»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg p6eval perl6: ... | irclog: irc.perl6.org/ | UTF-8 is our friend!
Set by sorear on 4 February 2011.
lichtkind [Coke]: so it say how utf will be normalized? 00:00
[Coke] s/utf/unicode/ 00:02
though I'd probably just say "Strings", instead of "Unicode"
lichtkind [Coke]: yes i was bit lazy just try to find a good explanation for the tablets, the index is slowly maturing 00:13
dalek blets: e358dfc | (Herbert Breunung)++ | docs/appendix- (2 files):
explain $?NF and repair some glossary links [Coke]++
00:29
dalek blets: 69f6d32 | (Herbert Breunung)++ | docs/appendix- (2 files):
more crosslinking for subrules
01:10
blets: d247077 | (Herbert Breunung)++ | docs/appendix-a-index.txt:
s/P5/perl 5/
blets: 3856104 | (Herbert Breunung)++ | docs/appendix- (2 files):
adding eulers number to glossary and linking to ops who using it
blets: 5b5069e | (Herbert Breunung)++ | docs/appendix-a-index.txt:
more crosslinks for $*UIDs
sorear colomon: hi 01:20
colomon \o
kind of curious how you knew I just got home from rehearsal. :)
sorear can you reproduce the oddity from earlier?
colomon: I just got home from dinner
oh, you don't have time yet then? 01:21
no rush
colomon give me a minute and I'll see if I can golf it for you.
(I can reproduce it, but it takes relatively a lot. Though I guess if you downloaded ABC I could send you the broken script which produces the error without golfing.) 01:22
sorear I have an ABC checkout
sorear I haven't done anything with it 01:22
colomon sorear: just e-mailed you everything you should need to reproduce the problem (long version) 01:26
want me to try to golf it as well, or no?
sorear No 01:30
dalek ast: ebb9ec5 | coke++ | S0 (13 files):
pugs unfudge
02:32
moritz \o 05:31
phenny moritz: 24 Apr 23:10Z <pmichaud> tell moritz in commit a83a4b the resulting text reads "The Rakudo developers would like to add the organizers from Oslo.pm". That doesn't read quite right to me -- what are we wanting to add them to? ;-)
dalek kudo/nom: 1a7e4df | moritz++ | docs/announce/2012.04.1:
[announce] wording fix from pmichaud++
05:43
dalek ecs: 053a580 | (Frederik Schwarzer)++ | S32-setting-library/Str.pod:
[S32] Grammar fix.
06:50
ecs: 17a1305 | (Frederik Schwarzer)++ | S32-setting-library/Str.pod:
[S32] Typo fix.
ecs: a835127 | moritz++ | S32-setting-library/Str.pod:
Merge pull request #11 from icwiener/master

Just some minor stuff I saw while reading.
frettled phenny: tell lichtkind Regarding the use of #foo to access an <a name> element in HTML: it is standard behavior that you're complaining about, it's what everybody expects to happen. There is of course no technical problem in having the <a name> element in another place than the header, you could put a hundred thousand lines of text inbetween the two if you wanted to. But that's bad design :) 07:16
phenny frettled: I'll pass that on when lichtkind is around.
eiro tadzik, ack 08:47
moritz r: gist.github.com/2484152 09:03
p6eval rakudo 1a7e4d: OUTPUT«94␤» 09:04
masak \o/ 09:04
of course, as soon as I went to bed, I thought of improvements. 09:05
arnsholt moritz: I've started sketching a fix to your sqlite problem 09:07
moritz r: my $c = 1; $c = 2 * $c + 1; for ^5; say $c
p6eval rakudo 1a7e4d: OUTPUT«===SORRY!===␤Missing block␤at /tmp/PlOpkKI8B5:1␤»
moritz r: my $c = 1; $c = 2 * $c + 1 for ^5; say $c
p6eval rakudo 1a7e4d: OUTPUT«63␤»
arnsholt It'll probably be a bit longer before it's all done, but it's mostly a matter of implementation I think
moritz arnsholt: ok. I don't plan to hold off the release for it
I'll release as soon as I've got a +1 from any <pmichaud jnthn> 09:08
masak moritz: it's ($c + 1) * 2
moritz r: my $c = 1; $c = 2 * $c + 2 for ^5; say $c
p6eval rakudo 1a7e4d: OUTPUT«94␤»
masak moritz: that's the inverse of $x / 2 - 1
moritz right
moritz finds his solution much mor straight forward and readable than masak++'s 09:09
masak the function replace-var is what I had to write to turn the AST "inside out" in the right way.
moritz: yeahbut.
moritz but of course it's much less interesting :-)
masak the objective of this exercise is not to compute the inverse yourself.
state the problem *forwards* and have the computer do the numbers.
