»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.
00:01 Adriaaaan left
raiph are atroxaper and Mouq the same person or at least coordinating their Pod::To::* efforts? 00:07
00:08 carlin left 00:12 _slade joined 00:16 hoverboard joined 00:24 myps joined
[Coke] jnthn++ lizmat++ 00:35
00:45 BenGoldberg joined 00:52 hoverboard left 00:53 hoverboard joined
[Coke] r: join ', ', [\//] Any, 0, 1 00:54
camelia rakudo-parrot b014db: OUTPUT«use of uninitialized value of type Any in string context in block at /tmp/tmpfile:1␤␤»
..rakudo-{jvm,moar} b014db: OUTPUT«use of uninitialized value of type Any in string context␤»
timotimo [Coke]: did you mean q[\//]? 00:57
oh, wait, what?
that's a triangle reduce %)
timotimo looks at the clock
ayup. brain's already asleep-o-clock.
[Coke] look for "orelse" in the test suite.
timotimo down.me/ >:( 00:59
01:00 Akagi201 joined
[Coke] m: "\x[c]" ~~ m/^ <[\C[FORM FEED (FF)]]>/ 01:01
camelia ( no output )
[Coke] m: say "\x[c]" ~~ m/^ <[\C[FORM FEED (FF)]]>/
camelia rakudo-moar b014db: OUTPUT«｢｣␤␤»
01:01 raiph left
timotimo m: say '\x[c]' ~~ m/^ \\ <[x]> '[' ~ ']' 'c' $/ 01:04
camelia rakudo-moar b014db: OUTPUT«｢\x[c]｣␤␤»
01:05 Akagi201 left
psch [Coke]: i guess those two tests in reduce.t want something like &make-string in S32-array/delete.t 01:16
because (Any) in lists is gone, isn't it 01:17
01:18 dayangkun joined 01:22 _slade left
timotimo gone? 01:25
no, i don't think so
there's still a difference
m: my @a = 1, 2, 3; @a[1]:delete; say @a.perl; say @a[1]:exists; say @a[3]; say @a[3]:exists;
camelia rakudo-moar b014db: OUTPUT«Array.new(1, Any, 3)␤False␤(Any)␤False␤»
timotimo m: my @a = 1, 2, 3; @a[1] = Any; say @a.perl; say @a[1]:exists; say @a[3]; say @a[3]:exists; 01:26
camelia rakudo-moar b014db: OUTPUT«Array.new(1, Any, 3)␤True␤(Any)␤False␤»
psch m: my @a = 1, Any, 3; @a.list.say
camelia rakudo-moar b014db: OUTPUT«1 (Any) 3␤»
psch right
so does join do the wrong thing?
01:26 _slade joined
timotimo i didn't even look what you're talking about %) 01:27
psch still the test [Coke] talked about
timotimo r: join ', ', [\//] Any, 0, 1
camelia rakudo-parrot b014db: OUTPUT«use of uninitialized value of type Any in string context in block at /tmp/tmpfile:1␤␤»
..rakudo-{jvm,moar} b014db: OUTPUT«use of uninitialized value of type Any in string context␤»
timotimo that seems correct
psch the tests are skipped with "implemented orelse"
timotimo oh, i think i did something to orelse recently .. didn't i? 01:28
psch err, todo
no skip :)
timotimo or did i end up not doing anything?
psch i wouldn't know 01:30
m: say [\//] Any, 0, 1
camelia rakudo-moar b014db: OUTPUT«(Any) 0 0␤»
timotimo i've watched too many speedruns recently because now i can't read "skip" without ... thinking of that 01:31
01:31 Vlavv left
psch but SGDQ is over for like two weeks already... 01:32
timotimo yes
i've watched a bunch live
and after that i've been watching bunches of vods
psch i've been rather selective about what i watch
timotimo got recommendations?
psch not really, actually 01:34
timotimo was all of it bad? ;)
psch i mostly watch the old sidescrollers
but those runs don't really change that much anymore
timotimo did you see the super metroid xray door glitch?
psch i don't even remember, actually 01:35
timotimo would you mind if i explained how the trick works?
psch no
timotimo wait, that means i can explain it, right? %) 01:36
damn english :D
psch i used to explain random people how q3 overbounces work; i understand
*to random people
so, yeah, you can explain
:9
timotimo i was expecting you'd prefer to see it first and then get the explanation
i didn't expect you'd have a hard time understanding
psch i'm not really passionate about it
timotimo aaw
so the basic thing that happens is you get out of bounds, fall down a whole bunch of screens and the game keeps interpreting its RAM contents as the level 01:37
so what ends up happening is you explore the out-of-bounds space and manipulate one specific piece of memory
and then ... i forgot how exactly it works, and i also forgot how the xray item is involved ... you end up in the planet self-destruct sequence at the very end 01:38
except there is no timer
psch :D
timotimo someone edited it into a very short video and uploaded it to youtube 01:39
psch "right, so it starts like this... and then there's a few steps, i don't know what they are... but anyway then this thing happens"
timotimo it'll do a massively better job than me %)
psch yeah, i think i'd have to watch it :)
timotimo www.youtube.com/watch?v=-GAymJt6KPM - 6 minutes of video to see it
out-of-boundsness starts at about 30 seconds 01:40
01:40 xenoterracide__ joined 01:42 atroxaper joined
timotimo oh damn, i think the actual explanation of why xray is so broken is *before* the start of that video?! 01:43
no, it comes afterwards it seems
Mouq jnthn++ lizmat++
sorry for unexpectedly extended afk
01:44 jack_rabbit joined, Vlavv joined, FROGGS_ joined
dalek ast: 0f90ebd | coke++ | integration/advent2013-day15.t:
skip all vms implicitly instead of explicitly
 01:45
ast: 4bf731f | coke++ | integration/advent2013-day19.t:
unfudge-passes on moar, marked as #conc
ast: d479879 | coke++ | S05-mass/named-chars.t:
Fix skip count
ast: ba56bb9 | coke++ | S05-mass/named-chars.t:
Add RT reference #
ast: d2f0d4f | coke++ | S05-mass/named-chars.t:
skip -> todo where possible.
01:46 klapperl_ joined 01:47 atroxaper left
psch timotimo: definitely interesting glitch 01:47
timotimo aye
psch the "the lost vikings" run was quite interesting too iirc, although not that glitchy
timotimo that was the one where the runner held both controllers at the same time, back-to-back?
psch i think there was more than one run with 1 player 2 controllers, but yeah 01:48
(the other was chip & dale, rescue rangers, says the schedule)
01:48 FROGGS left
timotimo it's kinda sad that games nowadays don't do such dirty things with their memory any more that you can't just walk outside the screen and start overwriting game memory 01:48
i was kind of disappointed at the octodad: dadliest catch race: misskaddykins got one of those very typical "this has never happened before" moments that forced her to fall back about half a minute or so :( 01:50
01:50 klapperl left
Mouq raiph: We're not quite the same person, no :) 01:50
01:51 _slade left
Mouq raiph: I haven't actually talked to him, but what he wants to do is one of the reasons I want to try to make Pod::Walker work 01:51
psch timotimo: i don't remember if i watched that, but "this has never happened before" did kind of come up a lot in the things i did watch; marathon as usual, i guess 01:52
timotimo yes :)
psch oh, also because (even horrendously off-topic) clear communication is important: i didn't mean to imply that i wouldn't understand the xray glitch, just that i understand the feeling of wanting to share cools one, hence my referal to q3 overbounces 01:57
psch .oO( ...but that's probably enough. )
01:58 chenryn joined
timotimo ah 01:58
yes, clear communication *is* important. thank you!
