»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.
TimToady skids++ 01:31
thou I'm making a tool to recursively clone module dependencies into a directory. I have a start, but I feel the grammar is still a bit clumsy. In particular, the actions for the pkginfo and info regexes are awkward. Is there a better way to write them? gist.github.com/softmoth/3bc7bd31ca65a5f7b3df 01:46
d^_^b paste.scsys.co.uk/368160 having issues building nqp for the jvm 01:59
raiph d^_^b: looks like you need to specify a memory option with more memory. 02:02
have you built it successfully before?
d^_^b raiph: would that be an option for passed to the configure script? nope 02:05
raiph searching clogs atm for the right reference 02:06
d^_^b raiph: I think I found what fixes it. 02:15
testing now
raiph cool, hadn't yet found what i was looking for
d^_^b it is in the readme.md actually 02:16
:D
raiph heh 02:17
d^_^b -Xms500m -Xmx2g as options passed to java
well tried that but that didn't work
d^_^b well I editted the nqp-j but I don't know if I did it right. 02:20
paste.scsys.co.uk/368179 02:21
raiph here's the best I've found from #perl6: 02:23
irclog.perlgeek.de/perl6/2013-11-02#i_7803667
irclog.perlgeek.de/perl6/2014-04-23#i_8629112 02:24
raiph did you still get an out of heap space error when using -Xms1000m ? 02:28
d^_^b: ^^
d^_^b yup
raiph is this while trying to build the new rakudo star? 02:30
d^_^b well I edited the file, but I don't see the commands being passed.
timotimo d^_^b: you most probably need to change it in your Makefile or even the Makefile.in from the tools/build/ folder 07:40
dalek ast: 0ab4c71 | (Elizabeth Mattijsen)++ | S17-supply/classify.t:
Also test Supply.classify( @mapper )
 07:41
timotimo otherwise the makefile will say "hey this file is out of date! i'll update it real quick" 07:45
or something like that
timotimo tadzik: i'm a bit annoyed that there's no good single repository from where i can grab steroids and improve upon it 08:01
and there's no history :( 08:02
tadzik timotimo: there's github.com/tadzik/steroids now 08:03
that's the newest iteration
I know, I screwed up a little
timotimo :) 08:06
FROGGS tadzik: you are bad! and you should feel bad! # 6gvts.wordpress.com/ 08:07
I still love that site :o) 08:08
tadzik :D 08:11
the existence of this website is amusing a bit
ABOUT 08:12
You probably suck.
FROGGS *g*
tadzik meh, I expected at least a dog
* doge
FROGGS hmm, translation for "doge" is "Doge"... not very helpful 08:13
lizmat "the chief magistrate of Venice or Genoa." 08:14
?
FROGGS ha! there is an easter egg about "doge meme" on youtube :D 08:15
but I see now that this is meant: www.google.de/search?q=doge+meme
tadzik yes, I'm a fan 08:19
our office is full of those :D
FROGGS /o\ 08:20
tadzik some of them hand-drawn 08:21
i.imgur.com/uK5SJOH.jpg 08:22
FROGGS -.- 08:24
:P
I only have that on my office door hezik.nl/wp-content/uploads/2012/03/ssh.jpg
tadzik :D 08:25
we have images.fineartamerica.com/images-me...rcell.jpg, but with "FIX EVERY COMPILER WARNING" at the bottom 08:26
timotimo tadzik: btw, $k eq any("Up", "Space") is much less optimized in rakudo than $k eq "Up" | "Space" is
tadzik oh, good to know 08:27
timotimo though really it should be doable to see that it's a list of constant strings and thus "non-dangerous"
and the same optimization could perhaps kick in here, too
m: say Int(5.7) 08:28
camelia rakudo-moar bd518a: OUTPUT«5␤»
timotimo oh, duh.
FROGGS *g* 08:29
add 0.5 :o)
timotimo i was about to say "i didn't know this exists!" 08:32
but i suppose it isn't that surprising
vendethiel m: class A { has $!a; submethod BUILD($!a = 5) {}; method p { say $!a; } }; A.new.p 08:36
camelia rakudo-moar bd518a: OUTPUT«5␤»
vendethiel tadzik: couldn't you use that here ? github.com/tadzik/steroids/commit/...77d5b79R17
tadzik oh, I could :) 08:40
vendethiel PRs 08:41
tadzik I don't quite trust BUILD generally 08:42
it's weird
vendethiel why are you using nqp::bindattr ?
(in add_state) 08:43
FROGGS is it a CStruct?
tadzik I probably have to
:>
if it works on all backends without it, then my fault
vendethiel probably so 08:44
(that it's a CStruct, not that it's your fault, that is ..)
