»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.
[Coke] (assuming you are invoking the sql multiple times) 00:00
dalek rl6-roast-data: a5d71dc | coke++ | / (5 files):
today (automated commit)
00:14 denis_boyun_ left 00:16 ajr_ left
[Coke] jnthn++ finding one of the OSX bugs. 00:20
colomon \o/ 00:22
00:24 KCL_ joined 00:27 KCL left 00:42 kurahaupo1 left 00:44 BenGoldberg joined 00:46 telex left 00:47 guru joined, guru is now known as ajr_ 00:48 telex joined, woshty left 00:50 kjs_ joined 00:51 Mouq joined 00:56 ajr_ left 00:57 guru joined 00:58 guru is now known as Guest15347, Guest15347 is now known as ajr_ 01:01 adu joined 01:03 woshty joined 01:05 adu left 01:13 teak1010 joined 01:16 kjs_ left 01:18 Mouq left 01:23 ajr_ left
timotimo o/ 01:37
01:46 adu joined
timotimo ah, indeed 01:46
i'm glad jnthn spotted that terrible little annoying buglet
but alas ... there's more to slay up ahead
01:51 prime left
adu hi timotimo 01:51
timotimo hello adu 01:53
adu how goes? 01:54
timotimo i'm all right, but motivation and productivity seem to have reached a local minimum
adu sounds like thanksgiving 01:55
timotimo i'ven't celebrated that in a long time
adu I usually work on WORK during the holidays, but this time I'm researching XQuery (and eXistDB)
timotimo i don't know what XQuery is 01:56
adu its an XML thing
timotimo i know XPath at least :) 01:57
adu there are 2 standardized XML "programming" lanugages: XSLT and XQuery, they're Turing-equivalent, but implementing XQuery in XSLT would probably take a million LOC, whereas implementing XSLT in XQuery would probably take 1 LOC 01:58
XPath is a subset of XQuery
timotimo ah, OK
i have had the fortune(?) to never have to play with XSLD 01:59
SXLT
XSLT*
adu don't, it's a bad choice for programming, which is what most people NEED it for, but they would never admit it to themselves
s/it/XSLT/ 02:01
02:02 Mouq joined
timotimo hm. 02:03
i'd like to build an android app that gains me a tiny but constant stream of money that exponentially grows in popularity
adu timotimo: PhoneGap. 02:04
timotimo i'm not terribly fond of html actually 02:05
i haven't had a proper go at it in a long time, though
adu the money making scheme is just: (1) find client (2) make prototype website (3) use PhoneGap to make Android app (4) show client until happy (5) make money 02:07
I thought HTML was the new English
I mean ... my grandma knows HTML, my parents know HTML ... 02:08
timotimo: a great place to start is H5BP 02:10
02:13 BenGoldberg left
timotimo what is that? 02:19
02:23 BenGoldberg joined 02:30 rurban left
hobbs html5boilerplate.com/ presumably 02:31
timotimo oh, interesting
most stuff i've been whipping up has been with bootstrap
i was hoping i could get into html5responsiveboilerstrap.js 02:32
02:46 Khisanth left 02:48 Khisanth joined 02:52 araujo left, beastd left 02:56 prime joined 02:59 prime left 03:06 rurban joined 03:17 noganex joined 03:20 noganex_ left 03:33 Akagi201 left 03:34 Akagi201 joined 03:38 raiph joined 03:58 adu left 04:00 fhelmberger joined 04:04 fhelmberger left 04:12 rurban left 04:22 gfldex joined 04:26 rurban joined
flussence 9 ;) 04:45
.oO( I remember the weirdest details... )
04:46
hobbs 9 is 4 + 5!
05:18 rurban left 05:19 teak1010 left 05:40 BenGoldberg left 05:43 gfldex left 06:03 kurahaupo left 06:06 kurahaupo joined 06:14 Mouq left 06:16 raiph left 06:23 raiph joined 06:26 Hor|zon_ joined 06:28 sven_123_ joined 06:30 mathw_ joined 06:31 zoosha_ joined 06:34 mr-foobar left 06:37 Alina-malina left, lue left, hugme left, sven_123 left, zoosha left, Hor|zon left, mathw left, Alina-malina joined 06:39 hugme joined, ChanServ sets mode: +v hugme, lue joined 07:08 anaeem1_ joined 07:09 anaeem1_ left 07:10 Akagi201 left, Akagi201 joined 07:15 Akagi201 left 07:29 Mso150_p left 07:42 Mso150_p joined 07:47 darutoko joined 07:49 raiph left 07:54 Akagi201 joined 08:00 Mouq joined 08:39 Alina-malina left, Alina-malina joined 08:42 lizmat joined
lizmat drops a pin 08:45
raydiak careful, someone might step on that... 08:46
lizmat picks up the pin
08:46 synopsebot left, tadzik left 08:59 ptc_p6 joined
bartolin FWIW, today I had clean spectests on Moar, JVM and Parrot (running on FreeBSD 10.0): github.com/usev6/perl6-roast-data/..._rates.csv 09:03
09:11 smls joined
smls o/ 09:12
09:14 spider-mario joined
moritz \o 09:16
09:17 bjz joined
smls Do you by any chance know if the «.new-from-pairs» method in the Setty/Baggy/Mixy roles is official puplic API, or just a Rakudo implementation detail? 09:18
09:18 PerlJam left
smls It's not mentioned in the specs. 09:18
In Rakudo, the default .Set coercer is implemented in terms of it, and it can also be called directly. 09:19
09:20 gfldex joined 09:23 rurban joined 09:33 vukcrni left 09:36 pmurias_ joined, pmurias_ left, pmurias joined
pmurias jnthn: what does the usecompileehllconfig op do? 09:37
09:37 vukcrni joined 09:42 mvuets joined
lizmat smls: please consider it an implementation detail subject to change without notice 09:52
smls ok
lizmat so please use .Set, which *is* part of the documented API
09:55 virtualsue joined 09:56 KCL_ left 09:58 KCL_ joined 09:59 Juerd left
pmurias is there a way to get the object from a compiled .moarvm file if I know the object id and serialization context handle? 10:05
10:05 hugme left 10:07 pmichaud left, Util left, masak left, sergot left, [Coke] left, woolfy joined 10:08 dalek left
itz masak++ # new post 10:10
10:25 anaeem1_ joined 10:33 hugme joined, ChanServ sets mode: +v hugme
lizmat does anybody know what happened to dalek this time ? 10:36
10:39 Alina-malina left 10:45 mvuets left 10:48 Mso150_p left 10:49 Juerd joined
lizmat Juerd: could it be that feather is in problems again? 10:50
10:50 Alina-malina joined
Juerd lizmat: Yea. Brand new server, installed last week, had a double disk crash :( 10:51
lizmat :-(
Juerd Two SSDs that were in RAID 1 vanished from the bus, roughly an hour apart
lizmat prefers DOA over this
10:52 anaeem1_ left
Juerd They're dead, not recognised by any of the three systems I tried them in. 10:52
lizmat that's really shit
10:52 anaeem1 joined
Juerd Yes, especially because they were used for writeback caching and had some 10 MB of dirty data on them. Not much, but someone is bound to run into missing or otherwise broken data. 10:53
Probably a year from now.
