pugs.blogs.com | pugscode.org | pugs.kwiki.org | paste: sial.org/pbot/perl6 | <stevan> Moose... it's the new Camel ":P | .pmc == PPI source filters!
Set by Alias_ on 16 March 2006.
00:00 SamB joined 00:29 trym joined 00:37 azuroth joined 00:44 bsb joined
audreyt TimToady, azuroth: The Ps/Pe + Bidi table is available at 00:49
svn.openfoundry.org/pugs/src/Pugs/Lexer.hs
scroll down a couple pages for "balancedDelim" 00:50
svnbot6 r10256 | audreyt++ | * run_pge.pir: set Pugs/PGE external-parrot bridge to 01:02
r10256 | audreyt++ | accept utf8 and output utf8, enabling unicode rules.
r10256 | audreyt++ | pmichaud++ for the hint on the treating-filehandle-as-arrays
r10256 | audreyt++ | trick of manipulating its layers.
audreyt (this doesn't fix embedded parrot) 01:03
?eval "\197" ~~ /./
evalbot_10255 Match.new( ok => bool::true, from => 0, to => 1, str => "\197", sub_pos => (), sub_named => {} )
audreyt ?eval "\197" ~~ /../
evalbot_10255 Match.new( ok => bool::false, from => 0, to => 0, str => "", sub_pos => (), sub_named => {} )
audreyt have fun :)
meppl gute nacht 01:04
01:20 Alias_ joined
Alias_ Wow those topics sick 01:20
errr
stick
Freudian :)
seen audreyt?
jabbot Alias_: audreyt was seen 17 minutes 14 seconds ago
Alias_ audreyt: So, auto_install()... it is safe to remove if there's no features or options? 01:21
s/safe/safe|preferable/
audreyt yes.
Alias_ And it's still needed for options/features 01:22
?
(since I don't see any other obvious user-interaction code)
audreyt it's needed so the user get to pick features 01:35
Alias_ ok 01:38
So the path to factoring it out is remove for no-interaction immediately, finish ::With, then reimplement the user-interaction somewhere 01:39
AutoInstall is feeling increasingly mitochondrial
Not part of M:I, just sort of lumped in there doin it's own thing
:) 01:40
audreyt was,, it _is_ mitochondrial :) 01:49
01:49 jserv-- joined
particle_ S02 specifies '::' as the sigil for package/module/class/role/subset/enum/type... 01:55
also included are grammar/regex/token/rule, correct?
audreyt grammar, yes. 01:56
regex/token/rule, no.
they are either just &
particle_ ah, of course.
audreyt or explicitly invoked as <rule> within rules
particle_ perhaps those should be added to the synopsis, for my edification :)
audreyt post a patch?
particle_ will do
can i svn co? is there a public interface? 01:57
obra audreyt: you're coming to oscon? 01:58
particle_ obra: you're coming to yapc::na?
obra particle_: I am 01:59
particle_ it'll be great to meet you 02:00
obra likewise 02:01
audreyt obra: yes 02:02
particle_: svn.perl.org/perl6/doc/trunk/design/syn
particle_ thx, just got doc/trunk 02:03
i'll submit diff from that root
audreyt k
obra audreyt: you seeing my /msgs?
audreyt sorry, forgot nickserv 02:04
obra no worries 02:06
audreyt pmichaud_: where is it specced that "self" means "$/" in rules? 02:10
it seems deeply wrong in many levels. 02:11
pmichaud_ audreyt: I don't think it's specced explicitly 02:15
audreyt: I agree it seems deeply wrong in many levels. But this is what was discussed at last summer's hackathon (more) 02:16
audreyt I remember that being discussed, but I didn't remember larry made a ruling... but maybe I wasn't there...
pmichaud_ and this is how it came across in some of the earlier versions of S05, as well as A05
audreyt but it's a very bad idea.
pmichaud_ PGE doesn't actually use it as "self", nor does it create rules as methods 02:17
audreyt unspecced one at that.
yup, neither does Pugs
pmichaud_ it was discussed a bit on p6l last fall
audreyt I think they are better off as class methods, or (even better) normal subroutines in mixin-like rule modules 02:18
I mean, grammar modules
but one way or another, having $.foo and $<foo> mean exactly the same thing
doesn't seems like useful to me
pmichaud_ ...is that the case?
I didn't think that $.foo and $<foo> were the same
audreyt inside a rule's production 02:19
they are the same.
which worries me.
pmichaud_ ...why is that?
