Camelia, the Perl 6 bug

IRC log for #mojo, 2013-06-10

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

All times shown according to UTC.

Time Nick Message
00:11 russum joined #mojo
00:20 good_news_everyone joined #mojo
00:20 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/gSVmww
00:20 good_news_everyone mojo/master 8e9af4f Sebastian Riedel: small optimizations
00:20 good_news_everyone left #mojo
00:21 good_news_everyone joined #mojo
00:21 good_news_everyone [mojo] kraih tagged v4.14 at 3db75d0: http://git.io/BCgogg
00:21 good_news_everyone left #mojo
00:40 HtbaaPi joined #mojo
00:57 inokenty joined #mojo
01:15 russum joined #mojo
01:49 russum joined #mojo
01:56 Meiermann joined #mojo
02:00 d4rkie joined #mojo
02:05 egopro joined #mojo
02:08 shmuel joined #mojo
02:13 * tempire tries to get selenium working with jasmine
02:35 russum joined #mojo
02:47 bpmedley tempire: Have you seen PhantomJS?
02:47 tempire yes
02:48 tempire don't see any use for it
03:02 russum joined #mojo
03:02 egopro joined #mojo
03:05 preflex_ joined #mojo
03:11 russum1 joined #mojo
04:51 egopro joined #mojo
05:14 denisboyun joined #mojo
05:22 egopro joined #mojo
05:29 egopro joined #mojo
05:29 d4rkie joined #mojo
05:49 hesperaux joined #mojo
05:49 basiliscos joined #mojo
05:49 Britzel joined #mojo
05:57 Meiermann joined #mojo
06:01 suy joined #mojo
06:09 basiliscos joined #mojo
06:34 ver joined #mojo
06:54 egopro joined #mojo
07:02 dpetrov_ joined #mojo
07:03 yakudza joined #mojo
07:09 Dandre good morning
07:09 sri happy wwdc day! \o/
07:13 egopro joined #mojo
07:15 Dandre hi sri. Have you seen my discussion last Friday about custom default renderer?
07:19 arpadszasz joined #mojo
07:34 sri Dandre: i think i commented on it
07:38 Dandre sorry I haven't seen
07:38 Dandre /me reading history log
07:39 sri if i understand correctly the feature you want existed already in the past and got removed again
07:40 Vandal joined #mojo
07:41 sri if it was the render(xml => {foo => 'bar'}) thing
07:41 Vandal joined #mojo
07:41 Dandre You gave hook advice I remember
07:42 sri yea, the hook doesn't exist anymore
07:42 Dandre look at this please: http://pastebin.fr/27651
07:43 sri that's a dead link
07:43 Dandre in Controler, if any=> is reached, the format requested by the client is deleted so a custom renderer can't get this information
07:43 Dandre oups
07:44 sri those are all topics that were up for discussion like a year ago, i'm not particularly interested in discussing it again tbh
07:44 Dandre http://pastebin.fr/27675
07:44 sri unless there's new information/proposals
07:46 Dandre you mean that deleting format for any renderer was a feature resquest? if so I'm sorry, didn't know
07:46 sri i don't know, would have to read up on code first
07:46 Dandre BTW I am just discovering mojo
07:49 Dandre ...  for any renderer
07:49 Dandre I meant 'default' renderer
07:50 sri if something is a feature there should be tests covering it
07:50 denisboyun joined #mojo
07:56 sri btw. i'm not opposed to reintroducing the before_render hook, but the rules are stricter now, every new hook needs at least two good use cases to be considered
07:58 Dandre I don't think that the before renderer could solve my issue. I may be wrong
08:03 jzawodn joined #mojo
08:05 judofyr joined #mojo
08:08 judofyr HEY MOJO
08:08 judofyr sri: dots looks nice
08:08 sri HEY JUDOFYR
08:09 * sri has his fingers crossed for dots and sub signatures in 5.20 :o
08:09 judofyr maybe Perl will finally become a proper language
08:09 * judofyr ducks
08:10 sri Ruby is a Perl!
