»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg camelia perl6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by sorear on 25 June 2013.
dalek kudo/nom: 0358bae | jonathan++ | src/vm/jvm/core/Threading.pm:
Clean-up and re-working of Promise.

The outside API remains largely the same, except the constructor is now only taking :$scheduler. The main change is that Promise is no longer tied to running code at all. Rather, that's just what the run factory method and then methods set up. Promise.sleep(...) is also mostly delegated to the Scheduler. This makes Promise more focused around its synchronization primitive nature, and puts the other Promise factories on a more even footing.
00:02
jnthn ...and also means you can create a TestScheduler to test your Promise.sleep-ing stuff. 00:03
japhb__ r: my $buf := buf8.new(); sub set_first(buf8 $buf, Int $offset, int $value) { nqp::bindpos_i($buf, $offset, $value) }; set_first($buf, 0, 42); $buf[0];
camelia rakudo-jvm 882e33: OUTPUT«(timeout)» 00:04
( no output )
japhb__ r: my $buf := buf8.new(); sub set_first(buf8 $buf, Int $offset, int $value) { nqp::bindpos_i($buf, $offset, $value) }; set_first($buf, 0, 42); say $buf[0];
camelia rakudo-jvm 882e33: OUTPUT«(timeout)»
..rakudo-parrot 91a1b0: OUTPUT«42␤»
japhb__ r: my $buf := buf8.new(); sub set_first(buf8 $buf, Int $offset is rw, int $value) { nqp::bindpos_i($buf, $offset, $value) }; set_first($buf, my $ = 0, 42); say $buf[0]; 00:07
camelia rakudo-jvm 882e33: OUTPUT«(timeout)»
..rakudo-parrot 91a1b0: OUTPUT«42␤»
japhb__ r-j: say "alive"
00:07 dansamo joined
camelia rakudo-jvm 882e33: OUTPUT«(timeout)» 00:07
00:09 dmol left
timotimo as a response to a recent conversation in this here channel, i bought some greek-style jogurt 00:09
with honey :)
colomon yum 00:12
japhb__ wants some of that now-ish, actually, but it's too close to dinner time. 00:14
00:17 stevan_ left, stevan_ joined
BenGoldberg Does the evalbot start a new jvm process for each request? 00:19
Or is there a rakudo-jvm server running, which the evalbot gives requests to? 00:20
japhb__ is finding with annoyance that inserting "say '';" before a line makes it stop working. 00:23
... on jvm.
00:29 dansamo left
diakopter r-j: say "\n"; say 'alive' 00:29
camelia rakudo-jvm 882e33: OUTPUT«(timeout)»
diakopter wat
BenGoldberg r-j: say 'alive' 00:32
camelia rakudo-jvm 882e33: OUTPUT«(timeout)»
00:33 benabik left
jnthn sleep & 00:39
dalek -pb: 63fc1c5 | (Geoffrey Broadwell)++ | lib/PB/Binary/Reader.pm:
Minor cleanups: sub order, vertical alignment, #=

  * Move decode subs to the top so that all the read-* subs can be together.
  * Vertically align some arguments that were bugging me as is.
  * Change a # comment to #=.
00:46
-pb: 4c2a971 | (Geoffrey Broadwell)++ | lib/PB/Binary/Reader.pm:
Add type and 'is pure' annotations to Reader subs
-pb: a71214a | (Geoffrey Broadwell)++ | lib/PB/Binary/Writer.pm:
Begin building low-level binary PB writer code

Approximately matches what the reader can do, except for length-delimited values, which are still WIP.
-pb: efdeed3 | (Geoffrey Broadwell)++ | t/binary-writer.t:
Tests for low-level binary writer

