visq Hardcoding coercion to the best of VInt,VRat,VNum in the context of primitive numeric ops isn't that hard 12:57
12:58 xdg joined 13:03 frodo72 joined 13:04 frodo72 is now known as froh-doh
nothingmuch gaal++ 13:05
cpanratings.perl.org/dist/Best
lambdabot Title: "Best reviews -"
gaal
.oO( A site dedicated for reviews of my module )
13:06
13:06 elmex joined
gaal The license incidentally should be MIT, but MM doesn't know how to eat that :( 13:06
visq: if you know what to do, go for it 13:08
visq gaal: btw: I fixed that catalog_tests perl script to run with windows path seperators some time ago - not much fun though ;)
gaal visq++ visq++ visq++ 13:10
visq hehe, for setting up windows ...
gaal indeed. BTW, have you added yourself to AUTHORS? please do 13:11
visq not yet; I will 13:12
svnbot6 r11085 | benedikth++ | AUTHORS: add myself to list 13:17
13:23 cognominal joined 13:25 neoesque joined 13:28 iblechbot joined 13:35 vel joined 13:45 frederico joined 13:59 cdpruden joined
spinclad where's the hackathon? 14:00
14:03 fglock joined 14:13 Aankh|Clone is now known as Aankhen``
clkao_ nothingmuch: data::bid::compmile was your nightmare? 14:13
14:14 shachaf_ joined 14:19 shachaf__ joined
gaal clkao_: I can attest that nothingmuch woke up saying something about Data::Bind 14:20
I think it was that he figured out how to do aliases
nothingmuch clkao_: yes =) 14:22
clkao don't read data::bind code before sleep 14:24
14:24 lisppaste3 joined
gaal audreyt: do we need our new PadEntry to still have the Bool freshness flag? 14:26
fglock does '$a.m1.m2' parse as '$a.m1;$a.m2' or '($a.m1).m2' ? 14:27
gaal the latter, I hope
spinclad the latter
fglock thanks
14:30 shachaf__ is now known as shachaf 14:37 diotalevi left 14:45 justatheory joined 14:53 sahadev joined 14:54 cognominal joined 14:55 shachaf_ joined 15:01 shachaf__ joined 15:07 shachaf joined 15:15 pmurias joined
svnbot6 r11086 | fglock++ | px/fglock/v6 - t/07-isa.t passes 2 tests 15:15
pmurias hi
Limbic_Region seen pmichaud 15:16
jabbot Limbic_Region: pmichaud was seen 45 days 23 hours 30 minutes 43 seconds ago
Limbic_Region hrm
pmurias is p6ast.hs supposed to be valid haskell?
Limbic_Region fglock do you know what, if anything, is the successor to Perl6::Rules for rules in p5?
s/fglock/fglock and/or anyone/
pmurias PCR
fglock yup 15:17
Pugs::Compiler::Rule
Limbic_Region ok, forgive me for sounding like the neophyte that I am but does the Pugs in the namespace imply I need pugs to generate the p5? 15:18
pmurias no
it's pure p5 15:19
maybe with PadWalker
fglock I think it means "Pugs-in-p5" as in "Pugs-in-Haskell"
15:20 cognominal joined
Limbic_Region no worries, I am going to pose a question to the list and wanted to be as focused as I could 15:20
thanks
fglock you can ask here too :)
pmurias is trying to use p6ast.hs for his haskell and TaPL practice 15:21
should i fix syntax errors in p6ast.hs? (1 found)
Limbic_Region fglock - well, you likely have more important things to do then to come up with an example of using P::C::R to parse (and evaluate) mathematical expressions 15:22
fglock Limbic_Region: actually, Pugs::Compile::Perl6 could do that out of the box 15:23
s/Compile/Compiler/
math expressions are parsed using a precedence parser instead
but Pugs::Compiler::Rule grammar actually use rules instead - it just takes more work 15:24
Limbic_Region fglock - feel free to tell me that I am a nimrod then in my post to the list 15:25
I still want to post it so as to potentially benefit others
fglock Limbic_Region: see svn.openfoundry.org/pugs/perl5/Pugs...le/Rule.pm - the last rules have looser precedence; the rules at the start have tighter precedence 15:26
ok
Limbic_Region fglock - sent 15:29
15:35 robkinyon joined
svnbot6 r11087 | fglock++ | px/fglock/v6 - 01-pod.t no longer needs eval() 15:35
Limbic_Region out of curiosity, did that message hit the list yet? 15:39
fglock yes 15:40
Limbic_Region I can never remember if a sent message comes back to me or not
ok - thanks
svnbot6 r11088 | fglock++ | px/fglock/v6 - pass t/07-tailcall.t 15:45
fglock bbiab & 15:48
Limbic_Region: search.cpan.org/~fglock/Pugs-Compil...cedence.pm 15:56
lambdabot Title: "Pugs::Grammar::Precedence - Engine for Perl 6 Rule operator precedence - search. ..."
Limbic_Region fglock - that doesn't help with the evaluation part 15:57
but thanks
15:57 elmex joined
fglock svn.openfoundry.org/pugs/perl5/Pugs...grammar.pl - doesn't do what you want, but you can see what the syntax looks like for evaluating rules in p5 15:59
it's a parser emitter written in rules 16:00
s/parser emitter/parser plus emitter/
lunch & 16:01
Limbic_Region fglock - I am off to lunch - would appreciate it if you replied to the email instead of just here (again - for the benefit of others) 16:02
lunch &
gaal mooseōæ½x9Cלkkkkk~. 16:04
16:11 shachaf_ joined
ingy audreyt: PCP6 was not passing tests 16:11
audreyt: so I didn't adjust
fglock Limbic_Region: sent 16:12
lunch for real &
clkao ingy: why are you playuing with that? 16:13
fglock++ # almost working
16:13 nothingmuch joined
spinclad pmurias: p6ast.hs looks to me like brainstorming -- a mix of example P6 syntax and the haskell data statements to support it 16:14
16:14 shachaf_ is now known as shachaf 16:16 shachaf_ joined 16:17 vel joined
gaal pmurias, spinclad: yes, it was a desifn sketch. the .hs name is just because we're lazy and don't want to type :setf haskell :) 16:18
spinclad thanks, my eyes can uncross now 16:19
ingy clkao: I moved the file execution code into M::C 16:20
hi fglock
rgs hi ingy
ingy hi rgs 16:21
rgs ingy: I've sent a patch for YAML via RT. 16:22
ingy rgs: thanks
I'll get to it in a few
pmurias gaal: i so the emmbeded P6 syntax,
rgs woot.
