svn switch --relocate svn.openfoundry.org/pugs svn.pugscode.org/pugs/ | run.pugscode.org | spec.pugscode.org | paste: sial.org/pbot/perl6 | pugs.blogs.com
Set by avar on 16 November 2006.
00:08 larsen_ joined 00:32 mjk joined 00:41 UWC is now known as UWC_ 00:43 UWC_ is now known as UWC 00:55 thepler joined, polettix joined 01:18 thepler joined
TimToady audreyt: was shocked will reading the events/threads paper at www.seas.upenn.edu/~lipeng/homepage/unify.html to find a diagram of my sandwich parser in the middle. only they'd mislabelled the top-down parser as "threads", the token parser as "events", and the opaque operator precedence parser in the middle as "scheduler". :) 01:25
lambdabot Title: Unifying events and threads
01:25 justatheory joined
Juerd Heh 01:26
TimToady s/will/while/
01:56 leed joined
pasteling "cmarcelo" at 200.232.236.201 pasted "build failure, missing instances?" (150 lines, 6.2K) at sial.org/pbot/22920 01:56
02:03 Caelum joined 02:08 Limbic_Region joined 02:15 GabrielVieira2 joined 02:19 dmq joined 02:29 GabrielVieira2 is now known as GabrielVieira 02:47 GabrielVieira2 joined 02:52 Schwern joined 03:01 Aankh|Clone joined 03:05 Squern joined 03:15 zgh joined 03:18 mako132 joined 03:33 UWC joined
audreyt @tell cmarcelo fixed thanks 03:45
lambdabot Consider it noted.
svnbot6 r15248 | audreyt++ | * Work around DrIFT's inability to deal with parallel array notation. 03:50
r15248 | audreyt++ | cmarcelo++ for reporting it.
03:51 Aankh|Clone is now known as Aankhen``
audreyt TimToady: as you have observed yourself in the Impossible Talk, it's but one instance of the universal trinity diagram... 03:51
03:54 poland joined
avar audreyt: do you know how to get %^H at the xs level? 04:42
audreyt PL_hints surely?
avar didn't know of that, I was trying to get_hv("^H" 04:43
audreyt ah. no. 04:44
avar PL_hints is $^H no?
audreyt wait... what are you going to do? 04:45
PL_compiling.cop_hints_hash ?
TimToady interestingly, parrot also wants to be async underneath. 04:46
audreyt the "get a hint" idiom is
Perl_refcounted_he_fetch(aTHX_ PL_compiling.cop_hints_hash, 0, "open", 4, 0, 0)
TimToady: yeah, but events are perthread least I heard
avar lexical imports in re::engine::Plugin don't work because I don't use %^H for sub callback lookup yet
peeking at cop_hints_vars
TimToady well, at least they profess to want continuations 04:47
and we can probably emulate the whole lower layer with a single event-driven thread, I expect 04:48
though perhaps not get multi-cpu benefit there
well, that's for later 04:49
audreyt yup.
TimToady and hopefully someone else. :)
audreyt so do you think it's sane to simply have
async {} # share everything in scope
fork {} # share nothing in scope 04:50
(except for those marked "is shared")
?
TimToady and ==> implies fork?
audreyt which ==> is that?
TimToady feed operators / pipes
audreyt I thought ==> carries no parallelism promises 04:51
TimToady the forgotten 3rd member of the parallism triumvirate
it's a pipe
audreyt I thought
@data ==> grep {...};
lambdabot Unknown command, try @list
audreyt is just sugar for
grep {...}, @data
and has no special parallelism semantics 04:52
oh that changed.
ok, hm.
TimToady S06:848 04:53
==> and <== now delimit statements
audreyt I think it's erlangish enough that we want share nothing.
i.e. process parallelism, not threaded parallelism 04:54
conceptually, unix pipes
TimToady Cell arch
audreyt yeah.
if so then maybe fork{} is not needed
fork() can just remain os-level fork
"heavyweight fork"
TimToady I think it would be confusing to use fork for internal 04:55
for the Unixheads
audreyt aye aye.
