6.2.12 released! | paste: sial.org/pbot/perl6 | pugs.blogs.com | pugscode.org | pugs.kwiki.org
Set by audreyt on 26 June 2006.
mdiep is the v6-pugs; source in the pugs tree? 00:07
(I guess it'd make sense that it is)
00:09 clkao_ joined
mdiep clkao_: is all the v6 code in the pugs tree? 00:11
audreyt mdiep: perl5/ 00:14
yes
perl5/PC-P6
mdiep audreyt: can I get a commit bit?
audreyt mdiep: email?
mdiep audreyt: [email@hidden.address]
audreyt welcome aboard, add your name to AUTHORS!
mdiep cool. thanks! :-) 00:15
audreyt dinner &
mdiep heh. my name's already in AUTHORS :-)
clkao_ mdiep: yes, i have a branch under pX/clkao, which does the data::bind-ism callconv and signature 00:16
meppl gute nacht 00:23
00:28 dolmen joined 00:35 buubot joined 00:39 buubot joined
TreyHarris so C<...> will returnan exception if use fatal is not set, but will throw an exception otherwise, right? 00:41
00:55 buubot joined 01:11 dolmen joined 01:12 coumbes joined 01:23 Entonian joined 01:46 weinig is now known as weinig|zzz 01:57 frederico joined
audreyt TreyHarris: right 02:06
02:31 scook0 joined 02:59 justatheory joined
nothingmuch performs an amazing spin kick 43 meters in the air, landing soundlessly on his head 03:06
03:08 markstos joined
markstos I'm migrating Class::ISA to Perl6 and am getting this error: No compatible subroutine found: "&length" 03:09
How should I call "length" now?
I think I found my answer in Differences.pod 03:10
nothingmuch ah 03:15
that's a pretty tough one
nothingmuch quits irssi to set timezone 03:16
03:17 nothingmuch joined
TreyHarris ?eval defined fail 03:20
03:20 evalbot_11061 is now known as evalbot_11065
evalbot_11065 *** Failed at <eval> line 1, column 1-13 03:20
PerlJam markstos: if you want "length in characters", you call .chars 03:21
markstos Perljam: Thanks, I found that. 03:22
Has something changed about how the ISA variable is used? That part of the code now runs, but causes an infinite loop....
nothingmuch markstos: the ISA variable is... errmmm. 03:23
you prolly wanna use the MOP for introspection
only we haven't gotten there yet
markstos Oh.
I'm basically trying to say "show me all the things I inherit from". 03:24
Does ISA even exist anymore? There' s not a single test which mentions it in the suite. 03:26
PerlJam markstos: not in that form, no. 03:28
perl6 will have lots of ways of introspection, but they're not all fleshed out. 03:29
markstos Ok. I guess my porting project is on hold for the moment, then. 03:30
03:32 stevan joined
nothingmuch markstos: short answer: you won't need Class::ISA in Perl 6 03:32
look in s12 (The bottom) for some previews 03:33
stevan markstos: you will be able to do something like: $foo.meta.class_precendence_list() to get all the parents in the correct ordering
that will be equivalent to Class::ISA::self_and_super_path 03:34
PerlJam I think I'll re-read the synopses before bed tonight. All sorts of stuff has changed and while I've kept up with some of it, a cohesive whole is better to look at that a bunch of parts :)
markstos Thanks for the answers everyone.
Is this a feature to expect to appear in the near term? Just curious. 03:35
stevan PerlJam: what I am talking about is not in Syn 12, Larry has basically said that the meta API is not part of the language, but instead part of the runtime (or at least this is how I understood it when audreyt told me)
audreyt markstos: Class::MOP backport? yes, that's when stevan arrives
stevan audreyt: I am here :)
audreyt in obra's house?
wow
somehow I missed you
come upstairs and say hi
stevan dont turn around,..
markstos You two... 03:36
PerlJam audreyt: you know how all geeks look alike
stevan the commisars in town ....
stevan keeps singing Falco in his head
its amazing the associations ADD will make
s/will/will cause you to/ 03:37
audreyt stevan has ADD?
I never knew that
stevan audreyt: if you dont know,.. I wont tell you now
PerlJam so ... what's the plural of moose? Is it meese?
markstos audreyt: Are you suggesting porting Class::MOP from p5 to p6? I'm confused.
stevan markstos: Class::MOP is a bridge to Perl 6
audreyt PerlJam: plural of mosse is moose
markstos: I'm suggesting implementing natively .meta
in the haskell runtime 03:38
nothingmuch but first it needs to be designed
markstos :)
nothingmuch *cough*
audreyt I thought that's what we did
in YAPC::NA
namely taking Moose and add Module/Package
that should be good enough no?
nothingmuch oh, so just CLOSish Class::MOP?
stevan stares blankly
nothingmuch i thought that was just for the bridge
audreyt no it's also for the hs runtime
nothingmuch what about the prototype object model?
audreyt should keep identical
markstos But right now, could do I the trick where I load p5 Class::MOP into Perl6? :)
stevan prototype, schmototype 03:39
audreyt markstos: you can certainly try :)
nothingmuch wants prototype object waaaaaaah
audreyt (but at which point you mise well write in p5 at this point)
nothingmuch: protometa can be done in p5 as well
nothingmuch waaaaaaaaaaaah
audreyt nothingmuch: just use anonclass gensym to emu
nothingmuch yeah
stevan markstos: yes,.. what audreyt said (damn these webbed fingers)
nothingmuch i think tha'ts what Moose does
sort of
stevan: maybe you should take a shower, so that the web can wash off? 03:40
markstos I'm really trying to port CGI::Application. So if I can port that Perl6, and replace Class::ISA with p5 Class::MOP, I'll be happy for now.