tadzik I may have just been wrong though
the thing being assigned (bound) is a cstruct, yes
FROGGS has Mu $!game;
I guess that bindattr is sane here 08:45
vendethiel I had to do a SDL game for school in C, and it was a bit painful. Animation abstractions seems good.. 08:46
tadzik they're a joy to use :) 08:47
(for me)
vendethiel Yeah, though the game I did was a map-based RPG, so maybe I'll create a class for a scrolling map
vendethiel with some nice DSL to create the map itself, maybe ? 08:48
tadzik aww yiss
the map could be loadable from a JSON-y thingy perhaps 08:49
in the other game I did a fun map, hold on
vendethiel tadzik: github.com/Nami-Doc/midgar/blob/master/map.txt that's how I did it in C :P
FROGGS I've got a top secret P5 thingy laying around about 2d maps and stuff... (secret of mana like)
tadzik github.com/tadzik/mario/blob/master/level
vendethiel hahaha
tadzik then somewhere else there's a mapping: s is sky, B is brick, g is grass
vendethiel I guess we did the same thing :>
except I just read the size
tadzik lol
FROGGS I wrote a proper map editor :P 08:50
with a palette of the tiles etc
tadzik nice :)
vendethiel FROGGS: don't get me wrong - I did so too : github.com/Nami-Doc/midgar/tree/ma...mapcreator
tadzik I didn't :O
FROGGS and the game is SDL+box2d+P5
tadzik anyway, I think maps are fine this way, then before you write the editor you can use a text editor :P
dialogues, thoguh, are probably better off with a DSL 08:51
because some of dialouge options want to be conditional, depending on quest status, items etc
FROGGS tadzik: I created a map editor to attach properties to tiles, like triggers
tadzik nice
FROGGS so you can stand before a sign and get its message displayed
damnit, now I want to work on that instead :/ 08:52
tadzik FROGGS: static.comicvine.com/uploads/origin...up.%2B.jpg 08:52
FROGGS *g*
tadzik so basically my plan was, after being done with gamestates (which means submitting a proper bugreport for nativecall/moar), to start building nice stuff on top of the foundation that there is 08:53
a nice menu generator, a thingy for tiled maps etc
if somebody jumped on it, that'd be super awesome
vendethiel tadzik: I'll probably try the tiled map thingie 09:00
tadzik vendethiel: \o/
vendethiel tadzik: is there something in steroids to deal with spritemaps ? 09:08
timotimo not yet, no
vendethiel tadzik: since I used that : github.com/Nami-Doc/midgar/blob/ma...prites.jpg
tadzik vendethiel: yes
well, to some extend
vendethiel with my hackingly hack : github.com/Nami-Doc/midgar/blob/ma...rite.c#L12
tadzik ok, no then :) 09:09
timotimo yikes.
vendethiel timotimo: you know you like it
tadzik what is supported is things like this: github.com/tadzik/steroids/blob/ma...nvader.png 09:10
timotimo i once made something fun where i arbitrarily defined that all sprites are equally sized and i put a list of names into the comment field of the svg, which ended up in the png as a comment and then i could load the png and at the same time parse out the comment tag to get the names
oh, you have support for animated stuff?
tadzik you tell steroids "there's images in there, 48x32", and it figures out how many are there and how to divide that
yes
vendethiel it "figures out" ? I see "Int $framecnt" 09:11
tadzik ok, maybe it doesn't figure it out :D 09:12
oh yes, it can't really
because in that example there's 7 of them, not like the size would indicate
vendethiel yeah, for the empty case
tadzik so you do self.load_spritesheet('invader', 'assets/invader.png', 72, 32, 7);
and then you can animate that or whatnot
FROGGS I know player sprites where you have 4 columns for the four walking directions 09:13
tadzik yes
FROGGS and the animations are the rows...
tadzik you can specify which frames go for an animation 09:14
so: self.add_animation($!player, 3..8, 200, True); or something
FROGGS that is what P5's SDLx::Sprite::Animated also allows I think
tadzik that's frames from 3 to 8, every 200ms, looped
FROGGS ohh, that will be nice in P6 09:15
tadzik it already is :)
FROGGS you can supply a range like that, or a list if the source image is not ordered well
:o)
tadzik exactly
FROGGS yeah, I have to play with it
timotimo someone on twitter objected to the periodic table of the operators :\
i kinda like it
vendethiel still 8 hours before age of mythology EX is playable :\ 09:18
tadzik oh, today is d-day too :D 09:19
vendethiel "project steroids not found in the ecosystem" 09:21
tadzik it's knot
I don't think it should be there (yet)
vendethiel how can I depend on it if it's not here ? 09:22
tadzik download it to your project directory, the lib/
vendethiel ew :( 09:26
FROGGS vendethiel: it will break your code anyway when it is in that early state... so better but a snapshot into lib/ as tadzik++ said 09:29
vendethiel sure, I expect stuff pre-1.0 to break
tadzik yeah, I break stuff all the time now 09:48
I realize my mistakes :)
it may be fun to try running Through the window on the latest engine 09:49
Woodi m: $t = Supply.interval(1).tap(&say); 10:13
camelia rakudo-moar bd518a: OUTPUT«===SORRY!=== Error while compiling /tmp/jp6A9Pd61C␤Variable '$t' is not declared␤at /tmp/jp6A9Pd61C:1␤------> $t⏏ = Supply.interval(1).tap(&say);␤ expecting any of:␤ postfix␤»
Woodi m: my $t = Supply.interval(1).tap(&say); 10:14
camelia ( no output )
timotimo Woodi: these interval taps won't keep your application alive
m: my $t = Supply.interval(1).tap(&say); sleep 5; 10:15
camelia rakudo-moar bd518a: OUTPUT«0␤1␤2␤3␤4␤5␤»
timotimo m: my $t = Supply.interval(0.1).tap(&say); sleep 1; $t.close; sleep 1;
camelia rakudo-moar bd518a: OUTPUT«0␤1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤»
Woodi timotimo: right, good to know. but I was just checking moar for .interval - start.04 saya NYI... 10:22
jnthn yes, 2014.04 didn't have the async bits 10:24
Woodi anyway, I think API is a bit strange for publish/subscribe service... sending something while getting with .more(...) is strange... 10:25
I expecting Supplies are inter-threads pipes and one end is producing things... 10:26
jnthn No
Woodi my thinkins is degenerated (case) ? :)
jnthn In fact, when you more something, it keeps to the same thread unless it has to. 10:27
uh, unless it has to change
jnthn It's just setting up a pipeline that values can be pushed through asynchronously. 10:27
timotimo Woodi: what you're thinking of is probably Channels. 10:29
Woodi hmm, .more isn't receiving end ?
