Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2011-11-28

Perl 6 | Reference Documentation | Rakudo

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:01 tadzik [Coke]: what time tomorrow?
00:07 [Coke] tadzik: I don't know when the extension is good until.
00:07 [Coke] checking dukeleto's email...
00:07 tadzik istr "Monday"
00:08 stepnem joined #perl6
00:08 tadzik well, it's Monday 1AM
00:09 [Coke] Monday, November 28th at 8:00 UTC
00:09 tadzik okay
00:10 [Coke] colomon: looking at colomon/ABC on github; a "here's how you run it" would be very helpful.
00:11 diakopter sorear: all the ones supported by libicu I guess..?
00:12 diakopter hm, nqp doesn't seem to support "\c[UNICODE CHAR NAME]"
00:25 * diakopter looks into adding it
00:26 thou joined #perl6
00:28 replore joined #perl6
00:31 wolfman2000 joined #perl6
00:35 cooper joined #perl6
00:46 colomon [Coke]: apologies!  I've had "proper readme for ABC" on my todo list for about three months now...
00:46 alvis joined #perl6
00:48 rsimoes1 joined #perl6
00:50 imarcusthis joined #perl6
01:00 diakopter sorear: why
01:01 am0c joined #perl6
01:14 diakopter std: /<-:!InArrows>/
01:15 p6eval std be1f10e: OUTPUT«ok 00:01 120m␤»
01:19 sorear diakopter: I didn't think Parrot supported very many char props
01:19 diakopter oh
01:20 imarcusthis joined #perl6
01:20 diakopter oh hm, I was wrong about nqp not supporting \c[]
01:21 silug joined #perl6
01:28 dalek roast: f199800 | diakopter++ | S05-mass/properties- (4 files):
01:28 dalek roast: change properties rx to the new syntax
01:28 dalek roast: review: https://github.com/perl6/roast/commit/f199800df6
01:31 diakopter who knows what I just broke
01:34 hypolin joined #perl6
01:51 wolfman2000 joined #perl6
02:37 donri left #perl6
02:37 orafu joined #perl6
03:14 * tadzik adds some new GCI tasks
03:17 djanatyn tadzik: :D
03:17 djanatyn Where's the GCI task listing page, again?
03:17 tadzik djanatyn: some are already on :)
03:17 djanatyn :O
03:18 djanatyn what should I search for, "perl 6"?
03:18 tadzik I... am not sure :)
03:18 tadzik yeah, the ones I added are all tagged "perl6"
03:19 djanatyn hmm, I guess they all got snatched up
03:19 djanatyn GCI doesn't have a very nice search feature, IMO
03:19 tadzik to say it gently
03:20 djanatyn I'll check back again tommorow morning, in ...let's see
03:20 djanatyn 7 hours
03:20 djanatyn Do you think there will be a bunch of perl6 tasks avaiable by then?
03:20 tadzik moritz: I'm adding you as a mentor to all tasks, I hope you don't mind :)
03:21 tadzik djanatyn: they are available now :)
03:21 satyavvd joined #perl6
03:21 tadzik like http://www.google-melange.com/gci​/task/edit/google/gci2011/7185260
03:21 tadzik this one could be tough
03:25 djanatyn hmm, it says that I can't edit it
03:25 tadzik oh
03:26 tadzik well, I'm not suprised :)
03:26 djanatyn I tried replacing s/task/view in the url, and it said that the page was inactive at this time
03:26 tadzik how about http://www.google-melange.com/gci​/task/view/google/gci2011/7195397 ?>
03:27 tadzik http://www.google-melange.com/gci​/task/view/google/gci2011/7185260 should be ok too
03:27 tadzik ok, time for me to get some sleep
03:29 djanatyn Both of those say "This page is inactive at this time"
03:29 djanatyn but, don't worry about it, I'll figure it out tommorow
03:29 djanatyn g'night
03:29 diakopter o/
03:29 colomon \o
03:37 envi_ joined #perl6
03:42 lestrrat joined #perl6
03:45 Chillance joined #perl6
04:53 Chillance joined #perl6
05:20 woosley joined #perl6
05:47 agentzh joined #perl6
05:53 packetknife joined #perl6
06:02 localhost joined #perl6
06:02 kaleem joined #perl6
06:43 moritz djanatyn: it looks like the tasks need to be approved by an admin first
06:46 diakopter moritz: could you review the last commit to roast?
06:48 moritz diakopter: I believe that some of them are wrong
06:48 hundskatt joined #perl6
06:48 moritz -ok("\x[8340]"  ~~ m/^<!isInArabicPresentationFormsA>.$/
06:48 moritz +ok("\x[8340]"  ~~ m/^<:!InArabicPresentationFormsA>.$/,
06:48 moritz diakopter: <!...> is zero-width, but I don't think <:!...> is
06:48 moritz or is it?
06:49 sorear <:!foo> means <-:foo>
06:49 diakopter ! there wasn't zero-width
06:49 moritz then why is tehre a dot after it?
06:49 sorear <!foo> is zero width, <-foo> is not
06:49 diakopter oh, hm
06:50 diakopter I guess all the :! should be !: then?
06:51 moritz no
06:51 moritz just remove the dots
06:52 diakopter how would that keep the meaning (sorry it's taking me so long to understand)
06:52 moritz well
06:53 moritz previously it was <zero-width>.
06:53 moritz then it's <one-character>
06:54 diakopter I still don't understand; I'm sorry
06:54 moritz the old tests did not test negative character classes
06:54 moritz but zero-width assertions, followed by a dot
06:54 moritz that combination emulated a negative character class
06:54 Woodi hi, functional programming arive some time ago, maybe time for something new ? games programing ! :) http://chessprogramming.wik​ispaces.com/Parallel+Search
06:55 moritz diakopter: are you with me so far?
06:55 diakopter moritz: no; the zero-width assertion followed by a dot doesn't seem to me it would emulate a negative character class
06:56 moritz diakopter: but it does
06:56 moritz it says "match if the next character is not isInArabicPresentationFormsA. Match one character"
06:57 diakopter oh; that's what ! means.  I see now
06:57 moritz that's the same as says "match the next character if it is not isInArabicPresentationFormsA"
06:57 diakopter I forgot ! meant negated.
06:57 moritz aye, <!foo> is short for <!before <foo>>
06:58 koban joined #perl6
07:01 dalek roast: 7ea256e | diakopter++ | S05-mass/properties- (4 files):
07:01 dalek roast: braino
07:01 dalek roast: review: https://github.com/perl6/roast/commit/7ea256e4e6
07:04 diakopter I haven't found any tests for things like <+ :HexDigit - :Upper >
07:05 diakopter and for &|^ combos and for parens grouping
07:05 diakopter seems kinda hard to implement anyway
07:09 diakopter nqp: say(1^^0)
07:09 p6eval nqp: OUTPUT«Confused at line 1, near "say(1^^0)"␤current instr.: 'nqp;HLL;Grammar;panic' pc 23429 (src/stage2/gen/NQPHLL.pir:6732) (src/stage2/gen/NQPHLL.pm:329)␤»
07:17 moritz diakopter: those features are relatively new
07:17 moritz and thus not tested yet
07:18 moritz maybe they are tested  as  <+alpha-hexdigit> or so somewhere
07:20 sorear &|^ combos are _extremely_ new
07:20 diakopter oh :)
07:21 sorear nobody implements them yet; I know how but lack the tuits atm
07:21 sorear they will very likely be in niecza v13
07:23 diakopter std: /<:a -:b>/
07:23 p6eval std be1f10e: OUTPUT«ok 00:01 120m␤»
07:26 diakopter nom: say "a" ~~ /<[ab] | [b]>/
07:26 p6eval nom 2154eb: OUTPUT«===SORRY!===␤regex assertion not terminated by angle bracket at line 1, near "| [b]>/"␤»
07:27 diakopter nom: say "a" ~~ /<[ab] + [b]>/
07:27 p6eval nom 2154eb: OUTPUT«=> <a>␤␤»
07:57 satyavvd joined #perl6
08:21 skangas joined #perl6
08:30 jnthn mornin'
08:31 koban left #perl6
08:42 moritz \o
08:47 mj41 joined #perl6
08:57 sorear o/
09:05 snarkyboojum o_
09:06 moritz )o.
