Check your feather email | spec.pugscode.org | paste: sial.org/pbot/perl6 | pugs.blogs.com | www.treehugger.com/files/th_images/paradigm.jpg [set by audreyt on 2006-08-29 05:26:57 -0700]
Set by Akwa|user on 14 September 2006.
rodi gnuvince: that commit bit is the single most insidious bit in the multiverse :-D 00:00
ahh, my lady is home home with Indian food. TTFN. 00:02
gnuvince So, how does writing tests work? I pick a section of a synopsis, write tests for that section (or improve those already existing)? 00:03
obra gnuvince: that will certainly work. do you have a pugs checkout yet? 00:05
gnuvince obra: it dates a couple weeks, I'm gonna update it and recompile.
obra: what happens if I write tests for something that doesn't exist yet?
obra that's ok.
that's one of the best ways to get new features implemented 00:06
checking in failing tests for new features energizes some of pugs' developers.
gnuvince ok
obra around release time, they may be 'todo'ed, but don't let that stop you
gnuvince todo means they can fail, but it's not marked as a failure? 00:07
markstos rodi: I enjoyed your journal post! 00:08
obra todo means they may fail, but that's because we know they're not done yet
markstos gnuvince: since we are preparing for a release, please use :todo<feature> for unimplemented features 00:09
Here's an HTML version of the p6 Test.pm docs:
feather.perl6.nl/~markstos/Test.pm.html 00:10
lambdabot Title: Test - Test support module for perl6, tinyurl.com/p2dor
gnuvince markstos: thanks.
markstos The structure of t/ roughly follows that of the Synopsis: try keep that pattern. So, builtins functions are spec'ed in Perl6/Spec/Functions.pod, aka "S29" and have tests in "builtins/" 00:11
00:11 [particle] joined
markstos (I'll avocate that it should be "Functions/" instead of builtins to match exactly, but that's a dicussion for another day). 00:12
test your test: prove6 my_test.t
(prove6 is in ./util )
And t/README is a good read, too. 00:13
00:13 lumi_ joined
gnuvince ok 00:16
svnbot6 r13316 | Darren_Duncan++ | incremented or decremented version numbers of my ext/ modules, brought Pugs' ChangeLog up to date regarding my modules
obra rodi: ping 00:19
dduncan now starting my first Pugs build since last month ... 00:23
00:24 shasbot joined
svnbot6 r13317 | markstos++ | [t/operators/reduce-metaop.t] 00:24
r13317 | markstos++ | New failing test:
r13317 | markstos++ | sub prod ($x,*@xs) { return $x unless @xs; return $x * prod([,] @xs); } print prod(2,3,4);
r13317 | markstos++ | Larry Wall and I thought the behavior seemed like a bug.
r13317 | markstos++ | I wasn't which language feature has the bug, so I blamed it on [,] and stuck the test here.
TreyHarris TimToady: In S01 diff just sent out, sentence "Mostly we're just a bunch ants" is missing article. :-) 00:27
TimToady anything else?
00:29 nekokak joined
obra TimToady++ # Doing the job I'm not ;) 00:30
TimToady: the year on the last-mod date is off by one 00:31
also: poeple -> people
clkao obra: would you like a commit bit ? 00:32
obra to the synopses? I don't think that's under pugs control ;)
TimToady obra++++ but too late. :) 00:33
gotta run now... & 00:34
obra *wave*
thanks for that update :)
00:35 jferrero joined
markstos Does Haskell have native reduction operators, like [*] ? 00:37
wolverian fold[lr]
markstos I know it's easy enough to work around them in Haskell, like: prod (x:xs) = x * prod(xs) ( == [*] )
rodi obra: pong 00:38
mauke product = foldr (*) 1
obra rodi: got your mail. we should chatter some time when I'm back home (after the 24th)
mauke or maybe foldl' (*) 1
obra I hadn't actually connected your irc nick to the guy I went to legal seafoods with ;)
markstos What does the "1" do there ?
rodi obra: sounds great- no rush on my part. Enjoying my time off, prolly more than I should ;)
obra Excellent. I'm jealous 00:39
mauke markstos: it's what the empty list reduces to
markstos Ah.
wolverian TimToady, "Another aspect of this is [that?] the Perl 6 compiler will be self-hosting."
mauke concat == foldr (++) []
wolverian TimToady, (I'm not sure if it is a bug, but I couldn't parse it at first without the 'that'. maybe it's a bug in english.)
markstos I was comparing the a little subroutine that does what [*] does in Perl5, Haskell and Perl 6. Perl 5 is pitifully verbose by comparison: 00:43
sub prod { my ($x,@xs) = @_; return defined @xs; return $x * prod(@xs); }
Perl 6 was an improvement, but Haskell looked the cleanest. :) 00:44
mauke return defined @xs;?! 00:47
also, sub prod { List::Util::reduce { $a * $b } 1, @_ } 00:48
markstos Oh. typo.
