The Return of the Journal : pugs.blogs.com/ | pugscode.org | pugs.kwiki.org | paste: paste.lisp.org/new/perl6 or sial.org/pbot/perl6
Set by GammaRay on 31 December 2005.
dduncan my smoke of r8786 on haskell runcore plus darwin (10.4.4) is done 00:15
Juerd "Usually ships: some day" 00:18
I hope they mean "same day"
00:28 Alias_ joined
Alias_ seen audreyt 00:28
jabbot Alias_: audreyt was seen 9 hours 28 minutes 48 seconds ago
Alias_ darn
Alias_ looks at watch, should be any time now 00:29
putter lol 00:32
00:46 bogey_lj joined
bogey_lj anyone who can answer some perl questions for me? 00:47
00:50 Khisanth joined
Alias_ maybe 00:51
purl, ask to ask?
Alias_ blinks
bogey_lj :)
Alias_ purl?
bogey_lj perl
Alias_ crap, wrong channel
bogey_lj says perl6
Alias_ And here am I thinking it's perl5 00:52
bogey_lj :)
hmmm
Alias_ on a different network
bogey_lj ah
00:52 vike joined
Alias_ with purl the bot 00:52
bogey_lj so no one in here knows perl
rhesa sure we do :)
bogey_lj oh 00:53
:P
I am having some ODBC connection issues
Alias_ Perl 6 ODBC support is available already?
rhesa are you using perl6? 00:54
bogey_lj I think at work they actually have 5.8 activestate
Alias_ So you are asking about Perl 5 in the Perl 6 channel?
bogey_lj possibly
qmole 5.8 is kinda close to 6
bogey_lj I am asking about win32::ODBC, regardless of which perl version
Alias_ bogey_lj: OK, just so you know, Perl 5 is as close to Perl 6 as JavaScript is to Java 00:55
bogey_lj k
active state even
?
Juerd "Unfortunately, your mailer does not support HTML mail."
Alias_ So a MUCH better and explicitly help channel would be just plain #perl
Juerd What do they mean by "unfortunately"?
bogey_lj fair enough 00:56
if anyone would be interested in helping me with some perl issues because the know perl, please message me, thank you
*they
Juerd bogey_lj: Join the #perl channel
putter bogey_lj: irc.perl.org #perl details here www.irc.perl.org/
rhesa Juerd: "unfortunately, perl5 does not support continuations"
bogey_lj I am in there...they are talking about religion 00:57
no one answered :)
Juerd bogey_lj: There are people who want to help you in that channel. Nobody wants to help you with Perl 5 questions in a Perl 6 channel. Or well, they shouldn't want to.
bogey_lj: Then wait a while and ask again.
theorbtwo bogey_lj: Try the chatterbox on perlmonks.org.
Alias_ Just like nobody would want to help with JavaScript questions in a Java channel :)
bogey_lj k, thx theorbtwo
Juerd rhesa: Well, that I can understand. But what is so unfortunate about not supporting HTML?
Alias_ Juerd: No pictures embedded ? 00:58
bogey_lj some programmers just like solve problems, I would help in a pm
putter bogey_lj: the freenode #perl channel is not what you are looking for. try irc.perl.org
Alias_ You MISSED OUT!
bogey_lj thx, though everyone
Juerd Alias_: I consider that quite fortunate
Alias_ putter: Oh yes it is
putter: He doesn't want to wander into irc.perl.org #perl
He'll be slaughtered
Juerd putter: Are you actually suggesting someone to ask a Win32 question on Magnet #perl?
Alias_ And they'll send him to Freenode #perl :)
rhesa Juerd: fwiw, i understand what you mean 00:59
putter is reminded that perl has a rather disfunctional community in some ways
bogey_lj I sure do want irc.perl.org
rhesa would #perl-help be better?
Juerd putter: What has bogey_lj done to you to deserve this? :P
Alias_ Juerd: I think the CamelPack and Vanilla Perl conversations have been the only Win32 Perl stuff in the last 3 years :)
bogey_lj :D
Iwhere is #perl-help
Juerd putter: It's functional, just not newbie friendly in all areas.
rhesa on irc.perl.org 01:00
bogey_lj k, txh
Juerd bogey_lj: Also consider efnet #perlhelp
rhesa so many places to get blasted
Juerd putter: Perl's community is old enough to have a separatist/elitist subculture.
Alias_ putter: Another way of saying it would be to have a community of module and book authors who spent the FIRST 10 years of Perl answering newbie questions and helping, and have burnt out and try to just yack and co-ordinate module development 01:01
Juerd putter: People who were once very friendly to beginners, and have helped hundreds of people, and just can't cope with doing that same thing any more much longer.
putter: And yes, there are also newcomers who like to join them because they're nice company sometimes. 01:02
Alias_ For example, this week in #perl we've helped one person with a problem, and released two new distributions of Perl for Win32 :)
Juerd putter: Those shouldn't be blamed. Many of them are *also* in the helpful channels.
putter txt for resp's. sorry, call
Alias_ The only questions welcome on #perl are those so fiendishly weird they are actually interesting :) 01:03
Juerd I myself have spent quite a lot of time helping people, but didn't manage to keep on doing that for very long. I have great respect for the people who have helped people much more than I have, even if they decide to no longer wish to be so forthcoming. 01:04
It is somewhat unfortunate that they are in channels called #perl.
Alias_ indeed 01:05
Juerd But that may be very hard to change.
Alias_ Personally, I prefer to help lots of people all at once by writing modules, rather than helping one at a time
Juerd I really like helping people individually.
Alias_ I gets hard after the first hundred times
Juerd The direct communication gives a nice satisfactory feeling afterwards, when the person is clueful and is actually helped. 01:06
However, after a thousand times, it gets tedious :)
Alias_ That said, after the first hundred CPAN modules it can get a little tedious too
:)
theorbtwo I can hardly claim a thousand times... but yes, it does.
Juerd I can imagine.
01:07 Cryptic_K left
theorbtwo When you start thinking to yourself "didn't I just answer that question", and it takes you a while to realize that it was somebody else, they don't talk to each-other... that's when you start to have issues. 01:07
Alias_ theorbtwo: Which is what's nice about the logbot
purl, seen obra? 01:08
jabbot Alias_: obra was seen 5 hours 43 minutes 4 seconds ago
theorbtwo wonders if hanging out on freenode #perl can help invograte him. 01:09
Juerd theorbtwo: I can claim that, and I have positive memories of the time when I spent most of my time helping people.
I wish repetition wasn't so boring.
(I'm heavily lagged, by the way)
</lag>
What is invograte?
Alias_ invigorate
Juerd Okay. What is invigorate? :)
Alias_ invigourate actually
from "vigour"
Juerd As you wish - What is invigourate? ;)
theorbtwo can't spell none too good.
Alias_ To replenish one's enthusiasm
theorbtwo Recharge. 01:10
Juerd (...! What's replenish?)
Ah, recharge I do know :)
Alias_ Juerd: To hunt for the koolaid
which is stupid that you will know :)
Juerd My problem with channels like freenode #perl is that they're TOO busy at times, and that they have too many newbies helping newbies 01:11
01:11 kanru joined
Juerd Alias_: Eh, that doesn't parse. Is it slang? 01:11
Alias_ I would have thought you would have encountered "Koolaid" before? 01:12
Juerd (Okay, get back to the regular technical or 6th grade (pick one) English, please)
theorbtwo To drink the koolaid is to trust your own marketing too much.
Juerd Alias_: Never
Alias_ "To drink the koolaid" is very common techie slang for believing in your own marketing
Juerd Hmm
Alias_ "To drink the Ruby Koolaid" etc
No idea where it comes from specifically
theorbtwo The term comes from a cult where everybody drank poisoned koolaid, beleving that their souls would be picked up by an alien ship. 01:13
stevan the electric koolaid acid tests
Alias_ ah
Juerd I'm in quite a lot of techie scenes and don't recall ever having seen it.
stevan theorbtwo: its older than that
theorbtwo Really, stevan?
Alias_ wikipedia to the rescue!
stevan I think it has to do with Ken Kesey and his Electric KoolAid Acid Test,.. where they drove a bus around in the 60s giving people LSD 01:14
theorbtwo: well the cult you are talking about was in the mid-90s,.. I think I have heard the term prior to that
but I wasnt in the techie community back then, so I might be wrong
Alias_: so... what does it say?
dont keep us in suspense man 01:15
theorbtwo stevan: I got my cults confused; the one in question is from 1978 (and actually drunk flavor-aid, a knock-off).
Alias_ One lasting legacy of the Jonestown tragedy is the saying, “Don’t drink the Kool-Aid.” This has come to mean, "Don’t trust any group you find to be a little on the kooky side" or "Whatever they tell you, don't believe it too strongly".
Juerd Let's have a holy war. I miss those.
stevan theorbtwo: ah yes,.. Jonestown
stevan decides to invade .nl in search of the P6 holy grail
Alias_ The phrase can also be used in the opposite sense to indicate that one has blindly embraced a particular philosophy or perspective (a "Kool-Aid drinker")
Hence, "drinking the Ruby Koolaid"
:)
stevan theorbtwo: you were right :) 01:16
Juerd stevan: Eh, I was more referring to technical holy wars, i.e. $product1 versus $product2
Alias_ Juerd: Where's purl when we need her
stevan Juerd: my p6 holy grail is better than your p6 holy grail then
Alias_ jabbot: start a holy war!
jabbot Alias_: Earlier you said your p6 holy grail is better than my p6 holy grail then.