timotimo nope 10:31
you use .more to "send" a value
and the recipients get invoked, so there is no ".receive" or similar
Woodi hmm :) 10:33
timotimo but i think there is a coercer from a supply to a channel
jnthn Think of it like a lazy list, but instead of the thing at the end pulling the next value through the pipeline of maps/greps etc, instead the asynchronously produced values are pushed through that pipeline to the tapper at the end. 10:34
Woodi probably my brain is damaged with (light lecture of) flow based programming and I want to have many stages of data processing everywhere... 10:36
so Supplies just construct parallel receivers ? 10:37
nvm, I have .Net book with reactive chapter, will consult it... 10:39
timotimo not parallel 10:42
unless you specifically make it multi-threaded the tappers will just be called in order
Woodi btw. some loops are implemented via map and in some Advent post I read map is implemented with gather. still is ? becouse lastly someon say gather is slow... 11:21
FROGGS map is implemented using MapIter
Woodi looking... 11:22
Woodi looks nqp::while or even assembly code :) 11:27
jnthn MapIter itself is quite hand-optimized. 11:28
lizmat m: my @m = ( $(), $(1,) ); .WHAT.say for @m 11:59
camelia rakudo-moar bd518a: OUTPUT«(Str)␤(Parcel)␤»
lizmat So how *do* I get an empty parcel into an array ? 12:00
moritz ().item
$() is special-cased to go into $/
lizmat aha.... ok
moritz++ 12:01
oddly enough, this doesn't work either: 12:02
m: my @m = ( Parcel.new, Parcel.new.item ); .WHAT.say for @m
camelia rakudo-moar bd518a: OUTPUT«(Parcel)␤»
lizmat m: my @m = (Parcel.new); say @m.elems' 12:03
camelia rakudo-moar bd518a: OUTPUT«===SORRY!=== Error while compiling /tmp/mkWHclIavZ␤Two terms in a row␤at /tmp/mkWHclIavZ:1␤------> my @m = (Parcel.new); say @m.elems⏏'␤ expecting any of:␤ method arguments␤ postfix␤ …»
lizmat m: my @m = (Parcel.new); say @m.elems
camelia rakudo-moar bd518a: OUTPUT«0␤»
lizmat was a bit of a WAT to me
FROGGS m: my @m = ( $(,), $(1,) ); .WHAT.say for @m 12:08
camelia rakudo-moar bd518a: OUTPUT«===SORRY!=== Error while compiling /tmp/FzwAhigGCp␤Preceding context expects a term, but found infix , instead␤at /tmp/FzwAhigGCp:1␤------> my @m = ( $(,⏏), $(1,) ); .WHAT.say for @m␤»
FROGGS >.<
jnthn lizmat: Parcel.new makes you a *flattening* Parcel 12:13
Like Array.new does with Array 12:14
lizmat so you would need Parcel.new.item
jnthn m: my @m = ( Parcel.new, Parcel.new.item ); say @m.elems
camelia rakudo-moar bd518a: OUTPUT«1␤»
jnthn There, the Parcel.new flattened away as expected, and the .item'd one remained. 12:14
lizmat I guess my main comfusion came from $() not being Parcel.new.item 12:15
jnthn so yes, .item
Ah, yeah...that's a...curious discontinuity I'm not entirely sure pulls its weight.
r: say $(,).WHAT # curious 12:16
camelia rakudo-{parrot,jvm,moar} bd518a: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Preceding context expects a term, but found infix , instead␤at /tmp/tmpfile:1␤------> say $(,⏏).WHAT # curious␤»
jnthn ah, of course :)
moritz r: say $( ).WHAT
camelia rakudo-{parrot,jvm,moar} bd518a: OUTPUT«(Parcel)␤»
jnthn oh my :)
r: my @a = $( ), $( ); say @a.elems
camelia rakudo-{parrot,jvm,moar} bd518a: OUTPUT«2␤»
jnthn hehe :)
lizmat wow, that whitespace sensitive ? 12:17
jnthn I'm...not quite sure $() and $( ) meaning a different thing is quite what the spec had in mind... :)
lizmat m: my @m = ($( )); say @m[0].WHAT
camelia rakudo-moar bd518a: OUTPUT«(Parcel)␤»
lizmat so it even doesn't flatten with the space 12:18
jnthn Right, that's what my .elems thing showed above 12:18
FROGGS jnthn: I'm sure I asked TimToady about '$()'/'$( )' before implemented only '$()' as the shortcut for $/.ast
jnthn FROGGS: Yeah; but this makes me wonder if $() as a shortcut for $/.ast is really wise. 12:19
FROGGS jnthn: I liked the idea when I implemented it :o)
I'm not so sure anymore though
vendethiel it does look surprising :-) 12:21
lizmat it was quite a bit of a WAT for me :-) 12:23
masak hi-from-a-train, #perl6 12:32
nwc10 good moving, masak 12:34
lizmat
.oO( he could be standing still )
masak: o/ 12:35
jnthn lizmat: Given he's probably on a train in Sweden, it could well be the case :P
masak surprisingly, the train is moving. forwards, even. 12:36
nwc10 jnthn: the trains in Sweden are happy reminders of home? :-)
colomon jvm issue blogged: justrakudoit.wordpress.com/2014/05/...weirdness/
lizmat masak: that's just a matter of orientation
colomon I wonder how many of the 30+ JVM-only module fails are because of this issue? 12:37
lizmat masak: is it going in the right direction?