arnsholt moritz: Yeah, getting it shipped will depend on my tuit supply, so definitely don't wait for it
moritz r: sub guard($x) { $x / 2 - 1}; sub all_guards($x is copy) { $x = guard($x) for ^5 }; for 0..* { if all_guards($_) == 1 { .say; last } } 09:11
p6eval rakudo 1a7e4d: OUTPUT«(timeout)»
moritz r: sub guard($x) { $x / 2 - 1}; sub all_guards($x is copy) { $x = guard($x) for ^5 }; for 0..100 { if all_guards($_) == 1 { .say; last } } 09:12
p6eval rakudo 1a7e4d: ( no output )
moritz r: sub guard($x) { $x / 2 - 1}; sub all_guards($x is copy) { $x = guard($x) for ^5 }; say all_guards(93)
p6eval rakudo 1a7e4d: OUTPUT«0.96875 0.96875 0.96875 0.96875 0.96875␤»
moritz oh 09:13
r: sub guard($x) { $x / 2 - 1}; sub all_guards($x is copy) { $x = guard($x) for ^5; $x }; for 0..100 { if all_guards($_) == 1 { .say; last } }
p6eval rakudo 1a7e4d: OUTPUT«94␤»
masak heh. 09:14
moritz r: sub guard($x) { $x / 2 - 1}; sub all_guards($x is copy) { $x = guard($x) for ^5; $x }; say (0..*).first: { all_guards($_) == 1 } 09:17
p6eval rakudo 1a7e4d: OUTPUT«94␤»
masak I'm in a superposition of thinking "straightforward" and "brute-force" ;) 09:19
I guess I have to make a module of my symbol handling, and then I can solve the problem in three lines ;)
moritz aren't you supposed to tweak the language until you can? :-) 09:20
masak modules are language tweaks ;)
moritz when I wrote Math::Expression::Evaluator, I had to constrain myself not to turn it into a CAS :-) 09:21
masak 'use Algebra::Symbolic; my Var $x .= new; my $y = $x; $y = guard($y) for ^5; solve($y == 1); say $x'
ok, five lines ;) 09:22
moritz plus the definition of &guard
masak I can make it three if I eliminate $y and write 'solve( guard(guard(guard(guard(guard($x))))) == 1 );'
right.
Ulti there is symbolic algebra in Rakudo??? 09:32
timotimo if you can implement it in perl6, it'll be in rakudo, sure 09:33
masak Ulti: well, the point is that you can make it easily as I do in gist.github.com/2484152
Ulti: it's not a complete solution by far, but it solves the problem at hand.
timotimo cute
moritz the basic idea is the same as "expression templates" in C++ 09:34
masak Ulti: what makes it easy is that you can easily co-opt all the operators to not calculate things right away, but to form syntax trees which you can then manipulate any which way you want.
moritz you overload operators to return a data structure that returns the operation and the operands
frettled guard⁵, as I wrote the other day, would be kindof cool .)
masak std: sub postfix:<⁵>($) {} 09:35
p6eval std fab64fc: OUTPUT«ok 00:00 41m␤»
moritz p6: say so '⁵' ~~ /<:L>/ 09:35
p6eval rakudo 1a7e4d, niecza v16-23-gaa61ed5: OUTPUT«False␤»
..pugs: OUTPUT«Error eval perl5: "if (!$INC{'Pugs/Runtime/Match/HsBridge.pm'}) {␤ unshift @INC, '/home/p6eval/.cabal/share/Pugs-6.2.13.20120203/blib6/pugs/perl5/lib';␤ eval q[require 'Pugs/Runtime/Match/HsBridge.pm'] or die $@;␤}␤'Pugs::Runtime::Match::HsBridge'␤"␤*** Can't locate P…
moritz yes, should work
masak frettled: to me, that's too sugary-sweet, but YMMV. :) 09:36
timotimo what does :L match?
masak letters?
moritz r: multi sub postfix:<⁵>($x) { $x ** 5 }; say 2⁵
p6eval rakudo 1a7e4d: OUTPUT«32␤»
moritz a single letter, to be precise
Ulti is letter defined as from any language? 09:37
moritz r: multi sub postfix:<⁵>($x) { $x ** 5 }; say 2⁵⁵
p6eval rakudo 1a7e4d: OUTPUT«33554432␤»
frettled masak: thank goodness, I was worried that you might suggest that it was aspartame sweet…
moritz that doesn't assoc right, does it?
masak r: multi sub postfix:<⁵>($x) { $x ** 5 }; my &id = -> $x { $x }; multi infix:<**>(&code, 0) { &id }; multi infix:<**>(&code, Int $n where * > 0) { -> $x { &code((&code ** ($n - 1))($x)) } }; say &guard⁵(94)
p6eval rakudo 1a7e4d: OUTPUT«use of uninitialized variable $x of type Nil in numeric context in method Numeric at src/gen/CORE.setting:653␤␤Method 'postcircumfix:<( )>' not found for invocant of class 'Int'␤ in <anon> at src/gen/BOOTSTRAP.pm:815␤ in any <anon> at src/gen/BOOTSTRAP.pm:811␤ …
masak :/ 09:38
oh!
forgot &guard, didn't I? :)
r: multi sub postfix:<⁵>($x) { $x ** 5 }; my &id = -> $x { $x }; multi infix:<**>(&code, 0) { &id }; multi infix:<**>(&code, Int $n where * > 0) { -> $x { &code((&code ** ($n - 1))($x)) } }; sub guard($x) { $x / 2 - 1 }; say &guard⁵(94)
p6eval rakudo 1a7e4d: OUTPUT«1␤»
masak \o/
Ulti moritz: this is cool perlgeek.de/blog-en/perl-6/physical...lling.html
moritz Ulti: I'm glad somebody likes it 09:39
masak likes it too
moritz Ulti: I presented it as YAPC::EU 2010 in Pisa, and the echo was mixed
Ulti really? surely the worst reaction was *meh* rather than negative?
moritz Ulti: the worst reaction was from somebody who hadn't read the abstract of my talk, and execpted something very different :-) 09:40
Ulti heh so the standard mixed reaction then
this doesn't solve all my problems??? why did I sit through this, you are worthless!