Juerd Where one could argue that numbering the products in such way is already an unfair bias towards one of the two products.
stevan: I don't have a holy grail. 01:17
stevan cant seem to find his either
Alias_ I have one
Alias_ chuckles
Juerd Damnit, holy wars aren't what they used to be. I blame purl for showing their silliness.
stevan Alias_: yeah, well,... .au is too far away
01:17 grayson joined
stevan and it doesnt have the same type of "coffee" shops 01:18
Juerd Alias_: Doesn't the definition of holy grail exclude any possibility of finding one? ;)
Alias_ steven: We don't need to contain it in shops
Juerd
.oO( damn tourists )
Alias_ steven: I live 15 minutes from Nimbin
www.nimbinmardigrass.com/2005/
theorbtwo I was thinking of en.wikipedia.org/wiki/Heaven%27s_Gate_%28cult%29 -- who had vodka and applesauce, not koolaid.)
Alias_ Juerd: true... 01:19
stevan wondered how one managed to write 100+ modules and still be sane ... :P
01:20 Amnesiac joined
Alias_ stevan: You make sure there are no bugs 01:20
theorbtwo stevan: What makes you think Alias is sane?
stevan theorbtwo: nothing actually :P
Alias_ stennie: The big trick is not to get trapped into writing code for vague non-well-defined problems
theorbtwo Hm, modules where you can be sure there are no bugs sounds like boring modules. 01:21
Alias_ stennie: Those are the ones that accumulate a bug a week and you spend lots of time patching
stevan looks around and wonders who stennie is?
Alias_ crap
tab-complete
stevan :)
Juerd crap complete
stevan thinks Alias_ should lay off the *cough* stuff *cough*
01:21 MrFarts joined
Alias_ theorbtwo: True, but the problem is you can only accumulate two or three of these sorts of "exciting" modules before you grind to a halt 01:22
Look at audreyt :)
theorbtwo Hm, point.
stevan Alias_: well the great thing is that very few people end up using them too
Alias_ M:I + PAR + a few other crazy things have stopped her release rate
mugwumpjism Anyone have the link to the online version of TAPL?
stevan points to some of his odder modules
mugwumpjism: it's online??
theorbtwo has three modules... and I only know of any non-me users for one of them.
mugwumpjism stevan: it's a .cn url :) 01:23
Juerd is currently working on a paid $job, where the result will in great part be open sourced when finished
stevan mugwumpjism: I have a copy right next to me, but I dont know the URL
Juerd But I don't get to design the stuff, and the design imposed sucks.
I think there'll be bugs because of workarounds that are needed because the design isn't solid.
Alias_ ew
Juerd And those are, perhaps, the worst kind of bug. 01:24
mugwumpjism ah, +site:cn found it
www.qiji.cn/eprint/abs/2761.html
Alias_ Juerd: There's one worse... specification bugs
Juerd: Get a copy of Code Complete
Juerd Alias_: What's a specification bug? When the specs are wrong?
Alias_ yes
Juerd Those are horrible, but at least you don't see them coming.
theorbtwo Where your choices are to be broken because you don't match the specification, or to be broken because you /do/ match the specification?
Alias_ But when they arrive they are far worse than a design problem 01:25
Juerd The bugs I speak of haunt me already, before the code that will have them even exists
Alias_ No chance to talk about design tweaks
Juerd Oh, there's talk already. Just hard to convince people who "spent months" designing something, if you've been working with them for less than a month only. 01:26
01:26 grayson` joined
Alias_ true 01:26
01:26 grayson joined
Juerd The most heard category of responses to my alternative design was "But why not merge that thing with that other thing?" 01:26
Where the two things are different. 01:27
The same category: "Why split that?"
They completely misunderstand the point of refactoring, it seems.
They want to end up with something monolithic 01:28
Alias_ yeah
01:28 vel joined
Alias_ Baby Syndrome? 01:28
Juerd Dunno - don't know the term (Tonight is my "learning English" night, I guess) 01:29
Alias_ In fact, we can call it Gollum Syndrome now!
"my precious!"
Juerd Ah, heh. Certainly.
Oh, and another form of response in the same category: "Can't we just create a single module that does all this?"
Alias_ Juerd: English slang, when a project becomes "your baby"
Juerd Or: "But then you end up with lots of small scripts and files, how can that be good design?" 01:30
Alias_: I was thinking the people having this syndrome were compared to babies, instead of the project :)
Also funny, in the weekly development meeting yesterday, code reviews were announced. Every developer is supposed to spend a few hours per week reviewing. 01:33
This is a good plan, of course.
But there are no guidelines by which to review. There was no clear objective for reviewing. There is no table of who should review what.
There are no coding standards in this company yet.
01:33 bogey_lj left
Alias_ scribbles "should have used CPAN module Foo... and here module Bar..." 01:33
Juerd And... their existing code sucks, and they admit that. 01:34
Alias_ Juerd: Buy some copies of PBP and slap "coding standard" on it
Juerd But code checked in also still contains goto LABEL like stuff.
Alias_ Perl::Critic :)
Juerd Alias_: I wish I could agree with PBP.
I have recommended that they get some copies and read it. 01:35
But recommending use of source filters, even if only during development, I do not condone :) 01:36
Alias_ There's some dangerous ones in there true
Feel free to mix in my style guide
theorbtwo IIRC, PBP's into is great, he says things like "focus on big things, not silly details", and "always think about why purticular standards, and what standards fit your environment", and then gives fiddly-detail recomendations, and the community takes his word as cannon...
Alias_ ali.as/devel/code.html 01:37
heh, yeah
But Perl::Critic does allow you to enable/disable specific rules
So at a practical level you can enforce the subset you agree with
Juerd Alias_: Thanks, I'll reread that and maybe use parts of it
Alias_ It's bits and pieces I picked up doing bigger projects than Damian does 01:38
So it might have more things applicable to you
Juerd Uhhuh; I've read it before.
And have commented on it.
Alias_ goodo
I should consider wikifying it
Juerd Or someone else has written a very similar one :) 01:39
putter back 01:42
my god, just a wee bit of conversation to backlog
Juerd putter: A "pause" button would be a useful addition to IRC :P
theorbtwo calls it a night. 01:49
putter that's the pause button ;) 01:52
01:52 drbean joined
Juerd rehi 01:53
putter: No, IRC continues :)
Alias_ TIVO for IRC
Juerd Hehe 01:54
"Cannot connect"
[details] --> "Check if the foo daemon is running" 01:55
I'm used to having techie information under [details] buttons
Or, details
putter mugwumpjism: so what's a .chm file?
Juerd putter: Eek, is that the windows help file thing? 01:56
putter ah, yes
Juerd Some form of used-to-be-html-but-was-binary-crippled-for-inoperability format
putter :) 01:57
01:57 Debolaz2 joined
putter ok, time for pizza. bbl & 01:57
02:29 n0m3rcy joined, n0m3rcy left
mugwumpjism putter: use extract_chmLib to convert to HTML 02:30
putter mugwumpjism: thanks! i actually bought tapl at a sale a while back. but was curious. :) 02:34
mmm, za. nothing like standing in the cold directing traffic around a fire scene (next to the pizza place) to build appetite. ;) 02:35
just never know what you are going to hit when you head out the door. neat. :) 02:36
bbiab
robkinyon seen stevan 02:42
jabbot robkinyon: stevan was seen 1 hours 19 minutes 28 seconds ago
mugwumpjism utsl.gen.nz/talks/perl6.does-ML/start.html 02:44
02:45 vel joined 03:09 dduncan joined
putter mugwumpjism: would you like comments? 03:15
mugwump: ping? 03:19
03:21 nothingmuch joined
putter oh, nz. utc+12. ;) 03:30
wait, that's...
beppu sometimes, I think we perl coders are to hard on ourselves and our favorite language. I just installed www.datenklause.de/en/software/qgreylistrbl.html on my friend's server, and it's blocking spam like a champ.
I'm pretty fucking happy.
putter spamblocking++ 03:31
beppu It might not be as sexy as rubyonrails or $whatever_the_fuck_people_are_doing_with_python but it really fucking useful.
s/it/it is/
greylisting is the best! 03:32
thousands of spam emails were hitting the server every day, but now it's but a trickle. greylisting is so simple and yet so effective. 03:34
03:37 gaal joined
beppu perl: 1 / spam: 0 03:37
03:37 Daniel_Nee joined
Alias_ beppu: You always want back where you used to be 03:38
03:38 kanru joined
Alias_ beppu: Personally, I think web is just a nightmarish battleground of hype at the moment, and I'm moving on to more interesting fields 03:38
especially low-end web
beppu Alias_, that's cool. I think branching out and diversifying the %things_that_perl_can_do is a great thing. 03:40
Alias_ plus there's some areas really ripe for it
Toolchain, testing, VBA replacement
swiss army integration 03:41
beppu My best friend works for Sony R&D (and at game companies in the past) and he's snuck a lot of perl into places where it was largely unheard of.
Alias_ :) 03:42
beppu A lot of game places write their tools in C/C++, but my friend (a good perl coder) looked at all that they were doing and thought to himself: I could do this in a fraction of the time ...in Perl. 03:46
Alias_ yep 03:47
beppu Many of his C++-coding coworkers were skeptical at first, but he showed them what was up.
Alias_ If it doesn't need to ship, why make it compilable
beppu That's like a foreign concept to some people.
Of course, not to us, but...
Game dev has a very different culture. 03:48
Alias_ we have the other problem :)
beppu Changing topics, what did you mean by "low-end web" ?
03:48 vel joined
Alias_ Budgets between zero and $20k 03:49
hell, between zero and $100k
beppu That's all I've ever worked on. 03:51
Alias_ That's all a lot of people have
There's where Perl and PHP and Python and Ruby are all duking it out
beppu Who has the high-end ? 03:52
Java?