09:12 alvis joined #perl6
09:16 masak o-
09:28 woosley left #perl6
09:38 dakkar joined #perl6
09:39 skangas joined #perl6
09:47 moritz http://www.perlmonks.org/?node_id=940242 might interest you... or not. I kinda think I'm wasting my time
09:49 masak moritz: typo: ss/most of use/most of us/
09:50 moritz masak: thanks, fixed
09:51 masak oh, it's a discussion with chromatic.
09:51 * masak stops reading out of self-preservation
09:51 moritz it didn't start as one.
09:51 envi_ joined #perl6
09:52 masak I think I know the terrain of opinion that chromatic maps out. I've read it many times, I've understood it, and reading it again doesn't give me anything.
09:52 masak it can lower morale unnecessarily, though.
09:53 masak I believe we are addressing his concerns as fast as we can with our resources, and have been all the time.
09:54 masak ...and I do sympathise. my code has often been the canary for breakages.
09:55 masak that is, after all, what it's like being an early adopter. things break when whole subsystems (in Rakudo or Parrot) get replaced.
09:57 gfldex i noticed :->
09:58 masak what I feel is unfair and silly in chromatic's comments is the "rewrites cause things to break, and take too long! I told you so!" component.
09:58 masak as if we're doing the rewrites just for the heck of it to see things get worse.
09:58 gfldex at the other head, it's christmas every time something starts to works again
09:58 moritz well, he does emphasize that he doesn't tell developers what to do...
09:58 moritz probably because he doesn't have a plan of action either that both works and avoids breakage
09:59 jnthn "There's also still no new Rakudo Star release." - OK, so let's get this straight. He complains about lack of stability/continuity/stuff keeping working, then complains about the fact that we didn't put out a new star release? The point of the Star ones is to provide something more stable. If that's his interest, why is he even looking at the compiler releases?
09:59 masak he does reserve the right to produce bad press, though, or the equivalent of it.
09:59 moritz ah well, a waste of time.
09:59 jnthn That we haven't rushed out a star relesae with tons of regressions is a *strength*.
09:59 masak :(
10:00 masak I'm starting to see why it's important to identify and isolate panics in individuals.
10:03 Trashlord joined #perl6
10:04 * masak wishes pmichaud were here
10:04 snarkyboojum there's always the good ol' line "the show must go on" :D
10:05 masak snarkyboojum: absolutely. let's continue delivering Perl 6 to the world. :)
10:08 snarkyboojum masak: I have nothing but respect and awe for perl6 hackers, and yes I agree :D
10:10 masak and while we're at it, let's think a bit more about chromatic's definition of "product". the Star releases are pointing in that direction, I believe, but surely we can do more, with our limited but nonzero resources.
10:15 mkramer joined #perl6
10:24 moritz nom: say $*VM
10:24 p6eval nom 2154eb: OUTPUT«("name" => "parrot", "config" => {"git_describe" => "RELEASE_3_10_0-30-ga8cb7da", "sha1" => "a8cb7da75c1c7e1db08a6d50330044b831347fe1", "a" => ".a", "ar" => "ar", "ar_extra" => "", "ar_out" => "", "archname" => "x86_64-linux-gnu-thread-multi", "arflags" => "cr", "as" …
10:24 moritz nom: say %*VM
10:24 p6eval nom 2154eb: OUTPUT«Failure.new()␤»
10:26 moritz std: my @a; $@a
10:26 p6eval std be1f10e: OUTPUT«[31m===[0mSORRY![31m===[0mâ�¤Invalid hard reference syntax at /tmp/quWOLnzCf7 line 1:â�¤------> [32mmy @a; $[33mâ��[31m@a[0mâ�¤    expecting any of:â�¤       desigilnameâ�¤    twigilâ�¤Parse failedâ�¤FAILED 00:01 120mâ�¤Â»
10:26 moritz why is that forbidden?
10:26 moritz why not make it the equivalent of $(@a) ?
10:27 masak sounds good to me, but I wouldn't be surprised if there's a reason not to that I don't see right now.
10:27 masak what does $(@a) mean, exactly? 'item @a'?
10:27 moritz yes
10:28 moritz (modulo precdence)
10:28 masak is there a difference between 'item @a' and '@a.item'? there is for .list
10:31 moritz if you override .item, yes
10:31 moritz but what's the difference for .list?
10:32 masak hold on, I'll get the relevant backlog.
10:35 masak http://irclog.perlgeek.de/​perl6/2011-11-18#i_4726855
10:41 moritz nom: class X::Foo { ... }; class X::Foo { }
10:41 p6eval nom 2154eb: OUTPUT«===SORRY!===␤The following packages were stubbed but not defined:␤    X::Foo␤␤»
10:42 moritz jnthn: ^^ that one is blocking me on the exceptions stuff
10:43 * moritz RTs it
10:43 masak moritz++
10:43 masak was just gonna ask :)
10:45 tadzik good morning
10:45 daxim_ joined #perl6
10:45 mls morning!
10:45 phenny mls: 26 Nov 20:01Z <jnthn> tell mls it seems that any infinite recursion in Rakudo generates a hang when trying to find an exception handler
10:45 jnthn nom: class Y::Foo { ... }; class Y::Foo { }
10:45 p6eval nom 2154eb:  ( no output )
10:45 jnthn er. :)
10:45 moritz there's a 'my package X { }' in the setting
10:46 moritz but if I put Foo into that package, the error stays the same
10:46 jnthn nom: package Y { }; class Y::Foo { ... }; class Y::Foo { }
10:46 p6eval nom 2154eb:  ( no output )
10:46 jnthn nom: my package Y { }; class Y::Foo { ... }; class Y::Foo { }
10:46 p6eval nom 2154eb:  ( no output )
10:47 sorear niecza: class X::Foo { ... }; class X::Foo { } # curious
10:47 p6eval niecza v11-27-gdaa9117:  ( no output )
10:47 * sorear out
10:47 jnthn moritz: Oddness. Guess you didn't golf it beyond having the setting involved?
10:47 moritz nom: class Y::Foo { ... }; my package Y { }; class Y::Foo { }
10:47 p6eval nom 2154eb: OUTPUT«===SORRY!===␤The following packages were stubbed but not defined:␤    Y::Foo␤␤»
10:48 moritz jnthn: I just did :-)
10:48 jnthn oh
10:48 jnthn ...do we actually do that?
10:48 jnthn (put the Y::Foo before the my package Y)?
10:48 moritz yes
10:48 moritz erm
10:48 masak that happens, yes.
10:48 moritz I don't know how else to do it
10:48 masak it's a legitimate use case.
10:49 moritz both X and X::Foo are defined in Exceptions.pm
10:49 jnthn nom: my class Y::Foo { ... }; my package Y { }; my class Y::Foo { }
10:49 p6eval nom 2154eb: OUTPUT«===SORRY!===␤The following packages were stubbed but not defined:␤    Y::Foo␤␤»
10:49 moritz and I need to throw typed exceptions earlier than that
10:49 jnthn moritz: Is there a reason for the "my package Y { }"?
10:49 jnthn Are you actually putting something in it?
10:50 jnthn It's a bug though for sure.