arnsholt moritz: Oooh, workaround for you (maybe)
Try doing @conn[1] = OpaquePointer and then calling sqlite3_open 09:41
frettled masak: so now you just need to generalize that code to work for all of ⁰⁻⁹ ;)
Ulti moritz: that style of programming would be really great for writing control systems
moritz Ulti: the talk was titled "Perl 6 and the real world", and that somebody expected it to be about real-life experience with Perl 6, not about modelling real life :-)
arnsholt Ah, nm. Won't work after all 09:42
masak frettled: I bet it could be done via a macro :P
arnsholt OTOH, it might give you a segfault =)
timotimo moritz: that's a little bit misleading, i agree
but that's what abstracts are for, so there's no reason to let that confuse you 09:43
moritz arnsholt: it doesn't change anything, apparently
frettled masak: \o/
Ulti with word modelling added between 'and' and 'the' might have avoided that I guess
frettled masak: A macro with a lookup table! ;)
arnsholt moritz: Right. I get a segfault locally if I just assign to index 1 and fetch index 0
moritz arnsholt: ah right, if I remove index 0, I get a segfault too 09:44
arnsholt I should probably try to fix that ^_^ 09:45
moritz Ulti: I was playing with people's expectations deliberately, I confess. But I expected them to read the abstract, so that their expectations and mine could go into sync before the talk :-)
frettled masak: writing the code for handling 2⁵⁵ correctly (and not as (2⁵)⁵) would probably be a nice kata 09:46
timotimo moritz: have you done anything with Math::Model after that? have you tried if it got any faster? or if it works on niecza? did anyone build that chaotic model? 09:47
moritz timotimo: well, for one the memory leaks stopped :-) 09:48
timotimo: and I think I made it work on rakudo/nom
it didn't run on niezca last I tried
but that was quite a while ago
masak frettled: I don't see how it could be without serious magic. 09:49
timotimo mhm
masak given that each ⁵ is one postfix operator, I mean.
moritz no problem with an 'is parsed' macro 09:50
which just matches all such numerical powers in one go, and code-gens them with the right associativity
frettled There was something like that I was hoping for.
timotimo is the syntax and semantics needed for that defined/modeled/explained yet? 09:51
moritz ok, niecza runs Math::RungeKutta (and does so fast :-) 09:52
timotimo: not in detail
niecza hiccups on one of the xml-writer tests 09:53
but that's just an issue with the tests, not the module (uses $str.match(:x(2), rx/.../)
daxim follow-up from yesterday: rt.perl.org/rt3/Ticket/Display.htm...xn-1097430 # happy end 10:00
masak this is very interesting: www.schneier.com/blog/archives/201...und_o.html 10:04
moritz, you need to read/watch that. :)
moritz masak: can't watch right now while at work :-) 10:05
masak: but the combination of prisoner's dilemma and games brought back memories of a very nice game, en.wikipedia.org/wiki/Hoity_Toity 10:07
masak moritz: I'm watching them now. it's fascinating. 10:08
moritz each round, the players can chose either between competing, stealing or playing detective. It's quite fun to play (though I don't have it at home :/)
masak moritz: I hadn't really grok'd how fragile trust is in this kind of contest. 10:09
honestly, I wouldn't want to play this kind of game against someone who holds a grudge, and who can't compartmentalize things that happen in a game. :)
moritz same here 10:10
tadzik eiro: ack as in ACK or as in "aargh"? :)
moritz so, I'll never play it against my wife :-) 10:10
masak ;)
moritz: I haven't watched all videos yet. but so far, all contestants say the same thing to each other: "let's be reasonable. let's both choose 'split'." 10:11
so clearly, the words are meaningless -- and most likely, everybody knows this.
and it's all in the nonverbal communication. 10:12
moritz and they have no way to enforce it :-)
moritz oh, I just got bitten by the lack of filling positionals by name :-) 10:13
masak :P
moritz the Math::Model tests relied on it
and since it's a method call, no error was thrown due to excess named args 10:14
masak interesting. there seems to be some correlation between contestant A being very trustworthy and contestant B defecting -- and then feeling genuine regret afterwards. 10:15
masak also interesting. there's only one reason to split (because you trust the other to do so), but two reasons to steal (either because you think the other trusts you to split, or because you think the other will double-cross you and steal). 10:18
moritz r: class A { has %.h; method x { %.h<a> } }; say A.new(h => { a => 5 }).x 10:20
p6eval rakudo 1a7e4d: OUTPUT«5␤»
moritz r: class A { has %.h; method x { %.h<a> } }; say A.new(:h{ a => 5 }).x
p6eval rakudo 1a7e4d: OUTPUT«5␤»
moritz and the second problem with Math::Model was that it used defined $v ?? ... !! ... ; 10:24
which stopped working after we switched &defined to be a listop 10:25
so, now it works again on nom
phenny: tell sorear it would be nice of niecza's lives_ok and eval_lives_ok could diag() the exception when one is thrown 10:28
phenny moritz: I'll pass that on when sorear is around.
masak moritz: I watched all the videos. make sure you watch them when you're $home. worth it. 10:30
dalek p: a04da69 | (Arne Skjærholt)++ | src/6model/reprs/CArray.c:
Make sure initial child_objs in CArray is zeroed out.
10:47
masak heh. p6 split <=> p5 split. but p6 comb <=> p5 m//g. 11:27
eiro tadzik, ACK: there is a chance you'll be there so i will try to spare some extra boxes
masak \o/ 11:30
tadzik yay 11:33
moritz masak: having some research and experience in the back of my mind, I'd be very sure to shake hands with the other player before they pick the ball 11:34
masak: at least in .de there seems to be much more reluctance to break word+handshake than just the word 11:35
masak ooh
but wait till you see the last video ;)
eiro is panda still used ? 11:38
tadzik sure 11:39
and quite actively developed these days :)
eiro cool
can panda install the perl6 toolset? 11:40
parrot+rakudo ? 11:41
tadzik nope
panda is written in Perl 6 and installs stuff from modules.perl6.orgj
masak probably better.
tadzik and it can install itself too :)
masak proto auto-installed Rakudo for you.
eiro panda isn't the proto successor ? 11:42
masak it's the neutro successor, which was the proto successor ;)
tadzik well, it's a bit like pls successor which was proto successor, because it didn't want to be like neutro, which was proto successor 11:43
or something
I think there's more than one axis
eiro well ... don't you think panda should be shipped with a perl6 installer in the bootstrap ?
tadzik no 11:44
do cpan clients install Perl 5?
