»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.
japhb_ Also, is there any project right now that builds each commit, runs all tests, installs panda, installs modules, runs their tests, etc. and then eventually tells you if the current Rakudo build is good on one or more backends and platforms? 01:43
s/Also, is/Is/
I seem to recall that's what Ementaller was supposed to be, though that's a fuzzy memory ....
japhb_ I'm basically looking for "reinstall locally on green builds", so I can stay mostly up to date, but not worry about bad rebuilds. Basically I want the 'debian testing' of Rakudo, where nom could be considered 'debian sid/unstable' 01:45
[Coke] japhb_: no one is doing that, no. 01:47
japhb_ [Coke]: What pieces do we already have? I know you do a daily build and spectest for the various compiler/backend combos. Is anyone doing any automated testing of Star or some portion of the ecosystem? 01:53
[Coke] I'm also testing star daily, yes. 02:08
I am in the middle of making it test rakudo * on non parrot vms, will probably hit early next week. 02:09
japhb_ Oooh, [Coke]++ 02:21
Actually, that gets us pretty close, doesn't it?
I guess that depends on whether Star uses Panda to do its installs these days .... 02:22
colomon japhb_: I've done a bit of fiddling with emmentaler and jvm / moar. Unfortunately, at the moment emmentaler is painfully slow on jvm and crashes completely on moar. 02:43
(mind, I don't think it's emmentaler itself that's slow on jvm; rather it appears to be compiling modules that is really slow.) 02:49
afk # sleep 02:50
japhb_ colomon: Hmmm, that's too bad. 03:37
TimToady maybe we should write a cross-compiler from moarvm to jvm :D 05:19
sergot_ for people in Poznan: today hackathon is at the same place that plpw - you can find more info here: act.yapc.eu/plpw2014/ 05:25
morning o/ btw
tadzik bleh, internet at home is at 60% packet loss... good thing I'm leaving for Poznan soon 05:38
TimToady when the internet has fast lanes, you'll be able to lose packets even faster 05:40
tadzik can't wait :) 05:44
tadzik oh, it's back \o/ 06:00
nwc10 dear experts at the hackathon. Currenty when I make spectest on x86_64 Linux, some fail. Please could they all pass, or be skipped? 07:23
lizmat experts are not at the hackathon yet, but are about to move there :-) 07:33
back in a few & 07:34
tadzik on rails o/ 07:54
Woodi hallo :) 08:00
Happy Hackatons ! 08:01
do Perl6 needs Memory Model or somethink ? :)
tadzik: is it possible to have office apps construction kit as side-effect of toying with SDL ? 08:03
masak interesting message on p6u. 08:03
(about "does Perl 6 have a way of ensuring method overrides?")
I know it has it for roles, but I haven't seen a similar mechanism for classes and inheritance. 08:04
Woodi methods "overrides" ? not overloading ? 08:05
masak that's what he asked about; see the message :) 08:09
I know how to do method overloading; just use "multi" on stuff. 08:10
it even works across an inheritance hierarchy.
jnthn Woodi: Certainly needs some think :P 08:11
jnthn Woodi: But yeah, will need to write up memory model. It's on my todo list at some point... :) 08:11
masak .oO( the missing think )
jnthn o/ masak
masak jnthn \o
tadzik Woodi: you could write sort of a gui toolkit on top of it, I guess
masak I would join you for the hackathon today, but I'm busy teaching... 08:12
jnthn masak: Method overrides? All dispatches are virtual... :)
masak I know, I know.
tadzik but it's probably more reasonable to just use gtk or something
masak but that's not what he's asking about.
see the message.
jnthn OK...link?
masak www.nntp.perl.org/group/perl.perl6....g1755.html 08:19
jnthn masak: Not build in, but doable as module that assd an "is override" trait 08:23
uh...adds
masak .oO( Not At All Freudian ) 08:27
moritz jnthn: I've tried to implement such a trait yesterday 08:34
jnthn: twice, in fact
jnthn: my first attempt was a trait is-declared-in(RoleName) 08:35
jnthn: but it seems that method traits can't receive arguments (yet?)
masak hm, do you have to specify the role?
seems a bit over-specific.
moritz jnthn: second attempt was to omit the argument, and walk the superclasses and roles
jnthn: but that failed, because at trait run time, the class isn't composed, and thus those lists aren't available 08:36
jnthn r: multi trait_mod:<is>(&r, :$omg!) { say $omg }; sub foo is omg(42) { }
camelia rakudo-{parrot,jvm,moar} 2e4961: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Can't use unknown trait 'is omg' in a sub declaration.␤at /tmp/tmpfile:1␤------> ␤ expecting any of:␤ rw parcel hidden_from_backtrace␤ pure default …»
masak so maybe delay the check until class composition time?
moritz masak: how do I do that in a trait? 08:37
masak maybe it's not possible.