masak lizmat: yes, because "away from work, towards home" is inherently right.
jnthn lizmat: It's going "in the direction of travel" :) 12:38
(Context: all announcements like "the platform will be on the left" are always qualified with "in the direction of travel" - 'cus apparently that's not obvious :))
dalek kudo-star-daily: fe49b71 | coke++ | log/ (5 files):
today (automated commit)
kudo-star-daily: 52fdef8 | coke++ | log/ (5 files):
today (automated commit)
rl6-roast-data: 3542cca | coke++ | / (4 files):
today (automated commit)
 12:39
kudo/nom: 286ead8 | (Elizabeth Mattijsen)++ | src/core/Supply (2 files):
Implement Supply.categorize
 12:41
ast: d3aaaf2 | (Elizabeth Mattijsen)++ | S17-supply/categorize.t:
Add tests for Supply.categorize
masak I hate that qualification. it turns an already unambiguous phrase into a longer but no less ambiguous one. 12:42
lizmat like it would be clear from a stationary train what the direction of travel is supposed to be, right? 12:43
jnthn Me too (this is the end of the sentence).
Woodi it's still better then a voice from megaphone announcing: Train from Koluszek to Koluszek... wtf I'm talking!!??
jnthn I don't know what they do it (it's a pondering).
masak just found en.wikipedia.org/wiki/Fumblerules and thinks jnthn would like those :) 12:47
dalek ecs: a375266 | (Elizabeth Mattijsen)++ | S17-concurrency.pod:
Spec Supply.categorize
kudo/nom: 3ba24fc | (Elizabeth Mattijsen)++ | t/spectest.data:
Make sure we test Supply.categorize
 12:50
kudo/nom: e64f6ef | (Elizabeth Mattijsen)++ | docs/ChangeLog:
Announce Supply.(classify|categorize)
jnthn rules for...fumbles? :P 12:51
jnthn looks
lizmat als: Muphry's law: en.wikipedia.org/wiki/Muphry%27s_law
jnthn hah :)
.oO( Don't accidentally verbs. )
 12:52
lizmat and don't stop mid 12:54
lizmat stops messing with Supplies for now 12:55
unless someone tells me I forgot / broke something 12:56
jnthn Put a article in a right place.
colomon: I think that the issue is that something is going on with importing symbols, meaning ABC::Note is accidentally considered a sub call, not a type name, which leads to the error you see.
lizmat j: $_ = IO::Path; say "foo" when IO::Path 12:58
camelia rakudo-jvm bd518a: OUTPUT«foo␤»
lizmat j: $_ = Test; say "foo" when Test 12:59
camelia rakudo-jvm bd518a: OUTPUT«===SORRY!=== Error while compiling /tmp/SbxGnZqpXB␤Undeclared name:␤ Test used at line 1␤␤»
lizmat j: $_ = Test; say "foo" when Test; use Test
camelia rakudo-jvm bd518a: OUTPUT«===SORRY!=== Error while compiling /tmp/JsE0mQbo3n␤Undeclared name:␤ Test used at line 1␤␤»
lizmat m: $_ = Test; say "foo" when Test; use Test 13:00
camelia rakudo-moar bd518a: OUTPUT«===SORRY!=== Error while compiling /tmp/2oBaJywLqc␤Undeclared name:␤ Test used at line 1␤␤»
jnthn lizmat: I suspect you need a more complex import chain to make it happen
lizmat j: use Test; $_ = Test; say "foo" when Test
camelia rakudo-jvm bd518a: OUTPUT«foo␤»
jnthn lizmat: And probably nested packages 13:01
lizmat yes, probably :-(
colomon jnthn: right, I figured it must be something like that. 13:12
colomon jnthn: just wanted to make sure this was on people's radar. :) 13:15
dalek kudo/nom: 6a97d13 | (Elizabeth Mattijsen)++ | src/core/Parcel.pm:
Emphasize specialness of "$( )" vs "$()"
 13:33
Woodi hmm, 'exit 0' is heavy op probably :) 13:33
masak looks at 6gvts.wordpress.com/ and feels a strange sense of pride 13:48
hm, do I boost the google juice of that site every time I link to it in the backlog?
jnthn hahaha... 13:49
You know you're doing something right when your blog ends up with a parody version of itself :P
masak with Zoidberg, even! 13:50
nwc10 is the parody going to get any new posts?
masak jnthn: be sure to check the "About" section.
nwc10: one can hope.
nwc10: but it'll be hard to top :)
jnthn :D
dalek ecs: 5262e8c | (Elizabeth Mattijsen)++ | S02-bits.pod:
Add paragraph about the Empty Parcel.