It worked before, but that version is broken.
Should have been: return $x unless defined @xs; 00:49
mauke why the hell are you using defined with an array?
markstos Something tells me you strongly prefer another solution. 00:50
mauke no, I prefer a solution
defined @array is not useful
(it's also deprecated) 00:51
00:51 justatheory joined
markstos which means the solution you prefer is... 00:51
mauke if you mean return $x unless @xs, that's still broken because it should return 1 for an empty list
<mauke> sub prod { List::Util::reduce { $a * $b } 1, @_ }
00:53 bsb joined
markstos The reduce function itself uses the construct: return $x unless @xs >= 1; (basically) 00:53
mauke yeah, but I'm passing an extra argument here 00:54
00:54 justatheory_ joined
markstos I can sense the extra argument in the room. 00:54
00:54 justatheory_ joined 00:55 justatheory_ joined 00:56 justatheory_ joined 00:59 justatheory joined 01:02 justatheory joined
dduncan compile of r13316 on my machine was successful 01:18
svnbot6 r13318 | markstos++ | [t/operators/reduce-metaop.t] 01:22
r13318 | markstos++ | Two more failing tests, directly from the spec:
r13318 | markstos++ | [*]() returns 1
r13318 | markstos++ | [+]() returns 0
06:51 ilogger2 joined 07:09 zgh joined
gaal mooses 07:15
hahaha, today's wordsmith.org word of the day: vamoose 07:20
07:20 zakharyas joined 07:24 jferrero joined 07:31 drrho joined 07:36 f0rth joined 07:39 f0rth joined 07:42 hareldvd joined 08:09 kane-xs joined 08:18 pdcawley joined 08:29 SuperL4g joined 08:30 yuval joined
SuperL4g yuval: got to re-extract the tree real quick. I logged out by mistake. 08:31
yuval: you pointed out my screenshots already, I take it? 08:32
www.linuxlooney.com/pugs_compile_process.png <-- that's where the process seems to hang 08:33
www.linuxlooney.com/top.png <-- that's self-explanatory
08:38 ruoso joined 08:39 b00t joined 08:47 ludan joined 08:54 SuperL4g joined 08:56 kisu joined
SuperL4g yuval: trying again, and I can already see >4 ghc processes spawned 08:59
I'm sorry... make that =4
09:03 xern joined
yuval SuperL4g: i have =4, too (and it compiles fine here) 09:06
SuperL4g yuval: do you get past the stage where mine seems to stop? 09:07
yuval SuperL4g: but they are killed/respawned pretty quickly :)
not yet, SuperL4g, hold on :)
SuperL4g: oh, oops, i did. went pretty smoothly :) 09:09
ruoso audreyt, once I saw some slides you made about jifty... do you still have them? 09:33
gaal yuval, SuperL4g: does gentoo have any patches to pugs' Makefile.PL? 09:40
SuperL4g: could you try with a checkout from HEAD?
make sure no ghcs are running before you start. 09:41
SuperL4g yuval: you would know this better than I, since this is my first time doing anything with perl6 09:42
gaal SuperL4g: svn co svn.openfoundry.org/pugs ; cd pugs ; perl Makefile.PL && make fast 09:45
lambdabot Title: Revision 13324: /
09:46 penk joined
yuval gaal: yes, we patch Makefile.PL (i think i reported this problem here a few times in the past?) 09:48
gaal yuval: the patch you sent me via IM doesn't look like it can cause the GHC problem.
yuval but we just turn my $build_dir = $ENV{PARROT_PATH}; (because the build dir reported by parrot-config is all wrong when compiled in Gentoo) 09:49
gaal: i know :)
gaal: fwiw, many people managed to emerge pugs using the same ghc/parrot versions, so it's weird...
gaal yuval: I have to leave again, but if you want to just go ahead and patch pugs' makefile go ahead 09:50
(re the gentoo patch)
if it makes sense on non-gentoo ssytems, that is
gaal mooses off & 09:51
09:51 kanru joined
yuval it doesn't... :) 09:51
09:51 pdcawley joined
yuval when co from HEAD, it gets stuck a tiny bit later: Compiling Data.ByteString ( ./Data/ByteString.hs, dist/build/Data/ByteString.o ) 09:57
one ghc-6.4.2 process is at 99.5% CPU :( 09:58
10:16 elmex joined 10:32 iblechbot joined 10:55 buetow joined 11:25 autark joined 11:46 mako132_ joined 11:53 pdcawley joined
audreyt ruoso: yes, they are part of jifty now 12:13
lambdabot audreyt: You have 1 new message. '/msg lambdabot @messages' to read it.
audreyt doc/talks/oscon.2006.xul
obra morning
audreyt ruoso: svn.jifty.org/svn/jifty.org/jifty/t...n.2006.xul 12:14
hi obra
lambdabot tinyurl.com/h3mtw
audreyt lambdabot: @messages
lambdabot putter said 14h 48m 1s ago: Is there a vision for how Prelude.pm is going to evolve? It seems there is no way to define globals but "is builtin", which seems to have the unfortunate sideeffects of
ripping things out of their defining module, and making multis non-multi.
audreyt mmm.