eiro ok ... so what's the easiest way to install perl6 ?
tadzik installing Rakudo Star
eiro tadzik, makes sense: question cancelled, your honnor
tadzik which gives you panda for free, iirc
eiro ahhh! this way :) 11:45
ok thanks
tadzik aye :)
mikec where does panda install modules to? 11:46
tadzik ~/.perl6/
but we're thinking about that one 11:47
mikec ah ok
in wondered if you could install rakudo star, upgrade rakudo and parrot, and still use panda to install modules 11:48
tadzik sure
but note that modules precompiled for the old rakudo won't work with new rakudo 11:49
and you'll see fireworks :)
mikec yeah, i found that problem
tadzik one should probably write something that rebuilds all modules if necessary
like panda --rebuild-ALL-THE-THINGS
mikec but i fetched them from source rather than using panda, which would have been easier
heh 11:50
tadzik if you precompile them the problem stays
you can always find ~/.perl6/ -name '*.pir' -delete
mikec aha
tadzik and lose plenty of speed :/
mikec how would a --rebuild-everything work? just go through everything in .perl6? 11:51
masak by the way, hakank++ built a MiniZinc model for the guards-and-apples problem: www.hakank.org/minizinc/guards_and_apples2.mzn 11:53
tadzik mikec: I guess so 11:54
mikec i guess there's no easy way to check whether the precompiled code is out of date
eiro this page is outdated, right ? rakudo.org/how-to-get-rakudo-nom/
i mean: nom is the current branch now 11:55
masak eiro: well, it says that.
do you have a really old cache?
it's even in the URL ;)
mikec heh 11:56
masak or, maybe you just mean that `git checkout nom` is rather redundant?
masak it is. it should be a no-op. 11:56
eiro redundant for you and disturbing for a newcomer 11:57
no ?
mikec i wouldn't say so. just makes it extra clear which branch you're getting? 11:58
eiro i mean: there is no "perl6 for impatients" in the perl6.org homepage
mikec although it's probably an accident
eiro there is no "welcome" banner (i really think like: and if i was a newbie) 11:59
mikec ah
eiro i'm installing perl6 (again :) ) so i'll try to provide a pr 12:00
masak "Programming language tutorials shouldn't be about learning languages. They should be about something interesting, and you learn the language in the process." -- prog21.dadgum.com/67.html
lichtkind hai 12:36
phenny lichtkind: 07:16Z <frettled> tell lichtkind Regarding the use of #foo to access an <a name> element in HTML: it is standard behavior that you're complaining about, it's what everybody expects to happen. There is of course no technical problem in having the <a name> element in another place than the header, you could put a hundred thousand lines of text inbetween the two if you wanted to. But that's bad design :)
lichtkind frettled: si ID is better? 12:37
frettled frettled: What is "si ID"? 12:38
hups
Nice typo :) 12:39
flussence well you don't need a dummy <a> element to use id=, so I'd say yes 12:41
frettled uh?
Ah, so lichtkind is asking whether you can use the ID attribute instead of <a name="foo"></a> to make example/#foo reach the foo section? No, that is not how it works.
frettled If you want _that_ to work, you need to employ CSS hiding/unhiding and javascript, which will certainly break navigation. 12:41
flussence are you talking about some sort of fancy thing using multiple pages? 12:42
frettled flussence: no
flussence: I'm not sure if you caught the original discussion yesterday evening, did you?
flussence I'm lost, I probably missed something
masak I had the impression that the 'id' attribute more or less supplanted named anchors.
flussence scrollbacks
masak at least in some best-practices sense. 12:43
frettled masak: the ID attribute provides a way of specifying an identificator for an element, which can later be used by e.g. CSS to specify properties, or by javascript to manipulate the DOM.
flussence: irclog.perlgeek.de/perl6/2012-04-24#i_5489786 12:44
masak frettled: yes, but they also function as anchors, no? 12:45
frettled masak: it is possible that this has been specified to work as an anchor as well
lichtkind frettled: i used previously name and they talekd so long into me that i started do <a id="Z-op"/> which works too
masak frettled: I'm pretty sure they do, and that it's quite a conscious feature. 12:45
lichtkind frettled: saying that that is the standard 12:45
frettled www.w3.org/TR/html401/struct/links.html#anchors
both seem to be standard now, I'd missed that
flussence I'm pretty sure even IE6 supports id= as a fragment identifier 12:46
frettled but it has nothing to do with lichtkind's problem
In fact, to get the behaviouf lichtkind desires, the ID attribute is wholly unsuitable, only <a name> will work
s/iouf/iour/
But then again, if I understood lichtkind's wishes correctly, those wishes are against BCP :) 12:47
lichtkind frettled: i just want normal inter noc anchors 12:48
lichtkind people just suggested here that ID is best practice 12:49
i mean inner doc anchors
frettled lichtkind: ah, you appeared to say that you wanted clicking on such a link to show you the lines _above_ the anchor as well 12:50
lichtkind: which is the opposite of normal
The ones on the webpage discussed _are_ behaving normally.
flussence <h1 id="w">W</h1> is more idiomatic than <h1><a name="w"></a>W<h1>, if that's what you're asking... 12:51
frettled flussence: well, that's debatable
(when I learned to write HTML, it was <a name="w"></a>\n<h1>W</h1>, though) 12:52
lichtkind frettled: allright i didnt know what your anserwing, that was my first discussiom , yes yesterday i wanted to move the anchors so that more of the page is visible, its just UI thing asking you as user
lichtkind if your not confused 12:52
if you see the previous item then you clicked on 12:53
flussence oh, I see what you want now...
frettled If there is a long list of things, and you use internal anchors to point to specific points in that list, that is a sign that your document is too long, and needs to be split into smaller documents. 12:54
masak er. I wouldn't move the anchor away from the element it's referencing. that way lies madness. 12:55
frettled There is nothing wrong with serving the whole document for certain uses, for e.g. print (which is something I hope will disappear).
