Parrot 5.1.0 "Zombie Parrot" | parrot.org/ | Log: irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC
Set by moderator on 3 March 2013.
00:54 contingencyplan joined, Patterner joined, slavorg joined, janus joined, perlite joined, particle joined, Timbus joined, cxreg joined, Hunger joined, jsut joined, PerlJam joined, dukeleto joined, prammer joined, drift joined, wagle joined, dngor joined, mj41 joined, ingy joined, sri joined, cosimo_ joined, moritz joined, Tene joined, allison joined, knewt joined, patspam_ joined, Coke joined, pjcj joined, pmichaud joined, p6eval joined, elmex joined 00:55 mtk joined, sivoais joined, atrodo joined, tokuhirom joined, japhb joined, cotto joined, rurban joined, chekkaa joined, Khisanth joined 01:06 kid51 joined 01:51 chekkaa joined 02:11 kid51_ joined 03:38 preflex_ joined 03:56 myhrlin joined 04:06 myhrlin joined 04:27 perlite joined 04:28 eternaleye joined
dalek rrot: d7db4f5 | util++ | MANIFEST:
Fix MANIFEST from merge of branch "ffa_sort"
 05:02
rrot: a3c2754 | util++ | / (2 files):
Fix POD from merge of branch "ffa_sort"
rrot: 5120d02 | util++ | src/io/utilities.c:
Revert "Merge pull request #943 from gerdr/fix-socket-readline"

This reverts commit fb2e9570d25950c568edc50c1f9f5b1e0275bc2a, reversing changes made to 83bb8639bc79699a16d594d53ca8abd92788ab03.
This is needed for the duration of the 5.2.0 release process, because the branch was merged prematurely, and causes test failures. When the branch is in a completed state, and ready to merge, it is important that a
  "revert-the-revert" is performed just before the merge:
   git revert ID-of-*this*-revert
For rationale, see:
  www.kernel.org/pub/software/scm/gi...-merge.txt
rrot: dfc3d7c | util++ | / (7 files):
Prepare for the 5.2.0 release
05:12 travis-ci joined
travis-ci [travis-ci] parrot/parrot#826 (master - dfc3d7c : Bruce Gray): The build passed. 05:12
[travis-ci] Change view : github.com/parrot/parrot/compare/d...c3d7c34f57
[travis-ci] Build details : travis-ci.org/parrot/parrot/builds/5752380
05:12 travis-ci left 05:25 cosimo joined, simcop2387 joined 05:28 arnsholt joined 07:20 xcombelle joined
dalek kudo/nom: b972ca3 | (Tobias Leich)++ | src/Perl6/ (3 files):
fixed spello as diakopter++ pointed out
 09:07
11:37 PacoAir joined 12:45 kid51_ joined 14:15 myhrlin joined 14:22 Psyche^ joined
ptc kid51: hi kid51, how's things? 14:32
kid51 ptc: Good. 14:35
Spending more time these days in perl5-porters than in parrot
But, since, things like parrot guts are not my specialty, and since perl5 components of parrot are stable, that's okay 14:36
ptc what are you working on in perl5-porters? 14:37
kid51 Much the same sort of stuff I did in Parrot
reviewing old bug tickets, trying to get them closed
trying to be a first responder on new tickets 14:38
ptc yeah, trying to see if I can do a bit of that in parrot :-)
there's quite a lot to do
kid51 Recently added a lot of tests to Data-Dumper to boost its test coverage
ptc I also need to get up to speed with where the project currently is... it's been a while
kid51 Well, our momentum really slowed from about Sept 2011 onward 14:39
ptc why?
kid51 Well, everyone has his opinion about that 14:40
ptc oh
kid51 For example, chromatic (modernperlbooks.com) has been proclaiming Parrot dead for > 1 year
ptc that's not good; chromatic was a big influence when I was still active in parrot. 14:41
kid51 He has some valid points, but the surprising thing was that as he was nailing the coffin shut, the decedent woke up
Well, here's what's probably important ...
Parrot is intended to be "a virtual machine aimed at all dynamic languages" 14:42
So, who are its human users?
Its customers?
Language designers and implementers.
A very small number of people.
ptc there are heaps of people involved in the devel of p5, p6, ruby, python etc... 14:43
is that then not large enough?
kid51 So, unless you get a large %age of those small people saying, "Our current VMs are inadequate," you don't have an inherently large demand for your product/project.
ptc ture
s/ture/true/ 14:44
it's a pity really. It's a great idea, and the project seemed to be moving so well
kid51 After all, the hassle of writing a new interpreter for your current HLL has to be less than the hassle of using your current interpreter for it to be worth your while
ptc of course. Hard to sell... 14:46
kid51 So that means that, when it came to anything other than Perl 6, it was difficult for us to attract n > 1 developers for any given HLL at any one point in time.
dalek rrot: 7c0d17e | paultcochrane++ | t/pmc/resizablefloatarray.t:
[GH #927] Added tests for sort() method of ResizableFloatArray PMC
rrot: 76e83f7 | paultcochrane++ | t/pmc/resizablefloatarray.t:
[GH# 927] Test output more specific wrt how float array was resized
rrot: fee04de | paultcochrane++ | t/pmc/resizablefloatarray.t:
[GH #927] Test sort() after unshift in ResizableFloatArray