@tell putter indeed. will think about it. 12:15
lambdabot Consider it noted.
gaal audreyt: yo. what would be the least painful way of getting ~ newval to dispatch?
(rather than be folded to "CCall" ++ etc.? 12:16
)
..I don't want the new parsers to wrap their results in Val . VV, though supposedly that would work 12:17
12:17 norageek joined
audreyt that seems like the best way actuall. 12:18
gaal aw :( 12:19
audreyt write a retVV helper perhaps?
12:19 Limbic_Region joined
gaal oh wait! return $ Val $ VV $ val $ case inv of .... 12:20
it already is an oldval
so we must either give up on that CCall dumper or maintain a list of safe methods to dispatch to 12:21
audreyt the ccall dumper was transient anyway 12:24
it's intended to delegate to the ResponderInterface 12:25
which will then lookup a concrete Method objec t from a MethodTable
# nothingmuch.woobling.org/MO 12:26
lambdabot Title: Index of /MO
gaal oh cool
with is moose for does, right? 12:27
sorry, "with" is Moose.pm for "does", right?
anyway, I gotta go once again. back in 7h or so. 12:29
audreyt okie 12:35
with is moose for does yes. 12:36
audreyt triages on
@pl \x y z -> x ++ (y ++ z) 12:37
lambdabot (. (++)) . (.) . (++)
audreyt @pl \x y -> moose ++ ( x ++ y)
lambdabot ((moose ++) .) . (++)
12:38 prefiks joined 12:43 avarab joined
svnbot6 r13325 | audreyt++ | * Parse for siglets: 12:45
r13325 | audreyt++ | :($, $, &)
12:47 zeriod joined 12:59 agentzh joined 13:02 pdcawley joined
zeriod what is the difference between svn.perl.org/perl6/pugs/ and svn.openfoundry.org/pugs/ ? 13:07
lambdabot Title: Revision 11986: /pugs
obra svn.perl.org is a mirror of svn.openfoundry.org (for pugs) 13:08
audreyt the first is readonly and the second is readwrite 13:09
zeriod: want a commit bit? :) 13:10
zeriod ok
audreyt your email?
zeriod i was going to help with the test smarklinks
i dont know how to use svn though yet
audreyt that's fine, search.cpan.org/dist/Perl6-Pugs/lib...c/Hack.pod has a quickstart 13:13
lambdabot Title: Pugs::Doc::Hack - How to hack on Pugs - search.cpan.org, tinyurl.com/nnepz
audreyt what's your email address? :)
zeriod send you a pm 13:14
kolibrie hmm, I'm trying to bring my pugs up-to-date, but on r12102 am getting the svk pull error: Incomplete data: Delta source ended unexpectedly 13:15
13:17 pdcawley joined
[particle] kolibrie: it may require an upgrade of SVN::Mirror 13:18
0.70 is latest iirc
kolibrie looks like I have SVN::Mirrir 0.68 13:20
audreyt zeriod: sorry, was disconnected 13:22
welcome aboard! add your name to AUTHORS as your first commit :)
after "svn co svn.openfoundry.org/pugs", edit the AUTHORS file, and then "svn ci AUTHORS".
13:25 dakkar joined
kolibrie [particle]: upgrading SVN::Mirror to 0.70 seems to have worked. Thank you. 13:35
[particle] clkao++ 13:36
kolibrie indeed
13:36 vel joined 13:37 lanny joined
lanny gaal or audreyt: any guidance on extracting the isof of a sub using a Siglist and .signature? Trying to implement sort() and KeyExtractor uses that info. 13:39
Well... I think it uses that info. Spec could be interpreted as checking the type on the return value but that doesn't have to be a single type. 13:40
audreyt lanny: it's still unspecced that .signature would actually return isof
it's possible that .returns returns naother signature 13:41
lanny I was thinking .returns might be the inner return
audreyt nod
lanny I know it's unspecced which is why I am looking for guidance rather than "the answer". 13:42
Right now I just check the type of the returned val and play with that.
audreyt I think that's safest right now
I think it boils down to whether
my $sig = :(Int --> Int)
lanny Ok. Guess it's time for me to learn some more Haskell and make the whole thing work then.
audreyt makes any sense 13:43
lanny: woot!
it's more complicated for perl6 as we allow multiple, even named, return values 13:44
lanny Well. Nobody ever said it was gonna be easy. :) 13:45
audreyt yeah :)
lanny my $sig = :(Int -> :(Int, Rational $phi))
oops. -->
audreyt if we allow that, then .signature encompasses both sides of 13:46
the function type
so you'd need to say
.signature.out 13:47
.signature.in
lanny But .signature is a Siglist. Siglist itself could have attributes .params, .returns, .isof
audreyt yes but what is the type of .params?