Tests writing varints, fixed32, fixed64, and pairs thereof.
00:53 benabik joined, jnap left 01:09 cbk1090 joined 01:23 dirk103 joined 01:26 dirk103 left 01:34 cbk1090 left 02:03 araujo left, araujo joined
retupmoca MIME::Base64 now has a pure perl version! 02:07
lue retupmoca++
retupmoca (but will still use the parrot library by default if running rakudo on parrot)
it also has a new interface, so you can encode a Blob or Buf directly 02:09
02:09 Guest19106 left
colomon \o/ 02:21
02:26 araujo left 02:27 araujo joined
retupmoca and now Email::MIME uses the new MIME::Base64 02:31
so it is now (in theory) feature complete
next step is documenting all this stuff, I suppose
03:01 cbk1090 joined 03:13 renormalist joined, cbk1090 left 04:16 nofks joined 04:25 nofks left 04:32 SamuraiJack_ joined 04:50 bloonix left 04:54 preflex_ joined, ChanServ sets mode: +v preflex_ 04:55 preflex left, preflex_ is now known as preflex 05:02 daniel-s__ left 05:03 daniel-s__ joined 05:06 xenoterracide left 05:07 BenGoldberg left 05:21 [Sno] left 05:47 logie joined 06:08 atta left 06:09 jtpalmer left 06:12 crazedpsyc joined, cotto joined, atrodo joined 06:13 ivan`` joined 06:15 atta joined 06:16 retupmoca joined, hoelzro joined 06:17 jtpalmer joined
moritz retupmoca++ 06:18
06:21 johnny5_ left 06:22 johnny5_ joined 06:23 johnny5_ left 06:24 johnny5_ joined 06:28 kaleem joined 06:33 logie left 06:39 denis_boyun joined
tadzik hello from Denmark o/ 06:58
07:05 nnunley left
dalek -pb: 60dadfc | (Geoffrey Broadwell)++ | lib/PB/Binary/Writer.pm:
Refactor with a WireType enum

Get rid of the magic wire type numbers in PB::Binary::Writer.
07:05
-pb: a41d773 | (Geoffrey Broadwell)++ | lib/PB/Binary/Writer.pm:
Add write-blob8() for length delimited fields

