Pugs t-shirts www.cafepress.com/pugscode | Pugs 6.2.9 released | pugscode.org | pugs.kwiki.org | paste: sial.org/pbot/perl6 | www.geeksunite.net
Set by stevan on 15 August 2005.
gaal morning! 05:04
dduncan evening! 05:10
however, I can't stay and chat, since sleep beckens
s/c// 05:11
adieu!
gaal autrijus, how's your passport situation? 09:32
hey szabgab :)
szabgab autrijus, while answering gaal, I would like to ask you if you have time after OSCON EU to come to visit the Hungarian Perl Workshop ?
hi gaal
new-b hello 09:33
gaal hi b 09:34
new-b hi gaal :) 09:35
nothingmuch morning 10:43
gaal hey nm 10:49
nice module
nothingmuch hi ho 10:51
grazie
nothingmuch really wants to finish the type inferrencer for Blondie, but is at work =(
all day i've been chasing silliness
gaal maybe you need to write a silliness chaser to do that for you 10:52
nothingmuch MSDEV took another 20 minutes out of my life because it failed to write a file but was displaying it as if it was written
gaal see if you can throw a type inferencer into that
nothingmuch our unit testing library got the TODO feature since someone finally asked for it
and some script broke because of a very stupid versioning scheme
how do you type infer silliness? 10:53
i think silliness is a bottom type
gaal yeah, and stupidity is a pervasive type
nothingmuch pugs is costing me lots of money =(
nothingmuch spent $170 on books so far
$1000 or so on a flight 10:54
$3500 on a new laptop
gaal add psychiatric expenses 10:56
nothingmuch oh my 10:57
gaal however, it'll get you a free ticket to yapc
nothingmuch really?
gaal because you'll have oodles of things you can talk about 10:58
nothingmuch oh, i see
nothingmuch always paid when he was a speaker
what will i talk about?
gaal whatever you like, of course
nothingmuch will think
gaal pitching MMD the way you think of it, for example 10:59
szabgab hi nothingmuch !
nothingmuch heh
hi szabgab
szabgab: so, do you have any suggestions for what I could talk about? 11:14
silly idea: Perl 6 is Here Today! Tomorrow Next Week!
essentially - far fetched ideas for perl 6's future 11:15
especially WRT implementation - all the crap I write p6l about
gaal yeah, for you the hard part doesn't seem to be thinking of "all the crap" 11:31
it's explaining it :)
szabgab nothingmuch that should be in a few month from now, not today 11:41
nothingmuch i'll find a topic
gaal fe 12:24
rafl_ Hello. 12:47
On 64 bit archs I get the following messages when compiling:
Triggering rebuild... done.
Generating precompiled Prelude... pugs: internal error: scavenge_one: strange o bject 68
Is this a pugs or a ghc bug? 12:48
gaal ghc 12:57
what ghc are you using?
do you get it in the same place every time?
how much ram do you have?
rafl_ gaal: ghc 6.4-4.1 on Debian. 12:59
gaal: I don't know how the buildd's hardware looks like.
s/how/what/ 13:00
gaal hmm, it happens when your pugs is eating lots of ram. i wonder if we can influence its runtime allocation.
s/your/their/
rafl_ So who needs to fix that? ghc I guess. 13:02
gaal there may be a way for us to set that. i don't have time to look into in now though 13:04
oh, RTS is available on the command line...
can you ask them to do something like
1. build a pugs with no precomp prelude (vi config.yml and s/true/false)
2. try to comp a prelude like this: ./pugs +RTS -M350m -RTS -CPugs src/perl6/Prelude.pm > /dev/null 13:06
if this helps, we can fix this
see haskell.org/ghc/docs/latest/html/us...ntrol.html 13:07
rafl_ tries 13:15
luqui just learned Curry 13:24
it makes me want to decouple parsing strategy from patterns in perl 6
but without a drastic reformulation of "let", I don't think it's possible
so, it's time to work on a drastic reformulation of "let" (just kidding ;-) 13:25
gaal rafl_, ping 13:47
rafl_ gaal: pong 13:48
gaal try also talking to putter, a random google shows he'd submitted a bug report to ghc about this.
