»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.
Chat1791 anyone in here from Canada 01:44
sjohnson charsbar: yes 01:45
oops.. wrong person
diakopter hunh 02:46
sorear good * #perl6 02:48
sorear dwhipp: hi 03:17
sorear dwhipp_: Hi 03:49
dwhipp_ hi
sorear perl6/specs commit bits are handed out quite freely
sorear if you want a rakudo commit bit, you need to mail in a waiver 03:49
dwhipp_ That OK, I can use pull requests 03:50
I was just a little surprised by the asymmetry
sorear www.perlfoundation.org/attachment/legal/cla3.pdf # I think this is the one 03:51
dwhipp_ colomon: thx for submitting my change 03:52
sorear: thanks for the link 03:54
Probably should do the legal stuff even without a commit bit -- patches are patches, however they get in 03:56
sorear If you plan to do anything on the Parrot side of the fence, know that PaFo has an independant CLA system 03:57
dwhipp_ It would be nice to do these things electronically -- physical docs seem so archaic 03:59
sorear you can scan a filled-out CLA and email it to (I forget where) 04:05
sorear (someday, the world will figure out that digital signatures are more secure than scanned handwriting. Someday.) 04:06
cotto README says that ICU isn't a hard requirement for Rakudo. Is that still correct? 05:11
sorear cotto: It should be 05:55
someday once we have more plumagy stuff icu will probably be made a hard dependancy
but as long as enabling parrot-ICU remains prohibitively difficult on Win32, Rakudo has to cope with sometimes not having it 05:56
AIUI
cotto sorear, ok. thanks 05:58
tadzik morning 07:36
mberends Here is an excellent thread that came up while researching "smoke testing". It brought back so many memories, particularly the "intermittent windshield wiper" that featured in my 1961 Mini 850. www3.telus.net/bc_triumph_registry/smoke.htm 08:56
jdhore mberends, you drive a 1961 Mini? 09:02
mberends drove, from 1975 to 1978 09:03
jdhore ah 09:04
Anyone who drives a classic Mini today is very worthy of my respect. 09:05
A classic Mini is a BRILLIANT car
mberends we once famously transported 8 people in it for about 5 miles - I was in the "trunk".
it wasn't 100% original, the previous owner had modded it with a Weber carb that managed to reduce the fuel economy from 50 to 35 miles per gallon (imp). 09:08
jdhore wow 09:09
jdhore With that carb, did it have about the same performance as the Cooper? 09:09
Or just that classic Rally Mini sound?
mberends it was a hack, the performance came out worse istr. The 0-60 time was either 27 or 34 seconds. 09:10
jdhore ouch 09:11
jdhore Even my "modern" (late 80's) 4-cylinder Diesel pickup truck has better performance than that 09:12
mberends so does my mid 90's 5-cylinder Diesel minibus 09:15
jdhore hmm...mid-90's...5-cylinder...minibus...I'm guessing VW EuroVan? 09:19
moritz_ dwhipp: just looked into the backlog... rakudo commit bit policy is a bit stricter than that of most other repos :(. After a few good patches and sending in a signed Contributor's License Agreement you usually get a commit bit 09:21
dwhipp: until then you're stuck with pull requests / patches. Sorry about that. 09:22
mberends jdhore: yes, Caravelle Club 8. most serious omission: aircon :( 09:32
jdhore I like aircon in a car, but it's not a must-have
mberends from personal experience I can confirm the old adage that if you have a (minibus|pickup truck|trailer), you'll always have friends :) 09:35
jdhore heh 09:38
I do have the: "Yes I have A pickup truck, no I will not help you move" T-shirt
mberends ymmv ;-) 09:39
lol 09:41
pmurias sorear: ping 09:54
moritz_ in perl 5, my $obj; sub f { $obj->x } keeps a reference to $obj, because sub f can be called from anywhere (it's in the symbol table) 10:20
in perl 6, could $obj be garbage collected once &f goes out of scope (being lexical and all)?
mberends moritz_: that makes sense 10:39
jnthn o/ #perl6
moritz_ \o 10:40
tadzik o/ 10:42
imho sub f should hold a reference to $obj, so it won't get collected before f is 10:43
moritz_ sure
the question is if &f's pad is going to get collected at all 10:44
and I hope the answer is "yes"
mberends \o jnthn: for polpolnik, are you willing to upgrade the VirtualBox on pivo from 1.x to 4.x? It would simplify maintenance. That means going ex-Debian and installing VirtualBox from source.
*popolnik 10:46
(I'm making a repo of build scripts) 10:47
jnthn mberends: I'm very tight on time today/tomorrow - is it something that I need to help with? 10:48
mberends: It's not being used for anything else though. 10:49
So no problems with upgrading.]
moritz_ tadzik: did you experiment with a :vtable('destroy') override in Mu.pir recently? 10:55
mberends jnthn: no rush, this weekend is fine. Unless you're happy to put my login into /etc/sudoers ;) 10:56
mberends I suppose jnthn and masak have to finish their homework before they're allowed to go out and play 11:03
jnthn :P 11:07
$dayjob and talk slides and such fun :) 11:08
At least I got most of one of my talks done last night :)
pmurias sorear: ping 12:56
colomon almost has smash's benchmarking code working on his system. 13:05
colomon anyone know how to use DBI / SQLite? 13:24
moritz_ in Perl 6?
colomon no, in Perl 5.
takadonet morning all
colomon trying to get smash's benchmark stuff running.
and he uses a perl 5 harness
moritz_ colomon: I'd look into the MiniDB test suite if there's an example
colomon error I'm getting is
DBD::SQLite::db do failed: no such table: bench at run_bench.pl line 13.