10:57 anaeem1 left 11:00 mvuets joined
lizmat m: class A { has int $.a; method BUILD(:$!a) {} }; A.new(a=>1) # do we have a ticket number for this yet ? 11:03
camelia rakudo-moar 28cfa0: OUTPUT«Cannot modify an immutable int␤ in method BUILD at /tmp/OO0_3ySwks:1␤ in method BUILDALL at src/gen/m-CORE.setting:961␤ in method bless at src/gen/m-CORE.setting:944␤ in method new at src/gen/m-CORE.setting:928␤ in block <unit> at /tmp/OO0_3…»
itz 24pullrequests.com/ 11:04
is perl6 or rakudo part of this? I don't seem to be able to search just view all 11:05
11:08 Isp-sec joined
itz oh there a tag cloud 11:09
Mouq oi, it's .new-from-pairs now? No wonder I couldn't do .new-fp anymore. Would've been nice to have had a longer deprecration cycle on that one :/ 11:13
(Wrt Setty/Baggy/Mixy)
11:14 gfldex left
lizmat Mouq: it was not part of the published API afaik, so no deprecation cycle... 11:14
and BTW, I did that almost a year ago already, if I remember correctly
11:15 denis_boyun_ joined
itz I added 4 projects to 24pullrequests 11:16
lizmat hmmm... apparently not... more like 6 weeks ago
Mouq Haha, alright 11:17
lizmat how time flies ... 11:18
11:24 Mouq left 11:30 Mouq joined
jnthn ...or doesn't, apparently :P 11:32
Afternoon, #perl6
lizmat jnthn o/ 11:33
jnthn "X is not in the spec" will not, sadly, mean folks won't discover/use it...
Especially given the introspection features we have...
lizmat well, I guess we should add a "official API" trait to the stuff we consider to be official API 11:34
jnthn Or just make private the things we don't :)
arnsholt That too. But I think at least some API documentation should be part of the spec 11:36
lizmat hmmm... then if we forget to do that for something, it is all of a sudden public
arnsholt In light of my recent kvetching over the Python spec =)
lizmat I'd rather have it the other way around 11:37
jnthn lizmat: My point was more that our userbase is going to want us hold us to de facto public, not de jure public. :)
*to want to
11:43 denis_boyun_ left
Mouq masak++ # I missed the new post! 11:43
arnsholt digs into the Rakudo code 11:47
Let's see if I can't figure out how this setting stuff works
Oh, wait. There's setting stuff in NQP too. That's probably easier to follow 11:49
11:49 JimmyZ_ joined
jnthn Yes, it will be :) 11:57
And I think common bits are factored out into some things in HLL::Actions, so you can re-use them a bit.
arnsholt Oh, awesome! 11:59
lizmat jnthn: nqp::getstdin gives you a PIO with a REPR MVMOSHandle, right ?
brb& 12:00
jnthn lizmat: On MoarVM, yes 12:01
lizmat ok, just double checking :-)
jnthn lizmat: In general, you should not assume anything about such a handle besides that you can use it with other fh-taking nqp:: ops.
lizmat yes, I was just refactoring opening STDIN and friends 12:02
arnsholt jnthn: Oh, and a follow-up question from a few days ago: What does :blocktype immediatestatic and declarationstatic mean? They're not documented in docs/qast.markdown
lizmat and it started complaining about the PIO not being an MVMOSHandle
so I wanted to be sure :-)
jnthn lizmat: Yes, the way it informs you of your mistake is platform specific, but the mistake will have cross-platform effects ;) 12:04
lizmat yup
jnthn arnsholt: It means "don't emit any closure-taking code"
arnsholt: Generally because you're going to manage it yourself or the thing is never a closure.
arnsholt Aha. So for example the blocks I emit for a for loop could be declarationstatic, since it'll never exist outside of its lexical scope? 12:05
jnthn No, it's not that simple 12:06
It cares about seeing the right outer things
I presume, anyway.
arnsholt Yeah. So a *static doesn't have an outer at all?
jnthn def foo:; omg = 'so implicit'; for syntax in i_cannot_remember_python:; print omg 12:07
It has an outer, but it is equal to the last outer that ran
arnsholt The last outer that ran? 12:08
jnthn In that snippet I just wrote, if you had that running recursively for for two threads, then whoever entered foo second would have their `omg` seen.
arnsholt Oh!
jnthn And the first would suddenly come to see a different outer 12:09
Yes, thus hwy it's rarely what you want ;)
*why
arnsholt I can see that now. Thanks =
=)
mvuets hello there! i have a question here (-: what would be a recommended getaway rakudo backend for a complete newbie? 12:11
colomon moar 12:12
12:14 tadzik joined 12:15 synopsebot joined, psch joined
psch good *, #perl6 12:16
colomon o/
psch colomon \o 12:19
[Tux] ==> Reinstalling
Could not download module metadata: Failed to connect: connection refused
in method update at lib/Panda/Ecosystem.pm:91
lizmat [Tux]: feather is down after a double SDD crash 12:21
Juerd is working on it afaik
*SSD
[Tux] ok
12:24 KCL joined
tadzik [Tux]: I picked it up just a few minutes ago 12:25
[Tux]: should be okay now 12:26
if it's still up that is, I decided to move my irc session elsewhere
botsnack!
hm
synopsebot: botsnack! 12:27
psch botsnack might be matched with /^botsnack$/
botsnack
oh
or
it's not here
tadzik $args->{body} eq 'synopsebot: botsnack!'