audreyt because "self" is "$/"
and $.foo desugars to self.foo as scalar
and $<foo> desugars to $/.<foo> as scalar
which turns out to be the same thing
pmichaud_ I'm not sure it is -- at least, it's not that way in PGE 02:20
audreyt assuming the A12 object-serving-as-hash idea
pmichaud_ I didn't catch the A12 object-serving-as-hash idea
I knew about the relationship of $.foo and self.foo, yes
audreyt in any case, we still have self.<foo> the same as $<foo> 02:21
pmichaud_ but I didn't think that $.foo would also be self.<foo>
audreyt and $.from the same as $/.from
pmichaud_: hm, come to think about it, Match defines its own "as Hash" methods
so $.foo won't become self.<foo> 02:22
pmichaud_ yes, exactly
audreyt it only does that, according to TT, for classes with no explicit hashification forms
mea culpa
pmichaud_ np
audreyt but still... what is the use case for $.from being the same as $/.from :) 02:23
pmichaud_ it's not so much that there's a use case, just that this is the means by which I've been able to get things to work 02:24
audreyt I understand :) it's just, I don't see the benefit of exposing this to userspace 02:25
I'll stop bugging you and let you do useful work in #parrot :) 02:26
pmichaud_ I don't plan to expose it to userspace. In fact, it's been this way for almost a year and nobody's noticed "how" it works until they start inquiring and the terms don't make sense :-) 02:28
audreyt ooh, that means Match needs boxing :) 02:29
(which is fine)
nod. it's jsut a bit confusing to think PGE's Match and Perl6's Match
as the same thing
but if it's clear they are drastically different
I think that'd be fine :)
02:45 Cryptic_K joined 02:56 weinig_ is now known as weinig 02:57 weinig is now known as weinig|sleep 03:28 dduncan joined 03:29 aeon__ joined
dduncan I was just looking at audreyt's new blog entry ... and from that I realized I had a few extra reasons to attend OSCON 2006 03:30
until reading that, I was on the fence as to whether it would be worth it, having gone to the 2005 one
still undecided, but now leaning further towards going
fyi
when I say "worth it", I mean that I enjoyed the 2005 one a lot, but was wondering how much new stuff there would be over that 03:32
fyi
audreyt nod :)
you not going to YAPC::NA?
merlyn <<== going to YAPC::NA and YAPC::SA, but not OSCON 03:33
audreyt I was like that
until yesterday
now I'm going to all 3 :D
merlyn oscon is no longer about perl
no point in me being there
audreyt exactly 03:34
hcchien mutt
merlyn I'm not spending money on any more parties, etc.
hcchien ouch, sorry. :p
audreyt I'm there to talk with mono and ruby and js people.
not particularly perl folks :)
dduncan as I'm still limited financially, relative to some of you, I have preferred OSCON to some other large events due to it being close enough that I can go there in a bus rather than needing to fly
merlyn anyone I could have influenced is already known to me.
dduncan but do want to go to YAPCs as well before too long
s/bus/car-pool 03:35
merlyn the YAPC::SA should be interesting
with lwall and audrey and me.
miyagawa <- going to YAPC::NA, OSCON and YAPC::EU. Not YAPC::SA :/
audreyt miyagawa: oh, btw. you okay with YAML::Syck and JSON::Syck be released under the same term as libsyck 03:36
aka the MIT (or BSD1) license?
dduncan merlyn, do you mean that OSCON has no interest to Perl folks, or that it just isn't as focussed on Perl as it was before?
miyagawa audreyt: sure
merlyn It's becoming dead to Perl
audreyt k, thx, it will go in effect next release
merlyn fewer talks, fewer special events
really no point for perl people to go
and this is partially because the OSCON organizers are openly hostile to Perl 03:37
dduncan I thought it was an official annual perl conference, with state of the onion talks and such
audreyt well, I wouldn't say "dead"
"the same status as python"
dduncan merlyn, is that so?
03:37 hikozaemon joined
merlyn Yup 03:37
audreyt is more like it
merlyn Not web 2.0-ish enough. :)
Perl - still runs half the known net-verse. But not interesting to O'Reilly any more
audreyt and instead of "hostile" I think it's just "not _hip_ anymore"
indeed. 03:38
03:38 KimJongIl joined
merlyn hopefully, p6 will bring back the ruby crowd 03:38
KimJongIl perl6 here yet?
audreyt evalbot_10255: say "I'm here"
?eval say "I'm here"
evalbot_10256 OUTPUT[I'm here ] bool::true
dduncan well, if I do go, I'll do the Haskell tutorial for sure
on the other hand, if I decided to go to YAPC::NA instead, is it too late to get reasonable flights and/or register? 03:39
audreyt KimJongIl: we have a prototype version running a few IRC bots. :) 03:40
so if "here" means #perl6, yes, the bots are here
(written in perl6 themselves, for months now)
KimJongIl prototype?
anything for production use?
i have run out of people wanting to mine the uranium deposits 03:41
and they refuses to be powered by rock soup
i need robotics powered by production quality perl 6
ETA?
miyagawa dduncan: I don't think so, since I booked a flight to Chicago yesterday
audreyt KimJongIl: by Christmas.
merlyn it's pretty null - thedailywtf.com/forums/71684/ShowPost.aspx
KimJongIl audreyt: that sounds familiar...
merlyn Perl needs to support "Null" :)
dduncan what meaning of "null"?
audreyt KimJongIl: that is the officially canonical answer 03:42
svnbot6 r10257 | audreyt++ | * Pugs.Embed.Parrot: make utf-8 strings when embedding
r10257 | audreyt++ | Parrot/PGE. This means proper unicode rule handling
r10257 | audreyt++ | in both external and embedded parrot.
audreyt we refuse to give other answers :)
dduncan Perl's "undef" is already like some senses of "null"
audreyt dduncan: no, undef+42 isn't undef
null+42 would be null
dduncan so that's a different meaning of null
KimJongIl audreyt: so when my friends say perl 6 will never come, is that the same thing as christmas? 03:43
audreyt KimJongIl: I don't know. they used to say that Mozilla will never come.
_I_ used to say that.
KimJongIl audreyt: no one uses mozilla
audreyt there are some firefox users, I hear.
KimJongIl 10% of power users use firefox
audreyt it's the same project, same engine.
right. if 10% of power programmers use perl6
I'd be very, very, very happy. 03:44
KimJongIl the other 90% of power user are too busy doing other stuff, like having sex
audreyt I think you answered your own question :)
KimJongIl dev.perl.org/perl6/faq.html
okay i should have heed your answer 03:45
there's no date
dduncan this isn't the same thing per se, but I once proposed on p6l that the behaviour of undef be changed to say null + 42 = null ...