Alias_ Java and Perl
Possibly in that order
Things get tricky when you want to serve a billion pages a day 03:53
beppu For the low-end, easy installation seems to be the order of the day. That's the only technical strength PHP has going for it, and look where that got it. 03:54
Alias_ Elaine's Law
www.google.com/search?&q=Elaine's%20Law
beppu I agree. 03:57
Alias_++ 04:02
04:07 vel joined
Alias_ huge part of what made debian so successful, despite it having a slower release cycle 04:07
and, less deservedly so, PHP :) 04:08
meppl guten morgen 04:20
04:25 feng joined
mlh_ Alias_: who is the Elaine in Elaine's Law? (nice law btw) 04:28
Alias_ aevil? happyfunball/hfb? Jarkko's wife? 04:29
one of the sysadmins for the master CPAN infrastructure
04:55 Amnesiac joined
stevan mugwumpjism++ # nice lightning talk :) 05:02
05:07 gaal joined 05:24 brentdax joined
brentdax Is Parrot embedding broken? 05:25
I'm getting an error about PARROT_PATH not being set correctly when it certainly seems to be. 05:27
05:28 audreyt joined
putter mugwumpjism: a couple of thoughts, then off to sleep. 05:30
05:30 scook0 joined
putter you might mention the FP community forsaking fragmentation, coming together to do haskell. also then monads, and the hairshirt becoming more armor than mortification. 05:32
Alias_ FP?
05:32 Skud joined
brentdax FP=functional programming, I assume. 05:34
Skud morning.
Alias_ "morning"?
Skud feels like morning.
anyway, it's morning in... *ponders* ... the middle east?
seen ingy? 05:35
jabbot Skud: ingy was seen 1 days 17 hours 39 minutes 7 seconds ago
putter re pre pugs p6, I cant speak to the p6l side of things. but for parrot, I think you mischaracterize it. it was one of the worst project management jobs I've seen, and that's included some real dozies. if pugs had been managed similar, well, audreyt being who she is, it would have gone further. but no further than that.
let's see, strictly typed is not equiv to does(ML). TCL is strictly typed. 05:38
I'm not sure I share stevan's enthusiasm for the St object model. Then again, it's nicer than what most people are using now, so at least in constrast... 05:40
but all in all, I liked the slides. :) good luck with the talk. (soon? i didnt see a listing on the local pm boards) 05:41
Alias_: yes, what brentdax said. :) 05:42
ok. *whistle blows*. next pugs shift. smokes on the smokeserv, ext/ fails to be characterzied, a TASKS file for the "dont feel like debugging right now". have fun. 05:45
:)
Alias_: re FP, my understanding is, to an approximation at least, the fp community started down the path of everyone doing there own little variant language, but the manage to instead pull together to build a common foundation, haskell. if they hadnt', they might today be worse of than even say the scheme and prolog folks, 05:48
Alias_ ah
putter whose versions of the "same language" are often, even pathologically, incompatable. and whose standardization efforts have a thickness of politics resembling... high-school student government? something pointless but loud and interminable. ;) 05:50
SamB and then somebody came up with type classes... and someone else stole monads from denotational semantics...
Alias_ denotational is a word?
putter I'm being a bit sloppy and unfair, but that's at least the flavor of my impression of it. 05:51
SamB Alias_: I'm pretty sure
putter denotational is a most excellent word
Alias_ I'm quite used to not knowing all the world
words
I just putter with lego and build things 05:52
putter the wikipedia entry for denotational semantics looks pretty poor, let's see...
Alias_ Give me some blocks and I'll make you some bigger blocks, if there's a denotational emantic monad involved, just call it something I don't need a math degree for :) 05:53
SamB Alias_: okay, we will go with the accepted (humerous) wisdom and call it a "Warm Fuzzy Thing" instead of a Monad 05:54
Alias_ Can we call it an Action... or an Imperative Block?
SamB and, well, monads are a lot like legos ;-)
Alias_ From what I can tell, they are mostly for doing C< imperative { ...; ...; ...; } 05:55
putter i haven't looked far or closely (just working down the google hit list), but www.risc.uni-linz.ac.at/people/schr...ensem.html looks maybe plausible
oooh, I haven't seen that tried. let's name monads and arrows after legos. that captures their broadest use - structuring library apis so all the pieces can be combined and simply plugged into each other... 05:57
Alias_ For gui assembly? 05:58
ooo
Or embed code fragments on RFID id chips inside REAL legos
putter "An introduction to software library legos. One bump is "return", the second..." 06:00
:)
the mit media lab has these beads on strings,
which contain various behavior, hmm, memory fuzzy, like reflect every other pulse, and then... fuzzy... light pulses move up and down the necklace, under the influence of the particular set and pattern of beads. kids can then trade, modify, etc. 06:02
web.media.mit.edu/~mres/papers/chi-...manip.html a bit old, but 06:05
06:05 Alias_ left 06:06 kakos joined
putter cant find anything better. eh. anyway, the idea is just to go overboard towards complete familiar objects/terms. not sure how to do monad function wraping. oh, maybe your lego with a chip. but a function chip. so the lego blocks are a way to easily plug together a bunch of function chips. how's that? ;) 06:11
1am. good night all & 06:16
06:16 drbean joined 06:25 putter joined
putter actually, Alias_? math is just legos. described by a community with the pathological concept that golfing should be applied the entirety of their written material. like a perl module which said "INSTALLation instructions: solve this puzzle OHnf08yhfb and it will be self-evident how to install this package." 06:29
Most all math writing is the antithesis of Elaine's Law. Various theories re the social mechanisms which go into that. But the upshot is, 06:30
imagine lego described in doggel latin verse, which compared them to... bits of foreign fruit and extinct animal behavior. ;) precision, but with little effort at introduction or accessible-by-approximation to non-collegues. survey articles are soo much better. but also seem less common than in sciences/engineering. 06:33
hopefully this will change some day. a greater role for computers in math literature will help. but it's going to stay a pain for a while. 06:35
that said, it would be nice if p6 ends up with an algebraic type hierarchy. groups and rings and such. affine geometry for graphics. what else? 06:41
hey, neat. a LOGO turtle graphics approach to affine geometry: 72.14.203.104/search?q=cache:YSUwb1...k&cd=1 06:42
& 06:43
07:31 beppu joined
audreyt resumes releng. 07:32
gaal hello!
audreyt yo! 07:33
I've been reading up on LINQ 07:34
<- filled with weird and nice ideas
but releng first, definitely :)
how's life for you, gaal?
gaal Not bad! I just worked with Ingy a bit on Perldoc 07:35
but so far my contribution has been minor
my morning looked like this: I woke up thinkging I should go finish rereading the awkward squad paper 07:36
but I coulnd't find it on my disk, so I started rereading the sexy types summary 07:37
this time around I actually didn't understand it
as opposed to stare blankly at it, which was what happened the first time :)
then ingy popped by and we discussed p5 export stuff. 07:38
then I got coffee :-)
anyway, what's on the plate in pugs now? :)
audreyt releng, mostly, I think 07:39
also solving the state.t mystery 07:40
(fails on make smoke but works with running by hand)
gaal different Test.pm ?
audreyt probably not 07:41
gaal svn ups
audreyt also I have a half-finished patch on pairs.t
or rather on pairs parsing
I'd like to get it correct this time
gaal I don't understand that stuff at all :-( 07:42
audreyt f(a=>1); # named
f((a=>1)); # positional
f("a"=>1); # named
07:42 iblechbot joined
audreyt f(("a")=>1); # named -- but currently parsed as pisitional 07:42
the reason is that it was looking at a '('
and force positional if it's that way
but in fact if the ( just surrounds the key part 07:43
gaal I remember, changelogging, I saw scook0's "demagicalize pairs" commit
audreyt then it is actualy named
yes
gaal this syntax is tricky.
audreyt indeed 07:44
gaal what magic was there that needed to be de-'ed?
audreyt anyway, I'll probably finish it up in ~30min
oh
it was previously specced
during toronto hackathon
that a Pair object causes named binding
much like a Junction does
however, it proved to be extremely bad
so the decision is reverted to a parser-level thing
gaal cause then how do you pass a Pair?
audreyt now named or not is a purely syntax patter 07:45
right, you can't pass a pair
without the function declared as Item|Pair
or some other nonsense
gaal I see
audreyt also, somehow "make ghci" is broken again
complaining missing syck symbols
havn't had time to look at it 07:46
gaal I'm afraid I can't help with that; it never worked for me on colinux :(
audreyt :/ 07:47
there's some triage left near the top half though
t/builtins/arrays/splice.t is failing last one
gaal what do you see with state.t? running by hand I see some failures and some okay
audreyt but you don't see "failed" in the last summary line
t/builtins/math/NaN.t need some work to sync with the p6l too 07:48
gaal that sounds like a wrong plan
audreyt *nod*
gaal oh, sorry, the fails I see on the cmdline are all TODOs.
could it be that the (#1) in the test description is confusing Test.pm? :-) checking. 07:49
yeah. 07:50
fixing.