I heard mumblings about a COMPOSE phaser a while ago...
moritz m: multi trait_mod:<is>(&r, :$omg!, |c) { say $omg; say c.perl }; sub foo is omg(42) { }
camelia rakudo-moar 2e4961: OUTPUT«===SORRY!=== Error while compiling /tmp/D0CIDeU58D␤Can't use unknown trait 'is omg' in a sub declaration.␤at /tmp/D0CIDeU58D:1␤------> ␤ expecting any of:␤ rw parcel hidden_from_backtrace␤ pure default DEPRECA…»
jnthn r: multi trait_mod:<is>(&r, $arg, :$omg!) { say $arg }; sub foo is omg(42) { } 08:38
camelia rakudo-{parrot,jvm,moar} 2e4961: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Can't use unknown trait 'is omg' in a sub declaration.␤at /tmp/tmpfile:1␤------> ␤ expecting any of:␤ rw parcel hidden_from_backtrace␤ pure default …»
jnthn grr, shoulda m
hmm, wtf..
moritz m: multi trait_mod:<is>(Routine:D $r :$omg!, |c) { say $omg; say c.perl }; sub foo is omg(42) { }
camelia rakudo-moar 2e4961: OUTPUT«===SORRY!=== Error while compiling /tmp/K7IHNH7xCY␤Variable '$omg' is not declared␤at /tmp/K7IHNH7xCY:1␤------> multi trait_mod:<is>(Routine:D $r :$omg⏏!, |c) { say $omg; say c.perl }; sub foo␤»
moritz m: multi trait_mod:<is>(Routine:D $r, :$omg!, |c) { say $omg; say c.perl }; sub foo is omg(42) { }
camelia rakudo-moar 2e4961: OUTPUT«42␤Capture.new()␤»
moritz oh
I expected an extra argument
gist.github.com/moritz/2c6ed01eef0029dabdeb works now 08:40
moritz it wouldn't hurt to have compose-time hooks in ClassHOW 08:45
masak \o/
moritz: ship it!
(and I agree about compose-time hooks)
actually, hooks in as many places as possible :) 08:46
everywhere it makes sense
moritz: s/mimi/mimic/ in the gist title
jnthn Typically I have written a meta-object and overridden compose and callsame'd to the usual one 08:47
moritz masak: typo-fixed 08:50
masak moritz++ 08:52
moritz ... and e-mail shipped :-)
dalek kudo/nom: b700214 | (Elizabeth Mattijsen)++ | src/core/Bag.pm:
Make Bag.WHICH on demand and remove old cruft
 08:53
masak "Feel inspired :-)" \o/ 08:55
moritz and the actual implementation is about 4 LOC 08:56
would be a bit more with proper typed exceptions
jnthn: if I were to go the way of custom meta class that overrides .compose, could I make the class use that metaclass from within the trait? 08:57
tadzik moritz++ 08:58
Woodi: btw, I think Gtk::Simple is already in the module ecosystem somewhere, or at least on github 09:02
lizmat moritz: now, to put this in a module, would adding "is export" to the trait mode be enough?
moritz lizmat: I think so 09:03
lizmat: well, and I'd need some docs :-)
tadzik and tests. Sounds like a nice hackathon task for a newcomer :) 09:04
nwc10 Mmmm, possibly PEBKAC 09:06
a.k.a. as git pull 09:07
or, failure to be on "this week"
masak ok, I checked out the hackathon in the next room 09:09
I approve :)
hackathon++
nwc10 they have beer?
jnthn masak: hah, you managed to drop by in the one moemnt I wasn't there :P
masak jnthn: it's the masak-jnthn exclusion principle!
jnthn moritz: No, typically your module just exports the meta-class along with the trait... 09:10
masak jnthn: have you ever seen them in the same room? suspricious, huh?