Larry strongly disliked it 03:46
and in retrospect, I like undef's current behaviour much better
KimJongIl i guess i should be using ironpython and learn to like "production-quality" commercial crap
dduncan where it acts like the given type's "empty" value
audreyt KimJongIl: fwiw, ironpython isn't much more production-worthy compared to pugs.
dduncan that works a lot better for me than undef + 42 = undef
audreyt esp. if you run it on Mono.
azuroth I concur 03:47
KimJongIl audreyt: i am not going to run on it mono
no offense
audreyt oh. you are on win32 exclusively.... then I recommend Visual Basic.
it's really nice. 03:48
(seriously.)
and production quality too.
KimJongIl well you see, visual basic is not really production quality, not many bots on this channel were written in vb 03:49
so it must suck
audreyt well, there are VB IRC bots
dduncan perhaps if I attend OSCON 2006, I"ll propose/give a lightning talk demonstrating how one can do SQL-like things in native Perl 6, using syntax that is both as similar as possible but better all the same
audreyt it's just we don't deploy it in #perl6 :)
dduncan who needs to embed another call layer when you can use Perl 6 itself
azuroth not many clients on this channel were written in windows, so ...? ;-)
audreyt dduncan: that would be very nice! 03:50
dduncan indeed
and you know from real life that this is one of the most common things people do with programs ... talk to databases
KimJongIl goes back to working with oss software as oppose to fan boying about it 03:51
dduncan Perl looks more "with it" if it can do things itself
I'll defer talking more about this in detail until later
buu dduncan: That sounds damn cool. 03:52
audreyt dduncan: you've read about LINQ?
dduncan I can't say I have
audreyt (the same thing for Visual Basic)
dduncan I know some languages like Coldfusion or Foxpro have db language built into them 03:53
I figure perl 6 should too
as is reasonable
audreyt oh. msdn.microsoft.com/data/linq/
the DLinq part corresponds to what you are talking about
azuroth are there decent free vb.net compilers? 03:54
audreyt no. mono can run the vb.net compiler 03:55
dduncan what I propose for databases in perl 6 is that perl 6 would include an interface for what looks like relational data structures in the language itself ... and it would have a default implementation that just has them in RAM like normal array/hash/set/etc variables ... but that one can tie these to other implementations that make them persist somewhere ... as one can tie hashes etc to DBM files
audreyt there is also SharpDevelop
dduncan and in perl 6, such DBM ties are no longer in the core, but third party
audreyt azuroth: there is a mono-vb compiler 03:56
but I'm not sure of its current state
dduncan: definitely read the LINQ material :)
dduncan okay
audreyt erik did exactly that, by making query methods work across a lot of classes
azuroth: www.mono-project.com/VisualBasic.NET_support 03:57
alpha-review status, seems to be useful to play with already 03:58
dduncan note that I only mention SQL for familiarity ... I'm not going by any means going to make the interface look like SQL for the sake of that ... more likely the interface will look very different, but wrappers would make it more SQL like
no reason to limit oneself
back in a bit
azuroth oh, cool. we were thinking of trying vb.net for a TAFE project, but now we're trying C# instead 03:59
audreyt TAFE? 04:05
azuroth turtle and four elephants. college 04:07
audreyt ah. so it's not turtles all the way down 04:08
KimJongIl audreyt: are progress being made in perl 6 at a worthwhile pace that one might imagine a near future date where asking when perl 6 might be done might be a meaningful question?
or are we looking at stagnation?
audreyt KimJongIl: we are definitely not looking at stagnation. 04:09
some would estimate _this_ christmas, some would estimate next.
of course, some may estimate year 3000 as well
oh wait, that's another language... 04:10
KimJongIl okay i give up
i tried, i really tried
i'll munch on some zen cookies instead
when it's here, i'll be here, and we'll all be taking five years to learn it and spending ten more years leveraging it 04:11
azuroth ?eval my @a = (1,2,3,4); @a >>*<< (1,2)
evalbot_10257 (1, 4, 3, 4)
KimJongIl until then, i guess it's just the state it is in and the bot writers, have fun
audreyt KimJongIl: alternately you can, well, improve the docs and examples and code :)
want a commit bit?
azuroth ?eval my @a = (1,2,3,4); @a >>*<< (1,2) xx * 04:12
evalbot_10257 Error: unexpected end of input expecting term
azuroth ?eval my @a = (1,2,3,4); @a >>*<< ((1,2) xx *)
audreyt azuroth: Whatever is not yet parsed
evalbot_10257 Error: unexpected ")" expecting term
azuroth ?eval my @a = (1,2,3,4); @a >>*<< (1,2) xx 2
audreyt or indeed supported ;)
evalbot_10257 (1, 4, 3, 4, 1, 4, 3, 4)
azuroth ?eval my @a = (1,2,3,4); @a >>*<< ((1,2) xx 2)
evalbot_10257 (1, 4, 3, 8)
azuroth ahh :-)
buu Horrifying 04:15
azuroth my eval..?
audreyt $job, bbl & 04:18
dduncan ditto 04:22
Alias_ KimJongIl: Perhaps I can give you your answers 04:24
You'll see an Alpha this year, and a Beta next year
I think that's a fairly safe call, right audreyt?
audreyt right. 04:25
Alias_ As to whether you'll see a beta this year, or a production release next year.... maybe
But then reading back over the last hour, he's almost a troll, so I'm not sure that answer helps 04:26
audreyt the way to tell trolls from non- is acceptance of a commit bit. 04:27
(to me)
FurnaceBoy_ ++(++audreyt)
Alias_ :)
So, when's the alpha due? :) 04:28
TimToady what is this "due" thing? 04:30
Alias_ Oh come on, you like stage shows. I'm sure somewhere in there you've got a "it would be nice to have something for $event" 04:31
TimToady: Also, you really need to ask people not to transcribe your talks... they come out weird
TimToady um... 04:32
Alias_ Video's would be great... but your talks don't translate so well to words. All the jokes dissapear
Videos
TimToady my wife married me because I make faces.