10:50 moritz jnthn: well, it's there because nested packages didn't work when I wrote the code
10:50 moritz erm
10:50 moritz x::foo package didn't work
10:50 jnthn moritz: Yeah, I thought I fixed that bit up :)
10:51 moritz jnthn: I'll try to change that
10:51 jnthn The whole notion of "my x::foo { }" is kinda weird though. The my only applies to the x.
10:53 jnthn What bothers me more is that if the setting does my X::Foo { }, and somebody in their programs creates a class X::Bar { }, then it's going and adding an extra entry into X.WHO
10:54 moritz erm
10:54 moritz that's... ugly
10:54 jnthn Indeed
10:54 moritz can't we just store X::Foo in the lexpad, and lookup X::Foo in the lexpad first
10:54 moritz and only look into X:: if we don't find it in the lexpad?
10:54 jnthn You mean, store the thing with the colons in the lexpad?
10:55 moritz ye
10:55 moritz s
10:55 jnthn eek :)
10:55 jnthn That will mean that
10:55 jnthn X::Foo
10:55 jnthn X.WHO.<Foo>
10:55 jnthn Would give different results
10:55 moritz aye
10:55 jnthn No, I don't think that'll fly.
10:56 jnthn What may work is that we create a new package in the lexpad and copy the .WHO
10:56 jnthn Just like a multi copies its outer's candidate list and then adds its own stuff
10:56 moritz hmmm
10:56 * jnthn isn't really happy with a lot of package-y stuff at the moment, tbh.
10:57 jnthn It's ended up pretty complex to try and get things to fall out correctly. I feel I'm missing some clean realization of package nested that does the things we want.
10:58 * moritz kinda thinks that sorear will agree there
11:00 masak :)
11:02 jnthn Maybe we want too much. :)
11:05 moritz Perl 6 -- Maybe we want too much.
11:06 masak ^_^
11:07 jnthn Well, I think part of it is that many, many things have become far more lexically oriented. So packages - in terms of their namespacing role - have become far less prominent.
11:07 jnthn sub dispatch doesn't even looking in the package any more, for example.
11:08 jnthn Exportation is package-ish, but actually relies on a lexically scoped EXPORT package in the UNIT of the module doing the exporting. And so forth.
11:09 jnthn I think the only thing that's packaged scoped by default is...packages. :)
11:09 moritz and our-variables :-)
11:10 moritz and enums, iirc (but they are pseudo packages, no?)
11:11 jnthn nom: enum Foo <a b c>; $Foo::x = 42; say $Foo::x
11:11 p6eval nom 2154eb: OUTPUT«42␤»
11:11 jnthn Seems so :)
11:12 jnthn Probably EnumHOW does Stashing or some such :)
11:12 moritz :-)
11:12 djanatyn Hmm, still can't find any perl6 tasks. :\
11:13 * tadzik asks on #gci
11:14 moritz std: my class Y::A::B { ... }; class Y::Z { }; class Y::A::B { }
11:14 p6eval std be1f10e: OUTPUT«ok 00:01 119m␤»
11:15 moritz nom: my class Y::A::B { ... }; class Y::Z { }; class Y::A::B { }
11:15 p6eval nom 2154eb:  ( no output )
11:16 djanatyn tadzik: \o/
11:16 djanatyn thank you.
11:16 moritz djanatyn: http://www.google-melange.com/gci​/task/view/google/gci2011/7177376
11:18 djanatyn Oh, sweet!
11:18 djanatyn Okay, I can see that task.
11:19 moritz http://www.google-melange.com/gci​/task/view/google/gci2011/7184225
11:19 moritz djanatyn: just search for 'Moritz Lenz' as a mentor
11:20 moritz though it seems only 3 are approved
11:21 djanatyn Okay, I claimed the "Style for perl6.org" task :)
11:21 moritz \o/
11:21 djanatyn I can do a bit of webdesign, and I really like the perl6.org site
11:21 djanatyn I've wanted to steal something from it a long time :)
11:22 djanatyn although I'll probably try to make it look like the perl6 module directory
11:22 moritz if you want, you can link to stylesheets on those pages directly
11:22 djanatyn Ah, that sounds like a good idea
11:22 moritz and add custom box stuff in the custom css
11:23 djanatyn okay, well, thank you for helping me :)
11:23 djanatyn I've gotta get ready for school, see ya.
11:23 moritz have the appropriate amount of fun :-)
11:34 fridim_ joined #perl6
11:43 donri joined #perl6
12:13 Psyche^ joined #perl6
12:25 replore_ joined #perl6
12:43 mkramer joined #perl6
12:45 Trashlord joined #perl6
13:09 bacek joined #perl6
13:10 JimmyZ joined #perl6
13:17 molaf joined #perl6
13:26 wk joined #perl6
13:30 mtk joined #perl6
13:32 Woodi hmm, I logged in but in other tab. but not want to be anonymous, http://www.perlmonks.org/?node_id=940401
13:33 wk joined #perl6
13:34 masak hm. for a post about how yelling does not help, I'm not sure that post helped either. :/
13:36 daxim I'm barely involved.  Woodi, I read your post/it mostly comes across to me as "I need to vent my frustrations"
13:36 masak yeah.
13:37 [Coke] I am bemused that chromatic is venting on perlmonks, which is clearly too slow to be usable for any practical purpose. ;)
13:37 masak I'm frustrated at chromatic too, but... not publicly, in posts.
13:38 masak [Coke]: chromatic visits #perl6 quite seldom. he's done it three times in 2011, every time after he wrote a blog post about Perl 6 and the #perl6 regulars misconstrued it in some way or other.
13:39 colomon I dunno, I think I'm kind of feeling the pain chromatic is describing.
13:39 [Coke] I felt all that pain just using parrot.
13:39 colomon I mean, nom is pretty awesome, but we're still struggling to get back functionality we had a year and a half ago in b
13:40 JimmyZ most by a wonderful vision.
13:40 [Coke] I cannot tell you how many times I had to re-write partcl, for example. (not including the most recent 'try to rewrite it in nqp' which is now already obsolete since nqp is no longer the newest/best thing)
13:41 [Coke] so, to be using something that is using parrot, and going through its own rewrite cycles, is no doubt frustrating. I am unsure who his target audience is, though.
13:41 colomon [Coke]: yeah, I surely don't think all the blame lies with the Rakudo developers!
13:41 [Coke] not the early adopters. we already know about the issues; Is it the people who are waiting for perl6 ? because they were already waiting, either in anticipation or uncaringly.
13:42 daxim James Russell is the target audience
13:42 colomon but I do kind of think that maybe in retrospect, nom was made the master branch about six months too early.
13:42 [Coke] the developers are just trying to make a decent product with limited tuits.
13:42 masak colomon: I'm not saying chromatic is *wrong*. it does feel like he's overstating facts he should know the causes of, though. but he's probably colored by recent disillusionment.
13:42 [Coke] colomon: if no one was working on master... what problem would keeping master master have solved?
13:43 [Coke] anyway, sunk costs, they is us: are we heading in the right direction from where we are now?
13:44 masak the ng refactor and the nom refactor were both quite big, and followed each other fairly quickly. but they've both led to more features and more stability. I'd argue they were inevitable.
13:44 colomon [Coke]: it certainly feels like we're moving in the right direction, just frustratingly slowly.
13:45 masak because the alternative would've been the sitution in 2008: tight Parrot coupling, partial features, no-where to develop Rakudo, instability.
13:45 colomon masak: on chromatic, if nothing else dismissing niecza out of hand is pretty crazy.
13:45 masak it's easy to forget that those refactors were made for sound architectural reasons.