TimToady Q: one end or the other of ==> is rooted in the current space
maybe both in certain circumstances?
avar audreyt: sial.org/pbot/22922 <- this is what I'm trying to do
lambdabot Title: Paste #22922 from Someone at 85.197.228.236
audreyt current space is consumer
that's standard I think 04:56
TimToady but you'd kind like to loop it
audreyt example?
TimToady maybe as long as you feed it something lazy
avar I'e getting a HV* for %^H in XS for re::engine::Plugin lexical import
TimToady 1..* ==> filter ==> @result
audreyt TimToady: lazy structures does that implicitly
it's called, I believe, fusion
so the filter cannot write to result? 04:57
TimToady anyway, I'm tired of answering the question of why we can't have two-ended pipelines
no, I'd like to feed it from my current space and retrieve the result back. 04:58
as long as it's all "pull" semantics, should be okay, I think
and as long as the user promises
audreyt my @result; 1..* ==> grep { @result.push(...); ... } ==> @result;
TimToady that the feeder doesn't depend on the result
audreyt what's the semantic here?
the .push is noop?
or rather, process-local?
that is, doesn't affect the @result at the end?
TimToady not sure what's sane there, but I'm not against enforcing some discipline on ==> if necessary 04:59
audreyt I think anything it closes over gets cloned
and it remains local absolutely
aka erlang
only sane model I'm aware of
TimToady well, let's plan to stay sane 05:00
avar audreyt: simnet.is/boaesd/re-engine-Plugin-0.01.tar.gz # if you want to take a look
TimToady can maybe loosen up later
audreyt it also means we don't get those ithread pains
since the amount of freevar is encouraged to be small
OUTER var that is
and as long as it's small there's no cost
avar audreyt: the up-to-date example is t/methods.t. But It's not cpan-ready until I can at least get lexical imports to work
I'm not exepecting an answer back to these mublings b.t.w:) 05:01
TimToady ==> stuff is more for event layer than for thread layer
audreyt sure, which is why explicit sharing is good
avar sleep &
audreyt we can get as far as saying you need to explicitly lexicalize package globals 05:02
==> grep { our $PkgVar; ... } ==> ...
otherwise you don't get to see $PkgVar
(because otherwise we always need to clone the stash)
TimToady sounds about right 05:03
audreyt symtable that is
TimToady don't mind another real use for "our" either. :)
audreyt heh, I use our all the time... even in p5
in places where "use vars" won't work
TimToady I use it where "use vars" would work. :) 05:04
audreyt s/in/even in/
TimToady but the p5er's continually carp about it
"what does it do that 'use vars' doesn't do?"...
audreyt well p5 has this sublanguage in it (p4) that hates lexical scoping... 05:05
TimToady there's still a bit of it in P6 too. O_o 05:06
see "temp"...
but I'm hoping 'is context' helps get away from that.
05:07 lambdabot joined
TimToady I suppose a thread sees the context vars in its spawning scope... 05:07
audreyt I thought temp doesn't hate lexicals 05:08
it's orthogonal
unlike in p5
TimToady well, but people will use it on globals
and they'll think globals are all you can use it on.
anyway temp is much more problematic when multiple threads are trying to do it 05:09
so there's probably gotta be some restriction on temporizing things outside your current thread's storage, or some way of only letting one thread manage any global location. 05:11
heh, one could almost turn temporization into the locking mechanism.
assuming one needed a locking mechanism.
hmm, can't temp a global outside a contend {}, maybe. 05:12
but that'll really screw up the P4 programmers...
audreyt well p4 has nothing like async{} 05:14
and I'd be fine if temp in async setting is _always_ transactional
TimToady and let, of course. 05:15
audreyt yes.