FurnaceBoy_ "everybody has something," they say. 04:33
TimToady she also married me because I would eat almost anything.
well, at least I can still make faces... 04:34
Alias_ Anyways... videos good, transcripts bad 04:36
Having only seen one talk in person, but read several transcripts 04:37
TimToady yeah, timing is everything when you're doing stand-up comedy.
Alias_ right
Standup comedy, without any timing, and without the aid of the visual props
That pretty much sums it up
04:38 dduncan left
Alias_ In fact, I recommend your next State of the Onion somewhere early includes an apology to all those reading the transcript :) 04:38
TimToady I could probably spend the whole half hour apologizing. :) 04:41
Alias_ hehh
There's probably a few of us that should be though 04:43
Anyone involved in Bundle::CPAN for a start
What amess
azuroth apologizing?
Alias_ For making such a cluster-fuck of the next-gen toolchain modules 04:44
By the standards of the previous generation
TimToady Some apologies are messier than others. :)
Alias_ audreyt: I've asked ANDK to pull Module::Signature support for CPAN.pm for a while 04:45
audreyt: More problems keep appearing, and there's nobody put up their hand yet to sort them out
(or at least, disable it by default, and not include it in Bundle::CPAN) 04:46
04:46 justatheory joined 04:48 Khisanth joined
TimToady how do we avoid a similar mess in Perl 6? 04:49
Alias_ Treat the toolchain to a different standard to $everything_else 04:50
If that means centralised control or something...
TimToady you mean, kinda like the Parrot model?
Alias_ Most of the problems seem to come from people not realising the impact of recursive dependency cascades 04:51
So they write some module, and release it, and 4000 modules break on Win32
Schwern's Test::Builder trick broke about 3000
Schwern++ # Exceptionall well recovered 04:52
Some of it is just not thinking 4th dimensionally :)
In fact, MOST of it is
TimToady is it something computers can help with, in theory?
Alias_ yep
We can go over this at the hackathon if you like 04:53
04:53 Khisanth joined
Alias_ But you should get a pretty good intro from the PITA talk 04:53
TimToady we've tried to address a few of the issues where they impact language design, but I'm sure that's the flag flying on the tip of the iceberg...
Alias_ nods 04:54
TimToady PITA?
Alias_ ali.as/pita/
I haven't talked about it much
Not till there's something to actually try out do I want people mucking with it
To be honest, MOST of the problem is located above the surface of the language, not below it 04:55
So there's no a lot you can do from a language design POV
There's certainly things to be done in terms of what you roll into the default "Perl 6" dist though
In fact, bootstrapping from the raw language up to a CPAN environment is half the problem 04:56
TimToady yes, I was thinking that earlier.
Alias_ And the other half is just quality control, which I think I'll have sorted out by the end of the year
CPAN Testers 2 et al
TimToady we need to be clear what the various definitions of "core" are. 04:57
Alias_ nods
rigorously so
With compliance tests
So RedHat can't do this missing-weaken shite
RedHat--
Some way of running the test suite against the binary package so you can stamp a "Perl" trademark on it 04:58
Similar to what Java has... that would be nice
I'd even commision the stamps for you if you did :) 04:59
TimToady unfortunately, at some point that takes both political and financial resources.
Alias_ Apparently the Perl Foundation has too much money...
TimToady: I don't think it's anything you'd want to spend effort legally enforcing 05:00
But if there's some script to lift the test suite off the build environment, so you can tell RedHat "run that", that would be nice
TimToady technological fixes are nice that way. 05:01
Alias_ right
TimToady and that's actually something that Perl 5 got a little bit right.
more out of necessity than anything...
Alias_ indeed 05:02
TimToady but we could do it so much better if we knew what we were doing.
Alias_ There's a reason that the best Documentation Management companies in the world come from South Africa
Because they need to be...
I see similar parallels to Perl
Anyways, see my talk after yours on Day 1 for the QA plan I want to use for both P5 and P6 05:03
And I hope to spend some time at the hackfest going over the 6PAN plans 05:04
TimToady sounds like a really good plan
Alias_ And I might solicit your help/name later this year if you don't mind, because I'd like to try and find a $big_company willing to donate a blade server or a small cluster 05:05
And my contacts in the relevant companies are negligable 05:06
TimToady well, I don't network much that way either... 05:07
but you can certainly whack people over the head with my name if you like.
Alias_ thanks
hmm... actually... I wonder if there might be some SGI stuff going cheap some time soon... 05:08
Anyways, I have an airport to finish 05:09
Alias_ &
TimToady ciao
05:11 mako132_ joined
azuroth what's a word for... saving something for later use? 05:20
I know, thesaurus 05:21
FurnaceBoy_ push :)
azuroth I think 'defer' works: "defer the sale and finish it later" 05:23
gaal postpone? 05:24
FurnaceBoy_ yeah, that's how postfix describes its "send later" queue
gaal { &something } ?