ingy I started writing a new YAML.pm 16:23
YAML3000.pm
rgs from scratch ?
ingy it's a port of PyYaml
so it will work perfectly
pmurias i wanted to know if haskell was used as a notation, or is the code intented to be also used as the haskell representation of the ast
kolibrie nothingmuch: that's okay about the photo. Exactly what happened to me. Had my camera out, didn't remember till I had turned in my keys.
nothingmuch kolibrie: =(
ingy because PyYaml is teh shizzle 16:24
kolibrie Maybe I can turn the picture of it in my brain into an oil painting
nothingmuch kolibrie++ 16:25
pmurias ingy: are you secretly a pythoner :)
svnbot6 r11089 | gaal++ | * set ignore property on spec docs, as they are slurped
r11089 | gaal++ | from an upstream repo
pmurias or pythonist 16:26
gaal pmurias: the .hs there is almost verbatim what we are using.
pmurias gaal: is the new ast actually used at the moment? 16:27
gaal one major difference from the existing ast is that pads don't live in Env any more
pmurias: we're moving to it, hopefully by the end of this hackathon
spinclad this hackathon has adjourned to Boston by this point? or in Chicago still? 16:29
gaal BOS
spinclad may i ask where? as i was planning on taking some sort of part... 16:30
gaal spinclad: 408 highland av, #416, 2nd floor, somerville 16:31
best practival office
s/v/c/
merlyn where can I read more about this mixed PGE that can parse top-down mixed with bottom-up? 16:32
I know there was the session last week, but I was still hung over. :)
so I didn't take very good notes 16:33
integral PM posted some yapc slides to parrot-porters
spinclad gaal: thanks, i'll make my way there
merlyn that'd help... I'll go look
pmurias integral: thats p6i? 16:34
integral pmurias: yeah
merlyn I don't see that in the past 50 messages
pmurias neither do i
clkao fglock: it seems that your branch is emitting perl5 up to where it can still parse the file ;) 16:36
integral hrm, well it was at least one of the lists :-/ 16:37
merlyn Yeah, if you can be more specific. :)
TimToady merlyn: I'm thinking about renaming ==> and <== to "prefeed" and "postfeed", with "feed" the generic term and "fed" the modifier on the target: "This is a prefed function."
integral subject, "Pm's YAPC::NA talk online", date: Tue, 27 Jun 2006 16:41:51 -0500 16:38
TimToady I think pipe indicates liquids, and a feed can be either solids or liquids.
merlyn Yeah, that's closer
TimToady and it's a verb.
merlyn the noun/verb-ing of pipe was a treat.
feed is nearly the same 16:39
gaal spinclad: cool
merlyn but prefeed doesn't imply a direction for me instantly
TimToady and not heavily overloaded yet in CS.
it's pre as in prefix
merlyn but what prefixes what?
TimToady considering the entire feed chain to be a prefix operator
merlyn ok
TimToady it's prefix to the fed thing
merlyn it'll take some getting used to.
ok
theorbtwo Erm, is that the same result as saying that pre-feed means the arrow goes on the beginning of it, as in <==? 16:40
TimToady no, that's a postfeed, specified after the fed func
and yes, it'd take some getting used to. still thinking about it. 16:41
don't much care about .uniq/.unique
theorbtwo I think I'd prefer "leftward" and "rightward", pointing out the direction of data flow, and the direction the arrow points.
TimToady not frequent enough to matter.
Yaakov Hey, it's Larry. Hello! 16:42
theorbtwo @foo ==> bar() is a rightward feed.
lambdabot Maybe you meant: faq todo yow
TimToady right/left feed would also work.
except that a right feed happens on the left of the fed func
so the "ward" is not superfluous
which makes it clunky.
feedinto/feedfrom maybe 16:43
Yaakov: howdy do.
Yaakov Pretty darned well, thanks for aasking. 16:44
16:44 robkinyon_ joined
theorbtwo Then you'll end up pronouncing things in the wrong order, with the preposition at the end of a prepositional phrase. 16:44
TimToady feed/feedback :) 16:45
spinclad @a feedsinto &f; &f feedsfrom @a 16:46
lambdabot Maybe you meant: all-dicts arr ask . v
TimToady audreyt: @?LABEL was intended to be all the labels visible outward in the dynamic scope.
::?CLASS is probably redundant with $?CLASS now 16:47
theorbtwo Hm, I guess that works.
TimToady perhaps with occasional circulocutions: ::($?CLASS)::($yow)
merlyn I saw something that showed expr ==> *; expr ==> *; expr ==> *; ... ***
TimToady they might just turn into feeds/fed
merlyn is there something that gets at the current list of built-up "whatevers"
in the interest of the debugger and other introspective things 16:48
TimToady @firstlist feeds &foo fed @secondlist
spinclad feeds/fedby ?
lambdabot Unknown command, try @list
merlyn like if I have code between the second and third of rhose
can I ask something if we have two whatevers so far
TimToady if bound to *@@args, yes.
spinclad & 16:49
merlyn ok, so there's some introspection available for the state of the world. good
TimToady each elem of @@args would be one "dimension"
but *@args means you want them flattened
(lazily, of course)
$BL ~~ s/circulocutions/circumlocutions/ 16:51
< merlyn> I saw something that showed expr ==> *; expr ==> *; expr ==> *; ... *** 16:54
I'd say that syntax is probably a placeholder for something else that we haven't thought of yet. 16:55
the problem it's trying to solve is foo ==> bar ==> baz
is trying to feed foo to bar, not foo and bar to baz.
merlyn yes, I see that. but if there's a place for a statement break, there should be some sort of introspection for "our story thus far..." 16:56
TimToady so we need some way to juggle the pointy ends of pipes and attach multiple of them at the correct spot.
might just rely on named arrays for that though.
ingy hi TimToady
TimToady named arrays would solve the story thus far problem.
ingy+- 16:57
:)
ingy: are you in Boston now?
obra ingy couldn't come to boston :/ 16:58
TimToady ah well. me too.
merlyn that's what happens when you have a warrant out for you. :)
16:59 nothingmuch joined
ingy I'm back in Seattle 16:59
hi nothingmuch
TimToady I prefer to specialize in unwarranted behavior.
merlyn o/~ back in... seattle, again... o/~
theorbtwo wonders if a Canada joke would be out of line. 17:00
TimToady the only straight line in Canada is the 49th parallel. :)
ingy I implemented lightweight generators in Perl using Module::Compile closures and goto
sun you can yield/next like in python
nothingmuch disagrees
TimToady nothingmuch is disagreeable. 17:01
ingy nothingmuch owes me
nothingmuch no i don't
ingy you break your promises?
nothingmuch you never listened =)
TimToady he's breaking is parameters with an argument.