PerlJam stevan: are you a platymoose?
stevan will make nothingmuch a small protomodel to have for his very own
audreyt markstos: I think it's best to start a p5 port to Moose
nothingmuch markstos: i'm not sure you wanna go there
audreyt and we will deal with syntactic conversion after that
nothingmuch agrees with audreyt
also, C::A should probably stay in p5 land, for maximum compatibility 03:41
while a p6 version makes use of the crazier features
stevan markstos: I have many things to show you then,.. like our Moose layer over CGI::App that we use for work
markstos Whoa.
stevan markstos: did you talk to robkinyon at YAPC?
markstos stevan: I didn't even see him there. 03:42
03:42 miyagawa joined
markstos stevan: I'm not sure I know his photo to recognize him. 03:42
stevan markstos: ah, oh well
stevan orders nothingmuch to produce a moose picture
nothingmuch markstos: flickr.com/photos/nuffin/179250393/
lambdabot Title: "DSC_3329.jpg on Flickr - Photo Sharing!"
markstos So the consensus seems to be that it makes more sense to migrate CGI::App p5 from Class::ISA to Class::MOP as a first step ? 03:43
nothingmuch i would bet my money on some role based wackiness
stevan markstos: yes,.. keeping in p5 land for now
yes, roles are good,.. they are the perfect plugin system
nothingmuch and rely on the perl 5 compatibility bridges for C::A apps
well, not perfect, but almost
03:43 petdance joined
markstos Ok. That's only ironic because I started the project to be able to dabble in Perl6. :) 03:43
stevan markstos: 90% of Moose features will be easily (if not automatically) portable to Perl 6 03:44
markstos Oh. I did meet Rob, I just didn't know it.
stevan markstos: dont let us stop you if you wanna play with perl 6 :)
audreyt to play with Moose is to play with perl6 :)
markstos He was explaining why OO is better than procedural programming to an innocent bystander... 03:45
petdance Hey markstos!
nothingmuch i'd imagine that was a painful process
stevan ah, you were there for that,.. I head about it
petdance What are you doin' here in perl6 world?
markstos audreyt: I'm sure you are right. I'll learn the new object moose meta model thingie better.
petdance I think I'm gonna be alester over here.
03:45 petdance is now known as alester
markstos alester: playin' and learning. Tinkerin' with a CGI::Application port 03:46
nothingmuch alester: how was your flight?
markstos which turned into a Class::ISA port... which turned into this IRC conversation...
alester nothingmuch: What flight?
nothingmuch err, antiflight
alester I live in McHenry. 2 hours away from IIT.
stevan flight from the hackathon
nothingmuch (stevan just corrected me)
i assumed you would fly because for some reason i decided you were further east
alester We fed Randal and Alias
stevan alester: he meant how was you pizza,..
alester it was delicious
also, cute waitress. 03:47
stevan mmmm, wonder-bread and cheese
stevan feed chromatic, gaal, nothingmuch, robkinyon and allison
alester Hackathon at my house tomorrow! Whoo! 03:48
stevan I will wave from the plane as we leave chicago airspace
nothingmuch you can retroactively feed us
03:49 perlbot joined
alester Stevan: You will probably fly over my house. 03:50
We were talking about that on the way out here.
svnbot6 r11066 | lwall++ | Latest thinking on STM verbiage.
03:51 jkanter joined
stevan markstos: if you have moose/Class::MOP questions, feel free to ask away on #moose as well 03:52
over on irc.perl.org that is
markstos stevan: Thanks.
kattana_ hmmm it would be neat if you could use seti@home like grid software to take advantage of peoples home/desktop/workstation/dev computers during hackathons/lans to speed up things like compile&testing since many people bring laptops and not heavy duty machines, or at least link the laptops as enough of them would be equivilant to such a machine. 03:53
03:53 kattana_ is now known as Kattana
nothingmuch Kattana: ccache and distcc 03:53
apple has that with zeroconf support, actually
a.k.a "Bonjour"
Kattana looks them up 03:54
those are only C/C++ 03:55
nothingmuch true that 03:56
but pretty easy to implement for typical preprocess-compile-link cycle compilers
Kattana well, at least i dont have to do my own proof of concept. 03:57
nothingmuch note that ghc tends to be a bit funnier thoug
h
Kattana I had something more general in mind to speed things up tho, get the processing out of the way and take advantage of the time people are together 03:58
TimToady audreyt: I think I'm done thinking about STM for now. I didn't like sync/resync because it implies different threads are doing the same thing at the same time. After glaring at words for most of the day, I've settled on contend/defer instead. alternates are still "maybe" though.
Kattana TimToady: did you get my pm I sent a while ago?
obra contend feels a little weird as a name. but I'm coming at this as "j random perl5 hacker"
TimToady pm?
audreyt TimToady: sure. so maybe needs to line up?
Kattana PrivateMessage
audreyt also I'm imagining these are imported from some package 03:59
use Threads::STM;
to avoid clashing
TimToady I doubt "contend" will class with much. "defer" is a little likelier.
*clash
anyway, I like the metaphor of having an intellectual argument between threads 04:00
audreyt yup
Kattana wonders when we will be advanced enough to have flamewars between threads 04:01
nothingmuch Kattana: STM is kinda that
they stupidly try to restate their point till something listens ;-) 04:02
Kattana isnt that more like trolling. 04:03
audreyt Software Troll Memory 04:04
04:04 Debolaz joined
TimToady Kattana: I don't recall getting a pm. 04:04
Kattana I will resend then. 04:05
alester TimToady is Larry, right?
TimToady that's what they tell me.
nothingmuch Kattana: make sure you're registered with nickserv 04:06
TimToady or just send email
Kattana hmm.. ty
TimToady larry at wall dot org 04:07
clkao_ nothingmuch! 04:08
nothingmuch clkao_: !
i think you friended the wrong nothingmuch on flickr, btw 04:09
clkao_: so are we going hiking?
i'd like to tell my mother "this is the guy that's going to prootect me from bears and rapists"
Kattana pm sent, not worth the effort of email. 04:10
TimToady got it, thanks! 04:11
Kattana good, now get with the times "email is only for old people" :p 04:14
clkao_ nothingmuch: what is the actual plan
nothingmuch clkao_: rent/bus to Franconia notch state park 04:15
then as per jwang's advice go via the loop
mt. flume, lafayette, twins, maybe bonds too
depending on progress and rate
i have maps + book to clarify details
and my del.icio.us page has some untagged pages 04:16
clkao_ nothingmuch: that's where you showed me with google map the other day? and which day
nothingmuch leave thursday, return monday
clkao_ dude, my flight is on sunday
nothingmuch oh 04:17
so return saturday?
if we have a car we don't need as much schedule padding
clkao_ *nod*
you arrive tomorrow afternoon?
nothingmuch yep
clkao_ let's decide over dinner. that's easier
so tired
nothingmuch okay
spleep!
oh
rob tought us an awesome trick for the SSV dorms
when the guy is not paying attention, pass the blipper several times 04:18
robkinyon++
too bad we're the only ones still here ;-)
TimToady no we're not... 04:22
nothingmuch TimToady: still about? 04:24
well, you don't have stowaways, right?