FurnaceBoy_ (deferred)
in abeyance :)
held over 05:25
gaal cryogenically treated?
azuroth postpone isn't bad. it needs to be imperative, for a button
I'm hoping to have half a year to work on pugs et al exclusively, soon. it should be cool. 05:30
FurnaceBoy_ wheeeee
christmas comes early this year 05:31
azuroth hopefully :-)
Alias_ wow, half a year? 05:32
eep, distracted, should be working
azuroth well, probably. might need to find a job to satisfy my parents, but hopefully they'll let me treat it just as more education or something 05:33
Alias_ How old are you? 05:35
azuroth just eighteen :-)
I'm at TAFE at the moment, doing a diploma 05:36
Alias_ blinks, TAFE? 05:37
Where?
azuroth tighes hill, newcastle
around three hours from sydney
Alias_ I know
I have family there 05:38
azuroth oh, wow
Alias_ Actually, I'll be visiting next week
Have to help my aging grandmother buy some headphones
azuroth heheh. that's a weird request. you're from sydney? 05:39
Alias_ Lismore
My grandmother is getting quite old now, she's well into her 80s
And she doesn't really trust anyone else to help her buy them
I helped her get her little CD player, and taught her how to use it
azuroth ahh 05:40
Alias_ Made sure she got one where she'd only need to ever press two buttons, play and eject
A lot of people born before the war struggle with technology 05:41
azuroth yeah, I suppose so
Alias_ I'm really pissed off with DVDs, because there's no way in hell she could navigate the menu systems
1. Put in video 2. Hit rewind 3. Wait 4. Hit play 05:42
So she can't watch movies any more now that everything is DVDs
Alias_ files that away for later under "Examples of why user interface design is important" 05:43
azuroth speaking of user interface design... want to do the gui for a point of sales system? *whince*
Alias_ I'm fully booked till YAPC 05:44
overbooked, actually
So much so I shouldn't be talking to you :)
azuroth hehehe
Alias_ will be in magnet #perl if anyone wants to talk 6PAN 05:45
05:45 Alias_ left
svnbot6 r10258 | scw++ | Pugs::Grammar::MiniPerl6: correct tests 05:57
TimToady ?eval "a" ~~ /a/; /^ <foo> $/; 06:11
evalbot_10258 *** Cannot parse PGE: ^ <foo> $ *** Error: Name 'foo' not found Match.new( ok => bool::false, from => 0, to => 0, str => "", sub_pos => (), sub_named => {} )
TimToady ?eval "a" ~~ /a/; "a" ~~ /^ <alpha> $/;
evalbot_10258 Match.new( ok => bool::true, from => 0, to => 1, str => "a", sub_pos => (), sub_named => { "alpha" => Match.new( ok => bool::true, from => 0, to => 1, str => "a", sub_pos => (), sub_named => {} ) } )
TimToady those both coredump with embedded... 06:12
my guess is that parrot is trying to reinitialize whatever it's using for $/. 06:14
that guess is based on the fact that the first ~~ must succeed to get the core dump. 06:16
Interestingly, changing /^ <foo> $/ to remove either of the spaces suppresses the core dump. 06:17
or changine <foo> to .
even changing ^ to ^^.
so it almost seems like a fencepost error somewhere. 06:18
azuroth this probably isn't the right place to ask, but anybody know any good books/resources about UI design? :x
TimToady I think most folks around here would rather just redesign the users... 06:19
azuroth hehe :-) 06:20
szbalint I wish. 06:23
06:33 ko1_away is now known as ko1 06:34 ko1 is now known as ko1_away 06:38 marmic joined 06:40 reZo joined 06:47 neoesque joined
TimToady pmichaud: by the way, my pugs smoke test is failing about 222 PGE tests because of the above coredump... 06:47
06:50 iblechbot joined
wolverian azuroth, _The Design of Everyday Things_, by Donald A. Norman 06:51
TimToady ooh, yes, good book.
"affordances". 06:52
wolverian yes. great examples. :)
azuroth thanks :-) I'm reading "the inmates are running the asylum" right now, and it interesting at least
s/it/it seems/
TimToady you'll never walk through the front door of a building without thinking about how the doors are misdesigned.... 06:53
wolverian and you'll comment about those misdesigns to your friends all the time, driving them crazy
azuroth, that seems like a good book too. thanks! 06:54
TimToady and if you're a language designer, you'll wish you could figure out how to apply some of the principles...
wolverian that's interesting. have you succeeded in that? 06:55
TimToady 'bout the closest I've come is "If a language construct is doing something weird, it should *look* weird..."
sort of negative affordances, if you will 06:56
wolverian that's a good idiom for doors as well :)
TimToady and just basically pay attention to all the people on PM that say "I tried to pull this thing but I should have pushed it instead." 06:59
kattana_ ahh so thats why so many of the FOSS people look the way they do.
TimToady well, if they consider themselves language constructs, I suppose...
But I think the perception that they're trying to do something weird is part of the problem, from the point of view of the Design of Everyday Things. 07:01
What seems weird now might be an Everyday Thing pretty soon.
azuroth is 'emotional design: why we love or hate every day things' as good/on the same boat? 07:02
Kattana you make "weird" sound like a bad thing, I find it rather synonymous with "interesting".