Will be used for both bytes and string field types.
-pb: 5eb0bb2 | (Geoffrey Broadwell)++ | t/binary-writer.t:
Test writing string and bytes fields, including utf8
japhb__ Hello to Denmark o/
;-)
07:07 darutoko joined
diakopter . 07:07
07:09 iSlug joined
diakopter tadzik: hi 07:21
07:47 iSlug left 07:49 iSlug joined 08:02 ingy^ joined
ingy^ o/ 08:02
08:04 Vendethi_ joined
frettled Hello, Earthlings! 08:12
08:12 xinming left
moritz Hello, frettling 08:12
08:13 xenoterracide joined 08:14 xinming joined
arnsholt o/ 08:14
08:21 denis_boyun left, nnunley joined
tadzik hrm, CPH wifi is weird 08:23
masak in what way? 08:25
08:25 nnunley left
masak (g'm', #p6) 08:26
tadzik it's randomly saying "There's been an error: Unknown error" or "Welcome back, you can now browse the web"
so, well, it's like a regular public wifi ;)
08:28 rindolf joined
mathw good $localtime 08:29
rindolf mathw: hello.
mathw I'm going to a wedding today! :D 08:31
moritz mathw: have fun! 08:32
mathw Halloween is clearly the best day to get married
08:32 dansamo joined
mathw Unfortunately I have to spend the day at work first 08:32
08:33 woolfy joined
masak lizmat: github.com/perl6/specs/commit/3150...17671a5640 -- "PDD" feels a bit misleading, as "P" (depending on context) might just as easily mean "Problem" or "Pwnage" or "Pet" or "Pirouette". 08:34
mathw hah
PDD drives half the stuff at my office
masak lizmat: also, one (joking) usage of something doesn't make it rise to the level of a term "used within the Perl 6 community"... had the same feeling with the "lmddgtfy" entry, which as far as I know was a one-off joke. 08:37
tadzik isn't it an actual website? 08:38
lmddgtfy.com/ sort of...
moritz thought that "PDD" was a "Parrot Design Document" 08:39
08:40 ingy^ left 08:41 tobyink joined 08:42 iSlug left 08:49 Rotwang joined
masak tadzik: let me error establishing a database connection for you :) 08:50
moritz: now, *that* would actually be a useful S99 entry. 08:51
08:53 woolfy left 09:01 dmol joined 09:03 tobyink left 09:08 Rotwang left 09:11 zakharyas joined 09:13 Vendethi_ left 09:14 Vendethi_ joined 09:15 iSlug joined 09:17 GlitchMr left 09:19 Vendethi_ left 09:23 GlitchMr joined, spider-mario joined
dalek ecs: 0a8c287 | masak++ | S99-glossary.pod:
[S99] PDD means Parrot Design Document
09:35
09:36 tobyink joined 09:38 Vendethi_ joined 09:41 donaldh joined 09:50 sqirrel joined 09:54 tobyink left 10:09 nnunley joined 10:11 xenoterracide left 10:13 nnunley left 10:20 fhelmberger joined 10:27 fhelmberger_ joined 10:30 fhelmberger left 10:37 daxim joined 10:43 odoacre left 10:44 sisar joined 10:48 donaldh left 10:59 iSlug left 11:03 pecastro joined 11:09 dmol left 11:15 rixx joined 11:20 Vendethi_ left 11:21 rixx left
sisar Hello. I'm unable to install rakudo. Full log: paste.ubuntu.com/6335241/ . What am I doing wrong ? 11:36
FROGGS ohh, you build on cygwin? 11:39
diakopter sisar: seems that library was built on line 400
FROGGS sisar: you can't use the windows msi by any chance?
sisar FROGGS: yes, Cygwin 11:40
diakopter .. however, it was copied to dynext, not /home/siddhant/rakudo/install/dynext
another prefix fail
sisar FROGGS: Cygwin would be preferrable because then I can build the latest rakudo 11:41
diakopter: what do you mean by 'prefix fail'?
FROGGS sisar: if you had ActivePerl and MSVC Express, you could build latest too
sisar: the dll is not in the right place 11:42
moritz that's the wrong approach
we need to fix rakudo to also build on cygwin
FROGGS moritz: true, but I wanted to offer something that would work right now 11:43
sisar FROGGS: are there instructions for that approach available anywhere ? 11:44
FROGGS sisar: in the README perhaps?
I think I've seen it somewhere
sisar also, the file INSTALL.txt in the rakudo repo needs to be updated to reflect recent changes 11:45
FROGGS but installing ActivePerl, Git, MSVC Express, and then configure + nmake it in the MSCV developer console should be all you need to do
moritz sisar: can you please do a grep PARROT_LIB_DIR Makfile|head -n1 and paste the result here?
sisar moritz: "grep: Makfile: No such file or directory" 11:47
moritz sisar: sorry, Makefile
sisar oh !
moritz: no output
moritz that is seriously weird 11:50
sisar: can you please put the entire Makefile somewhere where I can read it? 11:51
sisar ok
11:51 ivan`` left
FROGGS I'd suggest you would also compile parrot again... maybe it is build with another --prefix? 11:52
sisar: is that possible?
(the changed prefix I mean)
diakopter what about also incluing --gen-parrot