TimToady I'm still here till 10:44 tomorrow morning. 04:25
nothingmuch hmm
then I guess gaal's ride doesn't hel pmuch
TimToady I traded a 7-day card with someone who's card lasts till tomorrow, so I'm all set.
*whose 04:26
but have to leave here about about 7:30
Yaakov Hello, Larry. 04:34
TimToady howdy.
nothingmuch TimToady: good luck waking up =) 04:35
TimToady I'll have to find my emergency toothpicks for propping open my eyelids... 04:36
but I contend that I should defer that till morning, maybe... 04:37
audreyt zzz & 04:39
TimToady me2 & 04:40
04:40 Entonian joined 04:43 fglock joined 04:44 cmarcelo joined 04:46 lambdabot joined
gaal meows 04:49
stevan wonders if that is some kind of moose mating call
gaal no, it's what happens when you stand in a party next to a moose and somebody askes who wants another beer and the moose turns to look at you and bonks you with its antlers. 04:50
stevan :) 04:51
markstos Switching from Class::ISA to Class::MOP was two lines for CGI::Application. One for the 'use' line, and one for the method call. Very nice. 04:56
I'm pushing 4.07_01 to CPAN now with this change.
stevan markstos: nice! my Class::MOP Kwalitee will go up ;) 04:57
fglock hi! 04:58
stevan fglock++ 04:59
fglock: I meet many people at YAPC who wanted to talk to you :)
fglock hi stevan!
stevan lots of interest in Pugs::Compiler::*
fglock I wish I was there
stevan I told them all to go to YAPC::SA
fglock cool! I'm sorry I couldn't get sponsors interested :( 05:00
It would be cool to see you here! 05:01
stevan fglock: I would love to be a globe trotting perl superstar like audreyt, but alas the wife and kids would miss me 05:02
fglock yes I know :)
nothingmuch well, as a globe trotting superstar you could make more money 05:03
fglock I'm rewriting the grammar - the cool part is that I can now decide whether to write it in P6 or P5
nothingmuch so the kids will have a replacement for love
05:03 alester is now known as petdance
stevan actually the dog is the one who is most disturbed by my absence right now 05:03
fglock nothingmuch :)
stevan the kids have my parents to entertain them 05:04
clkao_ fglock! 05:06
fglock clkao_: hi!
clkao_ fglock: i am making &?ROUTINE work 05:07
do you plan to refactor the current emitter to different node classes or some sort? it's getting too huge
fglock clkao_: the plan is to migrate to nothingmuch's emitter - ask him :) 05:08
I'm refactoring the grammar to emit something closer to p6ast
clkao_ ya, using the emitter, of course. but i meant the tree traversal part 05:10
nothingmuch winks at fglock
stevan wonders enjoys the meta-circularity of using a Moose based module to emit code which targets Moose
fglock clkao_: I'm ok with refactoring the emitter 05:11
stevan: that's pretty close to happen
nothingmuch fglock: we'll have a chat tomorrow when i'm at obra's, deal?
i need a requirement list 05:12
the p6ast is slightly closer to the metal than what I was targetting
but is a pretty easy mapping
clkao_ fglock: also nnunley is interested in working on a c emitter
fglock nothingmuch: the ast emitted by v6.pm is close to the grammar
nothingmuch clkao_: you were msged on irc.perl.org
fglock: i'm not sure what you mean, but what i've seen from v6.pm is pretty high level 05:13
fglock clkao_: cool - I started working on one - I think it can be very useful for bulding low-level libraries
nothingmuch which *is* what I'm working on
although it's not that high up (parse tree)
fglock nothingmuch: the v6.pm tree is the requirement list, kind of - do you need blessed nodes? 05:16
nothingmuch fglock: i abuse them a lot for convenience
can we punt this till tomorrow? i need to pack and sleep
fglock what time? this is too late for me - about 10h earlier is ok 05:17
clkao_ ok, &?ROUTINE.name works
nothingmuch err
clkao_ now to make CALLER::CALLER etc work
nothingmuch hmm
we land at 16:30 or so, i think
or is that 14:30 or so?
probably the latter
hmm 05:18
i think i can get you ~6 hrs earlier
maybe even ~8 hrs earlier
fglock clkao_: may I add you to the cpan v6.pm maintainers?
clkao_ fglock: sure!
fglock nothingmuch: ok
05:19 petdance left
clkao_ btw, caller().file is still parsed funnily 05:19
oh, i accidentally fixed that along 05:20
s/parsed/emitted/
fglock clkao_: I'm refactoring the parser - I hope to fix methods and such
nothingmuch =)
fglock nothingmuch, stevan: may I add you to the cpan v6.pm maintainers? 05:21
stevan fglock: sure
although I dont think I have done anything on it (at least not yet) 05:22
clkao_ fglock: i am committing my fix. the "call" and "method_call" emittion is nasty
nothingmuch fglock: okay
fglock stevan: it helps when people start reporting bugs :)
clkao_ i hope i didn't break other things. now testing.. 05:23
stevan fglock: okay :)
fglock clkao_: better test before commit... 05:24
clkao_ it passes tests! 05:26
fglock the cpan thing is done
clkao_ fglock: i can make caller().package work next. mostly runtime stuff for me. can you make regex and flag capture parse? 05:27
svnbot6 r11067 | clkao++ | Support &?ROUTINE and &?ROUTINE.name.
fglock clkao_: I'm moving the statement syntax from Yapp to rules, it should be easier to maintain
clkao_ then we are down to 2 things to make test.pm work
YAYA
i will watch carefully and learn from that
but now i need some sleep
fglock I think I'll fork until it works 05:28
clkao_ in pX?
should i merge my data::bind support changes before you branch?
otherwise it can be nastier later
the data::bind support needs a way to tell if the callee is builtin or other native p5 code. 05:31
05:33 fglock_ joined
clkao_ fglock: did you see my irc lines? 05:34
svnbot6 r11068 | clkao++ | pull from perl5/PC-P6.