13:48 rafl_ is now known as rafl
rafl OK, thanks. 13:49
WTF is perl5/Perl6-MetaModel2.0/docs/p6_role_model.graffle ? 14:08
gaal rafl: I think it's a data file for pretty diagrams. 14:45
nothingmuch woo!! 15:26
App(Sym('&print'), Sym('$*OUT'), Val(42)) --> App(Val(Prim), Val(*STDOUT), Val(42))
then gets type annotated
App(Val(Prim::(GV -> PV -> IV)), Val(*STDOUT::GV), Val(42::IV)))::IV 15:27
since there is a conflict, the Val(42) is coerced from IV to PV
(the runtime provides this - in the perl runtime it's jsut sub { "$_[0]" } 15:28
so the tree is trasformed to
App(Val(Prim), Val(*STDOUT), App(Val(Prim), Val(42))
well, actually to the &cast builtin, so App(Val(Prim), Val(42), Val(PV), Val(IV)) 15:29
someone say yay or something 15:30
GeJ "yay".say 15:56
nothingmuch works on inter sub propagation of values 16:02
err, types
svnbot6 r7036 | stevan++ | Perl6::MetaModel 2.0 - 16:17
r7036 | stevan++ | * sketched out role(Role) in psyche
r7036 | stevan++ | * Role now calls resolve() on it's won in MiniMetaMode_w_Roles
r7036 | stevan++ | * added test to show that the model can host inside-out objects
r7036 | stevan++ | * other misc. minor tweaks
nothingmuch huzzah!!!!! '&say' now enforces the PV type on Val(42) when it's IV 16:18
pasteling "nothingmuch" at 82.81.247.180 pasted "typechecked App(Sym('&say'), Val(42))" (184 lines, 13K) at sial.org/pbot/13188 16:19
nothingmuch wonders wheather to clean out all that useful debugging or not ;-) 16:22
gaal maybe commented-out trace prints need to be tagged by feature, so that they can be nuked quickly later? 16:50
nothingmuch that's a nice idea 17:12
but if you're going that far, just Log::Log4perl
"nothingmuch.woobling.org/blondie_ty...notation." ~ any(<<png pdf>>) 17:59
nothingmuch ponders whether to do type inferrencing of generics, or to compile to a static language first 18:05
svnbot6 r7037 | nothingmuch++ | Type inferrencing Blondie: 18:11
r7037 | nothingmuch++ | * Blondie::TypeSafe does type inferrencing (no generics yet)
r7037 | nothingmuch++ | * example.pl simplified to demo it
r7037 | nothingmuch++ | * lots and lots of debugging info
r7038 | nothingmuch++ | explanation
r7039 | nothingmuch++ | Fix up edge colors for fully annotated AST 18:17
ods15 wow, sounds like larry wall and pugs guy are visiting in israel 18:31
unless i misunderstood this mail
Khisanth autrijus and larry are visiting isreal? 18:33
nothingmuch yep
Khisanth ods15: nope! you didn't misunderstand! :P 18:34
ods15 heh
nothingmuch btw, pugs guy has a name, y'know 18:35
it's our very own autrijus
ods15 its hard to type
nothingmuch you can use tab completion
ods15 i refrain from typing any weird names because i will probably mispell them
well i didnt know he was here
Khisanth a<tab> :)
ods15 Amnesiac:
nothingmuch for a long time jp-autark called himself 'autark-jp' so we would always call autrijus autark 18:36
au<tab>
Khisanth not here :P
nothingmuch autrijus: he isn't... he is in taiwan, leaving for estonia
Amnesiac ods15, yes?
ods15 ... nm
nothingmuch but YAPC::Israel is in feb
Amnesiac: ods15 was learning about his tab completion feature ;-)
ods15 ugh i hate ::
Amnesiac ah
:/
Khisanth hmm last msg field is broken 18:37
nothingmuch ods15: bummer
ods15 ? 18:38
nothingmuch WRT :: 18:42
ods15 whats wrt 18:43
nothingmuch with respect to
wolverian can object attributes have where { } clauses? 18:57
nothingmuch like has $.x where { } ?
wolverian can variables in general have where clauses? (basically, shortcut to STORE => { ... })
nothingmuch, right
nothingmuch i don't see a reason why not
as for variables:
my Any where { } $x; # subtyyped
i would guess
wolverian right.
?eval class Foo { has $.x where { 0 < $
evalbot_7039 Error: unexpected "w" expecting trait, "handles", ";" or "}"
wolverian er
?eval class Foo { has $.x where { 0 < $^x < 40 } } 18:58
evalbot_7039 Error: unexpected "w" expecting trait, "handles", ";" or "}"
wolverian oh well. :)
nothingmuch ?eval class Foo { has Any where { 0 < $^x < 40 } $.x }
evalbot_7039 Error: unexpected "w" expecting variable name
nothingmuch =)
wolverian nothingmuch, if an object attribute has a 'where' clause, is the clause checked on each assignment?
rafl gaal: Thanks. 18:59
wolverian I mean, does it apply to private assignments too, or only ones via the public accessor?
nothingmuch i would guess so
(always)
wolverian right, if the feature is general to all variables
nothingmuch if you want it just public, make an acessor on your own:
wolverian could you check this on p6l, or is there a reference I could read to believe it is so? 19:00
nothingmuch class Foo { has $.x; method x (Any where { ... } $.x) { } }
this automatically sets $.x, as documented in s06 or s12
wolverian I have to go shower now and I won't be back today
:)
thanks
nothingmuch uh, i'm not sure how to ask
wolverian I'm interested in the general property
nothingmuch if sxx doesn't help, you mail p6l
wolverian nothingmuch, right, but is this documented in Sxx? if so, do you have to have the Any there
? 19:01
nothingmuch, because that's pretty ugly
nothingmuch well, I guess you could omit Any
but that's a p6l question
wolverian (my $foo is Int where { ... }; is naturally not ugly, but not everyone Types their variables)
nothingmuch, right. want to ask that? :) shower, bye!