07:50 Cryptic_K joined
audreyt gaal: check /msg when you got a chance :) 07:51
svnbot6 r8788 | gaal++ | * "#" in the test description confuses Test.pm. 07:53
r8788 | gaal++ | Remove it now, and fix Test.pm later.
gaal audreyt: replied
audreyt not seeing it 07:54
nickserv?
gaal (error notices in different window)-- 07:55
audreyt and I fixed the parens/pairs parsing as well 07:59
gaal oh, there was the pil test issue, it wasn't resolved was it? let me find an old nopaste about it 08:03
gaal realizes his last commit message was bogus; it's the TAP harness that's confused, not Test.pm. 08:04
this is probably a bug in TAP though :-)
audreyt I think the correct way is to escape the # 08:05
see p5
gaal $desc ~~ s:g/#/\\#/ in the harness, or in the test? 08:06
audreyt in Test.pm I think
gaal er
audreyt s:P5:g for now
gaal I meant in Test or in the test :)
sure
ok
audreyt: please see sial.org/pbot/15479 08:16
svnbot6 r8789 | gaal++ | * t/var/state.t - revert r8788
r8789 | gaal++ | * ext/Test/lib/Test.pm - fix the problem where it ought to be fixed.
gaal context: it was meant as a patch to Test/lib/Test/Pil.pm
to fix prove t/pil when pil isn't found
audreyt nod
gaal when this is run, there's an error about &pil_is_eq not found (at runtime) 08:17
so I think there may be a problem in dispatch or in symbol construction 08:18
audreyt at the &pil_is_eq := {"";} line ?
gaal no, in the context of the .t file 08:19
(btw {"";} was just me being paranoid that {;} was still not being parsed as a closure) 08:21
audreyt ok, I'm still in the named thing 08:23
a sec
gaal sure. 08:24
i'll look at the array thing
regarding t/builtins/arrays/splice.t - it "fails correctly" here, that is, the last test fails and prove reports it as such. also m19s28.vlinux.de/iblech/stuff/pugs-...09cf9.html paints it red. 08:29
svnbot6 r8790 | audreyt++ | * Correctly implement parens as the disambiguator for named args.
08:33 wilx joined
svnbot6 r8791 | audreyt++ | * did I say "correct"? I said "approximately correct". 08:45
audreyt gaal: er, I mean, to triage it 08:46
i.e. fix the bug or mark it as todo
gaal fixing the bug I think involves getting the Prim spec right (remember our discusssion about List::foo?) 08:47
audreyt nod
gaal ergo, I'll :todo it :-)
audreyt so just todo it for now
gaal on it. (dies_ok does not take :todo; I think I'll fix that) 08:48
sub dies_ok (Code &code, Str $desc?, :$todo, :$depends) 08:53
is :$todo the syntax for a pair?
dduncan yes
gaal the code looks like it should already work, but doesn't. 08:54
hi dduncan
dduncan hi
but not for long
I also like to be explicit about about requirements, so I have an explicit ! or ? after ALL my parameter declarations 08:55
audreyt ?eval sub f (:$a!) { $a } f("this is not named")
08:56 evalbot_8782 is now known as evalbot_8791
evalbot_8791 \"this is not named" 08:56
dduncan that looks borken
gaal don't signatures have an automatic positional form if they can get away with it? it's a fewature 08:57
*feature
dduncan leaving out the : makes it positional
putting it in makes it named
the presence or absense of : is the only difference with an otherwise same parameter 08:58
audreyt yeah, that's broken.
gaal why *demand* your callers use named args?
audreyt fixed.
committing
gaal lol
so good to have you back, audrey :)
dduncan gaal, for forwards compatibility
gaal ahhhh yes
good reason
dduncan: thank you; your answer just helped me understand something in haskell 08:59
svnbot6 r8792 | audreyt++ | * required named parameters should be checked along 09:01
r8792 | audreyt++ | with required positional parameters.
audreyt gaal: oh?
?eval sub f (:$a!) { $a } f("this is not named")
gaal audreyt: I was wondering when to use trivial type aliases
audreyt answer is "always"
gaal if you have e.g.
evalbot_8791 \"this is not named"
audreyt unless you have newtype, which is even better
audreyt waits for evalbot to pick up r8792
dduncan how often does evalbot check for a new version?
gaal data YamlNode = MkYamlNode { tag :: YamlTag, anchor :: YamlAnchor }
though they are really both (Maybe String)
now other people's code won't compile if I change the order of these two elements, and they were using positional access. 09:02
which is, of course, good
audreyt dduncan: 5min
gaal because otherwise I'd have introduiced a bug
audreyt ?eval sub f (:$a!) { $a } f("this is not named")
evalbot_8791 \"this is not named"
audreyt gaal: aha, which is why you should've used a newtype :) 09:03
because if you use "type" aliasing, their code actually compiles.
sad, but true
gaal I was wondering about this, because I was reading the example in TaPL about how aliasing say Dollar and Euro prevents you form doing silly things like dollaramount+euroamount
audreyt nod 09:04
type alias in hs is mostly about readability and sensible errmsgs
to achieve safety, use newtype + deriving
which is guaranteed to not affect runtime performance either
gaal great, several clicks in one sitting :) 09:05
but that also reminds me, why do we sometimes get errors about EvalT blah blorgh foo instead of Eval?
I mean, when I make a programmer error and forget to "return Val" and instead just "Val" 09:06
the inferencer error there could be more friendly
09:06 Alias_ joined
Alias_ seen audreyt? 09:06
jabbot Alias_: audreyt was seen 2 minutes 4 seconds ago
audreyt Alias_: yes?
Alias_ audreyt: I hit a bug with install_share, and I need some input on the solution 09:07
audreyt gaal: indeed
obra Alias_: what are you running into?
Alias_ audreyt: When the libs install, an auto-dir ends up in BOTH of blib/lib and blib/arch
obra (As a user, I'm interested ;)
Alias_ But the actual share dir is installed into blib/arch 09:08
In all @INCs I know of, the lib comes before the arch
gaal audreyt: so why doesn't it? they are aliases, IIRC
Alias_ So if you do a dist_dir('Dist-Name') you see the (empty) blib/lib
But if you do a dist_file('Dist-Name', 'filename.txt') you get the path in blib/arch 09:09
09:09 Bit-Man joined
Alias_ Meaning that dist_dir('Dist-Name', 'file') no catfile( dist_dir('Dist-Name'), 'file' ), an extremely obvious assumption 09:09
s/no/ne
gaal okay, I looked into newtype, the problem is that for essentially trivial types it's a little clunky 09:10
audreyt gaal: I think it's just the desugarer didn't consult the type alias table
Alias_ So I'm wondering if either the files can be put into blib/lib instead (unsure of the implications within the lib naming scheme) or something else can be done
audreyt gaal: I can hack that into GHC, or you could too, but that'd be after this release :) 09:11
gaal :)
that "you could too" remark fills me with terror and anticipation
audreyt gaal: I can show you around the GHC tree when I'm in .il 09:12
it's really not that bad
especially compared to perl5 internals.
gaal heh.
audreyt Alias_: I think putting it to blib/lib/'s auto makes sense. 09:13
gaal I have to figure out how to set up a build environment for ghc though, it's a little tricky on debian unfortunately.
audreyt Alias_: the shared files shouldn't be arch dependent.
Alias_ audreyt: I would think not in 99% of cases
audreyt good then, move them off arch 09:14
Alias_ erm, how
audreyt carefully
Alias_ The instructions are in raw Makefile fragments
audreyt done
committing
Alias_ thanks, I'll package and release
(Changes, etc)
audreyt test first :) 09:15
Alias_ M:I has tests now?
audreyt no, I mean, test for interop with File::ShareDir first ;)
r91.
Alias_ My test case is that File::HomeConfig (new module) works if it uses lib, fails if it uses arch
audreyt ok
it's committed now, enjoy 09:16
Alias_ But I'll add an extra check to an appropriate place in File::ShareDir too
thanks
audreyt np :)
Alias++ # driving practical innovations
Alias_ obra: Should mean zero functional change except in the case where you want to recursively copy the contents of dist_dir('Dist-Name') to somewhere else
obra: In fact, dist_file will continue to work for both old and new locations
dist_dir will return the same too 09:17
It's just that the VALUE of dist_file might be either of the two for a while, wherever it finds the file
obra: Oh, also, is there any way in rt.cpan.org on a bug's page(s) to have a link to the queue it's in, so once a bug is resolve I can go back to the queue? 09:18
(without searching from scratch again)
svnbot6 r8793 | audreyt++ | * pairs.t now has two fails only (was 10+); fixing them both now.
obra Alias_: please use the address on the site 09:19
I no longer deal directly with rt.cpan.org. It's all delegated to minions 09:20
Alias_ There's a queue for itself?
obra Of course
on a differnet RT instance
Alias_ ooohhh...
ok, didn't see any obvious link
obra ...that you can't see ;)
Alias_ will look again
cpan-questions@? 09:21
obra upper right hand corner of every page 09:22
Alias_ So that's a yes :)
(and "except the front one")
obra on the front one, it's in the prose
Alias_ yep 09:23
So yes :)
svnbot6 r8794 | audreyt++ | * pairs.t: add extra tests for the currently broken behaviour: 09:27
r8794 | audreyt++ | $f(((:a))) # positional (correct)
r8794 | audreyt++ | $f((:a)) # named (incorrect)
r8794 | audreyt++ | so we can revisit this later.
audreyt looking at pil_is_eq now. 09:28
Alias_ (pause keys)++ # Makes life so much less fiddly
audreyt for a split second I thought you were talking about PAUSE2003.pub. 09:29
obra (pause keys)++ # Automatically sync my cpan mirror.
Alias_ oh god no :)
obra yes, actually
Alias_ well, yes too, but not the kind of fiddly I mean 09:30
audreyt nominates the paragraph above for the "context-free dialogues of the year" award 09:32
svnbot6 r8795 | gaal++ | * Data.Yaml.Syck - remove comment about pesky warning needing to be dealt
r8795 | gaal++ | with, bacause audretyt++ already did.
audreyt hm, I need to go picking up my eyeglasses. 09:33
which should make Reality even more realistic and fun.