It's probably a good idea to check if sorting after an unshift works as well as the currently available tests for simply sorting and for sorting after a push. This way we test the interaction between resizing and sorting.
rrot: 829fa8f | dukeleto++ | t/pmc/resizablefloatarray.t:
Merge pull request #947 from paultcochrane/ptc/resizable_float_array_sort_test

Added tests for sort() method in ResizableFloatArray PMC
kid51 This means that we essentially only had one customer for our potential product: Perl 6 14:47
Which leads to two questions:
1. How much does Perl 6 (Rakudo) want Parrot?
2. How much does anyone want Perl 6? 14:48
The answers to those questions will have a strong effect on how many volunteers each project attracts.
ptc true. and p5 is still so good and so actively developed
kid51 Yes, so much so that Perl 6 cannot be viewed as the *successor* to Perl 5, but only as the *sibling* to Perl 5. 14:49
But that sibling needs to develop a base of users whom it can serve better than P5 for certain applications. 14:50
Most open source projects fail.
They only survive 2-3 years. 14:51
By that criterion, neither Parrot nor Perl 6 is a failure.
But it's nonetheless true that after a certain passage of time, an OS project has to develop a user base to survive or thrive.
ptc so those were chromatic's arguments? Or has that sort of crystalised out of the discussion? 14:52
kid51 See modernperlbooks.com 14:53
This, of course, gets tangled up in Perl 5 vs. Perl 6 debates. 14:54
ptc oh, ok
kid51 So, Parrot's current challenge is to (re-)recruit volunteers who can provide a VM that meets some of Rakudo's needs ... even while Rakudo aims to develop on multiple VMs (JVM, .net) 14:58
Rakudo, in turn, faces its own challenges ...
... but I don't follow Rakudo closely enough to be able to comment intelligently on those.
But i want perl6/rakudo to succeed, if only because of my respect for TimToady and pmichaud 14:59
myhrlin I heard about this 15:03
it's the main reason I joined the channel just to see if it was true :-/
masak interesting discussion. 15:04
I can represent the Rakudo side of it, if anyone wants details from that group. :) 15:05
as a frequent *user* of Rakudo, I can see how Parrot has allowed Rakudo to progress/develop in some ways, and hindered development in other ways.
kid51 Yes .. 15:07
masak as a Perl 6 programmer, I believe Perl 6 has many advantages over Perl 5. they're not too apparent yet, and with the wrong sets of decisions over enough time, they never will be.
but plenty of things are making me hopeful for Perl 6's future.
as for Parrot, I'm less sure. but I'm also less involved.
Allison says that Parrot is a success in the sense that it instigated the JVM and the CLR to support dynamic languages. that's a sad kind of victory, though. 15:10
kid51 I didn't know that that was the case ... or that Allison had said as much.
If that's so, that explains why, e.g., Jonathan has spent so much effort on building Rakudo on those VMs 15:11
allison Not in the sense that it's the *only* success of parrot.
But, it was a huge impact parrot had. 15:12
Personally, I think Perl 6 has little hope of real success on JVM/.Net
it's the wrong market
Java and C# people scorn the very idea of Perl 15:13
masak we'll see.
allison and Perl people hate Java and C#
masak you could say the same of Python and Ruby people, I guess. and those have projects on those VMs.
allison it's a match made in some kind of 9th underworld
Python was quite welcoming to Parrot, and to alternate VMs in general 15:14
kid51 And Parrot has much more of an incentive to be responsive to the needs of Rakudo than do JVM/.Net.
allison Ruby pretty much ignores the alternate VMs
kid51 allison: It's good to have that cross-HLL perspective.
masak two years ago, I said "Parrot is a foundering project on top of a wonderful vision". it got perpetuated as a parrot-dev thread. two years later, I stand by that description.
allison I'd say Parrot is a foundering project dedecated to supporting a foundering vision 15:15
kid51 the foundering vision being "... VM aimed at all dynamic PLs" ? 15:16
allison neither is fatal, but it's certainly a warning sign worth looking at
yes, that, but also Perl 6
masak there are many beautifyl goals in there: an open-source VM for dynamic languages. an ecosystem of HLLs borrowing the best parts from each other. excellent threading and garbage collection. none of those have been achieved. (yet.)
kid51 allison: can you be more specific re Perl 6?
allison So, you know how when relationships fail, it's pretty much always the responsibility of both parties? 15:17
kid51 I'm told that's true ;-)
allison You can never say "It's all Bob's fault" or "It's all Martha's fault"
kid51 allison: I simply don't follow P6 enough to know what its status/vision is 15:18
allison I'm fully ready to say that Parrot has been part of the problem in the Perl 6/Parrot relationship trouble.
kid51 I don't have the tuits for that on top of everything else.
allison But, I really wish Rakudo folks would take responsibility for their own less-than-helpful actions.
And their own less-than-realistic ideas. 15:19
kid51 Well, I'm less focused on the Perl 6/Parrot relationship problems.
Those would be less critical if there were a larger user base.
allison kid51: leave aside the human tension, that mainly grows out of project tension 15:20
Just look at it purely as a project management problem.
kid51 P5P, I find, has some appalling problems, but the size of the user base means we don't get consumed by those problems
allison I was, for several years, the project manager of Perl 6.
And I look back at my mistakes and see the seeds of where we are today. 15:21
masak interesting.
allison The point when I took over from Nat was approximately when we realized Perl 6 wasn't going to be done in 6 months.
Or even a year. 15:22
But, we still all hoped it would be done in, say, 2 years.
masak there's a 2002 quote from TimToady saying, essentially "it gets done when it gets done".
kid51 That was in about 2003, IIRC?
masak: But that's not just a TimToady thing ... It's similar to Debian's "We release when it's ready" 15:23
allison kid51: I got involved in 2002, don't recall exactly when I took the PM hat
kid51: yes, but Debian's "release when it's ready" is every 2 years
like clockwork
they're talking delays in months
not decades
kid51 heard a talk by Debian project leader at Google in NYC this past week 15:24
allison I took a very "pacific" line as project manager.
"Don't worry about the setbacks, we'll get there"
"Keep at it"
The same with Dan and Leon
Leo
"It's okay, just try to work together" 15:25
over and over again, calming
kid51 And I, as a Parrot contributor/advocate, took exactly the same approach, FBOW
ptc those aren't bad premises
allison I sometimes wonder what would have happened if I took a whip to folks
ptc it's a huge idea, and was obvious that it would take ages to reach the final goal 15:26
allison If I told them "the only way this project will succeed is if you ship something *right now*"
kid51 Well, in 2010-11, I didn't take a whip to folks, but I did advocate structure.
allison It doesn't have to be perfect, just ship
kid51 That approach was ulttimately rejected.