sisar diakopter: that fails in the same way 11:53
moritz: paste.ubuntu.com/6335309/
moritz that's all not helpful
the output shows for example /home/siddhant/rakudo/install/lib/parrot/5.9.0-devel/tools/build/pmc2c.pl
which shows that parrot has the correct prefix
FROGGS k
moritz oh my 11:54
FROGGS does that lack the parrot specific makefile snippets?
moritz yes 11:55
because npq-p failed to build
moritz has been looking at the wrong end
sisar: sorry, I need nqp/Makefile too 11:57
11:58 nnunley joined
sisar ok 11:58
11:58 ivan`` joined
diakopter ヽ༼ຈل͜ຈ༽ノ 12:00
moritz UniArt
sisar moritz: paste.ubuntu.com/6335335/ 12:01
12:02 nnunley left
moritz sisar: in the nqp/ dir, please try running /home/siddhant/rakudo/install/bin/parrot.exe -L. -X. -Xdynext -o src/vm/parrot/stage0/nqpmo.pbc src/vm/parrot/stage0/NQPMO-s0.pir 12:06
sisar: if that works, I know how to fix it
sisar moritz: "error:imcc:loadlib directive could not find library `nqp_bigint_ops' in file 'src/vm/parrot/stage0/NQPMO-s0.pir' line 1" 12:07
moritz sisar: but dynext/nqp_bigint_ops.dll exists, right?
sisar moritz: yes 12:08
moritz rurban_: does parrot on cygwin look for dynamic extensions in .dll files?
12:11 darutoko left 12:15 ribasushi left, benabik left
sisar afk 12:20
12:21 BitPuffin left, BitPuffin joined, ribasushi joined
diakopter moritz: usually on wi dowz dkls have to br in ths path or same dir az the launched binaryy to be fou d 12:22
un 12:24
unless you ask with the dll'sfull path
12:25 Vendethi_ joined
diakopter moritz: u could ask sisar to try removing the .dll extension *shrug* 12:27
12:28 dmol joined
moritz diakopter: well, the -X option to parrot should control where it looks for these libraries 12:28
12:41 SamuraiJack_ left 12:42 SamuraiJack joined 12:45 SamuraiJack left, SamuraiJack joined 12:48 [particle]1 joined 12:49 [particle] left 12:50 dansamo left 12:53 dansamo joined 12:57 [particle]1 left 12:58 sisar left 13:01 iSlug joined
dalek rl6-roast-data: 507a1c5 | coke++ | / (5 files):
today (automated commit)
13:11
13:14 ajr joined 13:15 ajr is now known as Guest88233, nnunley joined, Guest88233 is now known as ajr_
[Coke] I screwed up and we missed a day. no real changes for days, though, so it was a good day to miss. 13:22
13:22 jaffa4 joined
masak [Coke]++ # tireless 13:23
FROGGS cygwin setup: 48% 13:25
masak FROGGS: that statistic sounds totally made up. :P 13:27
moritz FROGGS: but 87% of all statistices are totally made up! 13:28
(autopun!)
erm, meant to say that to masak
FROGGS *g*
jaffa4 FROGGS: any progress? 13:29
FROGGS jaffa4: not yet
had to repair the dishwasher this morning :o) 13:30
jaffa4 Can you do that?
masak moritz: don't worry, according to the Institute of Incomplete Information, 8 out of every 10 13:31
FROGGS jaffa4: repairing the dishwasher? well, it works now again :o) 13:33
moritz \o/
masak "Should array indices start at 0 or 1? My compromise of 0.5 was rejected without, I thought, proper consideration." — Stan Kelly-Bootle 13:38
:P
arnsholt =D
FROGGS hehe 13:39
mathw hah
FROGGS hi mathw 13:40
mathw hi FROGGS 13:42
PerlJam greetings
masak ahoj 13:43
FROGGS o/
masak so it has come to this.
moritz it has, indeed, come to this. 13:48
13:48 bluescreen10 joined
moritz why does youtube has to end playing the current video when I start a new search? 13:49
PerlJam foolishness
moritz it could just move it to a corner, make it a bit smaller, and does the search with ajax in the main area
masak moritz: that's the way the iPad YouTube client works, fwiw. 13:50
mathw that would immediately lead to people saying "why does YouTube keep playing the video when I do a new search?" 13:51
moritz mathw: to which the obvious answer is "because you didn't stop the video"
mathw the obvious answer to your question is why not open a new window and search in that? :) 13:52
moritz because that's a hassle
masak YouTube as a whole is a hassle. it keeps asking me to provide my "real name". as if watching videos has anything to do with that at all. 13:56
moritz thinks of "masak" as a kind of valid "real name" 13:57
masak oh, they apparently stopped doing that. now they ask "you have an unused channel -- can we just remove that?" -- yes, please. 14:00
14:02 darutoko joined 14:03 jnap joined
FROGGS cygwin setup: 2% 14:04
-.-
mathw I find it odd. Google integrate YouTube comments with their global identity system which is most manifest in the form of Google+ profiles. Half the reaction I see is "Google+ is a failure so obviously Google are trying to get higher user counts for it by the back door!" 14:07