TimToady I think it makes sense
audreyt TMVar, in GHC parlance
(transactional mutex)
since otherwise we can't reason about multiple dynamic scopes
TimToady so it really is more like "suppose $x = 2"
audreyt large bookkeepping overhead
but they deserve it. 05:16
05:17 luqui joined
audreyt luqui: greetings! 05:17
luqui hello audreyt
TimToady luqui: sorry, we just ran out of interesting topics. :)
audreyt actually I should get moving
luqui heh
audreyt ==> @parents
bbl
TimToady have the appropriate amount of fun 05:18
audreyt TimToady: would be nice to augment the S to say ==> clones closeover and you can't depend on any sharing _at all_
unless, of course, explicitly marked "is shared", not unlike liz's forks.pm
TimToady k
luqui: how are you doin'? 05:19
luqui bit sick 05:20
TimToady lotta that stuff going around
luqui also got a job...
TimToady not as a garbage collector, I presume
luqui no, but I applied for a job as one ('s author), didn't get it though 05:21
TimToady if that's the one the reference was for, they never asked...
luqui that was the one I wanted. I suppose I should count my blessings; working on a C++ compiler might not be as much fun as I thought. I mean, it's C++... 05:22
The reference just went on my resume. Even if they never asked, I'm pretty sure it helped.
TimToady dunno, they might have seen my name and puked
luqui (hmm, no bachelors degree, but he's actually done something, unlike these other candidates)
heh, maybe. I wouldn't be surprised if a C++ shop were full of zealots... 05:23
anyway
I like ruby sometimes, not others 05:25
luqui is struggling to do an Easy thing at the moment
TimToady would it be easy in P5? 05:26
or P6, for that matter? 05:27
05:30 luqui joined, lambdabot joined, Aankhen`` joined, TreyHarr1s joined, rfordinal joined, jdv79 joined, [particle] joined, kolibrie joined, spo0nman joined, stef_ joined, allbery_b joined, Khisanth joined, mugwump joined, integral joined, SamB joined, Odin- joined, SCalimlim joined, takesako__ joined, jamessan joined, Gothmog_ joined, LCamel_ joined, broquaint joined, nnunley joined 05:35 lambdabot joined
meppl gute nacht 05:37
05:42 buetow joined 06:14 BooK joined 06:35 justatheory joined 06:57 TimToady joined 07:00 MrTumnus joined, MikeJS left 07:17 jisom joined 07:37 dduncan left 07:48 TimToady joined
wolverian Juerd, the psycho test page doesn't come up on the first page on .fi google :( 07:58
08:01 poland joined 08:20 kanru joined 08:32 poland_ joined 08:45 devogon joined 09:10 elmex joined 09:16 iblechbot joined 09:28 Ziggy6 joined 09:55 MrTumnus left 10:03 Dr_Pi joined 10:05 marmic joined 10:26 Pat__ joined, GabrielVieira joined 10:27 VanilleBert joined
Pat__ hello 10:30
moritz hi Pat__ 10:31
Pat__ so what interresting/fun stuff happening in perl6 these days ? 10:34
I have been out of touch for a couple of years and just thought I'd catch up :-) Just for kicks.
moritz well, we can compile to javscript ;) 10:35
javascript
and pugs passes >95% of the test suite (but that number *might* be misleading) 10:36
Pat__ heheh yea I guess the 95% easy part is done
now the 5% hard part is left :-) 10:37
moritz ?eval [+] (1 .. 10)
10:37 evalbot_r15241 is now known as evalbot_r15248
evalbot_r15248 55 10:37
moritz some cool operators work ;)
Pat__ but about compiling to javascript where can i read more about it ... and playa little bit ?
Pat__ bows ot the eval bot :-) 10:38
moritz you can play at run.pugscode.org/
lambdabot Title: Run Perl 6 Now -- in your browser!
moritz and read... well... there is svn repository ;)
Pat__: but honestly, Im the wrong person to ask... I only joined the project a week ago 10:39
Pat__: and I didnt write a single line of code, just test cases, docs etc.