12:27 KCL_ left
tadzik no idea why it's broken 12:27
psch it's here, i should have checked /names :/ 12:28
tadzik yeah, it's just broken for some reason 12:29
psch S99:botsnack
synopsebot Link: perlcabal.org/syn/S99.html#botsnack
psch hm
tadzik hah
12:29 virtualsue left
psch jnthn: i'm understanding this correctly that i need to override createAdaptorMethod, am i not? i have a feeling i'm running into some kind of circularity there though 12:32
jnthn: as in, i have to defer the visitation of the code for the methods until i know which ones i want to treat as multis to have the dispatch logic inside the CompilationUnit, but i need to have created the CompilationUnit before i can see all methods i want to treat as multis 12:33
oh, i could override createAdaptor instead of createAdaptorMethod and run a first-pass through the methods to determine multis by their descriptor 12:35
12:42 JimmyZ_ is now known as JimmyZ
arnsholt jnthn: Looking at NQP::Compiler, I see that MAIN has to be @ARGS on Parrot/Moar, but *@ARGS on JVM. Is that a bug of some kind? 12:43
jnthn arnsholt: Yeah...
12:43 erkan left
jnthn psch: I don't quite remember how it works, but that sounds about right... 12:44
psch: That is, locate the things, then do the building.
12:45 ptc_p6 left
arnsholt jnthn: Is it related to something already in the issue queue, or something else? 12:45
12:46 ptc_p6 joined
jnthn arnsholt: Not sure if it's in the issue queue...I can't remember what it is, it was quite a while ago... 12:47
arnsholt: I think I was like "argh, just hack it and worry about it later" :)
It may be a bug in interaction of :main and slurpy stuff 12:49
arnsholt I'll add an issue for it. Doesn't look like it's there 12:52
And it's JVM that's buggy, right? =) 12:53
jnthn arnsholt: Well, we should probably follow what Perl 6's MAIN does, which is takes *@ARGS if there are many 13:00
That'd suggest NQP-JVM is the one in the right here.
bbiab
arnsholt Cool 13:01
13:03 Lasse_ joined
pmurias arnsholt: I think that's a case of "Just #ifdef it on parrot to avoid having to deal with it as much as possible" 13:03
13:04 kurahaupo left 13:05 girlperlie joined, girlperlie left
arnsholt Moar follows Parrot on this one, actually =) 13:05
lizmat alas, no clean spectest for me still... :-( 13:10
timotimo how close to clean is it for you? 13:11
lizmat trying a complete clean build again first
2 random files failing: t/spec/S32-array/end.t and t/spec/S32-array/push.t 13:12
are ok on their own
fwiw, I just pushed the NQP_REVISION bump 13:13
jnthn Well, 2 is a lot better than a dozen. 13:17
lizmat: What was the failure mode? 13:19
pmurias jnthn: is there a way to dump some info about an object in nqp without guessing what it is? 13:23
13:24 anaeem1 joined
jnthn pmurias: Well, you can $obj.HOW.name($obj) on pretty much anything in MoarVM. 13:24
pmurias it turns out my problems are being caused by some VMArray using a STable from the compiler setting 13:29
and the --bootstrap flag solves them for reasons I don't fully understand ;) 13:31
jnthn ;)
I suspect I'd have to read as much code to re-understand it and explain it, as you would to figure it out... :) 13:32
I remember thinking while writing it, "I'm SO glad Rakudo is not a bootstrapping compiler, only NQP is" :)
13:33 Guest97404 left 13:36 bowtie joined, bowtie is now known as Guest40300 13:38 denis_boyun_ joined 13:39 anaeem1 left, anaeem1 joined 13:40 kjs_ joined, ptc_p6 left
Mouq lizmat: What is IO::Dup short for? 13:46
13:47 telex left
jnthn How about you? 13:48
oops
ww
Mouq: I'd guess duplicate or something...
13:48 telex joined
lizmat well, I guess it would be an objectification of man dup 5 13:49
and another consumer of the PIO role 13:50
13:59 prime joined, prime left, prime joined 14:01 kjs_ left 14:09 denis_boyun_ left 14:10 Mouq left, anaeem1 left 14:12 denis_boyun_ joined 14:14 rindolf joined
smls "MixHash [... ] Keys with fractional weights are deleted if they go to 0.", according to spec 14:18
What does "go to 0" mean here?
moritz is assigned a 0 value 14:20
smls ok 14:21
lizmat m: my $m = (a => pi).MixHash; say $m.perl; $m<a>=0; say $m.perl 14:22
camelia rakudo-moar f3390f: OUTPUT«("a"=>3.14159265358979).MixHash␤().MixHash␤»
smls right, just like Set/Bag
I just wondered for a moment there, if "goes to 0" refers to something complicated with infinite fractions or limits... :) 14:23
14:23 denis_boyun_ left
moritz Stage mast : Unknown constant 'SIG_QUIT' 14:24
lizmat whenever the condition $value == 0 is true, it will be deleted
which backend ?
moritz moar
lizmat which OS?