 14:10
lizmat afk for a bit& 14:12
vendethiel (do we really want that $() meaning $.made ?)
vendethiel I think there's a small typo, as well 14:13
dalek ecs: 7e406c5 | Nami-Doc++ | S02-bits.pod:
Fix typo : `$.made` -> `$/.made`
vendethiel oh whoops
i meant to submit a PR ...
lizmat well, I'd be surprised about the typo, as it was copy n pasted 14:14
vendethiel lizmat: isn't that a typo ?
lizmat isn't $.made short for $/.made ?
vendethiel m: class A { has $.made; p { 'a' ~~ /a/; say $.made; } }; A.new.p; 14:15
camelia rakudo-moar e64f6e: OUTPUT«===SORRY!=== Error while compiling /tmp/I27hxPE4g7␤Variable $.made used where no 'self' is available␤at /tmp/I27hxPE4g7:1␤------> { has $.made; p { 'a' ~~ /a/; say $.made⏏; } }; A.new.p;␤ expecting any of:…»
vendethiel m: class A { has $.made; method p { 'a' ~~ /a/; say $.made; } }; A.new.p; 14:15
camelia rakudo-moar e64f6e: OUTPUT«(Any)␤»
vendethiel lizmat: I don't think so. That'd be really, really, really surprising
lizmat then the typo is in there twice 14:16
vendethiel is it ?
oh, I see it now
dalek ecs: f5d616e | (Elizabeth Mattijsen)++ | S02-bits.pod:
Fix typo spotted by vendethiel++
 14:17
dalek kudo/nom: 86db16f | (Elizabeth Mattijsen)++ | src/core/Parcel.pm:
Fix typo spotted by vendethiel++
 14:18
lizmat really afk now& 14:19
FROGGS jnthn: it looks like I am making progress with labels in nqp-j 15:22
jnthn \o/ 15:24
FROGGS++
jnthn is slowly moving forward with the tricky spesh work too
FROGGS jnthn++ # :o) 15:25
[Coke] (warehouse 13)++ 15:35
masak [Coke]: it amuses me a bit that Wikipiedia has a category named "Fictional government investigations of the paranormal" :) 15:38
[Coke] ok, now rakudo.parrot is failing 205 tests.. 15:43
jvm: integration/advent2013-day19.t aborted 6 test(s) 15:44
parrot is failing a ton of things in S32-temporal now.
esp: S32-temporal/calendar.t aborted 126 test(s)
'plan *;' could become 'plan 20;' in S03-smartmatch/signature-signature.rakudo.parrot
dalek ast: 04ed716 | masak++ | S03-smartmatch/signature-signature.t:
[S03-smartmatch/signature-signature.t] 'plan 20;'

Instead of 'plan *;'. Coke++		 15:51
masak [Coke]: on Rakudo Moar, S32-temporal/calendar.t flags up 43 potential difficulties, all related to leading zeros not indicating octal. 15:58
easy cage cleaning task for someone: either remove them, or insert #OK everywhere.
m: say 03
camelia rakudo-moar 86db16: OUTPUT«Potential difficulties:␤ Leading 0 does not indicate octal in Perl 6; please use 0o3 if you mean that␤ at /tmp/sVsC6g6giP:1␤ ------> say 03⏏<EOL>␤ Leading 0 does not indicate octal in Perl 6; please use 0o3 if you m…»
masak m: say 03 #OK
camelia rakudo-moar 86db16: OUTPUT«Potential difficulties:␤ Leading 0 does not indicate octal in Perl 6; please use 0o3 if you mean that␤ at /tmp/gKJxllOp8i:1␤ ------> say 03⏏ #OK␤ Leading 0 does not indicate octal in Perl 6; please use 0o3 if you me…»
masak oh, Rakudo doesn't honor #OK :/ 15:59
then I recommend their removal.
[Coke] masak: those warnings have been there since Day 0. 16:02
I've complained about them before, and they were (i think intentionally) not removed. 16:03
masak ok. who am I to disagree with someone's insistence that a test spew a lot of useless output? :) 16:04
[Coke] did someone recently change those tests or the temporal handling in r-p? 16:17
hoelzro [Coke]: we were talking about time vs now the other day 16:18
I don't know if someone changed it
PerlJam m: say 01;
camelia rakudo-moar 86db16: OUTPUT«Potential difficulties:␤ Leading 0 does not indicate octal in Perl 6; please use 0o1 if you mean that␤ at /tmp/5VyuKnExNs:1␤ ------> say 01⏏;␤ Leading 0 does not indicate octal in Perl 6; please use 0o1 if you mean …»
PerlJam Why does it output the warning twice?