what does overbouncing relate to?
i've seen videos of people "parkouring" with rocketjumps and the plasma gun and such 01:59
psch basically not accounting for the length of the collision trace vector when trying to determine if the player model is still falling
timotimo i don't think i've seen that yet 02:00
02:01 Akagi201 joined
psch www.youtube.com/watch?v=5DZIkZBTh94 02:02
if i remember the code correctly it comes down to the engine assuming that if the trace starts above a surface and ends inside one, it assumes the previous extrapolation already determined that the player must have landed 02:03
which means the player is assumed to not have a z component in his impulse, which messes up denormalization 02:04
timotimo oooh
"clever" optimization :)
it's quite nice that the movement vector ends up being "healthy" 02:05
02:05 hoverboard left
psch yeah, there's tons of entertaining videos around 02:05
timotimo i'm worried with the growing amount of games being developed using "big name" engines like unreal, source, unity ... they're not going to have as many "fun" bugs in (for example) their physich engines 02:06
02:06 potatogim joined
timotimo though physics engines can still get really glitchy even with "big name" engines given sufficiently pessimal collision geometries 02:06
02:06 Akagi201 left 02:07 huf joined
timotimo source engine 1 had this funny thing where you could run + jump backwards and become pretty darn fast :D 02:07
oh, it's already very late, i ought to go to bed
gnite!
02:08 potatogim is now known as PotatoGim
psch heh, good night :) 02:08
timotimo oh, btw: #perl6-gaming :)
gtodd the reactive programming talk by jnthn at NL perl workshop made me think of applications of perl6 in games video computer vision ... "grepping" inside streams of data 'n such 02:09
timotimo we'll have to see if moar can get sequences of supplies well optimized with - perhaps! - inlining and such 02:10
02:15 Akagi201 joined 02:17 Akagi201 left 02:18 noganex_ left 02:24 myps left 02:25 cooper_ left 02:27 Tene left, Tene joined 02:32 _slade joined 02:33 takesako left 02:36 cooper_ joined 02:45 jack_rabbit left 02:46 BenGoldberg left 02:48 takesako joined 03:00 jnap1 left 03:06 noganex joined 03:08 shbng joined 03:26 hoverboard joined 03:31 jack_rabbit joined
[Sno] FROGGS_, lizmat - HL before you talk then I can disturb better ;) 03:33
business hours: 7am-5pm :) 03:34
03:38 hoverboard left 03:42 synthkeys2 joined 03:48 xenoterracide__ left 03:50 SevenWolf left 03:53 [Sno] left 04:00 lue left 04:01 gfldex joined 04:03 kaare_ joined 04:07 lue joined
shbng Hey, could anyone give a guesstimate as to when a full (meets current spec and passes spec-test), unoptimized, implementation of perl6 might be ready? (Hoping within the year but I don't know if that's realistic, We seem really close) 04:10
perigrin there are a number of assumptions in that that are common but full of fallacies 04:11
the biggest being that the "spec" is somehow static.
shbng: perl6.org/compilers/features probably is the best answer to your question 04:13
most of the bits there that are red are bits where the spec itself is still somewhat fluid (afaict) 04:14
04:16 sftp left 04:19 sftp joined, rindolf joined, anaeem1_ joined
shbng Hey, could anyone give a guesstimate as to when a full (meets current spec and passes spec-test), unoptimized, implementation of perl6 might be ready? (Hoping within the year but I don't know if that's realistic, We seem really close) 04:31
oops
I realize the spec is rather fluid, actualy that is one of my biggest reasons for confusion. I have tried using perl6 again last week (after a 2-4 year break) and was happy to see it mostly usable with moarvm, but I am still having problems(mostly with threading) and am looking for a good time to retry. (so I guess, when do you think threads will be better implemented at least)
perigrin that I can't really say. perhaps leaving examples of the kinds of issues you're specifically having in a place (blog post maybe?) where the people working on it can look and provide feedback would help? 04:35
I know that jnthn recently (a few months ago) gave perltv.org/v/reactive-programming-in-perl-6 04:36
04:44 atroxaper joined 04:46 kaare_ left 04:48 atroxaper left 04:52 rurban2 left 05:00 _slade left 05:01 shbng left 05:03 kaleem joined 05:05 rindolf left 05:06 rindolf joined 05:10 kaare_ joined, rindolf left 05:16 [Sno] joined 05:23 treehug88 joined 05:31 kurahaupo left 05:37 dayangkun left 05:38 denis_boyun joined 05:39 treehug88 left 05:49 dayangkun joined 05:57 dayangkun_ joined 05:58 dayangkun left
sergot o/ 06:13
06:16 denis_boyun left 06:18 dayangkun_ left 06:28 dayangkun joined, dayangkun left 06:29 dayangkun joined 06:35 SevenWolf joined, dayangkun left 06:36 jack_rabbit left 06:38 ribasushi left 06:39 SevenWolf left 06:40 SevenWolf joined, teodozjan joined
FROGGS_ there's a regression in r-m*: [...]languages/perl6/lib/Math/Model.pm 06:40
===SORRY!===
When pre-compiling a module, its dependencies must be pre-compiled first.
06:41 ribasushi joined 06:42 jack_rabbit joined 06:48 dayangkun joined 06:49 dayangkun left, dayangkun joined 06:51 jack_rabbit left 07:01 pdcawley left 07:03 pdcawley joined, [Sno] left 07:06 [Sno] joined 07:11 denis_boyun_ joined 07:17 FROGGS_ left
mathw good morning! 07:19
07:27 dmol joined 07:30 FROGGS_ joined 07:32 synthkeys2 left 07:33 denis_boyun_ left
FROGGS_ that's weird: 07:34
[...]install/languages/perl6/lib/Math/Model.pm
===SORRY!===
When pre-compiling a module, its dependencies must be pre-compiled first.
Please pre-compile /home/froggs/dev/star/work-m/rakudo-star-2014.06/install/languages/perl6/lib/SVG.pm
but there is a SVG.moarvm in that path...
07:34 dayangkun left
FROGGS_ lizmat: is it possible that one of your patches did that? 07:35
07:37 ivanshmakov left, ivanshmakov joined 07:40 SamuraiJack_ joined
lizmat FROGGS_: possibly 07:43
FROGGS_ lizmat: I'm not sure how to provide a simple test case though...
lizmat my patches didn't break any spectest, maybe we need a spectest :-)
FROGGS_ I guess we have not a single test that tests precompilation 07:44
moritz lizmat: the problem is that precompilation isn't specced
well, Test.pm6 is precompiled
FROGGS_ moritz: good point
moritz we need to start rakudo-specific tests
but Test.pm6 is pretty simple, in terms of precompilation
most importantly, it doesn't use any other precompiled modules 07:45
lizmat moritz: I started making them in t/spec/S11-modules
and t/spec/S22-package-format
07:46 igorsutton joined 07:48 dayangkun joined
moritz lizmat: are those tests for packaging, or for precompilation? 07:48
lizmat in the future, both actually
FROGGS_ lizmat: the code here assumes that the :$from is always lowercase, that's not true for NQP for example: github.com/rakudo/rakudo/commit/c5...606b3edf80
that is not our problem here though 07:49
lizmat indeed not, as that behaviour didn't change in that patch
so you're saying it should always be :from<NQP>, or that it could also be :from<nqp> ? 07:51
FROGGS_ we are using NQP, and I use Perl5 in v5... 07:54
the reason is that this field is about a language, not a compiler or so... and 'use openssl:from<c>' just looks wrong :o) 07:55
lizmat ok, so do you always want "Perl5" then? and "NQP" 07:56
FROGGS_ I'm back in an hour or two...