nwc10 I can't remember if I'd read this before, but I re-read it today, and it feels an interesting analsys of the problems of migrating: blog.startifact.com/posts/python-2-gravity.html
jnthn masak: No, I've never seen myself in a room with you... :P
masak :P
nwc10: will read later. 09:11
nwc10 the IRC channel mentioned at the end is logged here: chat-logs.dcpython.org/channel/python2.8 09:12
related blog I found from the log is here: regebro.wordpress.com/2014/01/23/th...ython-2-8/
dalek kudo/nom: 94bd372 | (Elizabeth Mattijsen)++ | src/core/Mix.pm:
Make sure Mix.WHICH is value based, and on demand
 09:17
Woodi tadzik: I was thinking about SDL framework becouse thes days everything started to look Metro-like, even Qt. all that old-school buttons are missing ;) 09:57
techgage.com/viewimg/?img=/articles...htroom%204
and SDL is light and portable 09:58
tadzik well, you can style qt all you want, I think :) But it should be possible to write a windowing toolkit in SDL, sure 10:10
I'll probably need to do it myself at some point too, even for simple things like achievement popups :P 10:11
timotimo i believe steam actually uses sdl for their UI drawing needs 10:40
tadzik some of them, maybe, but a lot of it is HTML and JS 10:41
and you can style it with css and whatnot
timotimo they embed a webkit, yeah 10:44
but the stuff around it, i believe, is being rendered by qt
well, software rendering anyway.
lizmat lunch&
tadzik oh, there's beer festival here 10:45
lizmat: where are you going?
I'll be at the venue in 15 minutes or so
liztormato
.oO( tadzik has arrived at lunch )
 11:18
dalek p: 6a7aa4b | jonathan++ | src/NQP/Optimizer.nqp:
Make sure we callstatic the NQPCORE.setting subs.
 12:02
dalek Heuristic branch merge: pushed 460 commits to rakudo/eleven by lizmat 12:21
nwc10 does the hackathon dissolve into dinner at some point this evening?
FROGGS I hope so 12:22
the plan is not to starve but to survive until tomorrow
moritz starving doesn't sound like -Ofun 12:23
tadzik heheh, I have two pull requests with indentical patches :)
and they're from the same day 12:24
dalek nda: d47e254 | (Ben Tyler)++ | lib/Panda/Common.pm:
Track change in $*VM name access
nda: 05af361 | tadzik++ | lib/Panda/Common.pm:
Merge pull request #81 from kanatohodets/master

Track change in $*VM name access
kudo/nom: f57749e | (Elizabeth Mattijsen)++ | src/core/Str.pm:
Fix deprecation inside CORE
 12:26
colomon I thought the plan was to wait a bit on that fix, to maintain support for older Rakudos? 12:30
colomon tadzik: ^^ 12:32
colomon had been pondering solutions involving $*VM.?name ...
tadzik colomon: oh, I tagged panda a commit before, so people with older rakudos can check it out 12:33
colomon tadzik++
tadzik see github.com/tadzik/panda/releases
dalek kudo/eleven: 2c73b2a | (Elizabeth Mattijsen)++ | src/core/ (2 files):
Fix some more deprecations
 12:53
dalek kudo/eleven: 5090deb | (Elizabeth Mattijsen)++ | src/core/CompUnitRepo/Local/Installation.pm:
Tweaks
 13:27
colomon lizmat: did you mean for those last changes to go to rakudo/eleven instead of nom? 13:29
lizmat yes
preparing stuff for merge.. :-)
eleven bitrotted a bit since March 13:30
[Coke] (√(-shit))²
[Coke] "shit just got real" was the tagline I saw on that. :) 13:39
tadzik :D
sjn isn't that strictly "shit isn't imaginary any more now?" :) 13:40
I think I see something more there... 13:41
sjn you start out with shit, which everyone knows isn't good. it's a negative word 13:41
dalek ecs: ce1aeac | (Elizabeth Mattijsen)++ | S28-special-names.pod:
Spec $*VM.precomp-ext
sjn then you say "no, /this/ shit isn't negative, we're making it positive by negating it!" 13:42
but there's a problem!