*his
ingy: I think I know what a "goto" is, but what is a "Module::Compile closure"? 17:03
rgs commas aren't optional 17:04
ingy sorry... M::C, closures, goto
TimToady eats, shoots, and leaves...
ingy right
bad panda, no leaf 17:05
fglock clkao: re parse the file - sure - and we'll get precompiling for free
ingy hi fglock
fglock ingy: hi! 17:07
17:11 robkinyon_ is now known as robkinyon
fglock after a bareword space in an expression, everything else is slurp as parameters? 17:12
bareword plus space 17:13
like: ( name 1, 2 or 3 ) - is it parsed as name(1,2 or 3) or (name(1,2) or 3) 17:15
bbiab & 17:16
PerlJam fglock: are you assuming that perl hasn't yet seen the definition of "name"? 17:18
TimToady it will be taken provisionally as a listop, with listop precedence 17:19
so name(1,2) or 3
but it will fail compilation if name is not supplied by CHECK time.
it will also fail if name is declared as a unary or 0-ary func. 17:20
fglock TimToady: thanks 17:24
TimToady I guess the stuff that comes through a feed is "fodder". 17:30
gaal audreyt: can objects still have Dynamic opaque content or is it all slots now? 17:42
Limbic_Region speaking of feed and fodder - anyone know what I ate that is making me soooo sleepy
TimToady fodder? 17:43
gaal chocolate coated chocolate moose?
TimToady did you eat anything with tryptophans? 17:44
like, say, cow?
Limbic_Region apparently, that's an urban legend TimToady
Limbic_Region just looked for that very reason
www.snopes.com/food/ingredient/turkey.asp 17:45
lambdabot Title: "Urban Legends Reference Pages: The Big Sleep"
TimToady don't care if it's an UL, cow makes me sleepy...
gaal maybe you're just bored with the taste?
robkinyon meat in general is very heavy which can cause sleepiness
Limbic_Region it does say that eating solid foods (and likely some more than others) causes an increased blood flow to the abdominal region for digestion which might have something to do with it 17:46
robkinyon it's much harder to digest meat than vegetable matter
TimToady I think the belief that urban legends sites speak the truth is an urban legend. :)
Limbic_Region heh
in any case - my brain has decided to observe tomorrow's holiday prematurely and as such - the parsing experiment is on hold until Wednesday 17:47
thanks for the pointer fglock though I would still prefer a working example
TimToady of Perl 6?
Limbic_Region well - specifically Perl 6 rules 17:48
and if at all possible, without the use of Parrot or Pugs 17:49
but the latter is not really a requirement since I have both working quite nicely at work and at home 17:50
ingy TimToady: check out rafb.net/paste/results/BJZAbm91.txt (re: generators) 17:51
nothingmuch i still don't know how your goto works and mine doesn't 17:56
in ters of making the while recheck the condition
ingy nothingmuch: huh 18:02
gaal audreyt: where are objectIds registered? 18:04
some uplexy pad? 18:05
ingy nothingmuch: perl likes me 18:06
nothingmuch ingy: okay 18:07
rafl Which version of ghc is needed to build 6.2.12? 18:09
gaal I think 6.4.1 is still ok
on debian, which I'm assuming you're asking aobut :) 18:10
audreyt gaal: objspace
rafl So my 6.4.2 should work fine. Nevertheless I get some errors which seem to indicate some ghc or ghc-pkg b0rkage. 18:11
gaal: Of course it's about Debian. :-)
gaal audreyt: but is that global? i thought there was the possibility of masking packages
or do they nest?
rafl Ideas? nopaste.debianforum.de/3599 18:12
gaal anymoose mkType "foo" no longer makes sense
audreyt gaal: well, PIL^N models it as
type ObjectSpace = SeqOf (Weak NativeObj)
fglock audreyt: can you help me a bit with the parser? 18:13
ingy hi audreyt
18:14 ruoso joined
gaal is Data.Seq version dependent? I don't see it on hoogle 18:14
ingy fglock: why are there 2 v6.pm-s? 18:15
gaal oh, got it, src/Data/Seq.hs
ingy I need to patch one... which one?
or both?
audreyt fglock: sorry, I just woke up, gotta get some food first 18:16
and then I'll join the hackathoners
gaal audreyt: we were just discussing getting food
fglock ingy: perl5/Pugs-Compiler-Perl6/lib/v6.pm is the right one
Limbic_Region suggests audreyt avoid the fodder
gaal so if you aren't bent on breakfast-type chow you can join us here and eat
fglock ingy: the one in misc/px/ is a prototype for next version - I'll merge later 18:17
audreyt gaal: heh, I just reheated the rice form yesterday...
pasteling "rafl" at 217.246.159.40 pasted "Pugs 6.2.12 build failure on Debian" (363 lines, 31.2K) at sial.org/pbot/18189 18:18
ingy fglock: ok
fglock now I'm hungry - I'll get something to eat too &
audreyt gaal: so I'll finish it here I guess
ingy fglock: what is the basic changes?