lizmat or should "perl5" and "nQp" also be allowed
FROGGS_ lizmat: that is how the name of the language is written 07:57
lizmat ok
FROGGS_ Perl 6 is strict, so having case insensitive stuff feels weird
dalek kudo/nom: 18bb089 | (Elizabeth Mattijsen)++ | src/core/CompUnitRepo/Local/File.pm:
Also allow :from<NQP>
 07:59
kudo/nom: 6fab386 | (Elizabeth Mattijsen)++ | src/core/CompUnitRepo/Local/File.pm:
Only allow :from<Perl5 Perl6 NQP>
08:06 dayangkun_ joined
masak morning, #perl6 08:06
lizmat masak o/ 08:07
08:08 dayangkun left 08:13 darutoko joined 08:16 denis_boyun_ joined
masak is in a good mood today 08:20
08:23 dmol left 08:25 brrt joined
jnthn grumbles about the warm weather and being awake and stuff 08:30
uh, morning :)
08:31 virtualsue joined
nwc10 today, you're the anti-masak? 08:31
08:32 isBEKaml_mobile joined
jnthn I might improve with coffee :) 08:33
08:34 dakkar joined
nwc10 warm weather, warm coffee? That's illogical 08:38
You need warm *tea* :-)
lizmat probably going to rain all day here :-(
08:47 spider-mario joined
masak it is warm, I'll grant you that. 08:49
Mouq lizmat: What about :from<JVM>? 08:52
or <Java> or whatever it is?
lizmat once we support that, we will add it :-) 08:53
do we support that already then?
brrt morning jnthn :-)
it's quite nice here, just 17 degrees :-) 08:54
jnthn o/ brrt 08:57
lizmat: Yes, there is some basic JVM interop support using :from
masak LiveScript looks crazy, but possibly in a good way. 08:58
or, let's say, LiveScript looks like what CoffeeScript *could* have been, if it were ambitious/imaginative enough.
08:58 jack_rabbit joined 09:05 isBEKaml_mobile left 09:07 xinming_ joined
brrt livescript is? 09:09
brrt can't bother to renember all javascript copmilers out by now 09:10
09:10 xinming left
masak brrt: one of those compile-to-JavaScript languages. 09:10
brrt: livescript.net/
brrt i see
masak brrt: I believe au and vendethiel are both involved in it.
nwc10 wonderfully confusing
LiveScript was the previous name for JavaScript
brrt i wish them well :-)
masak nwc10: oh! I thought you meant the language itself :) it sure has a lot of syntax. 09:11
nwc10: but yeah, the naming is deliberate.
as in, whoever chose that name knows about the naming history.
livescript.net/#name
dalek kudo/nom: 3062bc1 | (Elizabeth Mattijsen)++ | src/core/CompUnitRepo/Local/File.pm:
Allow for :from<JVM>
 09:12
09:13 dmol joined 09:17 dmol left 09:18 dmol joined 09:19 sqirrel joined, dmol left
dalek c: 240dbdd | Mouq++ | / (14 files):
Call SUBTITLEs what they are
 09:20
c: eb191dc | Mouq++ | lib/Language/classtut.pod:
Merge branch 'master' of github.com/perl6/doc
09:21 dmol joined 09:22 dmol left 09:23 dmol joined 09:31 kkul joined 09:32 teodozjan left, dmol left 09:33 dmol joined 09:35 dmol left 09:36 kkul is now known as teodozjan 09:37 dmol joined
masak in 10 days, Perl 6 turns 14. 09:41
09:47 Psyche^ joined, dmol left 09:49 dmol joined 09:52 Psyche^_ left, dmol left
brrt we ought to have a birthday cake, then 09:52
09:53 dmol joined
masak or maybe we should get up and throw a mug. 09:54
09:54 dmol left
moritz we should throw mugnetic objects! 09:54
09:54 dmol joined 09:55 virtualsue left 09:57 virtualsue joined 10:06 jack_rabbit left
masak moritz: ...close to copper wires, generating some electricity... 10:07
10:08 xinming joined 10:10 xinming_ left, dmol left 10:11 dmol joined 10:14 dmol left, FOAD left 10:15 dmol joined 10:16 Akagi201 joined 10:28 sqirrel left, SevenWolf left 10:32 dmol left 10:33 dmol joined 10:45 atroxaper joined
colomon_ FROGGS_: 10:46
===SORRY!===
When pre-compiling a module, its dependencies must be pre-compiled first.
Please pre-compile /home/smoker/emmentaler/installlib/lib/NativeCall.pm6
FROGGS_ colomon_: yes, I see that also in my r-m* test
colomon_: I'm debugging it right now
colomon_ FROGGS_++
10:47 sqirrel joined, cognominal left
FROGGS_ lizmat: btw, I agree now that a single CUR in an @*INC slot makes more sense... 10:48
lizmat: we can have a CUR::Collection thingy in future, that has some rules to handle multiple CURs, especially when it has to pick an installation CUR 10:49
10:50 dmol left 10:51 cognominal joined, dmol joined, chenryn left
FROGGS_ ahh, found it 10:52
pretty sure at least
yes, that seems to work 10:55
colomon_ … tell us ...
FROGGS_ github.com/rakudo/rakudo/blob/nom/...pm#L51-L57 10:56
line 52 is bogus, the {$root} is missing
and we can reduce line 51 to 57 down to line 56
10:56 cognominal left
colomon_ Are you sure that's it? smoker's rakudo build doesn't even have that patch... 10:58
FROGGS_ colomon_: on what commit is your rakudo? 10:59
11:01 Adriaaaan joined
colomon_ FROGGS_: b014db2ebe0c5679b12c5bf709c98a9918f7f430 11:01
11:01 Adriaaaan left
colomon_ smoke run started 5-6 hours ago, I haven't updated it since 11:01
11:02 tadzik left
dalek kudo/nom: 3397a71 | (Tobias Leich)++ | src/core/CompUnitRepo/Local/File.pm:
fix lookup of potential comp unit candidates

In case we have already seen a basename of a candidate we dropped additional candidates. Therefor we missed a precompiled cu when we already had seen a source file. Since the content of a directory is not ordered the result depended on the phase of the moon, mostly.		 11:04
FROGGS_ colomon_: you have that bit of code: github.com/rakudo/rakudo/blob/b014...pm#L50-L56 11:05
colomon_: feel free to rerun :o) 11:07
colomon_ FROGGS_: oh! sorry, must have read github wrong.
FROGGS_ colomon_: no problem
colomon_ FROGGS_: I thought it was saying that was lizmat's patch from two hours ago. 11:08
must actually have been the commit level for the entire file.
FROGGS_ IIRC the bad commit was 15 hours ago
lizmat FROGGS_++ (good catch) 11:16
that was definitely a refactor gone wrong 11:20
.oO( need more tests )
colomon_ FROGGS_++ # that did i 11:22
*it
11:22 brrt left
colomon_ launches a full smoke test of the latest rakudo 11:22
11:26 sqirrel left
lizmat do we actually have a plan for the case there is a compiled version of a module, but it is out of date? 11:27
wrt to the current executor?