everyone knows that the root of positive shit is imaginary
unless you're a square. :) 13:43
tadzik . o O ( this discussion is full of shit )
sjn if you put this imaginary positive shit in a square, everything just becomes nice :)
real good shit, in other words 13:44
so what's the point with this story? 13:45
if you want to make real good shit, you need imagination. :)
brrt bloody netsplts 13:46
dalek ecs: 0852ee7 | (Elizabeth Mattijsen)++ | S28-special-names.pod:
Spec $*VM.precomp-target
 13:46
dalek kudo/nom: 8aa5c73 | (Elizabeth Mattijsen)++ | src/core/VM.pm:
Implement $*VM.precomp-(ext|target)
 13:58
volaj/tadzik/deprecations: 53317e5 | tadzik++ | / (2 files):
Get rid of deprecation messages
tadzik so this fixes zavolaj on new rakudos, but breaks it on old rakudos 13:59
dalek kudo/eleven: f57749e | (Elizabeth Mattijsen)++ | src/core/Str.pm:
Fix deprecation inside CORE
kudo/eleven: 8aa5c73 | (Elizabeth Mattijsen)++ | src/core/VM.pm:
Implement $*VM.precomp-(ext|target)
kudo/eleven: 2e87d47 | (Elizabeth Mattijsen)++ | src/core/VM.pm:
Merge branch 'nom' of github.com/rakudo/rakudo into eleven
tadzik regarding $*VM<name> vs $*VM.name etc
[Coke] tons of modules complaining about Please use $*VM.name instead. 14:03
tadzik yep, I fixed two
dalek kudo/loop_labels_test: 876bb54 | (Tobias Leich)++ | src/core/ (2 files):
implement labeled for loops on parrot
 14:04
dalek ecs: 14fb9cb | (Elizabeth Mattijsen)++ | S99-glossary.pod:
Some new lemma's: RC UGT
 14:15
lizmat note to self: next time, don't deprecate things immediately ? 14:18
tadzik well, it would happen at some point anyway 14:19
dalek p/loop_labels: 1655e4e | (Tobias Leich)++ | src/vm/jvm/ (3 files):
pass the label obj around instead of its addr
 14:35
p/loop_labels: eafcab0 | jnthn++ | src/vm/moar/QAST/QASTOperationsMAST.nqp:
Start using new getlexstatic_o op.
 14:36
nqp/loop_labels: 43fbe2b | jnthn++ | src/NQP/Optimizer.nqp:
nqp/loop_labels: Fix NQP analysis of when callstatic is valid.
[Coke] wonders if it's worth caching the repos for star/parrot/moar etc. on star when I have to grab all the modules over the interwebs anyway. 14:39
(er, on the daily star)
tadzik jnthn: Unhandled exception: Bytecode stream corrupt (missing magic string) 14:40
that's what I got from gdb, outside of gdb it just says "segfault"
jnthn tadzik: wow... 14:45
That's really corrupt....
Tht's the very first thing it reads in the file 14:46
tadzik :)
it must be really confused
yeah
Woodi zavolaj ? I thinked it evolved into Native Calls... and no such module on p6.org...
timotimo zavolaj is the name of the repository 14:47
tadzik yeah, it's binamual
timotimo NativeCall is the name of the module
Woodi ah, it's repo name
dalek kudo/eleven: 6a81129 | (Elizabeth Mattijsen)++ | src/Perl6/ModuleLoader.nqp:
Performance debugging tweak
 14:55
kudo/eleven: 90fdd07 | (Elizabeth Mattijsen)++ | src/core/CompUnitRepo.pm:
Use $*VM.precomp-ext
timotimo can we theoretically do something about the proxy being used in CArray? 14:58
timotimo now that we have assign_pos, for example? 14:58
jnthn Yes 14:59
But for native arrays we should be able to do even better I guess...
masak nwc10: really interesting post, blog.startifact.com/posts/python-2-gravity.html 15:00
nwc10: I thought the "gravity" metaphor wouldn't hold up -- but it does, amazingly.
dalek p: 96ef351 | jonathan++ | src/NQP/Optimizer.nqp:
Can also lower $_.

Before, this earned us a SEGV in the Rakudo setting compile, but now that issue is fixed, we can do this optimization.		 15:07
timotimo oh yay! :)
masak is "lower" a technical term? 15:09
also, what about $! ? :)
timotimo lower from a lexical to a local
might want an entry in S99 15:10
jnthn: which multi candidates should i implement for assign_pos? pos, assigner == int, int / Int, int / int, Int / Int, Int? 15:13
masak ah, that makes sense. timotimo++ 15:14
brrt (interesting post about py2 / py3) 15:18
jnthn timotimo: Could work out 15:20
nwc10 it seems to keep coming back to the point/problem that existing end user codebases don't get re-written. They get replaced. 15:21
brrt or! 15:23
masak nwc10: Rule 1 about community-wide backwards-incompatible upgrade: you can't upgrade the community. 15:26
PerlJam masak: does that mean we're ... the replacements?!?
nwc10 you need a better carrot 15:27
sticks don't work
masak carrot's don't work either.