13:46 * JimmyZ is -1 to Hague Grant Application: PL/Perl6 Infrastructure and Improvements
13:46 masak colomon: I cannot judge that. I don't live in the US, where patent fights seem a real threat.
13:46 dalek mu: d445465 | duff++ | misc/perl6advent-2011/topic-brainstorming:
13:46 dalek mu: Add R* topic
13:46 dalek mu: review: https://github.com/perl6/mu/commit/d4454657f8
13:46 jnthn I find it mostly depressing that people think we go through painful rewrites for the fun of it.
13:46 jnthn It's NOT FUN.
13:46 masak the nom one is meant to be the last big one.
13:47 jnthn Certainly for my tenure in Rakudo development, nom is the last big one.
13:47 daxim I don't believe it
13:47 [Coke] masak: you so funny.
13:47 masak [Coke]: no, really. we don't have more switchings out of grammar engine and object model on the horizon.
13:47 masak I'm not saying there won't be more refactors. of course there will be.
13:47 jnthn Sure there will be.
13:48 masak but they will be much smaller in scope and duration.
13:48 * [Coke] apologizes for doing pretty much nothing perl 6 related over his most recent break.
13:48 jnthn The thing with ng, in hindsight, was that it dealt with some architectural issues, but left others unresolved.
13:48 [Coke] does feel a lot less stressed, though. :)
13:49 Woodi hmm, then, again :) , maybe set points to achive in implementing spec parts - to avoid frustration. points defining is trick in busting morale i hear :)
13:49 Woodi so mark few features somehow...
13:49 Woodi becouse horisont is far far away....
13:50 JimmyZ rakudo is going very well, imo
13:51 JimmyZ but parrot ....
13:51 Woodi how many developers in parrot atm ?
13:52 masak ooh, misc/perl6advent-2011/topic-brainstorming! I had missed that until now.
13:53 PerlJam I sometimes think that chromatic purposely makes bold? loud? controversial? claims on various forums just to goad certain people into making things better or proving him wrong.
13:53 daxim I call it the reverse troll
13:53 Woodi sometimes it work
13:53 daxim yes
13:54 PerlJam but it's one of those things that can only be viewed in hindsight.  It's a the superposition of a rant and a "I meant to do that"
13:54 daxim I love a good rant.
13:55 dalek mu: b43b813 | colomon++ | misc/perl6advent-2011/schedule:
13:55 dalek mu: Claim a couple of days.
13:55 dalek mu: review: https://github.com/perl6/mu/commit/b43b813420
13:56 colomon darn, I love editing right on github.  :)
13:57 takadonet joined #perl6
13:58 PerlJam The most useful an disturbing item of note in that perlmonks thread is that most people seem to talk about perl 6 in the future tense.  :-(
13:58 PerlJam s/an/and/
13:59 PerlJam moritz++ for fighting the good fight
14:00 masak PerlJam: outsiders always seem to talk about Perl 6 in the future tense. "Perl 6 will do <something desired>", even in the cases when it already has the feature or does that thing.
14:00 masak PerlJam: what they mean is maybe "When I start using Perl 6, it'll do <something desired> for me."
14:01 PerlJam "Perl 6 Now!"  sounds like a good article for someone to write.  :)
14:01 masak sounds like a good button to wear, too.
14:01 PerlJam oh, there's scrottie's book called "Perl 6 Now" but that's about Perl 5 stuff
14:02 colomon except "Perl 6 Now!" sounds a little bit like you're demanding it rather than proclaiming it.
14:02 masak colomon: who's saying we're not? :)
14:02 PerlJam :-)
14:02 donri Perl 6 In Our Lifetime!
14:03 * jnthn imagines a bunch of protesters with "Perl 6 Now!" placards
14:03 jnthn .oO( Occupy Perl 6! )
14:03 PerlJam colomon: How about "Perl 6 in the Present Tense" ?  doesn't quite have a ring to it ;)
14:03 colomon PerlJam: I was thinking "Perl 6 is now" but that's not quite right either.
14:04 masak "Using Perl 6, and It's Great"
14:04 PerlJam jnthn: occupy <places where programming languages are popular>
14:04 donri occupy freenode
14:04 JimmyZ Is there any dynamic language's nci better than perl6's, which jnthn++ writes ?
14:05 daxim you're not focusing.  in terms of jwz, how can perl 6 get me laid?
14:09 PerlJam In any case, chromatic is entirely correct about one thing: people *are* waiting for "Perl 6 the Product"
14:09 colomon and he's correct it's not here yet, either, IMO.
14:09 PerlJam yep
14:10 colomon On the other hand, trying to use it now is one of the single most useful things you can do to help get it to where it is going.
14:10 jnthn Yes, but abrasive posts that irritate and suck up the time of people working on it will most certainly NOT help get "Perl 6 the Product" to be here.
14:11 PerlJam really ... it's all Perl 5's fault.  If it hadn't set the bar so high, we'd already be there ;)
14:11 masak no-one complained about chromatic's prolixity when he was on our side :P
14:12 PerlJam masak: do you mean to say that he's no longer "on our side"?
14:13 masak PerlJam: for all I know, he's moved to a monastery with excellent wireless, and picked up COBOL.
14:13 daxim you're all individuals!
14:13 colomon daxim: I'm not
14:13 masak :P
14:18 colomon is sorear back from vacation yet?
14:18 jnthn colomon: Believe so...saw him this morning
14:18 colomon \o/
14:18 colomon I need to talk to him about optimizing mandelbrot in niecza.  :)
14:21 masak best way to optimize mandelbrot in general, at least according to the Fractint school of thought, is to do it all with integer math.
14:22 colomon masak: I'd be surprised if that were true in a p6 context
14:23 colomon actually, I'd be a little bit surprised if it were even true on modern x86 platforms
14:23 colomon it certainly was awesomely true in the "good ole days"
14:24 masak aye.
14:24 colomon basically, what seemed screamingly fast in niecza for writing mandelbrot to a file seems painfully slow when you're waiting for a window to draw.
14:24 PerlJam colomon: For Dec 1, a godo retro-post would include how to get pugs ;)
14:24 PerlJam s/godo/good/
14:25 PerlJam (how does that happen?)
14:25 colomon PerlJam: brains -- they are untrustworthy things
14:25 masak finger un-sync?
14:26 PerlJam anyway ... time for the day-of-endless-meetings for me
14:26 * PerlJam &
14:26 JimmyZ how about NativeCall to be core setting? hehe
14:27 jnthn JimmyZ: It should need a "use" directive
14:27 jnthn JimmyZ: We'll include it in R*
14:27 jnthn Doesn't need/want to be in the compiler itself, imo.
14:27 JimmyZ core lib is fine too...
14:30 JimmyZ or https://github.com/quietfanatic/link-c , which can parse *.h files
14:35 colomon joined #perl6
14:44 uniejo joined #perl6
14:50 LittlePanic72 joined #perl6
14:56 PacoLinux joined #perl6
14:57 mishin joined #perl6
15:06 Woodi code like: ... method ping() { $!sock.send( "$4\r\nPING\r\n" ) or return False; my Str $resp = $!sock.get() or return False; return True if $resp eq '+PONG'; return False }; ... }; for ^10000 { $r.ping() }
15:06 Woodi takes 1.2s on nom atm. on Perl5 0.4s. who to blame ? parrot or rakudo ?
15:07 masak that question is age-old, and unanswerable.
15:07 masak it's like asking "nature or nurture?"
15:07 Woodi you like unanswerable questions :)
15:07 masak the architectures of Parrot and Rakudo mesh into each other.
15:09 jnthn Woodi: Am I reading that right? Just a factor of 3 different from Perl 5?