FROGGS_ lizmat: the current state is that it explodes
but the plan is to recompile it, in case we can write to the destination 11:28
11:28 anaeem1_ left
dalek kudo/nom: 9013d8f | (Elizabeth Mattijsen)++ | src/core/CompUnitRepo/Local/File.pm:
Only lookup / once
 11:28
FROGGS_ because a .moarvm file is a cache after all, and caches can get invalid and want updating
lizmat perhaps, but I would be against compiling a new version then 11:29
FROGGS_ with the CUR::inst we have enough information to do that actually
lizmat I'd rather see automatic re-compilation as part of the rakudo upgrade process
FROGGS_ well, such a cache can get only invalid when you recompile rakudo, or you upgrade a module using panda 11:30
in either case there is an installation process running
11:30 anaeem1 joined
FROGGS_ so, would not hurt to hook into that 11:30
lizmat agree 11:31
11:32 teodozjan left 11:36 isBEKaml_mobile joined
FROGGS_ but yeah, trigger re recompilation just because a .moarvm in your local lib folder is out of date would scare me as well 11:38
lizmat FROGGS_: what does the :$file paramater in CompUnitRepo.candidates mean ? 11:39
FROGGS_ lizmat: that we already know the path to a file to load
like in require "/foo/bar/baz.pm" 11:40
lizmat but it is not used anywhere ?
FROGGS_ it is used
in Perl6::Actions
lizmat aha, so we need to provide elsewhere as well, is what you're saying 11:41
FROGGS_ or better in: use Foo:file<blarg>
lizmat there is no string 'candidates' in Action ? 11:42
FROGGS_ no, it just calls load_module
which calls $*W.load_module which calls CUR.load_module which calls .candidates on the curs
lizmat ok, get that now (I think) 11:43
which feels to me it's rather backward: load_module shouldn't call candidates
by the time we call load_module, we should already know what module to load 11:44
FROGGS_ but yeah, the file param comes from require/use Foo:file<bar>
lizmat: no, 'use Foo' results in a call to load_module('Foo')
how do you know the candidates then already? 11:45
we are checking for candidates at the time somebody wants to load a module
lizmat in my view, load_module should be told which *file* to load 11:46
so it can be directly used for -require-
use Foo would first call candidates, find out the file name, then call load_module
FROGGS_ lizmat: it calls load_module because that is the API of the 'old' module loader 11:47
lizmat indeed
FROGGS_ and the new loader (CUR) just sits in front of it and provides the same API
it was made pluggable
lizmat and now the old module loader is obsolete... :-) 11:48
FROGGS_ so it is just called load_module
11:48 dayangkun_ left
FROGGS_ lizmat: no, we use it after we found a candidate 11:48
lizmat but is that a desirable situation?
FROGGS_ and pass it a :$file along
lizmat: yes, because we need to be able to load modules 11:49
you do not want to do that in the CUR
lizmat why not ?
11:49 isBEKaml_mobile left
FROGGS_ our module loader is written in nqp for one thing, and it delegates to NQP's ML or to MLs of a slang 11:49
you can try to port it to Perl 6, but I'm not sure this will work out 11:50
for example when we want to load BOOTSTRAP.pm, we need an NQP-level module loader to load it
because we are not in the setting yet 11:51
lizmat I don't actually want to port Perls/ModuleLoader to Perl 6
FROGGS_ note that we are registering our CUR at the end of the setting
lizmat I want to port Actions/load_module to Perl 6
FROGGS_ what is Actions/load_module? 11:52
also, we need to deserialize CUs when we load a precompiled module, that happens in $*W.load_module
11:53 rurban1 joined
lizmat World/load_module I meant 11:53
FROGGS_ we need the pieces where they are in order to do that
lizmat: I'm not sure you want to emit QAST in Perl 6 level code: github.com/rakudo/rakudo/blob/nom/...d.nqp#L332 11:54
lizmat I'm thinking along these ways: 11:55
1. CUR.candidates gives back one ore more CompUnit objects 11:56
2. if there is only 1, we try to load that by calling .load on the object
3. the CompUnit object should know which way to load itself (depending on pre-comp, source or external binary) 11:57
grrr... need to take our cat to the vet :-(
bbl&
FROGGS_ :/ 11:58
11:58 anaeem1 left 11:59 jnap joined 12:00 dayangkun_ joined 12:01 atroxaper left 12:05 atroxaper joined 12:09 rurban1 left, teodozjan joined, rurban1 joined 12:12 virtualsue left, erkan left 12:15 Ven joined
Ven o/ #perl6 12:17
12:21 sqirrel joined
masak ahoj, Ven. 12:24
atroxaper Hello, #perl6 ! 12:37
yoleaux 7 Jul 2014 20:01Z <lue> atroxaper: assuming L<foo> is a synonym for L<doc:foo> (S26 doesn't say anything about this explicitly), I could see a case for L<C<foo>> being equivalent to L<doc:foo> rather than L<doc:C<foo>>. However, I'd also say L<C<foo>> is wrong, and you deserve whatever broken output results :) You really should do either C<L<foo>> or L<C<foo>|foo>
7 Jul 2014 21:12Z <Mouq> atroxaper: github.com/perl6/Pod-To-HTML/commi...07f013a947
atroxaper Wow
r: (Nil, 0)[0].defined 12:40
camelia rakudo-jvm 9013d8: OUTPUT«(timeout)»
( no output )
12:41 carlin joined
atroxaper r: say (Nil, 0)[0].defined 12:41
masak lue: it feels rather obstinate for the parser to outlaw L<C<foo>> but allow L<C<foo>|foo>, IMHO. 12:42
camelia rakudo-{parrot,jvm,moar} 9013d8: OUTPUT«True␤»
masak lue: some part of me goes "do what I mean, darnit!"
lue: though I should have started by saying I agree with you in principle: L<C<foo>> feels "wrong". 12:43
atroxaper I thinks I will want to write L<httpC<s>://github.C<com>> for visibility, but without necessity to write |github.com 12:45
carlin github.com/rakudo/rakudo/pull/285
masak atroxaper: a very good point. that won me over to allowing stuff like that. 12:47
atroxaper: and (by extension) to providing the service of auto-stripping formatting inside of an L<>
atroxaper masak: actually, now we have L<C<email>|email> in AUTHOR section 12:48
masak heh.
I wouldn't be surprised if auto-stripping logic were found along some other code path in the Pod parser.
FROGGS_ carlin: wikipedia says: The offset from UTC is given in the format ±[hh]:[mm], ±[hh][mm], or ±[hh]
carlin: looks like we would not support ±[hh]
atroxaper masak: as I said a couple days ego, I'm writing system for writing Pod::To::{Something} modules. And I'm rewriting Pod::To::HTML to my system as proof :) 12:51
12:51 igorsutton left
atroxaper masak: Next step will Pod::To::Markdown 12:51
masak atroxaper++ 12:52
atroxaper masak: And maybe in future something like JavaDoc generator for Perl6. I feel a lack of documentation. And I'm Java developer and I can read java code with comments of developer any time. And can read html javadoc :) 12:53
12:54 fhelmberger joined
masak atroxaper: re JavaDoc-like generator: very, very yes. 12:54
especially if it's savvy to all the neatness of Pod <3
atroxaper masak: yes! 12:55
12:55 guru joined, guru is now known as Guest22875
atroxaper masak: I could not understand why there is no system to write comments under code in Perl 5. 12:55
12:56 igorsutton joined, Guest22875 is now known as ajr_
atroxaper And each developer should be fiction-writer and write some poem et the end of file) 12:56
masak I agree. 12:58
“Programs must be written for people to read, and only incidentally for machines to execute.” -- SICP 12:59
zengargoyle atroxaper: how much of Perl6 POD are you supporting? 13:00
awesomely pretty HTML of SICP: sarabander.github.io/sicp/ 13:02
atroxaper zengargoyle: all of it. I mean all things that current grammar can parse.