fglock_ clkao: I backlogged - sorry, the computer died - 3rd time today 05:35
clkao_: can you make data::bind optional? I'm on windows, can't compile right now (I've got mingw, but it is not working)
clkao_ oh ok. i will try to make it so tomorrow 05:37
svnbot6 r11069 | clkao++ | * No capture for return yet.
r11069 | clkao++ | * No binding yet.
r11069 | clkao++ | * Requires Data::Bind.
clkao_ though it makes it a bit hard to test
fglock_: how is it not working? what perl5 version are you using?
fglock_ activeperl - it is asking for some header files
clkao_ ya. you don't have those? 05:38
fglock_ maybe - I didn't look too hard yet - I had it working in PxPerl before 05:39
clkao_ ok. but i think we really need those to support named args 05:40
but that'd be tomorrow (did I say that a while ago?)
fglock_ I'll try to fix my installation tomorrow
clkao_: do you know when do you have to use \( ) syntax ? 05:41
is ($a):=($b) the same as \($a):=\($b) ? 05:42
clkao_ \() creates a capture
i am not sure it can be lvalue 05:43
wow, i am the top committer for the last 100 commits! 06:00
(svk log --filter 'stats')++
06:01 fglock joined
clkao_ Newest commit : 2006-07-02 06:03
Oldest commit : 2005-02-06
Commits: 11069
Commits per day : 21.7
Committer count: 157
06:03 penk joined
clkao_ Most active committers: 06:04
- autrijus (2593)
- iblech (1203)
- audreyt (1016)
- fglock (992)
- putter (519)
audreyt needs to work harder to beat autrijust
fglock clkao_: would you commit a test for the rule syntax you need?
clkao_ simply ~~ s:P5:g/XXX/YYY/; 06:05
i really should go to bed ;)
fglock ok
06:33 fglock left 06:46 cmarcelo joined 06:52 scook0 joined 07:02 shachaf joined 07:09 beppu joined 07:30 visq joined, Aankhen`` joined 08:09 lisppaste3 joined 08:19 marmic joined 08:33 kane_ joined 08:36 cmarcelo_ joined 08:37 cmarcelo_ is now known as cmarcelo 08:46 dduncan left 09:12 iblechbot joined
marcus_ lala 09:21
09:41 marmic joined 09:46 larsen joined 09:49 elmex joined 09:52 kanru joined 09:56 szbalint joined 10:42 shachaf joined 11:07 buetow joined 11:12 drbean left 11:13 drbean joined 11:25 chris2 joined 12:03 mauke_ joined 12:18 mauke_ is now known as mauke 12:58 elmex joined 13:13 ludan joined 13:30 tup joined 13:34 penk joined 13:53 cmarcelo left 13:54 nothingmuch joined 13:55 justatheory joined 14:27 leptonix joined, weinig|zzz is now known as weinig 14:30 DaGo joined 14:54 joshr joined 15:11 lollan joined 15:13 leptonix joined, iblechbot joined 15:34 leptonix joined 15:38 buetow joined 16:01 Debolaz joined, Limbic_Region joined 16:07 miyagawa joined 16:39 fglock joined 16:49 amnesiac joined, justatheory joined 16:52 diotalevi joined 16:54 diotalevi joined 17:24 MenTaLguY joined 17:31 Odin- joined 17:50 nothingmuch joined 18:07 lisppaste3 joined
svnbot6 r11070 | clkao++ | Quick hack to make '&?CALLER::CALLER::ROUTINE' work. 18:14
r11071 | clkao++ | r14948@ab (orig r11070): clkao | 2006-07-02 14:13:38 -0400 18:17
r11071 | clkao++ | Quick hack to make '&?CALLER::CALLER::ROUTINE' work.
clkao_ audreyt: \Foo->new()->something
audreyt (\(Foo->new()))->something 18:18
clkao i want \(foo->new->something)
audreyt you get that, sorry 18:19
Deparse was borken
?eval 'abc' ~~ 'bc' 18:22
18:22 evalbot_11065 is now known as evalbot_11071
evalbot_11071 Bool::False 18:22
merlyn audreyt - did you see that I got pugs working? was a matter of chown -R on .ghc 18:24
I'm going to be writing an article on p6 for linuxmag this weekend
audreyt merlyn: oh cool 18:25
excellent!
merlyn basically just a textual form of the learning perl talk we gave
quick intro, using the things we can already play with
18:26 tup joined
merlyn probably also starting with C<print "Hello, World!\n"> to show that not much has necessarily changed. :) 18:26
wolverian then introduce say() and everyone will rejoice 18:27
merlyn exactly 18:28
I want to focus on how much perl5 really isn't going away, it's just getting a facelift, and most of everything everyone wants to keep is still there.
like damian's exegesis, but in my style. :) 18:29
audreyt :D 18:30
merlyn++
merlyn I'm more hungry for p6 to arrive now. It still scares me as a businees owner, but at least I know more of why I want it to arrive sooner rather than later. 18:31
I wish I had another N minutes per week to specifically contribute to a project, but I'm afraid it'll have be limited to doc review where I point out that .uniq is hard to pronounce 18:32
obra merlyn: that's why getting p6 semantics into CPAN modules is so important.
it'll provide an actual gentle upgrade path
merlyn Now that brian and I are building training materials, we'll be looking specifically at things that don't get "spoken" well.
hopefully, we can correct some of these design items before they get committed in hard stone, and everyone doing a code review will rejoice 18:33
theorbtwo I don't see a problem pronouncing "uniq" as "unique"... or for that matter just as "uniq".
audreyt yes.
ZuluOne indeed
audreyt Synopsis reviews, as the ones uri has been doing in the hackathon, is also a good way
obra . o O { ubiq }
merlyn You can pronounce it as "unique". then I will type ".unique" and it won't work
ZuluOne especially given the precedent
merlyn that's the problem
audreyt I like .unique, actually, since it's not unixy.
merlyn if you're going to use an english word, it needs to be spelled that way 18:34
and "pipe". argh. we're gonna wrestle on that one for sure.
MenTaLguY I don't think there's a pressing need to unix-spel things, yeah
merlyn pipe means "|" to millions of unix hackers.
even if you say "pipe operator"
they'll still think |
ZuluOne You now seem to be arguing for the opposite viewpoint.