moritz I ran Configure.pl, it didn't complain about NQP being too old
lizmat I guess that's the reason I always nuke my install dir :-) 14:25
moritz huh
$ cat tools/build/NQP_REVISION
2014.11-27-g3254b59
$ ./install/bin/nqp-m --version
This is nqp version 2014.10-33-g86f1841 built on MoarVM version 2014.11
wth doesn't it complain about it being too old? 14:26
lizmat if it is because I did something wrong, please let me know
moritz sounds like a Configure.pl bug
smls In Rakudo: role Mixy does Baggy, and role Baggy does QuantHash 14:29
In S32/Containers: «role QuantHash[::T, $default = Any] does Mixy does Associative {...}»
lizmat smls: the spec says use QuantHash 14:30
smls use for what?
lizmat however, currently private attributes defined in a role that is consumed in another role, are not visible in the class that consumes that role 14:31
that's why we have the rather strange order of Sets/Bags/Mixes atm
smls I was pointing out the inconsistency between Rakudo and spec, of which role does which
lizmat yes, it is not up to spec wrt to roles it should use
smls oh, so it's a temporary thing?
lizmat it *should* be up to spec wrt to functionality 14:32
14:32 Alina-malina left, kjs_ joined
smls but on p6doc we show type graphs as well... 14:32
(currently, based on Rakudo.) 14:33
14:33 Alina-malina joined
moritz dalek, lazy bastard, where are you? 14:35
14:36 raiph joined, dalek joined, ChanServ sets mode: +v dalek 14:38 Alina-malina left 14:39 kjs_ left 14:40 spider-mario left 14:42 anaeem1_ joined
dalek c: 77dda07 | Mouq++ | lib/Language/operators.pod:
Missed in last commit: add dumb example for .Foo::bar

Something more in-depth would be better, but this at least shows the syntax of the op.
14:43
smls Mouq: 'does Bar' should be 'is Bar' in that commit. 14:46
14:46 anaeem1_ left
smls or alternatively, 'class Bar' ---> 'role Bar' 14:47
14:48 Alina-malina joined 15:02 spider-mario joined 15:06 araujo joined 15:12 brrt joined
raiph hi brrt, please see stackoverflow.com/questions/2717859...ile-to-exe 15:18
yoleaux 23 Nov 2014 11:14Z <nine> raiph: thanks. Of course there'll always be edge cases that need some workarounds. But from my practice using Perl 5's Inline::Python I can say, that there are always workarounds and most code does just fine
arnsholt "While looking for 'SNAKE.setting.moarvm': no such file or directory" that's a step in the right direction at least =D 15:22
15:23 Akagi201_ joined 15:26 Akagi201 left
timotimo brrt: i'd be interested in a status update on the static moar binary generator thingie 15:28
15:39 kjs_ joined, brrt left 15:44 zakharyas joined
arnsholt jnthn: ping? 15:50
smls m: .say for SetHash(5).pick.perl, SetHash().pick.perl, SetHash(5).grab.perl, SetHash().grab.perl 15:54
camelia rakudo-moar c11b2a: OUTPUT«5␤Nil␤(5,).list␤().list␤»
smls lizmat: ^ What's with grab (but not pick) list-ifying its return value?
timotimo TimToady: should i invest a bit of time into building a "NFA to graphviz" tool? 15:58
psch m: .say for SetHash(1..5).pick(2).perl, SetHash(1..5).grab(2).perl 16:01
camelia rakudo-moar c11b2a: OUTPUT«(2, 5).list␤(3, 1).list␤»
16:03 masak joined
masak aloha, #perl6 16:03
psch hi masak \o
masak feather did another reboot? 16:04
psch masak: irclog.perlgeek.de/perl6/2014-11-29#i_9735384 16:05
16:06 kjs_ left
jnthn arnsholt: I'm here now 16:08
raiph: Short answer is "there's not a way to do that yet", and --target=mast is just a debugging aid, so it's really not a suitable thing to use. 16:09
arnsholt jnthn: I think the setting stuff is mostly there, but when it runs the fixup to set the outer on the compiled thing, it looks like getcurhllsym("ModuleLoader") returns null
OTOH, looks like it's working in the World code 16:10
I guess it's not running in the right HLL, but I'm not entirely sure what's causing it
That's set by the CompUnit's :hll, right?
jnthn Yes 16:11
Also, do you have a module loader?
NQP iirc emits code to loadbytecode one.
timotimo hm. would it make sense to somehow visualize the contents of the serialized blob as a graph or html page of some kind?
jnthn I suspect you can crib its one initially.
arnsholt Aha. That's probably it
16:12 zakharyas left
jnthn You can do a gethllsym('nqp', 'ModuleLoader') to grab it even if you're in the other hll 16:12
arnsholt I do have a ModuleLoader class, but I guess my World might actually be getting NQP's ModuleLoader
16:15 prime left 16:20 zakharyas joined 16:27 Grrrrrrr joined, Grrrrrrr left
dalek ake: 8d2700e | (Arne Skjærholt)++ | / (7 files):
Initial setting implementation.