DBD::SQLite::db prepare failed: no such table: bench at run_bench.pl line 14.
moritz_ ah, seems like you need to create a table first 13:25
colomon knows basically nothing of database handling in perl. 13:32
or anywhere else, for that matter.
moritz_ if there's no schema in the benchmark repo, you have to try to reverse-engineer the table layout from the queries 13:33
takadonet ouch
colomon it looks like it should be very simple
but then, I don't understand why he didn't just dump the results in a text file. 13:34
moritz_ in particular the code for inserting things into the db should contain all rows (except mabe autoincremeting IDs)
colomon my $sth = $dbh->prepare("INSERT INTO bench VALUES (?,?,?,?,?)");
moritz_ urks.
colomon moritz_: if I get you a simple comma-separate file of benchmark results, can you use your amazing powers of SVG to make graphs? 13:37
takadonet colomon: Smash not around to help u? 13:38
moritz_ colomon: SVG::Plot can (it needs arrays of arrays, but it's an easy transformation; there are scripts that do something similar in the examples/ dir)
mberends colomon: I know very well how to write such code, but I don't want to spend time on your instance because it is covered by what I plan to do this weekend with popolnik.
colomon does not have smash's graph code, just the benchmark running code
moritz_ it's the old "I want to polish this code before I publish it" thing 13:39
moritz_ usually prefers unpolished code to unpublished code
colomon takadonet: he's presumably busy, else he'd be running these benchmarks himself. 13:40
poplnik? 13:41
jnthn
.oO( Popolnik has something in common with Zavolaj and Blizkost. )
 13:42
moritz_ a jnthnproject?
mberends popolnik: an ashtray (Slovak): a smoker results consolidation database
colomon ooo, smoker 13:43
[Coke] is rakudo's config system smart enough to bail out it parrot available is too old?
(after switch to git?)
moritz_ mberends: is that your agenda for the hackathon?
jnthn moritz_: Language the name is taken from. :)
moritz_ [Coke]: yes
mberends moritz_: yes, one of the items, under the general umbrella of "ecosystem". 13:44
moritz_ [Coke]: the code is a bit convoluted, but stealable
[Coke] moritz_: ... Was just going to comment on the convoluted bit. :P 13:45
moritz_ [Coke]: I planned to do a single function that reads PARROT_REVISION, and returns the current revision (if any), the required, the release version (if any) and the information if it's new enough 13:46
[Coke]: but after the decision to target nqp (which then targets parrot) I kinda lost motivation for doing so 13:47
though now i realize that if it's not rakudo that needs that step, nqp will
jnthn :) 13:48
moritz_ it's just always a PITA to test 13:51
moritz_ hates fragile systems without unit tests 13:53
colomon Life is a fragile system without unit tests. ;) 13:57
moritz_ life has very rigid unit tests 13:58
moritz_ called "survival" 13:58
and if you fail, you die
every commit to your DNA is check in post-commit hooks 13:59
moritz_ that why we still live 14:01
[Coke] O_o 14:03
mberends "life's a test, and then you die" 14:09
dalek kudo: 9bca538 | moritz++ | build/lib/Parrot/CompareRevisions.pm:
[build] make parse_parrot_revision_file a bit more reusable
 14:10
colomon phenny: tell masak Your svg module is still using done_testing. 14:31
phenny colomon: I'll pass that on when masak is around.
moritz_ colomon: it does not :-) 14:33
colomon finds it more than a little bit sad that he still has to install modules by hand, and fix things as he goes.
ooo, moritz_'s svg module.
I thought masak wrote it? 14:34
colomon feels very confused
moritz_ we renamed masak's SVG module to XML::Writer or so
colomon phenny: tell masak Never mind.
phenny colomon: I'll pass that on when masak is around.