PerlJam neither carrots nor sticks work
masak nothing works.
see Rule 1.
PerlJam desire (self-impulse) works. If you're not "self", you can't affect much change. 15:28
nwc10 OK, I was meaning, make something sufficiently better that people will want to move it
move to it
but, in any non-trivial setup, that move wont' be a flag day
PerlJam
.oO( ergo, we need more trivial setups ;)
 15:29
masak nwc10: I teach Python these days. my clients are squarely in 2.7, because their code is. some still use Python 2.6. (mostly by accident, not so much by intent.) 15:29
nwc10: we offer to give the course using Python 3 as basis. so far, no corporate client has taken that bait. they want Python 2.x
TimToady if those durn language designers would only get it right in the first place...
nwc10 TimToady: yeah, the pesky Rule 2 getout 15:30
TimToady ...then we could all be programming in Lisp :)
masak nwc10: Perl course -- same deal, except even more so. the most promising I've had is one group that were so happy about the three days they got so far, that they clamored for the final "Perl 6" section at the end of the course. but I don't think they'll be doing any Perl 6 development at work. 15:31
TimToady oh wait...
nwc10 masak: no-one sane is claiming that Perl 6 is production ready yet (even if the concurrency stuff is already good enough to make it useful for some tasks) 15:31
nwc10 whereas Python 3.1 was 15:31
masak *nod* 15:32
PerlJam People still keep waiting for this magical perl 6.0.0 release before even *thinking* about using Perl 6 for anything.
[Coke] To sell it here, I need a web stack that works with Oracle, preferably java, adequate docs, and commercial support.
masak well, fwiw, I don't consider "production-ready" to be a single global Bool.
which is another way of saying that I am probably insane, at least here and there, for certain uses.
[Coke] (and if you have to just meet adobe's level of support for coldfusion, that's not hard. :P) 15:33
masak [Coke]: interesting.
[Coke]: having a potential end user is an enticing prospect. 15:34
[Coke]: maybe we should meet up sometime (preferably on IRC) and talk requirements?
PerlJam What, exactly, was Ian Hague's insterest in Perl 6? Was there some project in the future that wanted it? Or just general, "I think this is the future, so I'm gonig to support it"? 15:35
nwc10 I believe the latter. In that he wanted to give something back, and did so, very generously 15:37
PerlJam was hoping for another "end user" that wasn't on #perl6 :( 15:38
[Coke] masak: sure. 15:39
I also have one project that is doing a lot of text generation that I'd love to be able to switch to a threaded perl6. 15:40
dalek ecs: abb8d13 | (Elizabeth Mattijsen)++ | S28-special-names.pod:
Small typos fix
 15:48
[Coke] (REPR: Null, cs = 0) - someone just opened a ticket with this. check out #81060 also. 15:50
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...l?id=81060
dalek ecs: 9294dd4 | (Elizabeth Mattijsen)++ | S28-special-names.pod:
Spec $?PERL/$*PERL and the Perl class
 15:58
dalek kudo-star-daily: 2c5ab39 | coke++ | / (2 files):
grab copies of repos to work with locally
 16:05
kudo-star-daily: c85f012 | coke++ | / (2 files):
move script to bin dir
kudo-star-daily: 86b7004 | coke++ | bin/ (2 files):
Begin to use local repos.
kudo-star-daily: f71fdf1 | coke++ | bin/star.sh:
use parrot-latest, not parrot-recommended
kudo-star-daily: 556da35 | coke++ | bin/star.sh:
Don't conflict with parrot's VERSION
kudo-star-daily: df86539 | coke++ | log/ (5 files):
today (automated commit)
[Coke] ^^ ok, we're now using parrot-latest, not parrot-recommended for the daily star testing. should let us know sooner of any issues. 16:06
dalek p: a438ec8 | jonathan++ | src/NQP/Optimizer.nqp:
Optimize simple 'for' loops.

