6.2.11 released! | pugs.blogs.com | pugscode.org | pugs.kwiki.org | paste: sial.org/pbot/perl6 | < audreyt> my fingers juxpatosed
Set by nothingmuch on 24 February 2006.
00:34 nnunley joined 00:53 kanru joined 00:55 dduncan joined 01:06 kanru2 joined 01:10 kanru2 is now known as kanru 01:14 scw joined 01:34 lypanov left 01:39 kanru joined 02:05 kanru joined 02:25 stevan_ joined 02:35 scook0 joined 04:11 mako132_ joined 04:27 macli joined 04:39 hcarty joined 04:43 hcarty left 04:49 K_ is now known as K 04:58 lao joined 05:04 kanru joined 05:06 anatoly joined 05:33 kanru joined 06:23 anatoly joined 06:47 Cryptic_K joined 07:08 Aankhen`` joined 07:13 cognominal joined
audreyt greetings from OSDC.il! 07:14
dduncan greetings from Victoria.bc 07:15
gugod greeting from chupei.pm 07:16
07:31 levin__ joined, nothing_pasta joined
nothing_pasta hola 07:31
from OSDC
07:40 Khisanth joined
audreyt mm, lwall is on stage 07:53
(no electricity, brb)
nothing_pasta you lie 07:59
i just took your electricity
beppu_ is virtually and belatedly applauding for our $Larry; 08:09
dduncan I also applaud that which I can neither see nor hear 08:15
nothing_pasta actually he didn't get starting appluase
only one person
because nobody said "and now... LARRY WALL!!"
he just came up
dduncan well, the real fans need no introduction ... they can see him and know who he is 08:16
nothing_pasta but he still didn't get applause =) 08:18
beppu_ dduncan, sometimes, you just have to believe. ;-)
08:30 elmex joined
Khisanth dduncan: the weirdest looking one in the crowd? :P 08:30
dduncan I'm not at the event; I'm at home
but I have met Larry in person at OSCON2005 08:31
my point earlier was that I consider Larry distinctive enough that no one else at an IT conference would look like him, most likely 08:32
and all Perl fans would have seen his picture multiple times
08:32 K_ joined 08:33 iblechbot joined 08:53 christo joined
audreyt actually, David Roundy and another darcs developer looks 80% like lwall 09:08
except perhaps a bit younger
dduncan is ... that ... so ?
audreyt yeah, I was very surprised during ICFP 09:09
ayrnieu You should have killed the doppleganger while you had the chance.
audreyt dopplegangers, you mean
nnunley . o O ( There can be only one )
audreyt . o O ( doppler gangsters ) 09:10
dduncan why ... the dopplegangers can become useful to Larry if someone targets him ... more false targets
sounds kind of like a spy story 09:11
maybe
and you know Larry likes spy stories
especially Man From Uncle
or Get Smart
anyway, I have to get gone 09:12
09:21 Cryptic_K is now known as K 09:27 ghenry joined 09:28 beppu_ joined 09:34 KingDiamond joined 09:40 gaal joined
gaal audreyt: do you happen to have a version of Spork that can do .pretty for colorizing code? 09:42
audreyt gaal: p5 code? 09:43
p6 code?
shell?
gaal i can colorize it myself
the manpage mentions ".pretty" "coming soon"
audreyt don't, use VimMode
gaal and put in verbatim html?
audreyt .vim 09:44
gaal cool
audreyt #!perl6
...
.vim
search.cpan.org/dist/Kwiki-VimMode/
gaal cheers
audreyt may need some tweaking to work for Spork, as this is a kwiki plugin
gaal okies
audreyt just steal the to_html wafl
09:47 nothingmuch joined
nothingmuch audreyt: on feather, osdc/darcs_slides 09:47
09:47 TimToady joined
audreyt cheers 09:47
nothingmuch naturally it's a darcs repo
i'm darcs pushing cached version s of the images
so pull them in a few
this sucks, i wanted to go to the python tut =(
audreyt TimToady: I want you slides... $0/$1 gets confused a lot :)
TimToady audreyt: not surprising, considering the talk was originally written before the great $1/$0 shift... 09:49
nothingmuch webcal://ical.mac.com/nothingmuch/OSDC::Israel::2006.ics
םםפד
oops
audreyt nothingmuch: I got your slides
nothingmuch grazie
darcs pull again
the image cache dir is in 09:50
audreyt nothingmuch: put it in your public_html
otherwise how do I pull
nothingmuch darcs pull
audreyt via ssh?