moritz_ and I've create a new SVG that is actually SVG specific (masak's wasn't) 14:34
colomon moritz_: yes, I installed that one first. :)
btw, I have a massive benchmarking run going on my 64-bit Linux system. 14:35
moritz_ now I've updated SVG::Plot too
colomon and just before I could type "make test" in it! moritz_++ 14:36
colomon btw, what I find sad isn't that we haven't fixed everything yet, but that our ecosystem is not well-enough used or automatically tested to detect build-breaking changes until months after they happen. 14:37
moritz_ agreed 14:38
colomon okay, so I've got a comma-separated data file, and svg-plot installed. 14:39
moritz_ colomon: look at examples/parse-spectest-progress.pl 14:41
colomon just what I had open. :)
moritz_ it uses Text::CSV::Simple to turn things into a graph
but you might just was well split on ',' or so
colomon it is a very simple file format, after all. :) 14:42
moritz_ though you might want to use :lines instead of :stacked-bars
colomon errr... Text::CSV::Simple isn't in the ecosystem? 14:43
colomon is trying to run the example. 14:44
moritz_ colomon: might be in the Text::CSV repo 14:44
colomon bah, I'm just going to jump in and start coding. 14:45
moritz_ +1 :-) 14:45
fwiw the tests from the SVG::Plot test suite are woefully incomplete 14:46
mostly because I don't have any clue how I should test it
colomon moritz_: believe me, I understand how that works. 14:46
colomon afk # toddler dressing time 14:49
ruoso rakudo: say "hello from android irc client" 14:54
p6eval rakudo 9bca53: OUTPUT«hello from android irc client␤»
moritz_ niecza: say 'hello from another compiler :-)' 14:55
p6eval niecza v3-28-ge477540: OUTPUT«hello from another compiler :-)␤» 14:55
moritz_ ruoso: how are you?
flussence_ ouch, just got bitten by Javascript treating null and undefined as separate values :( 14:56
PerlJam flussence_: but ... they are separate values :) 15:03
moritz_ is all for having null, nil, undef, undefined and Mu as separate values in the language 15:11
flussence_ don't forget NaN :) 15:13
moritz_ of course, NaN and -NaN
PerlJam moritz_: I'd be curious about the distinction you see between undef and undefined 15:13
moritz_: other than that, I agree with you entirely :) 15:14
moritz_ PerlJam: undef is from Perl 5, undefined from ECMAscript :-)
moritz_ forgot Nil
cotto I'm running spectest_regression against the mmd-op-cache branch of parrot. When I run make spectest_regression I see a failure in test 13 of t/spec/S02-builtin_data_types/instants-and-durations.t, but when I run that test alone it works. 15:16
flussence_
.oO( why isn't the opposite of "defined" called "fined"? )
moritz_ cotto: ignore that one
cotto moritz_, will do. Thanks.
moritz_ cotto: it's a time dependent floating point inaccuracy in some p6 level code
cotto is it known-wonky?
sounds exciting 15:17
moritz_ well, the 'time dependent' really comes from the testing
because it always uses 'now' as one of the reference values
and it's triggered in about one of 10 cases or so 15:18
but I haven't got around to investigate it in detail
and I don't want to skip it because we don't have many more time interval arithmetic tests 15:20
colomon std: gist.github.com/852937 15:29
p6eval std 4608239: OUTPUT«===SORRY!===␤Unrecognized regex metacharacter , (must be quoted to match literally) at /tmp/1joZnaM28B line 3:␤------> $benchmark, $time) = $line.split(m/\s* ,⏏ \s*/);␤Couldn't find terminator / at /tmp/1joZnaM28B line 3:␤------>
..$benchmark…
colomon ah
std++
moritz_ knows why he wrote "split on ','" :-) 15:30
rakudo: /,/
p6eval rakudo 9bca53: OUTPUT«===SORRY!===␤Confused at line 22, near "/,/"␤»
colomon std: gist.github.com/852937
p6eval std 4608239: OUTPUT«===SORRY!===␤Variable $tie is not predeclared at /tmp/1BDCKpwEzr line 5:␤------> chmarks{$benchmark}.push($rakudo => $tie⏏);␤Function 'lines' needs parens to avoid gobbling block at /tmp/1BDCKpwEzr line 2:␤------> for lines⏏ ->
..$l…
colomon apparently needs to get std working locally again 15:31
std: gist.github.com/852937
p6eval std 4608239: OUTPUT«===SORRY!===␤Function 'lines' needs parens to avoid gobbling block at /tmp/pOBeupqHRm line 2:␤------> for lines⏏ -> $line {␤Missing block (apparently gobbled by 'lines') at /tmp/pOBeupqHRm line 8:␤------> <BOL>⏏for
..%benchmarks.key…
moritz_ for lines() -> $line
colomon std: gist.github.com/852937 15:33
p6eval std 4608239: OUTPUT«ok 00:01 127m␤»
colomon ouch!
just realized the trans-sprintf benchmark has actually gotten slower, not faster. :( 15:34
jnthn :/ 15:37
colomon more on that once I've got the graphs displaying. 15:39
domidumont Hello. Is anyone here involved with packaging parrot or rakudo for Debian ? (I'm willing to help even if I'm don't have much tuit) 15:49
moritz_ domidumont: parrot people have been working on that, occasionally 15:50
over in #parrot on irc.perl.org
domidumont: there's a ports/debian directory in parrot which should contain all the necessary files, albeit a bit outdated
domidumont: also don't miss docs/project/debian_packaging_guide.pod in the parrot repo 15:51
domidumont moritz_: ah, was this used to produce the current (outdated) package in Debian ?
colomon moritz_: can you take a look at gist.github.com/852937 and see if there is anything obvious I'm doing wrong? the result graph is pretty useless...
moritz_ domidumont: I think so
domidumont moritz_: ok. Thanks. I'll ask on #parrot channel for more information
colomon moritz_: resulting SVG is gist.github.com/852974 15:52
moritz_ colomon: values expects an array of arrays - are you passing that along?