Can avoid a lot of block invocations in some cases, making things a bit cheaper.
donaldh nqp-m: my $ok := 1; try { nqp::die($ok); $ok := 0; }; nqp::say($ok);
jnthn That should be another little saving on CORE.setting compilation (and other things)
camelia nqp-moarvm: OUTPUT«␤»
donaldh nqp-m: my $ok := 1; try { nqp::die($ok); $ok := 0; 1 }; nqp::say($ok);
camelia nqp-moarvm: OUTPUT«1␤»
donaldh something weird with nqp-m try blocks ^^^ 16:07
donaldh nqp-m: my $ok := 1; try { nqp::die($ok); $ok := 0; 1 }; nqp::say($ok); 16:28
camelia nqp-moarvm: OUTPUT«1␤»
donaldh nqp-p: my $ok := 1; try { nqp::die($ok); $ok := 0; }; nqp::say($ok);
camelia nqp-parrot: OUTPUT«1␤»
donaldh nqp-j: my $ok := 1; try { nqp::die($ok); $ok := 0; }; nqp::say($ok);
camelia nqp-jvm: OUTPUT«(signal )#␤# There is insufficient memory for the Java Runtime Environment to continue.␤# pthread_getattr_np␤# An error report file with more information is saved as:␤# /home/p6eval/hs_err_pid29070.log␤»
donaldh oh, latest nqp commit is broken on JVM backend. 16:43
donaldh P6opaque representation does not implement elems 16:54
in commit a438ec8b70
dalek p: 0ea245b | (Donald Hunter)++ | src/vm/jvm/ (2 files):
Added nqp::seekfh on JVM.
 17:08
p: d4c8ae3 | (Donald Hunter)++ | src/vm/parrot/QAST/Operations.nqp:
Added nqp::seekfh, nqp::readfh on parrot backend.
p: 616e461 | (Donald Hunter)++ | t/nqp/19-file-ops.t:
Add tests for seekfh.
p: 53eae59 | (Donald Hunter)++ | src/vm/jvm/runtime/org/perl6/nqp/io/FileHandle.java:
Fix seekfh failures on JVM backend.
p: 612cf2d | (Donald Hunter)++ | src/vm/jvm/runtime/org/perl6/nqp/sixmodel/reprs/P6OpaqueBaseInstance.java:
Add elems to P6opaque.
 17:09
thou masak: my experience is that it's mainly lib compatibility that keeps people away from python3. I think we're closing in on that state for perl6 (a ways off yet, true, but I'm impressed w/ the progress). 17:12
vendethiel o/ 17:15
japhb .ask lizmat In spec commit 9294dd4c7945be95b56ee991532abaa608917cb5, why just a build-date at Date precision? Why not a DateTime? 17:49
yoleaux japhb: I'll pass your message to lizmat.
timotimo jnthn: that for optimization looks very clever! 17:50
let the other parts of the optimizer see if it can inline if we can replace decl with param and if it didn't turn out to work (or be necessary or whatever), we just turn it right back 17:51
japhb timotimo: reverse param and decl in your description :-) But yeah, that's a cool idea -- kindof speculative optimization 17:55
timotimo uh yes 18:11
timotimo i wonder if it'd be possible to do something very similar on top of rakudo rather than nqp 18:16
timotimo if we know lots of stuff, a for ^10 { } could become as cheap as a loop 18:17
japhb Anyone have any idea of the intent of github.com/rakudo/rakudo/blob/nom/...L1186-1189 ? I have several ideas of what the author was intending to write, but none of my explanations so far allow the code to be WAI. 18:18
timotimo japhb: if you want a cheap performance win there, foo | bar | baz is optimized, whereas any(<foo bar baz>) is not 18:19
but i don't think this function is used in anything hot
japhb timotimo: Fair enough, but in this case, I just can't see the code as being *correct*. But since I'm slowly recovering from something that makes my brain very fuzzy, I'm compensating by assuming a much greater likelihood that I'm just missing something obvious. 18:21
timotimo ah 18:24
i didn't even look into it much :)
vendethiel (is `is override` gonna make it into the core ? seems like something people might want) 18:25
timotimo m: say "0x99".Numeric;
camelia rakudo-moar 8aa5c7: OUTPUT«153␤»
timotimo the <= is indeed weird. 18:26
timotimo okay, i have no idea what that code is supposed to exactly do under which circumstances and what is considered correct input 18:27
japhb timotimo: Yeah, that's pretty much where I was with it. 18:28
And since I wrote Str.Numeric, it was making me question my sanity.