nothingmuch oh
yes
one sec
okay
public_html/osdc/darcs_slides 09:51
pasteling "audreyt" at 192.118.100.254 pasted "Reworked S06 macro patch: require COMPILING <$x>, q:code(:COMPILING), etc." (62 lines, 2.4K) at sial.org/pbot/16025
audreyt TimToady: the patch above (based on nothingmuch's) should summarize the separate subchunks we produced in the hackathon 09:52
nothingmuch this is based on my patch, right? i don't need to throw it away from svk
(i mean, will it merge correctly?)
oh wait
no it won't
sorry
audreyt: got the image cache?
audreyt nothingmuch: HTTP here blocks .gz 09:54
I'll just rsync
otoh, HTTP here supports DAV
nothingmuch fsssscccckkkkkkkkkkkkkkkkkk11112$!$%!!!!!!!!!!!!
audreyt so SVN>Darcs here
:D
nothingmuch heh
ssh it then
they opened that up
audreyt doing so
nothingmuch that's why i couldn't download that file
assholes
i hate firewalls, and even more so the admins that abuse them 09:55
TimToady audreyt: looks good to me. The require is fancy, but kinda "use vars"-ish when my $x := COMPILING::<$x> would probably be what most people would use. 09:57
audreyt TimToady: the generalization here is
=head1 Importing from a pseudo-package
You may also import symbols from the various pseudo-packages listed in S02, 09:58
which behaves as if all their symbols are in the export list:
# same as my ($IN, $OUT, $ERR) ::= ($*IN, $*OUT, $*ERR)
use GLOBAL <$IN $OUT $ERR>;
(if okay with you, this goes into S11)
TimToady Yes, but what's a pseudo package? I think any require has to know whether it still needs to slurp a .pm, and these just happen to be born unslurpable.
audreyt as you cannot declare a module with that name and expect it to work 09:59
so the only part that needs to be speccted is that ::GLOBAL::EXPORT is always just ::GLOBAL
TimToady Yep. It basically looks fine to me, and as long as you understand the require requirements, I don't mind if it's somewhat terse. 10:00
10:03 azuroth joined
TimToady wonders what he should say about p5-to-p5 on Tuesday... 10:05
Isn't there a cheat sheat in pugs somewhere that lists p5-to-p6 transformations?
pasteling "audreyt" at 192.118.100.254 pasted "S11: Importing from Pseudo Packages" (26 lines, 858B) at sial.org/pbot/16026 10:06
audreyt TimToady: looks good?
TimToady: docs/other/porting_howto
but they are somewhat outdated 10:07
(and needs to be ported to docs/Perl6/Perl5/Porting.pod)
anyone up for such a task? :)
10:08 larsen joined
TimToady audrey: s/attempt to load C<GLOBAL.pm>/attempt to load, for example, C<GLOBAL.pm>/ 10:12
Other than that, looks fine.
audreyt TimToady++ # committed 10:13
mauke good morning 10:14
10:15 chris2 joined
audreyt hm, bsb is not here 10:15
mauke is svn pugs broken for anyone else?
audreyt no... how come? 10:16
nopaste the error?
mauke rafb.net/paste/results/BaXrPv32.html
integral is that building from clean with the same PUGS_EMBED options that you normally use? 10:17
mauke that's with PUGS_EMBED=perl5 10:18
audreyt that looks incremental
I'll attempt a rebuild to confirm
mauke I can make clean and run it all again but it'll fail at the same point
audreyt hm, something tripped _stub.o generation 10:19
can you try
./Setup configure
./Setup build
?
(it builds here)
mauke sec, rebuilding from make distclean now 10:20
audreyt sorry for the pain 10:21
try unoptimised
it's really fast now
as opposed to thrashing on .Run
mauke hey, that's what my new pc is for :-) 10:23
10:24 anatoly joined
audreyt anatoly: hey 10:24
TimToady I just know someone's gonna suggest that require GLOBAL <$IN> be made mandatory rather than falling back to $*... 10:25
audreyt yeah, that's the logical next step :)
though I'm working on my talk, so that's gonna wait
use GLOBAL <$IN> # works too
anatoly hey audreyt 10:26
audreyt anatoly: sanity-check xrl.us/j7no ? :) 10:27
10:28 GeJ joined
gaal yo anatoly :) 10:28
anatoly yo gaal :)
audreyt: sure, looking 10:29
gaal audreyt: how do I svk praise, but look back into the parent repo?
arcady mauke r9188 unoptimized compiles just fine here
audreyt gaal: svk praise --remoterev /pugs/trunk/path/to/file 10:30
gaal thanks
audreyt np :)
mauke audreyt: ./Setup build seems to succeed (last line: /usr/bin/ar: creating dist/build/libHSPugs-6.2.11.a) 10:31
audreyt mauke: okay...
then "make" again (without clearing) 10:32
I think it's just a bad incremental build startpoint
arcady incremental builds don't always work so well...
mauke failed
audreyt mauke: hunt for 10:33
$(RM_RF) dist/build/src/Data/Yaml
mauke $ find . -name Syck_stub.o
audreyt in Makefile
mauke ./dist/build/Syck_stub.o
audreyt and remove them
hmmmm
mauke the file is there, it's just not in the right directory
audreyt have you upgraded Cabal?