08:12 sri http://instagram.com/p/Z616sag_1u/ # ;p
08:13 judofyr Ruby.is_a?(Everything)
08:15 nicolaas joined #mojo
08:31 sri infix subroutines would be cool too :)
08:33 kwa Infix subroutines? *jumpst on Google*
08:33 kwa s/jumpst/jumps/; # that's how excited I was
08:34 sri you're not following me on twitter!1
08:34 sri http://www.xray.mpe.mpg.de/mailing-lis​ts/perl5-porters/2013-03/msg00389.html
08:35 kwa heh, I am, but only in the last couple of months. :P
08:37 sri infix subroutines is how perl6 operators work i believe
08:38 sri with multi subs
08:39 kwa I was grokking some perl6 code last week and it made my head hurt. Was like looking at perl for the first time, not understanding Perl context and seeing the topic variable and other crazy operators.
08:40 sri http://doc.perl6.org/language/operators # it's pretty crazy, but i like it
08:41 kwa Good lord. :)
08:41 sri "a space age language" as stevan would say
08:42 maxhq joined #mojo
08:44 kwa I can imagine. It'll make your mojo one liners more interesting, too.
08:46 Dandre sri: I have this small patch: http://pastebin.fr/27679
08:46 Dandre before calling 'any' render, Controler puts in stash an array reference to the formats requested by the client. This doesn't break any test in Mojo distribution. Can you comment please?
08:46 sri i actually like the ruby way for operator overloading more than multi subs
08:47 sri but the infix, postcircumfix... stuff rocks
08:48 sri Dandre: i don't see the value i'm afraid
08:48 Dandre what do you mean?
08:50 sri Dandre: it's an additional line of code, without context, no visible use case
08:50 Dandre the whole modified respond_to subroutine: http://pastebin.fr/27680
08:52 Dandre sorry
08:53 kwa Ah, respond_to. I know that beast well.
08:53 Dandre here it is http://pastebin.fr/27681
08:53 sri i'm very hard to convince though, it's usually better to try another core dev first and make him convince me ;)
08:54 bpmedley Mojo::IOLoop->recurring(5 => \&IOLoop::Feeder::the_tick);  <-- A line such as this seems to be called in every hypnotoad process?   If so, how can I limit to one instance running at a time?
08:55 Dandre sri: I've seen ;-)
08:58 Dandre I just want to use mojo and I don't think that what I intend to do is particularly specific: I get one data structure from my model ans I want to present it to rest client in either supported format it requests without duplicating code. May be I haven't understood how to do that this stock mojo
08:59 batman bpmedley: the ioloop does not start in the parent. Mojo::IOLoop->start is called in each child process.
08:59 sri for a change like this where you add a new rendering feature my first thought is always "how would i explain that in the rendering guide?"
09:00 bpmedley batman: Hrmm.. that would explain it.  Is there a way for just one child to process the Mojo::IOLoop?  Does that make sense?
09:01 * sri often designs features documentation driven
09:02 Dandre ok
09:02 batman bpmedley: i'm not sure. i actually got the same problem. I'm thinking about starting a completely different worker for that event loop.
09:02 bpmedley batman: How?
09:02 batman bpmedley: since it sounds like it's not that you want "one child" to do it, but rather "one process"
09:03 bpmedley batman: Yes, I think you're correct.
09:05 batman bpmedley: https://gist.github.com/jhthorsen/5747415
09:05 bpmedley batman: Will that stop and start with hypnotoad?
09:06 batman no. that's the whole idea. are you sure you want to start hypnotoad for that "job" ?
09:06 bpmedley batman: I want my worker processes to be self contained in the app and stop/start with the app.  Does that make sense?
09:07 batman yes. and i think you're making things a lot more complicated than you need to
09:07 sri i'm not saying it's a good idea... but you could just use a mutex to make only one worker do stuff
09:07 bpmedley Understood.  How can i simplify?
09:07 batman and if only one child can do the task...then it doesn't sound like a job for a hypnotoad worker
09:08 batman make a separate process.
09:08 sri like the lock file hypnotoad uses as an accept mutex
09:08 bpmedley Hrmm.  How can I stop and start the process when hypnotoad stops and starts?
09:09 batman bpmedley: why do you need to start/stop the procees when hypnotoad stops and starts?
