The Return of the Journal : pugs.blogs.com/ | pugscode.org | pugs.kwiki.org | paste: paste.lisp.org/new/perl6 or sial.org/pbot/perl6
Set by GammaRay on 31 December 2005.
00:03 dduncan joined
putter Hmm. Prim.hs thinks reduce -->List, but S29 has it --> Scalar, which seems more plausible. 00:07
luqui: multi Array::join (@values, Str $delimiter --> Str ) is another one which should perhaps be @values :
good night all & 00:14
00:31 pfenwick joined 00:41 theorbtwo joined 00:46 offby1 joined 00:47 daja77 joined 00:55 hcarty joined 00:57 lisppaste3 joined 01:05 hcarty left 01:14 bsb left 01:23 Cryptic_K joined 01:38 Ovid joined, wilx joined 01:52 revdiablo joined 02:11 Amnesiac joined 02:27 vel joined 02:41 petdance joined 02:46 petdance joined 02:57 avar joined 03:13 petdance left 03:27 vishku joined
meppl guten morgen 03:48
dduncan $acknowledgement 03:52
03:52 luqui joined 04:09 putter joined
putter luqui: ping? 04:10
luqui pong 04:12
er, putter, pong 04:13
putter hi luqui. I had another question about S29draft.pod, that I wasnt sure enough about to declare a typo. 04:17
luqui shoot
putter should this
multi Array::grep (@values, MatchTest $test --> Lazy )
be
multi Array::grep (@values: MatchTest $test --> Lazy ) 04:18
luqui again, my own reservations about the multiness of grep etc. aside, yes
putter ? ( , -> : )
luqui hmmm... that multi seems just to be declaring a method
in which case my reservations don't apply ;-) 04:19
putter ok, thanks. there were several like that that. Prim.hs also uses an invocant.
luqui makes sense
putter so your reservations are for declaring the global subs multi? 04:20
luqui yeah
in fact, IMO most builtins should not be multi
putter because on can $old_value := &foo; sub foo { ... $old_value() } ? 04:21
s/on/one/
luqui more tersely spelled "wrap", but no
the reason is because if you want a different behavior for "say", you should import a different "say"
defining different global multis is like overriding CORE::GLOBAL::*...
putter (which can be very useful in ruby, great for metaprogramming, but not the thing for large systems) 04:23
luqui well, the thing is, in ruby do you ever need to do it at runtime?
perl 6 gives you so much control at compile time that such things shouldn't be necessary at runtime
putter in ruby, you do _everything_ at runtime ;)
luqui oh, right, I think I remember hearing about that 04:24
admittedly I don't have a very strong argument against them
but it's just the "in control" feel that I like to have
the difference between map @foo: {...} and map {...} @foo should not be decided at runtime 04:25
based on the types of @foo and {...}; that's much more like overloading than like multi dispatch
putter "define method" is a regular function. very flexible. ruby now my favorite metaprogramming language, above p5 and CL. p6 will take 1st place as long as its "staticness" doesnt turn out an unavoidable burden. 04:26
hmm...
luqui well please argue if you see us going too far on the static side 04:27
putter one reason to do a ruby frontend soon is to turn up any problems as we go along. 04:28
re, map, what about additional args? map @foo:{...},FrobMap that have an ok feel? 04:29
luqui what does that mean? 04:30
just passing args to the map method? 04:31
that's okay. i'm just saying that map @foo: {...} and map {...} @foo should desugar to either the exact same thing or calls to two different functions 04:32
putter having a map/2 (array, code adverb, plus extra bit) without stomping on the existing map/1. (numbering args looks odd when some of them are optional named args...)
luqui but map @foo: {...}, FrobMap is just @foo.map({...}, FrobMap)
it's hard to explain I suppose
it's the difference between "implementation selection" and "meaning selection"
multis, IMO, are for the former alone 04:33
if you want the same name to mean two different things (a dubious practice in the first place), it should be a compile time selection
putter err, I thought :{...} was just an adverb. so map @a :{} _is_ the same as map :{} @a. no? does it become a method call at some poing? putter squints and scratches head. 04:34
s/poing/point/
luqui do you see a colon in the second form? 04:35
also, the colon in the first form isn't an adverbial colon, it's an indirect object colon
(yes, colons are very overloaded)
putter ahh.