18:35 justatheory joined
merlyn no - I'm arguing for common meanings for things 18:35
the "uniq" command in unix is actually fairly unknown
but nearly everyone uses "|" at one point
ZuluOne never
Who hasn't heard of uniq?
merlyn ZuluOne - surely you're jesting
MenTaLguY This probably isn't the best place to ask that question
ZuluOne no
merlyn if not, you have a pretty narrow view
MenTaLguY Probably everyone HERE has, but I don't think we're particularly representative 18:36
ZuluOne Maybe just a pretty narrow set of friends.
merlyn I work with "casual unix users" all the time
almost all of them do "|" and call it "pipe"
very few have ever typed "uniq"
ZuluOne typing it isn't the problem
merlyn I think we're talking an order of magnitude difference
possibly even *two* orders of magnitude 18:37
trust me on this as an instructor
and as someone who has been around unix since 1977
ZuluOne I can count the number of times I have *used* uniq on one hand, probably, but that doesn't mean I haven't heard of it.
merlyn again ZuluOne - you're the rare one here
ZuluOne nah, I'm well done
merlyn heh
merlyn has another 3.6 hours to kill at CVG 18:38
merlyn patpats his EVDO card
MenTaLguY what is it in p6 that uses "pipe" spelled out, btw?
merlyn the specs
they call ==> as pipe
MenTaLguY oh.
merlyn which means that everyone i scalling it that
MenTaLguY ee, that is going to be brutal
merlyn which meanss it'll get into error messages as that
theorbtwo I suspect the "uniq" spelling comes from math originally, and unix got it from there.
merlyn We need another name that means what ==> and <== does 18:39
"tube" ? :)
MenTaLguY channel?
ZuluOne I don't think people will be calling ==> pipe when reading out programs.
I certainly don't, when reading them in my head.
merlyn ZuluOne - you've never been in code reviews or worked with someone over the phone then
you need NAMES for all OPERATORS
trust me
MenTaLguY Zulu: what do you call it then?
ZuluOne Especially given they go both ways. 18:40
lumi Piping? Arrow-pipe? Channel?
theorbtwo connector?
ZuluOne I can't bring it to mind right now. Probably feedto and feedfrom
merlyn pipe is nice because it's both a noun and a verb
it'd be good to do another thing like that
ZuluOne I will pay more attention next time I am talking to myself. :->
merlyn that's just my gut
lumi Like "channel"?
MenTaLguY channel's the best alternative I've heard so far, but it'd be nice to distinguish ==> and <==
ZuluOne like "I'm channelling $larry"? 18:41
fglock audreyt: I'm rewriting the Yapp grammar to rules - it'll be only be used for precedence, so that it is closer to p6-on-parrot grammar
MenTaLguY the other thing is that you can't really google for "==>"
it needs a distinct name for that reason :P
theorbtwo Left channel and right channel works fine, I think. 18:42
audreyt fglock: good good
merlyn it'd have to be "channel left"
ZuluOne wait
merlyn to be verbed
theorbtwo Or channel in and channel out.
merlyn ugh "in" means which? :)
ZuluOne Which way does a channel left go? I don't find it sufficiently unambiguous.
merlyn I could see in meaning ==> or <==
yeah
those suck too
ZuluOne indeed
MenTaLguY hmm, actually googling for "==>" seems to break something
lumi foo ==> bar :: foo channels in bar?
merlyn heh
MenTaLguY you don't get the normal "no results" page
merlyn into
18:42 Aankhen`` is now known as Aankhen-EP
theorbtwo @foo ==> bar() -- channel @foo out to bar. 18:43
ZuluOne to and from work, I think
lambdabot Maybe you meant: faq todo yow
svnbot6 r11072 | clkao++ | * Emit for simple ~~ case (quotemeta exp2).
r11072 | clkao++ | * Don't bind yet for := when exp2 is not a scalar.
r11073 | clkao++ | r14950@ab (orig r11072): clkao | 2006-07-02 14:42:01 -0400
r11073 | clkao++ | * Emit for simple ~~ case (quotemeta exp2).
r11073 | clkao++ | * Don't bind yet for := when exp2 is not a scalar.
merlyn too bad "push" and "pull" mean something
theorbtwo @foo channel out bar parens.
lambdabot Maybe you meant: faq todo yow
MenTaLguY ==> and <== as the to and from operators
hmm
maybe
merlyn see - I could read either of those as "to"
ZuluOne not if you are reading code from left to right, surely
merlyn maybe I'm just weird
MenTaLguY heh. or "to" and "fro" [sic]
merlyn fro!
g'fro dat data!
MenTaLguY well, you know, "to and fro"? 18:44
merlyn yeah I know
I'd rather be home right now, but at least I had a 4 hour connection instead of a 32 minute connection
MenTaLguY besides, the notion of a "fro" operator is perversely delightful
merlyn and since I got in 90 minutes late, I'm pretty sure my flight later today would have been nearly that too
and I would have missed the last flight home 18:45
ZuluOne at result fro grep left curly dollar underscore greater than 2 close curly fro at input semi
merlyn so anyway, I'll probably be popping in here a bit more often to ask questions and suggest nomenclature changes
or at least bringing up the issues.
theorbtwo Anyway, there's no reason we can't make .unique an alias for .uniq.
ZuluOne or vice-versa
theorbtwo (Or at least not much reason.)
fglock merlyn: have you tried v6.pm?
merlyn fglock - I just barely got pugs to work again yesterday :) 18:46
I built pugs with an embedded parrot. I presume that's the normal model.
fglock merlyn: no pugs needed :) 'install v6' from cpan
merlyn Oh - no, I don't think I have
merlyn scurries off
MenTaLguY is aliasitude symmetric? 18:47
merlyn fglock - do you have a cvs/svn that I should be tracking?
fglock it's in pugs svn - perl5/Pugs-Compiler-Perl6/
merlyn ok - I have that already
wolverian theorbtwo, I think ruby shows that aliases aren't the solution. 18:48
merlyn I'm building git-svn mirrors of all svns so that I can see what's been changing at 30K feet
aliases actually suck
because peopel who don't know the alias wonder what the difference is
it's pretty distracting
theorbtwo Hm.
merlyn it also messes up googling 18:49
ZuluOne true, but you're going to make *someone* unhappy whatever
wolverian aliases are a slippery slope.