I see.
mauke no idea
looks like I have dev-haskell/cabal-1.1.3 10:34
audreyt yeah, we only tested on 1.1.2 and 1.1.4 I think, hmm
it's like the old MakeMaker days
my talk is coming up 10:35
but I'll resume fixing it
meanwhile
bbl &
TimToady audreyt: you plannin' to eat lunch?
gaal TimToady: did you pick up vouchers? 10:37
for lunch
TimToady Yep, got vouchers.
anatoly ah, audreyt's gone. gaal, how long is her haskell talk supposed to be? :) 10:38
gaal anatoly: 60 min 10:39
anatoly haha :)
gaal TimToady: audreyt's helping nothingmuch set up the laptop 10:40
anatoly: that's fine, under the AudreyT the audience is overclocked as well
anatoly gaal: it is, but not nearly enough to grok these 275 slides. That's one slide every 15 seconds. 10:42
gaal TimToady: <audreyt>: TimToady: yes 10:43
10:43 clkao joined
gaal anatoly: <audreyt>: that's plenty 10:43
clkao oops. jsperl5 prelude no longer compiles
pugs: src/Pugs/AST/Internals/Instances.hs:(684,4)-(691,63): Non-exhaustive patterns in function showPerl5
10:46 levin___ joined
TimToady well, I'm using up battery, so I better shut down and go to lunch. 10:46
Oh, it's not lunch time yet... 10:47
anatoly audreyt: great slides. if you're looking for errors and typos, I didn't see any. If for suggestions, best thing I could think of is putting laziness, lazy lists, fibs and maybe a bit more about laziness right after referential transparency stuff and before monads 10:49
audreyt: then you have something to keep them going, as it is when you reach monads they still have no clue what good any of this is for, and their brains shut down and never boot back up after that. 10:50
gaal poor sheep 10:57
anatoly: <audrety>: okay
TimToady: we're in nothingmuch's talk, to end 13:05 10:58
lumi 13:15 I think? 10:59
TimToady I'm out in the lobby, feel free to glom me en passant.
gaal ooh camel katamari 11:01
anatoly audreyt: for "more laziness", would be nice to show a few slides on how a lambda is passed around and only executed lazily when needed. examples of map with toUpper etc. are not enough because they're felt to be builtin stuff which is allowed magic. the general power is not felt. and if you have a lambda being passed around, there's the great place to stress how it's a closure like perl's anon sub holding a my var, only much more general :) 11:03
gaal lumi: i think today's schedule is different by 10 min from tomorrow's 11:07
lumi oic
Although my model seems to fit reality better 11:10
gaal apparently the printed sched needs adjustment as well, though perhaps -- too late 11:12
12:10 mauke_ joined 12:20 ilya666 joined 12:26 mauke_ is now known as mauke 12:30 TimToady joined 12:34 KingDiamond joined 12:35 elmex joined 12:40 xinming joined 12:41 rantanplan_ joined 12:44 gaal joined
gaal rehi 12:44
audreyt's giving her haskell talk
12:46 nothingmuch joined
nothingmuch lumi: ping 12:46
12:47 lypanov joined
lypanov nuts 12:47
gah
lypanov must stop typing while joining
12:48 levin__ joined 12:49 xinming joined
anatoly rehi 12:49
how's the talk going?
nothingmuch i'm learning python 12:51
anatoly hey nothingmuch
nothingmuch audrey's talk seems like it's going to be wonderful
gaal anatoly: going well, we're in QuickCheck already
nothingmuch (i had a rehearsel)
anatoly how was darcs received?
lumi I\m learning QuickCheck
nothingmuch quickcheck?
wtf? i didn't get to see that
lumi No? 12:52
It's the opposite of HUnit :P
gaal nothingmuch: tests w/random test cases
nothingmuch i guess the rehersal was cut short ;-)
i know what it is, i just didn't see it in the slides
gaal we skipped a few cool monads
but are entering Parsec now
anatoly :)
nothingmuch this talk is very amusing
some java guy told the speaker:
"it's stupid! why are there no copies in your stupid python thing! this is very non standard" 12:53
(python defaults to binding for assignment, it seems)
gaal no copies?
oh
lumi Java defaults to binding for assignment as well
gaal how'd the speaker resond?
nothingmuch ofcourse, there *are* copies
lumi Reference semantics
nothingmuch he is ast the ignoring step
gaal denial!
bargaining!