ptc allison: agreed
allison ptc: that wasn't obvious, in fact the first idea was that it would only be a few quick fixes to Perl 5
ptc that's effectively agile devel... 15:27
allison ptc: that "would take ages" story was added later
kid51 Yes. A shippable product would have led to more users. More users would have led to more contributors. More contributors increase chance of solving remaining problems.
allison ptc: as a justification for the delays
ptc aha
allison ptc: but it was never the original idea
ptc i meant the idea of parrot in the beginning was a huge idea. 15:28
masak allison: from my perspective, when Parrot got started, we simply didn't have enough "domain knowledge" to build Perl 6. (we got much of that 2008-now. )since Parrot needs to support Perl 6, it also needs to factor in that domain knowledge. I wonder how much of the team friction came from that knowledge crunching changing the goalposts for both teams.
allison ptc: not really, even Parrot was just supposed to be Perl 5 + threading and unicode, at first
ptc i think you did a great job taking over from Dan and Leo
oh
allison masak: but that knowledge could have been developed on a *shipping product* 15:29
masak: and that would have totally changed the game
masak: look at Perl 5 today
masak: and everything that's changed since it first launched
masak: that's totally the "academic" persepective, that you have to know everything before shipping anything 15:30
masak it's possible I completely don't understand what this great "shipping product" of yours is. I assume it isn't as simple as making monthly releases, which both Parrot and Rakudo have been doing for ages.
allison (which is why academics don't ship products)
masak no-one more than us wants our software to find actual use. 15:31
allison masak: it requires shipping something that's stable and usable in production
masak: and doesn't keep ripping the carpet out from under its users
masak ...and in order to get there, you need users who use the product, submit bugs, and polish it.
I am a firm disbeliever in magical robustness.
allison masak: which Parrot tried to do in 2009 with the 1.0 release, but Rakudo couldn't make the transition 15:32
masak wow, things really do look different from there, don't they? :)
allison and, to be honest, neither could most of the Parrot developers :)
masak: how do you support production users then, if not by giving them stability? 15:33
masak I agree stability is a very important desirable. it's right up there with features and performance.
allison a small set of stable features is more useful than a large set of unstable features 15:34
masak all I'm saying is that any product will need to gradually reach stability. its velocity to reaching stability is the integral under the curve of alpha users.
allison And I'm saying it should be stable from the start, release early, release often.
Stability is not a finish line, it's a daily practice.
masak your model of the actual state of affairs is so different from mine that I don't even find a point where I disagree with you... :) 15:35
allison Where would Twitter be if they said "oh, we can worry about stability later, let's just add more features"
masak I don't believe the Rakudo team misspent their time adding features in the 2007-2010 period, though. 15:36
and a great deal of the lack of stability was to be found in the Parrot layer.
(and it got fixed over time, because we had those features)
allison No, and to be fair, Rakudo entered the party with a 5 year legacy of delays that weren't their fault at all :) 15:37
(that is "No, the Rakudo team didn't misspend 2007-2010")
masak *nod* 15:38
when I look back over the past decade, I see a clear progression of ideas that make each layer better with time.
allison but still not actually usable
Which, makes it a really neat academic project. 15:39
masak the layers being runtime, type system, OO model, MOP, grammar, VM.
allison: you're neither completely wrong nor completely right.
allison: be aware that you make my blood boil by saying that a language which I profitably use every day is "not actually usable". 15:40
allison: that feels deeply unfair. especially as we're working day-to-day at making it *more* usable, and a comment like that *slows us down* by altering the perception of others.
kid51 masak: How do you profit by it? How many other people do so?
allison sadly, you don't really count as a user, just a bleeding edge enthusiast
masak allison: granted, but it's still a sliding scale.
arnsholt What's a user, if not someone using the language?
masak kid51: I profit by it by having code execute which pays my bills. there aren't enough other people doing so. 15:41
arnsholt Admittedly, some users are more advanced than others, but in the end we're all users if we use the thing
allison arnsholt: someone who is using the language, not because they're a developer on the language
masak allison: saying "you don't really count as user" does nothing to cool me down. I don't like when you say that.
allison arnsholt: Using my own software doesn't count as "getting users"
moritz of course if you find reasons to discount all the users, you're left with a count of zero
allison masak: the point is not to make you angry 15:42
masak allison: please consider being kinder to a developing project. a *still* developing project, after 13 years.
allison masak: the point is to make you *calmly* realize that there is work to be done
masak allison: we're making nice progress, on many levels. please don't say "Perl 6 has zero users".
moritz we're well aware that there's work to be done
masak that's why we're doing it ;)
moritz diminishing the current state doesn't help at all
allison masak: If you stop at "it makes me so angry when people say Perl 6 has no users", you won't get to "hey, really, why don't we have more users and how do we fix that?" 15:43
masak I'd like to point out that even though both I and moritz have nicks starting with 'm', we're two distinct people. both of us are Perl 6 users, by the way.
moritz and fwiw we have users that aren't also developing the language
so, please stop the trolling that we don't have users. Once and for all.
masak allison: no, it makes me angrey when *you* say that Perl 6 has no users. you should know better.
allison: so should chromatic. 15:44
allison masak: look, I'm a business owner, I know about selling products
kid51 moritz: I don't know who those users are. Do they, e.g., post blogs about how they use Perl 6?
allison masak: and so is chromatic
moritz kid51, allison: fwiw we receive about two new submissions of Perl 6 modules written by people we have never heard of
who never submitted patches to rakudo, niecza or any other Perl 6 compiler project
kid51 is *not* a business owner, but even I know you got to have customers
allison masak: we would both like to see a whole lot more of the "let's sell this" perspective in Perl 6 15:45
moritz and then they come forth and publish a whole MODULE written in Perl 6
allison masak: and instead, we see a lot of "no really, it's all good"
moritz so, these are certainly users.
masak I wasn't saying it's all good.
allison masak: which doesn't inspire confidence