colomon why an array of arrays?
or, what is the array of arrays?
moritz_ wait, I was still thinking of stacked bars 15:53
let me re-check
(array of arrays because each data series should have one y value per x value)
colomon what should the arrays look like, then? 15:56
moritz_ colomon: you'll need :values([@data.keys], ) or :values([[@data.keys]]) I think
colomon I thought I was passing a Y value to go with each label.
moritz_ that's what i would expect if it plotted only one data series per graph, ever
colomon with :values([[@data.keys]]) I get a Divide by zero error from SVG::Plot. 15:57
moritz_ it expects ([y1-for-x1, y1-for-x2, ...], [y2-for-x1, y2-for-x2, ...], ...)
colomon: I can't really investigate now, but will later on if you also provide a data file 15:59
colomon I'll poke around at it a bit more, and then send you the data file if I don't make any progress. 16:00
!!! 16:02
colomon ah. 16:02
:values([@data>>.value]), not >>.values
duh.
moritz_ oh.
colomon bitten by p6 16:03
's urge to do something.
hmmm... would just plain .values work there? 16:05
nope.
moritz_ .values on an array returns the array (or a copy thereof) 16:07
colomon right 16:08
that makes sense
is it possible to put a label on the entire graph? 16:09
and how do I embed the SVG in an HTML file? ;)
moritz_ that would be a title
<img src="path/to/graph.svg" /> 16:10
moritz_ there's also "real" embedding, but it requires xhtml, not html 16:10
colomon gotcha.
moritz_ hopes colomon doesn't want IE support 16:11
colomon I need to get this stuff bundled up into a package and in the ecosystem, so people who are better at the graphics bits can hack on it.
colomon can't inkscape convert SVG to PNG or something? 16:12
moritz_ it can
colomon what should I use as the package name for this? 16:13
moritz_ inkscape --export-png=target.png -D --file=input.svg
Benchmark::Chart or so?
or Benchmark::Plot # to keep consistent with SVG::Plot 16:14
colomon I was actually thinking of tossing smash's benchmark harness in there too.
Benchmark::Plot would be simple and straightforward, though. hmmm.
Layla_91 Howdy world o/ 16:27
jnthn Layla_91! \o/
Layla_91 jnthn: =)
jnthn Layla_91: Hope you're less feverish than last time I saw you here? :)
Layla_91 jnthn: yup.. now am so fresh and ready for some work B-) 16:28
jnthn: How's perl6 doing ? :D
jnthn Layla_91: yay! Glad you're better 16:29
Layla_91: Perl 6 is doing OK. Though I ddin't have enough hacking time recently. :( 16:30
Layla_91: Though on the up side, I'm writing some new slides. :)
Layla_91 jnthn: yay! :D I am so hungry for knowledge! :D I spent 3 days doing nothing in the hospital.. I hugged my laptop when came home! :D 16:31
jnthn Aww
Hospital sounds un-fun. :( At least you've got your laptop back now. :) 16:32
Layla_91 jnthn: yes :) but I have tons of pending work,,, as I Figuring out from the e-mails... :\ 16:33
jnthn: And they are moving me to a PHP team... (they are taking away my Mac :'''( ) 16:34
jnthn Oh noes!!! :(
Layla_91 jnthn: its ok I will buy one.. someday =) 16:35
jnthn: plus php is easier to me :) 16:36
jnthn :) 16:37
Layla_91 jnthn: can you give me the link to your slides again? can't find my bookmarks :S 16:39
Layla_91 jnthn: Somebody must put them on perl6 site... seriously :\ 16:40
moritz_ Layla_91: how about you do it? :-)
Layla_91 moritz_: mmm.. Can I put my name and picture there too? :P 16:41
jnthn layls: www.jnthn.net/articles.shtml
Do we not have links to presentations and stuff on perl6.org already?
Maybe not. :)
Good idea though. 16:42
www.perl6.org/documentation/ could be a good place to link to a page of some good Perl 6 presentations to see. 16:43
Layla_91 www.jnthn.net seems cool :-) Like an entire library! :) 16:44
Layla_91 moritz_: can I really add content to the perl6 website? 16:49
jnthn Layla_91: pivo.jnthn.net/ is cool (and very slow loading) 'cus it's written in Perl 6. :)
Layla_91 jnthn: omg looool! ^_^ 16:49
Layla_91 * Taking a dive into the slides..... back sooon......... 16:49
Layla_91 jnthn: Russia not mentioned in photos :( 16:55
donri In Soviet Russia, photos mention you!