anatoly "java. where every man _is_ an island"
nothingmuch the *other* members of the audience told the java guy to shut up =)
haha
ast = at 12:54
gaal even if he does say jeh***<boulder>
lumi Is it aan ignoring step, or phase?
nothingmuch phase, i guess
he argued before
then he said "ok, whatever"
and at the "your stupid python" event he just didn't naswer
this is second-row-guy syndrome, only much worse 12:55
lumi Learn anything new yet?
nothingmuch ?
lumi Of Python
nothingmuch yes
the higher order help function
it takes any function and prints it's help
we talked about doing this in perl 6
as a trait
or attr 12:56
sub foo : doc("....") { }
lumi Yeah, ST has that
gaal or $=POD
modulo twigil?
nothingmuch $=POD is lexical data
attr is semantic
i guess we want sub foo : doc($=POD) or something, right?
or an easier macro
or is $=POD smarter than that?
gaal that would make it top-heavy, on the syntactic level it can't be so unweildly 12:57
nothingmuch like &function.pod ?
yeah, i guess
they have it as a void context string as the first expr
lumi sub add($x,$y) is doc "I add two numbers" # ?
nothingmuch i guess we could do the same if pod directives are rvalues
gaal ingy had some ideas for Perldoc
nothingmuch lumi: same thing... it was all discussed
ingy has a lot of ideas ;-)
gaal yay for STM in perl 6 :) 12:58
nothingmuch yayayayayayay!
lumi I like SmallTalk's Fight Club style comments
nothingmuch if you got to STM in that talk
gaal just did
nothingmuch and audrey's slides/speaking is as clear as it was in the rehersal than she is an even better speaker than I thought
then again, i didn't get to see takahashi employed in practice, just in theory/video 12:59
the sxip guy had a lovely takahashi-esque talk
on identity 2.0 or something like that
gaal it is -- though i wonder what it's like for people who've never seen haskell
nothingmuch gaal: from the way bindings/monads/laziness were explained it sounds good
i was a bit weary of pattern matching
anatoly nothingmuch, that was a wodnerful talk yes, on identity 2.0
nothingmuch it was explained too vaguely IMHO
gaal takahashi looks much less shallow-populistic than i thought 13:00
nothingmuch anatoly: i wonder how much time he had to rehearse it
gaal: yeah... i thought so too
then i remembered that the speaker doesn't read out the slides
like audrey said yesterday, takahashi forces you to explain
anatoly nothingmuch: yeah.. it was pretty light on substance too when you thought about it. but great acting and setup.
gaal i'm more convinced black on white is much better than the reverse now
nothingmuch anatoly: not really... for someone without identity mgmt background (the target audience was non programmers, IIRC) i think it was pretty impressive 13:01
but i agree that for our kind, people who tend to generalize automatically it was pretty moot
anatoly nothingmuch: hmm... maybe if non-programmers, ok
gaal nothingmuch: toly's a mathematitian, everything he understands is trivial :)
nothingmuch gaal: it's more boomish
gaal: =)
anatoly gaal: shut up :)
nothingmuch anyway, that's why i messed up the perl slides
err, darcs slides 13:02
i removed all useless english scaffolds (grammar)
gaal first cell phone call in the day!
anatoly maybe it's me.
gaal people are well behaved here
lumi There was one earlier
gaal HSP is cool
lumi iirc
anatoly ah, you mean cellphone ring, not your cellphone.
nothingmuch the java guy had a phone call ;-)
HSP? 13:03
gaal heh
haskell server pages
anatoly yeah
nothingmuch ah.. right
never saw that
gaal quite spectacular
nothingmuch just heard of it
i should have stayed
i thought it was just a haskell tutorial
gaal but like a lot on the web it probably ends up sucking in practice
nothingmuch mutters something about Catalyst and how it doesn't ;-) 13:04
gaal she's now on the slide demonstrating gui and web have a lot in common
nothingmuch this python stuff is uninspiring
it's at the intellectual level of a language i would implement as a toy
gaal and how if you write in HSP you can do either with the same code (!)
nothingmuch except it's very broad
and a complete implementation
lumi Broad like an accent?