masak also, it seems to be a clear mixing-up of content and presentation.
frettled masak: yup
masak: regarding moving the anchor away: the element it's referencing is the element it's closest to :) Sometimes you may want a graphical element before the sub heading or whatever, and then it makes sense to have the anchor before the image, not after. 12:57
lichtkind frettled: the document is long because perl 6 has so many builtins, talk with larry :) 12:58
frettled lichtkind: I was speaking generally. 12:58
lichtkind: but are you saying that TimToady is maintaining that web page? :) 12:59
lichtkind no that i just write what he pulles out
frettled So splitting the document into smaller documents, e.g. one per letter of the alphabet, is for you to do, should you choose to take that piece of advice :) 13:00
flussence if you *really* want to show some scrollback then there's a nasty CSS hack for that: give each heading a big top padding and an equal-but-negative top margin. 13:02
frettled ooh 13:02
flussence (but if you go that route, I'd recommend you put <section id=> tags in instead and put the spacing on those, so it doesn't mess up any the headings already have 13:02
frettled I'd rather put some context in the form of forward and back links with minor snippets of what you're missing out on
flussence *)
.oO( in an ideal world we'd only need "h1:target { anchor-offset: 5em }" or something )
13:06
lichtkind frettled: this way i get easily over 100 pages which doesnt sounds nice to me
frettled lichtkind: why? There are 26 letters in the alphabet, and you have one section for specials 13:08
That looks like 27 documents to me, by separating in the same way that you use internal anchors today.
bonus: quicker loading/display times
lichtkind frettled: we have 7 apendices +10 tablets if splitting them all with that logic i get over 100 pages very fast 13:09
frettled lichtkind: okay, but why is that a problem?
lichtkind we using html now which is much faster than socialtext was 13:10
frettled: i see no problem in some kb of text, i just asked if its more pleasent if the item you click to is not on the uppermost border 13:11
frettled lichtkind: and it generally is not more pleasant
lichtkind frettled: allright so it stays on top boder, thanks :) 13:13
frettled lichtkind: np, glad to be of annoying assistance :) 13:16
lichtkind haha 13:17
frettled: please have a look tablets.perl6.org 13:18
frettled lichtkind: any change since five minutes ago?
lichtkind no :) 13:20
but soon :)
dalek blets: a43824e | (Herbert Breunung)++ | docs/appendix-a-index.txt:
polish entries for chrs, ords, orelse, andthen
13:23
blets: 339d632 | (Herbert Breunung)++ | docs/appendix-a-index.txt:
add trim and some crosslinks
blets: 3d73921 | (Herbert Breunung)++ | docs/appendix-a-index.txt:
bringing more of S32/Str
masak lichtkind: re tablets.perl6.org/tablet-6-blocks.html -- "2.3.2 continue" -- continue doesn't exist anymore. maybe you're thinking of 'next' for jumping an iteration ahead in loops, or 'proceed' for not automatically breaking out of a 'when' block. 13:25
[Coke] learns something new (almost) every day here. I had no idea the #foo in urls now worked with ids.
masak [Coke]: once you know it, it's easy to see why they're spelled '#foo' in CSS too, though. 13:26
I think it's a really established feature. like, HTML 4. 13:27
or something.
lichtkind masak: thanks but i didnt looked into that since years, i currently busy bringing index uptodate and then going on with lower nimber tablets
masak lichtkind: ok. just letting you know there's an error there. do with that information what you will.
lichtkind masak: i checked index b i would but that out now, but it wasnt in there , the blocks tablets which will be rewritten anyway, it just has a sceleton that i cen set the right links into it 13:29
masak it's a bit hard to understand what you mean when you're writing underwater. but I think I got the gist of it. 13:31
dalek blets: 53a56f1 | (Herbert Breunung)++ | docs/appendix-a-index.txt:
more crosslinks+small format fixes
13:52
lichtkind masak: i under blub you soblub either
masak .oO( the Blub paradox ) 13:58
ab5tract likes the Perl 6 Tablets :D 14:01
masak ab5tract! \o/ 14:02
dalek blets: a91b786 | (Herbert Breunung)++ | docs/appendix-a-index.txt:
no more ucfirst but titlecase
14:03
lichtkind ab5tract: thank you
ab5tract lichtkind++ :)
wrote my first perl 5 in a decade. was fun to get back into a perl way of thinking. 14:04
lichtkind ab5tract: how much you read?
ab5tract though i had completely forgotten about the referencing system :/
lichtkind, just started on the history section 14:05
lichtkind ab5tract: perl 6 has no reference :)
ab5tract: this i consider not even started :9
but thanks for double karma
but i have to go
shopping time
ab5tract lots of stuff to add to history (where i might actually be of some use) 14:06
lichtkind some food, and cycle parts
ab5tract: that would be grat
ab5tract: im most pouf of tablets.perl6.org/appendix-a-index.html
but just look into our git repo and clone it or get commit bits
ab5tract lichtkind, sure. 14:08
dalek blets: 1f8d1f4 | (Herbert Breunung)++ | docs/ (2 files):
added link to github repo
14:09
masak lichtkind: in tablets.perl6.org/tablet-1-language-design.html -- "But every context is its own domain, optimized to solve the problems of that domain." -- what does this mean?
how is a context its own domain? 14:10
lichtkind masak: im gone and dont remember anyway :)
masak it sounds almost, but not quite, completely meaningless to me.
lichtkind masak: maybe just bad translation, these are the parts i wrote in geman 14:11
masak a bit like "Every situation is its own state." or "Every country is its own region."
flussence
.oO( something something lexical context )
14:12
masak under "One pass parsing" on the same page: "As a nice side effect: syntax errors can only damage the understanding of small units." -- I'm not sure that follows from one-pass parsing. 14:13
what TimToady keeps putting forward as the immediate advantage of one-pass parsing is: "You always know which language you're parsing." 14:15
the classical counterexample from Perl 5 would probably be a m//x expression with a '#'-style comment with an accidental slash in it.
since Perl 5 only searches blindly for the closing '/', it will find the one in the comment. hilarity ensues. 14:16
Perl 6 doesn't do blind searches forward to the next closing delimiter. ever. it always keeps track of how far nested it is down the sublanguage stack.