16:30
arnsholt jnthn++ # Halp
masak setting? you mean __builtin__, right? 16:32
arnsholt Yeah 16:33
Not actually accessible through that name yet though
Although now that you mention it, that's probably the correct way to do it once modules are implemented 16:34
Just implicitly load and import all the stuff from the builtins module
masak I think so. 16:36
arnsholt Although wrapping stuff in the lexical scope of __builtins__ might avoid a bit of computation on startup. Hard to tell a priori 16:38
Anyways, that requires a fair chunk of infrastructure I don't have yet
16:40 JimmyZ left, ggoebel111111113 left 16:52 Lasse_ left
dalek kudo/newio: b78a931 | (Elizabeth Mattijsen)++ | src/core/ (2 files):
Move STDIN/OUT/ERR init to a more logical place
16:57
kudo/newio: 1db1c24 | (Elizabeth Mattijsen)++ | src/core/ (2 files):
$*ARGFILES initialization in a more logical place
kudo/newio: c11b2a7 | moritz++ | Configure.pl:
Configure.pl: Complain if nqp-m is too old
16:58
kudo/newio: 0b64cc4 | (Elizabeth Mattijsen)++ | Configure.pl:
Merge branch 'nom' into newio
raydiak o/ 17:21
17:23 mr-foobar joined
ugexe did feather change dns settings again? im getting Connection Refused on one specific machine now for nickl@localhost:~/perl6/panda$ perl6 -e 'IO::Socket::INET.new(:host<feather.perl6.nl>, :port(3000))' 17:26
dalek kudo/newio: 3be8fcd | (Elizabeth Mattijsen)++ | src/core/PIO.pm:
Streamline IO::ArgFiles.get
17:27
kudo/newio: 17e652f | (Elizabeth Mattijsen)++ | src/core/Str.pm:
Str.path should make an IOU, not an IO::Path
moritz ugexe: might be an ipv6 problem 17:28
ugexe moritz: any hints on how i can tell? my sysadmin skills are lacking 17:29
tony-o ugexe: there is no ipv6 on that box, i have it disabled 17:30
oh that's on your machine
ugexe the problem isnt on that box
moritz ugexe: as root, try echo 1 > /proc/sys/net/ipv6/conf/all/disable_ipv6 17:31
ugexe: that will disable IPv6
ugexe: if it works afterwards, you know it's really the rakudo ipv6 bug 17:32
tony-o any help with Unhandled exception: When invoking is_bindable, Provided outer frame 0x7fb7ae4c0be0 (MVMStaticFrame bind_one_param) does not match expected static frame type 0x7fb7ae4c09a0 (MVMStaticFrame ) ?
ugexe connection till refused 17:33
still
moritz huh
ugexe: any firewall in the way?
ugexe i have iptables setup. but ive been able to connect fine for the last couple days 17:34
ive been working on panda and wonder if i somehow tripped some fail2ban type deal?
moritz I get the same error from panda
but nmap shows the port as open
ugexe m: IO::Socket::INET.new(:host<feather.perl6.nl>, :port(3000)); 17:35
camelia rakudo-moar c11b2a: OUTPUT«Failed to connect: connection refused␤ in method initialize at src/gen/m-CORE.setting:22518␤ in method new at src/gen/m-CORE.setting:22502␤ in block <unit> at /tmp/0ERLnFfkDT:1␤␤»
17:36 BenGoldberg joined
moritz ok, now nmap shows the port as closed 17:36
ugexe yeah now its not working on other machines (when it was ~15 minutes ago)
raydiak gets same from his web browser even
moritz seems that server process died
tony-o full dump for MVMStaticFrame error is here: gist.github.com/tony-o/6de78fe91379a2501415 17:38
dalek ast: 4d13580 | (Elizabeth Mattijsen)++ | S10-packages/basic.t:
s/.path.directory/.IO.dirname/
17:40
17:43 dwarring joined
TimToady timotimo: NFA graph might be interesting, though dunno how useful, unless to motivate someone to implement DFA 17:45
speaking of implementing DFA, why does the NFA never remove epsilons from the NFA, except over and over again in the results 17:46
seems like it'd be more efficient to guarantee there are no epsilons, and then we wouldn't have to check for them at NFA run time
and also, coincidentally, make the NFA readier for DFAish caching 17:47
17:48 gfldex joined
TimToady I guess it's just a DTSTTCPW. 17:48
mvuets i cannot get panda working: feather.perl6.nl:3000 refuses connections 17:49
is it down for someone else?
ugexe m: IO::Socket::INET.new(:host<feather.perl6.nl>, :port(3000)); 17:50
camelia rakudo-moar c11b2a: OUTPUT«Failed to connect: connection refused␤ in method initialize at src/gen/m-CORE.setting:22518␤ in method new at src/gen/m-CORE.setting:22502␤ in block <unit> at /tmp/2Ba4oymwvJ:1␤␤»
raydiak mvuets: yes, it's down for the moment
17:52 FROGGS_ joined
mvuets is there a mirror by chance? 17:53
raydiak mvuets: not that I know of, but you could d/l a module from github via links on modules.perl6.org/, then install locally w/"panda install ." in the module's dir, I believe 17:54
17:55 Sqirrel left 17:56 FROGGS left
mvuets ah, good idea! 17:56
ugexe you might have to use panda-install, as panda install now does Panda::Reporter.new( :$bone ).submit; which may not work without internet 17:57
psch it submits without the env var?
17:59 cognominal joined
psch ah, it always calls &Reporter::submit but &submit has if %ENV<PANDA_SUBMIT_TESTREPORTS> { #`[[ submitting happens here ]] } 17:59
ugexe oh, so yeah if its not set it should be fine
18:00 anaeem1 joined
psch i think cpandatesters lives on a different host as well, but i'm not sure about that 18:00
mvuets do you happen to know what generates that projects.json? 18:02
ugexe if you used panda recently you should have a current projects file 18:03
18:03 perltricks joined
mvuets sadly not, i just got it 18:04
18:04 anaeem1 left
colomon mvuets: what's your e-mail? (send message privately if you want) 18:04
or heck, wait a moment 18:05
perltricks hey y'all. This 1 liner's Syntax is OK: perl6 -ce 'state %l;.say if ++%l{$_}==2' 18:06
but this one hangs: perl6 -MTest -e 'plan 1;is q:x{perl6 -ce q/state %l;.say if ++%l{$_}==2/ example.txt}, "Syntax OK\n";'
any idea why?