tadzik also, there is Perl 6 DarkPAN
there are modules on github which we've never heard of
masak I was saying *we have users*.
kid51 By contrast, during the lifetime of Perl 6/Parrot, we have had new technologies in the Perl 5 world -- Moose, Catalyst, Dancer, etc. -- which we know have many production users.
allison Great, so when do we see ActivePerl 6?
When do we see Mojolicious 6? 15:46
moritz we are well aware that our numbers are more in the dozens or hundreds than in the millions
allison dozens or hundreds doesn't count
that's ALGOL
kid51 moritz: What can you do to get those dozens or hundreds talking about their use of Perl 6?
moritz why doesn't it count, as long as it's groing?
allison that's "it was a neat idea"
is it? 15:47
do you have statistics?
moritz kid51: we can ask them
sri allison: as soon as perl6 can do everything perl5 (core) can do at about the same speed
sri is quite interested in a perl6 port of mojolicious actually
allison sri: but see, that's just the *beginning* of the curve. Python 3 has shown that 15:48
Python 3 actually did "just ship it"
and, they're still not getting adoption, years later
masak they are, just slowly.
sri many big python projects have still not been ported
allison yup, and Perl 6 can expect no greater speed
masak if I were a Python 2.6 user, I would also wait at this point, especially if not all of the dependencies to my project weren't ported yet. 15:49
sri or only recently, like django
allison which is exactly what people are doing, waiting
moritz python 3 has the problem that they don#t have a very good answer to the "what do you offer that Python 2 doesn't have?" question
allison and the dependencies have dependencies
moritz Perl 6 doesn't have any problem with that particular question
allison well, really, it does
moritz it's really about speed, stabilty and modules 15:50
allison it has some cool syntax improvements
masak it goes deeper than that.
allison but Perl 6 has none of those
sri allison: don't underestimate the appeal of killer features though, perl6 can be substantially better than perl5, especially when it comes to threads
python3 has no real killer features
allison sri: yes, perl6 has *potential*, but hasn't actually delivered any advantages... yet
arnsholt There are questions Perl 6 has problems answering currently, but "what can you offer that's new" sure isn't one of them 15:51
sri right
allison arnsholt: except that people don't just want "new" they want something that will improve their bottom line
arnsholt: so it needs to be a cold, hard "this will enable you to work faster, stronger" 15:52
arnsholt: a few bits of syntax don't offer anything there
arnsholt Don't confuse what you want and what everyone else wants. Sure, that's where we need to be to get a good ecosystem going, but first adopters have different motivations
And let's not kid ourselves here. That's where we are at the moment, like it or not
allison arnsholt: (and I say this with full awareness that some of those bits of syntax were created by a much younger me) 15:53
arnsholt: oh, no, I'd love it if the whole world were as fascinated with syntax as I am
arnsholt: I'm a linguist, like Larry :)
arnsholt: but syntax doesn't sell a language 15:54
arnsholt: getting things done is what sells a language
kid51 Will Perl 6 have presentations at YAPC::NA::2013 in Austin?
moritz kid51: yes
sri the type system alone will make it rather easy to sell to perl5 folks
masak allison: I know both you and chromatic have Perl 6's best interests at heart. but standing by the sidelines and saying "no, you're not doing it right. your priorities are wrong. if you did statistics, you'd see that you are irrelevant and not growing fast enough" is *not helping*. it's your choice what you choose to do, but I'd prefer if you didn't not help.
kid51 moritz: Can you be specific?
allison masak: well, I did cancel my astrophysics class this year to work on Parrot 15:55
masak kid51: I can be specific. but I'm not sure how much I can say yet.
allison masak: but then got sucked into 80 hour weeks at work
moritz kid51: I'm pretty sure that jnthn, pmichaud and diakopter at least will be talking about Perl 6
and i guess masak++ too
masak yes.
moritz but there's no official schedule yet
masak people are working on it.
allison masak: If I had any indication that you're actually listening to what I'm saying, I'd be much more comfortable settling down and leaving you to get on with it
masak: it's the denial that makes me angry 15:56
masak allison: yes, we're not getting across to each other.
kid51 moritz: That would be good. pmichaud's lightning talk last year was inspiring ... but it should have been transcribed into a blog post or something (not just a video)
masak allison: I'd just like to add that I respect you, and seem to get along with you well in person.
allison: maybe IRC isn't the medium for this.
Twitter sure isn't with chromatic ;)
allison masak: I don't understand what's so difficult about it.
masak allison: you say "it's the denial that makes me angry", I say "stop saying that Perl 6 doesn't have users". 15:57
allison masak: "Perl 6 is not a successful language"
masak: is that untrue?
masak no, it pretty much sums it up.
allison masak: it's not a personal attack
masak didn't say it was.
I said "stop saying that Perl 6 doesn't have users". 15:58
allison but it doesn'
doesn't
masak aarrrrgh
moritz but it does
allison truth is truth
moritz it doesn't have so many that you consider them as "counting"
but it still *has* users
masak I did that rather than kicking allison ;)
allison: Perl 6 *does* have users!
allison: please, *please* stop saying it doesn't!
allison masak: that's semantics
masak NO. 15:59
moritz semantics are important
allison I can say "Perl 6 has 100 users"
moritz allison: that's much better than saying "Perl 6 has no users"
masak the definition of "no users" is *pretty clear*.
allison and everyone in the world will say "that's no users at all"
kid51 masak: Then, let's hear about your users ... your users who are not also P6 developers and who are using P6 in production situations in businesses or governments or such.
masak kid51: that's not fair and you know it.
allison masak: it is fair
kid51 masak: It is fair. 16:00
allison masak: it's a completely logical definition of the problem
kid51 And even if it weren't fair, it *is* the question I get asked about Perl 6
masak Perl 6 will get different users at different stages.
allison we have to identify the *problem* before we can create a *solution*
masak it's not yet at the stage where it gets users in government.
kid51 masak: Tell us about its *current* users
allison masak: right and *that* is the problem
masak its current users develop modules and put them on modules.perl6.org
allison it has developers and bleeding edge enthusiasts 16:01
masak its current users think about how to use grammars and the MOP and macros to do cool things.
moritz and those are also users.
allison which is great progress
masak its current users write scripts at work to do parsing problems that would be possible but harder with Perl 5.
kid51 masak: Can you get those people to blog about that parsing ... or give talks at conferences? 16:02