Layla_91 donri: what do you mean? 16:57
jnthn Layla_91: Plenty of Russia photos from travel.jnthn.net/ :-) 16:58
donri I mean only the meanest of things.
jnthn Layla_91: Sure I've used a Russia photo in some talk :) 16:58
jnthn If not, I promise to soon :P 16:59
sjn o/
seen masak 17:00
aloha masak was last seen in #perl6 20 hours 48 mins ago saying "I can't believe we didn't have that one already. tadzik++".
sjn aloha: thanks! ^^
Layla_91 jnthn: jnthn++ :)
colomon tried to update his local git, and now has completely broken it. :( 17:01
moritz_ colomon: did you have any success with generating SVGs? 17:27
colomon moritz_: yes 17:28
moritz_ \o/
colomon but I've had very little success getting my script pushed to github. 17:28
colomon currently refreshing my entire macports ecosystem in hopes of getting a working and reasonably-up-to-date version of git. 17:29
dalek kudo: 09461fc | moritz++ | / (3 files):
[build] Perl 5 cleanup

  * move read_config file to module
  * remove parrot_ name parts		 17:41
kudo: 191305f | moritz++ | build/PARROT_REVISION:
bump PARROT_REVISON for testing
kudo: 4e55077 | moritz++ | Configure.pl:
[build] if --gen-parrot-prefix is present, ignore all other possibly available parrot_config executables
masak good evening, dear zebras. 17:53
tonight I'm committed -- not to mention dedicated -- to bring you a p5 review post. 17:54
masak just because I like you so much, #perl6. :) 17:54
phenny masak: 14:31Z <colomon> tell masak Your svg module is still using done_testing.
masak: 14:34Z <colomon> tell masak Never mind.
moritz_ \o/ 17:55
jnthn o/ masak 17:58
moritz_ wow, Configure.pl and gen_parrot.pl use different mechanisms for reading config information from parrot 18:09
jdhore moritz_, BTW, i don't know if you saw, but i believe bacek said that gms would probably not be the default GC with the parrot 3.2.0 release and that perhaps Rakudo should default to it in configure..... 18:16
tadzik heh, I'll have to re-revert my commit :)
PerlJam gms?
PerlJam oh 18:17
generational mark sweep
bacek++
[Coke] jdhore: correct. if rakudo wants that, they should continue to specify it for --gen-parrot 18:17
I believe current concensus is can't be default until after 3.3 (but by all means, use it now.) 18:18
PerlJam [Coke]: because of policy?
[Coke] it's not a drop in replacement - you have to write more code in your project to safely use it. 18:19
rakudo already has that code, but joe user doesn't. 18:20
PerlJam gotcha
Layla_91_ jnthn: are you still here? 18:29
moritz_ Layla_91_: (re perl6.org website) yes.
masak Layla_91_: I think jnthn went for a walk. 18:29
Layla_91_ masak: hi! =) 18:30
masak Layla_91_: \o :)
Layla_91_ moritz_: how?
moritz_ Layla_91_: first step: tell me your github ID 18:31
masak (zeroth step: get a github account) 18:32
moritz_ Layla_91_: then clone the perl6/perl6.org repo, apply and push your changes, wait.
it's updated from source every 15 min or so
PerlJam gotcha 18:33
Layla_91_ masak: (the real teacher.. :D ), moritz_: I will need to create an account, give me few minutes..
PerlJam oops 18:33
masak PerlJam: in Soviet Russia, I've got YOU. 18:34
Layla_91_ masak: what is everybody talking about soviet russia? I can not understand the jokes.. :S this is second time for today.. 18:34
PerlJam Layla_91_: In soviet russia the jokes understand you. 18:35
moritz_ Layla_91_: uncyclopedia.wikia.com/wiki/Russian...%28joke%29
masak Layla_91_: en.wikipedia.org/wiki/Slashdot#Culture 18:36
also, autopun spotting on what PerlJam just said :)
Layla_91_ kinda did not like the essence a lot.. Russia is both oppressive and bad :( I think america is the bad guys :P 18:37
But sounds funny ^_^ 18:38
moritz_ Layla_91_: that doesn't matter for jokes :-)
PerlJam Layla_91_: But we have Chuck Norris!
Layla_91_ PerlJam: ah.. we have Yeltsin xD 18:39
masak Yeltsin is kinda badass.
masak there's probably some law of Nature that prevents him from ever ending up in a fight with Chuck Norris. 18:40
Layla_91_ masak: hehehehe :D 18:41
Layla_91_ Oh totally forgot! github! :P 18:41
moritz_ masak: repulsion of singularities! 18:42
masak moritz_: that's probably it. if there were two singularities in the same place, it'd be a duality. :P
jnthn Layla_91_: fear not, I'm back :P 18:50
Layla_91_ jnthn: \o/ 18:50
jnthn: creating a github account based on the Mr.moritz_ request :P I needed to ask you something but maybe later today :) 18:51
jnthn Layla_91_: yay...GitHub account = can hack on lots of Perl 6 projects \o/ 18:55
Layla_91_ moritz_: layla-91 =) 18:56
moritz_ Layla_91_: added. Welcome! 18:58
moritz_ oh I forgot, many people in the channel can add you via hugme 18:58
Layla_91_ moritz_: thanks =) 18:59
moritz_ ...except that hugme seems to have gone pinin' for the fjords...