TimToady which is why we have to ask ourselves questions like "are scan and reduce operators actually going to be used frequently?"
azuroth: dunno
webmind good morning
TimToady Kattana: I'm just attempting to point out that, while "weird" is culturally determined, the culture itself is evolving to make some weird things not-so-weird. 07:03
Kattana Of course, things on the fringe always naturaly gravitate towards the center. 07:04
TimToady In that sense, design is an art, not a science, because you have to be a futurist.
Kattana Well we are artists if the code is beautiful. 07:05
TimToady well, I dunno if there's a "the center", but things on the fringe often end up defining new centers.
07:05 phpError joined
TimToady the big design problem is how to simultaneously be fringe friendly and center friendly. 07:06
wolverian oh, by the way, do we need a fold/fold1 distinction in perl6, since you can just do '($initial, @foo).reduce...'? just thinking about typing, as haskell needs the distinction. 07:08
07:08 elmex joined
TimToady don't need to do that at all for operators that define an identity multi. 07:09
Kattana I take the easy way out and settle for not being friendly :)
TimToady yes, well, we've seen plenty of that kind of design in the world too. :) 07:10
wolverian I wasn't thinking of operators, actually. just your generic :(Foo, Bar --> Foo) fold.
Kattana heh, sometimes thats the best way, if the fringe were too friendly with the center a new fringe would appear.
What we really need is a bigger mass in the center pulling stuff in, get the interesting ideas into everyday things quicker. 07:11
TimToady now I'm doing free associations with Oklahoma: "a pretty little surrey with a fringe on top". 07:12
means I should probably be in bed.
or turn the whole mess into a black hole quicker...
Kattana event horizons are interesting... 07:13
svnbot6 r10259 | scw++ | Pugs::Compiler::Rule compile_p6grammar.pl: deal with backslashed
TimToady especially cultural event horizons.
wolverian oh, well, this is sort of different than in haskell, given mmd. I have to get rid of the static mindset. 07:14
TimToady But, yes, Kattana, getting enough mass in the center to discourage random experimentation without having so much mass that you discourage useful experimentation is another way to think of the design space. 07:15
wolverian and making the mass a sphere instead of a cube.. 07:16
TimToady Certainly with P6 we're trying to simultaneously provide all the hooks you might ever need, while taking away nearly all the motivation for using them.
Kattana Theres the advantage of a black hole, it would shrink the center and let the fringe stretch out in its place. 07:17
TimToady tidal effects can be difficult
Kattana hmm its also hard to get out of the black hole once you are in it. 07:20
TimToady that's not nearly as uncomfortable a problem as when your feet are already in it and your head isn't.
Kattana its harder to fix, weird and interesting things often require change and people are loath to do that most of the time. 07:22
TimToady there's a sense in which you're allowed to do weird stuff after you've established the center.
we could only work on P6 because P5 was so "centered" already. 07:23
wolverian oh hm. I don't actually see how the identity multi helps with the :(Foo, Bar --> Foo) case. I mean, you want a :(Bar, Bar --> Foo) initial call instead. which, I suppose, isn't a problem, given MMD. heh. I'm seeing the light.
TimToady well, the lights are going out here, so I'd better retire. 07:24
nite all. &
wolverian good night!
Kattana 'night.
07:34 neoesque joined 07:47 drrho joined 08:02 Aankhen`` joined 08:15 bsb left 08:20 lisppaste3 joined
wolverian hm, S03 still talks about references with === 08:46
azuroth www.dogthebountyhunter.com/ 08:48
szbalint I wish people would stop those annoying flash detect pages... 08:58
*stop using
azuroth I'm more worried about the content than the medium. what the hell is it?! a tv show..? 08:59
szbalint well its a bit hard to worry about the content if you can't see it because of the medium. :) 09:02
azuroth good point 09:03
09:10 ghenry joined
ludan szbalint: flashblock.mozdev.org/ 09:16
wolverian can you give array elements individual types? I can see subscript types in S09, but they all return the same type. 09:20
azuroth like.. tuples? 09:21
wolverian yes, except arrays.. :)
I'm just wondering what kind of a type you'd infer for @bar in: 'sub foo (*@bar) {...} foo 1, "a"' 09:22
szbalint ludan: Yes, I'm using exactly just that. :)
ludan szbalint: it's my salvation :) 09:27
09:54 neoesque joined 10:03 r0nny joined 10:31 reZo joined 10:46 Aragone joined, Aragone is now known as Arathorn 10:58 iblechbot joined 11:10 reZo is now known as phpError
svnbot6 r10260 | scw++ | Pugs::Grammar::MiniPerl6 11:17
r10260 | scw++ | * Temporary variables now available
r10260 | scw++ | * MiniPerl6 spec updated
11:18 meppl joined 11:32 weinig|sleep is now known as weinig 12:02 mako132_ joined 12:21 Limbic_Region joined 12:24 hexmode joined 12:26 jsiracusa joined 12:37 kolibrie joined 12:44 macroron joined 12:48 Aankh|Clone joined 12:59 GeJ joined 13:09 fglock joined
svnbot6 r10261 | fglock++ | PCR - :ratchet update 13:09
13:14 Aankh|Clone is now known as Aankhen`` 13:15 jsiracusa joined
nothingmuch anybody got a machine where i can run perl5.005/perl5.004 ? 13:29
13:30 weinig_ joined
audreyt my laptop ;) 13:31
but no, not at the moment, I neeed to crash. maybe 10hr from now
you can build 5.005_04 on feather surely?