vendethiel PerlJam: because there's two numbers, duh :P 16:20
TimToady maybe someone wanted to make it twice as irritating in hopes it would be removed? 16:21
PerlJam I was thinking it was because of backtracking in the grammar but I don't see it. 16:24
TimToady p6: say 01; #OK octal
camelia niecza v24-109-g48a8de3: OUTPUT«1␤»
..rakudo-jvm 86db16: OUTPUT«(timeout)»
..rakudo-{parrot,moar} 86db16: OUTPUT«Potential difficulties:␤ Leading 0 does not indicate octal in Perl 6; please use 0o1 if you mean that␤ at /tmp/tmpfile:1␤ ------> say 01⏏; #OK octal␤ Leading 0 does not indicate octal in Perl 6; please use …»
PerlJam r-j timed out?!? 16:25
TimToady not fast enough to do it twice :P
n: say 01; #OK division 16:26
camelia niecza v24-109-g48a8de3: OUTPUT«Potential difficulties:␤ Leading 0 does not indicate octal in Perl 6; please use 0o1 if you mean that at /tmp/vTHqq4tLhd line 1:␤------> say 01⏏; #OK division␤␤1␤»
PerlJam maybe rakudo implements <!!{ ... }> in a way that causes it to execute twice? (there are two !, right? :) 16:49
er, I guess that'd be NQP
TimToady could be, actually 16:56
since <{...}> is evaling regex code, maybe <!!{...}> is evaling by (false) analogy, which is should just call the closure
*when
*it 16:57
TimToady what's wrong with my brane, that I think "when it" and type "which is"? 16:57
TimToady maybe I'm evaling something when I should call a closure... 16:58
PerlJam I tried a simple test in nqp and couldn't get it to execute twice.
TimToady yes, well, it wouldn't really explain the double warning either, I suppose 16:59
we need something that parses twice
flussence
.oO( this slowness in the p6 syntax highlighting is forcing me to become better at vim... )
 17:00
TimToady maybe I should, like, actually look at the code, or someth'n
nah...
PerlJam TimToady: src/Perl6/Grammar.nqp line 2908 :) 17:01
github.com/rakudo/rakudo/blob/nom/....nqp#L2908 (if that lowers the barrier to looking) 17:04
TimToady well, not really, since I've been doing recursive searches since long before 'ack' was written :P 17:19
I just wish I could train my fingers to 'make m-install' instead of 'make install'
or get it to use multiple cores without me sayin' -j3 or so 17:20
though it did give me a chance to backlog a bit more...
btyler flussence: heh, same here :) my usual "scroll scroll scroll-->word jump" habits don't work so well 17:24
TimToady it's calling into the integer rule twice for some reason
it's not internal to the integer rule
TimToady hopes if we find this bug the compiler will run twice as fast :D 17:25
TimToady actually thinks we'd see a lot more duplication if that were so...
PerlJam maybe dec_number or rad_number somehow eventually calls integer (but again, I don't see it) rad_number calls radint which calls integer, but only after it's parsed a '<' in rad_number 17:30
TimToady here's your clue 17:32
m: 01i
camelia rakudo-moar 86db16: OUTPUT«Potential difficulties:␤ Leading 0 does not indicate octal in Perl 6; please use 0o1 if you mean that␤ at /tmp/V3a1ub9NSD:1␤ ------> 01⏏i␤»
TimToady it's complex backtracking
TimToady 01 doesn't find the i, so tries again with <integer> 17:33
obviously complex is not including the i in the LTM, though that seems to be the intent 17:34
so <complex> and <integer> are tied, <complex> runs first and fails
PerlJam ah
does the '\\'? in complex stop the LTM? 17:35
TimToady it's possible std and/or niecza deal with this sort of thing simply by de-duping the warnings
or the re-collection after the [] in <integer> 17:36
we tend to give up on LTM at that point because the possibilities explode
TimToady well, on this one, it's possible that we should simply not even look for the i anymore, and rely on constant folding instead 17:37
it's really kinda bogus to hardwire a postfix there, since someone could add a prefix tighter than i 17:38
that's why we have <1i> instead as a literal syntax
I wonder what happens if we simply delete the rule... 17:39
yup, deleting number:complex fixes the double warning; spectesting 17:43
TimToady to answer you question, I don't think a ? term will stop LTM, but I think the end of [a|b|c] will, since [a|b|c] [d|e|f] ... [x|y|z] has 3 ** N possibilities, where N is the number of alternations 17:49
TimToady though a regular language might handle that okay, I suppose 17:51
so maybe it's not an issue these days
anyway, neither of those was the issue here
PerlJam TimToady++ 17:52
TimToady so no double speedups on parsing programs, alas, unless your program was all non-complex numbers
PerlJam But you halved the number of warnings from S32-temporal/calendar.t 17:54
:)
or ... will halve 17:55
[Coke] will halve halved 18:02
if I was going to do it, we could say
will will halve halved!
PerlJam heh 18:03
[Coke]: btw, I was on a flight to New Orleans on Sat. and I sat behind a guy who strongly reminded me of you. Same height/build and same 'hair cut' :) 18:04
(of course I may be goind a little crazy as I saw someone at the grocery store who at first from slightly behind at 3/4 view looked like lizmat to me) 18:06
s/goind/going/
lizmat can confirm she was not there
[Coke] WE ARE ALL AROUND YOU
colomon AND WE ARE FOLLOWING YOU 18:06
dalek kudo/nom: 362f3fd | TimToady++ | src/Perl6/Grammar.nqp:
remove bogus literal complex rule

We should not have a special literal rule for 42i, for several reasons: First, it's incorrect in the face of language extensibility, since someone could add a different affix that overrides i in either precedence or semantics. Second, the rule was incorrect anyway, since it didn't check for a word boundary after the i. Third, we can and should rely on constant folding these days to reduce 42i to a literal value when it's appropriate (and refrain from doing so when it's not). Fourth, a literal complex ought to be written with angles anyway, if you don't want the constant folder to get involved. That's why that construct is there. Fifth, this fixes the double warning on
  (non)-octals beginning with 0, since the complex rule was calling the
integer rule and then backtracking in the absence of an i.