<- never wore eyeglasses before
Alias_ audreyt: The field of vision restriction gets annoying after a while
But you can read!
audreyt Alias_: my eyeglasses has style="border: none" 09:34
Alias_ doesn't matter
optical edge cases at the boundaries create border: solid 1px #99999 09:35
audreyt true
though it's more like ridged in my case
audreyt sighs at the general geekiness
anyway... be back in ~20mins
&
Alias_ yeah, and still accurate 09:36
which is more annoying
svnbot6 r8796 | gaal++ | t/builtins/arrays/splice.t - mark test TODO for release 09:38
09:39 kane_ joined
gaal wonders if there's a good way to prevent bugs like f(a), b; when you mean f(a, b); 09:39
apart from getting into the habit of not using parens 09:40
I just spent 15 minutes debugging dies_ok because I didn't notice the parens in the original call :/ 09:41
09:48 pfarmer joined
dduncan good night folks 09:49
I look forward to that 6.2.11 ... 09:50
10:01 dada joined 10:02 drbean joined 10:09 chris2 joined
Alias_ crap... 10:15
enoaudrey
10:19 G2 joined
gaal is t/rules/from_perl6_rules/charset.t working for you guys? it's causing ghc errors for me here. 10:24
10:24 Alias_ left 10:33 nnunley joined 10:37 elmex joined 10:43 Alias_ joined
Alias_ Module::Install 0.55 uploaded 10:43
That is all. Carry on
10:43 Alias_ left
gaal what's our current parrot version requirement? 10:54
I'm suspecting the problem is related to that. 10:55
11:28 putter joined
putter 0.4.1 11:28
theorbtwo gaal: In theory, the answer is always "the latest released version", I think.
putter answers from his sleep
gaal well, parrot looks totally borked now :( 11:31
I tried both embparrot and external
embparrot segfaults
audreyt back
gaal external claims no rule support
audreyt hmm? weird, it worked for me 11:32
gaal parrot HEAD
audreyt HEAD I havn't checked.
check 0.4.1 instead?
gaal do you have the r for that?
putter re charset, a couple of Cannot parse PGE: on things which say ok. odd. but otherwise looks like it did the right thing
audreyt gaal: download it off cpan
gaal k
audreyt so I got my eyeglasses. then I looked into a mirror. that makes me depressed :/ 11:33
putter ?
audreyt my skin sucks and I was blissfully unaware ;)
putter lol
sorry ;)
audreyt nah, that gives me another project to work on ;) 11:34
putter :)
doh. no odd. just failure test which failed for not quite the reason anticipated. 11:36
putter goes back to sleep 11:37
&
svnbot6 r8797 | audreyt++ | * hexadecimal.t: add the missing =cut, so all its tests pass now. 11:38
audreyt sleep well :)
11:41 raar joined
gaal well. external parrot just doesn't work for me :( 11:45
audreyt hm. win32?
gaal colinux 11:46
let me try again with embedded / parrot release
11:50 lypanov joined
gaal nope, segfaults. 11:50
trying to get a stacktrace.
audreyt both embed and external workshere. 11:51
svnbot6 r8798 | audreyt++ | * NaN.t: In this release, 0**0 and Inf**0 are both 1, not NaN,
r8798 | audreyt++ | until we find a better solution.
gaal looks related to parrot_split_path_ext 11:53
leo: ping
nnunley ?eval NaN ** 0 11:55
11:55 evalbot_8791 is now known as evalbot_8798
evalbot_8798 1 11:55
nnunley That doesn't seem right. 11:56
11:56 ruz joined, drbean joined
audreyt that seems wrong. fixing 11:58
ooh, new spj paper, again directly related to p6. 12:07
s/related/applicable/
# research.microsoft.com/%7Esimonpj/p.../index.htm 12:08
svnbot6 r8799 | audreyt++ | * Test::Pil - skipping PIL tests when ./pil is not present. 12:15
wolverian wow, nice. 12:17
(the paper)
he keeps stealing our ideas ;)
gaal audreyt: r8799 works, I guess, though skip_all ; exit looks cheeky... did you see at all why my version wasn't working? 12:19
wolverian audreyt, what do they mean with "non-strict" when describing Haskell? 12:20
12:20 Blicero joined
gaal summay of that paper: 10 Curry-Howard 20 GOTO Brain explosion 12:29
nnunley wolverian: I believe it means strict evaluation order. 12:33
audreyt gaal: no, not at all :) 12:34
wolverian nnunley, ah, thanks. I don't know what that means, exactly, either, but I'm closer now :)
audreyt wolverian: it means that Hs programs can be evaluated in any order :)
i.e. in [f x, g y, h z] 12:35
wolverian ahh. thanks!
audreyt it doesn't matter which one of the three are evaluated first
which is not the case in other languages, e.g. if "f" contains a $::global++
and "g" contains a $::global--
svnbot6 r8800 | audreyt++ | * nnunley caught that 0**NaN should be NaN, not 0. 12:36
audreyt Juerd: got the T-shirt, it's lovely, thanks! :) 12:40
12:45 pfarmer left 12:47 joepurl joined
stevan putter: Just FYI - I have no love for the Smalltalk 72/78 object models, mugwumpjism was a little off there,.. the mm is based on Smalltalk 80 and CLOS :) 13:19
morning all
stevan runs off to do errands and get coffee
audreyt heya stevan 13:20
svnbot6 r8801 | audreyt++ | * perl2.t: test cleanup, proper unTODO and TODOing, etc.
audreyt heard anything from tpf yet re yapc tokyo?
13:21 kolibrie joined
Juerd audreyt: Yay :) 13:24
gaal I have to go now, back in... 20 hrs? unfortunately even after make clean embparrot still doesn't work here. 13:31
audreyt gaal: we'll figure it out later. I'll also try to get colinux running
?eval 1 ~~ /1/
13:31 evalbot_8798 is now known as evalbot_8801
evalbot_8801 Match.new( ok => bool::true, from => 0, to => 1, str => "1", sub_pos => (), sub_named => {} ) 13:31
lypanov audreyt: colinux is fun :) 13:35
audreyt :)
audreyt is resmoking
the tree, not chemicals
lypanov used it for just on 6 months when he got sick of the lack of wireless support for his laptop in linux
hehe
can you resmoke chemicals? :P
audreyt sure could :) 13:36
lypanov wonders when vmware will announce something for the macbook pro
13:37 stevan_ joined
stevan_ audreyt: re: tpf on yapc tokyo - no, not yet they review on Feb 1 13:38
stevan_ crosses his fingers and prays to the gods of Perl :P 13:39
audreyt :D
I'm about to sleep (as usual)... been a very long day with various pre-lunar-new-year errands
tomorrow is the new year's eve :)
wolverian I wonder why people dislike POE.. any clues?
ohh, have fun :)
stevan_ yummy mooncakes :)
audreyt mooncakes are midautumn, which is many months from now :) 13:40
stevan_ ??
audreyt s/are/are for/
stevan_ whats the new years treat then?
gaal should look up this lunar new year thing. It sounds like something that ought to happen every month? 13:41
audreyt rice cakes
gaal also, I must go now. see you later :-) &
stevan_ hmmm,..they dont sound as tasty
audreyt indeed they are not
stevan_ notes that white anglo-saxon americans enjoy mooncakes this time of year, probably because the chinese are not eating them so there are leftovers :) 13:43
audreyt lol 13:44
stevan_ ok, time to take out the dogs and run some errands 13:46
adios all &
lypanov www-users.york.ac.uk/~fj102/photo/b...0china.htm 13:50
13:59 kanru joined 14:01 rantanplan_ joined 14:05 Qiang joined 14:12 nothingmuch joined 14:14 putter joined
putter hey. looks like shift change coming up? ;) 14:14
14:15 iblechbot joined
audreyt ;) 14:16
hey putter :)
putter re NaN ** 0, I don't remember if were still doing 754-1985 or are speculating, but if the latter, than everything ** 0 == 1, including NaN. but whatever.
audreyt is, indeed, going to sleep now -- journaling is deferred to tomorrow
putter hi audreyt
audreyt putter: I am speculating here :)
putter anything in particular you would like to see happen as you sleep?
audreyt 754-1985 thinks NaN**0==NaN, I think 14:17
putter: hm, more follow-up changelogging? :)
putter yyyes, putter pauses, I think so.
changelogging. ok, anything else?
audreyt also, TODOing t/pugsbugs/* that are not regressions (i.e. those that fails at the end, or are checked in after last release) 14:18
I think it's safe to focus only on fixing regressions now
so newly reported pugsbugs, unless they are misfeatures or really trivial, should simplky be TODOed.
putter ok. :todo<> non-regression pugsbugs. makes sense. what else?
audreyt I think that's all :) 14:19
have fun :)
(also look at regressions if you have cycles)
putter ok. thanks. great to have you back. sleep well. :)
ok, and linq, another time. 14:20
audreyt also the irreconsilable gaps between eager and lazy evaluation :)
and between typechecking and "where" style dynamic assertions 14:21
putter you mean the not^2 ML paper? yes. i just skimmed it a while back, but looked fun.
audreyt ("Haskell is not not ML" and "Hybrid Type Checking", both via LtU, on my reading list now)
putter LtU++
audreyt the hybrid typecheck thing is more directly applicable to PIL* though. 14:22
we'd really need that.
putter ooo, haven't seem HTC. add to list. bottom of list implodes. office anti-black-hole safeties kick in. putter wonders what paper he just lost. oh well. 14:23
err, s/list/pile/
good night audreyt & 14:25
audreyt g'nite :)
putter stevan: want to read HTC together?