Pat__ heheheh well I did not specify english so I guess the svn repo is a fair answer :p
hey no problem, I'm just here to chat 10:40
no life saving decisions envolving perl6 (yet)
10:40 TimToady joined
Pat__ so thatks for your help 10:40
moritz ok, maybe somebody else can give you a more accurate summary
np :)
moritz is bored because my only HD died today, and all I can do is chat (booted from knoppix) 10:41
Pat__ heheh Everyone else seem so busy rushing perl6 out so door ;-) I don't want to bother anyoen :-) 10:42
moritz you exagerate... nearly everybody fancies a half-technical chat now or then ;)
Pat__ well things is I had almost forgotten about perl6 ... thinking well it'll be ready when it'll be ready 10:43
but then I just happened to read this steve-yegge.blogspot.com/2007/02/ne...guage.html 10:44
lambdabot Title: Stevey's Blog Rants: The Next Big Language
Pat__ and thought ... hey could it be perl 6
and here I am enquiring about the state of the perl6 union. 10:45
moritz well, it is active and progressing (I think thats important ;) 10:46
and of course it will be *so* cool once it is finished ;)) 10:47
Pat__ It is already too cool and it is far from being finished :-) 10:48
moritz but hey, it will be finished `til christmas ;) 10:50
(because when it`s finished, every day is x-mas ;)
Pat__ hahahaha
moritz not my idea dude ;) 10:51
I have to go and get my laundry, brb 10:52
Pat__ ok thanks fo rthe pointer ... I'll be playing around with the javascript thingy. Bye
10:52 Pat__ left
Juerd wolverian: It does on the .com and .nl 11:11
wolverian: Maybe there are too many .fi psycho tests. What a country! ;) 11:12
moritz or there are to many .fi psychos *eg* 11:14
;) 11:15
11:17 marmic joined
Juerd I tried to imply that :) 11:18
moritz after all HUMPPA! music comes from .fi, and Star Wreck and other cool but crazy stuff ;) 11:20
wolverian humppa :( 11:21
moritz wolverian: uc "humppa", I guess it's case sensitive ;) 11:23
Juerd There was someone in #perl yesterday who claimed they were parsing XML with regexes because their XML was... case insensitive. 11:25
Needless to say, this person eventually ended up on my ignore list.
moritz ;) 11:27
mugwump xrl.us/casesensitive 11:31
11:33 GabrielVieira2 joined
wolverian Juerd, the page has been spammed, btw :) 11:47
11:59 amosrobinson joined
amosrobinson hiya 11:59
moritz hi 12:00
amosrobinson how's it all going? 12:01
moritz my harddisk is dying... but that's probaply not what you wanted to hear ;) 12:02
amosrobinson eek 12:03
well that's OK, I'm partial to spreading bad news myself
moritz d'oh 12:04
12:06 prism joined
Juerd wolverian: Severely. I try not to care :) 12:06
amosrobinson I suppose, in backthought, that trying to understand the grammar while drunk is a bad idea 12:12
Juerd Is there anyone who plans to go to the German Perl Workshop this year? 12:13
'cause Mark Lehmann needs to be reminded of his promise to put the presentation "Was stimmt mit Parrot und Perl 6 nicht?" online :)
moritz where and when is it?
Juerd ("What's wrong with Parrot and Perl 6?")
moritz I'm back in Germany from july on...
Juerd moritz: Muenchen, 21, 22, 23 feb. 12:14
moritz Juerd: thanks, damn
Juerd Hm 12:19
I might be able to go there myself
$client has paid their invoice now
(Unexpectedly)
rafl: Where are you staying sleep-wise during the GPW? 12:20
12:27 ludan joined
ludan ola 12:27
Juerd Hi 12:29
12:32 VanilleBert left 12:37 b_jonas joined 12:43 explorer joined 12:52 pdcawley joined 12:54 ekelund joined 13:09 cmarcelo joined, buetow joined 13:28 elmex joined 13:35 web_knows joined 13:37 GabrielVieira joined
b_jonas /msg MemoServ DEL 1 13:40
13:46 Schwern joined 13:50 chris2 joined 13:59 Ziggy6 left 14:01 stephang joined 14:04 Squern joined 14:47 cnhackTNT joined, marmic joined 15:10 rashakil_ joined 15:13 explorer joined
TimToady Just got a note from audreyt++ that she is hospitalized with serious hepatitis. 15:20
Juerd Damn. 15:21
15:21 web_knows left
rafl Juerd: youth hostel muenchen neuhausen 15:25
Juerd rafl: There's a small chance I'll join you :) 15:26
rafl Juerd: cool!