SimpleSignature?
FirstOrderSignature?
13:48 marmic joined
lanny heh. Would love to just say Siglist again but that's not gonna do it 13:48
audreyt right.
well, we can say it's a subset of signature
but it feels very nonclean
I'd rather have Signature to be --> free, and have Code to have three Signature on it. 13:49
lanny agreed
audreyt and --> in declaration position is just shorthand to define isof
13:49 enantiodrome joined
audreyt and in :() literals it is bogus. 13:49
lanny also agreed
Not sure I agree
audreyt but what would be the type of :(Int --> Int) then? 13:50
a CodeSignatureSet ?
lanny my $func; my $sig = :(Int --> Int); $func is $sig;
$func ~~ $sig;
Probably more useful
audreyt not really
$func is $sig is outside spec a lot 13:51
lanny Although $func.signature.isof ~ $sig gets there easier
audreyt since isTrait is named
13:51 enantiodrome left
lanny ah gotcha 13:51
audreyt sub f is signature(...)
is canonical
so I think &code.signature, &code.of, &code.returns
is easiest (and makes slot allocatio agree with trait label)
13:52 cjeris joined
audreyt &code.of ~~ Int 13:52
is not bad at all.
lanny &code.of ~~ :(Str, Num, @x); 13:53
Also not too bad.
Not sure about mixing typed and untyped. So used to systems that are all one or all the other., 13:54
Ok. Thanks for the guidance. I'll use &code.of ~~ as the leading possibility.
audreyt 'k. a S06 diff to p6l would also be much appreciated :) 13:56
13:56 stclare joined
lanny Ok. Will do. Gotta take my girls to school and start the day right now. 13:57
audreyt yay!
audreyt is about to sleep...
though I'd like to make
while f() -> $x {}
work ere I doze off.
13:58 stclare left 14:02 pdcawley joined 14:36 penk joined
audreyt Limbic_Region: looks like www.haskell.org/ghc/dist/current/di...w32.tar.gz is out 14:37
lambdabot tinyurl.com/g7otr
audreyt havn't tested it myself though.
agentzh audreyt: is there a haskell builtin like perl's join? 14:49
unwords == join ' ' :)
audreyt @type intersperse 14:56
lambdabot forall a. a -> [a] -> [a]
audreyt > intersperse ',' "moose"
lambdabot "m,o,o,s,e"
svnbot6 r13326 | audreyt++ | * Don't parse :a as ::a.
TimToady reblink
agentzh audreyt: thanks! 14:57
audreyt TimToady: while/until doesn't bind $_ implicitly, but can do that explicitly right? 14:58
while f() -> $_ { ... }
repeat while f() -> $_ { ... } # also this?
for that latter form I have arbitarily bound the first unconditional loop with the value True
and "repeat until" with the value False 14:59
sensible?
15:01 MacVince joined
TimToady the first is correct, and the second is probably sensible, though my own sensibility at this hour is suspect 15:01
nothingmuch awa 15:02
TimToady on the other hand, if it makes more sense to simply disallow binding on the repeat, that would be okay too. 15:03
esp if you think of the "repeat while" as a macro that does code motion of the while to the end. 15:04
where it makes no sense to say ->
audreyt you mean as postfix?
TimToady (at least as -> is currently designed) 15:05
audreyt but that introduces no scope
and doesn't quite apply
TimToady I mean if you remap repeat while X -> $_ {...} to repeat {...} while X -> $_ it makes little sense 15:06
under current scoping rules
audreyt ah.
right, exactly. though
repeat -> $_ {... } while X
can be made perfectly fine no?
TimToady yes
audreyt then I guess I'd prefer a bindable repeat :) 15:07
TimToady maybe it's "undef but True". :)
nothingmuch lumi: ping 15:08
audreyt been thinking that :) though for enums "undef but" makes very little operational sense...
TimToady but an argument can be made that it's simply undef the first time
audreyt yes.
I think either concrete True or simply undef
undef might be better as it would allos
TimToady yes 15:09
make it undef
audreyt while f_ret_moose() -> Moose $x {}
lumi nothingmuch: pong
audreyt to not bindfail
TimToady right
audreyt yay :)
TimToady and it gives a good conditional to use to determine if you're after the first itme 15:10
*time
audreyt nodnod.