luqui the reason those two maps mean different things is because the first one means "map the second arg over the first" and the second means "map the first arg over the second" 04:36
it seems like you picked up on that already
anyway... </rant>
putter so map @a: {} === @a.map({}) but map {} @a ... err, how does that parse without a comma? {} taken as implicit adverbial? (feel free to punt discussion at any point - this is just non-bottlenecking curiosity on my part) 04:38
luqui commas are implicit around blocks 04:39
in a very handwavey sort of way
putter ah. so to rephrase to check my understanding, the issue is f($x: $y) and f($x, $y) shouldnt call distinct versions of the same (multi) function f. yes? 04:43
luqui uh, yeah, that's the specific case of my complaint
the more general case is that multis shouldn't be used to do meaning selection 04:44
but anyway, gtg
luqui &
putter bye luqui
tnx for your help
04:46 offby1 left
putter hmm, anyone: f($x) is different than f($x : ) , yes? (another S29draft oddity) 04:48
dduncan okay, so I just sent to CPAN the (presumably) final versions of the Rosetta framework from the first / pre-rewrite code base 04:49
just in time for friday the 13th
whenever Pugs 6.2.11 goes out, I'll publish the first release of the second major code base, concurrently to it
putter luqui: re f($x: $y) vs f($x, $y), I think it will be interesting to see if people are confused by it in practice. if I was in "scheme mindset", then of course they should be normalized/minimized. but an advantage of perl over scheme, is there are a lot more "hooks" to shoehorn in what you need to do. so I end up 04:52
reluctant to give up hooks in p6 without knowing what keeping them will cost. random thought.
dduncan++
putter is finding reading Apple announcements strange... "hey, Rosetta! oh, wait, no again." 04:53
;)
dduncan I hear thta
that
Apple probably won't mention theirs after a year or two, as its usefulness fades when apps are universalized 04:54
heck, I don't know for sure, but they may have just killed their Mac OS 9 compatability environment
at least, I heard it won't run on the intel machines, though I don't see why not 04:55
BlueBox should be emulatable like any other Mac OS X app for PPC
I know many people still have 9 apps that were never nativeized, like games 04:56
putter hmm... ah.
err, do optional arguments need to be declared with ? now even in the presence of a default? ie, is sub f($x = 3) ok? (more S29draft oddities) 04:57
or only sub f($x? = 3) ? 04:58
svnbot6 r8655 | putter++ | docs/AES/S29draft.pod - tweaks. 05:10
r8655 | putter++ | Some Array methods were using "," instead of ":" - fixed.
r8655 | putter++ | Adjusted whitespace for alignment in some signatures.
r8655 | putter++ | Finished off luqui's conversion of CALLER::<$_> to $+_ .
r8655 | putter++ | QUESTION: some arguments with default values end in "?", others not. Is this a typo? Eg, f($x = 3) vs f($x? = 3).
avar $ perl -CO -le '$_=shift;s/\D//g;print chr hex' $(perl -CA -e 'printf "U%x\n", ord shift' ā‰ ) 05:11
ā‰ 
does perl 6 have an easy way to do the former thing? 05:12
I.e. an in-place s/// that returns the mutated string
pseudocode: print shift ~~ s/..
luqui avar, specced is $string.subst(/pattern/, "replacement"/
avar guess that kind-of-works 05:14
b.t.w. is pugs -e 'say shift' arg not supposed to work?
shift @ARGS does
putter avar: not sure, sorry. 05:23
05:25 Bit-Man joined
putter actually, I do know. fiddled with its code earlier today. :) shift defaults to the caller's @_. unless in a global context @ARGS is somehow magically @_ (it may be, I just don't know), then shift() won't get you shift(@ARGS). 05:26
05:27 vel left
putter luqui, anyone: S29draft has Array::grep etal in the List section. That's just to keep them next to List::grep, and they are really Array methods, yes? 05:30
05:37 rantanplan_ joined 05:55 gugod joined, comet^ joined, clkao joined, LCamel joined, kcwu joined, buu joined, arguile joined, ingy joined, mtve joined, typester joined, Grrrr joined, hcchien joined, integral joined, gaal joined, jiing joined, calvin__ joined 05:56 geoffb joined, sahadev joined, obra joined, spinclad joined, tcliou joined 05:57 pasteling joined 05:58 miyagawa joined 06:00 Jooon joined, sky joined, dvtoo joined, xern joined
svnbot6 r8656 | putter++ | src/PIL/misc/prelude/ - more fun converting Prim.hs to p6 for piln. 06:20
r8656 | putter++ | Array.pm - Added additional methods. Now parses. Still some TODOs.
r8656 | putter++ | Lazy.pm - Begun.
r8656 | putter++ | Extra eyeballs encouraged.