How about fosdem?
Juerd rafl: I plan on going to fosdem only the second day
My office is rewired the 24th
That also means that if I'm going to the GPW, I have to leave during day 3
And drive 10 hours. 15:27
rafl ok
Juerd "There are no parking facillities directly in the youth hostel grounds, however there are a number of free parking possibilities in the direct vacinity of the youth hostel." 15:28
Hmm
I wonder how direct that is :)
b_jonas TimToady: that sounds bad 15:29
Juerd rafl: Do they have wifi at that hostel?
rafl checks 15:30
Juerd: their website says: "internet access nearby", so I guess they don't have wifi 15:31
Juerd Darn
That'd mean I have to do all work stuff from the workshop venue
Why did you pick this hostel? :) 15:32
15:33 prism joined
Juerd Oh, it seems to be close to the venue 15:35
Hm, € 30 per night :( 15:38
15:52 pdcawley joined 15:54 chris2 joined
Patterner I plan to go to the GPW, but as long as I don't have a confirmation I won't book flights and hotel... 15:55
Juerd rafl: There are hostels that charge half that amount, and provide free wifi and breakfast... 15:56
Patterner: Where would you be flying from? 15:57
Patterner Juerd: Hamburg
Juerd Patterner: No car?
Patterner Nope. Flying is cheaper and faster.
Juerd I looked at flying options. For me it would be faster, but much more expensive. 15:58
And it has the cost of inflexibility.
rafl Juerd: oh no.. I didn't search for a hostel myself. My co-travelers did.. :-(
Juerd Driving there and back again would cost me approx 80 € for fuel.
rafl: Have you already booked?
rafl Juerd: yes 15:59
Juerd rafl: Can you still cancel? :)
Patterner 80-100 Euro for both flights... 1 hour flying time :)
rafl Juerd: I think so..
Juerd rafl: Talk to your cotravelers then :) 16:00
www.hostelworld.com/availability.ph...unich-4274
rafl Juerd: will do :-)
Patterner Bring a whip :)
Juerd www.hostelworld.com/availability.ph...unich-8249
Patterner: "free wifi in lobby" should suffice :) 16:01
As should "half the price"
And it's not that much further away from the venue
And it's closer to the Hackerbruecke. Hacker bridges are always good. 16:03
maps.google.nl/maps?f=d&hl=nl&a...5&om=1 16:04
16:06 justatheory joined
Juerd Some negative comments about the paid-for internet access, weird. 16:08
www.hostelworld.com/availability.ph...unich-8249 # better ratings 16:10
Same street :)
16:22 thepler joined 16:39 cnhackTN1 joined 16:44 larsen_ joined 16:58 cnhackTNT joined 17:10 _marco joined
svnbot6 r15249 | kudra++ | Added a note about fetching posts 17:43
17:47 putter joined 17:57 xinming_ joined 17:58 buetow joined
putter :( 18:03
18:04 justatheory joined
putter audreyt: O O 18:07
audreyt: O@O O@O
audreyt: O\/ O
audreyt: \// Get well soon. 18:08
(fixed with font required)
18:13 Schwern joined 18:25 gnuvince joined 18:34 thepler joined
nothingmuch yo yo yo homies yo 18:42
nothingmuch actually saw some OGs at the train station
note to self: get timberland boots and an ugly biege parka
dmq and a gun?