13:02 kaare_ left
masak zengargoyle: wow. I just started re-reading the original HTML, but this is much nicer to read. thank you! 13:04
zengargoyle sicpebook.wordpress.com/2014/01/16/...5-version/ # this was the page about that version, mentions an EPUB version as well. 13:05
zengargoyle thinks that's what Pod::To::HTML should look like :) 13:06
atroxaper zengargoyle: Current Pod::To::HTML version has a hint that it would be nice to have some system that iterates on Pod and call some callbacks. I'm trying to make that system and make it comfortable. And make code of that callbackers more readable and mainteined.
[Coke] (callbacks) like a pod::sax ? 13:09
atroxaper zengargoyle: I will try to do so ;-)
[Coke]: No. Actually, I don't know about pod:sax. 13:10
zengargoyle i tried to play with POD in code a couple of days ago and got --doc all confused. think i was the :option<> type things or the = :more<options> continuation lines or something.
eventually decided to stick with simple stuff ala Perl5 style 13:12
atroxaper [Coke]: Imagine, that you write a bunch of functions (callbacks) and register it on Start, Finish and Content of any Pod's objects. Off course you have shared storage, array for result and some other data. 13:14
moritz sounds like inheriting from a visitor class to me 13:15
atroxaper [Coke]: And each callback will get all data that he want. Callback just need to write name of nessesary data in the header of sub 13:16
moritz: yes. With features :)
moritz: nothing rocket-science. 13:17
13:18 virtualsue joined
atroxaper r: say (Nil, 0)[0].defined # btw, why? 13:19
camelia rakudo-{parrot,jvm,moar} 9013d8: OUTPUT«True␤»
psch r: say (Nil, 0).elems 13:20
camelia rakudo-{parrot,jvm,moar} 9013d8: OUTPUT«1␤»
psch good afternoon #perl6 13:21
atroxaper psch: hello) 13:22
psch: hm...
jnthn I think Nil is no longer meant to flatten away. Once the spec had Nil as vanishing in List context, which Rakudo's list model relied on.
And in places still does.
13:22 dayangkun_ left
jnthn It was a rather unfortunate/painful spec change. 13:22
Anyway, I think these days it wants .elems to be 2 there. 13:24
psch jnthn: yes, that's what S02#Nil says at least: "Nil [..] does not disappear in list context"
jnthn Yeah. It hadn't used to be that way.
13:29 rurban1 left 13:30 rurban1 joined 13:34 rurban1 left 13:38 teodozjan left 13:42 firefish5000 left, rurban1 joined 13:50 dmol left
carlin FROGGS_: you're right, I've pushed another commit to that pull req 13:51
FROGGS_ carlin++ 13:52
13:53 Ven left 13:54 Akagi201 left
dalek kudo/nom: 78e1fa3 | Carlin++ | src/core/Temporal.pm:
[Temporal] accept colon in the timezone offset

This is valid in ISO 8601
  www.w3.org/TR/NOTE-datetime
  en.wikipedia.org/wiki/ISO_8601#Times		 13:54
kudo/nom: db37b6f | Carlin++ | src/core/Temporal.pm:
[Temporal] accept timezone offsets with only hours

specifying the minutes is optional eg. 2014-07-09T01:50:05+12 is a valid ISO 8601 timestamp
kudo/nom: 597f35f | (Tobias Leich)++ | src/core/Temporal.pm:
Merge pull request #285 from carbin/nom

  [Temporal] accept colon in the timezone offset
13:54 Akagi201 joined, [Sno] left
carlin thanks FROGGS_++ 13:54
I've discovered a new level of laziness 13:55
rather than change my code, patch the compiler to make my code work
FROGGS_ carlin: :o)
13:55 ajr_ left, dmol joined
FROGGS_ carlin: exactly 13:55
13:55 tadzik joined
jnthn I love that kind of lazy :) 13:56
13:57 Akagi201_ joined
psch that reminds me, was there anything missing still for tr/// jnthn? 13:57
(or anyone else, of course :) 13:58
13:58 Akagi201_ left
jnthn psch: Is there a latest PR that needs somebody to review it? 13:58
jnthn missed it if so 13:59
13:59 Akagi201_ joined
psch it's the same PR ( github.com/rakudo/rakudo/pull/284 ), i've changed the p6store to bind/decont as per your latest review and moritz++ mentioned some debug output 13:59
i'm still not sure if the Bool{anon} is a good solution though, but i also don't feel like i can competently judge that :) 14:00
for Str.distance that is
14:01 Akagi201 left 14:06 brrt joined, guru joined 14:07 igorsutton left, guru is now known as Guest52360, Guest52360 is now known as ajr_, igorsutton joined, kaare_ joined 14:11 isBEKaml_mobile joined
[Coke] atroxaper: I wasn't implying a specific pod::sax module, i was trying to imply a SAX like parser as opposed to a DOM style (from XML parsing land) 14:12
14:12 igorsutton left, bluescreen10 joined, MilkmanDan left 14:14 MilkmanDan joined
atroxaper [Coke]: Oh... Yes. Something like that... And there is good name, I think! Can I get it? ;-) 14:14
14:16 btyler joined 14:17 FROGGS_ left
[Coke] you can write it, sure. :) 14:22
14:22 _slade joined 14:25 sqirrel left 14:28 Akagi201 joined 14:30 Akagi201_ left
hoelzro morning #perl6 14:39
14:40 FROGGS joined
psch jnthn: maybe i missed some other feedback? i only remember the p6store leak, which i fixed 14:41
masak DOM-based parsing results in a tree structure in memory. SAX-based parsing results in a lazy list of "events".
there is also at least one hybrid between DOM and SAX. 14:42
timotimo i would strongly vote for a flag that you can use to prevent "automatic recompilation" if your compile process is more complicated than just "compile to bytecode", for example if you do lots of BEGIN blocks 14:43
14:45 rurban1 left 14:47 treehug88 joined 14:53 telex left 14:54 telex joined, virtualsue left
dalek ast: 8cc408b | Carlin++ | S32-temporal/DateTime.t:
[S32-temporal] add tests for more timezone offsets

Per ISO 8601 timezone offsets can be colonated and minuteless		 14:56
atroxaper masak: timotimo: I will show you when finish :-)
14:57 tadzik left
timotimo i'm looking forward to it :) 14:57
i remember my feeble attempts at making the pod stuff better; it only worked partially
atroxaper I'm getting Pod by current grammar. I just transform Pod to something. For example we still can't parse FormattingCode in Tables by grammar :( 15:01
15:03 hoverboard joined
timotimo yes, it's a hard problem 15:03
i threw some doodled code tadziks way for parsing tables "simpler", but it wasn't good enough to get him to the "whole thing"
15:04 virtualsue joined 15:05 raiph joined, SamuraiJack_ left
atroxaper timotimo: yes... I've tried to understand the Grammar for improoving that (yes, i'm too presumptuous), but I couldn't... 15:06
timotimo i don't think that's presumptuous :)
atroxaper :)
raiph didn't someone conclude Pod table syntax requires multi-pass parsing? 15:07
timotimo it *can* be done in one pass 15:08
but it's ... quirky
masak why?
timotimo to implement, anyway
masak oh, because of things like comments above stuff?
prediction: it will be non-trivial to do it in one pass, but it will be worth it. 15:09
timotimo what does "comments above stuff" mean? 15:10
this is only about tables right now
masak I meant something like #| -- I *think* that's the syntax we landed on. 15:11
timotimo ah, that's something else
masak why do tables want to be two-pass? I don't quite see it.