masak its current users sometimes package the solutions that they had use for personally as a module and add it to the ecosystem.
allison can we agree that Perl 6 needs to have a goal of getting "real world, production users"
?
masak kid51: sure. but the ratio between people who do stuff and people who talk about stuff stays pretty constant.
allison I don't care what we call them
moritz we can agree that Perl 6 needs more "real world, production users"
allison but Perl 6 doesn't have them, and it needs them
Is there any company, anywhere, running their live web service on Perl 6? 16:03
That is actually profitable?
moritz not their live web service
but there's a startup doing business logic with Perl 6
in the background
allison okay, then lets stick that as my definition
moritz no idea if they are profitable
allison Perl 6 needs a "profitable live web service" as a user. 16:04
moritz ok. But please continue to call it that way when you talk about it
masak what moritz said.
allison Really, you'd get a lot more credibility if you'd just say "yup, we have no users, and we need to work on that"
sri wonders if perl6 i/o has improved in the last few months
masak allison: it may be just semantics, but it matters greatly to us in the trenches what you call it.
allison than quibbiling over whether your enthusiasts count as "users" 16:05
it looks petty
masak it matters because it becomes part of the perception.
if you publish stuff online saying "Perl 6 has no users", then it becomes more difficult to attract users.
allison stop worrying about perception
arnsholt allison: Your sophistry over what counts as a user and what doesn't looks equally petty, mind 16:06
allison the whole world things Perl 6 is a joke, you have to ignore them
thinks
kid51 masak: But to correct the perception, you need people who can step forward and say, e.g., "I am running a profitable live web service using Perl 6"
allison they're wrong, it's a great idea
but, you have to have a tough skin
don't get distracted
kid51 Similarly, those of us in Parrot need to be able to say, "Production users use a HLL which runs on Parrot"
allison you will never change their perspective by arguing 16:07
sri takes back his statement from earlier that he wants perl6 to do everything perl5 does... the node.js feature set is actually a much better benchmark nodejs.org/api/
allison the *only* way to change their perspective is to prove them wrong
with rock-solid production users
that wildly successful startup, built on the back of Perl 6
that is what will prove them all wrong
nothing else will
moritz ok. What are you going to do about it? 16:08
allison work on parrot, when I can
masak allison: none of what you say is really news to me.
allison: we're working on making Rakudo a better platform for production users.
and working on things like documentation and packaging.
I'd like to think that I spend a good deal of time on productizing. 16:09
allison <sigh> it feels like trying to push jello
it always has
it just springs back to "it's okay, by Christmas" 16:10
it's not okay
Perl 6 is hurting Perl 5
masak no, things won't magically fall into place unless people work towards realizing them.
allison if Perl 6 isn't going to ship, I wish it would just go away
shut the doors 16:11
masak "Perl 6 is hurting Perl 5" feels like an antiquated world view at this point.
allison let Perl 5 go on to Perl 7
moritz *sigh*
rakudo ships monthly
it doesn't ship they way you want it to ship
masak allison: clearly you're not in touch with the Perl community when you say that.
moritz but please don't imply it doesn't ship.
allison masak: Perl 5 is my bread and butter
masak: I don't know what universe you're in
sri speaking as an outsider... i get the impression rakudo development is too focused on polishing the core language, and doesn't pay enough attention to the stdlib
moritz it's such unqualified FUD which really hurts Perl 6
sri: that may well be the case 16:12
kid51 masak: While I don't agree with Allison on P5 vs P6, there are *many* people in the P5 world who say precisely that, i.e., let's skip over to perl 7
moritz sri: we really need to find somebody who takes IO into his hands
allison moritz: "shipping monthly" doesn't mean anything
masak kid51: assume for a moment you're right. so what? it's not going to happen.
moritz allison: likewise "doesn't ship" doesn't mean anything 16:13
without further qualification
kid51 But I do think we need to hear from people who can forthrightly say, "Perl 6 is my bread and butter."
16:13 not_gerd joined
allison moritz: I didn't say "doesn't ship" 16:13
moritz 17:10 < allison> if Perl 6 isn't going to ship, I wish it would just go away
kid51 Just as at YAPCs, we hear from people who say, "Catalyst/Dancer/Mojolicious is my bread and butter."
moritz that kinda implies to me that you think it doesn't ship 16:14
no?
moritz really wastes his time here that he should spend hacking instead
16:14 moritz left
allison moritz: let's not get tangled in semantics 16:14
moritz: you know what I mean
okay 16:15
that is not a helpful attitude
16:15 ruoso_ joined
allison If it weren't for the fact that Perl 6 has had far too many forks, I'd be highly tempted to fork it again. 16:17
not_gerd sri: there were some io fixes in parrot 5.1.0 and 5.2.0, which Rakudo currently doesn't take advantage of as they default to 4.10.0
sri: I've got a branch that fixes readline() for sockets
(which needs more polish)
kid51 builds master and hopes that test failures have been fixed 16:18
ptc kid51: they have; there was a revert earlier on today 16:19
not_gerd kid51: should be
if it's still broken, I'm not to blame ;)
kid51 ptc, not_gerd: Many thanks
not_gerd kid51: thanks go Util, who also did the release 16:21
kid51 The weird thing about the above discussion is: Parrot has fewer users than Perl 6, but we're not in denial about it.
16:23 tadzik left
kid51 Success: smolder.parrot.org/app/projects/rep...ails/39317 16:24
(linux/i386) 16:25
allison kid51: yes, there's a zen in that, accepting where you are 16:30
kid51 all g++ build: smolder.parrot.org/app/projects/rep...ails/39320 16:35
16:37 moritz joined
moritz when I do a .send on a socket, is there any buffering going on? 16:38
because it seems the IRC server doesn't receive the JOIN #channel lines 16:39
not_gerd moritz: looking at the code, it appears to be unbuffered by default (but of course I might have missed something) 16:51
socket initialization is a bit messy
moritz not_gerd: thanks
I'm now trying to get netcat to dump me a copy of the traffic
so that I can be sure of the problem
not_gerd you could also check the write_buffer attribute on your PIO handle 16:53
16:55 xcombelle joined
moritz ok, turns out it's not a problem with with parrot sockets at all 17:01
masak moritz: ooc, what was the problem? 17:14
moritz masak: see #perl6 17:15
masak oh! 17:18
lazy for loops strike again... :)
ptc is there a policy as to what should go into .gitignore and what shouldn't? 17:20
dalek rrot: 216da7f | paultcochrane++ | t/pmc/resizableintegerarray.t:
[GH #926] Add a test for sort method of ResizableIntegerArray PMC
rrot: f361d1a | paultcochrane++ | t/pmc/resizableintegerarray.t:
[GH #926] Extended sort() tests of ResizableIntegerArray PMC