& :)
stevan HTC? 14:27
stevan backlogs enough to decifer HTC 14:29
putter sorry, let's see...
stevan putter: sure... my errands have been delayed for about an hour
putter there it is, www.soe.ucsc.edu/~cormac/papers/popl06-hybrid.ps 14:32
stevan cool
ohhh,.. this one looks good too www.lfcs.inf.ed.ac.uk/reports/98/EC...CS-98-389/
putter stevan: have you seen a critque of smalltalk-80? i vaguely remember seeing a nice one years ago, but my googling isnt turning it up... 14:33
stevan putter: I know the ST 7* metamodel designs were tossed out because of the "turtles all the way down" issue
but '80 has been around for 26 years without much change 14:34
the p6mm was closer to '80 when it had eigenclasses,.. which we are not using anymore since Larry wants class methods and instance methods to be the same thing
putter yeah, but I thought I once saw a "in retrospect, wish we had done x y z differently paper". maybe. 14:35
stevan I am thinking a mm 3.0 is going to be needed after we have the new PIL runcore in place anyway
putter: hmm... that would be good to read
putter putter pops briefly over to #squeak . then list list audreyt's todos for this shift. then, whatever. 14:36
stevan has found that few people agree on the best approach to metamodels
stevan follows putter to #sqeak
14:37 Grrrr_ joined
stevan hmmm, very quite in there 14:37
I think I would like to see Roles have a similar model to the ML module system 14:38
putter yes. ah well. todo:
stevan :)
putter more follow-up changelogging; TODOing t/pugsbugs/* that are not regressions (i.e. those that fails at the end, or are checked in after last release); look at regressions. Also "Hybrid Type Checking"; "Types For Modules". And is the a st80 postmortem paper. 14:40
stevan: anything in particular you'd like to do pre-errands? 14:41
stevan putter: nope, just looking for the smalltalk stuff 14:42
putter ok, non-regressions seems like a dirt-brushing banana. checking...
stevan users.ipa.net/~dwighth/smalltalk/bl...p_toc.html 14:43
its not what you are looking for.. but an interesting bit of stuff anyway
14:43 chris2 joined
putter anyone know what the rNNNN of the last release was? 14:46
err, or what the last release N.N.N was?
lypanov giggles 14:47
i have a new tshirt idea:
$
=>
FIXED
stevan putter: lambda-the-ultimate.org/node/276
stevan is following some of these links
putter stevan: re bluebook, neat. you've seen Slate, yes?
stevan Slate? 14:48
putter ST derivative? putter looks for link... may have name wrong....
slate.tunes.org/
putter bookmarks bluebook chapters, doesn't quite get lypanov's t-shirt - puzzled... 14:49
chris2 lypanov: ? 14:50
putter stevan: oh, awesome. classics-online++. now if only we could get the ACM digitial library to open up, open source would actually have access to the literature... 14:51
stevan putter: slate.tunes.org/documentation.html 14:52
lypanov putter / chris2: $$$ == dollars, FIXED as in bugzilla resolution
putter still trying to figure out rNNNN of last release... oh I know - grep captured #perl6 log... ;) 14:53
chris2 lypanov: i'm too clever for this world
lypanov what'd ya do now?
chris2 i thought of $ as being apply in haskell and fix-points :P
lypanov :P
umm 14:54
www.smallscript.org/
wish that was oss and worked with mono
14:57 avar joined
putter [01:13] <svnbot6> r7520 | autrijus++ | * This be Pugs, version 6.2.10. 14:58
stevan: re slate docs, anything in particular? it's a neat system. 14:59
lypanov: so create a smallscript front end to p6? 15:00
15:02 Limbic_Region joined
svnbot6 r8802 | audreyt++ | * slurpy_param.t: unTODO two now-passing tests thanks to the 15:03
r8802 | audreyt++ | named arg refactoring.
lypanov putter: mainly interesting from the perspective of for the speed and integration
coding in her sleep again :)
putter :) 15:04
stevan fscript is kind of cool too ,.. like Smalltalk meets APL IIRC www.fscript.org/
lypanov ooo
stevan: thx for the link. shall test this out tonight
stevan :)
stevan thinks #perl6 should start authoring papers 15:05
lypanov www.fscript.org/news.htm
yum
putter caution - I just now svn up, and make broke on 8801. ext/Locale-KeyedText makefile issue.
Juerd stevan: Just dump IRC logs ;)
stevan: Publish those.
stevan Juerd: but its not a *real* paper unless its formatted with LaTeX :P 15:06
Juerd stevan: \begin{verbatim} 15:07
svnbot6 r8803 | audreyt++ | * operators/binding/scalars.t: chase the change that restricts 15:08
r8803 | audreyt++ | $CALLER::a to only see $a declared as "env" in caller's scope.
putter ooo, pyinjector looks neat. (top of news.htm)
stevan awards Juerd a PhD from the University of Perl 6
Juerd Woohoo!
Juerd orders new business cards immediately. 15:09
putter :)
Juerd wraps a verbatim block around this part of the IRC log, prints it, and frames it. 15:10
putter lol
stevan :D
svnbot6 r8804 | audreyt++ | * inherit_from_role.t: todo<feature> all 4 subtests as those
r8804 | audreyt++ | depends on working Roles in the metamodel.
Juerd What use is a PhD without a certificate? :)
binary42 You will need to generate a References section though. Quick get all the host names! 15:11
stevan yes, that all important "peice of paper" :)
Juerd binary42: The references are mentioned in angle brackets just after each time stamp.
binary42 but how will I go about looking up more information about that source without something more that a name? 15:12
Juerd Nicknames are more stable and identifying than hostnames :)
(culturally)
binary42: This paper is dynamic. You can join and participate at any time.
binary42 waves to the future LtU community.
svnbot6 r8805 | audreyt++ | * data_types/num.t: chase S02's more restricted definition 15:13
r8805 | audreyt++ | of underscores between numeral digits; now multiple underscores
r8805 | audreyt++ | are not ok, and neither are beginning/trailing underscores.
Juerd binary42: So if you think some information is lacking, you can ask the authors and it'll automatically update! :)
lypanov putter: very (pyinjector) 15:19
rafl Juerd: Do you have some copies of the s:g/5/6/; shirts? If yes, can you bring one of them to fosdem for me? :-) 15:21
Juerd rafl: No, none 15:23
I haven't even seen a cafepress one yet 15:24
rafl :-(
Juerd As far as I know, Audrey is the only one on this planet who already has them :)
rafl I think it would be a good idea to bring some of them to GPW.
Juerd I agree, but I don't like the financial risk, and the lots of work to ship shirts, and that's why I chose cafepress. 15:25
lypanov rafl: hehe. neat tshirt idea
Juerd I can have them mass produced much cheaper than Cafepress.
But I'd first need some assurance that they'd be sold, 'cause I don't want more than two of them in my own wardrobe :) 15:26
(Approx EUR 13 per shirt, high quality shirt and screenprinting) 15:27
audreyt EUR 6~7 here :) 15:32
Juerd Wowie
audreyt (for the perl6 timeline short you have)
s/short/shirt/
Juerd audreyt: Are black shirts an option?
(That fits better with the cryptic nature of s:g/5/6/;, in my opinion)
audreyt I can ask, probably yes 15:33
Juerd Neat. At that rate, I don't mind getting a few
audreyt maybe I should print a bunch and bring it to osdc/gpw
Juerd That'd be great, I think
See also my p6l post; consider replying if you want to do this
Have to run; afk 15:34
svnbot6 r8806 | audreyt++ | * change the "sub *foo" in Prelude.pm into "sub foo is builtin", 15:35
r8806 | audreyt++ | and rework Scalar::as etc into multis.
r8806 | audreyt++ | That makes sprintf_and_as.t pass.
r8807 | putter++ | :todo<bug>ed all t/pugsbugs tests which are non-regression failures. That is, failing tests added since r7520 aka 6.2.10. 15:40
r8807 | putter++ | As a result, t/pugsbugs is now entirely clear of failures.
audreyt heh. 15:41
15:41 hexmode joined
audreyt in that case, maybe rename the all-pass-without-todo parts of t/pugsbugs/*.t to the correctly-named subdirs? 15:42
putter audreyt: not that I don't enjoy and appreciate your presence, but... why aren't you asleep? ;)
audreyt that's a good question
I don't have good reasons 15:43
audreyt vanishes in a puff of logic
putter dusts hands
15:44 cognominal joined 15:53 Eimi joined 15:55 vel joined, DaGo joined 15:56 Eimi joined 15:58 vel joined
putter So, now we have a shift todo list of: more follow-up changelogging; move all-green files out of t/pugsbugs/ (but not rules_with_embedded_parrot.t I think); Reading suggestions: "Hybrid Type Checking"(audreyt), "Types For Modules"(stevan). 15:58
Anyone want to tackle changelogging? 15:59
from ./ChangeLog: ** to r8657 - needs editing ** XXX big things not mentioned here: - objectspace / metamodel - PILN - sixpan / JIB - PIL2JS 16:01
16:04 vel joined
putter so three distinct tasks/task-sets: 1- read through svn log -r8658:HEAD , adding entries to ChangeLog. You don't have to do the whole thing, just raise the ** to r8NNN **; 2- edit/clean what's already there; 3- write lines for some of the "big things not mentioned". Again, with no need to do all of them, just make progress. 16:04
Volunteers encouraged. 16:05
I'll start moving files out of pugsbugs, unless someone else wants that.
None of these should repair or break anything, so if someone wanted to run a smoke on HEAD, and smokeserve it, now would seem a good time. 16:06
helloooOOOOooo. there is someone else besides me around here isn't there? ;) 16:09
masak :) 16:11
putter Juerd: how about ChangeLog'ing svn log -r8658:8675? seventeen measly log entries, most of which can no doubt be ignored.
hi masak :) got a few spare cycles. five minute chunks fine. audreyt left a list of a couple of things she'd like done this "shift", while she sleeps. game? 16:12
16:13 vel joined
putter ie, t/pugsbugs/parsing_my.t is no longer failing. so how about check that the tests appear in the usual my test files, or copy them over, and svn rm parsing_my.t 16:14
we can all use m19s28.vlinux.de/iblech/stuff/pugs-...c13c1.html to see which pugsbugs tests are all light-green. 16:16
16:18 elmex joined 16:20 vel joined, iblechbot joined
putter elmex, vel: got five minutes and commit bits? like to knock off one of the pre-release todo items? 16:22
avar cool s:g/5/5/ t-shirt 16:24
elmex ? 16:26
putter: ??
