Camelia, the Perl 6 bug

IRC log for #mojo, 2012-03-16

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

All times shown according to UTC.

Time Nick Message
00:12 abra joined #mojo
00:15 tm good night all
00:32 zivester joined #mojo
00:36 tempire Mojo::IOLoop->stream($self->t​x->connection)->timeout(300) is indeed scary
00:37 tempire what's the problems with connection->timeout?
00:38 sri tempire: try to make it work ;p
00:39 tempire well I could try.  but I would rather sit at the alter of your wisdom
00:39 sri but to make it short, Mojo::UserAgent depends on ->connection being an id
00:40 sri storing the stream in ->connection also has *a lot* of potential for new leaks
00:40 sri not just our leaks, but also leaks in user code
00:40 tempire tangent: namecheap has an option to automatically set your mx records to google apps.
00:44 tempire storing stream in the connection makes a lot of sense.
00:44 tempire but I can't comment on the potential leaks.
00:51 GitHub37 joined #mojo
00:51 GitHub37 [mojo] kraih pushed 1 new commit to master: http://git.io/E6Ea6w
00:51 GitHub37 [mojo/master] removed experimental status from many classes, helpers, attributes and methods (closes #295) - Sebastian Riedel
00:51 GitHub37 left #mojo
00:51 sri huge commit
00:51 tempire whoa
00:51 tempire it's like a whole new day
00:51 sri but now we are down to things being experimental that *really* need attention
00:53 xaka joined #mojo
00:56 sri before anyone asks, ioloop layer is still experimental
00:57 sri but it should be pretty clear now what actually needs to be reviewed
00:59 sri ooooh, community is back
00:59 sri \o/
00:59 tempire it's about time
00:59 sri did you reach season 3 yet?
01:01 d4rkie joined #mojo
01:01 tempire oh yeah, I finished watching them months ago
01:01 sri :o
01:01 tempire when I watch a series, I go through 2-3 episodes per day
01:02 tempire ack
01:02 tempire sketch doesn't open eps files
01:02 tempire and neither does inkscape
01:02 tempire D:
01:03 sri maybe convert with imagemagick?
01:03 sri sketch 2.0 will be interesting, they are aiming for a fireworks replacement now
01:05 tempire how is fireworks different from illustrator?
01:06 sri very, fireworks targets pixels
01:08 tempire dernit
01:08 tempire are there any other options?
01:08 tempire must I purchase illustrator?
01:15 sri tempire: https://github.com/kraih/mojo/blob​/master/lib/Mojo/UserAgent.pm#L255 # that part is the big problem
01:17 scott joined #mojo
01:21 tempire hey you weren't joking.  ep 11 is available RIGHT NOW!
01:23 sri and some new show called missing
01:23 tempire I'm going through supernatural right now
01:23 tempire 1/3 through season 3
01:24 sri i never really started with supernatural
02:44 scott joined #mojo
03:56 hshong joined #mojo
04:40 mattastrophe joined #mojo
06:36 kanishka joined #mojo
06:55 cosmincx joined #mojo
07:02 mattastrophe joined #mojo
07:10 Foxcool joined #mojo
07:13 batman joined #mojo
07:21 pfig joined #mojo
07:22 pfig_ joined #mojo
07:55 marcus 'ello
07:55 spleenjack joined #mojo
07:56 marcus sri: illustrator can do pixels too.
07:56 marcus our design director uses it for like everything.
07:57 tempire I caved and downloaded the trial of illustrator
07:57 tempire I remember it being amazing when I learned it in 2008
07:57 tempire and that nothing else compared to its capability
08:04 marcus sounds right.
08:06 metys can anyone tell me, how can I set user/pass to Mojo::UserAgent, same as wget --http_user/--http_password?
08:07 tempire http://mojolicio.us/perldoc/Mojo/UserAgent - it's in the first couple lines of the synopsis
08:07 tempire well, 8 lines
08:08 metys sorry! thx :-)
08:11 marcus tempire: did you open source your irc thingie?
08:12 tempire not yet.  I want to refine the way it handles events a bit.  And have it show the channel members.
08:24 batman joined #mojo
08:26 tempire wow.   DBIx::Class::InflateColumn::Currency is a nifty tool.
08:44 pfig joined #mojo
09:33 preflex_ joined #mojo
09:35 tempire hmm, interesting.  using gist.github.com as a markdown blogging platform.
09:40 marcus shiny happy people
09:51 gbacon joined #mojo
09:57 mire_ joined #mojo
09:59 spleenjack joined #mojo
10:15 lammel2 joined #mojo
10:25 soulchild joined #mojo
10:51 dpetrov_ joined #mojo
10:54 * sri yawns
10:55 sri marcus, tempire, crab: so none of you has a better idea than Mojo::IOLoop->stream($c->tx-​>connection)->timeout(300);?
10:56 sri just don't complain when we can't change it anymore
10:56 marcus sri: I've never used ->stream. What's wrong with it?
10:57 sri to quote tempire "Mojo::IOLoop->stream($self->t​​x->connection)->timeout(300) is indeed scary"
10:59 sri some have been asking for $self->tx->connection->timeout(300), but nobody wants to actually work on it