nothingmuch but it's really very.... err.... boring
ugh
i want to go back
gaal but of course you are limited to only the basic widgets
nothingmuch too late though 13:05
only 10 mins left
lumi The interesting languages are in the *other* room
gaal nothingmuch: work on your slides then
nothingmuch no, i'm tired
i need a bigger break
i'll start in the next slot
nothingmuch is still stressed
the talks make me really nervous
gaal the guy in front of me is running aptitude update from hamakor, heh
nothingmuch stage fright and all
hamakor have a nice mirror =)
it's pretty fast
gaal esp if you're on actcom ;) 13:06
nothingmuch python has an amazing itneractive shell though 13:08
syntax highlighting
readline support
function autocompletion
intellisense
syntax *ERROR* highlighting
lumi Steal that, then :P
nothingmuch (very clear)
yes =)
i think i will download a dashboard widget with a timer 13:09
and call it talkdoom
i.e. the point in time by which your slides have to be ready
mauke OCaml highlights errors, too
anatoly bbl 13:12
nothingmuch one nice thing about python the language is that so far (i don't know how synthetic this is) most of the examples were not very ugly 13:13
that is, even when people asked simple questions, the examples looked pretty good
in perl 5 that doesn't always happen
in perl 6 otoh, it's beautiful, but people won't understand it immediately
in python the only one way to do it seems good enough for most things, if it's sugary
and if not you can also use the other (dumb) way to do it, i guess, but it'll be easy to explain 13:14
i'm surprised that haskell doesn't highlight in ghci 13:15
lumi There's no thesis in it 13:16
gaal everything will come out whatever Function is colored as :)
nothingmuch ?
oh
err, don't be silly =) 13:17
we have haskell vimmode, that's good enough
lumi Yeah, there's types, too
nothingmuch and do
gaal colorcode every variable by its type?
nothingmuch i have a nicer idea
color code every variable based on it's lexical scope
like a map colorer
gaal that does sound useful, yes 13:18
nothingmuch so that if there is a binding, it's colored the same way as all it's uses
but if it's masked, it'll be a different color
hmm
i guess, i gave python it's chance, i don't think i'll study it
is this fare to diss it now?
fair
gaal well done, audreyt 13:21
13:23 KingDiamond joined 13:25 xerox joined 13:31 bsb joined
audreyt gaal: thanks :) 13:32
13:32 nothingmuch joined
nothingmuch hola 13:32
ilya666 Finland -- Sweden 1:0 after first period 13:33
audreyt mm, this FUSE talk is actually interesting 13:34
too bad I don't get to use it 13:35
(yet)
integral do we have a pugs/p6 FUSE binding? 8-)
audreyt SyntaxNinja has a Haskell binding
so we should really revive use haskell:Module
nothingmuch =) 13:36
audreyt (or use c:header)
xerox Which FUSE talk?
nothingmuch wow, that would be a slick FFI interface
my dermatitis is erupting =(
audreyt xerox: we are in osdc.org.il/schedule.html
nothingmuch stress of talks is != good
xerox Ah, okay.
lumi Ah, you write PHP in lisp! 13:37
nothingmuch yes =)
lumi and I are in the codegen talk
lumi Nonono, never touch generated code by hand! 13:38
nothingmuch i agree
but he said something slightly different
he said "my generator is not good enough some of the time"
"it's too much effort to fix it"
"i'd rather fix it by hand when it breaks, which is not often"
gaal so, can we mount perl6 modulespace via a fs implemented in p6? :-P
nothingmuch yes 13:39
lumi Okey, but then you need to read its output, every time, instead of getting it right once
nothingmuch the module autoloader from CPAN thing could be just a virtual file system in Perl 6
lumi Heh
gaal use Meta;
nothingmuch lumi: yes, but his point is good: it's good enough, no need to push it
if it was *really* a problem, he would fix it
at least that's what he implied 13:40
(like, an in practice problem)
lumi There's "indent" everywhere.. He should read the Haskell Pretty
nothingmuch =)
we'll let him know
gaal has anybody seen Ran around? I just noticed he has a talk today (yay)
nothingmuch maybe he's high
gaal (ran's talks)++
lumi nil
nothingmuch yeah, they rock
13:45 wilx joined
lumi Generated JS can be size-optimized (translates into $$ not paid for bandwidth) 13:49
gaal $$ means vretical cat in Pretty :) 13:54
lumi I like cats in any orientation 13:59
audreyt mm landscape cats 14:00
nothingmuch =)
50 uses for a dead cat: pretty printing
(spread vertically)
lumi Heh
Ew
You just apply the cat virtually to two docs?
Vertically (sorry) 14:01
nothingmuch sorry 14:02
yes
that's what the butcherknife operator is for
lumi Ew -> Ew 14:03
nothingmuch or the schnitzel hammer one: []---
xerox schrƶendinger's cats
gaal my first guess when I saw $$ was that ($$) f = (iterate f) !! 2 14:05
xerox heh
Maybe ($$) = deepSeq
gaal ($$$$$$$) = y -- dammit 14:06
fuse is really nice
xerox gaal, ($$) f = (!! 2) . iterate f
audreyt hm, you mean (!! 2) . iterate f
xerox Dang.
gaal argh
audreyt heh
gaal well in that case, pl it 14:07
audreyt they even align
xerox Indeed :-)
xerox high-fives audreyt
audreyt smiles
gaal: ((!! 2) .) . iterate
gaal shares the high-five in realspace
xerox It is a pattern. 14:08
To compose a unary function with an n-ary function, you compose n (.) together.
gaal is reminded that we need a bot here
whee!
xerox++ -- haskell carg^H^Hpatterns
audreyt mm, cargo patterns 14:09
xerox Yay.