The tests of the sort() method weren't testing that after resizing the array, that the sort() method still works (the array is resizable after all
  :-) ). This commit adds these extra tests.
ptc I grepped the sources but couldn't find any policy...
rrot: bcc8fac | dukeleto++ | t/pmc/resizableintegerarray.t:
Merge pull request #946 from paultcochrane/ptc/resizable_int_array_sort_test

  [GH #926] Add a test for sort method of ResizableIntegerArray PMC
ptc ... and does anyone have objections if I add *~, *gcov, *gcda, *gcno and *html to the main .gitnore? 17:22
17:23 dukeleto_tmp joined
dukeleto_tmp ptc: hola 17:24
not_gerd ptc: ooc, why *html?
dukeleto_tmp ptc: don't think we have a formal policy, feel free to propose one :)
ptc: do you have questions? I am about to head out somewhere, but I was reading IRC logs and saw you asking questions :)
not_gerd also, aren't there some dots missing?
ptc not_gerd: because 'make cover' generates a lot of them... Maybe I can just add better regexps for them 17:25
dukeleto_tmp: hi
dukeleto_tmp: I asked a couple of questions regarding tickets on github, but they can wait 17:26
dukeleto_tmp ptc: i either answered or merged :)
ptc dukeleto_tmp: aha. cool, thanks :-)
saw the merges, just haven't got around to checking the other stuff yet. 17:27
Am working on sort benchmarks mentioned in GH #175 atm
dukeleto_tmp ptc: i haven't been on irc much lately, so feel free to email me if you have a question. duke at leto dot net
ptc: awesome!
ptc dukeleto_tmp: cool, no worries
dukeleto_tmp goes back into the meatspace
not_gerd bye, #parrot 17:30
17:31 not_gerd left 18:22 xcombelle_ joined
dalek p/vmarray-list: 2de3300 | (Arne Skjærholt)++ | src/6model/reprs/VMArray.c:
Implement copy_to in VMArray.
 19:02
p/vmarray-list: ee613a6 | (Arne Skjærholt)++ | src/ (2 files):
Remove dependencies on get_number VTABLE for nqp::list.