PerlJam maybe everyone has a doppelgänger in Texas? 18:07
(except for the people that already live here, we have doppelgängers elsewhere)
lizmat ah, but only in ASCII :-) 18:08
TimToady hah, if I change my name in the local repository to my nick, it comes through here :) 18:10
so I'm no longer my own double, except maybe in Texas 18:11
m: say 42ior43 18:12
camelia rakudo-moar 86db16: OUTPUT«===SORRY!=== Error while compiling /tmp/XryjrdWrLL␤Confused␤at /tmp/XryjrdWrLL:1␤------> say 42i⏏or43␤ expecting any of:␤ whitespace␤»
colomon lizmat: right, in Texas they're doppelga:ngers 18:13
TimToady okay, I guess <ws> catches it
[Coke] PerlJam: uh, my doppelgänger in texas was me. 18:15
TimToady you can say that again 18:17
FROGGS jnthn: that is what I have about labels in nqp-j: gist.github.com/FROGGS/5f4dd9417cecc972a5ad
PerlJam [Coke]: You were sitting in front of me on a flight from Dallas to New Orleans? 18:18
[Coke] maaaaaybe.
[Coke] I'm not willing to rule it out, especially after having just listened to: pseudopod.org/2013/09/06/pseudopod-...low-house/ 18:20
FROGGS jnthn: I do a $ATHROW when the handler got an exception for another label, but it just seems to explode rather then to look for another (outer) handler 18:20
hmmmm, perhaps I should put my stuff in another unwind_check... 18:26
jnthn FROGGS: Is it a vm exception? May need a higher level Ops.rethrow then? 18:30
FROGGS jnthn: no, it is a unwindexception that carries a vmex in the result
the explosion also states: Exception in thread "main" org.perl6.nqp.runtime.UnwindException 18:31
but I am refactoring the label stuff into unwind_check now... I think that will be cleaner
masak today's mini-challenge: from bash, what's the golfiest way to chop the first line off a given text file? 18:32
you can use anything you want: perl6, Perl 5, awk, sed, tail...
jnthn FROGGS: But wait...an unwindexception has a particular target, iirc?
geekosaur sed 1d 18:33
FROGGS jnthn: yes... do I need to update that? O.o
jnthn FROGGS: So if you rethrow it...the target won't match 18:34
FROGGS hmmmm
jnthn FROGGS: Well no, I just think you want to grab the vmexception out of it
FROGGS: And them do a high-level rethrow
FROGGS okay
and this the origin was a higher level op also, I might get an unwind exception back later 18:35
s/this/since/
s/op/ex/
gorram!
[Coke] masak: github.com/coke/perl6-roast-data/b...in/doit#L1 does it 18:37
I like sed 1d, though. :) 18:38
masak [Coke]: `sed 1d` is a really good start. it doesn't write back to the file, tho'. 18:41
Perl 6 question: have variables `my ($class, $message) = "C", "hi";`, how do I generate (through the MOP) `class C { method message { "hi" } }` ? 18:43
timotimo if nobody gives you a satisfactory answer quickly enough, have a look at my ADT module 18:45
:)
PerlJam masak: just use a macro! ;) 18:46
timotimo :(
[Coke] masak: sed -i\ 1d <filename>
masak PerlJam: :P 18:47
[Coke]++
jnthn m: my ($class, $message) = 'C', 'hi'; my $c = Metamodel::ClassHOW.new_type(:name($class)); $c.HOW.add_method($c, 'message', anon method message() { $message }); $c.HOW.compose($c); $c.message 18:48
camelia rakudo-moar 362f3f: OUTPUT«(signal )»
jnthn uh..
[Coke] jnthn: nice!
jnthn That worked first time locally.
jnthn m: say 'alive?' 18:49
camelia rakudo-moar 362f3f: OUTPUT«alive?␤»
jnthn hmm
r: my ($class, $message) = 'C', 'hi'; my $c = Metamodel::ClassHOW.new_type(:name($class)); $c.HOW.add_method($c, 'message', anon method message() { $message }); $c.HOW.compose($c); $c.message
camelia rakudo-jvm 362f3f: OUTPUT«(timeout)» 18:50
..rakudo-parrot 362f3f: OUTPUT«Nominal type check failed for parameter '$expected'; expected Any but got Mu instead␤ in block at /tmp/tmpfile:1␤␤»
..rakudo-moar 362f3f: OUTPUT«(signal )»
jnthn wonders why on earth that works locally, but not here...
masak why 'anon method'? 18:52
oh, I see why.
it acts as a kind of quoting mechanism.
jnthn yeah
uh...it explodes if I put it in a file. what.
masak "Bus error" here. 18:53
moritz "Cannot invoke this object (REPR: Null, cs = 0)" (with not qutie the same code
masak did I just, um, uncover some kind of bug? or something? 18:54
:)
I'm kind of new at this.
oh wait.
mine said "hi" and then "Bus error".
that's good, right?
masak hopes "Bus error" has nothing to do with "Bus number" 18:55
moritz m: gist.github.com/moritz/1f58e20a98d7b0f5219d
camelia rakudo-moar 362f3f: OUTPUT«gist not found»
moritz not the message I was hoping for :( 18:56
timotimo aye, it doesn't like private gists unfortunately 18:57
moritz can I pulicify a gist?