audreyt finds it curious that haskell.org got more than twice SoC applications than perl.org 13:32
obra . o O { Powered by PhD candidates }
audreyt indeed so 13:33
though we have our PhD candidate here (fglock) :)
13:34 vytautas joined 13:36 particle_ joined
szbalint I hope though that perl.org ones were of quality, rather than quantity :) 13:38
13:39 jserv-- joined
audreyt the quality are on par :) 13:39
fglock audreyt: I wish it was CS related :)
audreyt everything is CS related :) 13:40
so, the Meeting went well. I should start having slices of time on Capturizing starting tomorrow. yay.
but, sleep first :)
*wave*
Arathorn fglock: what it is it, if not CS, out of interest? 13:41
night audreyt
fglock Arathorn: neurology - some kind of wearable wireless device to record EEG 13:43
Arathorn fglock: very cool 13:45
a friend of mine was doing a Ph.D in a similar direction on computational neuroscience at UCSD
(although he gave up and went to work for google :\) 13:46
szbalint interesting.
13:46 Southen joined
Arathorn tries to make an EEG once with an ancient valve-based HP oscilloscope with two probes in differential mode, and lots of electrode gel. didn't get much other than mains noise, though... ;) 13:47
fglock Arathorn: heh - I had an interview with google
13:47 justatheory joined
fglock I once made an ECG using a ZX80 and lots of TTL, 741, resistors and assembler - it was very cool 13:49
Arathorn hehe, that rocks 13:50
13:52 cdpruden joined
fglock good times - I had to remap the char generator to RAM, in order to draw the graphics; and the A/D converter resistors were calibrated with sand paper 13:53
Arathorn did the ZX80 even have a graphics mode? 13:55
still, at least Z80 was a relative sane processor to hack on
13:57 stevan_ joined
fglock Arathorn: it was char-mode only, but the char generator was in EPROM - so I changed the address to RAM and wrote some software to allocate chars to draw the pixels and lines 13:58
Arathorn lovely hack :)
fglock can I use [[+]] to force list context in a scan? 14:02
OTOH, how about using [[...]] for scans, instead of the implicit context on reduce? 14:06
Juerd fglock++ 14:13
14:14 Qiang joined 14:15 flounder99 joined 14:25 theorb joined 14:36 nnunley joined 14:44 theorb is now known as theorbtwo 14:45 FurnaceBoy_ joined 14:57 drbean_ joined 15:00 penk joined
scw fglock: you added \r on eol in t/09-ratchet.t and lib/Pugs/Emitter/Rule/Perl5/Ratchet.pm ... 15:05
kolibrie scw: fglock: just run util/add-svn-props.sh on the file and checkin 15:11
15:25 iblechbot joined 15:28 penk joined 15:47 wilx joined 15:59 ghenry joined
Arathorn kolibrie: doesn't that clobber history? 16:10
16:10 bernhard joined
kolibrie Arathorn: I don't think so, I think it just allows newline conversions and such on checkout 16:14
Arathorn hm, i thought that as you re-check-it-in, it causes every line to change, as the repo standardises on its own canonical line ending style (i.e. \n under unix) 16:21
and so you then have to use an external diff tool (e.g. gnu diff -b) to ignore the line endings when diffing
at least, that's what happened to me, and i've yet to find way of working 'round it :( 16:22
particle_ it should just affect the svn properties, which are metadata
svnbot6 r10263 | fglock++ | PCR - fix t/09-ratchet.t eol - scw++
fglock scw: thanks 16:24
Arathorn how do you save the properties (given you can't set them remotely) other than committing after setting them locally, though?
Arathorn goes looks at the util script in question 16:25
particle_ it's true, you must commit.
Arathorn tests 16:26
pasteling "Arathorn" at 87.194.32.52 pasted "svn EOL fun" (42 lines, 1.2K) at sial.org/pbot/17128 16:33
Arathorn ^ particle_
if you (or anyone else) can suggest the magic required to avoid that... 16:34
Arathorn would love to know
PerlJam Just set svn:eol-style before you commit. 16:35
(the first time)
Arathorn sure - but if you're correcting a file which someone committed incorrectly in the first place? 16:36
PerlJam Then ... you do what you did. 16:37
Arathorn and the history gets melted
particle_ don't think there's a way around that, unless you put a trigger on the svn server
Arathorn minging
Arathorn nods.
(or tell everyone to populate up their .subversion files to autoset svn:eol-type for certain extensions or mime-types, like Apache do) 16:38
PerlJam particle_: that may be a good idea. Have it carp if the property isn't set on certain files (which you could glean from the suffix)
particle_ yep, parrot does that
pj: that is a good idea. 16:39
PerlJam particle_: make it so ;)
particle_ my delegation skills are a perfect fit for this job ;) 16:40
meppl guten abend 16:42
16:51 shachaf joined 16:53 Aankhen`` joined
svnbot6 r10264 | fglock++ | PCR - the 'Match' class is emitter-dependent 16:56
r10265 | fglock++ | PCR - :ratchet creates a Pugs::Runtime::Match::Ratchet 16:59
r10266 | fglock++ | PCR - MANIFEST update 17:08
bernhard On 'perl Makefile.PL' I get: Could not find module `Data.Yaml.Syck': it is hidden (in package Pugs-6.2.11) 17:16
Do I need to have that installed?