Layla_91_ moritz_: now I understand nothing O_o .. pinin huh?!
moritz_ Layla_91_: "pining for the fjords" is like "dying" 19:00
Layla_91_ moritz_: undersood.. :)
Layla_91_ Must have a nice avatar first ^_^ 19:01
moritz_ Layla_91_: www.youtube.com/watch?v=4vuW6tQ0218 prominently featured here
Layla_91_ moritz_: my dad used to watch this show :) 19:03
I like British Accent :) must get monty python DVDs :) 19:06
jnthn There's quite a few ways to have a British accent. :) 19:07
slavik1 Layla_91_: one of the 5 people was american
Layla_91_ slavik1: I am not that expert (since I am not american nor british) but I guess it is the british ones who speak more funny :) 19:08
slavik1 true
Layla_91_ jnthn: like what? I want to speak BRITISH! :D 19:10
moritz_ Layla_91_: living in britain helps a lot, my deary 19:11
masak or looking at British sitcoms. 19:12
Layla_91_ moritz_: I wish I could.. :'(
masak: you are great today! ^_^ 19:13
jnthn Layla_91_: I have some cousins who speak "British" like me, but I find their accent decidedly hard to follow. Britsh accents can vary quite a lot. And lots of regional slang too, just for fun. :)
Layla_91_ jnthn: We had a customer from Ireland 2 months ago.. I think his accent was too different from british one.. but funnier! =) 19:15
jnthn Yes, they're fun, but often very nice to listen to. :)
moritz_ a thick Irish accent is *very* hard to understand 19:16
I had a flat made in Scotland... quite impossible to understand
i usually gave up three "excuse me?" iterations :-)
Layla_91_ jnthn, moritz_: he kept making his voice louder and louder.. I guess he thought I am deaf not that I could not understand his accent :S 19:17
Layla_91_ ah... I have to create a wordpress or Gravatar.com account to have an avatar.. :\ 19:21
oh god I promissed someone to setup a joomla server for him and completely forgot.. !
leaving!
o/
jnthn Layla_91_: o/ 19:23
masak hm. getting a strange error involving infix:<~> 20:07
moritz_ die "strange error involving infix:<~>" ? 20:08
masak no, a multi dispatch failure.
if only I knew what the types of the operands were... :) 20:09
jnthn :/
masak someone should totally patch that error to say "you passed in ..."
jnthn That'd be...odd. Since ~ works on (Any,Any)
colomon is still trying to get git working again on his system. (Though there was a long nom break in there.) 20:10
moritz_ rakudo: say infix:<~>(Mu, 1)
p6eval rakudo 191305: OUTPUT«No applicable candidates found to dispatch to for 'infix:<~>'. Available candidates are:␤:(Any $x = { ... })␤:(Any $a, Any $b)␤␤ in main program body at line 22:/tmp/M4PLaGZ2b2␤»
masak I'm currently updating and rebuilting Parrot/Rakudo. see if the problem persists.
moritz_: that's the error.
moritz_ so, a Mu involved?
masak but I was doing something totally innocent with a .substr and a literal string.
masak don't see where the Mu would come from. 20:10
moritz_ I remember substr out of range returning Mu or so 20:11
masak that's just... wrong.
moritz_ rakudo: say 'abc'.substr(5).perl
p6eval rakudo 191305: OUTPUT«Mu␤»
masak is there an RT ticket for that?
moritz_ I think so
masak moritz_++
that must be it, then.
jnthn What should it do? 20:12
moritz_ fail('substr out of range')
jnthn That'd make sense.
colomon rakudo: say infix:<~>(fail('substr out of range'), 1)
p6eval rakudo 191305: ( no output )
jnthn masak: You should harrass somebody at a hackathon to help get the arg listing on errors implemented.
masak what does &fail return in a typed subroutine?
jnthn: oh, good idea.
jnthn masak: iirc, Failure is somehow always magically acceptable. Or something. 20:13
Otherwise failing wouold cause a type check error :/
masak right. 20:14
so Failure is a bottom type, basically.
"PARROT VM: Could not load bytecodeNULL current PMC at 2 in thaw" 20:16
we're back to sucky error messages when Rakudo and Parrot versions are out-of-sync. :(
I'd like to see something like "You tried to use this HLL compiled for Parrot x.y.z, but this is Parrot i.j.k" 20:17
colomon seems to have git version 1.7.4.1 installed on his MBP again! 20:18
errr.... first time for that version, actually.
colomon as a bonus, it looks like I've managed to make perl 5.12 my default! \o/ 20:20
colomon wonders how much stuff he broke in the process....
moritz_: are you fixing the substr returns Mu bug? it certainly appears to be LHF... 20:21
moritz_ colomon: I fear that the might might make things worse, considering how crappy the failure implementation is 20:23
colomon but will you let yourself be governed by your fears? ;)
moritz_ yes
masak is reminded of the lyrics of Deteriorata: "... Give up." 20:26
ooh, and Deteriorata is on Spotify! \o/
[Coke]: ping 20:27
colomon moritz_: github.com/colomon/Benchmark-Plot 20:41
now in ecosystem, too. 20:47
masak colomon++
dalek osystem: 49ae828 | (Solomon Foster)++ | projects.list:
Add Benchmark::Plot.
 20:48
colomon It's graphs are pretty crude, but it's a start.