since normally it'd be undef,False,False,False,True
15:10 vel joined
TimToady or versa vica 15:10
well, you never see the final one 15:11
inside the loop, but yes
audreyt k, I amended repeat.t 15:12
will commit shortly and sleep :)
TimToady nnito
audreyt would be cool if you can hack S04 and make the new test smartlinkable :)
TimToady ryokai! 15:13
audreyt arigato gozaimasu! 15:14
TimToady these smartlink thingers are really good for what ails both ADD and autistic folk... :)
nnito = not necessarily in that order, btw 15:15
nothingmuch TimToady: have you looked at the MO thing? 15:16
i'm thinking of proposing it formally as Perl 6's object metamodel
but my emails to p6l are always warnocked ;-)
TimToady ENOTYET
nothingmuch makes a puppy face 15:17
brb, phone
TimToady it's been a crazy week here, between jury duty and plumbing and jet lag and more plumbing and weird orientations for lewis and jet lag and mail backlog and design backlog and jetlag 15:18
nothingmuch wonders which part of that is worse ;-) 15:19
15:23 kanru joined
Limbic_Region completely misparsed "weird orientations for lewis" 15:23
svnbot6 r13327 | audreyt++ | * t/run/12-script-args.t: Remove "pugsrun" pugsism.
r13328 | audreyt++ | * more removal of "pugsrun".
r13329 | audreyt++ | * repeat.t: test for the soon-to-be-specced: 15:26
r13329 | audreyt++ | repeat while f() -> $x { ... }
r13329 | audreyt++ | that binds undef to $x first.
r13330 | audreyt++ | * POD.t: %=POD is not for this release, so TODO them.
r13331 | audreyt++ | * while.t: Fix bogus logic in "while f() -> $x {...}" tests.
r13331 | audreyt++ | (now all subtests pass for this one.)
TimToady not to mention irc backlog... 15:27
svnbot6 r13332 | audreyt++ | * Implement bindable while/until loops: 15:28
r13332 | audreyt++ | while f() -> $x { ... }
TimToady Limbic_Region: that would be school orientations. :)
audreyt ?eval 1 15:32
evalbot_r13318 pugs: user error (*** *** Unsafe function 'use' called under safe mode at -e line 1, column 1 at)
Limbic_Region yeah, I got it the third time I read it 15:39
TimToady he's in "Middle College", a strange cross between high school and junior college, and the orientation was to go to an organics farm and pick tomatoes for three days. 15:40
and all the taking and fetching adds up... 15:41
15:41 marmic_ joined
nothingmuch did a bit of that in highschool 15:42
it was an agricultural school
before i could pick cinema i had 1 day of work per week
nothingmuch worked in the greenhouse
the big win was that people who majored in biology related fields had various experiments 15:43
TimToady beings as this is Northern California, I'm sure the emphasis was more on the "organics" than on the "farm". :)
nothingmuch my favourite was strawberries =)
heh
when i was there it was organic
by the timie i graduated they started using pesticides for the veggies and chemical fertilizers for the flowers 15:44
TimToady but mostly it was just something for the whole class to do together as a team-building experience, I suspect.
nothingmuch i think it was to keep people like me from eating the experiments
=)
TimToady organic farming would be easy if it weren't for those darn pests...
nothingmuch aye 15:45
especially since todays market is so intolerant to produce that doesn't look perfect
TimToady reminds me of software design...
15:48 xdg joined
svnbot6 r13333 | fglock++ | v6 - more WHAT conversion 15:48
TimToady heh, one more week and I will have been running the same irssi process for 6 months.
nothingmuch ever since OSDC? 15:49
TimToady since Mar22
nothingmuch oh right, it's september
TimToady I guess that's pretty much the uptime on my wall.org server.
nothingmuch my uptime got screwed when the second ethernet card popped out of the PCI slot =/ 15:50
TimToady obviously the power has been pretty stable here, considering I don't have a UPS on that machine any more
nothingmuch one would think power would only become stable after you buy a UPS
TimToady maybe it doesn't work like umbrellas 15:51
nothingmuch well, that's the case here anymoose
15:51 bernhard joined
TimToady I think the power becomes more unstable so as to fry your UPS more quicly. 15:51
*quickly
nothingmuch oh... hmm. that's a good point.
TimToady In any event, the only UPS that has remained functional for me is the one on my fishtank. 15:52
the rest seem to have a halflife of about 6 months. Maybe I'm just jinxed when it comes to UPSs though.
nothingmuch are they tropical or something? like, they need heat or they die?
TimToady just for aeration 15:53
nothingmuch ah
TimToady there's a great deal of thermal stability in a 100gal fishtank
nothingmuch wonders how much that is in m2
TimToady and they can live for a while without the main filtration pumps as long as the bubbler keeps bubbling oxygen into the water
nothingmuch .3 15:54
that's a lot
.37
TimToady It's about 400 kg of water.
nothingmuch *nod*
like a bathtub and a half?
TimToady and you can bet I worried alot about how to anchor that sucker down for an earthquake. 15:55
it's not something you want landing on your foot, or any other part of your anatomy. 15:56
nothingmuch not something you'd want to clean up either
TimToady well, that probably can't be helped if it sloshes. 15:57
but slosh is better than crunch
nothingmuch i'd figure it would smash
or is it plexiglass or something?