not "I don't want to tie my YouTube activity to my Google+ account"
masak FROGGS: that's a lot fewer percents than last time.
mathw not privacy concerns
but worry about Google claiming Google+ is more successful than it is
very odd
masak I don't want to tie my YouTUbe activity to my Google+ account. I don't want a Google+ account. 14:08
mathw I think some people think that one day it might become more popular and they'll suddenly be forced to stop using Facebook.
14:08 rurban joined
masak well, I'd rather be in Google's silo than in Facebooks. but my first choice is neither of them. 14:09
mathw that's a viewpoint I can understand 14:10
I don't understand this apparent fear some people have of Google+ appearing to be anything more than an abject failure (which it isn't, but they like to pretend)
moritz well, facebook and google both have way too much private data of folks 14:11
imagine all that data in one hand
that's truly scary
14:13 Vendethi_ left
PerlJam ... and that's how skynet starts its attack. 14:14
mathw moritz: I'm mostly seeing these viewpoints from avid Facebook users, so their data's already out there. 14:15
moritz mathw: but not out there at google
14:15 logie joined
masak PerlJam: Skynet seeding its first attack on social user accounts? now *that's* scary. 14:15
moritz most avoid facebook users are *also* avid google users 14:16
(because most everybody is)
masak yeah, most of us have our feet stuck really deep in the cloud.
FROGGS masak: yeah, after about 95% it decided that the file transfer had aborted, so it asked nicely about choosing AGAIN what to install 14:18
mathw FROGGS: fun!
FROGGS yeah :/ 14:19
geekosaur notes that "avoid" and "avid" are almost but not quite antonyms... 14:20
PerlJam avoid avid ovid video 14:23
14:25 rurban left, pernatiy joined 14:26 denis_boyun joined, rurban joined 14:27 kaleem left 14:28 bluescreen100 joined 14:31 bluescreen10 left 14:32 pernatiy left
masak I think many ovid^Wavoid^Wavid language design people on this channel will enjoy exple.tive.org/blarg/2013/10/22/citation-needed/ 14:37
14:38 lizmat joined
PerlJam I mildly enjoyed some of the comments. It's like they didn't even read the article. 14:39
masak sounds like most comments on the web :/
japhb__
.oO( Hey! I object to that! You're completely wrong! Wait, what's the topic again? )
14:40
geekosaur there is a reason I use the Shut Up extension in my browsers :) 14:41
14:42 Vendethi_ joined
japhb__ geekosaur: What does it do? Remove comments like ABP removes ads? 14:42
geekosaur also, they seem to have missed perl (< 6)'s own $[, granting that it is deprecated in perl5
yep 14:43
mangles CSS
one unforutnate side effect is it also tends to remove bugzilla (and some other bug tracker) comments, so you need to turn it off there
but as it stores state per-site, that's a one-time thing per bug tracker 14:44
masak geekosaur: you imply bugzilla comment are somehow less bad than comments in general? :P 14:47
geekosaur at least a *slightly* higher signal to noise ratio, usually 14:48
given that anything over 0 qualifies
jaffa4 is there perl6 parser, short example? 14:49
geekosaur (besides, the dev responses to garbage comments can also be interesting in their own way :)
14:50 risou_awy is now known as risou, risou is now known as risou_awy, risou_awy is now known as risou
masak jaffa4: are you thinking of something like github.com/moritz/json/blob/master...Grammar.pm ? 14:51
lizmat good *, #perl6, from Salzburg, Austria
moritz \o lizmat 14:52
jaffa4 masak: maybe
14:52 kaleem joined, jeffreykegler joined
lizmat afk to see sights & 14:53
jaffa4 I mean perl 6 language perser. there is std.. but is there a short example of using std?
masak jaffa4: then maybe that example will serve you well. :)
jaffa4: oh! using STD.
jaffa4: I can't think of any. 14:54
moritz jaffa4: yes, the 'tryfile' script in the STD repo
14:54 tobyink joined
masak oh, nice. 14:54
TimToady ./viv -e is what I usually use
14:54 iSlug left
jaffa4 and the std repo isM 14:55
moritz ... easy to find
jaffa4 github.com/perl6/std/? 14:56
moritz yes 14:57
15:04 benabik joined, kaare_ joined
jaffa4 STD_P5.pm6 does not compile, there is noe TD 15:07
there is no STD. 15:08
15:08 rurban left 15:10 rurban joined
jaffa4 I am getting a weird error 15:13
Could not locate compile-time value for symbol P6
15:14 Vendethi_ left, rurban left 15:15 pernatiy joined, rurban joined 15:17 rurban1 joined
dalek ecs: fb6536e | masak++ | S99-glossary.pod:
[S99] PDD means two things