merlyn but as a designer, you *choose* to make some people unhappy
that's normal
MenTaLguY most aliases in Ruby exist because the original name was determined to suck (or was ungrammatical english), but had to be kept for backwards compatibility
clkao_ fglock: hey. i pretty much got all runtime support for test.pm done. only need :$negate in capture and ~~ s/// now.
theorbtwo How hard is it to just say "unique, that's u-n-i-q" -- or just pronounce "uniq".
merlyn a langauge cannot be all things to all people
MenTaLguY You don't really want to be designing in aliases from the beginning
clkao_ but the s// is tricky, involves '#' in it ;)
theorbtwo I like uniq, but possibly that has something to do with the fact that I use the unix command uniq constantly. 18:50
merlyn fglock - lots and lots of "use of uninitialized value in length" errors
is that normal?
perl is for far greater than unix geeks
fglock - I've never seen more errors during installation and test where it still says "ok" :) 18:51
fglock merlyn: no - try reinstalling Pugs-Compiler-Rule - check if tests pass 18:52
merlyn I think you set the record. :)
all of these from Pugs::Compiler::Rule line 24 and 48
s/Compiler/Grammer/
s/er/ar/
when I installed it last, the tests passed just fine 18:53
have you tested against 5.8.8?
heh... 90 lines of warnings, then "... ok" :)
fglock maybe PC-Rule Makefile.PL was not run? 5.8.8 here (windows)
merlyn when I installed PCR, I didn't have a working pugs before 18:54
so it used your files
anyway, this is humorous
I'll retest PCR in a second
after this completes the tests successfully
fglock Makefile.PL needs to be run in order to bootstrap the grammar (.pmc file)
merlyn fglock CPAN.pm does what it does 18:55
I didn't cheat on anything
and yet I have PCR installed
so Makefile.PL probably said "oh, you don't have pugs, we'll use this"
MenTaLguY silent "oh, you don't have X, we'll just use Y" things tend to be dangerous 18:56
or at least maddening
fglock v6.pm doesn't need pugs - there must be a different problem - Makefile.PL should report common problems, like an old PadWalker or an invalid .pmc file - maybe CPAN.pm can't see the errors? 18:57
clkao_ fglock: got a chance to see if data::bind works for you?
merlyn well in any case, the tests here are passing 18:58
just geting lots and lots of warnings
as I said, almost comical 18:59
I presume if things were broken, I would not be getting test success?
fglock clkao_: not yet - I hope I can finish the grammar soon, than I'll try again
merlyn: no messages in Pugs::Compiler::Rule Makefile.PL? 19:00
merlyn Uh - I can't test that yet
I did think I got something like "you don't have pugs installed, so I'm not recompiling anything"
I'm still waiting for v6 to finish installing
since I'm getting 10s of thousands of warnings. :) 19:01
once that's done, I'll go back and retest PCR
fglock Do you have an older v6.pm in your system? I've been bitten by this 19:02
merlyn I don't know
but the test harness is supposed to push the real thing there
perl -Mv6 didn't do what I wanted. :)
it says "perl 6 wanted, you have only 5.8.8" :) 19:03
ahh - perldoc -l v6 says nobody home
fglock v6 is special - it doesn't follow the normal rules
merlyn still waiting for millions of error messages to scroll by
theorbtwo Do it again with 2>/dev/null 19:04
?
merlyn so far, no test failures
no point in restarting
it's almost done I hope
theorbtwo Ah.
theorbtwo is off to pick up some chinese food.
TreyHarris merlyn: your discomfort with "to" and "fro/from" for the pipes--does the same discomfort exist with /<before ...>/ and /<after ...>/? 19:05
merlyn up to test t/07-try
TreyHarris - a bit
how many more tests are there?
I can't rememeber if its the text before me, or me before the text
ahh there. installed! 19:06
fglock merlyn: I don't think it makes sense to wait for the tests if you are getting these messages
merlyn all tests passed
now I'll try PCR again
Ugh - what's the name?
fglock Pugs-Compiler-Rule 19:07
merlyn got it
I had 0.07 installed already
and the tests there also get hundreds of warnings
I don't think you've tested this on 5.8.8
TreyHarris merlyn: i think, then, that Perl just needs to make a decision, make it consistent, and use it. without making all the ops visually mnemonic, and naming them complicatedly ("channel from thing to left to thing to right"), at some point a decision should just be made and stuck to. the problem in P5 and most other languages is that we have both left-associative and right-associative things that cognitively collide
fglock would you reinstall PadWalker?
merlyn "PadWalker is up to date 1.0" 19:08
fglock is your Perl compiled with pmc support? see v6.pm pod
merlyn probably not
I have a stock perl
5.8.8 bleed 19:09
soon to be 5.8.9 I hope :)
all tests pass on padwalker too
warning on t/closure because someone wrote a bad test
TreyHarris how does one test for PMC support? 19:10
merlyn squawk like a pirate. :)
fglock rt.cpan.org/Public/Bug/Display.html?id=20152
lambdabot Title: "#20152: can&#39;t find own compile method"
fglock TreyHarris: looking
you should not have 'DPERL_DISABLE_PMC' in the perl config 19:11
merlyn perl -V:'.*' | grep -i pmc
I don't have that for sure
... Compile-time options: PERL_MALLOC_WRAP USE_LARGE_FILES USE_PERLIO 19:12
svnbot6 r11074 | clkao++ | v6.pm - Support caller().package. 19:13
clkao_ i think module::compile won't pass test if there's disable_pmc
merlyn And I just installed Module::Compile
Oh - does that make .pmc?
that's cool
clkao_ yes.
merlyn makes note to use *that* for the SysAdmin column 19:14
clkao_ ingy's revenge on everyone hating spiffy
19:15 DaGo joined, Odin- joined
merlyn so now that I have v6, what can I do? :) 19:15
TreyHarris so, i don't understand, why does Perl 5 even need to know about PMC's? is it futureproofing against ponie somehow?
diotalevi The acronym predates parrot's pmcs.
fglock TreyHarris: pmc = precompiled perl module 19:16
audreyt TreyHarris: parrot .pmc files are never installed
merlyn yes unfortunate collision of terms
TreyHarris oh, ok.
clkao_ fglock: what grammar changes are you working on?
fglock audreyt: any ideas for merlyn problem with PCR? 19:17
merlyn perhaps the regex engine works differently
but oddly, it still passes tests
clkao_ merlyn: perl -Ilib lib/v6.pm -e 'say "foo"'
fglock clkao_: I'm moving the Yapp grammar in Operator.pm to Perl6.pm, and rewriting to Rule
clkao_ all of them?