[Coke] 10:14 <@Coke> huh. autofudge is running, not emitting any output, and I don't have any perl6 processes running.
10:14 <@Coke> I think this means /fudge/ is hanging.
10:15 <@Coke> argh, and autounfudge doesn't respond well when you kill fudge 14:17
also, "make spectest" should not "git pull", but "git pull --rebase". no?
masak I generally like 'git pull --rebase', but I'm inclined to say that "make spectest" should die if you have unpushed commits. 14:19
[Coke] masak: then how do I test my work in progress? (commit. at which point I've now got a merge commit) 14:21
masak [Coke]: my point is this: I'd rather have 'make spectest' fail fast and leave it up to me to take action, than it silently introducing merge commits in my local copy. 14:26
this scenario only affects core contributors anyway, so it's easy to teach everyone what to do if 'make spectest' fails fast. :) 14:27
[Coke] if I have local commits and do a git pull, will that not introduce a merge commit silently? 14:28
masak if you have local commits and there are remote commits and you do a git pull without --rebase, you'll get a merge commit.
(so don't do that)
I just mean, I don't think the solution is for 'make spectest' to do something else silently. 14:29
[Coke] masak: so, silent merge commits trump other silent things. got it. ;) 14:34
pmichaud good morning, #perl6 14:35
masak [Coke]: no, failing fast trumps silent behavior.
pmichaud! \o/
masak autopun: twitter.com/sixthformpoet/status/19...7042370561 14:35
pmichaud perhaps 'git pull' could be run with --ff-only 14:40
that prevents a merge commit, while still doing the right thing in most cases
tadzik like pull --rebase? 14:41
pmichaud I think pull --rebase requires that there be no local uncommitted changes
tadzik possibly
masak +1 14:42
[Coke] masak: but it... you're not.. nevermind. 14:46
[Coke] catches up with pmichaud. Ok, that sounds better.
tadzik ah, I didn't backlog enough 14:49
masak [Coke]: I understand that we were talking past each other, but not exactly how or why. anyway, nevermind. :) 14:51
[Coke] masak: exactly!
hugme: hug masak ... TO DEATH 14:52
hugme hugs masak
[Coke] ... no, you... stupid bot.
masak :P 14:53
also, *phew* :)
hugme: you can let go now. 14:54
masak tries to shake off hugme
hugme: hey!
tadzik like a Face Hugger 14:59
masak eww 15:00
tadzik I hope I didn't spoil hugme for you 15:01
hugme: hug masak
hugme hugs masak
tadzik see? Harmless
masak :)
I should've thought of that myself. 15:02
[Coke] hugme: hugs? 15:04
hugme [Coke]:
[Coke] r: try {die 'eek'} ; say $! 15:08
p6eval rakudo 1a7e4d: OUTPUT«eek␤ in block <anon> at /tmp/C5ap9F9Itg:1␤␤»
[Coke] r: try {die 'eek'} ; say $::("!");
p6eval rakudo 1a7e4d: OUTPUT«eek␤ in block <anon> at /tmp/jI5eszEGvt:1␤␤»
[Coke] r: say %*ENV 15:09
p6eval rakudo 1a7e4d: OUTPUT«("PERLBREW_VERSION" => "0.42", "MAIL" => "/var/mail/p6eval", "USER" => "p6eval", "SHLVL" => "1", "HOME" => "/home/p6eval", "OLDPWD" => "/home/p6eval", "PERLBREW_ROOT" => "/home/p6eval/perl5/perlbrew", "PERLBREW_MANPATH" => "/home/p6eval/perl5/perlbrew/perls/perl-5.…
[Coke] r: say ::%*ENV
p6eval rakudo 1a7e4d: OUTPUT«===SORRY!===␤Confused␤at /tmp/7oMHWEQKOc:1␤»
[Coke] r: say ::("%*ENV")
p6eval rakudo 1a7e4d: OUTPUT«Symbol '%*ENV' not found␤ in method gist at src/gen/CORE.setting:8667␤ in sub say at src/gen/CORE.setting:6555␤ in block <anon> at /tmp/HsksPDCmzw:1␤␤»
pmichaud global dynamic vars tend to live in the PROCESS:: namespace, iirc. 15:10
r: say %PROCESS::ENV
p6eval rakudo 1a7e4d: OUTPUT«("PERLBREW_VERSION" => "0.42", "MAIL" => "/var/mail/p6eval", "USER" => "p6eval", "SHLVL" => "1", "HOME" => "/home/p6eval", "OLDPWD" => "/home/p6eval", "PERLBREW_ROOT" => "/home/p6eval/perl5/perlbrew", "PERLBREW_MANPATH" => "/home/p6eval/perl5/perlbrew/perls/perl-5.…
[Coke] pmichaud: ok. old pugsian test, then. 15:13
brrt is there a perl5-ism for reduce? 15:15
(this is a p5 question, really)
moritz o/
brrt: I think List::Util or List::MoreUtil have a reduce function
PerlJam List::Util::reduce 15:16
brrt thx 15:19
spider-mario I have not been able to find an equivalent for scan, however. 15:22
(Perl 6’s [\…]) 15:23
I’ve tried Language::Functional::scan but it seems buggy 15:24
producing only two elements for a list that has six
masak couldn't you implement scan easily-ish in terms of reduce? 15:46
masak hm, maybe not. 15:49
masak decommutes 16:00
Juerd Please don't ever make the whitespace around infix / mandatory. 16:02
I discovered that I tend to leave it out in simple divisions.