TimToady given it's a saltwater tank, one also must worry about electrocution.
yes, plexiglass 15:58
so that circuit has a ground fault interrupter. 15:59
which as also been known to kill my server since it's the same kitchen circuit for both...
15:59 integral joined
TimToady in fact, what usually kills the server is the refrigerator kicking in while both computers are up and we're making toast and coffee at the same time. 16:00
nothingmuch my dad once accidentally put the washing machine on the lighting grid
... with a space heater
TimToady I've been meaning to split that circuit for about 15 years now, and even have the wires, but the cable routing in the attic is rather problematic. 16:01
nothingmuch how so?
too much stuff there?
TimToady getting through that last 4 feet to the circuit box involves getting under where the roof comes down really low. 16:02
nothingmuch ah
that's what children are for
TimToady and the existing conduits seem too full to fish something through.
nothingmuch is it std for circuitry to go through the attic? 16:03
we don't have attics here... it never occured to me
TimToady around here, yeah 16:04
plumbing tends to go down below in the crawlspace
well, the hot circuitry goes in the attic, but all the phone and ethernet tend to be below.
basically, if you spring a plumbing leak you don't want to be crawling around in well-grounded mud with lethal voltages in handreach. 16:05
nothingmuch in .il plumbing generally goes below the floor, and cabling is either dug into the wall and then covered in plaster, or is embedded in irrigation pipes 16:06
the soft plastic ones
usually they get bent in the concrete
so they make them redundant
quite hackish ;-)
nothingmuch wonders how you tell P::G var nodes that their sigil is changing due to a subscript 16:11
16:13 kanru joined, vel joined 16:14 justatheory joined 16:25 pbuetow joined 16:26 Psyche^ joined 16:28 Psyche^ is now known as Patterner
xinming_ hmm, for contextual variable, what will it return by default if the key doesn't exist? 16:32
eg, $*ENV<not_exist> or $+not_exist
an error or undef? 16:33
nothingmuch TimToady: in p5, are labels for loops in an entirelyt different category than labels for statements?
agentzh i expect the former to be undef and the latter to be an error.
xinming_ shakes hands with agentzh. 16:35
agentzh :) 16:36
PerlJam I expect them both to be undef except in the presence of pragmata used to modulate such behavior.
agentzh is still playing with jifty on win32. 16:37
[particle] plays with jifty on win32 also :)
agentzh wow
[particle] i wrote an app to manage use case definitions 16:38
xinming_ PerlJam: But the problem is, %*ENV is env hash, and $+contextual access...
hmm
not the problem
just the suggestion. because, they are not the same in some sense. :-)
[particle] thinks $+foo de-sugars to $*ENV<<foo> 16:39
xinming_ [particle]: No, $+foo is declared by env IIRC, and %*ENV is built-in, 16:40
then, $+foo and %*ENV<foo> would be better to become the seperated thing, except searching for $+foo fails. It would equal %*ENV<foo> 16:41
TimToady and therefore a non-existent $+foo should return undef 16:45
well, it would be an "unthrown exception", actually, which would fail if you try to use it, or fail to test it. 16:46
nothingmuch: no, they're the same as far as the parser is concerned. It's just that loops pay more attention to statement labels than ordinary statements do. 16:48
16:49 hexmode joined
svnbot6 r13334 | fglock++ | v6 - 'Any' compile-time class 17:04
TimToady ooh, that number is extra 13334! 17:05
17:10 putter joined
putter the other 6... the r6rs draft is out, after, hmm, several, years. www.r6rs.org/ woot. 17:13
lambdabot putter: You have 1 new message. '/msg lambdabot @messages' to read it.
Title: R6RS.Org
17:13 turrepurre joined
putter the reference implementations will be a nice aid to making scheme-on-p6 tightly compliant. 17:16
svnbot6 r13335 | fglock++ | v6 - join 'Value' compile-time classes into single file
17:18 ludan joined 17:23 zeriod joined
putter talk & 17:36
nothingmuch nothingmuch.woobling.org/Perl-Gener...s_tree.rtf 17:39
lambdabot tinyurl.com/la73u
nothingmuch someone please go over this if possible
i can export to other formats
zeriod is svn.perl.org/perl6/doc/trunk/design/syn where i should be getting the newest synopsis 17:42
lambdabot Title: Revision 12083: /doc/trunk/design/syn, tinyurl.com/evycd
[particle] zeriod: yes
nm: what other formats can you do? the output looks funny for me 17:43
nothingmuch [particle]: omnioutliner 2 or 3, opml, html, plain text (tab separated, expanded tabs, or something else that i can't remember) 17:44
[particle] text or html would be great
agentzh zeriod: try out perlcabal.org/syn 17:45
lambdabot Title: Official Perl 6 Documentation
nothingmuch nothingmuch.woobling.org/Perl-Gener...tree_html/
lambdabot Title: , tinyurl.com/elxco
agentzh particle: does jifty's autocompletion works for you? 17:50
*work
[particle] nm: did you generate the topic list?
nothingmuch [particle]: no, i wrote it by hand
i'm rehauling Perl::Generate's AST
[particle] agentzh: haven't tried that yet
agentzh when i hit the enter key, it commits the page instead of completing the text in the textbox. 17:51
that is very annoying.