fglock all 'statement' 19:18
almost done :)
merlyn can't open perl script "lib/v6.pm"
did you mean something different?
fglock merlyn: in Pugs-Compiler-Perl6/ 19:19
merlyn is there something I can just "do" in general?
from a normal command line? :)
fglock merlyn: would you paste the errors?
merlyn ... Can't open perl script "lib/v6.pm": No such file or directory
that error?
or when testing PCR?
fglock merlyn: PCR errors
merlyn OK hold one 19:20
clkao_ perl `perldoc -l v6` -e 'say "foo"'
pasteling "merlyn" at 70.217.181.166 pasted "from t/09 onward (the rest are the same)" (19 lines, 2K) at sial.org/pbot/18158
fglock see the synopsis here search.cpan.org/~fglock/v6-pugs-0.002/lib/v6.pm 19:21
lambdabot Title: "v6-pugs - an experimental Perl 6 implementation - search.cpan.org"
audreyt merlyn:
echo 'use v6-pugs;' > test.p6; echo 'say "Hi!"' >> test.p6; perl test.p6
merlyn that clkao line is giving me hundreds of the same errors
Debolaz audreyt: Any luck with the local installation issue? 19:22
merlyn But it did finally say "foo"!
19:22 MenTaLguY left
audreyt merlyn: you have some env that turns -w on all the time? 19:22
Debolaz: no, sorry, I havn't looked into prefix.
merlyn No I don't
I'm using 5.8.8 bleed though 19:23
clkao_ maybe the pcr on cpan is funny
audreyt Debolaz: can you backlog into the summary in irc and collect it and post it to perl6-compiler?
Debolaz: or add it to the TODO file 19:24
merlyn Yeah, I can't think of anything unusual I'm doing 19:25
I'm compiling bleed about every third or fourth day
fglock you should have a hello_world.pl in the installation dir - can you run it? perl hello_world.pl
merlyn in PCR?
fglock PC-P6
merlyn uh - be clearer 19:26
fglock or perl -Ilib hello_world.pl
merlyn I don't have hello_world.pl in my cpan build area
fglock cd Pugs-Compiler-Perl6; perl -Ilib hello_world.pl
merlyn ahh v6-pugs
fglock yup
merlyn yeah, that's giving me hundreds of errorrs again
fglock does it generate a hello_world.plc file? 19:27
merlyn yes, it didi
and it says "hello, world" at the end
audreyt s/errors/warnings/ :)
merlyn you just gotta get rid of those warnings. :)
who could be turning on warnings globally? 19:28
now that I think of it, that *does* seem odd.
nothing is supposed to have that power except -w
Debolaz audreyt: I don't subscribe to email lists and usenet posts are filtered afaik.
audreyt merlyn: you alias perl to -w?
merlyn No I don't
audreyt Debolaz: commit to TODO, then? or send me a hiveminder ticket 19:29
merlyn in fact, I generally hate -w
Debolaz audreyt: Will see if it can be done. :)
audreyt ok :)
Debolaz But do I have commit access?
merlyn this is a perl I built two days ago
audreyt merlyn: upgrade Text::Blanced?
merlyn from this rsync://ftp.linux.activestate.com/perl-5.8.x/ 19:30
I have the latest of that
the latest of Text::Balanced too
audreyt merlyn: ok, try svn up pugs
svnbot6 r11075 | audreyt++ | * stopgap measure to allow Text::Balanced to return undef stuff
audreyt and install the perl5/Pugs-Compiler-Rule/ there 19:31
Debolaz: what's your email addr?
merlyn that's in svn.perl.org/perl6 right? 19:32
I'm not seeing anything new
audreyt merlyn: the mirror takes some time; best to switch svn checkout
to
svn.openfoundry.org/pugs
lambdabot Title: "Revision 11075: /"
merlyn ahh - I'm not tracking that
lemme see if I can switch it 19:33
audreyt please recheckout
merlyn wait - is /pugs there the same as /perl6 ?
or just some part of it?
audreyt the perl6/pugs/ part of it
merlyn Ugh - I can't switch easily then 19:34
audreyt just recheckout...
merlyn svn co http:// ... ? 19:35
ok that's working
But that now explains why i waswn't seeing the updates to docs as quick as they were showing up here. :) 19:36
audreyt Debolaz: did you give me your email addr?
merlyn pulling now 19:37
does this archive contain anything that isn't in svn.perl.org/perl6 ? 19:38
audreyt Debolaz: note if you /msg me without identifying to nickserv I can't see the msg :)
merlyn does this archive contain anything that isn't in svn.perl.org/perl6/pugs ? [i mean] 19:39
audreyt merlyn: no, they should be identical
merlyn but just delayed
ok
merlyn watches many thousands of filenames whip by
merlyn is happy he has his EVDO card
the airline lounge where I'm located has a special drink during 4-7pm only 19:40
I saw the sign and said "too bad it's not 4pm"
he said "it's 4pm *somewhere*" :)
theorbtwo (Jimmy Buffet)++ 19:41
merlyn checked out
no META.yml :) 19:42
"make all test install"
no warnings!
audreyt good then.
merlyn farkin' a you did it
what's different?
audreyt no warnings uninitialized.
audreyt shrugs
merlyn heh
theorbtwo Did you put it in a small scope? 19:43
merlyn perhaps some warnings are on by default now
perhaps p5p needs to know this is a Bad Thing
audreyt theorbtwo: in that file all subs use Text::Balanced and trigger warnings
and there are already no warnings once on top
theorbtwo Oh. Nevermind, then.
merlyn maybe they broke the warnings mechanism
19:44 nothingmuch joined
fglock audreyt: btw, you are PCR comaintainer 19:44
audreyt cool
danke
theorbtwo merlyn: IIRC, there was a patch recently to fix certian cases where an uninit warning wasn't thrown due to a bug.
fglock merlyn: perldoc v6 should work now 19:45
audreyt theorbtwo: that sounds definitely like it.
merlyn I still don't understand how that commandline there could work
perl -Ilib lib/v6.pm
unless you were in the RIGHT directory 19:46
is that presumed?
audreyt yeah.
merlyn in which case it should say so
otherwise, it's a broken synopsis
fglock sure - it needs to be replaced by a normal script
merlyn I know there's no lib/v6.pm from my home directory
so that really isn't going to work at all
clkao `perldoc -l v6.pm`
;)
audreyt there needs to be a way to make 19:48
perl -Mv6-pugs -e ...
work
but I can't think of a way yet.