timotimo formatting codes inside tables 15:12
it's not actually quite "two-pass"
well, actually i'm pretty sure i don't have a complete understanding of what "one-pass" vs "two-pass" entails
the problem is that each cell can span multiple rows and formatting codes inside those tables are allowed to span multiple rows inside that cell, too 15:13
and formatting codes can lead to cells being merged, which - i believe/suspect - should influence the other cells in those rows, too
masak ok, I see the challenge/issue now. 15:14
raiph irclog.perlgeek.de/perl6/2013-06-25#i_7247199 # timotimo discusses some table parsing challenges
15:15 denis_boyun_ left
masak raiph: gracias 15:16
15:16 isBEKaml_mobile left
timotimo raiph: the amount of helpfulness per line written to the channel that you have reached astounds me in a really good way 15:16
atroxaper There is annual conversation xD
timotimo :D
hoelzro how can one determine whether a type is Mu? $type eqv Mu doesn't seem to do the trick 15:18
timotimo "is" mu in what way?
hoelzro my $type = Mu;
timotimo ah
hoelzro like the Mu object
dalek kudo-star-daily: e145f55 | coke++ | log/ (14 files):
today (automated commit)
hoelzro well, type object 15:19
15:19 fhelmberger left
timotimo m: my Mu $foo = Mu; say $foo eqv Mu; say $foo.WHAT eqv Mu; say $foo === Mu; 15:19
camelia rakudo-moar 597f35: OUTPUT«Type check failed in binding <anon>; expected 'Any' but got 'Mu'␤ in block at /tmp/y7jLZIOxMn:1␤␤»
masak m: my $x = Mu; say $x.WHAT === Mu
camelia rakudo-moar 597f35: OUTPUT«Type check failed in binding <anon>; expected 'Any' but got 'Mu'␤ in block at /tmp/J8Mo6t1Jq5:1␤␤»
timotimo which little anon is that?
[Coke] m: my $x = Mu; say $x ~~ Mu
camelia rakudo-moar 597f35: OUTPUT«True␤»
[Coke] m: my $x = 234; say $x ~~ Mu
camelia rakudo-moar 597f35: OUTPUT«True␤»
timotimo [Coke]: that will give true for basically anything ever
masak [Coke]: ...but anything will give True on that ;) 15:20
[Coke] that's what i was just checking. :)
psch m: my Mu $x; $x.^mro.elems.say # hackish
camelia rakudo-moar 597f35: OUTPUT«1␤»
hoelzro so far, $type.WHERE == Mu.WHERE works
but it's dodgy
timotimo i realize i'm kind of a topic-nomad ... i go from topic to topic and leave behind unresolved issues and partial implementations of things :S
[Coke] m: my $x = Mu; say WHAT($x)
camelia rakudo-moar 597f35: OUTPUT«(Mu)␤»
timotimo hoelzro: at least Mu never ought to move :)
hoelzro =)
[Coke] since I know that jnthn just fixed that ^^
hoelzro is 'Mu eqv Any' throwing an exception a bug? or should I not expect so much from Mu? 15:24
raiph irclog.perlgeek.de/perl6/2013-08-03 # I'd love to hear masak's take on the Pod tables discussion and solution and problem gists from this day
15:27 atroxaper left
[Coke] here's a not an autopun for masak: "We all have two lives. The second begins when you realize we have only one." 15:28
15:28 atroxaper joined
dalek rl6-roast-data: 8ac5962 | coke++ | / (6 files):
today (automated commit)
 15:28
timotimo bbl 15:29
15:32 atroxaper left
hoelzro if one wanted to do something with an accessor's new value after it's been called, I'm guessing one would have to implement the accessor as 'is rw' and return some sort of container that does something special when storing a value? 15:37
15:37 sjn_roadie joined
raiph github.com/rakudo/rakudo/tree/pod-table # most recent attempt at Pod table parsing? 15:38
15:41 atroxaper joined 15:42 Akagi201 left
dalek ast: da034fd | Carlin++ | S32-temporal/DateTime.t:
[S32-temporal] negative offsets and +/-00 offsets
 15:42
15:44 kaleem left 15:49 thou joined 15:56 dmol left 15:57 virtualsue left 15:59 [Sno] joined 16:01 denis_boyun joined 16:09 denis_boyun left, ajr_ left, ajr joined 16:10 virtualsue joined, ajr is now known as Guest1355 16:11 sjn_ joined, sjn_roadie left 16:13 kaleem joined 16:14 Guest1355 is now known as ajr_ 16:20 isBEKaml joined 16:29 sjn_roadie joined 16:30 sjn_ left 16:32 molaf joined 16:33 rindolf joined 16:41 atroxaper left, dmol joined, atroxaper joined 16:42 brrt left, virtualsue left, Akagi201 joined 16:43 _slade left 16:46 atroxaper left 16:47 hoverboard is now known as cholby 16:48 _slade joined 16:49 kaleem left
hoelzro what's the easiest way in NQP to figure out what type an object is? 16:51
jnthn $obj.HOW.name($obj) 16:52
hoelzro oh perfect 16:53
thanks jnthn
16:53 Akagi201 left 16:56 dakkar left 17:00 cholby is now known as hoverboard 17:01 fhelmberger joined 17:02 treehug88 left 17:05 Rotwang joined 17:07 fhelmberger left 17:10 _slade left, _slade joined 17:15 rindolf left
FROGGS ahh, latest perl6_pass_rates.csv looks really really good :o) 17:19
colomon_ FROGGS: also host07.perl6.com:8080/report
FROGGS hoelzro: and when you want to check that something is of certain type you do: nqp::istype($x, $type)
colomon_: yeah 17:20
though I should perhaps check why Inline is broken
colomon_ FROGGS++
17:24 ajr_ left
carlin FROGGS++ # merging all my pull requests 17:27
FROGGS carlin: thanks for the PR :o)
and carlin++ for contributing!
[Coke] yay, now to fudge/RT the remaining rakudo issues.
PerlJam y 17:28
rurban FROGGS: Is rakudo-p fixed now? 17:36
FROGGS rurban: sort of
17:36 molaf left
rurban I'll check 17:36
FROGGS rurban: it is relying on an optimization, and this happens again 17:37
rurban we need more realistic benchmarks for our new improvements. I doubt that we only got 2%
FROGGS some blocks need to be inlined, otherwise very strange things happen
rurban yes, I saw :)
FROGGS would be awesome if you guys achieve more than 2%, aye :o) 17:38
rurban sure. I'm sceptical, but we only sped up method calls, last time with the WBs we did vtable and method calls 17:43
and we still have to check the arity at run-time. bad specs unfortunately 17:44
17:45 ren1us left 17:46 ren1us joined
ren1us is there any sort of profiler system in place, or at the very least, something that can help me track down where i've apparently got a massive memory leak? 17:46
(honestly i suspect the memory leak is a bug in moar's garbage collector, rather than my code) 17:47
17:48 Alina-malina left 17:52 kaleem joined
jnthn It's extremely unlikley to be a GC bug; it'll only keep objects alive if they're actually referenced. 17:53
Could be some non-gc-managed resource leaking.
ren1us it's very weird... i would suspect it to be my code if the memory usage popped down to something reasonable after my main recursive stuff hit the base case, but it doesn't. it just constantly increases over time 17:55
17:56 Alina-malina joined
jnthn Either way, there's not some good profiling support yet. 17:57
ren1us unfortunate. i'll try running through parrot just to see what happens.