putter good night & 06:27
gaal night putter! morning #perl! 06:29
s/l/l6/ 06:30
foundry is slooooow
06:32 iblechbot joined
dduncan gaal, I agree, and it has been for a week plus 06:44
svnbot6 r8657 | Darren_Duncan++ | r1950@Darren-Duncans-Computer: darrenduncan | 2006-01-13 22:51:09 -0800 06:53
r8657 | Darren_Duncan++ | /ext/Rosetta : in Rosetta.pm, removed placeholder class Rosetta::Interface::fubar
07:01 elmex joined
dduncan it seems that Pug's version fetching code is busted ... with all the current updates, it says the version is 8580 07:17
perhaps this means that it broke after r8580
er, actually I will try a make clean et al and try again, since smoke says some people's are still working 07:19
maybe
gaal ChangeLoggeth 07:20
dduncan a quick look in my system with svk info etc, and looking at version_h.pl, indicates that nothing changed in the version handling parts 07:27
07:30 lisppaste3 joined 07:32 Debolaz joined
dduncan so everything is right up to and including the setting of pugs_version.h ; that file contains the right version 07:32
but the compiled pugs does not
maybe this is a make clean issue ... or a build bug
did make realclean, now trying rebuild ... 07:34
08:17 vishku left
gaal I'm taking a break from changelogging 08:23
svnbot6 r8658 | gaal++ | ChangeLog towards release - got to r8131; doesn't include objectspace stuff
gaal audreyt + stevan: I have ignored MM/OS stuff completely, please update yourselves
everyone: please check your changes haven't slipped 08:24
bbiab &
08:50 integral left
audreyt woke up from a 16+ hours sleep with a terribly starved stomach 09:27
audreyt downloads the backlog and went out for some food... bbiab 09:29
09:33 elmex joined 09:42 ilogger2 joined 09:51 G2 joined, meppl joined 09:56 bernhard joined 09:57 iblechbot joined 10:04 r0nny joined 10:10 kanru joined
dduncan the clean rebuild seems to have worked; pugs now displays the correct version 10:13
10:14 cognominal joined
svnbot6 r8659 | Darren_Duncan++ | r1956@Darren-Duncans-Computer: darrenduncan | 2006-01-14 02:20:29 -0800 10:26
r8659 | Darren_Duncan++ | small ChangeLog correction; moved items from Updated Modules to New Modules
10:42 GeJ joined, kane_ joined 10:48 xinming joined
kane_ audreyt: codebase1 down for scheduled maintenance, but not up as expected -- no ETA yet, just FYI 10:51
dduncan well, naptime for me, but I should be back in about 10 hours 10:53
11:10 chris2 joined
gaal rehi 11:19
audreyt rerehi 11:20
gaal yo audreyt!
audreyt gaal++ # chglogging
yo!
gaal plans for release? 11:21
audreyt I ate too much and am feeling sleepy again... *pout*
gaal aw :)
audreyt plans for release is to work on Emit.PIR a bit
till we regain 01-sanity
and PGE
and then chase down test failures 11:22
gaal okay. I'm out of ideas re: multi export :-(
audreyt any outstanding regressions? multiexport?
ok, will fix that first in the testfail batch then. which test should I look for?
gaal ext/Algorithm-bucket or other; anything that does "skip".
I left comments in the code with what I'd tried. 11:23
audreyt 'k
gaal I'll continue chlogging
audreyt that'd be lovely
gaal only ~500 rs to go :)
audreyt oh btw, have you sent me the invitation letter? 11:24
gaal cripes, no, sorry
audreyt I'm going to book airplane ticket and work on visa next week
'sok
gaal Ah I need a fax/snailmail address too
audreyt msged 11:25
gaal got it, thanks :)
audreyt :) 11:26
audreyt starts a smoke run in the bg 11:28
gaal fixes his colinux to have more ram and not die with evacuate errors 60% into the smoke :(
audreyt so, nothing definitive with use_ok? 11:31
audreyt ponders the equivalent of ok.pm in p6
gaal hmm, no, I need to fix that too
audreyt k 11:33
mm, the &skip looks like a showstopper. I'll tackle it now
ingy audreyt: hello 11:42
I need to ask some questions of you 11:43
audreyt ingy: yes? 11:47
brb 11:53
ingy: oh, when do you leave .tw? I need to send the wristwatch to you
ingy audreyt: hi sorry, I leave on 26th 12:05
I may come to Taipei next weekend
I need to either release my YAML module unsigned, or register my key 12:06
I could use some direction there... 12:07
I think that goes for the other module releases I did 12:08
svnbot6 r8660 | gaal++ | tests in ext/ and t/ - workaround the breaking of 'use_ok' 12:09
gaal sorry big change to tests, audreyt you may want to resmoke
ingy hi gaal 12:10
gaal heya ingy, new YAML? yum! :)
ingy it's still not as tasty as it should be, but thanks 12:11
gaal "the breaking of 'use_ok'" sounds like "the hunting of the snark"
ingy ha
ingy is done hunting snarks
gaal did you sell all your railway-share? 12:12
ingy yes
audreyt ingy: ok, we'll meet this weekend 12:13
ingy audreyt: also i need to know if I register my key how long that will take to become effective.