timotimo ah! :) 18:30
nwc10 TimToady: I'm not sure about the need for a liquid nitrogen overclocked ARM $whatever 18:31
(something with MOAR RAM would be fine) 18:32
acutally, right now, I'm blocking on a single thread on x86_64
because I'm insisting on building with ASAN and assertions
nwc10 and I've not figured out how to get MOAR's Configure.pl to build with the optimiser *and* assertions 18:32
the single thread is the Rakudo setting 18:33
but NQP's build doesn't paralleise much either
parallel tests for the win
backlash_jack hey guys, can anyone tell me how to contact nwc10 from this post p6weekly.wordpress.com/2014/05/12/2...-invaders/ ? 19:00
[so i can send him some MIPS hardware] 19:01
dalek kudo/nom: 6413980 | (Geoffrey Broadwell)++ | src/core/Temporal.pm:
Fix incorrect order of operations in sleep-timer()
japhb backlash_jack: He watches this channel. 19:02
btyler backlash_jack: nwc10 is often active in this channel, he was last around about 30 minutes ago. I imagine you'll see him soon
backlash_jack ok, i'll just idle until then ;)
btyler (also, cool!)
timotimo backlash_jack: cool! 19:03
backlash_jack hrm, maybe i shouldn't idle with the web client 19:06
i'll be back later though
later guys 19:07
liztormato .tell japhb DateTime sounds fine to me. I guess I got distracted by the attribute name 19:21
yoleaux liztormato: I'll pass your message to japhb.
dalek ast: eeac0ec | (David Warring [email@hidden.address] | integration/advent2013-day19.t:
disable test which should be using tap_ok Test::Tap
 20:10
dalek ast: 7e38376 | (David Warring [email@hidden.address] | integration/advent2010-day14.t:
adding 2010 advent day 14
 20:13
backlash_jack re 20:26
dalek kudo-star-daily: 60a9995 | coke++ | / (11 files):
start keeping backend-specific logs
 20:38
kudo-star-daily: 9fde22f | coke++ | log/parrot- (5 files):
today (automated commit)
thou backlash_jack: you can use the bot to .tell nwc10 something whenever they are next active 21:33
backlash_jack awesome thanks 21:36
which bot tho?
geekosaur just use .tell
japhb .tell backlash_jack Like this 21:37
yoleaux japhb: I'll pass your message to backlash_jack.
19:21Z <liztormato> japhb: DateTime sounds fine to me. I guess I got distracted by the attribute name
backlash_jack does it go out to the channel? i'd rather not have my work email address on the internet
yoleaux 21:37Z <japhb> backlash_jack: Like this
japhb You can /msg yoleaux the command
geekosaur but you can't prevent it from being read back in public... 21:38
backlash_jack i'll just wait then
japhb But yeah, the problem is that when yoleaux notices him come back, it will speak publicly anyway. :-(
backlash_jack i know, i have been on irc before ;)
japhb You can .tell nwc10 to contact you when he's back, that will at least let him know you're waiting before he slogs through the backlog 21:39
backlash_jack: There's no reason a bot couldn't notice that you sent a message privately, and pass it on privately ... this one just doesn't. :-/
backlash_jack it's ok, i'm patient, also at work and not going anywhere
japhb he
backlash_jack i used to run a bot in #punk years ago 21:40
timotimo the proxies for store in CArray return self rather than the assignee; is that correct? 21:41
timotimo oh wait 21:42
the self in there most probably refers to the proxy, not the CArray
timotimo jnthn: github.com/jnthn/zavolaj/pull/40 would appreciate a hint if this is correct or should be changed 21:50
timotimo bbl 21:57
thou on occasion, something inside and something outside seem unusually connected; like deja vu's persuasive illusion, a tangential conversation or a brief series of moments seems compellingly relevant. tell backslash_jack Like this. I'll pass your message. Oh, it sounds fine to me, I guess I got distracted. Like this. You can message the command, but you can't prevent it from being done with in public. You're not 22:09
in control. I'll just wait, then. I know, I have been here before. #punk
the self in there most probably refers to the proxy 22:10
is it all, really, poetry; and I'm mostly blind? 22:11
backlash_jack omg i'm caught in a markov chain 22:13
tadzik blogs.perl.org/users/ovid/2014/05/a...n-vim.html wow, this is awesome :) 22:21
jnthn timotimo: Well, may want num candidates too 22:22
timotimo: but looks sane
lizmat m: gather { await start { take 42 } } # jnthn: golfed threaded gather/take pb 22:39
yoleaux 17:49Z <japhb> lizmat: In spec commit 9294dd4c7945be95b56ee991532abaa608917cb5, why just a build-date at Date precision? Why not a DateTime?