17:57 Akagi201 joined 17:58 guru joined
jnthn dinner; bbiab 17:58
ren1us because there's really no reason for moar to be getting up to about 4.5 gigs of memory in half an hour
17:58 guru is now known as Guest31724 17:59 Guest31724 is now known as ajr_
jnthn Well, depends what you're doing, but no, that does sound like a leak of some kind. I just suspect it'll turn out not to be the GC. The last time I fixed leaks, they were of stuff allocated along the way. 18:00
(At C level)
dinner; bbiab
18:02 Akagi201 left 18:04 isBEKaml_ joined 18:05 isBEKaml left 18:07 isBEKaml_ is now known as isBEKaml, isBEKaml left, isBEKaml joined 18:10 erkan joined, Akagi201 joined, Akagi201 left
ren1us wow what's with me and breaking literally everything. hooray for a segfault. 18:11
18:15 isBEKaml left 18:19 teodozjan joined 18:20 kurahaupo joined 18:22 kaleem left
TimToady cool, I get bubble+cryosurgery in half an hour for a detached retina 18:31
may or may not be visually active during that time...we'll see
ren1us that sounds really unpleasant and painful. best of luck, get better soon, etc etc
TimToady I get to sit in a funny position for five days to keep the bubble in the right place...
nwc10 you keep living in interesting times :-(
I hope it goes well 18:32
TimToady well, in other times I'd've been dead by now several ways :)
so I'll take interesting :)
afk &
jnthn TimToady: Hope all goes as well as possible. Take care.
colomon_ TimToady: best wishes! 18:33
[Coke] interesting++
retupmoca TimToady: good luck, I hope it goes well 18:34
[Coke] best wishes. 18:35
hoelzro FROGGS: oh, thanks! that's good to know 18:44
timotimo best wishes from me, too! that does sound like an amazing feat 18:50
[Coke] plays the latest guardians of the galaxy trailer on loop. 18:51
timotimo oh, so there's a gas bubble inside the eye where the "eye fluid" also is?
i mean: a gas bubble is put there
five days is a long time to keep your head in a particular position ... 18:52
woolfy Oh man, detached retina... interesting times and all are fun, but this is no fun. Best wishes. TimToady++ 18:56
PerlJam seems to recall that this is not the first time TimToady has had a detached retina. :-( 18:57
timotimo is there something that predisposes you to get that? 18:58
lizmat having had it before? in any case, TimToady++ for hanging in there and not giving up on us 19:00
FROGGS what the...
TimToady: I wish you all the best!
hoelzro TimToady: godo luck! 19:01
*good
19:03 SamuraiJack_ joined 19:08 treehug88 joined 19:09 darutoko left
timotimo i wonder how some surgeries are ... "discovered" 19:13
19:13 bluescreen100 joined
PerlJam After trying it out on cadavers for a while, eventually someone says, "Now let's try it on a live one!" 19:13
Mouq TimToady++ 19:14
19:17 bluescreen10 left
masak TimToady: best wishes. 19:18
19:22 sftp_ joined
Mouq m: class A { method @.[*@abcd] { say "ok @abcd[]" } }; say A.postcircumfix:<[ ]>(42) # can we have this sugar for .at_pos instead? same with %.{...} 19:22
camelia rakudo-moar 597f35: OUTPUT«ok 42␤True␤»
19:22 SamuraiJack_ left
lue .tell atroxaper just so you know, I've already followed that hint in Pod::To::HTML (here it is: github.com/lue/pod-walker), and AFAIK Mouq is using it to reimplement Pod::To::HTML too. I'll be interested in seeing what your solution to that hint is! :) 19:23
yoleaux lue: I'll pass your message to atroxaper.
19:23 sftp left, sftp_ is now known as sftp
Mouq lue: So, on that topic, what do you think of making Pod::To::HTML actually act like a class instead of a module? and maybe having Pod::Walker be a role it inherits? 19:24
lue Mouq: perhaps, I couldn't say off-hand what route would be best.
Mouq lue: Ok, I'll try it out. Pod::To::HTML uses a lot of global variables for each document, it seems like some sort of encapsulation somewhere would be for the best 19:26
lue Sounds good. :) 19:27
19:30 bluescreen100 left 19:31 itz_ joined 19:34 cognominal joined
Mouq spectests @.() -> at_pos(), %.() -> at_key() 19:37
19:41 hoverboard left 19:42 Akagi201 joined, hoverboard joined 19:46 Akagi201 left
timotimo "Over time the vitreous changes, shrinking and developing pockets of liquefaction, similar to the way a gelatin dessert shrinks, or detaches, from the edge of a pan over time." - this is a wonderful simile 19:48
19:50 xdoctor joined, teodozjan left 19:51 Akagi201 joined
FROGGS timotimo: in case you're working on the weekly: jnthn (and me) have tracked down the perl6-p borkage that introduced 1800 spectest fails 19:52
dalek kudo/nom: 548a9b8 | (Elizabeth Mattijsen)++ | src/core/CompUnitRepo/Local/File.pm:
Add :no-precomp flag

This should allow for searches for a non-precompiled version of a module.		 19:55
19:55 Akagi201 left
ast: 3aa17a3 | (Elizabeth Mattijsen)++ | S22-package-format/local (3 files):
Some more CURL::File tests
 19:56
20:06 cognominal left
Mouq Hmm... My spectest turns up fails for S32-list/pick.t and S32-temporal/local.t 20:06
20:07 orevdiabl left
timotimo i still didn't have an opportunity today, but i think i'll be able to do it in the next 2h 20:07
20:07 revdiablo joined 20:14 japhb joined
Mouq (i may just have a stale Moar, but the dog is insisting I take her for a walk before I find out) 20:15
itz_ Brazil is still in with a chance! 20:25
:)
carlin that S32-list/pick failure is impressive 20:26
*** Error in `/home/carlin/rakudo/install/bin/moar': free(): invalid size: 0x00000000088017e0 ***
FROGGS scary, aye 20:28
20:31 raiph left 20:37 kurahaupo left 20:38 anaeem1 joined 20:39 Alula_ left, kurahaupo joined, kurahaupo left, Alula_ joined
Mouq Oh? So it's not just me :) 20:40
20:41 kurahaupo joined, kurahaupo left
carlin pick.t is giving me fireworks, it's very spectacular 20:43
dalek kudo/nom: fda3461 | Mouq++ | src/Perl6/Actions.nqp:
Use at_{pos,key} for method {@,%}.()
Mouq Oh, wait 20:45
timotimo waits 20:46
Mouq perl6 -e'class A { method %.(*@abcd) { say "called" } }; say A.{42, 43}'
called
(should only be called once, reverting my commit)
jnthn uh, no 20:47
at_key is called once per key to access
20:47 kurahaupo joined
vendethiel thinks it's kinda confusing it's not postcircumfix:<{ }> :o) 20:47
well, there's <> to take into account I guess 20:48
jnthn That's a sub, not a method, like all the other operator-y things. :)
(used to be a method, but consistency and optimizability won over)
vendethiel right-o
yeah, I think somebody already told me 'bout that, subs for inlining
Mouq jnthn: Yes, so having %.() call at_key creates the wrong behaviour 20:49
jnthn Mouq: Bit tied up with $dayjob so didn't read the commit, but I'd gueesed you were desugaring %.() to at_pos() 20:50
uh, at_key
20:51 hoverboard left
dalek kudo/nom: 41574ab | Mouq++ | src/Perl6/Actions.nqp:
Revert "Use at_{pos,key} for method {@,%}.()"

This reverts commit fda34618fa69e331bfc1ea44785cf4c40843b193.		 20:51
20:51 Akagi201 joined
Mouq "@.[]" really implies the whole postcircumfix, rather than just a single-element access method 20:52
jnthn OK, well...maybe they don't make sense to exist any more, then. 20:53
Mouq Is 'method postcircumfix:<[ ]>' never coming back? 20:54
FROGGS kinda unlikely I guess
jnthn I don't believe so.