TimToady don't define a postfix beginning with / then 16:03
Juerd (e.g. $foo + 1/2)
TimToady: I promise I won't
TimToady we are also unlikely to :)
Juerd I know, but there's also the thing that happened to -
Juerd Even without any postfix beginning with that character :) 16:04
TimToady .u minus
Juerd For - I don't mind, but for / I'd have a hard time dealing with it :)
phenny U+2216 SET MINUS (∖)
Juerd $foo-bar 16:04
TimToady .u −
phenny U+2212 MINUS SIGN (−)
TimToady we need to make everyone use that instead of - :−) 16:05
Juerd My keyboard has only 86 keys.
TimToady that's why my smiley has a long nose
Juerd Apparently it has 87 keys. Hm. 16:06
Nose optional!
Hm, compose - n could be endash, compose - m could be emdash...
And compose - - minus :) 16:07
TimToady I hate my compose key file, it doesn't have any way to make U+2212 16:08
last I checked, I can't swap out my compose for a saner one and also keep anthy, alas 16:09
it gives me MINUS SIGN WITH DOT BELOW but no MINUS SIGN, grrr 16:11
TimToady perl6: sub infix:<−> ($a,$b) { $a - $b }; my $x = 43; say $x−1; 16:14
p6eval pugs, rakudo 1a7e4d, niecza v16-23-gaa61ed5: OUTPUT«42␤»
TimToady the nice thing about the - situation is that if you do leave out the same, the next thing is unlikely to start with a letter, and if it does, you'll get a compile-time error 16:16
s/same/space/
that was a weird braino
PerlJam starts writing the Acme module that will make / valid in identifiers ;) 16:25
sorear TimToady: for me, I can't use compose *at all* with anthy running 16:28
phenny sorear: 10:28Z <moritz> tell sorear it would be nice of niecza's lives_ok and eval_lives_ok could diag() the exception when one is thrown
sorear good * #perl6 16:29
TimToady sorear: even in Input method Off state?
do you have "IBus/Preferences/Advanced/Use system keyboard layout" checked? 16:34
TimToady or do you have a ~/.ibus/tables/compose-user.db file? (I don't know if I'm using that) 16:39
sorear TimToady: how do you access Input method Off state? is that "Enable or disable:" ? 16:56
use system keyboard layout is checked
I do have a compose-user.db 16:57
symptomatically, "Compose" is simply ignored while ibus is running
I press compose ' e, the latter two chars echo instantly
dalek rl6-bench: 177960d | (Geoffrey Broadwell)++ | TODO:
TODO: put two up ...
17:17
rl6-bench: 6c541cd | (Geoffrey Broadwell)++ | / (2 files):
TODO: ... and knock two down. With individual test skipping now available, perlito6.js/v8 is now reenabled, with just rc-forest-fire skipped.
ab5tract perlito6.js ? 17:22
is that a js implementation of perl 6?
sorear ab5tract: what about it? 17:22
ab5tract: yes 17:23
ab5tract curious :)
ab5tract sorear, where can i find out more? 17:23
sorear ask fglock or pmurias
ab5tract found the github page 17:26
interesting stuff
dalek ecza: 6ef6b2c | sorear++ | lib/Test.pm6:
Diag the exception in lives_ok et al (moritz)
17:29
dalek ast: 1cc6e47 | coke++ | S0 (18 files):
rakudo (auto) (un)fudge
18:04
[Coke] moritz, jnthn, pmichaud: that was a pretty mechanical autounfudge, with some minor undos for things that were clearly marked as "if you todo this, it will give false positives" (there is an autounfudge marker you can put in the skip message to help enforce that, but I forget) 18:07
masak home 18:08
[Coke] wonders if cabin in the woods is playing out where jnthn and co live. 18:15
diakopter hugme: hug masak...TO_DEATH 18:26
hugme hugs masak...TO_DEATH
flussence wonders if hugme is unicode-whitespace-aware... 18:27
[Coke] # 04/25/2012 - rakudo++ ; niecza (94.32%); pugs (42.27%) 18:35
"niecza", 20362, 2, 749, 1532, 22645, 24154
"pugs" , 9125, 2, 3341, 1553, 14021, 24015
"rakudo", 21586, 30, 644, 1932, 24192, 24164
diakopter what are the columns there?
[Coke] "Impl", "pass","fail","todo","skip","plan","spec" 18:37
[Coke] diakopter: percentages in the comment line are of the "lead" implementation. 18:38
benabik Why does the spec column vary? 18:40
moritz because we don't really know how many spectests there are 18:41
and for those test files with dynamic test plan, our guess depends on what the implementation does when running the test file
[Coke] mainly it varies because each impl runs different spec test files. 18:56
AIUI.
[Coke] sorear: gist.github.com/2492335 # niecza spectest failure 19:04
[Coke] or colomon 19:06
colomon eh? 19:06
colomon that's new since ... two days ago? yesterday? 19:07
I'll take a look when I get back from the post office. 19:08
wow, that is a lot of roast files changed since I last did a spectest.
dalek p: c2ae263 | moritz++ | VERSION:
bump VERSION to 2012.04.1
19:26
spider-mario oh
I hadn’t noticed 2012.04 was there. 19:27
TimToady April Fool's
spider-mario :D 19:28
dalek kudo/nom: 3447164 | moritz++ | tools/build/NQP_REVISION:
bump NQP revision to 2012.04.1
19:36
blets: 62aa1b7 | (Herbert Breunung)++ | docs/appendix-d-delta.txt:
clear delta and add ucfirst->titlecase
19:42
blets: db5a45a | (Herbert Breunung)++ | docs/appendix-a-index.txt:
sanitize and polish string methods
lichtkind i hope i dont insult tadzik
moritz moritz.faui2k3.org/tmp/rakudo-2012.04.1.tar.gz # release preview, testing welcome 19:43
lichtkind moritz++ 19:51
masak .oO( a star is born ) 19:52
lichtkind moritz was always cool, you mean he is hot now?