09:09 batman sri: do you think that's a good idea? even though it's possible?
09:09 sri "i'm not saying it's a good idea..."
09:10 bpmedley batman: I want a web app that can be self contained with no external requirements other than "stop or start the server".
09:10 Dandre sri: http://pastebin.fr/27682
09:10 bpmedley sri: How to make it a good idea?
09:10 batman i hear what you're saying. but i don't see any good reason for why you want that.
09:10 bpmedley batman: To make it super easy for the user.
09:10 batman sri :)
09:11 batman bpmedley: provide an init script or something similar
09:12 sri bpmedley: i only know how to make it a worse idea ;p
09:12 batman https://metacpan.org/source/JHTHORSE​N/Toadfarm-0.10/etc/init.d/toadfarm <-- i bundle an example on how to start my server...
09:12 bpmedley batman: Understood.  However, that's not the route I want to take.
09:12 bpmedley sri: Gaw.  Maybe that's what we need.. :)
09:12 sri Dandre: that doesn't show a use case though, which for me is the important part, something needs to get easier
09:12 batman bpmedley: ok. then i don't want to help you furter. reason: i've already been down that road, and it blew up :/
09:13 bpmedley batman: I would love to hear what happened.
09:13 sri i think you could actually hijack one worker process and make it not accept any connections :)
09:14 sri depends on multiple side effects though
09:14 bpmedley sri: Sweet.  Then I could just increase the hypnotoad pool by my number of worker processes?
09:14 batman bpmedley: i just suck at ipc programming
09:14 batman bpmedley: and i really like to make the simplest solution.
09:15 bpmedley batman: Practice makes perfect.. :)
09:15 batman i also suck at making simple solutions though, but i'm getting better :)
09:15 sri it's a very very bad idea and will prolly break in the future
09:16 bpmedley Understood.  Do you think the rationale of a worker being started/stopped by the web server is useful?
09:16 batman no
09:16 sri use a lock file to decide which worker gets to be the one, and disable the accept mutex
09:16 sri no i'm not gonna say how to do it ;p
09:17 sri if you can figure it out you know enough to be aware of the risks ;)
09:17 bpmedley Cool, thx.
09:18 sri for what it's worth, i'm not sure i'm fully aware of the risks
09:20 bpmedley Hrmm.  Is there a way to run a sub on server exit?
09:28 fhelmber_ joined #mojo
10:22 batman END {} ?
10:22 batman :P
10:28 bpmedley https://gist.github.com/brianmed/5747787  <-- Thoughts on this solution to stop/start workers when hypnotoad stops/starts?
10:50 cpk joined #mojo
10:55 cpk i'm trying to deploy a web on windows
10:55 cpk some people suggest me to use apache/nginx reverse proxy
10:55 cpk what I don't understand is how does it solve the single process execution of the app ?
10:58 yakudza joined #mojo
11:34 mire_ joined #mojo
11:45 Vandal joined #mojo
11:50 cpk left #mojo
11:59 libsysguy joined #mojo
12:18 preflex joined #mojo
12:28 ryozi joined #mojo
12:37 d4rkie joined #mojo
12:41 asarch joined #mojo
12:42 preflex joined #mojo
12:57 Dandre how can I register a custom renderer in a Mojolicius Lite application?
13:10 BeDa joined #mojo
13:13 moltar joined #mojo
13:13 depesz anyone can explain to me why i'm getting "Bad name after error' at template flash_messages.html.ep line 14." error in this template: http://depesz.privatepaste.com/2eb125d281
13:19 inokenty joined #mojo
13:20 Dandre sri: here is a short testcase for my patch: http://pastebin.fr/27693
13:33 gryphon joined #mojo
13:34 arpadszasz joined #mojo
13:42 * marcus craves the dots pragma
13:43 btyler joined #mojo
13:46 libsysguy Has anybody looked into using AnyEvent::DBI::MySQL?