nothingmuch nope =)
but i may
oh my... this is going to be tough 19:27
Juerd has suspend to memory and suspend to disk working 19:54
for the first time in my life :)
IBM ThinkPad X41 ++
Kubuntu ++
nothingmuch =) 20:03
gaal ThinkPad++ indeed 20:23
does anyone know how the new company is, whatever they're called? 20:24
rafl Juerd: Yes! Does suspend to ram work for you, btw? 20:46
Khisanth gaal: new company for what? 20:48
rafl Juerd: Oh, you wrote it works. How did you get STR working? 20:54
wolverian nothingmuch, oh, that reminds me (and it's the next day already so I didn't lie ;) - I wanted a way to set attribute accessors in the 'has' statement, like C# 21:03
nothingmuch, has $.foo get { ... } set { ... }; or so
nothingmuch doesn't know C#
that sounds cool though
but perhaps overly sugared
wolverian perhaps. the point is to attach that to the 'has' statement 21:04
(for easier refactoring etc)
nothingmuch delays the nested call placeholder synchronicity bug in the inferrencer till tomorrow
*yawn*
time to emit some C
Juerd rafl: What's STR? :) 21:05
wolverian I'll mail this to p6l, if no-one has objections 21:06
nothingmuch wolverian++
Juerd 23:04 < wolverian> nothingmuch, has $.foo get { ... } set { ... }; or so
isn't that just has $foo will FETCH { ... } will STORE { ... }?
$.foo even 21:07
wolverian Juerd, oh, that works? :)
I obviously skipped some synopses then
Juerd++
Juerd I don't know if it WORKS 21:09
But IIRC, that's how it's done
Is p6l dead, by the way? 21:10
Last message I received was 3 days ago
wolverian same here 21:12
cognominal same here 21:27
and I am worried about Nicholas being Warnocked
obra hmm? 21:32
21:33 khisanth_ is now known as Khisanth
nothingmuch 's brain is tired 21:37
s/tired/liquified/ 21:38
obra nick has ways of getting the answes he needs 21:39
and if any implementor isn't getting what they beed thry should be mailing me 21:40
nothingmuch obra: i need cookies 21:41
to implement Blondie::Backend::C and Blondie::Backend::PIR
am i considered an implementor?
do I need to mail you? or is IRC enough?
obra email me :)
email is logged and threaded
Juerd IRC is threaded too. Just not in any usable way :) 21:46
nothingmuch obra: sent
damn amazon
they're buttfucking the environment
they shipped my two books yesterday
and my CD today
Juerd Isn't everyone?
nothingmuch so that the books will "get there sooner" 21:47
Juerd That's for the global economy, it keeps postal people working.
nothingmuch at least i'll pay less customs
Juerd Worse: one of my suppliers sends invoices in one envelope per page
That means that if you ordered enough from them to get a two-page invoice, you get two separate envelopes. 21:48
That's not just bad for the environment, it sucks to handle too.
nothingmuch what do they supply?
Juerd Computer stuff
15 or so products fit on one page 21:49
nothingmuch wonders if C wouldn't have such a bad name if the std c library was designed to be slightly easier to use 21:50
obra: well?
obra dude i'm owalking around town o n irc with my phone 21:53
Juerd rafl: Oh! Suspend To Ram!
rafl: thinkwiki.org/wiki/Problem_with_dis...ter_resume 21:54
rafl: And I had to remove cardctl -eject from the preparation script, because that crashes at a distance, when my EDGE card is inserted
nothingmuch obra: so why did you ask me to mail you? 21:56
obra er. youb said you were blocking on something right? 21:57
nothingmuch i said I wanted cookies
obra oh. I was asuming pmcs
heh
nothingmuch oh, no... i don't grok those yet 21:58
obra i'll get in the nex elal flight out of logan
nothingmuch nono, that'll take too long
just mail them
nothingmuch calls it a night after having made 0 progress in the last 2 hours 21:59
obra heh
nothingmuch but all in all twas a productive day
svnbot6 r7040 | nothingmuch++ | r8165@syeeda: nothingmuch | 2005-09-19 01:03:46 +0300 22:09
r7040 | nothingmuch++ | Blondie:
r7040 | nothingmuch++ | * Minor refactoring of prim handling (more to come to :::Builtins)
r7040 | nothingmuch++ | * Improvements to pretty printer readability
r7040 | nothingmuch++ | * Better tracing in type engine - still have to fix >1 call level and placeholder synchronicity.
dduncan nothingmuch, is there a Dagwood module to go with the Blondie one? 22:53

Powered by Cro and the Raku® Programming Language.

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

Thank you!