The two meanings now homestead together under one entry, warming themselves under each other's feathers.
15:18
15:19 rurban left 15:26 kaleem left 15:27 FROGGS left 15:28 lizmat left, jaffa4 left, pernatiy left, pernatiy joined 15:32 pernatiy_ joined 15:35 pernatiy left 15:36 PZt left, [particle] joined 15:39 FROGGS joined 15:43 dmol left 15:46 lowpro30 joined 15:58 denis_boyun left, tobyink left 16:00 sqirrel left 16:05 PZt joined 16:14 rindolf left 16:15 rurban1 left 16:17 rurban joined 16:22 pernatiy_ left, pernatiy joined 16:23 denis_boyun joined 16:27 rurban left 16:29 pernatiy left 16:30 pernatiy joined 16:34 dmol joined 16:35 jnap left 16:37 jnap joined 16:38 zakharyas left 16:40 rurban joined 16:45 jnap left 16:51 iSlug joined
japhb__ I have been wondering for some time how to describe the greater part of my Perl 6 development work to someone who isn't in the community (I get asked by friends and relatives all the time), and I just came up with the moniker "mantle developer". Not a core developer, but not a crust developer either. I figure it will be more fun to explain that term than the intracasies of Str.Numeric. :-) 16:51
16:51 pernatiy left
arnsholt That's a good word. I may very well steal it =) 16:52
16:52 jnap joined
moritz japhb__++ 16:53
colomon japhb++
colomon hopes there is only one japhb.
japhb__ There Can Be Only One.
(In fact, that was the whole point of the nick. ;-)
moritz wants to create more subduction zones that draw crust developers into the mantle and mantle developers into the core 16:54
jnthn wonders what a mantle plume would be... :) 16:56
16:58 jnap1 joined 16:59 jnap left 17:03 jnap1 left
benabik Lots of developers do tend to get ejected by energetic eruptions... 17:05
17:07 iSlug left
jeffreykegler I definitely will steal the core/mantle/crust meme 17:12
japhb__ moritz: jnthn++'s two day training was an excellent subduction zone. I think it just needs marketing.
.oO( Come visit the Marianas Trip and say you went deeper than ever before! )
17:13
s/Trip/Trench/
Sheesh.
s/Marianas/Mariana/ 17:14
japhb__ rolls his eyes
17:16 ajr joined 17:17 ajr is now known as Guest41749, ajr_ left, Guest41749 is now known as ajr_ 17:25 jeffreykegler left 17:28 pernatiy joined
TimToady Perl uses 0-based indexing because modular arithmetic is 0-based, and computers are all about modular arithmetic. 17:31
17:36 rurban left
japhb__ I use 0-based indexing because I can't *not* think about the CPU opcodes that implement standard compiled language data structures. 17:38
IOW: I use 0-based indexing because the CPU does.
17:39 stevan_ left
japhb__ Although I do admit to using 0 as an in-band sentinel for some algorithms. 17:39
17:40 rurban joined
TimToady what we need is a numeric type that is (1 but 0), (2 but 1), (3 but 2) ... * 17:42
TimToady had better go decorate for halloween before doing any more psychic damage 17:43
17:43 pernatiy left
geekosaur postfix º / ª ? 17:44
17:44 sqirrel joined 17:45 rurban left 17:46 fhelmberger_ left 17:48 ajr_ left 17:50 bluescreen10 joined 17:52 stevan__ joined 17:53 bluescreen100 left 17:58 daxim left 18:03 [Sno] joined 18:08 risou is now known as risou_awy 18:10 risou_awy is now known as risou 18:11 ajr_ joined 18:15 rurban joined 18:16 cognominal__ joined 18:18 darutoko left 18:19 FOAD_ joined 18:25 cotto left, mathw left, cognominal left, FROGGS[mobile] left, markov left, ggoebel13 left, woosley left, Celelibi left, _ilbot left, rafl left, FOAD left, _sri left, FOAD_ is now known as FOAD 18:26 Celelibi joined 18:27 risou is now known as risou_awy, iSlug joined 18:31 tobyink joined 18:32 PacoAir left 18:37 sqirrel left
diakopter masak: you've already got a virtual google identity/profile 18:38
and they probably know all your personal info too..
18:38 iSlug left
diakopter masak: I guess I'm a meteorite developer 18:41
18:44 dansamo left 18:54 SamuraiJack left 18:55 iSlug joined
dalek albot: 92be771 | moritz++ | runtime/run-rakudo-jvm-evalserver:
no need to use negative PIDs for killing the JVM
18:55
18:57 lizmat joined
moritz j: say 42 18:59
camelia rakudo-jvm 882e33: OUTPUT«42␤»
19:04 woolfy joined 19:05 SamuraiJack joined 19:10 SamuraiJack left 19:18 cotto joined, mathw joined, FROGGS[mobile] joined, markov joined, ggoebel13 joined, woosley joined, _ilbot joined, rafl joined, _sri joined 19:20 jaffa4 joined, Rotwang joined
FROGGS /home/leicht/rakudo/install/bin/parrot.exe -L. -X. -o src/vm/parrot/stage0/nqpmo.pbc src/vm/parrot/stage0/NQPMO-s0.pir 19:23
error:imcc:loadlib directive could not find library `nqp_bigint_ops'
moritz: I get the same 19:24
lizmat is back from seeing sights and catching up on backlog
FROGGS moritz: that nqp_bigint_ops.dll seems to be fine... (checking with ldd and nm) 19:26
19:27 slava joined, slava is now known as slava-work
lizmat [Coke], timotimo: re irclog.perlgeek.de/perl6/2013-10-30#i_7786286 and tests of Mix.roll: I can now replicate the problem 19:27
and am looking at a solution
moritz FROGGS: is there something like strace on cygwin? 19:28
slava-work TimToady: hate to ask here, but it appears that there was a video where you were show casing a module that showed what the regex engine in perl 5 did as it tried to match a regex against a string
moritz FROGGS: if yes, please try to run it, and see what file names it looks for
slava-work: perl -Mre=debug
FROGGS moritz: how do I use it? 19:29
moritz FROGGS: strace $command
slava-work moritz: ahh, thanks
moritz or strace -o file $program
19:30 Rotwang left 19:31 jeffreykegler joined 19:32 slava-work left 19:34 FROGGS left 19:35 xinming left
lue hello world o/ 19:36
19:37 xinming joined 19:39 sqirrel joined 19:44 Rotwang joined 19:49 FROGGS joined
dalek ast: fa28a3b | (David Warring [email@hidden.address] | S05-grammar/protoregex.t:
added fudged test for RT120146
19:49
lizmat [Coke], timotimo: testing fix for Mixy.rolll now
dalek kudo/nom: 1dd6720 | (Elizabeth Mattijsen)++ | src/core/Mixy.pm:
Fix Mixy.roll, to not use negative weights