camelia rakudo-moar 641398: OUTPUT«Unhandled exception: No exception handler located for take␤ at <unknown>:1 (/home/p6eval/rakudo-inst-1/languages/perl6/runtime/CORE.setting.moarvm:throw:4294967295)␤ from src/gen/m-CORE.setting:18257 (/home/p6eval/rakudo-inst-1/languages/perl6/runt…»
jnthn lizmat: That's arguably an issue, but not the one that bothers me. 22:40
m: await do for 1..4 { gather { .take for 1..10 } } 22:41
camelia ( no output )
jnthn m: say await do for 1..4 { gather { .take for 1..10 } }
vendethiel dammit
camelia rakudo-moar 641398: OUTPUT«(timeout)»
dalek ecs: 013a6e0 | (Elizabeth Mattijsen)++ | S28-special-names.pod:
Build-date is a DateTime, japhb++
vendethiel I went back to ST (home) after working with emacs all week ... I keep hitting ctrl-x T_T
donaldh github.com/rakudo/rakudo/pull/269 Start removing backend specifics in IO.pm
jnthn m: .say for await do for 1..4 { gather { .take for 1..10 } }
camelia rakudo-moar 641398: OUTPUT«(timeout)»
jnthn hmm 22:42
lizmat jnthn: there is no start in there, so what are you awaiting?
lizmat good UST, #perl6 22:45
sleep&
vendethiel uh, doesn't use lib '.' work ? 22:47
backlash_jack there's an emacs plugin for ST that lets you use ctrl-x
vendethiel backlash_jack: I just used emacs beause I had the computer for one week and didn't want to install everything for only one week
not because I find it the best :) 22:48
backlash_jack i woulda used alpine lol
if i had to switch for a week, anyway
once i went to a job interview and the asked me if i was familiar with the 'six' editor 22:49
vendethiel well, I had used emacs (for maybe 3 months) but I really didn't like it so went back to ST
backlash_jack and i was like "umm, no, i've actually never heard for that one."
so after it was over i realized that they meant 'vi'
and i was all :q! 22:50
vendethiel o_o
huf six... ;D
backlash_jack i'm absolutely dependent on emacs auto-indent, and i don't think that feature is coming to any other editors anytime soon 22:51
vendethiel oh, there are certainly ST plugins for that
backlash_jack that work the same as emacs?
vendethiel well, I can't actually judge them. First time I talked about linux's V system, I said V as the letter ¬¬
I think I've seen a plugin that does
backlash_jack so, if you forget a semicolon or something, your indenting gets screwed up? 22:52
vendethiel like in emacs, i guess
backlash_jack the main reason i never learned python very well was my dependence on emacs knowing perl ;)
vendethiel { "keys": ["f12"], "command": "reindent"} <- pretty limited tho 22:53
backlash_jack i like ST, but i only use it in windows.
vendethiel you have to select the line. I guess you could make a macro ...
oh right : { "keys": ["ctrl+k+d"], "command": "reindent" , "args": {"single_line": false}}
backlash_jack i had some config setting that let you move forward and backward with the tab or shift tab, 22:54
but it's just not the same. emacs is reading what you type
vendethiel yeah, and sometimes I have to indent manually because it can't quite figure out what I'm doing
esp. with `if ();`
timotimo jnthn: num candidates for the Int CArray? 23:01
vendethiel m: my $a; for (^15) { LAST $a = $_; last if $_ > 5; }; say $a;
camelia rakudo-moar 641398: OUTPUT«Cannot call method 'gist' on a null object␤ in sub say at src/gen/m-CORE.setting:13622␤ in block at /tmp/gMAO9lFuvq:1␤␤»
vendethiel mh. 23:02
vendethiel m: my $a; for (^15) { LAST { $a = $_; }; .say; last if $_ > 5; }; say $a; 23:05
camelia rakudo-moar 641398: OUTPUT«0␤1␤2␤3␤4␤5␤6␤6␤»
vendethiel nice~
jnthn timotimo: no, the Num one 23:14
Where bindattr_n is used
m: .say for await do for 1..4 { start { gather { .take for 1..10 } } } 23:15
camelia rakudo-moar 641398: OUTPUT«(signal )1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤Unhandled exception in code scheduled on thread 139833294665472␤Unhandled exception in code scheduled on thread 139833286272768␤1␤2␤This type does not support positional operations␤ in method rei…»
jnthn lizmat: Thanks :)
lizmat: It's almost as if I drank a few nice beers this evening :P
timotimo i thought i added those 23:55
i gotta run now, though
ttyl :)

Powered by Cro and the Raku® Programming Language.

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

Thank you!