(What would 'carg..' mean?)
audreyt cargo culting^W^Wpatterns 14:10
gaal: how's your talk?
gaal um, ah, eh, oh...
slide 14 :)
xerox "The Cargo Cults believed that manufactured western goods ('cargo') were created by ancestral spirits and intended for Melanesian people."
gaal i should go to more boring talks 14:11
so i have time in vi
xerox: Feynman coined that, I believe
audreyt gaal: maybe we move out from the talks after this one
so I can _also_ work on my talk tomorrow
of which I've produced 0
xerox gaal: whoa! Do you have any reference about that? 14:12
audreyt and is keynote
gaal xerox: it's in Surely you're Joking, I think
xerox "ncidentally, Richard Feynman wrote a great essay on cargo cult science."
+I
audreyt sees bsb looking at her slides again and smiling mysteriously in her front row
gaal my $talk is keynote;
well
audrey's talk is
not mine happily :)
we can add a scoping keyword that only installs a symbol in somebody else's pad 14:13
their $talk is keynote;
audreyt my $talk is keynote is export(:THEIR);
xerox fails to decypher 14:14
gaal that also leaves $talk in this scope :)
bsb I was just wondering how the hell <img src=a /> works as Haskell code
gaal bsb, preprocessors :)
azuroth legal_alcohol++ 14:15
lumi I guessed preprocessor.. 14:16
gaal yay, there's ran
ilya666 Finland -- Sweden second period 2:2 14:24
14:24 cognominal__ joined 14:26 KingDiamond joined 14:36 nothingmuch joined
nothingmuch audreyt: ping 14:36
wrt scrhodinger's cat: between hitting it and removing the blood from your eyes you don't know if it's alive or dead
until you make an observation ;-) 14:37
gaal nothingmuch: she's w/me working on slides
nothingmuch sah
ah
gaal but is disconnected
nothingmuch where are you guys?
gaal if not urgent, later :)
one floor up
nothingmuch hmm
i'll join you, my slides are related
(this is for gaal's talk, right?)
if you don't mind ofcourse
gaal cool 14:38
nothingmuch it's just that I was going to ask for help myself
okay
see you in ~10 mins
gaal k i really need to concentrate
nothingmuch /devoice nothingmuch 14:39
14:44 michimueller joined 14:49 KingDiamond joined
ingy hi nothingmuch 14:52
15:02 hoowa joined
hoowa tang u still here? 15:04
Juerd "OpenOffice has *multiple thousands* of regression tests", the guy says :) 15:24
My guess is that Pugs wins easily ;) 15:25
xinming hmm, class A { has $.a = 3; method a { return 5 }; method t1 { $.a }; method t2 { $.a() } }; will there any difference between t1 and t2? 15:38
15:45 orafu joined
TimToady xinming: those should be equivalent. All $.foo are virtual accessors. 15:46
xinming TimToady: hmm, so, there is no way to get the attribute value $.a? 15:47
TimToady sure, say $!a instead if you really mean the attribute. But if you're going to do infrastructure like that, you should probably be using a submethod instead. 15:48
xinming thanks.
audreyt TimToady: nothingmuch raises this point: 15:49
15:50 bsb joined
audreyt my &code := sub {...} # see, no semicolon 15:50
my $var = %foo<bar>; # this is mandatory
my $var = %foo{'bar'} # can we make this mandatory as well?
I think only closing '}' for circumfix gets the EOL treatment, not postcirfumfix
TimToady Hmm. well it *is* a closure, if you use it like one... 15:51
audreyt %foo{my $x;}
TimToady you can say %foo{ my $x = $y } for instance.
audreyt indeed
so maybe that is all well after all
TimToady the tricky thing will be if someone says 15:52
%foo{$x}
+ 42
audreyt what about the "use GLOBALS"-under-use-strict-so-we-get-$x-always-lexically-bound idea? :)
mm, but actually 15:53
TimToady I've probably had enough time to get used to it. :-)
audreyt ooh :)
TimToady++
my &code := sub { $x }
.assuming();
same problem
so probably same solution? 15:54
(i.e. both gets implicit EOLed)
TimToady at one point I was reserving _ for that, so %foo{$x}_ would eat whitespace. 15:55
audreyt that's exactly like VB
clkao audreyt: got time to see why jsbackend can't compile prelude anymore?