(problem is that @ARGV info isn't available to -M) 19:49
or we giveup and install bin/perl6 ;)
(would be bad, so j/k)
fglock how about bin/lrep 19:50
Debolaz audreyt: Sorry, switched window. :) 19:51
audreyt: My email address is [email@hidden.address] (Which also is the explanation for why I don't sign up for mailing lists unless most posts are interesting) 19:52
audreyt fglock: that would work, but I somehow still want to keep v6
but lrep is the best name we've got so far. 19:53
fglock sure
clkao fglock: if you are killing the yapp grammar - i have changes to the signature strucutre ,and splat => is_slurpy
fglock I mean, sure == v6
clkao other than that it's all in emitter
fglock clkao_: ok - I'll check it before merging - it is taking more time than expected 19:54
19:56 shachaf joined
audreyt Debolaz: both hiveminder and openfoundry intivations are sent 19:57
clkao_ audreyt: is there a pugs group on hiveminder? 19:58
audreyt clkao_: yes and you are in it
or at least invited to it 19:59
Debolaz: you can eg add the context information needed to get the --prefix done form irc to "up for grabs" in hiveminder.com/groups/38/unowned_tasks
lambdabot Title: "Hiveminder - Get busy!"
TreyHarris i never got a hiveminder invite... 20:01
i got an openfoundry one.
audreyt sent 20:02
TreyHarris hm. i guess i gotta contact my email admin and get him to check the spam quarantine. 20:04
ingy hola 20:05
fglock ingy: hola que tal 20:11
Debolaz Where are the channel logs at? 20:31
audreyt miyagawa: balloon.hobix.com/
lambdabot Title: "Balloon. (Scripty postcards.)"
audreyt Debolaz: irc.pugscode.org
20:41 DaGo joined
lumi How does one join this hiveminder? 21:09
audreyt very carefully 21:10
lumi Oh my
audreyt trying to do a batch invite
21:14 elmex joined
audreyt lumi: invitation sent 21:19
21:22 gaal_ joined
lumi Thank you 21:29
21:53 DaGo joined
audreyt np :) 21:59
22:02 cognominal joined 22:11 Debolaz joined 22:19 kanru2 joined 22:40 ramzi joined 22:51 MenTaLguY joined 22:56 fglock joined
fglock clkao: finished rewriting the grammar - it's very slow to load, will need to add precompilation soon 23:00
audreyt TimToady: ::?CLASS -- is it still needed? 23:13
isn't it just the same as $?CLASS?
svnbot6 r11076 | fglock++ | px/fglock - v6.pm statement grammar rewritten from Yapp to Rule
audreyt gaal also noticed there's no ::?GRAMMAR though there is $?GRAMMAR.
fglock++ fglock++
TimToady: also what does @?LABEL mean? 23:14
23:17 Debolaz2 joined 23:26 hubertb joined
clkao fglock! 23:28
svnbot6 r11077 | gaal++ | * p6ast - reasonably close enough to start hacking?
clkao i might look into doing precompiling
fglock clkao: I committed to px - I'm going through the tests now 23:29
audreyt gaal++
yay
clkao_ fglock: OOOH you are not branching with svk?
fglock clkao_: I could not get installed here 23:30
one sec - I'll see what's the error
clkao_ fglock: and you are not using a svn cp either... this will make merging a bit painful 23:31
fglock I don't have svn - I use tortoise-svn 23:32
clkao_ (!)
nothingmuch fglock: btw, i'm all yours
fglock windows is hard to use...
nothingmuch windows--
maybe you can pull it off with cygwin?
oh wait, instead of checking for you if svk supports windows 23:33
i'll let clkao
okay
he's too slow
svk.elixus.org/?SVKWin32
lambdabot Title: "SVKWin32 - Kwiki"
clkao there's windows binary
i will try to manually add a merge ticket for you, before we all lost track 23:34
fglock I've got a slow computer, with windows, and internet over phone - if it works here, it will work anywhere 23:35
clkao_ fglock: do you know which revision you based on
before my changes - r11062? or after any of them? 23:36
nothingmuch fglock: if you have dialup you *REALLY* want svk
pasteling "fglock" at 200.102.126.175 pasted "SVK install error" (32 lines, 1.5K) at sial.org/pbot/18168
fglock clkao_: checking 23:37
clkao_ fglock: just use the binary
fglock clkao_: downloading 23:38
clkao_ fglock: i think you are based the last change you had. 23:39
fglock: are the t-old things dead? 23:40
nothingmuch fglock: ping me
fglock ping 23:41
clkao_: I just removed the things I haven't changed 23:42
sorry for the mess
nothingmuch: pinging you 23:44
nothingmuch fglock: okay
so ping me again in ~2 minutes because I need to pee
and then we'll talk about Perl::Generate and see how I can make it not suck for you
fglock ok
svnbot6 r11078 | clkao++ | Stick a svk:merge ticket so we can track fglock's branch for merging.
fglock download will take about 1h 23:45
23:45 mdiep_ joined
clkao orz 23:46
audreyt oy, initial sync will take forever
nothingmuch fglock: you may ping me now 23:47
clkao_ fglock: don't worry, cleaned up a bit for you
fglock: what's the parsing state? i see 04-if not parsed 23:48
svnbot6 r11079 | clkao++ | Correct Makefile.PL to use module::install from the right place so we
r11079 | clkao++ | can actually run the tests.
fglock clkao_: tests are running real slow - I'm working on t\03 now 23:50
clkao fglock: oh ok. but it looks good. byebye yapp 23:51
23:53 buubot joined
fglock nothingmuch: ping! 23:56
nothingmuch pong 23:57
fglock re Perl::Generate - it would be nice if it could follow the v6.pm structure, but it is more important to be compatible with the other backends too 23:58
meppl gute nacht
nothingmuch i don't know 100% what is the v6.pm structure is right now 23:59
i need to go over this with you
fglock it generates a statement list, which has

Powered by Cro and the Raku® Programming Language.

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

Thank you!