next weekend you mean?
gaal fglock: ping 12:14
rafl: ping
kane_: ping
ingy audreyt: are you highly distracted? 12:15
gaal this is like ping 255.255.255.255 :)
rafl gaal: pong 12:16
audreyt ingy: "register" your key?
ingy yeah
rafl gaal: Hurry up. I need to get a train.
audreyt ingy: yeah, today is my off-ritalin day since 5 days of it :)
gaal rafl: hi! I'm not going into any detail about JIB/sixpan stuff in the changelog, please prepare a summary of what you think needs to be mentioned.
audreyt ingy: I don't know about registering... probably just "gpg --send-key 0xDEADBEEF" 12:17
gaal rafl: you don't have to do this right now, but please do it before the release
ingy hmm. people are having trouble installing YAML because of the gpg signing
rafl gaal: OK. I think I can do that this evening.
gaal: When will the release be?
ingy which module::install forced on me
gaal rafl: thanks. probably much later tonight our time, or tomorrow. 12:18
audreyt ingy: huh? I think you have an old M::I
ingy does a newer version not be so nazi?
audreyt the forced signing was a bug that's been resolved a while ago
ingy ok
rafl gaal: OK. Thanks for the reminder.
ingy do I need a new Module::Signature?
gaal rafl: sure :)
audreyt yeah, you were releasing with .47
0.54 would do much better
that's optional 12:19
ingy or just a new Module::Install?
audreyt just a new M::I would do
gaal audreyt: do we have other showstoppers except multis and PGE?
ingy ok, thanks...
audreyt: also i want to add YAML.pm tests to YAML::Syck if that is ok
audreyt gaal: PIR targetting 12:20
ingy: sure, just committ
gaal: -BPIR
gaal: but that's all
gaal audreyt: unfortunately it looks like you're in the critical path for all three of these :)
ingy audreyt: thanks. you can go back to your distractions :) 12:21
audreyt++
gaal I have today for misc. releng tasks like changelog and bughunting
but none of tomorrow I'm afraid.
audreyt sure, chglogging + bughunting for regressions would be much help 12:22
I'll solve all three in the upcoming 24hrs :)
gaal cool. there are a few yellows in the smoke too I've noticed, btw. 12:23
audreyt mmm the color of unexpectedness 12:24
gaal PJF: ping 12:26
(anyone with cygwin): ping 12:27
PJF++ fixed the port, I want to make sure it's still fixed for the release.
audreyt I can test 12:29
which would require a reboot... so bbiab :)
gaal :) 12:30
12:54 GeJ_ joined
gaal it would be nice to get variable exports in for this release too, but I wouldn't bet on my completing it :/ 12:55
12:56 mncharity joined, mncharity is now known as putter
gaal hey putter 12:57
putter hey gaal 12:58
12:59 GeJ_ joined
putter gaal: re variable export, something I didnt appreciate until just now reading the responses to stevan's p6l post re the object hierarchy, is just how much in flux the whole thing is. To manage the "big p6 prelude" in the face of such flux, one really wants powerful (say haskell like) imports, either in use() or does(). 13:02
audreyt indeed.
s/either/both/
audreyt is, amazingly, still rebooting win32 thanks to various secutiry updates.
it's been a long while since I booted here. 13:03
s/secutiry/security/
gaal putter: I'd love to add that stuff, if I knew what it looked like :) 13:06
13:14 elmex joined 13:15 chris2 joined
putter "Security updates are available. To install them (and thus maintain system security for another few hours), press Ok. To be immediately compromised, press Skip." 13:18
both++
putter tries to think of a language which handles it well, to use as a prototype... 13:20
audreyt Scala? 13:21
svnbot6 r8661 | autrijus++ | * cygwin build fix -- integrated from M::I 13:24
putter re both, definitely both. use() is not sufficient because it would basically ahve to take over from the inheritance/mixin system. package and kids (modules, roles) are not sufficient because you want to be able to inherit/mixin and thus modularize things which impact multiple namespaces. 13:25
audreyt gaal: cygwin works.