nothingmuch no, my assignment is simply to know how to be an OG 18:43
not to go around OGing people
18:43 poland_ is now known as poland
nothingmuch (also, sardonic tone was lacking from that comment-- those gangsta's looked anything but original ;-) 18:43
dmq heh 18:44
nothingmuch my first thought on seeing the second one was "is that a different coat?" and then i realized it's a different human altogether
dmq "human" :-) 18:45
nothingmuch freud might have something to say about why I dind't think of using "person' ;-)
dmq might?
nothingmuch well, he's dead now
dmq ah, that type of might. :-) 18:46
nothingmuch hehe, i stole his ethernet ;-) 18:51
18:53 Khisanth joined
avar yawn 19:00
19:02 Squern joined
nothingmuch avar: for the past week or so dipsy has kept claiming we are one... how do you feel about this? ;-) 19:08
avar nothingmuch: o/~ we must never be apart o/~ 19:10
dmq . o ( an interesting solution to "there can only be one") 19:11
19:12 xwrn joined 19:15 cmarcelo joined 19:39 jisom joined
putter is there any p5 construct which can take multiple statements and grab the value of the last (eg, do{...}) _without_ introducing a lexical scope? 19:45
actually, what really matters is 'without introducing a scope for an local()s among the statements'. 19:46
s/an/any
sigh. in p6, this will be a small macro. in p5... 19:54
putter goes to explore cpan modules for something which pretends to parse p5 19:55
suggestions most welcome
20:00 lambdabot joined 20:06 Dr_Pi_ joined, nipra joined 20:13 elmex joined 20:26 justatheory joined
Khisanth putter: PPI? 20:35
20:38 rashakil joined 20:43 justatheory joined 20:53 VanilleBert joined 20:58 nperez joined 21:04 larsen_ joined 21:11 larsen__ joined 21:16 ruoso joined 21:20 Limbic_Region joined
putter Khisanth: thanks. :) yes, that looks like the best option. 21:23
lambdabot putter: You have 2 new messages. '/msg lambdabot @messages' to read them.
putter unfortunately it looks hairy to do right, and PPI is testfailing on my platform... I'm not sure it's on the critical path so I'm punting for now. 21:25
thanks again
prism killing 72.20.26.1:8088 from streaming, as its adding load to the streamer, use alchemy.occult-oracle.net:8088 instead 21:27
putter topic 21:29
svn switch --relocate svn.openfoundry.org/pugs svn.pugscode.org/pugs/ | run.pugscode.org | spec.pugscode.org | paste: sial.org/pbot/perl6 | pugs.blogs.com | dev.pugscode.org 21:30
putter svn switch --relocate svn.openfoundry.org/pugs svn.pugscode.org/pugs/ | run.pugscode.org | spec.pugscode.org | paste: sial.org/pbot/perl6 | pugs.blogs.com | dev.pugscode.org/ 21:31
putter audreyt: re wikirename, ah, sorry. I guess I misread the trac ticket suggesting it was still pending. Perhaps only for non-admins. 21:33
@tell cmarcelo A Commentary is indeed a better alternative to having a TALK associated with particular .hs files. Agreed. :) 21:36
lambdabot Consider it noted.
21:52 Aankhen`` joined
GeJ Morning lambdam􏿽xF8􏿽xF8ses 22:06
rgs: hey theyre :)
22:06 stevan_ joined 22:07 cmarcelo_ joined
GeJ morning stevan_ 22:07
stevan_ $localtime GeJ 22:08
22:15 Aankh|Clone joined 22:23 Psyche^ joined 22:30 cmarcelo_ is now known as cmarcelo 22:32 Aankhen`` joined 22:39 Psyche^ is now known as Patterner 22:42 poland left 22:44 thepler joined 22:56 justatheory joined 23:06 UWC joined 23:17 buetow joined 23:18 gnuvince joined 23:37 sunnavy joined 23:45 coke joined 23:46 coke is now known as Coke
Coke random 23:47
moritz ?eval rand() 23:48
23:48 evalbot_r15248 is now known as evalbot_r15249
evalbot_r15249 0.3501304171164824 23:48
moritz ;)