simpler version: perl6 -MTest -e 'plan 1;is q:x{perl6 -ce "state %l;.say if ++%l{$_}==2"}, "Syntax OK\n";' 18:08
colomon mvuets: gist.github.com/colomon/a16b64f27f10010afed3 18:09
dalek osystem: 3ea5b89 | (David Warring)++ | META.list:
Added CSS::Writer

Reserializes AST data, as produced by CSS::Grammar etc
colomon mvuets: that's from midnight EST or so
dalek kudo/newio: 873e88a | (Elizabeth Mattijsen)++ | src/core/IO/Local.pm:
Add obsolete .directory for API consistency
18:10
kudo/newio: fd1ee99 | (Elizabeth Mattijsen)++ | src/core/IOU.pm:
Allow "foo".IO.open/mkdir if "foo" doesn't exist
kudo/newio: c4d1f1f | (Elizabeth Mattijsen)++ | src/core/io_operators.pm:
Make mkdir() work again
18:10 xiaozeng joined
mvuets colomon: sweet, thanks! 18:11
xiaozeng what's going on here 18:12
lizmat xiaozeng: welcome, we're working on perl 6 here 18:14
xiaozeng What can I do to help
18:15 woolfy left, Mso150_p joined
xiaozeng May i join you, or where can i get to know more? 18:17
lizmat xiaozeng: you probably should start with learnxinyminutes.com/docs/perl6/ or so 18:19
18:19 sftp left
lizmat that should be a good introduction into perl 6 18:19
mvuets xiaozeng: another nice page you may want to check out: perl6.org/getting-started/ 18:20
dalek kudo/newio: 21731f5 | (Elizabeth Mattijsen)++ | src/core/PIO.pm:
Apparent "bin" means don't set encoding

Rather than "binary". Wonder if it's possible of needed to be able to change the encoding of a PIO during its lifetime. If it should be possible, I don't think we can revert to "binary" at the moment.
18:21
xiaozeng thanks : ) 18:22
18:23 SevenWolf joined 18:28 j4janicej left, lizmat left, j4jackj joined 18:30 j4jackj is now known as j4janicej, sftp joined 18:31 xiaozeng left 18:32 perltricks left
dalek kudo/nom: 62c092f | (Elizabeth Mattijsen)++ | src/core/IO/Notification.pm:
Fix deprecated Supply.more (to .emit)
18:32
ast: ab7cddd | (Elizabeth Cholet)++ | S17-supply/watch_path.t:
Fixed deprecated "uniq" to "unique".
18:34
ast: 8167450 | lizmat++ | S17-supply/watch_path.t:
Merge pull request #50 from zrusilla/master

Fixed deprecated "uniq" to "unique".
18:36 lizmat joined 18:38 vike left
dalek ast: 98d86f1 | (Elizabeth Mattijsen)++ | S32-io/io-path-unix.t:
Fix obsolete .directory -> .dirname
18:40
kudo/newio: 62c092f | (Elizabeth Mattijsen)++ | src/core/IO/Notification.pm:
Fix deprecated Supply.more (to .emit)
18:43
kudo/newio: f4496d2 | (Elizabeth Mattijsen)++ | src/core/IO/Notification.pm:
Merge branch 'nom' into newio
18:43 vike joined 18:45 ilbot3 left 18:46 ilbot3 joined
mvuets cosimo: ping 18:47
18:51 lizmat left, denis_boyun_ joined 18:56 Khisanth left 18:58 Khisanth joined, adu joined 18:59 adu left 19:08 psch left
dalek ake: 5bc1c25 | (Arne Skjærholt)++ | t/ (2 files):
Remove most nqp::op use from test files.
19:10
19:11 prime joined
mvuets is META.info obsolete in favor of META6.json? 19:31
19:32 denis_boyun_ left 19:35 guru joined, guru is now known as Guest38251
dalek href="https://perl6.org:">perl6.org: 65b9998 | raydiak++ | source/style.css:
Fix broken nav styling
19:36
19:36 Guest38251 is now known as ajr_, ajr_ left 19:43 anaeem1_ joined, mullagain left 19:46 perltricks joined 19:49 SevenWolf left, perturbation joined 19:50 anaeem1_ left 19:52 anaeem1_ joined 19:57 kjs_ joined, kjs_ left 19:58 kjs_ joined
raydiak is anyone with the proper access trying to get feather.perl6.nl:3000 back up? 19:58
FROGGS_ raydiak: only tadzik can do that 19:59
raydiak ah
20:02 smls left 20:04 rindolf left
masak what depends on feather.perl6.nl:3000 being available? 20:07
moritz panda install
20:07 zakharyas left
moritz ok, I've restarted the app 20:08
(which involves logging in as tadzik, which is why I don't like doing it)
colomon … installing panda, or installing modules using panda?
moritz colomon: the latter 20:09
20:09 darutoko left
colomon huh. I thought module installation just used your local projects.json and github or wherever. 20:09
masak is it wise to tie something as common as panda install to something as seemingly fragile as feather.perl6.nl:3000 ?
moritz masak: nope; the plan is to make it a "proper" service 20:10
masak: at least that's my plan as soon as I have the new server
(and run it on port 80, so that fewer firewalls complain) 20:11
FROGGS_ +1
colomon wonders if it might be possible to have a backup on github or something.
20:11 KCL left
moritz but I won't invest more time on that on the old feather 20:11
20:14 anaeem1_ left, anaeem1_ joined
hoelzro o/ mvuets 20:17
good to see a familiar face here =)
dalek p: 0cf0677 | TimToady++ | src/QRegex/ (2 files):
enforce invariant of correct # of NFA fates

The mergesubstates method now always guarantees that states[0] has a size reflecting the number of fates used by this NFA. Previously it relied on its caller to set the fates list, and since /@array/ didn't do that (oops), the fate count was off, resulting in reallocations in the NFA engine. Since alternatives just use the fate numbers directly, they no longer have to set the fates. Protoregexes, on the other hand, still override the fate number with rxname, but now do so by setting it after mergesubstates sets it to the number.