audreyt: is PCP supposed to pass tests? 18:19
fglock ingy: the grammar is all new
ingy fglock: oh, so not changing v6.pm
I see
ll 18:20
erp
fglock ingy: yes - I'm rewriting the grammar in order to try to merge with Parrot/P6 later
ingy I mean the v6.pm wrapper is not changing
fglock: btw, your code seems to have DOS line endings :(
fglock no
audreyt yeah, please someone run utils/add-svn-props.sh over all fglock code 18:21
:)
fglock ingy: sorry
I use windows in the weekends
audreyt that's fine, that's what add-svn-props.sh is for 18:22
(to normalize line endings so DOS people see crlf and UNIX people sees lf)
fglock cofee & 18:23
18:28 rashakil joined
rafl Does third-party/fps somehow differ from www.cse.unsw.edu.au/~dons/fps.html? 18:34
lambdabot Title: "Data.ByteString / FPS"
fglock re parser: is it better to have the dot in a method call parsed as an operator, or is the dot a part of the '$object.method(param)' term?
audreyt rafl: it's a snapshot that works with pugs, and I pull from upstream every now and then 18:35
fglock: I parse the ".method()" as a postfix 18:36
fglock audreyt: you mean, including the parameter list? 18:37
audreyt right, that's "method postfix"
the secondmost tight prec
rafl audreyt: There's no configurable way to not install those third-party modules somewhere else than in the users home directory, I guess?
audreyt rafl: currently, no, though I can see installing them to system may work 18:38
rafl: but then that means that step needs to happen as root before pugs gets built
Aankhen`` What are .hi files? 18:39
audreyt Aankhen``: they are like C's .h files that describes what a .o file contains
rafl audreyt: Why not use a temporary packages.conf somewhere under the build tree to register them and install them with the rest of pugs after the build?
audreyt rafl: we tried that
Aankhen`` I thought C .h files are header files? :-S
rafl audreyt: What was wrong with that approach? 18:40
audreyt Aankhen``: and .hi are "interface" files
rafl: that will require "make install" stage to readjust the .conf
Aankhen`` Ah.
rafl audreyt: What's the problem with that? Except someone may need to write some code that does it? 18:41
audreyt rafl: it could be made to work, I think, and is the correct way.
rafl: that requires a ghc-pkg wrapper 18:42
ingy clkao: ping
audreyt since cabal doesn't take ghc-pkg flags
rafl audreyt: I've done that several times while packaging haskell libraries for Debian.
audreyt rafl: and the wrapper needs to be written in Haskell and compiled to .exe much like Setup.exe (or merge both? maybe)
rafl audreyt: Why does it need to a .exe?
audreyt rafl: because Win32 would require that 18:43
I mean an executable, not the suffix
fglock audreyt: I was confused with '.meth listop' in S03 - I was reading it as 2 words - think it needs quotes
audreyt i.e. a #!/usr/bin/perl wrapper wouldn't do
rafl How about perl? It's a prereq anyway?
audreyt yes but you can't pass arguments into it
--with-ghc-pkg=perl
that doesn't work
rafl I see.
audreyt I suspect you can fiddle with PERL5OPTS or something to cheat 18:44
rafl Luckily I recently had some haskell lessons at university. :-)
audreyt but the cleanest way is just build a ghc-pkg wrapper from .hs into .exe
the same time Setup.exe gets built
gaal audreyt: have you compiled PIL^N recently? I'm getting compiler errors on Data.Seq
audreyt rafl: so if you can Solve This Tangled Mess then wow, you++
rafl audreyt: At least I'll give a try. 18:45
18:46 coumbes joined
svnbot6 r11091 | ingy++ | r713@ingy: ingy | 2006-04-14 21:47:37 -0700 18:46
r11091 | ingy++ | Fixed so that pod2html works
gaal audreyt: sial.org/pbot/18190
lambdabot Title: "Paste #18190 from Someone at 66.92.67.231"
ingy svk-- # hanging... 18:48
Limbic_Region is there anyway with svn/svk to ensure every single file in my local rep matches that of the repo?
audreyt gaal: fixed 18:49
Limbic_Region on more than one occassion I have gotten a message about foo not being updated and to run with merge track something or other
gaal cheers
18:49 vel joined
svnbot6 r11092 | audreyt++ | * repair "make pil" to work with newer libs and GHC. 18:52
r11093 | ingy++ | r714@ingy: ingy | 2006-04-15 04:42:01 -0700
r11093 | ingy++ | S26 first sweep of changes
ingy wonders wtf svk is doing 18:53
18:57 elmex_ joined 18:59 shachaf joined
svnbot6 r11094 | audreyt++ | * Another set of fps changes to make "make pil" happier 18:59
gaal audreyt: objectIds can't live in objectspace, they aren't objects. where do I keep a Str -> objectId mapping and is it global? 19:03
audreyt Str? 19:05
audreyt is confused 19:06
gaal mkType "Moose"
should return the id of Moose.meta
audreyt instead of a VType value?
gaal do we stil have VType? 19:07
audreyt we do
it's the prototypical object for something named that
inside the scope
gaal is it still represented the same way? as a wrapped string? 19:11
audreyt I think so. 19:12
otoh, it can alternately be represented as a instance
with no slots
gaal and I still don't understand how the lookup is to work. data Object no longer has a VType field, we changed it it the ojectID of meta
TreyHarris pugs has no float support?
audreyt and .meta as the id
gaal ?eval 0.234234 19:13
19:13 evalbot_11071 is now known as evalbot_11094
evalbot_11094 117117/500000 19:13
TreyHarris heh
gaal ?eval 0.234234.as("%f")
evalbot_11094 \"0.234234"
svnbot6 r11095 | fglock++ | px/fglock/v6 - fixed .meth
TreyHarris so it has float support, so long as the floats are strings :-)
gaal it's just very happy to use Rats when it can 19:14
19:14 vel joined
audreyt gaal: let's talk f2f :) 19:15
obra: please pick me up in... 15mins?
Limbic_Region oh, is anyone taking pics at the hackathon?
svnbot6 r11096 | audreyt++ | * repair "make pil'.