15:10 Woodi just tested. will check
15:10 tadzik oddness :)
15:10 jnthn We...normally have a bigger gap ;)
15:10 [Coke] jnthn: a factor of 3 and he's complaining! ;)
15:11 Woodi 1.26s, will extract to separate script and test
15:12 masak how much of that is startup time?
15:13 Woodi $s = now; after startup
15:13 Woodi count just code time
15:14 Woodi must test, maybe i mess something
15:14 masak rakudo: my $s1 = now; my $s2 = now; say $s2 - $s1
15:14 p6eval rakudo 2154eb: OUTPUT«0.0161627759174326␤»
15:15 masak well, a "now" call takes ~0.016 seconds :P
15:17 jnthn That seems excessive...
15:17 Woodi sending 10k pings one by one takes now 1.15s but all script execution 6.14s real and 5.78s user
15:19 Woodi problem is that redis optimization - pipelining - should work better but work few time longer... probably my code is to nested :)
15:20 jnthn Woodi: Well, best way to find out where it spends time is profiling :)
15:21 mls jnthn: please pull https://github.com/rakudo/rakudo/pull/43
15:21 Woodi Pure Perl5: 0.35s; XS: 0.2s and pipelined: 0.018082857131958s  total
15:22 Woodi jnthn: it is what you asked before and it is I want to ask how to do this :)
15:22 Woodi C lvl or Parrot or Rakudo ?
15:23 jnthn Woodi: C level and Rakudo/NQP level are both interesting in general.
15:23 jnthn mls: done
15:23 Woodi any manual ?
15:23 jnthn mls: Does this fix the infinite recursion thing?
15:23 Woodi blogpost ?
15:23 dalek rakudo/nom: bd2f8df | mls++ | / (2 files):
15:23 dalek rakudo/nom: invoke_catchhandler: tweak call chain after the invoke call, so that the recursion_limit is set correctly
15:23 dalek rakudo/nom:
15:23 dalek rakudo/nom: Also added myself to CREDITS ;)
15:23 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/bd2f8dfa4c
15:23 dalek rakudo/nom: f4ee614 | jonathan++ | / (2 files):
15:23 dalek rakudo/nom: Merge pull request #43 from mlschroe/newexcept
15:23 dalek rakudo/nom:
15:23 dalek rakudo/nom: invoke_catchhandler: tweak call chain after the invoke call, so that the...
15:23 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/f4ee614451
15:24 mls jnthn: yes
15:24 jnthn mls: Is there a guide to invoking the sub profiler? :)
15:24 * jnthn just memorized the command by now... :)
15:24 Woodi --invokeSubProfiler ? :)
15:24 mls as perl6 doesn't support a "--profile" option yet, you need to do 'parrot -R subprof perl.pbc ...'
15:25 jnthn mls: Yes, like that. How hard would it be to add a --profile?
15:25 mls I think parrot supports switching runcores somehow, letme check the sources...
15:26 jnthn Woodi: It's basically the incantation mls++ just wrote out
15:26 Woodi ok
15:26 Woodi now how to compile to pbc ? :)
15:27 jnthn Woodi: It's already built
15:27 jnthn Woodi: Did you "make install" or are you running Rakudo from build directory?
15:27 Woodi installed to /opt/perl6
15:27 jnthn Woodi: OK...there should be a perl6.pbc installed in there somewhere :)
15:28 Woodi ok
15:29 mls hmm, I didn't find a way. I'll talk to the parrot guys.
15:37 Woodi no perl.pbc in /opt/perl6/bin
15:38 Woodi some flag during configure is needed for that ?
15:38 jnthn Woodi: May be in /opt/perl6/lib or some such iirc
15:39 Woodi found
15:41 Woodi got PackFile_Header_validate: This Parrot cannot read bytecode files with version 12.0.
15:42 jnthn Is it the correct Parrot?
15:42 Woodi I removed Debian Star two day ago
15:42 Woodi should be
15:42 jnthn (the one in /opt/perl6/bin)?
15:45 Woodi hang and memory 350mb+, ok ?
15:46 Woodi maybe 10k loop iteration is not needed for trace ?
15:47 jnthn oddness, it normally executes only a little slower than normal for me.
15:47 jnthn That sounds....unusual.
15:48 mls it should only be 10% slower
15:48 jnthn yeah, that's what I've tended to observe
15:48 Woodi want 4.7Gb file ?
15:48 Woodi parrot.pprof.5457
15:49 mls pprof?
15:49 mls that's the wrong one
15:49 jnthn Woodi: Did you do -R subprof, or something else? :)
15:49 mls Do not use -R profile, but -R subprof
15:50 Woodi -R profile
15:51 Woodi -R profiling - was in parrot manual :)
15:52 mls yes, that's the old profile. It writes a line for every executed op.
15:52 mls not really useful for a HLL like perl6
15:53 Woodi ok, doing subprof
15:53 Woodi it realy want to put result on screen, stderr
15:54 jnthn Woodi: 2> callgrind.foo
15:55 dalek rakudo/nom: 04a1158 | coke++ | t/spectest.data:
15:55 dalek rakudo/nom: Track failure modes
15:55 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/04a1158770
15:58 Woodi put that file somewhere ?
15:59 jnthn Woodi: It's viewable with kcachegrind
15:59 jnthn Woodi: If you want to take a look at it
15:59 jnthn Otherwise, you can put it somewhere and I can take a look later on :)
16:00 Woodi apt-get install kcachegrind eta: 3min
16:01 arnsholt Be warned: kcachegrind is awesome
16:01 arnsholt You're gonna want it for all your profiling from now on =)
16:01 jnthn arnsholt: I know! I've spent some hours in that thing :)
16:02 arnsholt I've not even used it that much! =D
16:02 Woodi i'm on putty yet. ^c it ?
16:03 arnsholt But for my research (dang, it's weird to say that!) I'm working on some code in Common Lisp, and the profile output not being trivially inputtable to kcachegrind does annoy me a bit
16:03 Woodi eta 18s
16:04 Woodi ok, eta job in 30 mins, witt see that visualizations :)
16:10 thou joined #perl6
16:14 mkramer joined #perl6
16:19 patch_ joined #perl6
16:25 ethndbst joined #perl6
16:38 * TimToady is home again, fwiw, which ain't much...
16:39 jnthn o/, TimToady :)
16:39 jnthn Hope you had a nice trip, wherever it was :)
16:39 TimToady ate some turkey
16:40 * jnthn still has another month to wait until he gets turkey :)
16:41 Woodi subprof file is in http://lunski.pl/~sylwek/perl/  mainly Grammar.pm and Cursor.pir shows up
16:41 [Coke] jiggity jig.
16:41 TimToady on the negative side, been battling a shingles flareup on my tongue, and the parts of my brane connected to it, so am rather out of sorts still
16:41 jnthn Woodi: oh...hm
16:41 jnthn Woodi: I fear that the benchmark is being rather skewed by compilation time.
16:41 jnthn TimToady: Ouch. Doesn't sound like fun. :( Hope it clears up soon.
16:42 TimToady it's on the mend, but it just takes time
16:43 masak TimToady! \o/
16:43 TimToady on the positive side, having my trigeminal nerve fried by herpes zoster means my migraines are relatively painless, but I still get all the other effects...
16:44 masak :/
16:44 alester joined #perl6
16:44 TimToady and I have a job interview, sort of, this afternoon
16:45 takadonet TimToady: good luck!
16:45 * TimToady is kinda looking around to see if now is a good time to switch to a different Perl 6 vendor :)
16:46 Guest2721 joined #perl6
16:47 TimToady or who will support at least one full-time Perl 6 dude :)
16:47 masak we're hiring :)
16:47 * TimToady also doesn't want to move to Sweden :)
16:47 masak dang. well, worth a shot.
16:48 [Coke] TimToady: good luck.