Spotted by [Coke]++ and further investigated by timotimo++
19:57
19:57 lizmat left
19:58 lizmat joined
dalek ast: e195bd8 | (Elizabeth Mattijsen)++ | S02-types/mix.t:
Small adaptation to Mixy.roll test

To make sure that if it fails, it will fail both on parakudo and jakudo
19:59
20:02 lowpro30_ joined 20:05 lowpro30 left 20:11 risou_awy is now known as risou 20:13 iSlug left 20:14 jeffreykegler left 20:27 risou is now known as risou_awy 20:32 bluescreen10 left
moritz tadzik: irclog.perlgeek.de/perl.pl/ is now fixed 20:32
dalek ecs: e27f008 | (Elizabeth Mattijsen)++ | S02-bits.pod:
Add :chop adverb to quoting adverbs

The idea being that in combination with :to, a .chop could be done at compile time already if there is no interpolation happening at the end of the result.
20:33
20:33 bluescreen10 joined 20:38 stevan__ left 20:47 sqirrel left, Vlavv` left 20:55 benabik left
dalek p: 4f69419 | dwarring++ | examples/rubyish/ (5 files):
rubyish slurpy params
20:56
kudo/nom: 2bf7cbc | (Elizabeth Mattijsen)++ | docs/compiler_overview.pod:
RLHF fixes

Wow, is this document outdated now.
20:58 lizmat_ joined, lizmat left
lizmat_
.oO( Wifi at hotel really sucks )
20:58
20:59 Vlavv` joined
jnthn Balcony wifi here is very slow, but stable. :) 20:59
itz /last salz 21:01
oops
moritz does the JVM backend translate QAST straight to JVM bytecode? 21:02
or is there an intermediate format?
21:04 rurban left, rurban joined, telex left
lizmat_ at one point, I was sure there was an intermediate step, but right now I'm not so sure 21:05
jnthn will know :-)
just found this in docs/deprecations:
Str.capitalize and &capitalize are deprecated in favor
of the the Str.wordcase and &wordcase routines.
They will uncondtionally warn in 2012.10, and be removed in 2012.11.
jnthn moritz: JAST 21:06
moritz jnthn: takk
21:06 stevan_ joined
jnthn moritz: Which is pretty close to JVM bytecode 21:06
moritz tries to update docs/architecture.svg to multiple backends
21:06 telex joined
jnthn moritz: For Moar it's MAST, which goes straight to bytecode (no textual phase) 21:07
moritz I wonder if the serializer should be a separate box
like the backend-dependent part of World
jnthn The format is the same but it's gotta be implemented per backend...
moritz and the QAST compilers don't use World, right? 21:10
21:10 lizmat_ is now known as lizmat
lizmat so should I remove capitalize ? 21:10
moritz lizmat: +1 21:11
21:11 zwut00 joined, kaare_ left
dalek kudo/nom: 09a2ec1 | jonathan++ | src/vm/jvm/core/Threading.pm:
A Lock primitive, for use in low-level things.
21:12
kudo/nom: b83de34 | jonathan++ | src/vm/jvm/core/Threading.pm:
De-confuse start/run on Thread.

We use .run on the Scheduler to mean "run this thing", and .run on a Promise to mean "promise to run this thing", so Thread.run should really mean "run this on a new thread".
kudo/nom: c1064c7 | jonathan++ | src/vm/jvm/core/Threading.pm:
Implement Promise.Bool.
ast: 5051341 | jonathan++ | S17-concurrency/promise.t:
Test Promise boolification.
ast: 1f4fd4a | jonathan++ | S17-concurrency/lock.t:
Tests for Lock.
ast: 6f65a8b | jonathan++ | S17-concurrency/thread.t:
Update thread.t for API change.
kudo/nom: f5f8e47 | (Elizabeth Mattijsen)++ | src/core/Cool.pm:
Remove capitalize, as per docs/deprecations

Originally scheduled for 2012.11
21:20
ast: 46006ae | (Elizabeth Mattijsen)++ | S02-types/deprecations.t:
Remove deprecation tests for capitalize