PerlJam avar: why would anyone want a no-op shirt like that? ;) 16:27
avar ?eval $x; 16:28
16:28 evalbot_8801 is now known as evalbot_8807
evalbot_8807 Error: Undeclared variable: "$x" 16:28
avar ?eval my $x = "aaa"; $x ~~ s:g/a/b/; $x
evalbot_8807 \"bbb"
avar ?eval my $x = "aaa"; $x ~~ s:xg/a/b/; $x
evalbot_8807 \"baa"
avar ?eval my $x = "aaa"; $x ~~ s:gx/a/b/; $x 16:29
evalbot_8807 \"baa"
avar odd
what does x do now?
putter elmex: do you have permissions to edit pugs files? would you like to spend 5 min helping toward the release?
:)
elmex hm, i don't know, i guess i checked in a test around 1 year ago 16:30
putter if the latter but not the former, can fix the former now...
great!
elmex but i've no idea where my checkout is 16:31
tewk putter: I can run a smoke if you need it. 16:32
putter one task is to "look at a test in pugsbugs/ which is now all happy, say t/pugsbugs/postincrement_in_subscripts.t, make sure the test inside of it is present in whatever file usually tests such things (or copy it over), and svn rm the pugsbugs file. pugsbugs is full of files for things which are no longer bugs...
elmex: just svn co svn.openfoundry.org/pugs/ openfoundry is being fast today, so no problem! 16:34
elmex it's just that i don't really have time right now
putter tewk: that would be great! and did you see the todo list back a 1/2 hour or so? 16:35
tewk I'll go look
putter re ChangeLog, gaal says in the r8658 log 16:36
ChangeLog towards release - got to r8131; doesn't include objectspace stuff
so there's a ways to go. could do svn log -r8131:8150 ! :) 16:37
tewk++
elmex putter++ 16:38
elmex--
putter colabti.de/irclogger/irclogger_log/...=701#l1034 16:40
elmex: why elmex - - ? 16:41
putter puzzled
tewk elmex is a little hard on on him/herself 09:23 < elmex> it's just that i don't really have time right now, hence - - 16:45
elmex because i'm coding perl5 right now
svnbot6 r8808 | putter++ | Added some blank lines between the current release and the finished entry for the last one. 16:47
putter ach, that's fine. pugs is a long haul thing. plenty of ongoing opportunities to contribute. new TASKS file. :) I'm just trying to stir up activity. 16:48
tewk putter++ stirring up activity 16:49
elmex perl6 seems like a long haul thing too 16:51
16:55 vell joined
putter tewk: thanks. :) when audreyt has "things she would like to see done" which we can do, it seems very worth while getting them done, so she can tackle the harder bits. 16:58
svnbot6 r8809 | putter++ | Deleted the now working t/pugsbugs/parsing_lexical_code_vars.t, moving its test to t/var/my.t.
putter elmex: re p6, yeah, but I suspect we are seeing the slowest period right about now. with the piln switch, I suspect there will be a cascade of "oh, now that works so we can do that, and now that works so we can do that, and now...". with associated increase in attention and personnel. 17:00
17:01 ycheng joined
putter looks at t/pugsbugs/parsing_if.t, maybe some ChangeLog next if someone doesnt beat me to it. 17:01
tewk putter: where should pugsbugs/postincrement_in_subscripts.t go ? 17:13
putter tewk: looking...
svnbot6 r8810 | putter++ | Deleted the now working t/pugsbugs/parsing_if.t . One test was added to t/statements/if.t. A second was already there. The third was an always working test, a minor variation, which was present only for contrast - so it was discarded.
putter so candidates are operators, arrays, and parsing... 17:14
theorbtwo Put it in array-derefence, with a comment in postinc, or vice-versa? 17:15
putter reading the test more closely, it looks like it was an array issue...
tewk there is t/builtins/arrays 17:17
putter oh, I forgot step one. find . |grep -v svn|xargs grep ']++' ... there are already similar tests in t/operators/inc.t so that likely the best place. 17:18
err, find t/ ... 17:19
theorbtwo: is there a postinc somewhere? checking...
theorbtwo t/operators/inc.t sounds about right for that. 17:20
theorbtwo was just guessing.
putter ah. :) maybe just inc.t.
17:20 eric256 joined
putter wonders how much cross referencing is going on... 17:20
17:20 eric256 left
putter so a little bit, but not a lot. mostly "moved this" comments. 17:23
putter wonders if the current test organization/tech can survive another 2x or 3x of size and still be manageable. maybe with Test::Base?? 17:24
theorbtwo Putter: I don't think it's manageable now. 17:25
putter tewk++ theorbtwo++
;)
yeah
have there being thoughts on how to proceed? Test::Base? a test generator? 17:26
something else?
putter tries to think of the largest language validation suite he has ever seen... (reasoning that p6's will have to be larger;)
theorbtwo: in what way(s) is it not manageable? 17:27
theorbtwo "I found a problem, is there a test for it?" "I have no idea." 17:28
"Let me do a test run..." <three hours later> ...
putter todo item: check out scala and slate and see what they are doing test wise. also pypy. mono?
:) 17:29
theorbtwo The kennel idea is a very good one -- that is, be able to track the history of a purticular test -- when it started succeeding or failing.
17:29 vel joined
putter so better organization... maybe an index? or a "table of contents"? giving a decision tree for what goes where? something else? 17:32
and... the ability to do a quick check, rather than a full "all the edge cases anyone has ever thought of" test run, to get faster feedback? 17:33
and something like the... mumble... test failure tracking system. tinderbox? 17:34
theorbtwo Tinderbox would be kind of cool, in a way. 17:37
More a system of reporting in semi-realtime changes in tests.
<tinderbot6> t/foo/bar/baz.t # 26 "Frobble's correctly" failed between r2743 and r2746 -- blame theorbtwo or putter! 17:38
(Tinderbox just tracks if it still compiles.)
putter yeah. good point. 17:47
actually, the whole "blame" aspect doesnt quite fit, since we're find with folks breaking the build. but knowing the rNNNN number would be great 17:48
17:48 hexmode joined
Juerd putter: query? 17:48
Oh, I'm not identified yet
Fucking freenode 17:49
theorbtwo My thinking was that it'd always be working on the latest thing. If there's a new checkin in the middle of building, the build is killed, svn up, restart building. 17:50
If it's killed partway through a test run, it can report thus far.
putter hi Juerd, ralf
colabti.de/irclogger/irclogger_log/...=701#l1034
theorbtwo Also it'd possibly be nice for it to be able to automatically handle pugsbugs. 17:51
That is, there should never be actual tests in pugsbugs, just symlinks to tests that have recently been failing. 17:52
17:52 brentdax joined
putter some of the pugsbugs have been done (but not all. I think tewk is working on that?) 17:52
rafl Who's ralf?
putter no progress yet on the ChangeLog. note we're up to r8131 I think, rather than the much later rev I mentioned above. 17:53
oooops. sorry. typo.
putter blushes
rafl will do changelogging about JIB/sixpan after getting something to eat. 17:54
putter awesome. :)
putter reviews t/README to see what exactly pugsbugs is... 17:56
doesnt mention it. a "audreyt please fix this" directory? ;) a "this one is getting in the way more than usual"? something else? 17:57
18:01 jdv79 joined
putter Oh, no, I was right the first time. The ChangeLog is up to r8657. gaal++ 18:04
18:05 rantanplan_ joined
putter tewk: are you still plugging away on postincrement_in_subscripts.t ? 18:12
svnbot6 r8811 | putter++ | ChangeLog up to r8670. Hey, that was easy!
putter bbiab 18:18
tewk putter: got distracted, I can move it over. 18:20
18:21 cPbrian is now known as br4in
putter great 18:30
thanks tewk. after lunch, I'll try to clean up whatever hasn't been gotten to. 18:31
bbl 18:32
18:34 bernhard joined
svnbot6 r8812 | tewk++ | t/pugsbugs/postincrement_in_subscripts.t moved to t/operators/inc.t 18:36
18:43 Khisanth joined
kolibrie I built pugs and am now getting './pugs: error while loading shared libraries: libparrot.so.0.4.1: cannot open shared object file: No such file or directory' 19:08
I have parrot/blib/lib/libparrot.so.0.4.1
and /usr/local/lib/libparrot.so.0.4.1
any ideas what I need to do so they are seen?
rafl kolibrie: What does ldd pugs give you? 19:10
kolibrie rafl: lld: command not found 19:11
od
oh
rafl ldd, not lld
PerlJam kolibrie: ldd, not lld :)
kolibrie: also, perhaps you need to set your LD_LIBRARY_PATH to the appropriate directory(ies) 19:12
kolibrie ok, that got something. What is relevant?
libparrot.so.0.4.1 => not found
rafl Or add /usr/local/lib to /etc/ld.so.conf 19:13
kolibrie hmm, everything else is in /lib, not local/lib
rafl kolibrie: Update ld.so.conf or install parrot into /usr instead of /usr/local. 19:14
kolibrie rafl: ok, /etc/ld.so.conf updated 19:15
how do I make it take effect?