Limbic_Region would love to be able to put faces to handles since he couldn't be there to see ppl f2f 19:16
obra audreyt: food just appeared. I'll want a few more minutes to actually eat 19:18
audreyt sure! 19:19
gaal Limbic_Region: nobody seems to have a camera here, but see nothingmuch's moosepage on flickr 19:21
19:22 elmex_ is now known as elmex
miyagawa just arrived NY 19:25
Limbic_Region gaal - link? 19:27
gaal Limbic_Region: www.flickr.com/photos/nuffin/sets/7...183758406/ 19:28
lambdabot Title: "YAPC::NA::2006 MĆøĆøse photos - a photoset on Flickr"
stevan M???????????ssseeeee 19:29
gaal stevan!
you are moosed
stevan gaal: Alias seems to think we are doing an "underground marketing campaign:
Limbic_Region err, lots of photos but apparently not lots of names/tags
stevan no love for just pure moose ness
gaal no, just taking over the world
stevan Limbic_Region: the roll overs have the names on them
gaal Limbic_Region: there are tags in many of them (annotations) 19:30
Limbic_Region roll overs?
Limbic_Region recognizes Randal and always knew he wasn't human
stevan Limbic_Region: careful with the ones of ingy, they are not 100% work safe ;)
Limbic_Region: the little flickr tags.. they appear when you roll over the pics 19:31
Limbic_Region ahhh - roll over
got it
PerlJam wonders what the odds are of two people named "Patrick Michaud" attending YAPC ... 19:32
TimToady 100%
stevan yeah exactly,.. I thought that too
We are working on cloning audreyt as well, so this should be more common in subsequent YAPCs 19:33
audreyt TimToady: hi. is ::?CLASS just $?CLASS?
PerlJam stevan: awesome!
audreyt TimToady: and is @?LABEL just $?LABEL?
Aankhen`` LMAO stevan.
TimToady I answered that earlier. 19:34
audreyt hm.
I didn't see them in backlog-machine
gaal missed the answer too
audreyt checks her backlog reader
audreyt: @?LABEL was intended to be all the labels visible outward in the dynamic scope.
::?CLASS is probably redundant with $?CLASS now
gaal colabti.de/irclogger/irclogger_log/...l=340#l511
lambdabot Title: "#perl6 2006-07-03,Mon"
audreyt TimToady: "dynamic" scope?
not lexical?
gaal audreyt: obra on his way to you 19:35
audreyt i.e. OUTER or CALLER's label?
gaal: oh. ok.
TimToady lexical then dynamic, I guess.
clkao audreyt: obra just went out 19:36
19:36 dduncan joined
TimToady same order they'd be searched for. But it's probably not important 19:36
audreyt i.e. anything that can be goto'd to
TimToady I was thinking of loop labels. 19:37
finding gotoable labels seems a lot harder.
specially if we make line numbers gotoable. :)
audreyt wait...
do you mean @?LABEL only remembers the labls associated with Code
rather than with statements? 19:38
also, is .ref still used to return the prototype object
or is it .class now?
Dog.new.ref; # Dog 19:39
TimToady doesn't "remember" anything. Just constructs the list of labels it'd find if you tried "next NoneSuch".
audreyt 'k
TimToady don't really need it, probably.
I want to kill .ref.
audreyt hard to find a use case.
ok. then .proto? .class? .object_class?
TimToady .type maybe, or .proto
audreyt still need something that's not .meta
.type is fine too
TimToady yes 19:40
can't use .meta
merlyn and I were discussing this very thing
.type is probably the most transparent
and available
TreyHarris what's wrong with .meta?
TimToady it's the metaclass object, not the type 19:41
TreyHarris beside being CS-y, which i know merlyn doesn't like? oh, so .meta will exist, it just can't be used for that.
TimToady you could have a single metaclass object managing multiple types, and vice versa.
yes 19:42
PerlJam TimToady: BTW, I was rereading the Synopses last night and I noticed that, in general, it doesn't say much about when things go wrong. The only example I remember is in S06 where it talks about "attributive parameters". It doesn't say what should happen when the object doesn't have an attribute with the name used in the submethod.
TimToady plus it's just a really old-school meaning of "type" for it to be the prototype object.
"Abraham is a type of all who believe". 19:43
PerlJam s/it doesn't/they don't/
TimToady well, in general, if something goes wrong, you get an exception. Sometimes it's thrown, and sometimes it's just returned. 19:44
what it does with an attributive parameter that isn't supplied would presumably be just the same as an ordinary parameter that doesn't have a default. 19:46
(where optional means defaulting to undef)
19:47 pmurias joined
pmurias nothingmuch: ping 19:47
nothingmuch pmurias: pong
TimToady but in general, not supplying a required parameter would result in a die rather than a fail.
pmurias nothingmuch: does the new wemail interface work for email sending? 19:48
TimToady and yes, that's not specced anywhere.
nothingmuch die == your fault
fail == maybe your fault
but probably not?
pmurias: i tested it and it seemed to, yes
TimToady fail == huh? die == fu! 19:49
nothingmuch TimToady: i think that's what I meant =)
as in, bad input vs. bad circumstances
there should be a style guide for this
TreyHarris my browser's off in la-la-land for the moment and i don't have my browser's off in la-la-land for the moment and i don't have the .pod files handy, so apologies if this is in the synopsis, but is croak built-in now? 19:50
nothingmuch also, when does 'use fatal' and 'no fatal' actually take effect?
pmurias nothingmuch: used the old one and it worked (worse is better:)
nothingmuch ialways assumed that whenever an error object passes through (implicitly or explicitly) a use fatal block it dies
and that no fatal has an implicit try { } around every call to out of scope code
TimToady croak is currently unspecced. 19:51
nothingmuch pmurias: okay, i'll look at the smtp logs
19:51 rindolf joined
nothingmuch croak should be Carp::Clan like, IMHO 19:51
rindolf Hi all.
Hi nothingmuch
nothingmuch Hi
rafl Is there a ghc equivalent to perls Config.pm?
rindolf nothingmuch: I need help with sial.org/pbot/18191 - it's in Haskell.
lambdabot Title: "Paste #18191 from "rindolf" at 88.155.89.47"
nothingmuch create_empty doesn't need to be in the IO monad 19:52
rindolf nothingmuch: OK.
nothingmuch: so how do I use it with main?
nothingmuch just use empty and a non binding
rindolf nothingmuch: thanks.
nothingmuch instead of <- have a where/let
rindolf Couldn't match `IO' against `Map k' 19:53
nothingmuch or just putStr (to_str empty)
right, because empty returns values from the type 'Map k' (k is an unbound type variable (s))
rindolf nothingmuch: yes, but how do I add more elems to it?