*Its
masak rakudo: my $time-difference = now - now; say $time-difference.fmt("%s") 20:49
p6eval rakudo 191305: OUTPUT«Method 'fmt' not found for invocant of class 'Duration'␤ in main program body at line 22:/tmp/pDKYRoRLvu␤»
masak yup. still hating Duration.
can we please kill it now?
it's basically a cumbersome Num.
colomon Regardless of whether it should be killed (I've no dog in that fight), shouldn't that work anyway? 20:52
masak depends if it's Cool, I guess.
colomon but it should be, shouldn't it?
masak I don't think the spec has an opinion about that. 20:53
colomon I thought almost all builtin types were supposed to be.
masak I'm not sure where that point of view comes from :)
in my mind, only a select few builtin classes derive from Cool.
they're common, yes. but they're not that many. 20:54
jnthn They're common in core but uncommon in userspace.
masak I wouldn't call Str and Int uncommon in userspace...
jnthn I meant declarationally.
masak oh, agreed. 20:55
but I could definitely see someone inheriting from Cool in userspace, too.
colomon Wow, the spec is actually completely silent on the question. 20:59
tadzik ~~ 21:01
masak tadzik! \o/
tadzik: you know what? this time tomorrow, we'll have met. :)
tadzik yay! I'm looking forward :) 21:02
I need to pack my things and ponies
masak and zebras.
I have a whole compartment in my bag just for the zebras.
tadzik oh my, my wannabe-ISP is hilarous 21:03
have I told the story of the ISP who isn't able to connect the cables for 2 monts? 21:04
flussence_ pff, I spent 6 months with my connection dropping every 2-3 minutes because mine wouldn't admit the problem was on their end 21:05
tadzik more than a month after breaking the deal (accepted by both sides) I get a bill for .85 PLN (~20 euro cents)
tadzik to care or not to care... 21:05
dalek tpfwiki: (chris denorch)++ | www.perlfoundation.org/perl6/index.cgi?perl_6 21:06
masak I'm benchmarking the p5 solutions now.
colomon \o/ 21:10
masak colomon: interesting. fox's solution is the naivest. it beats yours for the small string lengths, but your solution gets back at it for the large ones.
colomon That's actually a pretty common thing with algorithms, no? 21:11
masak :P
well, after having just run yours I figgered that wasn't what I was going to see. 21:12
gist.github.com/853570
colomon seems to recall that bubble sort is the best sort for arrays of length 4...
tadzik oh wow
colomon has no idea what that gist means. 21:13
masak looks like fox's solution is to be preferred up to string lengths of about 50 chars.
colomon: oh, sorry. first column is length, of both strings.
second column is whether there is a common substring at all.
(I try both without and with a common substring for all lengths) 21:14
colomon huh. I'm surprised mine is so much slower when there is a common substring
masak third column is run time in seconds.
colomon in the 320 case.
masak colomon: that might just be GC, mind.
I should run it once with the Y/N reversed.
colomon I'll bet I'm doing something grotesquely inefficient, actually.
masak or that :P 21:15
yeah, the difference is a bit too large to be ascribed to just GC, actually.
but I'll still do the reverse test, I think.
masak chromatic++ # www.modernperlbooks.com/mt/2011/03/...moose.html 21:18
colomon I'm not sure exactly what, but I'm betting that's it. Once I had the SuffixTree built, I didn't really look too closely at the efficiency of building substrings from it. 21:19
masak fwiw, fox's solution exhibits the same increase even for the same lengths. 21:21
colomon huh. I guess I should actually time the bits before jumping to conclusions. 21:25
That is to say, I don't know the effect duplicate substrings has on SuffixTree construction time. 21:28
masak oh yes. suffix trees definitely beat the naive algorithm. 21:29
I'm relieved to see that.
moritz_ \o/ my first contribution to a scientific conference has been accepted 21:36
colomon moritz_++
masak yay. moritz_++ 21:37
jnthn moritz_++
tadzik great!
congratulations!
moritz_ thanks 21:38
masak argh. these benchmarks are taking too long. won't be able to finish today. 21:41
oh well. at least now I have an approach.
tadzik masak: have you checked the weather maybe? 21:43
masak tadzik: no...? 21:44
tadzik just asking
colomon masak: how are you getting your two strings to compare?
masak colomon: hold on, I'll nopaste you the data-generation script.
colomon danke
masak gist.github.com/853570 21:46
colomon perfect 21:47
hmmm... you realize this may not be the most realistic test. 21:48
?
masak sure. 21:51
I'm mostly interested in the general trend.
but, hm. seems I'm hitting memory/swapping considerations at the upper end. :/
colomon ah, the joys of Rakudo!
masak :/ 21:52
jnthn Rakud'oh!
moritz_ :-)
jnthn is working on his talk :)
masak interesting that the naive solution had swapping issues, but colomon's with lots of objects didn't, not as much at least. 21:54
that suggests that reducing runtime of medium-long programs could be an end in itself :)
colomon how is the naive solution implemented?