[particle] eew
nm: you're definitely missing some special block (PRE/POST...) 17:52
nothingmuch [particle]: that's perl 5, not perl 6 =)
[particle] ah
nothingmuch perl 6 can have a cleaner AST
i'm trying to model something that is easy to write by hand
[particle] __DATA is missing trailing __
nothingmuch given enough type info nodes can be coerced correctl 17:53
y
and then we get higher order macros at the price of not having perl's syntax
but for Moose & friends this is OK
[particle] i don't see 'no'
agentzh particle: it seems that the autocompletion problem is specific to my firefox. ie 6 works fine for the same page. 17:57
[particle] agentzh: i just upgraded ff to 1.5.0.7 today... maybe that'll help?
agentzh i'm already at the latest version. firefox updates itself automatically. 17:58
[particle] nm: need 'tie' or a 'tieable' role?
missing ? regex quantifier 17:59
nothingmuch ? is not a quantifier, it's more of a meta quantifier
i put it as a flag on the quantifiers
[particle] yeah
oh
nothingmuch hmm... what is tieable? 18:00
[particle] s/oh/ok/
nothingmuch any var, no?
any deref too
hmm... i suspect this is just LV
i need to clean up the string stuff a bit
i haven't yet decided, but i think it'll have it's own sub ast
that can be reused
and must be present in an appropriate quasiquoting op (q, qq) 18:01
instead of being just some kind of "constant" thing
nothingmuch needs food 18:02
agentzh @tell obra jifty's autocompletion functionality does not work properly on my windows machine. when i hit the enter key, the whole form is committed instead of completing the text in the box. but my ie works fine for the same page. am i missing something obvious? 18:03
lambdabot Consider it noted.
[particle] agentzh: the #jifty folks are quite helpful
agentzh particle: i've just asked there, but no body has replied. 18:04
18:04 mauke joined
obra agentzh: I've seen that locally myself recently. 18:04
lambdabot obra: You have 1 new message. '/msg lambdabot @messages' to read it.
agentzh obra: i'm happy to see you can reproduce it. 18:05
obra agentzh: I believe it's part of the "don't submit the page, but actually click the right form submit field when hitting return" bug fix
the javascript there needs to be smarter
agentzh oh, so it's a known issue. :)
obra yeah.
agentzh hopefully it will be fixed *pretty* soon. ;) 18:06
in the meantime i can use ie as a workaround. :)
[particle] or mouse instead of kb 18:07
agentzh yeah
zeriod in cygwin i get: "Can't locate Pod/Simple/HTML.pm" when i do: "~/pugs/util/smartlinks.pl --check --fast *.t"
agentzh zeriod: you need to install Pod::Simple. 18:08
zeriod how do i go about doing that?
agentzh preferrably via the cpan utility.
[particle] > cpan Pod::Simple
lambdabot Not in scope: type constructor or class `Simple'
obra or hitting the right arrow or tab seems to do the right thing 18:10
agentzh okay
indeed 18:11
obra: thanks for the help.
jifty is lovely.
obra I'm glad you like it. what are you building?
agentzh i'm just trying out wifty. :) 18:12
i'm still new to jifty.
and also new to mason.
[particle] wifty i couldn't get working. i'll try again after i get my new laptop and install half of cpan again ;) 18:13
agentzh hehe.
particle: maybe you'd like to try out the jifty ppm package.
18:13 elmex joined
[particle] is just glad that half of cpan installs on win32 18:13
agentzh that is very helpful.
[particle] agentzh: i would!
agentzh :D 18:14
TreyHarris agentzh: hi :-)
agentzh TreyHarris: hi :)
TreyHarris i think it's been a week or more since we've both been awake and on channel at the same time :-)
agentzh TreyHarris: i'm about to leave.
TreyHarris course :)
agentzh i've just said "g'night, folks" on #jifty. :) 18:15
2:12 AM here...end of day for me...
TreyHarris lol, well, take care, it was nice to see you briefly ;-)
agentzh me too :D
night &
18:16 agentzh left 18:20 BooK joined
svnbot6 r13336 | trey++ | [t/operators/reduce_le1arg.t] 18:29
r13336 | trey++ | Better smartlinks.
18:33 mdiep_ joined
nothingmuch returns 18:41
[particle]: any more comments? 18:49
[particle] i don't see void context 18:50
nothingmuch [particle]: that's just a non last statement 18:52
[particle] i think it looks pretty complete, then. 18:54
oh, qr? qw?