TreyHarris croak being built-in would be nice, imho. "my fault" (die) versus "your fault" (croak) is a good duality.
nothingmuch you don't, it's a purely functional data structure
the add function returns a new map object 19:54
s/object/thingy/;
gaal FWIW, Carp::longmess has been in Prelude.pm for ages now
rindolf nothingmuch: I see.
nothingmuch whcih contains all the values from the previous one and this one too
there's prolly an IORef like map object that can work for you
i just don't know about it
maybe even JudyHS
since Judy can be abused very effectively for caching
err 19:55
hashing, not caching
doesn't the map class have an instance of Show, btw? 19:56
also, i believe maps should be sorted by default since IIRC they're implemented as trees
gaal @hoogle Data.Map
lambdabot No matches, try a more general search
nothingmuch yes
insert :: Ord k => k -> a -> Map k a -> Map k a
k, the key, needs to be of the class Ord 19:57
so maps use ordering to search in log n instead of constant time
or n time
pmurias is there an official static type system for perl?
nothingmuch pmurias: Perl 6?
gaal wtm? Map is not core?
pmurias yes
nothingmuch there is a fairly well known type annotation system
and a probable approach to how to infer (it'll probably be local type inferrence) 19:58
gaal haskell.org/ghc/docs/latest/html/li...ml#t%3AMap
rindolf Hi gaal
gaal instaces: ... (Show k, Show a) => Show (Map k a)
Hi there
so if the key and value are Show, so is the map 19:59
nothingmuch @hoogle showTree
lambdabot Data.IntMap.showTree :: Show a => IntMap a -> String
Data.IntSet.showTree :: IntSet -> String
Data.Map.showTree :: (Show k, Show a) => Map k a -> String
nothingmuch that's a debugging function
you can use that
haskell.org/ghc/docs/latest/html/li...ap.html#21
pmurias how much does type_meta.pod apply? 20:00
nothingmuch if you annotate and it's mostly by me or luqui than probably noit much ;-) 20:01
oh
hmm
that actually explains fairly well how there is no real type system yet =)
pmurias luqui was supposed to have a big influence on the type system 20:02
nothingmuch yes, but this stuff never really materialized yet 20:03
it's too early to know
we need typing in pugs to figure out the details
and then make smarter decisions
luqui has stepped back from the project since 20:04
pmurias shower& 20:06
nothingmuch pmurias: okay, something was less strict before
probably spam related
20:06 larsen joined
nothingmuch i'll reconfigure some stuff and it should work soon 20:06
pmurias: done 20:07
svnbot6 r11097 | fglock++ | px/fglock/v6 - 3 tests to go - t/07-simple-multisubs.t, t/10-moose1.t, t/11-capture.t 20:13
nothingmuch rindolf: how are you managing?
20:14 petdance joined
pmurias nothingmuch: did you read TaPL? 20:17
20:17 ludan joined
pmurias nothingmuch++ #non-sucking-mail-services 20:17
20:18 fglock joined
nothingmuch pmurias: i lent my copy to lumi 20:18
i didn't get it back yet 20:19
and I haven't gotten to it before
so no =(
nothingmuch basks in his karma 20:21
pmurias how do you get your karma level? 20:22
get info abouy
@karma nothingmuch 20:23
lambdabot nothingmuch has a karma of 0
pmurias @karma pmurias
lambdabot You have a karma of 0
nothingmuch lambdabot has a different syntax
pmurias ?
?karma nothingmuch 20:24
perlbot karma nothingmuch
perlbot Karma for nothingmuch: 43
nothingmuch @karma+ and @karma-
pmurias perlbot karma pmurias
perlbot Karma for pmurias: 5
nothingmuch jabbot: karma nothingmuch?
jabbot nothingmuch: nothingmuch? has neutral karma 20:25
nothingmuch jabbot: karma nothingmuch
jabbot nothingmuch: nothingmuch has karma of 79
lambdabot nothingmuch has a karma of 0
nothingmuch purl on irc.perl.org wins though
lambdabot and's karma raised to 1.
pmurias jabbot karma pmurias
nothingmuch nevermind, i'm glad you like my mail services =)
fglock perlbot karma fglock
perlbot Karma for fglock: 130
Juerd perlbot karma juerd
perlbot Karma for juerd: 66
Juerd 600 to go.
nothingmuch heh
ruoso perlbot karma ruoso 20:29
perlbot Karma for ruoso: 6
ruoso heh
rindolf nothingmuch: not well at all. 20:36
nothingmuch: it takes too much time.
nothingmuch: with Perl, I would have finished it by now.
Anyway, I'd like to report some Mandriva bugs I found in Cooker.
Juerd perlbot karma rindolf
perlbot Karma for rindolf: 15
rindolf nothingmuch++ # Helping me a bit. 20:37
nothingmuch rindolf: well, you don't *have* to use haskell 20:38
as audrey says haskell is good for when you know what you exactly what you want to do 20:39
while perl is good for when you're poking around 20:40
both probably more so than any other language
rindolf nothingmuch: I see.
nothingmuch: well, I'm trying to write a script I want to write in a differnet language than Perl.
gaal "Perl 6's model is essentially like Perl 5's except that it treats classes as first class objects, not strings" ~~ audreyt 20:41
rindolf nothingmuch: my candidates were Smalltalk (that I'm learning at the moment), Haskell (which I learned back then) and Common Lisp that I'd like to learn.
nothingmuch okay 20:42
rindolf I'm also refactoring the Graham Function Perl 5 script so it wil be easier to translate to ST and CL.
nothingmuch haskell is probably not the best scripting language
Juerd What is it called when something happens often, but at irregular intervals? 20:43
ruoso periodic
?
rindolf nothingmuch: tell that to wli. 20:44
dduncan I wonder if mine works yet ...