When nqp::list was an RPA, this was fine, but due to a mismatch between sixmodel and how Parrot expects PMCs to behave, it's easier to use nqp::elems explicitly instead.
p/vmarray-list: 85bea1d | (Arne Skjærholt)++ | src/pmc/sixmodelobject.pmc:
Make sixmodelobject.pmc's VTABLE_does handle some Parrot types.

Specifically, objects with the VMArray REPR return true for "array", and those with VMHash true for "hash". This way, an nqp::list (and nqp::hash when that's actually implemented) can be flattened into a subroutine call using Parrot's
  :flat syntax.
p/vmarray-list: 6892934 | (Arne Skjærholt)++ | src/HLL/Compiler.pm:
Use list_s instead of list for string lists in HLL/Compiler.pm.
p/vmarray-list: 95084d2 | (Arne Skjærholt)++ | src/QAST/Compiler.nqp:
When converting Regex nodes in as_post, get PMC from cursor_start_all.

This means we use a few more registers, but the new nqp::list only supports fetching PMCs, so now it fetches PMCs from the list and then converts to int or string by setting the appropriate kind of register.
p/vmarray-list: 22ecde9 | (Arne Skjærholt)++ | src/ (2 files):
Implement smart numification similar to how it works on JVM.
p/vmarray-list: a82ef51 | (Arne Skjærholt)++ | src/ (2 files):
Implement some of the scaffolding needed for proper HLL config.

This also makes hlllist get the list type from the HLL config, rather than being hardcoded to BOOTArray.
p/vmarray-list: d52f45d | (Arne Skjærholt)++ | src/ (2 files):
Implement sethllconfig.
p/vmarray-list: d331b8c | (Arne Skjærholt)++ | src/core/NQPArray.pm:
Set NQPArray's boolspec and add it to the HLL config.
p/vmarray-list: 2c41d73 | (Arne Skjærholt)++ | src/ops/nqp.ops:
Tweak nfa_from_statelist to work with VMArray objects, rather than RPAs.
19:09 benabik joined 19:52 contingencyplan joined
Util ptc: For policy, see docs/project/committer_guide.pod section "IGNORED FILES". 20:21
*.gcda, *.gcno, *.gcov are already in .gitignore; I do not object to the dotless forms being added, but I am curious why it is needed.
*~ is not generated by the parrot build process; you could add it to your personal ~/git/info/exclude .
ptc Util: hi
Util: the gcda, gcno etc files turn up in the base dir. Acutally, when I mentioned the files in irc I didn't include the dots, but in the patches I added them 20:22
Util: thanks for the tip wrt git/info/exclude :-) 20:23
arnsholt For editor files, I recommend creating a file in ~ ignoring those and pointing git's core.excludesfile at it 20:24
See also the FAQ section of github.com/tpope/vim-pathogen 20:25
ptc Util: thanks also for the link to the pod. I did see that document, however, there wasn't any mention of how vim, emacs, etc. backup files were handled.
arnsholt: thanks :-) 20:26
20:50 kurahaupo joined
ptc Util: comment about the *.gcda etc already being in .gitignore (finally) made me think. I looked in .gitignore, but the files which are ignored are '/*.gcda' etc and not simply '*.gcda' etc. This is most likely the reason that these files are turning up in my 'git status' output and is actually the change that should be implemented. What do you think? 21:10
dalek rrot: 38f41b7 | paultcochrane++ | examples/benchmarks/sort.pir:
[GH #175] sort.pir benchmark reports how many elements were sorted

This gives us output which can be checked in the benchmarks.t test, such that the sort.pir test can be added to the benchmarks test suite		 21:28
rrot: d3c828a | paultcochrane++ | t/benchmark/benchmarks.t:
[GH #175] Add the sort benchmark to the benchmark test suite
rrot: 43c83b1 | paultcochrane++ | examples/benchmarks/sort_ffa.pir:
[GH #175] sort_ffa.pir benchmark prints number of elements sorted