(using ghc 6.4.0 deliberately)
will retest with 6.4.1
but should work as well
putter doesnt have a clear idea of what a win would look like. going to look again a scala...
audreyt actually
JS2.
www.mozilla.org/js/language/js20-19...kages.html 13:26
it's an early draft
hunt a newer one?
www.mozilla.org/js/language/js20/co...kages.html 13:27
hm, but it doesn't handle mutual-use 13:28
putter and scala.epfl.ch/docu/files/ScalaReference.pdf sec 4.7 13:29
physical page 42 13:30
13:34 bsb joined
svnbot6 r8662 | iblech++ | * Usual svn props. 13:37
r8662 | iblech++ | * util/file_to_hs.pl: "DO NOT EDIT THIS FILE" is added to the output files now.
r8663 | gaal++ | ChangeLog updates to r8657 13:39
gaal rehi 13:40
audreyt: re: cygwin - yay! I thought we required 6.4.0?
svnbot6 r8664 | autrijus++ | * cygwin CRLF fix for gen_prelude
putter iblech! :)
audreyt gaal: sure, we require 6.4.[01]
gaal err I meant I thought we required .1 :)
glad to be wrong, fixing changelog. 13:41
putter definitely need .1 on x86-64.
svnbot6 r8665 | gaal++ | ChangeLog - 6.4.1 recommended, 6.4.0 possible on some platforms 13:43
gaal okay, the brunt of ChangeLogging is done. 13:44
audreyt woot 13:45
gaal++
gaal audreyt: you and stevan in particular still have lots of work to do there :)
audreyt GHC 6.4.1 vs cygwin is also confirmed as working. 13:46
vs 6.5-head is broken due to ghc bug (getDirectoryContent) 13:47
no tuits for a core fix now
I'll boot back to a far saner system :)
gaal lol
gaal start a build for smoking 13:50
s/t /ts/ 13:51
putter dylan oop.rosweb.ru/dylan/book.annotated/ch5.html www.gwydiondylan.org/books/drm/Program_Structure eh. proto and goo (dylan kids) went overboard towards simplicity. 13:52
oh, actually, goo did. rechecking proto. 13:53
13:58 Southen joined
chris2 mmmh, goo. what a nice language it was 13:59
audreyt heh, people.csail.mit.edu/jrb/goo/wiki/i...hp/GooNews 14:00
boo.codehaus.org/Language+Features 14:01
putter nice biography with ps/pdf papers readscheme.org/modules/
chris2 i'd be satisfied with a sexpr dylan, actually :-)
audreyt putter: heh, I was just looking for that link 14:03
14:05 SamB joined
chris2 can your syck round-trip {[]=>""} ? 14:06
audreyt chris2: for PIL^N p6hash types, yes, but the normal runcore still has Str keys 14:07
s/normal/current/
chris2 i mean, does it really work? because there is a bug in ruby 1.9 wrt that
audreyt let me write a quick test 14:08
rafl gaal: Is it OK to edit the changelog directly? 14:10
gaal rafl: yes.
audreyt emit works 14:11
gaal audreyt: you're emitting yaml from PIL^N? wow
kane_ gaal: pong 14:12
ah, rafl... long time no see
audreyt loading works
gaal kane_: hey. please talk with rafl re: updating the changelog with your changes in the past ~month; we're planning a release soon.
audreyt chris2: parseYaml "--- { []: \"\" }" looks good to me
gaal++ # nice binding
kane_ gaal: exclude JIB for now :) 14:13
gaal audreyt++ # loads of help
chris2 and that's what gets generated?
because it should be ? [] : "", i think
kane_ rafl: on that note, RL has caught up with me in a bad way... no time for JIB for a while it looks like :(
audreyt oh.
audreyt turns inlining off
gaal chris2: isn't that the same thing?
audreyt aha. 14:14
chris2 what is {"[]" => ""} then ?
audreyt trunk syck under noinling generates
? []\n: ""
but the release syck (which pugs now uses) gens
rafl kane_: I won't have much time until february as well.
audreyt "--- \n[]\n: \n"
gaal kane_: please write with rafl whatever you think appropriate directly.
audreyt which is indeed a parsefail
chris2 okay, so it's a syck bug, not a ruby one :) 14:15
gaal hmm, we can update to trunk syck, but doesn't that have stylistic output problems? or were thouse fixed?