18:54 araujo joined
[particle] and 'local' 18:57
nothingmuch qr is a quasiquoting op 19:12
missing
qw is under constants
local... hmm
i guess it's a sub decl
err, var decl
no, it isn't
you can localize any lv
[particle] agreed, any lv 19:13
nothingmuch cheers for local @hash{qw/foo bar gorch/} = ($moose, $elk, $gopher) 19:16
19:21 zeriod joined
zeriod doing the test smartlinks, would it be correct to link pugs/t/oo/can.t to S12/Introspection? 19:21
TreyHarris zeriod: yes, why not? 20:08
zeriod just checking i understand what im supposed to be doing :D
20:17 sockstat joined 20:30 zeriod joined
TimToady ~. 20:34
TreyHarris cu: disconnected
20:37 zgh joined 20:47 jferrero joined
TimToady interesting, I only had a half-duplex ssh connection. 20:50
20:54 weinig joined 21:03 weinig joined 21:06 fglock joined 21:10 KingDiamond joined 21:11 fglock joined 21:14 fglock joined 21:19 larsen_ joined 21:24 Limbic_Region joined, polettix joined 21:28 cjeris left 21:36 hareldvd joined
svnbot6 r13339 | fglock++ | v6 (misc/pX/fglock) 'emit( $ast )->perl' generates code; needs more work 22:20
22:22 fnave joined
[particle] fglock++ 22:22
fglock [particle]: the code is very ugly, because I'm translating the existing emitter - but misc/pX is a scratchpad area 22:23
it can do simple compile-time things like ' 1.WHAT.say ' compiles to ' print "Int\n" ' 22:26
[particle] fglock: it's been too long since i built pcr and friends. what's the make target?
fglock make target? 22:27
clkao need to fix those tests being disabled!
fnave Is someone still doing perl6 list summaries?
[particle] fnave: yes. subscribe to [email@hidden.address] for weekly summaries
or visit pugs.blogs.com 22:28
fnave Thx!
[particle] fglock: to build and test and use v6 and friends, what do i do? 22:29
fglock [particle]: just install as a plain cpan module - install v6
[particle] cpan v6's
fglock mm - there are some failures from cpan-testers - it looks like a dependency version problem, but where? 22:34
[particle] fglock: line 632 if ($subname eq ... ) would be better as a map, i think
s/map/grep/
fglock [particle]: yes, I know - it will be a hash, some day :) 22:35
[particle]: I'm currently rewriting other things 22:36
the emitter just grew without a plan
[particle] looks like a lot of this could become multi-level dispatch
fglock sure 22:37
[particle] i'm still installing v6 and deps (new perl install, missing many modules) 22:38
fglock the v6 project works like: make it work -> refactor -> repeat
clkao but things are broken! 22:39
fglock clkao: we are in the 'repeat' part
[particle] :) 22:40
clkao by definition refactor should keep what's working working.
fglock clkao: yes, sorry 22:41
clkao: I'm trying to split the emitter into manageable chunks 22:44
nothingmuch
.oO( // is cool, but with an && ish counterpart we could have something nicer...
take the expr defined($a) && defined($b) && $a == $b
i'd like a meta chaining operator
$a <<< &&, == >>> $n 22:45
$a <<< &&, == >>> $b
where && is defined-and
22:45 Khisanth joined
nothingmuch and <<< >>> is some funny syntax 22:45
) zzzzz
(somebody remind me i dreamt of this when I wake up
fglock @tell nothingmuch remember the // && dream 22:46
lambdabot Consider it noted.
dakkar nothingmuch: write a macro!
mauke liftM2 (==) $a $b 22:47
using the Maybe monad or something
nothingmuch mauke: that's the inspiration 22:48
lambdabot nothingmuch: You have 1 new message. '/msg lambdabot @messages' to read it.
nothingmuch lambdabot: tell me later?
;-)
dakkar: i meant std
zzzzzzzz
22:48 markstos joined
[particle] @tell nothingmuch he had a <<< crazy >>> dream 22:50
lambdabot Consider it noted.
fglock clkao: are you running a v6 smoke?
TreyHarris anyone have a suggestion of a good way to communicate a code review to someone else over the internet? a unified diff with annotations?
fglock TreyHarris: you can work it online with gobby 22:51
[particle] gobby++
sobby is installed on feather, even
TreyHarris too much timezone mismatch, we don't have time to do it in real time 22:52
obra has gobby gotten visible cursors yet?
fglock going home &
22:52 fglock left
dakkar TreyHarris: that's what I did when I had to do it... 22:53
I added comments and code fixes, and sent the diff
TreyHarris dakkar: yup. thanks 22:55
just wanted to be sure there wasn't some obvious tool i was missing
23:14 justatheory joined 23:26 weinig is now known as weinig|bbl 23:29 bcorn joined 23:34 larsen_ joined 23:44 mako132_ joined 23:51 weinig|bbl is now known as weinig

Powered by Cro and the Raku® Programming Language.

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

Thank you!