TimToady it it?
is it?
audreyt clkao: will do after tomorrow's keynote of mine
clkao :)
audreyt TimToady: pugscode.org/osdc/vb.xul?page=41 15:56
100% identical :)
15:56 KingDiamond joined
audreyt lumi: .xul and safari does not mix 15:57
Juerd 17:03 < audreyt> my $var = %foo{'bar'} # can we make this mandatory as well? 15:58
TimToady audreyt: I'm not sure VBishness is a recommendation... :/
Juerd You kind of need that, for postfix if, right?
Isn't }-implies-; more useful if the corresponding { wasn't on the same line? 15:59
Oh, I get another blob in now (connectivity here is very, very bad) 16:00
You already discessed this...
I don't like _.
(at all!)
Juerd would, to be honest, prefer to always type ;, if the alternative is to in surprising cases have to type _. 16:01
lumi Sacred code?
audreyt but I dislike the A03 lonely-curly-in-a-line idea 16:06
as it will require the parser to remember things for each line
i.e. whether it has seen some token already for this line 16:07
as compared to EOL impliedness, which is only referring to forward information
(and hence good for one-passness)
16:08 nothingmuch joined
audreyt Juerd: the implied-; is useful for 16:08
sub f { 3 }
even though they were on the same lnie
16:08 nothingmuch joined
nothingmuch my conclusion about ran: 16:09
either he knows everything
in advance
or thinks really fast
that was a very elegant definition
"meta configuration"
and it took him 3 seconds and no retries
gaal he did think of that one
in advance
nothingmuch ah
oops 16:10
gaal i really like how he presents stuff
nothingmuch iis dyslexic, please remember that reading is not a natural process for $?SELF
gaal oh, I know that only because I remember discussing it with him a few months ago ;-)
16:11 justatheory joined
nothingmuch hehe 16:11
audreyt can't help but think of Harmony, which can solve this elegantly in a one-pass fashion 16:14
TimToady If we really wanted to drive people nuts, we'd say } assumes ; only if the next line is not more indented than this one. 16:15
audreyt alliance.seas.upenn.edu/~harmony/cg...n/demo.cgi
TimToady: aw, I prefer layout to be a separate pass
having to remember how each line indented is a pain in an onepass scanner
(just ask ingy) 16:16
nothingmuch mooseness is good
the moosness property pattern is like the DWIM language
you should be able to say
foo.property = moose in your cfg
and get DWIM
pasteling "audreyt" at 192.118.100.254 pasted "S03: Make non-qualified variables lexical-only by default." (44 lines, 1.5K) at sial.org/pbot/16030
audreyt TimToady: looks good? 16:17
s/limit the lookup the/limit the lookup to the/
nothingmuch isn't /}\s*$/ ending a block eq '};', and '}\s*$' ending anything else eq '}' ?
i mean, blockwise is the only way I'd expect it 16:18
it's a simple rule
(for humans)
what I'm afraid of is that when it does bite someone, you need to explain why it did that
and indentation sensitivity, or anay other heavy DWIM is.... err... hard to explain 16:19
especially if you use e.g. perltidy
audreyt the current rule is that it adds a semicolon iff a semicolon can terminate a statement there.
i.e. if we are in a statement-list position.
nothingmuch (although arguably perltidy with Perl::Grammar is easier to write in a way that is not going to screw stuff up)
hmm
that's a good DWIM, i guess
audreyt glad you agree
TimToady so you can always parenthesize to hide the DWIM. 16:20
nothingmuch i.e., try to parse '}' as '};', and as long as the next statement (till a real ';') is OK don't backtrack?
audreyt yup, which is probably the most inexpensive disambiguator
nothingmuch yes, but probably good enough 16:21
since we assume caching and separate compilation anyway, if we're going to be fast and stuff
audreyt nothingmuch: more like, don't do implicit-semicolon if we are inside some expression-level bracketing construct.
(which can be succintly defined in both Parsec and OpTable)
nothingmuch (optimizing is probably going to be much more expensive than this DWIM, but Perl 6 almost *requires* it based on it's arcitecture) 16:22
hmm
fair enough
sounds easier to implement
audreyt it better be :D
TimToady I'm not really concerned with parse speed--I want it to look clean (on average)
nothingmuch it still means the same thing
TimToady: we didn't expect otherwise ;-)
audreyt nothingmuch: crucial difference is that this is determined at semicolon-newline site 16:23
so it can commit there
instead of waiting for some backtrack
nothingmuch yeah
okay
audreyt which is extensive as it'd set up a prompt
s/extensive/expensive/
nothingmuch i'm thinking this:
';' is a disambiguation
it says "this is definately the end of stmt"
do we have the opposite? 16:24
"this is definately not the end fof stmt?"
does it even make sense?