20:19
20:20 colomon left
masak TimToady++ 20:20
FROGGS_ that sounds nice indeed 20:21
20:21 anaeem1__ joined
TimToady you just like it because I used the word "invariant" :P 20:21
20:22 anaeem1_ left
TimToady and while it sounds great to avoid reallocations, only the /@array/ construct was actually doing it :) 20:22
FROGGS_ I like it because it fixes an oops :o) 20:23
masak .oO( de-oopsing ) 20:26
speaking of nothing: when a class A keeps calling into class B in a way that makes you feel that logic isn't properly encapsulated in B but has leaked into A, what design principle is broken, or does that kind of code smell have a name? 20:28
moritz liked it because it was a detailed commit message (and saves allocations)
dalek href="https://perl6.org:">perl6.org: cef6e11 | raydiak++ | source/ (2 files):
Fix broken layout in /documentation

Also rearrange box colors to compensate for layout changes
20:30
arnsholt Argh! Contextual, y u no contextualize! 20:37
20:39 Mso150_p_l joined 20:40 Mso150_p left 20:43 kjs_ left 20:50 anaeem1__ left, anaeem1 joined 20:54 Sqirrel joined
dalek c: 300883d | (Rob Hoelz)++ | lib/Type/List.pod:
Document grep-index
20:55
c: 7fe985f | (Rob Hoelz)++ | lib/Type/List.pod:
Document first-index
c: b28737d | (Rob Hoelz)++ | lib/Type/List.pod:
Document last-index
20:55 anaeem1 left
timotimo o/ 20:58
dalek ake: 9680eaf | (Arne Skjærholt)++ | src/Snake/Grammar.nqp:
Allow unlimited whitespace inside enclosures.
arnsholt \o
21:03 Mso150_p_l left, Mso150_p_l joined 21:04 raiph left
moritz hoelzro++ # docs 21:20
21:21 kaare_ left 21:22 adu joined
timotimo panda refuses to work if it can't connect to the metadata server? >_< 21:39
like, it won't even install packages locally?
masak :/ 21:40
timotimo in that case, i'll ufo it up :)
21:41 raiph joined
timotimo ufo has some deprecations in it 21:41
masak haven't used ufo myself in a while.
21:41 raiph left
masak compilation is so fast now. 21:42
ugexe you can still hit the url directly 21:44
feather.perl6.nl:3000/projects.json
timotimo i can?
arnsholt masak: Yeah, rebuilding NQP after a year off was a bit of a revelation... 21:49
21:51 perltricks left
raydiak timotimo: no, you couldn't :) but moritz++ fixed it for us, all should work normally again 21:52
timotimo oh
it doesn't work for me 21:53
ugexe yeah, i can access it through a browser on my desktop but not on another machine through panda
raydiak hm...works here, now, even via panda
ugexe m: IO::Socket::INET.new(:host<feather.perl6.nl>, :port(3000)); 21:54
camelia rakudo-moar 62c092: OUTPUT«Failed to connect: connection refused␤ in method initialize at src/gen/m-CORE.setting:22518␤ in method new at src/gen/m-CORE.setting:22502␤ in block <unit> at /tmp/D4WneolcE0:1␤␤»
raydiak panda also working from my server 21:55
timotimo ah, must be because the browser falls back to ipv4 if it gets a connection refused from the port over v6 21:56
raydiak oh right, the ipv6 thing
ugexe m: IO::Socket::INET.new(:host<193.200.132.131>, :port(3000)); 22:00
camelia rakudo-moar 62c092: OUTPUT«Failed to connect: host is unreachable␤ in method initialize at src/gen/m-CORE.setting:22518␤ in method new at src/gen/m-CORE.setting:22502␤ in block <unit> at /tmp/ksDXWgsvqd:1␤␤»
ugexe wrong ip
m: IO::Socket::INET.new(:host<193.200.132.135>, :port(3000));
camelia ( no output )
mvuets hoelzro: o hai! (-: 22:01
22:02 lizmat joined, rurban left, kurahaupo joined
timotimo i can't build GTK::Simple any more: No STable at index 47 22:03
damn, has been broken in the smoke test reports for a while now 22:07
and on cpandatesters, too :( 22:08
masak 'night, #perl6
timotimo gnite masak
lizmat gnight masak 22:09
hoelzro mvuets: playing with Perl 6? 22:10
timotimo i'm still not satisfied with the performance of my shooter game on top of sdl2 :(
raydiak timotimo: but this one runs good on my computers :) 22:21
22:21 adu left 22:27 [Coke] joined
raydiak timotimo: of course by "this one" I mean example.pl in SDL2::Raw, which I assume is not the current state of your actual sdl shooter...???... 22:36
wow...windows on our media box failed updates and miraculously still boots... 22:40
mvuets hoelzro: yes - for the first time since 2009 (-: 22:43
speaking of which - i want to be helpful. so i tried to install Task::Star and it failed because LWP::Simple tests would not pass 22:45
the test in question relies on a server script which is gone 22:46
hence a call returns Nil, and a next call on the returned object, presumably, dies 22:47
[Coke] $ panda install LWP::Simple 22:48
Segmentation fault: 11
... huh
hobbs also wants to be doing something useful for/with perl6
mvuets not sure, maybe it isn't Nil: No such method 'match' for invocant of type 'Any'
so i wanted to fix the test and i wasn't sure how. would very appreciate any input so i could fix it 22:49
[Coke] mvuets: can you open an issue here: github.com/cosimo/perl6-lwp-simple/issues ?
arnsholt With Perl 6 is intrinsically useful for Perl 6, at this stage, IMO
Also, you may very well run into a bug or something that lets you contribute directly as well
bartolin m: 1, 2 ... 4, 6 ... 8, 10 ## there seems to be something wrong with chained sequence operators
camelia ( no output )
[Coke] if you want to supply a fix, I'd fork that repo and issue a pull request when done.
bartolin m: say 1, 2 ... 4, 6 ... 8, 10 ## there seems to be something wrong with chained sequence operators 22:50
camelia rakudo-moar 62c092: OUTPUT«1 2 3 4 6 8␤»
hobbs oh hi arnsholt :)
mvuets [Coke]: that's what i aim for, actually
[Coke] m: say 1, (2...4), (6...8), 10
camelia rakudo-moar 62c092: OUTPUT«12 3 46 7 810␤»
mvuets arnsholt: that's kinda my hope as well (-:
[Coke] bartolin: I suspect your 8,10 is the RHS of that last seq.