rafl kolibrie: ldconfig 19:20
19:21 nnunley joined
kolibrie rafl: ahh, much better. Thank you 19:22
PerlJam: and thank you, too
19:22 sahadev joined 19:29 khisanth_ joined 19:30 khisanth_ is now known as Khisanth, jp-autark joined 19:33 khisanth_ joined 19:43 khisanth_ joined 19:44 khisanth_ is now known as Khisanth
svnbot6 r8813 | putter++ | TASKS+=1 Explore how other langauges are structuring their validation tests. 19:49
r8814 | kolibrie++ | small maintenance to keep this example running
r8815 | putter++ | Moved the passing t/pugsbugs/numbered-match-variable.t file to t/rules/numbered_match_variable.t. 19:58
19:59 Cryptic_K joined 20:02 lypanov joined
svnbot6 r8816 | putter++ | Deleted the now working t/pugsbugs/parsing_do_my.t. Tests moved to t/var/my.t. 20:10
r8817 | putter++ | Deleted the now working t/pugsbugs/parsing_my.t. Test moved to t/var/my.t. 20:19
putter I think that's it for passing pugsbugs files. 20:20
20:21 dduncan joined
putter I'm going to start on ChangeLog. Doing it chunks. Anyone who wants to take over, or work in parallel, that would be great. 20:36
avar putter: generate it from commit messages? 20:42
svnbot6 r8818 | putter++ | ChangeLog to r8700. 20:46
putter avar: yes! :) svn log -v -r8NNN:8MMM > deleteme
usually one doesnt actually have to look at the patch. unless the log entry is a puzzle. 20:47
20:47 hexmode joined
putter want to take r8700:8730 and I'll take r8730:8760 ? 20:48
the "where to put it in ChangeLog" is pretty simple, there's a section for t/ docs/ examples/, another for misc/, another for ext/, and the mainstuff gets broken down as feature or fix, and within those, pugs or js. 20:50
avar me?
putter the what to say is easy to, something simple, that doesnt lose much information. can be crufty too, as there will be a cleanup pass. 20:51
avar: ya, you :)
dduncan allo
putter hey dduncan
dduncan is changelogging all that's left to do?
putter like to help ChangeLog'ing
20:52 stevan joined
avar I'm uh busy .. doing nothing whatsoever;) 20:52
putter nope, but its the last of audreyt's 'hings she'd like to see done while she slept'.
ext/ also needs to be worked through 20:53
dduncan for my part, I've completely changelogged my ext/ modules 20:54
and of course, this is a delta since 6.2.10, 4-5 months ago 20:55
putter oh, I meant debugging the rest of the failing ext/ modules. ;)
dduncan the problem is "no such method/sub" errors
which I think may be an omission in pugs itself not handling private ones 20:56
putter right
dduncan at least that's the case with my own ext/ modules, though mine don't fail tests because they only test for compiling right now
putter would you like to share in bringing the ChangeLog up to date. split with you alternate blocks of 50 revs? 20:57
dduncan perhaps ... though unless the solution is fixing pugs, I'm more inclined to track down ext/ and fix or todo the failures there 20:58
maybe
as I recall about the changelog
a pass was done through it recently
20:59 avarab joined
dduncan and certain sections were skipped as they should be done by people more intimitely aware of the sections they referred to 20:59
20:59 vike joined
putter starts on -r8700:8750 . something like 120 revs to go after that. 20:59
yep. but the scan has a ways to go before it reaches present. about 200 revs to go. 21:00
dduncan hm, I'm getting a "build failed" on my last make ... 21:01
will try again
putter i had a svn up; make failure. a rebuild from scratch worked. 21:05
pasteling "dduncan" at 24.108.164.7 pasted "link failure on 'make', undefined yaml symbols" (289 lines, 16.9K) at sial.org/pbot/15636 21:06
dduncan did you get the same error?
also, by "from scratch", do you mean with a 'make realclean' in the middle? 21:07
21:07 vike left
putter no. I only have two modes, simple make, and blow away the entire directory, and literally start from scratch. 21:08
s/tory, and/tory and/
dduncan are all the normal t/ passing or skipping now, with only ext/ left? 21:09
putter that (the paste) does look like what I saw. 21:10
dduncan and you fixed it with a blow-away
putter re tests, don't know. audreyt did assorted commits, but I don't think I've seen a smoke on the other side.
kolibrie dduncan: I did a make realclean and build successfully today 21:12
putter i just (cd SVN/pugs; tar cf - .)|(cd pugsxNN; tar xf -) or something similar 21:13
but a make clean of some sort should also do it, unless that's borked too.
dduncan well, I'll do a clean rebuild and smoke now ... will take about 3 hours
okay
blow away for me works like this 21:14
an advantage of svk is that the whole thing is backed up on my own machine, so no network pulls necessary
21:14 stennie joined
dduncan I just cd to the root pugs directory, delete all its contents, then svk revert -R it 21:14
21:15 Debolaz2 joined, integral joined
dduncan and it restores from the svn server based backup on my machine 21:15
putter I should be able to one a bit faster than that. we'll see. oh, drat. oh well, I'm going to get a make test too. ;) 21:17
svk++
dduncan that said, I'll try the less severe 'make realclean' first 21:18
21:24 avarab is now known as avar 21:29 _SamB_ joined 21:32 _SamB_ is now known as SamB
putter dduncan: nope, there are still core failures. split, arglist, array_mmd, array_ref, ... 21:43
svnbot6 r8819 | putter++ | ChangeLog to r8750.
putter takes a break from ChangeLog'ing 21:44
stevan: could you do the ChangeLog object model entry(ies)? 21:46
oh, wow. the phrase "best verification suite" appears nowhere in the googleable universe. 21:48
stevan putter: i will try, I am somewhat short on tuits lately 21:55
Juerd Everyone is :(
21:59 lisppaste3 joined 22:08 lisppaste3 joined
wolverian hmm, slate looks interesting 22:22
putter yes, doesn't it 22:23
slate on p6 anyone? p6 on slate?
nnunley More, porting mobius is interesting. 22:24
And perhaps the pmd stuff for multiple dispatch.
Slate itself is a little bit of a pain to program in, IMNSHO.
dduncan now starting a smoke of r1819 22:31
I mean, 8819
it should be done in about 2 hours 22:32
22:37 soam joined
putter mobius is slate's implementation. vm. slate.tunes.org/doc/ 22:45
nnunley: what didn't you like about programming in slate?
putter shakes out his brain. ada 95 test suite. eeps. archive.adaic.com/compilers/acvc/95acvc/acvc2_3/ 22:52
23:00 Shachaf left 23:05 drbean joined 23:12 Cryptic_K joined 23:25 soam left, avar joined 23:28 lightstep joined 23:31 r0nny joined
putter does apple still use : as a path sepaeator? 23:32
putter notes svn.myrealbox.com/viewcvs/trunk/mcs...ss/corlib/ is much clearer than the equivalent would be in p5, because a hierarchical name need not be expressed hierarchically in the file system. 23:34
hmm, though here it may just be the equivalent of PIL-Run (aka PIL::Run), a name for humans not understood by the system. dont know (mono). 23:35
re apple, no. osx is / 23:38
so, should we start permitting/using file names of the form mumble/Foo::Bar::Hee/frotz in the pugs distribution? 23:39
rafl What's the benefit from doing so? 23:42
putter flatter, unhidden, is clearer. ls giving Foo::Bar Foo::Hee::Qux Zim::Zog rather than just Foo Zim 23:44
dduncan: looks like my smoke ran faster. r8818 is up as m19s28.vlinux.de/iblech/stuff/pugs-...3430f.html 23:46
still lots of red.
rafl: no?
rafl putter: Sure that every single filesystem for every single platform we want to support can live with that?
putter no 23:47
especially if that includes pre osx apples. macintosh.
rafl I don't even know what that includes. And I know even less about the filesystems. So I fear doint that.
But what I would be interested in is a directory layout that supports multiple installed versions of a module. 23:48
putter yes. essential. 23:49
I may be insufficiently risk averse because i _really_ dislike the Foo/Bar/Hee "smash stuff down across the filesystem" game. 23:50
Juerd rafl: Have you sent the CF?
putter CF? 23:51
Juerd Yes.
putter saw "seen" for "sent". thought it was a nifty new thing to look at. never mind. ;)
Juerd Re filesystems, I once considered for a project of my own, to use Foo/Bar/Baz/0.13.foo kind of file names, with Foo/Bar/Baz/default being a symlink to the latest version. 23:52
rafl Juerd: Yes.
Juerd rafl: Okay; thanks. I hope it arrives before you do :)
rafl ;-) 23:53
Juerd 00:53 < putter> flatter, unhidden, is clearer.
putter Juerd: has there been any p6l discussion of this kind of thing?
Juerd Flatter isn't necessarily better. In fact, I think the opposite is true. 23:54
putter: Not that I know of.
rafl We should hack something together in docs/AES/, imho.
Juerd putter: I like the Foo/Bar.pm, and would dislike Foo::Bar.pm, because when it's Foo/Bar/Baz/Quux/Xyzzy, the use of "../Blah" makes life easier. 23:55
23:55 lightstep left
Juerd I wouldn't even mind changing :: to / ;) 23:55
use Foo/Bar;
Foo/Bar/BAZ++;
(I *know* it's not possible)
rafl I also get along with Foo/Bar.pm quite well. 23:56
putter Juerd: if I have a lump of functionality Languages::Lisp::Other::Lisp1::OnPerl6, I'd like to have a directory of that name. not a long dangling tail hanging off of the completely uninformative Languages/. no? 23:57
Juerd putter: "tree"
putter: "find"
You can already get any view you want 23:58
Mapping one hierarchical structure to another just makes sense.
:: is the path separator
putter hey, we can file it under L/a/n/g/u/a/g/e/s/-/L/i/s/p/... and the same tree/find argument still holds. :-P
rafl Beside that it's faster on most file systems to have files seperated into directories over multiple levels than to put lots of files into a single dir. 23:59
putter (but not admittedly the hierarchy <-> hierarchy argument)

Powered by Cro and the Raku® Programming Language.

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

Thank you!