If you really wanted to do custom slicing then you'd need to write a sub 20:55
tbh, though, I really do not think the sugar is worth it.
It's quite a special case.
20:56 Akagi201 left
Mouq I only just realized that you can only define new postcircumfixes in a sub 20:56
But what about the inconsistency of 'method postcircumfix:<( )>'
zengargoyle spent a few hours trying to get a class attribute to «handle postcircumfix<{ }>» something when it seems it should have been «handle <at_key» 20:58
jnthn Mouq: I think we should rename that INVOKE or something
Or jsut invoke
20:58 silug left
masak +1 20:59
21:01 treehug88 left
moritz same capitalization as at_key, IMHO 21:11
21:11 silug joined
timotimo zengargoyle: i think you need to "does Associative" 21:12
21:19 ivanshmakov left 21:22 ivanshmakov joined
zengargoyle timotimo: tried that on the class at least. rakudo/src/core/Associative.pm role has a single method of(). 21:28
timotimo oh 21:31
oops
yeah, that's a different kind of thing; i think that's the role we use to apply type restraints to hashes 21:32
21:32 gfldex left
zengargoyle i really thought: class A does Associative { has SetHash $!h handles Associative = SetHash.new } # would have DWIMd 21:32
and grepping the specs for clues didn't help much. :P 21:33
timotimo handles is only for methods ;( 21:37
zengargoyle is that NYI? spec has things like: has $.payload handles <Str Numeric>; 21:39
and: has $.fur is rw handles Groomable; 21:40
21:40 anaeem1 left 21:41 anaeem1 joined
zengargoyle i wanted to do something like: my Flags $flags = Flags.new(one=>0x1,two=>0x2,three=>0x4); $flags<one>=True; $flags<three>=True; # $flags.Int == 5 21:43
where the keys are restricted to the enum of valid flags. 21:44
psch doesn't an enum on its own do that already? 21:45
zengargoyle enum is immutable. 21:46
21:46 anaeem1 left
zengargoyle was thinking: class Flags { has EnumMap $!em; has SetHash $!sh; # enough logic to do stuff } 21:46
psch oh 21:47
i guess the accessors could get some logic to restrict, but that's clunky
zengargoyle where EnumMap holds keys=>values; and SetHash holds the desired subset of available flags.
psch github.com/perl6/roast/blob/master...09-day11.t # here's handles examples, btw 21:48
zengargoyle but wanted $flags to behave just like a SetHash would, $flags.keys, etc.
psch so each kv from the SetHash should be one kv from the enum? 21:49
retupmoca discovers trello: trello.com/b/8vpgVKdK/retupmoca-s-...-6-modules
psch the use is something like "which flags are of the available flags are set for this instance" i figure
zengargoyle SetHash hides the values, all set keys are True, keys set to False are deleted.
21:50 dayangkun joined
zengargoyle so it's like: em = (one=>1,two=>2,three=>4), sh = (<one four>) 21:51
'tho underneath it may be sh = (one=>true_value,four=>something_true)
if you ask $sh<one> you'll only get True back. 21:52
21:52 Akagi201 joined
zengargoyle i finally gave up because i was only going to support *one* of the available flags. :P 21:53
psch i don't think i fully understand what you're trying to do :) 21:54
(not to mention i might not even be able to help)
zengargoyle named bitflags.
21:54 dayangkun_ joined
psch on the surface exporting the enum and having an &addflags(*@) method that checks if the supplied arguments are in the enum and throws otherwise seems enough to me 21:55
21:55 itz_ left
psch although i think not-from-the-enum barewords would throw on their own in the mainline anyway 21:55
which makes it not really desireable...
where "it" is "this way of building it" 21:56
zengargoyle well yeah, but not as pretty as. :P
psch right :)
21:56 Akagi201 left
psch i guess you could build an accessor for $!sh that checks keys that are to be added for some kind of equality with the enum keys 21:56
which might be Stringy equality or something else 21:57
zengargoyle yeah, i had the basics working, just not the make it look like a plain SetHash parts.
21:58 dayangkun left 22:03 Rotwang left 22:14 kaare_ left 22:28 wamiks joined 22:29 _slade left 22:32 jnap1 joined 22:34 jnap left 22:36 ajr_ left 22:38 hoverboard joined 22:41 grondilu joined, pdcawley left
grondilu r: constant primes = grep &is-primes, 2 .. *; # this times out locally on Moar 22:42
camelia rakudo-jvm 41574a: OUTPUT«(timeout)»
..rakudo-{parrot,moar} 41574a: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Undeclared routine:␤ &is-primes used at line 1. Did you mean '&is-prime'?␤␤»
grondilu r: constant primes = grep &is-primes, 2 .. *; say primes[^10] # but this does not
camelia rakudo-{parrot,jvm,moar} 41574a: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Undeclared routine:␤ &is-primes used at line 1. Did you mean '&is-prime'?␤␤»
grondilu r: constant primes = grep &is-prime, 2 .. *; say primes[^10] # but this does not
camelia rakudo-{parrot,jvm,moar} 41574a: OUTPUT«2 3 5 7 11 13 17 19 23 29␤» 22:43
grondilu r: constant primes = grep &is-prime, 2 .. *;
camelia rakudo-{parrot,jvm,moar} 41574a: OUTPUT«(timeout)»
grondilu see?
r: constant integers = 1 .. *; 22:44
camelia ( no output )
grondilu r: constant even-numbers = grep * %% 2, 1 .. *;
camelia rakudo-{parrot,jvm,moar} 41574a: OUTPUT«(timeout)» 22:45
grondilu r: constant even-numbers = grep * %% 2, 1 .. *; say even-numbers[^10]
camelia rakudo-{parrot,jvm,moar} 41574a: OUTPUT«2 4 6 8 10 12 14 16 18 20␤»
Mouq r: constant primes = grep &is-prime, 2 .. *; 1 22:46
camelia ( no output )
grondilu it's like the compiler is bored of having nothing to do so it decides to evaluate the lazy list completely :) 22:47
Mouq lol 22:48
22:49 pdcawley joined
psch m: constant n = 1 .. Inf # apparently the compiler feels teased by the grep 22:50
camelia ( no output )
psch m: constant n = 1 .. * # ... i think? 22:51
22:51 btyler left
camelia ( no output ) 22:51
22:52 paraxor left 22:53 Akagi201 joined, dayangkun_ left
grondilu r: constant even-numbers = grep * %% 2, 1 .. *; say even-numbers.elems 22:53
well it can not know the grep will return an infinite number of elements, I guess 22:54
camelia rakudo-{parrot,jvm,moar} 41574a: OUTPUT«(timeout)»
grondilu so I'm actually not sure about the lazyness of a constant defined with grep. I supposed I should read the spec about these stuff. 22:55
22:57 Akagi201 left 23:02 dmol left 23:25 mr-foobar left 23:31 spider-mario left 23:37 broquaint left
jnthn Yay. Got a bunch of stuff off my plate and tomorrow I can spend all day on Perl 6 things. \o/ 23:38
psch jnthn++ # finish his plate for good weather 23:42
eh, *finishing
as in, weather suitable for butterflies :) 23:43
23:48 xragnar left
psch is still trying to find places where the outer chain could break... 23:50
but no more today i think
23:50 xragnar joined 23:51 SevenWolf joined 23:52 rurban1 joined 23:54 Akagi201 joined
psch g'night o/ 23:55
23:55 sjn_roadie left 23:57 psch left 23:58 Akagi201 left

Powered by Cro and the Raku® Programming Language.

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

Thank you!