13:50 Bender joined #mojo
14:12 toebu joined #mojo
14:13 bluescreen joined #mojo
14:15 mattastrophe joined #mojo
14:32 Vandal joined #mojo
14:38 ajmrch joined #mojo
14:55 whitebook joined #mojo
15:03 sh4 joined #mojo
15:04 delias joined #mojo
15:11 mire_ joined #mojo
15:19 btyler joined #mojo
15:48 jberger ETOOMANYBACKLOGCONVOS
16:20 judofyr_ joined #mojo
16:20 mire_ joined #mojo
16:43 asarch__ joined #mojo
16:45 sri oh, wwdc livestream starts soon
16:50 denisboyun_ joined #mojo
16:58 yakudza joined #mojo
17:11 mire_ joined #mojo
17:22 basiliscos joined #mojo
17:37 btyler joined #mojo
17:42 mattastrophe joined #mojo
17:50 * marty had to turn off wwds livestream cuz the applause after each sentence invoked his fan-boy gag reflex.
17:51 mattastrophe joined #mojo
18:02 sri haha, new mac pro looks like a trashcan
18:06 sri oooh, a tiny trashcan :)
18:13 btyler joined #mojo
18:18 whitebook joined #mojo
18:28 rem_lex| joined #mojo
18:41 cpk joined #mojo
18:46 cpk hi there
19:07 cpk left #mojo
19:15 libsysguy "A tiny trashcan, and we think you're going to love it"
19:21 jberger http://www.xray.mpe.mpg.de/mailing-lis​ts/perl5-porters/2013-06/msg00316.html
19:21 zackiv31 joined #mojo
19:21 preaction holy hell
19:22 preaction so it IS true, Perl 5 will become Perl 6
19:23 libsysguy 0_0
19:23 preaction so E3 is going on. WWDC is going on. and now this. this is like, the best day ever.
19:33 whitebook joined #mojo
19:39 jberger hehe, as long as Perl5 doesn't go to static sigils I'm cool :-)
19:48 sri i've made the mistake of bringing the topic up in #p5p... i think there might now be a stronger opposition to adding use dots :/
19:50 libsysguy joined #mojo
19:51 jberger :o
19:51 preaction i thought that was what #p5p was for: arguing against evolution of Perl 5
19:52 sri yea, i really should know better
19:52 preaction i mean, a lot of people are in agreement that proper method signatures should be in core, and even then it gets quite a lot of vitriol
19:53 * jberger forbids sri from using strict for one month in penance
19:53 sri you only have to look at the perl language family
19:53 sri which parts are consistent among all of them
19:54 sri dots should really be a no brainer
19:55 jberger truth is, I don't have any opinion on dots personally, however, if it helps Perl adoption and retention (from new users or old) then I'm all for it
19:56 sri problem is that it's hard to argue for aesthetics in language design
19:56 sri 10 people, 10 different opinions
19:56 batman off topic: this is really weird: some guy has pointed his domain to my ec2 instance, and running requests against my hypnotoad server
19:56 jberger but if it can be done with lexical scope, then who cares
19:57 batman not sure if i should be worried or not :P (http://iotools.net is not my domain)
19:57 preaction or maybe they're just forging the HOST header
19:58 batman what does that mean?
19:59 preaction i can connect to your IP, and ask for whatever domain name I want. it doesn't have to be a domain name that points to that IP in the DNS system
19:59 batman i mean... i understand the "host header" part, but why..?
19:59 preaction looking for open proxies
19:59 preaction not "why forge the host header", but "why not?"
20:00 batman preaction: i just checked. there's no forging. it's actually pointing at the same ip
20:00 preaction then someone screwed up. i'd do goatse
20:00 batman goatse?
20:00 preaction nevermind. don't google it
20:00 sri ban for the first one to post a link ;p
20:01 batman i got a lot of GET http://iotools.net/images/morfeoshow/ze​nzanon_80m-5475/thumbs/Untitled-20.jpg 200 in my log...
20:01 batman i've set up a default route to always respond with 200 and "This is not the resource you're looking for. Move along" :)
20:02 preaction 404 would be more appropriate. or 403
20:02 batman yeah. normally. i got a bit weird config...
20:02 sri be nice and redirect them to cuteoverload or so
20:02 batman :)
20:07 batman https://github.com/jhthorsen/mojo-apns # anyone who wants to review this before i push it to cpan?