16:48 TimToady there are several possibilities in Sili Convalley
16:49 TimToady partly it's just that my current company got acquired, and I have no sense of whether I'll have a place in the new scheme of things
16:52 Woodi jnthn: Str is heavy red, maybe just becouse I use string protocol...
16:52 TimToady sigh, only five days of backlogging to do...
16:52 jnthn TimToady: It was quietish here on some of them, as I remember it.
16:53 * TimToady blames turkeys
16:54 * masak decommutes
17:03 * jnthn ponders how on earth SJ has ended up selling first class tickets for cheaper than standard class ones on a normally rather busy Friday evening train...
17:04 jnthn ...sales optimization algorithms clearly do interesting things.
17:05 cognominal_ When going to the Bruxelles peerl workshop with Thalys, first class was cheaper than second class. I suppose they don't want to waste food.
17:07 japhb TimToady, if you have enough working brane, I could really use a ruling on http://irclog.perlgeek.de/​perl6/2011-11-25#i_4755358 and the continuation at http://irclog.perlgeek.de/​perl6/2011-11-25#i_4756773 and http://irclog.perlgeek.de/​perl6/2011-11-25#i_4757166 ... would you mind taking a look?
17:07 cognominal_ there was wifi too but not connected to the larger world. Better check that kind of detail if want to upgrade for wifi.
17:07 jnthn cognominal_: huh, there was wifi but only between people on the train? :)
17:08 mkramer joined #perl6
17:14 Woodi jnthn: maked one milion calls in 111s + 11s profilling. file available on net. looks chomp takes a lot of time
17:15 cognominal joined #perl6
17:15 jnthn Woodi: ooh
17:15 jnthn Woodi: file URL?
17:16 jnthn Woodi: oh, I see it
17:16 tokuhirom joined #perl6
17:18 TimToady japhb: my brane isn't talking to me this morning, at least not on that subject...maybe later today I can think about it
17:18 jnthn Woodi: yeah, that is a lot of time in chomp
17:18 japhb TimToady, roger that, thanks
17:18 jnthn Woodi: I improved it a bit before, but seems it needs some more.
17:18 TimToady and as masak said, it might need a community design if I can't come up with a Gordian sword
17:19 TimToady but I rather suspect I'll have an opinion after I think about it enough
17:21 TimToady if anyone wants to do some thinking for me, they could consider what downsides there might be to putting Failure outside of Any
17:21 mkramer joined #perl6
17:23 envi_ joined #perl6
17:24 PerlJam .oO( Failure is not an Any!)
17:25 jnthn Possibly that using Mu to suppress junction auto-threading now also means you can get passed Failure. But if we consider that stuff outside of Any is meant to be a general place for special things, maybe having at lesat two of them will be better anyway.
17:25 wolverian joined #perl6
17:25 mkramer joined #perl6
17:27 TimToady I think of the outside of Any as "concept" space, so this might be the right direction, but an off-kilter brane doesn't weigh things accurately
17:27 fsergot joined #perl6
17:27 jnthn I have an easier time thinking of upsides than downsides on this one.
17:28 jnthn One is performance...junctions are basically zero impact on the non-junction case because they're just a binding failover.
17:29 TimToady could get the same benefit from .foo.bar.baz not having to worry about dispatching Failure
17:29 jnthn That said, it means that my $x = foo(); really wants to keep meaning that $x has a type constraint of Mu (we've sometimes pondered that it should be Any) since it's good to be able to assign then test in many cases.
17:30 TimToady yes, Any is the default, not the constraint
17:30 jnthn my $x = foo(); if $x { ... } # guess we want this to end up OK, if foo() returned a failure
17:30 TimToady defaults make even more sense on arrays and hashes
17:30 jnthn TimToady: It is now; I've had trouble sensing whether we see that as 100% decided or not. :) I do prefer it with Mu the constraint, Any the default. Just flagging up that putting Failure outside of Any probably ties us more.
17:32 jnthn (ties us more to that decision)
17:32 TimToady I think I'm comfy with distinguishing constraint from default now, because of KeyHashes wanting to delete their default value, which might well be a value, not a type
17:32 jnthn aha :)
17:33 jnthn Yeah, that fits in well.
17:33 mkramer joined #perl6
17:33 kaleem joined #perl6
17:33 TimToady well also have the postulated Each concept outside of Any
17:38 MayDaniel joined #perl6
17:45 jnthn dinner time...got plenty of $dayjob done, so should have some 6 tuits this evening :)
17:45 jnthn bbiab
17:46 dalek nqp: bf55068 | mls++ | src/ (2 files):
17:46 dalek nqp: add experimental support for --profile and --profile-compile
17:46 dalek nqp: review: https://github.com/perl6/nqp/commit/bf55068c6b
17:53 TimToady sorear: I am as yet undecided on whether STM can be mandated across all implementations without being too fiddly for some implementations to do efficiently
17:53 TimToady so I'm staying one remove from that, much like continuations
17:54 TimToady doubtless some wag will point out that continuations and STM are isomorphic...
17:56 mkramer g, e
17:57 mkramer left #perl6
18:11 ksi joined #perl6
18:16 hoelzro joined #perl6
18:22 dalek nqp/unirx: 1f5883b | diakopter++ | src/QRegex/P6Regex/ (2 files):
18:22 dalek nqp/unirx: support ! inverted Unicode properties. Only supports the simple cases, no +/-/&/|/^/(/) combos yet.
18:22 dalek nqp/unirx: review: https://github.com/perl6/nqp/commit/1f5883b112
18:22 diakopter but it's enough for a thousand tests or so for rakudo
18:22 jnthn yay
18:22 jnthn diakopter++
18:23 wolverian nice work
18:23 diakopter heh, I just ported Pm's work from rakudo to nqp
18:24 moritz diakopter++
18:24 jnthn ...and implemented the new syntax, etc :)
18:24 diakopter o_O I wonder how much of rakudo will end up being ported to nqp.
18:24 jnthn diakopter: Probably not much. This just felt like something that belonged with the other regex engine stuff.
18:24 diakopter hm, now how to merge this branch.
18:25 jnthn git checkout master
18:25 jnthn git merge unirx
18:25 diakopter augh. merge conflict
18:26 diakopter src/ops/nqp.ops
18:26 * diakopter has no clue how to proceed.
18:26 diakopter as usual, I'm helpless
18:26 moritz diakopter: lemme check
18:27 moritz diakopter: I think just both branches add ops
18:27 diakopter ergh, I should've pulled before I merged.
18:29 diakopter so, I see these ======= and >>>>>>> in the file
18:30 * diakopter googles
18:31 moritz diakopter: if the file is changed in two branches in different ways, the merge adds those markers
18:31 moritz and the different changes between <<<<<< and ====== the part from one branch
18:32 moritz and the part from the other branch between ====== and >>>>>
18:32 moritz and you just edit that whole seciton, make it look like it should, and then   git add src/ops/nqp.op
18:32 moritz s
18:32 moritz git commit
18:32 moritz diakopter: I've successfully resolved that conflict here locally, should I push it?
18:32 worr how much still needs to be done for rakudo?
18:33 worr I haven't really followed development since before nom
18:33 diakopter moritz: sure; then I get to figure out (I mean, ask :) how to reset my local checkout to your master once you push :)
18:34 moritz worr: that's hard to answer. There's still a whole lot to be done before it implements full perl 6, but much less before it is well usable
18:34 moritz worr: http://perl6.org/compilers/features gives a short overview over what is implemented, and what not
18:35 dalek nqp: 1f5883b | diakopter++ | src/QRegex/P6Regex/ (2 files):
18:35 dalek nqp: support ! inverted Unicode properties. Only supports the simple cases, no +/-/&/|/^/(/) combos yet.