Karma for dduncan
20:44 petdance joined
rindolf perlbot: karma for dduncan 20:44
perlbot Karma for dduncan: 24
jabbot rindolf: for dduncan has neutral karma
rindolf Hi petdance 20:45
petdance hi shlmoi
dduncan perlbot: karma for Darren_Duncan
perlbot Karma for Darren_Duncan: 6
jabbot dduncan: for Darren_Duncan has neutral karma
nothingmuch rindolf: by "scripting" i mean "forgiving, adhocish"
dduncan unfortunately, mine's split due to the latter being my SVN username
rindolf nothingmuch: yes, I know.
nothingmuch it's not so much the type of program, but the methodology
rindolf nothingmuch: thing is wli is writing his scripts in Haskell.
nothingmuch: I know. 20:46
nothingmuch that's good for wli
rindolf nothingmuch: do you know Common Lisp? Can you recommend an implementation?
nothingmuch but wli is probably not learning haskell right now
rindolf nothingmuch: do you know who wli is?
nothingmuch nope,nope, and nope
well, i kinda know cl, i know scheme
but i haven't more than played around 20:47
petdance oh wait, I don't want this channel on autojoin.
I just like parrot. Perl6 is too real! :-)
20:47 Odin-LAP joined
rindolf nothingmuch: I also know Scheme. 20:48
nothingmuch: or at least knew. Now my knowledge is somewhat rusty.
nothingmuch scheme and cl are similar languaages, but subtly different. however, cl is much more than just a language
Juerd ruoso: Not what I'm looking for...
audreyt TimToady: this:
Because Perl 6 uses a false C<.id> to signify a non-instantiated prototype,
all instances should arrange to return a C<.id> that boolifies to true.
nothingmuch it has 20:49
ack, typing
audreyt I take that it means Int.id and Dog.id is both 0
rindolf Maybe I'll write a TAP runner in CL.
audreyt (or False)
nothingmuch rindolf: okay 20:50
ruoso Juerd, in which sense you mean?
Juerd Wish I knew.
rindolf nothingmuch: since MikeSchwern wants two different implementations and stuff. 20:51
So there will be 3 or 4.
audreyt TimToady: so I wonder if there should be a separate (.meta.id, .id) tuple
that is good enough for (say) hashes that use objs as keys
dduncan One thing I've never been quite clear on ... are .id each integers?
and nothing else? 20:52
s/nothing/something/
forget that
Juerd I hope they will be integers.
It would be a nightmare if they could be any object.
nothingmuch rindolf: good luck
dduncan my reading of them so far didn't indicate any type, though a few examples suggest integers 20:53
rafl audreyt: While writing the ghc-pkg-wrapper, would it be safe to use components in src (namely System.FilePath)?
Juerd Strings may be useful too.
Then they can include hostnames in network cluster builds.
dduncan that said, I wonder about container types
Juerd But the performance difference is huge.
dduncan eg, what is the .id of a Pair object?
how is that determined 20:54
Juerd 42
dduncan I need to know things like how a .id is determined for a Pair or Seq or Mapping etc, so I know the best way to generate such for complex things like my Relation
I'm assuming it can't just be a hash, since sometimes multiple things would hash to the same value, and two unique .id are supposed to guarantee two identical values, regardless of the type, and for immutable types, even with different actual objects 20:56
rindolf OK, reported one Mandriva bug - now trying to recall others. 20:57
20:57 spinclad_ joined
rindolf Found one - akregator won't open links using konqi. 20:57
21:05 Limbic_Region joined
chip audreyt, all: hi 21:08
audreyt: Q about PIR
21:09 xdg joined 21:13 vsoni joined
svnbot6 r11098 | fglock++ | px/fglock/v6 - pass t/10-moose1.t 21:13
vsoni chip: hi chip
chip rehi :-) 21:14
21:14 unobe joined
obra hey chip 21:15
chip obra: hi jesse 21:16
rindolf Hi chip 21:18
chip: didn't the King of Prussia whatever press charges against you?
chip: in any case, I really appreciate the work you've done to maintaining perl5, etc. 21:19
dduncan hi chip, how's the old block? 21:21
Limbic_Region nothingmuch ping 21:22
nothingmuch Limbic_Region: pong
spinclad_ cheeep!
21:22 coumbes joined
Limbic_Region very cool moose pics 21:23
Limbic_Region is from Maine where moose are common place
Limbic_Region isn't sure which are better looking though
Limbic_Region especially liked the traitor shot 21:25
dduncan url? 21:26
spinclad_ chip: audrey's talking with gaal atm 21:27
fglock clkao: ping
Limbic_Region www.flickr.com/photos/nuffin/sets/7...183758406/ 21:28
lambdabot Title: "YAPC::NA::2006 MĆøĆøse photos - a photoset on Flickr"
Limbic_Region anyone know what "1 file skipped, you might want to rerun merge with --track-rename." from svk means and how to fix it
spinclad_ offline borg 21:29
clkao fglock: pong 21:33
ingy clkao: ping
clkao ingy: hey
ingy clkao: two things
first I changed M::C to include the file hacks with your package hack 21:34
fglock clkao: almost ready for merging - can you point me to svk instructions?
ingy but I changed caller(4) to caller(2)
not sure if it's right, so please test
Secondly, I hate svk ;)
well right now I even hate svn! 21:35
why won't anything work today??!!
ingy pouts
21:36 jiing joined
fglock wonders what "Automatically merge" would do 21:40
svnbot6 r11099 | fglock++ | px/fglock/v6 - ready to merge back to perl5/
chip spinclad_: ok 21:46
spinclad_: hey there
dduncan: yo dude
clkao fglock: svk sm -C //path/to/px/fglock/pc-p6 //path/to/perl5/pc-p6
ingy: why did you change it to caller(2)?
fglock clkao: starting... 21:47
installing Devel::Caller in order to test perl5/PCP6 - "Module::Build is not configured with C_support" - any ideas? 21:48
ingy clkao: well it's in a different place
I'm not even sure that caller isn't too fragile... 21:49
anyway let me know if it works or not
dduncan hello chip
ingy and I'll fix it clkao
clkao fglock: resintall module::build after you have something extutils:: something c. i forgot 21:52
hate module::build
fglock k
clkao or you can do the merge down by swapping the two args, and make sure it works. then merge back 21:54
fglock perl5/PCP6 tests don't pass - checking... 21:55
"die _emit" :) 21:57
clkao uhm :p 21:58
fglock error - Repository /home/fglock/.svk/local does not exist, create? (y/n)y / No such depot: home. 22:00
svnbot6 r11100 | ingy++ | r1931@ingy: ingy | 2006-06-26 13:29:00 -0700 22:02
r11100 | ingy++ | perl5/
clkao er? i thought you have svk checkout already/ you haven't used svk at all? 22:04
svnbot6 r11101 | ingy++ | r2126@ingy: ingy | 2006-07-03 11:44:37 -0700
r11101 | ingy++ | Refactor v6.pm to use M-C-0.17 which does the .xxxc file check.