17:19 szbalint joined
gaal bernhard: update your source tree 17:26
bernhard No change. I already was at r10266. 17:28
17:29 theorb joined
gaal weird. any chance you have some stale files around? 17:32
rm Pugs.cabal for example 17:33
if that doesn't work try a clean checkout 17:34
bernhard I did a 'make realclean', removed extra file shown by 'svn status'.
No change with fresh checkout. 17:43
17:50 DaGo joined 17:52 theorb is now known as theorbtwo 17:54 weinig_ is now known as weinig 17:58 aufrank joined
gaal hmm I don't know, this worksforme 18:07
aufrank scw: ping?
PerlJam bernhard: start with ./configure 18:09
bernhard building GHC 6.4.2 18:32
18:37 chris2 joined
svnbot6 r10268 | fglock++ | PCR - from/to need to be calculated in 2 steps, it seems 18:44
18:44 hexmode joined 18:51 m joined 19:08 SamB joined
svnbot6 r10269 | fglock++ | PCR - :ratchet passes 2 tests 19:11
aufrank fglock++ 19:12
fglock "perl -Ilib t/09-ratchet.t | less" for some funny-looking perl6->perl5 code 19:16
19:17 p5evalbot joined 19:26 Odin-LAP joined, FurnaceBoy_ is now known as FB|afk
svnbot6 r10270 | fglock++ | PCR - :ratchet - 7 tests pass 19:29
aufrank fglock: when I try that I _do_ get lib in my @INC, but perl can't find PadWalker.pm
fglock aufrank: are you sure you have PadWalker installed? 19:33
(it's a separate CPAN module)
aufrank probably don't ;) 19:35
aufrank checks
fglock bbiab & 19:37
19:41 Arathorn is now known as Aragone 19:48 FurnaceBoy_ joined
kolibrie fglock: pretty cool looking 19:50
aufrank fglock: very cool :) 19:53
20:05 nnunley joined 20:06 nothingmuch joined 20:10 ruz joined 20:38 Odin-LAP joined 20:43 vell joined 20:49 kanru joined 20:58 qu1j0t3 joined 21:06 qu1j0t3 is now known as FurnaceBoy, vell joined
audreyt bernhard: try 21:11
ghc-pkg unregister Pugs
first
if that helps
bernhard Looks like I had ghc 6.5, from ubuntu repos, on my Linux machine. 21:12
audreyt does ghc-pkg unregister work? 21:13
bernhard After getting ghc 6.4.0 from ubuntu repos and building 6.4.2 things look fine
bernhard running make test
21:15 knewt_ joined
audreyt ooh. 21:15
if you are not too deep into it, try "make smoke upload-smoke" instead :)
21:16 knewt_ is now known as knewt
bernhard About halfway trough, being optimistic 21:18
audreyt k. :) 21:19
fglock audreyt: hi 21:21
audreyt hi 21:22
fglock audreyt: PCR :ratchet generates fully inlined code - it seems to be 10x faster than the plain PCR 21:24
this may be a good idea for actual p6->p5, not only rules
audreyt elaborate on the 'not only' part? 21:25
fglock there may be parts of p6 than can translate to inlined p5 code - maybe 'Int', 'Str', 'Num' for example 21:27
audreyt 'int', 'str', 'num' surely 21:31
not sure about boxed variants
maybe via autoboxing
fglock such as Int stringification could generate something like ($_ == $inf ? 'Inf' : $_)
I was thinking about autounboxing :)
audreyt right. I think generally the "is inline" flag can be on the prims
nod
21:34 vell joined
nothingmuch do we have a seen bot? 21:34
21:34 beppu joined
audreyt seen nothingmuch 21:34
guess not 21:35
stevan_ audreyt: he is floating around,.. lemme get him for you 21:37
buu jabbot did it 21:38
fglock & 21:40
svnbot6 r10271 | fglock++ | PCR - TODO update
21:40 fglock left 21:41 FurnaceBoy is now known as FB|afk, davidfetter joined 21:49 penk joined
audreyt nothingmuch: I've painted a porcelain dish for you :) 21:51
# pugs.blogs.com/photos/porcelain/lambdamoose.jpg 21:52
nothingmuch hugs audreyt 21:53
nothingmuch bounces
audreyt (there are other ones, but IrDA is really slow for getting pictures from this phone, so I'll upload them later)
nothingmuch hugs balas too for good measure 21:54
audreyt :)))
21:57 froh-doh joined
bernhard Failed 41/521 test scripts, 92.13% okay. 101/9252 subtests failed, 98.91% okay. 22:03
audreyt gaal: and this one is for you :) # pugs.blogs.com/photos/porcelain/moose.jpg 22:04
davidfetter waht teh m0053? 22:05
er, m00se
audreyt davidfetter: it's... the new camel
search.cpan.org/dist/Moose/
davidfetter resists urge to start telling jokes about ed meese & hunter thompson 22:06
audreyt bbiab 22:08
Toaster heh, the lamdamoose rocks 22:24
nnunley Moose and Squirrel. 22:26
22:28 penk left
davidfetter pages nina 22:28
22:41 hexmode joined 22:46 vell joined 23:02 OutThere joined, OutThere is now known as SubStack 23:10 cdpruden joined 23:19 AtomicStack joined 23:28 larsen joined 23:29 Cryptic_K joined
svnbot6 r10272 | audreyt++ | * Signature.pg: brief answers to aufrank++'s Qs. 23:32
23:33 shachaf joined 23:37 mako132_ joined 23:39 Quell joined 23:47 jsiracusa joined 23:50 mako132_ joined, Ymmv joined 23:59 kanru joined