audreyt that's what _ would be
nothingmuch i'm not sure...
audreyt but I agree it looks a bit too vbesque
(VB, btw, is as of 2005 finally parsable by a one-pass BNF.) 16:25
nothingmuch . o O ( i hope they don't propose \\\n
audreyt )
lypanov hehe 16:27
16:27 kanru joined
TimToady audreyt: snippit looks good--do we need to find the fallback verbiage elsewhere and countermand it? 16:29
audreyt it's on the para directly above it
TimToady are people going to the cafeteria again for dinner?
audreyt so I think it's safe
(already looked elsewhere)
I'll look again at dinner
TimToady audreyt: nothingmuch informs me that there's wifi in the restaurant. (and food) 16:30
nothingmuch recants
audreyt ok. I've committed the S02, and we can go now
TimToady I'm probably thinking of Apocalypse $IN then. 16:31
16:41 kanru joined 16:46 Limbic_Region joined 17:01 elmex joined 17:09 marmic joined
Limbic_Region hmmm - anyone know why every program I run results in "user error (Error loading precompiled Prelude: Left "syntax error: line 2, column 2" 17:38
then it reloads from source and works fine?
17:41 hcchien joined 18:08 anatoly joined 18:13 Arathorn joined 18:15 cognominal__ left 19:28 gaal joined
gaal Limbic_Region: can you nopaste the first lines of src/Pugs/PreludePC.yml 19:33
?
Limbic_Region sure 19:34
perlbot nopaste
perlbot Paste your code here and #<channel> will be able to view it: sial.org/pbot/<channel>
pasteling "Limbic_Region" at 24.35.57.240 pasted "First few lines of src/Pugs/PreludePC.yml" (59 lines, 1.3K) at sial.org/pbot/16037 19:35
gaal looks okay 19:36
Limbic_Region this happened after a nmake realclean; svk pull; perl Makefile.PL; nmake 19:38
though the svk pull did result in over a 100 revision updates
Limbic_Region ducks
gaal sorry, no immediate idea why this should happen and we're still all writing slides 19:40
Limbic_Region okies 19:41
19:48 nothingmuch joined
gaal nothingmuch: yo 19:49
nothingmuch yo yo yo
gaal it's faaaat albert
nothingmuch what was the mc joke?
19:50 lypanov_ joined 19:51 lypanov left 20:08 lypanov_ is now known as lypanov 20:09 salty-horse joined 20:10 lypanov joined 20:11 justatheory joined
xerox audreyt: how is it called that tool to do XML inside .hs ? 20:13
gaal HSP? LINQ? 20:17
xerox Hmm.
...the one that lets you do f <IMG ..> = .. 20:18
gaal xerox: let me peek at the slides to refresh my mem
20:18 justatheory joined
xerox Danke. 20:18
20:18 salty-horse left
gaal XTC? 20:21
xerox Any url? 20:26
gaal it's hsp that does it. www.cs.chalmers.se/~d00nibro/hsp/thesis.pdf 20:27
omg he was like asking me for a citation and I was like go read this thesis lol
xerox Thank you.
20:33 mj41_ joined 20:48 macli joined 21:40 r0nny_ joined 22:07 _SamB_ joined 22:14 nothingmuch joined
nothingmuch does anybody want to help me group thing? 22:15
think?
avar the borg?;)
22:15 elmex joined
lumi Group think what? 22:17
nothingmuch you're awake =D 22:21
my slies, ofcourse
slides
lumi I'll help with your sly slides 22:23
azuroth a horse, of course 22:24
22:25 hew joined
nothingmuch okay 22:25
lumi: did you get darcs at the end of my talk or before that? 22:26
;-)
hew Reading over [EAS]5, I'm wondering for a grammar like: 22:33
rule pattern { <alternation> }
rule alternation :w { <conjunction> [q(|) <self>]? }
rule conjunction :w { <catenation> [q(&) <self>]? }
rule catenation :w { <term>* }
22:33 nothingmuch_ joined
nothingmuch_ oops 22:33
i pulled out the cable
hew How would I flatten the degenerate(but common) case of 'term term' to just Pattern=>{Catenation=>[term, term]} 22:34
22:34 nothingmuch_ is now known as nothingmuch
lumi I had darcs earlier 22:39
So I'm in category one, yes? 22:40
22:40 macli joined 22:50 chris2 joined 22:52 larsen joined 22:59 diotalevi joined 23:06 calanya joined, Shillo joined 23:18 rhesa joined, rhesa left, calanya left 23:42 macli joined 23:48 macli left
elmex will perl6 be delivered with hurd? 23:58
obra Definitely not. 23:59
Perl 6 will ship at least a year before Perl 7.
Hurd won't ever ship ;)

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!