audreyt gaal: it always generates ?
probably to fix exactly this bug 14:16
gaal heh.
audreyt as it doesn't srop the ? even on simple strings
putter Mixin modules for dynamic rebinding citeseer.ist.psu.edu/730012.html (not on the readscheme list)
audreyt s/srop/drop/
chris2 oh :)
audreyt gaal: maybe we should consider using trunk syck and turn on inlining
audreyt ponders
chris2 ? "bla" : foo is icky 14:17
audreyt yeah. definitely need adaptable stuff
chris2: ping Why to fix it?
chris2 he reads ruby-core
gaal he did mention he had a release coming up... 14:18
audreyt gaal: 14:19
pugs> eval(/1/.yaml, :lang<yaml>)
{rule}
it worksforme
parrot 0.4.1
what was the problem?
(external parrot)
trying with embparrot now
gaal audreyt: I think the problem isn't in .yaml at all, it's in embparrot PGE 14:20
audreyt ok, looking
gaal 0.4.1 also
brb, getting food
shapr++ # "Haskell separates Church from state" 14:31
audreyt shapr++ 14:32
putter wishes the individual... clauses/implementations/mumble (what is the jargon?) of multis were themselves first class objects. or at least, given multi f, f ::= assemble_a_multi(f.disassemble.filter(...)); 14:36
gaal 's machine crashes again :( 14:38
putter !
SamB wasn't Church already seperated from state? 14:39
gaal this happens a lot. make smoke... ghc evacuate errors ... look at temperature... kablooie
SamB gaal: are you actually talking about smoke/temperature?
gaal SamB: make smoke pushes the cpu, causing high temperatures etc. 14:40
SamB gaal: ah
gaal yeah, this chip runs hot :(
SamB is glad to have a PII
putter throttle chip in bios?
gaal I should shop for RAM for the old laptop so I can make optimized /preludepc builds with it 14:41
putter add memory quota to force swapping? ;)
gaal putter: I haven't found the option. it's a leet gamer bios, only good for overclocking :(
putter: :)
maybe I should submerge the thing in oil like those crazy hackers 14:42
SamB CPUs that overheat at 100% usage sound like a pain.
putter gaal: linux? 14:43
gaal actually, is there a way to isolate the culprit? maybe it's faulty ram or bios? 14:44
putter: uh, sorta. debian on colinux on winxp
SamB especially if the schedular doesn't give them idle time to keep the temperature down...
putter k. I've a fuzzy recollection there is a hard cpu quota available, so you could say max 50% cpu, and will get 50% idle if nothing else is happening on system. 14:45
gaal putter: at what level?
SamB but I don't think Pentium IIs ever have this problem, so I am happy ;-)
gaal i'd just lower the clock speed by 15%
but I don't know how. 14:46
SamB I mean, unless the fan breaks or something...
gaal SamB: how long does a pugs build take you?
SamB it takes a long time
haven't done it lately though 14:47
putter gaal: or not. couldn't find it. hmm. suspend ghc for m seconds every n seconds? 14:50
gaal loop { sleep 1 ; kill STOP, $ghc; sleep 1; kill CONT, $ghc }
heh.
putter :)
hmm... dust accumulated in case/power supply/fan blades? 14:53
heat sink?
audreyt hm, running into parrot pthreading bugs.
or maybe it's freebsd pthreading bugs.
putter starts build... 14:54
audreyt decides to sleep some more and do the releng with a wakeful mind 14:57
aha, parrot only fails when built with parrot_is_shared
putter someone said something about a load library path(?) issue on #perl6 within the last few days. 14:58
gaal?
gaal putter: no dust/cat fur, and the cpu itself was replaced 14:59
putter ohhhhh! maybe a bad cpu-heatsink bond?
gaal putter: no, that was just in running; export LD_LIBRARY_PATH=.../blib/lib worked.
putter is the heatsink itself very hot?
gaal putter: how do I tell?
putter touch it? or is this a laptop? 15:00
gaal i'm on the laptop now; the problem machine is a desktop. but what's considered hot? is hot ok?
chris2 dont you have any sensors? 15:01
gaal anyway there's a huge fan sitting on the heatsink, this is a p4
chris2: it can get to 80C.
chris2 ooh
gaal prescott :(
chris2 my athlon xp always crashed around 72 15:02
pentium at least has overheating protection :P
gaal it's supposed to be designed to run this hot
i think
putter another possibility is to look at how fast it cools down. once the load is gone, the temp should drop quickly, limited only by thermal mass of heat sink. if it doesnt... 15:03
gaal it does, actualyl.