which test is the failing one? 22:51
(mv)
mvuets t/custom-headers-and-content.t
bartolin [Coke]: mom
[Coke] mvuets: ok. Step one, ping FROGGS_ to see if he can fix the server. 22:52
22:52 telex left
bartolin [Coke]: actually it's a skipped test in S03-sequence/misc.t 22:52
[Coke] step 2, we shouldn't rely on something outside the ecosystem there.
bartolin [Coke]: but there also is this example in S03: 22:53
FROGGS_ [Coke]: what's up?
bartolin m: say 1, *+1 ... { $_ == 9 }, 10, *+10 ... { $_ == 90 }, 100, *+100 ... { $_ == 900 }
camelia rakudo-moar 62c092: OUTPUT«use of uninitialized value of type Nil in numeric context in block <unit> at /tmp/Beh76NAMxB:1␤␤1 2 3 4 5 6 7 8 False False False False False False False False False False False False False False False False False False False False False False False F…»
[Coke] FROGGS_: see github.com/cosimo/perl6-lwp-simple...-content.t
FROGGS_ ups
[Coke] it points at a froggs.de URL that is dying.
22:54 telex joined
raydiak but what about github.com/cosimo/perl6-lwp-simple...tent.t#L11 > 22:54
?
bartolin m: say 1, 2 ... * > 3, 6 22:55
camelia rakudo-moar 62c092: OUTPUT«1 2 3 4 6␤»
bartolin m: say 1, 2 ... * > 3, 6 ... 10
camelia rakudo-moar 62c092: OUTPUT«Too few positionals passed; expected 1 argument but got 0␤ in block <unit> at /tmp/89Sg_IwM_r:1␤␤»
bartolin m: say 1, 2 ... { $_ > 3 }, 6
camelia rakudo-moar 62c092: OUTPUT«1 2 3 4 6␤»
bartolin m: say 1, 2 ... { $_ > 3 }, 6 ... 10
camelia rakudo-moar 62c092: OUTPUT«use of uninitialized value of type Nil in numeric context in block <unit> at /tmp/41i58xbXxW:1␤␤1 2 3 False False False False False False False False False False False False False False False False False False False False False False False False False…»
bartolin I think, I should open a ticket for that .. 22:56
FROGGS_ [Coke]: I fixed froggs.de/cgi-bin/test/test.cgi
mvuets FROGGS_++ 22:58
[Coke] the skip was there in case the site didn't respond; if it's responding now, it shouldn't skip.
mvuets you broke my use case q-:
FROGGS_ mvuets: my server had a hardware failure and it got moved... apparently not without a hickup :o) 22:59
raydiak obviously I don't understand the details of testing semantics, but wasn't the skip there so if the server had a hiccup, things *wouldn't* break? 23:01
FROGGS_ raydiak: I bet the code just dies before reaching that point
mvuets where does Test.pm come from? 23:02
rakudo?
raydiak FROGGS_: no the weirdest thing is that it's complaining about the call to match on L13 *after* that test should have skipped at L11 23:03
FROGGS_ :/ 23:04
mvuets: aye
raydiak FROGGS_: but then it goes on to say 1 passed (0 okay, 1 skipped)...and then at the end it complains 'no plan in tap output blah blah' 23:06
so color me confused :)
FROGGS_ hehe
yeah
23:11 colomon joined
hobbs sounds like: it printed a skip, then it died, and it was a plan-at-the-end TAP document so when the plan didn't come out the whole thing was malformed 23:13
skipping 1/1 is okay, skipping 1/? is a fail :) 23:14
tony-o timotimo: btyler: got timeout working on http-async, i know timotimo wants 'keep-alive' to work better, btyler anything you're angling for? 23:15
by work better i mean actually be handled properly
bartolin submitted RT #123329 wrt chained sequence operators 23:17
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...?id=123329
mvuets hobbs: sounds right! that ok() would be indeed skipped if evaluation of one of its arguments didn't die 23:18
mvuets just realized Test.pm is a part of language specification - neat! 23:19
S24 reads: The skip() function is called instead of the some tests (usually because they would die), and emits $count SKIP markers in the TAP output.
btw that test still fails, but due to some other reason now. 23:25
23:26 Isp-sec left
mvuets paste.scsys.co.uk/448715 does it look very lame? 23:26
raydiak mvuets: looks great, and now I understand why I was confused too :) 23:29
23:30 prime left
dalek kudo/newio: e84fb7e | (Elizabeth Mattijsen)++ | src/core/IO.pm:
Fix CHANGE-DIR (don't make an IO::Path)
23:31
kudo/newio: 7ed1673 | (Elizabeth Mattijsen)++ | src/core/IOU.pm:
Some whitespace fixes
kudo/newio: 4697967 | (Elizabeth Mattijsen)++ | src/core/io_operators.pm:
Make use of primitives without IO::Path
lizmat and on that thought, I'm calling it a night 23:33
23:33 lizmat left
Juerd - 23:34
s/-//
raydiak FROGGS_: that test.cgi just seems to print out a few env vars from the server...is supposed to be a json-rpc service providing an 'echo' method 23:35
FROGGS_ hmmm 23:36
raydiak: can you test it now? 23:40
mvuets i have a trivial patch (single char typo) for perl6/specs. how do i send it? what's the preferred way? 23:41
raydiak FROGGS_: works like a charm :)
FROGGS_ raydiak: \o/
mvuets FROGGS_: works here too!
FROGGS_ mvuets: a pull request
mvuets ok
23:43 spider-mario left 23:53 rmgk_ joined, rmgk is now known as Guest25135, rmgk_ is now known as rmgk 23:54 Guest25135 left 23:56 cognominal left
raydiak mvuets++: feel free to submit a PR for your LWP::Simple patch as well 23:57
mvuets raydiak: will do with pleasure after some rest - i am used to make stupid commits when sleepy 23:59