20:10 perlite_ joined #mojo
20:11 sri neat
20:11 sri only thing i don't like is the sent event
20:12 sri when would i ever want to know thay *any* message has been sent? instead of a drain event to knwo that *all* messages have been sent?
20:13 sri or is it just a diagnostics event?
20:14 batman i'm not quite sure... i think maybe i added it too late yesterday
20:14 batman i'll remove it.
20:14 batman easier to remove stuff before release than later :P
20:14 sri :)
20:14 batman what about drain?
20:14 batman is it really useful?
20:15 sri imo it is essential
20:15 batman then i'll keep it ;)
20:15 batman and now: SHIP THE SHIT!
20:15 sri how else would you know when the buffer is getting too big
20:15 batman no idea
20:16 sri for all you know the apple server is super slow accepting only one byte per minute
20:18 batman right
20:20 batman *fixed*
20:22 bluescreen joined #mojo
20:33 gtodd joined #mojo
20:33 gtodd
20:36 batman sri: https://metacpan.org/module/Mojo::APNS # thanks! :)
20:37 sri batman: oops, i think the error event has a problem too :S
20:38 sri just noticed it mentions message specific errors
20:38 batman is that bad?
20:38 sri that's usually a no no for async apis
20:38 batman i was thinking the "timeout" event as error was bad... but i still went with it :/
20:38 sri you want to know which message the error was for
20:39 batman ah! crap.
20:39 * batman ponders
20:39 sri personally i would go with a confirmation callback
20:40 sri $apns->send(... => sub { my ($self, $err) = @_ }) or so
20:41 batman yup
20:42 batman crap. need to try to save a domain for my brother
20:42 batman i like send(...,sub {}) since it works nicely with delay
20:42 batman but then i think i will do sub { my($apns, $error) = @_; ... }
20:43 batman and let the $cb be called after each message is sent
20:43 good_news_everyone joined #mojo
20:43 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/hHR0WA
20:43 good_news_everyone mojo/master 060a775 Sebastian Riedel: fixed a few error reporting bugs in Mojo::IOLoop::Client and Mojo::IOLoop::Server (closes #503)
20:43 good_news_everyone left #mojo
20:46 * sri had no idea IO::Socket::INET clears $! and only leaves $@
20:47 marcus apple server is at certainly super slow now :-/
20:47 sri the livestream was terrible
20:47 * marcus is trying to ruin his phone with ios 7, failing
20:48 marcus And safari claims it is *I* who's offline.
20:51 sri http://www.apple.com/mac-pro/ # soooo, who
20:51 sri ...'s going to buy me one?!!!
20:52 sri DAT UNIFIED THERMAL CORE
20:52 preaction that is quite possibly the ugliest thing they've ever made
20:52 preaction except the original imac
20:54 preaction why did they go rounded corners on the airport base station, but cylinder on the mac pro?
20:55 marty inefficiency of thermal cooling has always been my biggest gripe with towers.    I really like the new mac-pro.  If I needed one I'd buy it.
20:56 preaction i need the upgrade to os 10.10. multiple displays really sucks on 10.9
20:56 * preaction NEEEEEEEEEDS IT
20:56 marty I used to buy all my servers from quietpc and they work well.  I've got one here, all 44 lbs of it.  :)
20:59 marty To embed or not embed in mongodb, that is the question.
21:00 marcus I also want working multiple displays
21:00 marcus But I *need* the new ios.
21:00 KindOne joined #mojo
21:00 marcus I'm not so sure I want to upgrade my work mac to the first beta.
21:01 marcus but I cary two phones now anyways, so it's not a crisis if the ios beta is buggy.
21:35 lukep joined #mojo
21:53 Adurah joined #mojo
21:56 mire_ joined #mojo
22:01 gryphon joined #mojo
22:26 denisboyun_ joined #mojo
22:49 batman sri: do you think the callback to send() should be called once(drain => ...) or when the data has been written?
22:49 batman i'm leaning against drain
22:54 batman https://github.com/jhthorsen/mojo-apns/comm​it/91ddf685cb65749accfccdfdae624a8f4018f7c6
23:05 ryozi joined #mojo
23:34 mire_ joined #mojo
23:39 popl joined #mojo

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