putter heat sink rated for the cpu?
gaal iirc goes to about 50 or 40 even
15:03 xern_ joined
gaal putter: how do I check? 15:04
chris2 that said, my ibook fan only spins when i do haskell related stuff :P
putter heat sink spec sheet?
ha, cant say / box at the beginning of an irc line. ;)
gaal takes a break also 15:07
bbiab &
putter basically the chain goes something vaguely like; clockspeed/voltage too high, bad cpu/heatsink bond, heat sink too small, insufficient airflow over sink (eg, sink fan dead, or a temp->speed control not doing the right thing), case hot. ;) 15:08
bye gaal. good nap audreyt.
audreyt see ya :) *wave* 15:10
putter gaal: re use/does, the low-hanging immediately tasty fruit would be use() preserving order of multi overwrites. Ie, two
&
ingy eats a bowl of mueslix with milk using a metal spoon and pines for apple pie 15:11
(which I'm sure they probably have at the corner Taiwanese 7-11) :\ 15:13
mmm milk... *sigh* 15:14
putter multi f(Num $n) {$n+1} and multi f(Num $n) {$n+2} in the same file, the first definition wins. (on pil2js with an warning). But if you put them in separate files, and use() them, the order of the two use()s does not matter. Oh my. The file name determines the order. 15:15
ingy notices for the first time that gugod has no oven... 15:17
putter If that were changed to use() order, we could start creating a unified pil2js/piln/etc p6 prelude.
I think.
ingy stares at the prebuttered garlic bread just procured from the grocery 15:18
putter is being made hungry by ingy. this is not a good thing. ;)
ingy reluctantly walk over to the microwave :\ 15:19
hi putter
putter hi ingy
ingy how's New England? 15:20
putter I'm told it was supposed to get nasty today. 20degF drop from yesterday. Still looks fine. Maybe later.
ingy yucky 15:21
we been having a most unfortunate streak of sunny days in the 70s here in Taiwan
putter boooooorrrrriiinnngggg. ;)
my sympathy. 15:22
ingy well I deserve your tears because I sleep all day, so it does me no good
putter has (re)discovered the joy of getting up a few hours pre-Sun. City veerrrry quiet. (good for hunting wabbits) 15:24
bbiab
ingy intewesting
15:30 lisppaste3 joined 15:51 nothingmuch joined
audreyt spent far too much time 1)joining acm.org and 2)reading papers 15:51
16:00 putter joined
putter papers++ acm-digital-library-policy-=12 16:01
piln, piln, ra!
got to go. bye all. happy hacking. & 16:02
nothingmuch seen gaal? 16:04
jabbot nothingmuch: gaal was seen 56 minutes 23 seconds ago
16:06 gaal joined
gaal is dazed from a misconfigured nap, searches for caffeine 16:08
16:12 justatheory joined
gaal nothingmuch: you pang? - got your email 16:19
obra morning 16:21
gaal hey obra
nothingmuch yeah, it was about the email 16:23
sorry =(
got your reply though
gaal rereplying 16:24
gaal wanders off to wake up again & 16:27
16:35 imperator joined 16:54 feng joined, bsb joined
leo_ audreyt: I saw your freebsd troubles - still problems? 16:57
17:09 vel joined 17:19 nothingmuch joined 17:21 vel left 17:22 kane_ joined 17:38 nnunley joined 17:43 mjl69 joined 17:52 justatheory joined 17:57 nnunley is now known as boris, boris is now known as nnunley 18:39 stennie joined 18:40 phredmoyer joined 18:41 sapper joined 18:43 landover joined 18:49 elmex joined 18:59 justatheory joined
wolverian hm, any idea if -a and @F still exist in perl6? 19:18
19:22 sapper left 19:32 integral joined
gaal wolverian: the command line is almost completely unspecced 19:47
wolverian right. well, given laziness, I was just thinking it could be @F[line; column]
that seemed a tad more elegant. 19:48
gaal you mean for line number? interesting, but only efficient if it's used monotonically rising. 19:49
afk & 19:51
19:53 elmex joined
wolverian gaal, more efficient than using -n? if it's equally efficient it doesn't matter. 19:56
20:09 elmex joined
wolverian also (just brainstorming): perl6 -e'@ARGSĀ».openĀ».linesĀ».splitĀ»[0].grep(/^129\./).uniq.elems' /var/log/apache2/access.log* 20:14
20:30 xinming joined 20:35 elmex joined 20:36 Grrrr_ joined, elmex joined 20:38 beppu_ joined 20:51 elmex joined 20:52 dduncan joined 21:08 avar joined 21:34 xinming joined 21:36 putter joined
putter Ok, here's an odd thought. 21:38
class A{#A1} package P{class A{#A2} class B is A;} 21:39
my recollection, which may be out of date, is that this is specced as defining A, P::A, and P::B, where P::B is P::A. That's A2. All references to A in P::B are understood as P::A. 21:41
This isnt the current pugs behavior. But let's say it was.