I mean, broadly speaking.
masak out loop over chars of one string. inner loop over chars of other. 21:55
masak from those two positions, check how many from that point forward are shared. 21:55
colomon actually using the chars, or some sort of sub-strings?
masak compare with maximum so far.
jnthn That...should be memory cheap...
masak it should.
jnthn wtf. 21:56
colomon we may be finding rakudo glitches here, too...
jnthn wonders if we're accidentally keeping things around when iterating that we shouldn't be.
colomon oooo, yeah, if the inner loop's iterator hung around, that would be beyond ugly....
masak fox's statistics online: gist.github.com/853570 21:58
I'll stop there for tonight. will probably have some time to continue tomorrow.
jnthn colomon: A quick test produces no evidence for that, but I agree, it would be. 21:59
Will have to look more closely.
dalek p/lexical-setting: a32c5f9 | jonathan++ | src/NQP/Compiler.pm:
Make sure if we try to use a setting twice it'll work out.
 22:00
masak 'night, #perl6. 22:01
colomon o\
tadzik see you tomorrow :)
jnthn Yay, Holland Hackathon \o/ 22:02
mberends \o/
jnthn Are we all meeting up at the airport?
tadzik I'm just trying to figure that out :) 22:03
jnthn tadzik: What time you land?
tadzik jnthn: the ticket says 18:55 22:04
jnthn Ah, OK
masak and I get in at 21:00
tadzik you'll have a Greeting Comitee then :)
jnthn \o/ 22:05
colomon is kind of jealous. :)
tadzik is kind of excited
slavik1 wait, so you are all going to code while getting high? 22:11
I am diff in the wrong job
jnthn Yes. We're going to smoke LOADS of weed and then, like, commit man! 22:12
PerlJam jnthn: you're getting married?!? 22:13
jnthn ...weed is a marriage ritual?
PerlJam jnthn: no, but committing is.
colomon can't say much of anything after several cider-fueled hacking runs this winter.
jnthn Thanks to Rakudo, I've commited hundreds of times. :P 22:14
[Coke] wonder what masak wanted. 22:16
colomon rakudo: say time 22:23
p6eval rakudo 191305: OUTPUT«1299191029␤»
colomon rakudo: say time
p6eval rakudo 191305: OUTPUT«1299191038␤»
colomon phenny: tell masak My timings runs using your data-generation script are showing very significant differences from one test run to another. For 80, for instance, I've gotten average (N, Y) runs of (10.8, 14.2), (9.4, 10.8), and (9.6, 11.4). Maybe difference has to do with different random strings? 22:41
phenny colomon: I'll pass that on when masak is around.
sjohnson rakudo: say 3 22:52
p6eval rakudo 191305: OUTPUT«3␤»
sjohnson rakudo: say (undef // 4)."pig"
p6eval rakudo 191305: OUTPUT«===SORRY!===␤Unsupported use of undef as a value; in Perl 6 please use something more specific:␤ Mu (the "most undefined" type object),␤ an undefined type object such as Int,␤ Nil as an empty list,␤ !*.defined as a matcher or method,␤ Any:U as a type constraint␤ or
..fail() as a fail…
sjohnson rakudo: say (undef // 4)~"pig"
p6eval rakudo 191305: OUTPUT«===SORRY!===␤Unsupported use of undef as a value; in Perl 6 please use something more specific:␤ Mu (the "most undefined" type object),␤ an undefined type object such as Int,␤ Nil as an empty list,␤ !*.defined as a matcher or method,␤ Any:U as a type constraint␤ or
..fail() as a fail…
sjohnson defined-or in perl6? :3 22:53
tadzik please read the error message :)
sjohnson rakudo: say (nil // 4)~"pig" 22:56
p6eval rakudo 191305: OUTPUT«===SORRY!===␤Null regex not allowed at line 22, near " 4)~\"pig\""␤»
sjohnson argh, i've forgotten how to concat strings in the perl6 world. i'm not long for this [perl6] world
tadzik owait 22:57
rakudo: say (Nil // 4) ~ "pig"
p6eval rakudo 191305: OUTPUT«4pig␤»
tadzik rakudo: say (nil // 4) ~ "pig"
p6eval rakudo 191305: OUTPUT«===SORRY!===␤Null regex not allowed at line 22, near " 4) ~ \"pig"␤»
tadzik std: say (nil // 4) ~ "pig" 22:58
p6eval std 4608239: OUTPUT«===SORRY!===␤Null regex not allowed at /tmp/3ti6xoF7Gw line 1:␤------> say (nil //⏏ 4) ~ "pig"␤Two terms in a row at /tmp/3ti6xoF7Gw line 1:␤------> say (nil // ⏏4) ~ "pig"␤ expecting any of:␤ bracketed infix␤
..infix or meta…
tadzik std: nil
p6eval std 4608239: OUTPUT«===SORRY!===␤Undeclared routine:␤ 'nil' used at line 1␤Check failed␤FAILED 00:01 117m␤»
tadzik I wonder why does it parse as a regex 22:58
jnthn tadzik: nil is parsed as a list op 22:59
tadzik: That means we're expecting a term after it, not an operator.
tadzik jnthn: so isn't nil an undeclared routine? 23:00
jnthn tadzik: yes
tadzik: any time you have "x y" and x isn't a known term, it treats it as a call to x and parses y as a term. 23:01
And we call x a list op in this case, to make ourselves feel it's not a violation of TTIAR. ;-)
tadzik :) 23:03
sjohnson i broketed it. 23:09

Powered by Cro and the Raku® Programming Language.

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

Thank you!