18:35 dalek nqp: review: https://github.com/perl6/nqp/commit/1f5883b112
18:35 dalek nqp: 3e87ffd | moritz++ | src/ (4 files):
18:35 dalek nqp: Merge remote branch 'origin/unirx'
18:35 dalek nqp:
18:35 dalek nqp: Conflicts:
18:35 dalek nqp: src/ops/nqp.ops
18:35 dalek nqp: review: https://github.com/perl6/nqp/commit/3e87ffd04a
18:35 moritz diakopter: if you are on branch:   git fetch; git reset --hard origin/master
18:37 diakopter moritz: thanks for all your help
18:37 cognominal_ joined #perl6
18:37 moritz diakopter: you're welcome
18:38 worr moritz: thanks! I'd love to get involved in development again :)
18:38 worr I've been away for far too long
18:38 lumi_ Is qx{} supposed to capture stderr?
18:39 moritz lumi_: don't think so. It's not specced that way, and perl 5 doesn't do it
18:39 worr is there a bug tracker? rt.perl.org seems to be kind of old
18:39 TimToady generally not, unless you've got a 2>&1 inside and a shell to do it
18:39 moritz worr: rt.perl.org tracks rakudo's bug. Other projects have other bug trackers
18:39 lumi_ moritz: Thought so. Then I think it's broken
18:40 TimToady but the fact that $*ERR is a dynamic var means we nearly have the necessary semantics for IO redirection within Perl 6
18:40 TimToady my $*ERR = $*OUT; qx{}; or some such
18:41 TimToady assuming we make fork/exec track that
18:41 TimToady which is kind of a big assumption
18:42 benabik joined #perl6
18:43 diakopter moritz: I'm curious to see whether rakudo with latest nqp can pass roast/S05-mass/properties*
18:43 diakopter oh, hm
18:44 diakopter I still need to remove a bunch of "is" prefixes from those tests
18:44 diakopter from the test output I mean
18:45 wk joined #perl6
18:46 diakopter oh, found another problem with my previous patches to those tests
18:46 tty234 joined #perl6
18:47 Exodist joined #perl6
18:48 TimToady lumi_: but the basic idea is that qx is for pulling in well-formed data, and interspersing error messages does not contribute to well-formedness
18:48 cognominal___ joined #perl6
18:48 dalek roast: 0b6c46d | diakopter++ | S05-mass/properties- (4 files):
18:48 dalek roast: remove "is" from test output
18:48 dalek roast: review: https://github.com/perl6/roast/commit/0b6c46d445
18:49 TimToady we might contemplate a form of qx that can harvest out-of-band messages and make them available somehow
18:49 TimToady but dumping them to the terminal is a reasonable default, if there is a terminal...
18:49 moritz I'd like to propose a different approve
18:49 moritz I'd rather have qx (as an operator) simple
18:50 moritz rather have shell() and run() some :capture options
18:50 TimToady sure, I was using "qx" generically
18:50 sorear good * #perl6
18:51 moritz and we can have them return objects that hold all the important information
18:51 lumi_ How do I open a rakudobug for it?
18:51 moritz like .out, .err, .status
18:51 moritz lumi_: email to rakudobug@perl.org
18:51 colomon \o/ sorear!
18:51 sorear niecza: my class A::B { } # for jnthn
18:51 p6eval niecza v11-27-gdaa9117: OUTPUT«[31m===[0mSORRY![31m===[0mâ�¤â�¤Pathed definitions require our scope at /tmp/UQ24fKCXvd line 1:â�¤------> [32mmy class A::B [33mâ��[31m{ } # for jnthn[0mâ�¤â�¤â�¤Unhandled Exception: Check failedâ�¤â�¤  at /home/p6eval/niecza/boot/lib/CORE.setting line 880 (die @ 2)…
18:53 jnthn sorear: Then how do we declare X::Base in the setting?
18:53 jnthn (I forbade it initially too, fwiw)
18:53 sorear colomon: (on chromatic and patents) I can sort of see where he's coming from, although I think it's a red herring and he's undermined his own argument for reasons I can explain if you want.  OTOH I am planning a niecza/C.
18:54 colomon sorear: I'm curious to hear how you think he's undermined his own argument, if you don't mind.
18:54 colomon sorear: niecza/C  !?!!
18:56 sorear colomon: [Mandelbrot/gtk] I rather suspect (without seeing a profile) that the problem is the overhead of repeated marshalling calls.
18:56 moritz diakopter: S05-mass/properties-block.t passes now \o/
18:56 colomon sorear: I was hoping that when you get a chance, you could provide optimization suggestions for the core loops (in ExposeEvent and mandel) in https://github.com/colomon/mandelbro​t/blob/master/bin/gtk-mandelbrot.pl  .... and you're ahead of me here.  :)
18:58 colomon sorear: I did a bunch of timings with various changes to the code.  I think mandel is still taking the most time at this point.  (I've already done a bunch of optimizations to the Gtk calls to get to where it is today.)
18:58 colomon sorear: anyway, no rush at all on that, just hoping for some advice, because I'd really love it to be 5-10x faster.  :)
18:59 * colomon is in the middle of his yearly billing session for $work
18:59 diakopter moritz: try the latest patches to the other properties tests
18:59 sorear jnthn: X::Base is a member of a package, so it only makes sense to be declared 'our'.
19:00 sorear colomon: I'll have to help a bit later ...
19:00 moritz properties-scripts also passes
19:01 sorear colomon: you saw "Microsoft sues TomTom" right?  if you actually read the link, you'll see that Microsoft has a bunch of random patents "dashboard-mounted computer" etc
19:02 sorear colomon: I think it's a safe bet that both Parrot and Mono infringe a large number of Microsoft patents, since Microsoft has written several VMs and HALs over the years and has probably patented a lot of the territory
19:02 sorear colomon: TomTom makes money, Parrot and Mono don't, which is probably why they picked TomTom
19:03 moritz and it also likely that both sue lots of patents by IBM, Google and any other big software patent holder
19:03 moritz s/sue/infringe/
19:03 moritz because so many trivialities are patented
19:03 sorear colomon: now, if you compare Parrot and Mono, there's one difference that stands out - Microsoft has explicitly (albeit not as a binding contract) granted a usage license to all patents required for a third-party implementation of the CLR
19:04 diakopter sorear: I fervently agree with "I think it's a safe bet that both Parrot and Mono infringe a large number of Microsoft patents"
19:04 colomon sorear: oh.  so you're saying that (at least in theory) Parrot is more vulnerable than Mono, not less?
19:04 sorear colomon: yes.
19:04 sorear IANAL, of course.
19:05 colomon dang
19:07 diakopter sorear: have you thought about what GC scheme Niecza/C might use?
19:08 diakopter generational concurrent :)
19:09 sorear diakopter: I am a huge fan of generational garbage collectors with copying nursery eviction.  I'll probably implement one of those first
19:09 sorear Dunno about concurrent though.  Majorly annoying locking issues there.
19:11 sorear I say "implement" not "use" because I'll almost surely start with Boehm-Demers-Weiser
19:12 molaf joined #perl6
19:12 colomon sorear++
19:12 [Coke] .u ���
19:12 phenny U+FFFD REPLACEMENT CHARACTER (�)
19:12 phenny U+FFFD REPLACEMENT CHARACTER (�)
19:12 phenny U+FFFD REPLACEMENT CHARACTER (�)
19:14 * sorear really likes thinking about GCs, but they're too time intensive to play with as much as ey'd like
19:16 sorear any Windows users (jnthn, diakopter) available and willing to take a look at v12 for me?