masak who are you talking to? :) 19:55
moritz is always away from equilibrium
diakopter too much Hawking radiation 19:56
masak my wakeup times are always at disequilibrium. :/
moritz turns up the equalizer a bit 19:58
tadzik lichtkind: heh, no, pay no mind :) 20:05
moritz ok, 2012.04.1 release tagged 20:14
masak didn't have time to run the spectests yet, but presumes everything's OK 20:15
PerlJam moritz: Who is doing the R* release?
masak more importantly, who is doing the May release of the Rakudo compiler? 20:16
there ain't no-one signed up for it yet. 20:17
tadzik I can do it
PerlJam tadzik++ 20:17
tadzik yay, karma
masak tadzik++
masak then I'll take July. 20:20
dalek kudo/nom: f61253e | masak++ | docs/release_guide.pod:
[docs/release_guide.pod] volunteered for July
20:21
Heuristic branch merge: pushed 51 commits to rakudo by masak
kudo/nom: 7ad3711 | masak++ | docs/release_guide.pod:
[docs/release_guide.pod] volunteered tadzik++ for May
masak I don't like that merge commit there :/
hm. I don't see a merge commit locally. 20:22
nor do I see one on github.
I conclude that dalek is charmingly misinformed.
dalek kudo/nom: efbc48c | duff++ | docs/release_guide.pod:
[docs/release_guide.pod] claim June release
20:23
moritz masak: where do you see a merge commit from dalek?
benabik dalek infers a merge from large numbers of commits.
PerlJam moritz: < dalek> Heuristic branch merge: pushed 51 commits to rakudo by masak
benabik "Heuristic branch merge" basically means "I got too many commits to display them all"
masak yeah.
that concerned me, because I hadn't merged anything locally.
as far as I can see, it's reality against dalek on this one. 20:24
moritz oh, dammit. I managed to forget to push rakudo's VERSION bump
PerlJam benabik: but why did dalek think there were many commits instead of just the 2 that there were?
actually only one at that point
dalek kudo/nom: 5a65af3 | moritz++ | docs/release_guide.pod:
add 2012.04 release to release_guide.pod
20:25
kudo/nom: 4ec5106 | moritz++ | VERSION:
[release] bump VERSION to 2012.04.1
benabik PerlJam: I think it only knows what github sends it. It must have received a resend for some reason.
dalek kudo/nom: 6bee9b6 | duff++ | docs/release_guide.pod:
[docs/release_guide.pod] claim birthday month release too
20:27
moritz 56
pmichaud the merge commit was because of a push to the master branch, I think. 20:29
PerlJam claims every Oct to gie a little birthday present to himself :)
s/gie/give/
pmichaud I got email about the merge commit -- it shows a bunch of commits from May 2011
PerlJam
.oO( I guess "gie" works too though :)
20:30
moritz ok, tag 2012.04.1 re-created, this time with correct VERSION
sorry for the inconvenience
pmichaud I thought we had dropped the master branch? 20:31
masak pmichaud: you're right.
pmichaud: I'll remove the master branch now.
there. removed locally and from github. 20:32
pmichaud++ masak--
pmichaud masak++
masak dalek++
dalek blets: 798b40c | (Herbert Breunung)++ | docs/appendix-a-index.txt:
added string method fc
21:07
blets: 1f26a05 | (Herbert Breunung)++ | docs/appendix-a-index.txt:
more crosslinks for string methods
blets: bb5aa7c | (Herbert Breunung)++ | docs/appendix-a-index.txt:
snippet for substr
[Coke] fudge now hanging on a niecza autounfudge run. 21:31
[Coke] should really track that down at some point.
sorear good * #perl6 21:48
dalek ar: e1fe704 | moritz++ | Makefile:
update version in preparation of the 2012.04 release
21:50
ar: 05d8730 | moritz++ | Makefile:
parrot 4.3.0 is a stable release
ar: 5237f98 | moritz++ | skel/tools/build/Makefile.in:
update version in Makefile.in
21:51
moritz phenny: tell pmichaud basing a star release on rakudo 2012.04.1 seems to give star the version 2012.04.1 too. Should we try to separate that? or just go with it? 21:53
phenny moritz: I'll pass that on when pmichaud is around.
moritz phenny: tell pmichaud never mind, doesn't seem hard 21:54
phenny moritz: I'll pass that on when pmichaud is around.
dalek ast: 7becae6 | coke++ | S0 (9 files):
niecza autounfudge
21:55
masak moritz: haven't there been similar situations before? what was the Star version number those times? 21:57
sorear o/ masak
masak \o 21:58
tadzik \o/ 21:58
masak 'night, #perl6 22:34
spider-mario good night, masak 22:42
sorear night masak.
o/ spider-mario
spider-mario (oh, I’m a little late. sorry) 22:43
\o sorear
dalek blets: 26c92d0 | (Herbert Breunung)++ | docs/appendix- (2 files):
minor fixes and crosslinks
22:54
dalek rl6-bench: ee8d57a | (Geoffrey Broadwell)++ | TODO:
Reorg TODO a bit, and add another feature TODO
22:59
dalek blets: 34aada7 | (Herbert Breunung)++ | docs/appendix-a-index.txt:
expand nav menu so that chapter fit a large screen (less searching)
23:27
dalek blets: 63e3ae0 | (Herbert Breunung)++ | README.md:
add rules for glossary anchors
23:39
blets: 3e5f0c4 | (Herbert Breunung)++ | docs/appendix- (2 files):
tiny fixes
sorear japhb: ping 23:50