r11101 | ingy++ | All add clean_files to Makefile.PL
ingy svk++ # decided not to segfault this time
fglock clkao: do I need an import or checkout first? (I was only using svn until now) 22:08
clkao fglock: ya. 22:13
fglock clkao: can I reuse the svn directory, or will it be mixed up 22:15
(still reading the svk instructions) 22:17
clkao uhm no. but someone suggested something to turn svn checkout into an initail svk mirror and checkout
fglock: ask nothingmuch for quick tour. i am working on something now
TreyHarris lol. message from one of our 'developers' to the internal Perl list: "i have program that needs to check every few hours whether condition is met. right now, i go into infinite loop and check time() to see if enough has past. any way to just tell program to go idle for some hours? thx" i should tell him to sleep on it, maybe the problem will come to him in the morning.
fglock clkao: ok
chip irc.perl.org connection went down 22:18
22:20 spinclad_ left
TimToady society is to blame. 22:21
svnbot6 r11102 | gaal++ | * remove unneccessary method so that Data.Seq compiles on GHC 6.4.2 again 22:22
r11101 | ingy++ | r2126@ingy: ingy | 2006-07-03 11:44:37 -0700
r11101 | ingy++ | Refactor v6.pm to use M-C-0.17 which does the .xxxc file check.
r11101 | ingy++ | All add clean_files to Makefile.PL
nothingmuch fglock: clkao has assigned me to you 22:23
ingy raises an eyebrow at svnbot6
fglock nothingmuch: I have 2 dirs in svn, I want to merge using svk 22:25
nothingmuch how did you make the two dirs?
svk cp in your checkout?
fglock no, but clkao said he fixed it already 22:26
I only installed svk afterwards
nothingmuch svn cp
is also the same 22:27
okay, so if clkao fixed it
fglock can I convert the my svn checkout to svk? 22:28
nothingmuch then it's probably going to be svk smerge -f path_to_branch -t path_to_trunk
you canj just check it out again
fglock with 'svk co' ?
nothingmuch yes 22:29
svk co svn.openfoundry.org/pugs
lambdabot Title: "Revision 11075: /"
nothingmuch or just the dir
that will ask you to mirror
you can accept the defaults
but you probably don't want all
because that will take hours 22:30
fglock I'm trying with 'svn co -r HEAD svn.openfoundry.org/pugs/ .
ingy IT'S FASTER TO ORDER THE PUGS CD FROM AMAZON
lambdabot Title: "Revision 11075: /"
fglock done! 22:31
nothingmuch i dunno how to merge with svn
fglock nothingmuch: I'll retry with 'svk sm -C //path/to/px/fglock/pc-p6 //path/to/perl5/pc-p6' - clkao recommended 22:32
ZuluOne Juerd: frequent
nothingmuch clkao knows $n times better than me 22:33
Juerd ZuluOne: That doesn't have to be random intervally
ZuluOne doesn't have to be
it has no connotations one way or t'other
Juerd I'm looking for a word that does have the connotation of random intervals.
ZuluOne I don't believe such a word exists, but I'll check. 22:34
fglock nothingmuch: I get 'No such depot: home.'
nothingmuch oi
that seems beyond me
did you check //path/to/px with svk?
fglock oops - I type svn instead of svk before - one sec.
ZuluOne Recurrent has a very slight connection of irregularity. 22:36
Intermittent might be OK: depends on the context.
Fitful is quite a good fit, but would sound weird in a technical context. 22:38
Of course, the obvious solution is just "irregular."
What actually is the context? 22:39
fglock nothingmuch - I'll finish later - thanks! & 22:40
nothingmuch fglock: ok 22:41
svnbot6 r11103 | gaal++ | * Add ghci quickfix-able mode for vim 22:43
22:44 fglock left
gaal audreyt: cp ~gaal/.vim/plugin/Align* to your dir 22:44
every other .hs-writing vim user: vim.sourceforge.net/scripts/script....ipt_id=294 22:45
lambdabot Title: "Align - Provides commands and maps to help produce aligned text, eqns, declarati ..."
ZuluOne Juerd? 22:46
dduncan coming up fast ... 22:47
so who wants to do commit #11111?
the likes of which only comes about every 11111
lumi Juerd: Sporadically? 22:49
Juerd Not often enough :)
ZuluOne Juerd: What is the context? 22:50
Juerd I think I was looking for intermittent. 22:51
ZuluOne k
Juerd ZuluOne: An alarm system that beeps.
ZuluOne yep, sounds like intermittent
Juerd thesaurus++ # found it, beginning at irregular 22:52
Is English the only language that has this phenomenon (thesaurus)?
ingy rgs: YAML-0.62 uploaded
ZuluOne There is no problem that Roget's cannot solve.
Juerd ZuluOne: The halting problem? 22:53
ZuluOne If the halting problem is that you can't find a synonym for it, then sure :->
lumi It turns it into a hesitant bother 22:54
ZuluOne 145 Cessation: change from action to rest 22:56
Juerd Good night
ZuluOne gnight
rafl needs testers for files.perldition.org/pugs_local_pac..._conf.diff 23:13
audreyt rafl: commit it 23:14
and you get testers
automatically
rafl audreyt: I don't have the bandwidth to update my svk co and merge it, sorry.
audreyt: Could you commit it, please?
audreyt: The patch currently works fine for compiling pugs. I'm not sure what happens if it get's installed, yet. 23:15
audreyt ok 23:17
testing
rafl goes for a cigarette in the meantime. 23:20
ingy stevan: ping 23:23
svnbot6 r11104 | gaal++ | * handle multi-line errors (which ghc is likely to produce :-) 23:24
23:29 diotalevi joined 23:31 cognominal joined
vsoni hi audreyt 23:40
xerox pokes audreyt 23:44
audreyt pong 23:45
23:45 Entonian joined
rafl goes to bed 23:50
Good night everyone