19:16 diakopter moritz: if you bump the NQP_REVISION, you can greenify Unicode properties on the feature matrix
19:16 diakopter sure
19:17 sorear diakopter: http://feather.perl6.nl/~sorear/n12draft.zip is what will be the v12 binary release
19:17 sorear diakopter: the last time you helped me, you ran into the problem that niecza chokes on CRLF
19:18 sorear diakopter: I still haven't fixed that.  But I want to know if it works, if you use _Unix_ line endings in your source code and CORE.setting
19:18 sorear I'd rather have "works with major proviso" than "doesn't work"
19:18 dalek rakudo/nom: 9d79fef | moritz++ | src/core/Exceptions.pm:
19:18 dalek rakudo/nom: get rid of nested packages X packages in Exceptions.pm
19:18 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/9d79feff03
19:18 dalek rakudo/nom: a8dfd47 | moritz++ | tools/build/NQP_REVISION:
19:18 dalek rakudo/nom: bump NQP revision to get Unicode propertiy fixes by diakopter++
19:18 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/a8dfd47838
19:18 [Coke] I have a windows box. what do you need?
19:19 sorear [Coke]: 1. unzip linked file 2. make sure lib\CORE.setting contains \n and not \r\n 3. run\Niecza.exe; say 2 + 2
19:20 [Coke] 1 and 3 work fine.
19:21 moritz phenny: tell jnthn that with http://moritz.faui2k3.org/tmp/x-buf-as-str.patch setting compilation still fails with "The following packages were stubbed but not defined: X::Buf::AsStr"
19:21 phenny moritz: I'll pass that on when jnthn is around.
19:22 [Coke] if I gvim -b the file on windows, no extraneous ^Ms.
19:22 jnthn sorear: just tried it; it works
19:22 phenny jnthn: 19:21Z <moritz> tell jnthn that with http://moritz.faui2k3.org/tmp/x-buf-as-str.patch setting compilation still fails with "The following packages were stubbed but not defined: X::Buf::AsStr"
19:22 moritz jnthn: I haven't been able to golf that one to non-setting code (note that the 'package X { }) is now gone
19:22 sorear awesome
19:22 moritz )
19:23 sorear how about run\Niecza.exe -c -e 'say "Hello world"'; obj\MAIN.exe ?
19:24 * sorear leaves for a bit
19:24 [Coke] http://nopaste.snit.ch/97024 --parsefail
19:25 jnthn sorear: Run.MAIN.exe works
19:25 diakopter [Coke]: try this instead
19:26 diakopter run\Niecza.exe -c -e "say ""Hello world"""
19:26 diakopter run\Niecza.exe -c -e "say 'Hello world'"
19:26 diakopter windows doesn't take single quotes as quoters
19:26 [Coke] first one gives "unsupported use of bare say"
19:26 [Coke] second one works.
19:27 diakopter hm, maybe it's three " to make it work
19:27 jnthn diakopter: heh, I did that automatically without noticing I had :)
19:27 [Coke] diakopter: doh, didn't even think to check that, I assumed he was cut and pasting from something that was expected to work. ;)
19:27 dalek nqp/nci: 6e9d75d | jnthn++ | / (4 files):
19:27 dalek nqp/nci: Add CPointer REPR, which represents a raw pointer that we don't attempt to manage at all.
19:27 dalek nqp/nci: review: https://github.com/perl6/nqp/commit/6e9d75df2a
19:27 dalek nqp/nci: a52b468 | jnthn++ | src/ops/nqp_dyncall.ops:
19:27 dalek nqp/nci: Implement CPointer handling in native caller.
19:27 dalek nqp/nci: review: https://github.com/perl6/nqp/commit/a52b4681ef
19:30 wk joined #perl6
19:33 masak TimToady: the sword is Alexandrian, not Gordian. if they'd had the idea about the sword in Gordium, they wouldn't still have a knot. :P
19:34 TimToady it was a Gordian sword afterwards
19:34 TimToady regardless of whether Alexander took it away from there
19:35 hoelzro left #perl6
19:36 TimToady and the flying mugs now belong to Perl 6, even if they originally belonged to the hotel
19:44 masak ;)
19:53 dalek rakudo/nci: 61db3e8 | jnthn++ | src/core/traits.pm:
19:53 dalek rakudo/nci: Implement exporting of types.
19:53 dalek rakudo/nci: review: https://github.com/rakudo/rakudo/commit/61db3e8447
19:54 dalek zavolaj/v2: 4561a4f | jnthn++ | lib/NativeCall.pm6:
19:54 dalek zavolaj/v2: Support opaque pointers in NativeCall.
19:54 dalek zavolaj/v2: review: https://github.com/jnthn/zavolaj/commit/4561a4f4c8
19:55 jnthn OK, only thing that new Zavolaj doesn't support that old one did is passing arrays. And the way it was done before was one of the most disgusting hacks I ever did...
19:56 colomon hack!  hack!  hack!
19:56 colomon jnthn++
19:56 jnthn colomon: I may wonder to the sofa and actually *design* it this time before I hack. ;)
19:56 colomon designed hack!
19:57 jnthn :)
19:57 diakopter talk about an armchair expert
19:57 diakopter er, sofa
19:57 diakopter er, nm
20:00 dalek nqp/nci: 8c94248 | jnthn++ | .gitignore:
20:00 dalek nqp/nci: Update .gitignore.
20:00 dalek nqp/nci: review: https://github.com/perl6/nqp/commit/8c94248c61
20:00 masak also, it's a sofa you "wonder" to :P
20:00 masak not a regular sofa, then.
20:00 tty234 joined #perl6
20:13 vmspb joined #perl6
20:13 diakopter pmichaud: ping
20:17 tokuhirom joined #perl6
20:24 takadonet left #perl6
20:40 cooper joined #perl6
20:50 cooper joined #perl6
20:50 cooper joined #perl6
21:04 aindilis` joined #perl6
21:05 dalek roast: feb17e7 | moritz++ | S05-mass/properties-general.t:
21:05 dalek roast: fix a thinko in S05-mass/properties-general.t; rakudo refudges
21:05 dalek roast: review: https://github.com/perl6/roast/commit/feb17e7502
21:09 envi_ joined #perl6
21:10 ethndbst joined #perl6
21:10 ethndbst joined #perl6
21:11 dalek roast: b05364c | moritz++ | S05-mass/properties-general.t:
21:11 dalek roast: more thinkos
21:11 dalek roast: review: https://github.com/perl6/roast/commit/b05364cb61
21:13 simcop2387 joined #perl6
21:14 rdesfo joined #perl6
21:21 thou joined #perl6
21:33 tadzik 'evening
21:35 tokuhirom joined #perl6
21:36 aindilis joined #perl6
21:37 tokuhirom joined #perl6
21:40 pmurias joined #perl6
21:42 dorlamm joined #perl6
21:44 Eth4n joined #perl6
21:55 jnthn o/ tadzik
21:57 rgrau` joined #perl6
21:59 PacoLinux joined #perl6
22:07 masak tadzik! \o/
22:18 am0c joined #perl6
22:18 [Coke] moritz: are working on adding those to t/spectest.data also?
22:18 mj41 joined #perl6
22:19 [Coke] *you
22:20 [Coke] (if not, I'll get to them sometime tonight, probably)
22:43 tokuhirom joined #perl6
22:46 rdesfo left #perl6
22:51 * sorear returns
22:55 whiteknight joined #perl6
22:58 lestrrat joined #perl6
23:22 jferrero joined #perl6
23:28 sivoais joined #perl6
23:37 wolfman2000 joined #perl6
23:42 tokuhirom joined #perl6
23:43 masak 'night, #perl6
23:45 colomon \o
23:48 tokuhirom joined #perl6
23:50 tokuhirom joined #perl6
23:52 tokuhirom joined #perl6

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

Perl 6 | Reference Documentation | Rakudo