It's good to see that a program gives some output to the user, so that we know that *something* happened. Now we have this output and can add the relevant search in the benchmarks.t test suite.
rrot: ae87849 | paultcochrane++ | t/benchmark/benchmarks.t:
[GH #175] Added sort_ffa.pir benchmark to main benchmark suite
rrot: 63affbc | paultcochrane++ | / (3 files):
[GH #175] Renamed sort.pir benchmark to sort_fia.pir

This makes it more obvious that this benchmark is for sorting FixedIntegerArrays in much the same way that sort_ffa.pir benchmarks sorting FixedFloatarrays
rrot: bd184ce | paultcochrane++ | examples/benchmarks/sort_fia.pir:
[GH #175] Correct typo in sort_fia.pir POD
rrot: 7922f60 | paultcochrane++ | / (2 files):
[GH #175] Added a sort benchmark for ResizableIntegerArrays
rrot: 8b8110c | paultcochrane++ | / (2 files):
[GH #175] Added a sort benchmark for ResizableFloatArrays
rrot: 6a7f23f | dukeleto++ | / (6 files):
Merge pull request #950 from paultcochrane/ptc/sort_benchmarks

  [GH #175] Slightly reorganise current and add new sort benchmarks
rrot: bff647b | paultcochrane++ | config/gen/makefiles/root.in:
Remove gcov-generated files in root parrot dir with 'make cover-clean'

The *.gcov, *.gcda, *.gcno files, which are generated by gcov, are also output in the base parrot directory. 'make cover-clean' didn't remove these files. This commit corrects this issue.
rrot: c027f67 | paultcochrane++ | config/gen/makefiles/root.in:
Removing cover-generated html files in 'make cover-clean'
rrot: ce65db1 | paultcochrane++ | config/gen/makefiles/root.in:
Added frontend/pbc_disassemble to list of coverage directories
rrot: 54c8b86 | dukeleto++ | config/gen/makefiles/root.in:
Merge pull request #949 from paultcochrane/ptc/cover_fixups

Clean up better after 'make cover'
rrot: 09c8e2f | paultcochrane++ | .gitignore:
Moved vim-relevant lines to its own section in main .gitignore
rrot: 59d249f | paultcochrane++ | .gitignore:
Ignoring vim backup (~) files
rrot: aa80534 | paultcochrane++ | .gitignore:
Ignoring gcov and gcov2perl automatically generated files

These files turn up when one runs 'make cover' and don't need to be in the output from 'git status'
rrot: 28c4c0c | paultcochrane++ | .gitignore:
Ignoring html files generated by 'cover' in 'make cover'
rrot: b685611 | dukeleto++ | .gitignore:
Merge pull request #948 from paultcochrane/ptc/gitignore_extras

Ignore more automatically generated files in .gitignore
21:40 travis-ci joined
travis-ci [travis-ci] parrot/parrot#829 (master - 6a7f23f : Jonathan "Duke" Leto): The build was broken. 21:40
[travis-ci] Change view : github.com/parrot/parrot/compare/b...7f23f4b2f5
[travis-ci] Build details : travis-ci.org/parrot/parrot/builds/5765828
21:40 travis-ci left 21:42 kid51 joined
kid51 Successful build on Darwin/PPC: smolder.parrot.org/app/projects/rep...ails/39328 21:53
commit 829fa8f119e3ccf3506a3f5d1af07e9d7164b48a
21:54 travis-ci joined
travis-ci [travis-ci] parrot/parrot#830 (master - 54c8b86 : Jonathan "Duke" Leto): The build was broken. 21:54
[travis-ci] Change view : github.com/parrot/parrot/compare/6...c8b867886d
[travis-ci] Build details : travis-ci.org/parrot/parrot/builds/5765836
21:54 travis-ci left
dalek rrot: ddb7b3f | jkeenan++ | MANIFEST (2 files):
Update MANIFEST.
 21:57
ptc doh! I forgot to add the new files to the MANIFEST.... sorry!
22:02 travis-ci joined
travis-ci [travis-ci] parrot/parrot#831 (master - b685611 : Jonathan "Duke" Leto): The build was broken. 22:02
[travis-ci] Change view : github.com/parrot/parrot/compare/5...856115e78f
[travis-ci] Build details : travis-ci.org/parrot/parrot/builds/5765851
22:02 travis-ci left
dalek rrot: 35d28ae | jkeenan++ | t/steps/init/hints/darwin-01.t:
Update test of error message to reflect new output under verbose => 1.
 22:10
kid51 We're okay now. 22:12
I just got a PASS on 'make fulltest'
It may take Travis a few cycles to catch up.
22:12 travis-ci joined
travis-ci [travis-ci] parrot/parrot#832 (master - ddb7b3f : jkeenan): The build was fixed. 22:12
[travis-ci] Change view : github.com/parrot/parrot/compare/b...b7b3f01c2b
[travis-ci] Build details : travis-ci.org/parrot/parrot/builds/5766312
22:12 travis-ci left 22:22 travis-ci joined
travis-ci [travis-ci] parrot/parrot#833 (master - 35d28ae : James E Keenan): The build was fixed. 22:22
[travis-ci] Change view : github.com/parrot/parrot/compare/d...d28ae1b45a
[travis-ci] Build details : travis-ci.org/parrot/parrot/builds/5766528
22:22 travis-ci left
ptc goodnight * 23:01
afk

Powered by Cro and the Raku® Programming Language.

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

Thank you!