timotimo oof
dunno
moritz m: my ($classname, $message) = ('C', 'hi'); my $class := 1.HOW.new_type(:name($classname)); $class.^add_method('message', anon method message() { $message }); 18:58
camelia rakudo-moar 362f3f: OUTPUT«Cannot invoke this object (REPR: Null, cs = 0)␤ in block at /tmp/pqCI2HXXZR:1␤␤»
FROGGS m: my ($classname, $message) = ('C', 'hi'); my $class := 1.HOW.new_type(:name($classname)); $class.HOW.add_method('message', anon method message() { $message }); 18:59
camelia rakudo-moar 362f3f: OUTPUT«Not enough positional parameters passed; got 3 but expected 4␤ in any add_method at src/gen/m-Metamodel.nqp:404␤ in block at /tmp/jz3At_r9q0:1␤␤»
timotimo i think you need to pass in the class with add_method, too 19:00
FROGGS m: my ($classname, $message) = ('C', 'hi'); my $class := 1.HOW.new_type(:name($classname)); $class.HOW.add_method($class, 'message', anon method message() { $message });
camelia ( no output )
timotimo if you use .HOW
FROGGS the .^ seems borken
timotimo hm, maybe because it's not composed yet?
moritz ah right, .^ just dispatches to a method in Mu in rakudo 19:06
I wonder if that's kosher
moritz now it works for me 19:06
masak moritz: you can't publicify a gist, but anyone you share the URL with can see it. 19:11
oh! apparently not camelia...
moritz maybe she needs glasses 19:14
(or classes? :-)
masak 'night, #perl6 19:50
colomon \o
retupmoca colomon: is there a public link to your ecosystem smoke tests? 19:53
thou m: my $s = "abc/def/ghi.jkl".subst(rx/ ^ .* '/' (.*) '.jkl' ? /, {"GOT ($0)"}); $s.say; 20:35
camelia rakudo-moar 362f3f: OUTPUT«GOT (ghi.jkl)␤»
thou m: my $s = "abc/def/ghi.jkl".subst(rx/ ^ .* '/' (.*?) '.jkl' ? /, {"GOT ($0)"}); $s.say; 20:36
camelia rakudo-moar 362f3f: OUTPUT«GOT ()ghi.jkl␤»
thou The REPL doesn't handle this case of subst with closure, but it seems to work fine in a script. Is this a known issue with the REPL (moar)? pastie.org/9157118 21:10
geekosaur that looks familiar-ish? recent report that the repl is not handling the second one properly 21:14
retupmoca yes, .subst doesn't set $/ properly in the REPL 21:25
FROGGS ohh, and I know why... 21:27
every line is in its own scope, due to EVAL 21:28
retupmoca iirc, last time I looked at it nqp::getlexcaller couldn't find '$/'
FROGGS and .subst looks a few caller frames to find the $/
and in case of an EVALd line, it is one caller too few/less
retupmoca note that the very first REPL line you enter will set $/ properly 21:30
FROGGS yeah 21:34
itz the Czech Act Site list of talks seems to be broken 21:51
(which I mention in the hope that someone here might know about it)
jnthn Hmm...so it is 21:53
Guess I could email the organizatorka I've been in contact with... :) 21:54
itz: done 21:56
colomon retupmoca: 75.134.170.198:3001/report 21:57
itz cool we may even find there was someone working on it currently :)
FROGGS something happened in the last three days 21:58
retupmoca colomon: is that perl6-p ooc? 22:06
colomon retupmoca: that's the parrot results, yes. 22:12
thou OK, another bug (I think), with subst and $/ in closure: pastie.org/9157367 22:42
If it's inside a for loop, it doesn't get $0 in the closure.
timotimo aye, i think $/ in subst is b0rked :(
thou s[foo (.*) bar] = "$0" doesn't work right either 22:43
OK. 22:44
thou tadzik: I think panda's .work should be renamed to .panda-work or something to avoid accidentally removing a non-panda .work dir. 23:14
timotimo +1 23:15
thou OK, my p6 module dependency clone tool works: pastie.org/9157490 23:22
Not sure if it'd be useful for anyone else....
timotimo ah, good, you're not doing it with a big grammar any more to parse panda's output :) 23:23
thou no, i knew that was stupid, just was having fun practicing w/ grammars
timotimo ah, good :) 23:24
thou OK, good night, #perl6. 23:25
flussence aagh... that's the first time I've had to kill -9 _vim_... 23:36
timotimo perl6 syntax highlighting again? ;) 23:37
flussence mispasted a few lines of ascii table in without doing a :syn off first :( 23:38
timotimo ouch :) 23:39
flussence (at this point I'm starting to wonder if a perl6-based vim clone would run faster...) 23:40
timotimo lolnope
jnthn says profiles say we spend too much time in the multi-dispatch cache, so the newest spesh features are working towards doing spesh-time MD :) 23:41
though i don't think that'll make parsing much faster
it would be pretty cool if perl6 regexes/rules were blazing fast
flussence they feel like they're within an order of magnitude of vim's now, at least (and p6's are readable :) 23:43
timotimo hah 23:44

Powered by Cro and the Raku® Programming Language.

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

Thank you!