As capitalize is no more
21:21
jnthn
.oO( capitalize punishment issued )
21:29 araujo left
dalek kudo/nom: 0c62dab | (Elizabeth Mattijsen)++ | docs/ROADMAP:
Remove mention of KeySet/KeyBag

As they're respecced and implemented
21:33
kudo/nom: cf66175 | moritz++ | docs/architecture.svg:
update architecture.svg to multiple backends

it contains the MoarVM backend in good faith
21:34
21:38 araujo joined 21:43 logie left 21:47 jaffa4 left
lizmat tries to get an early night 21:47
21:48 lowpro30_ left 21:51 woolfy left 21:54 colomon left 22:00 colomon joined 22:02 lizmat left 22:05 denis_boyun left, Ulti joined, denis_boyun joined, bluescreen10 left 22:08 BenGoldberg joined 22:12 risou_awy is now known as risou 22:25 denis_boyun left 22:26 denis_boyun joined, FROGGS[mobile] left 22:30 risou is now known as risou_awy 22:32 colomon left 22:33 xenoterracide joined 22:34 colomon joined
dalek kudo/nom: adf8221 | jonathan++ | src/vm/jvm/core/Threading.pm:
Provide a Thread.yield method.
23:03
kudo/nom: 7409806 | jonathan++ | src/vm/jvm/core/Threading.pm:
Fixes/improvements to select.
kudo/nom: 15a4103 | jonathan++ | src/vm/jvm/core/Threading.pm:
Add a CurrentThreadScheduler.

Always schedules things on the current thread. Useful for forcing synchrony on things that default to asynchrony.
23:03 spider-mario left 23:04 rjbs joined 23:05 BitPuffin left 23:06 BitPuffin joined 23:07 sftp left, sftp joined, Celelibi left 23:09 grondilu joined, rurban left 23:11 Celelibi joined
japhb__ jnthn++ # Kicking concurrent butt 23:19
jnthn *butts # :) 23:20
dalek p: 1d2a068 | dwarring++ | examples/rubyish/ (3 files):
rubyish slurpy hash params
23:23
japhb__ jnthn: When you're thinking you could really use a break from threads, but are still in a hacking mood, ability to convert between (u)int32/64 and float/double on JVM is near and dear to me. :-) 23:33
japhb__ attempts to enqueue a job on jnthn.scheduler ... 23:34
23:34 zwut00 left
japhb__ Come to think of it, arnsholt would know how to do that too, I suspect. 23:34
23:35 PacoAir joined
jnthn japhb__: Where convert means "re-interpret bytes"? 23:35
japhb__ jnthn: exactly 23:36
The problem is that Protocol Buffers intrinsically see data on two levels: chunks of raw data of certain lengths, and typed data. When you read a field from the PB, you read a tag indicating field ID and a "wire type" that says how to calculate the length to read. 23:37
Then after that you read the raw bytes, and encode/decode the bytes as appropriate according to the field type you look up for that field ID. If you don't recognize the field ID (because you have an old copy of the message definition), you are required to pass the raw bytes through for other processing elsewhere. 23:39
The inverse occurs while going the other direction; dumb filters are supposed to be lossless, except for possibly rearranging fields on the wire for canonical packing. 23:40
So I've got code that sees a "64 bit" wire type, loads 8 bytes and turns them into a uint64, and passes them on to the code that tries to map a field type to determine whether to reinterpret as a double. 23:41
And I'm trying to do all this without creating and passing a ridiculous number of tiny blob8's around. 23:43
23:45 Vendethi_ joined
jnthn Thus why you'd like to be able to cheaply get a subblob too that doens't copy, I guess... 23:46
japhb__ Yeah. 23:47
23:48 denis_boyun left
japhb__ There's lots of places I'd like to do work directly on subblobs/subbufs -- like encoding/decoding strings, for instance. 23:48
23:50 Vendethi_ left
tadzik greetings, #perl6 23:50
dalek kudo/nom: 9c52be0 | jonathan++ | src/vm/jvm/core/Threading.pm:
Assorted Channel API tweaks and improvements.

Change "finish" to "close", since it is more clear. Also "completed" becomese "closed" to match, removing that naming oddity. There is now also a peek method, and a bug fix to closed.
jnthn japhb__: Maybe when I'm done with my current round of concurrency work, I should spend some time on our binary data handling :) 23:51
dalek ast: 6013403 | jonathan++ | S17-concurrency/channel.t:
Some Channel tests.
jnthn japhb__: I think I've figured out what kinda direction to go in for subblob just referring to the underlyuing blob. 23:52
o/ tadzik 23:53