package Test1 { use BigP6Prelude; &bootstrap(); } could give you a oo system in a box. Test1::Object, Test1::Array, Test1::infix:<+>, etc, etc. 21:43
And you could simultaneously have a Test2, etc. And they could all be differently structured systems. Item in one but not the other, etc. And they are relatively self contained. Only if one says ::*:: does one pop out of the box. 21:45
The underlying p6 system in which the packages TestN are defined, is serving the role of the underlying backend runtime. 21:46
One can execute code in any of them. With package Test2 { blah } , blah sees the Test2 object system. 21:47
Does this hang together? Am I missing something? 21:48
It raises the question of, can we have chroot? :) chroot Test1 { ::*::Foo } with Foo now referring to Test1::Foo. 21:49
21:55 avar is now known as avar\away
putter And of how scoping for the parser is handled. Can each of the TestN have their own current parser state, set of macros, etc. If so, then we could have multiple simultaneous p6 systems running. 21:55
stevan, audreyt, anyone: thoughts? 21:56
stevan: among the many languages you've written bootstrap code for, is p6 among them? :)
Hmm, is the pil bootstrap just p6 plus backtick? Could one macro backtick and get a p6 bootstrap? 21:57
Oh, on rereading, one can of course "pop out of the box" (in the absence of chroot) by simply calling something defined in the underlying p6 and not shadowed by the Test one. 22:00
22:04 justatheory joined
putter goes back to trying to think of ways to have a single shared big agile prelude. 22:09
putter again considers attempting to hack in full-file source filtering. Sign by highway: "if you lived here, you would be home now". if we could write a little filter, we could easily dodge classes not quite working. our current macros just aren't quite powerful enough. 22:16
could just prepend $?PACKAGE:: to all class names in the prelude... ;) 22:18
(currently #A2 lands on the top level A. #A1 ::*::A ) 22:20
ah well. small goals. get piln. flesh it out with any sort of prelude. giving working objects. and then we can have fun. 22:26
of course, everyone is welcome to have fun before then too. ;) 22:27
22:43 sub_chick joined 22:47 iblechbot joined
putter hi iblech(bot) 22:48
22:49 vel joined
nothingmuch does anyone here camp a lot? 23:03
elmex you mean idle ? 23:04
elmex++
perlbot What kind of idiot karmas himself? Your kind of idiot!
elmex perlbot--
putter ok, re source filtering. how about a new ruleWithCurrentInput which defines a keyword/statement with_current_input <identifier>; which getState and setState's with a new input of <identifier> <<ENDD old input ENDD so...
macro f($input) { my $src = $input; $src ~~ s/Foo/Bar/g; $src } with_current_input f; blah Foo blah 23:05
hi nothingmuch 23:06
nothingmuch elmex- err nope, like in nature 23:07
hi putter
putter anyone have api comments? or understand Parser guts enough to say "sounds vaguely (im)plausible"? 23:08
nothingmuch me backs out carefully 23:09
putter wishes there were more people here UTC night
lol
"but wait, there's more. for only $9.99USD, you get..."
what was it Alice (in Wonderland) said? Believe N impossible things before breakfast? So I'm off to dinner. ;) 23:14
audreyt, anyone: comments on source filter idea would be appreciated.
good night all &
23:17 mjl69 joined 23:28 putter joined
putter hey, at least currently, one could have an api of quote_current_input or somesuch. 23:29
pugs> quote_current_input foo bar hee
" foo bar hee"
23:30 lisppaste3 joined
putter so, macro f($in){my $s=$in; $s =~ s/Foo/Bar/g; $s} ... f quote_current_input ... all this gets handed to f ... . ie, input "quote_current_input everything to eof" -> "<<END everything to eof END". 23:32
&
23:37 xinming joined 23:38 putter joined
putter hey, its just a normal-ish quote operator, q-with-eof-terminator! :) 23:39
&

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!