11:04 sri i'm actually close to the point where i stabilize what we have now
11:07 sri marcus: well, now is the time to review the code
11:07 marcus sri: I have no time atm. working for 2 more hours, and then leaving for the mountain whole weekend.
11:09 batman left #mojo
11:09 sri just don't complain later, the api has been areound for 6 months now and i want to stabilize it finally
11:10 sri and i mean everybody here!
11:12 sri same goes for this ticket https://github.com/kraih/mojo/issues/258
11:12 sri speak up or it goes stable
11:15 sri or i just remove all traces of unstable api from the docs
11:19 marcus sri: I'm very happy with app_url
11:20 marcus I'd be fine with stabilizing that
11:22 sri hmm
11:29 GitHub13 joined #mojo
11:29 GitHub13 [mojo] kraih pushed 1 new commit to master: http://git.io/08Ruww
11:29 GitHub13 [mojo/master] removed experimental status from a few more classes, helpers, attributes and methods, also removed experimental examples from stable documentation - Sebastian Riedel
11:29 GitHub13 left #mojo
11:29 sri now it's just the Mojo::IOLoop layer that's still experimental, nothing else
11:30 sri i've removed examples for how to change the inactivity timeout per connection though
11:31 sri if it becomes an FAQ again someone will hopefully solve it
11:41 mire_ joined #mojo
11:48 Ultali joined #mojo
11:51 Ultali hi all, I'm currently trying to do some FB auth using Mojolicious::Plugin::OAuth2 from github, I'm getting the following error from code that previously worked: Can't use string ("") as a subroutine ref while "strict refs" in use at script/../lib/Mojolicious/Plugin/OAuth2.pm line 81.
11:51 Ultali protips or general ridicule are welcome :)
11:52 Ultali this all kicked off when I tried to add email as a scope parameter... reverting my code the problem persists
11:54 Ultali also since its my first time here, thanks for Mojolicious its making my life a lot easier!
11:57 Ultali ok so its because I didn't have an error handler defined... :S
12:10 * sri waves
12:11 sri http://mojolicio.us/perldoc/Mojoli​cious/Guides/Cookbook#Event_loops # not sure how to classify this example, EV support is kinda experimental, but there's no api associated with it
12:22 batman joined #mojo
12:25 jnap joined #mojo
12:42 daxim joined #mojo
13:13 CromeDome joined #mojo
13:36 sri marcus: which desktop twitter client are you using these days?
13:42 inokenty joined #mojo
13:43 sri the lack of timeline sync between the official apps is annoying
13:45 preflex_ joined #mojo
14:00 vel joined #mojo
14:01 mire_ joined #mojo
14:09 sri hmm
14:09 sri perhaps a new method in Mojolicious::Controller makes sense
14:10 sri $self->timeout(300)
14:12 sri i'm not particularly thrilled about it, but it's an option
14:20 inokenty1 joined #mojo
14:34 batman sri: will it set inactive timeout?
14:34 sri it would
14:35 sri or do nothing if the deployment environment has no equivalent
14:35 batman not sure if i like it. nut i'm also not too fond of my around dispatch hook which sets it on the $tx :)
14:36 sri adding methods to Mojolcious::Controller feels wrong anyway :/
14:37 batman i agree
14:44 kaare joined #mojo
14:54 pfig joined #mojo
15:18 Psyche^ joined #mojo
15:28 mire_ joined #mojo
15:46 Psyche^ joined #mojo
15:52 sri http://f.cl.ly/items/263H2l2m0f0q0​u1C081Z/sketch2_demo_uncropped.mov # sketch 2 is really pretty
15:53 vel joined #mojo
16:42 xaka joined #mojo
17:09 random left #mojo
17:20 tempire sri: don't stabilize the stream thing yet.
17:20 tempire give it another month, I'll take a look.  there's no rush.
17:21 marcus Sri: i use yorufukuru.it doesnt sync.
17:21 sri tempire: sure
17:21 tempire sketch2 looks a lot better
17:21 marcus I desperately want tweetbot for the mac.
17:21 tempire sketch1 definitely has some usability issues
17:22 * sri agrees with marcus
17:22 tempire I tried yorkufurofkru, but it's just not pretty enough
17:22 * marcus is in a car.
17:22 * tempire hasn't logged into twitter in almost a week as a result
17:22 sri marcus: i can tell by the pixels
17:22 marcus 3g pixels?
17:23 * marcus dreams of the new ipad.
17:23 sri cause i've seen a lot of 3gs in my time
17:23 tempire whoa.  css generation in sketch2
17:23 marcus Shiny
17:24 * sri dreams of the new thin macbook pro
17:24 * marcus dreams of it having a retina display.
17:27 * nuba dreams of a cymek kit
17:29 * tempire dreams of nordaaker action figures
17:29 nuba not now, still having fun. but later. much later.. :)
17:29 marcus Tempire: with big heads?
17:30 tempire hmm.  bobbleheads are a good option.  I prefer interchangeable though, 'cause bobbleheads are the first to be destroyed in an all out war with the lego people.
17:30 tempire interchangeable *heads
17:36 sri maybe we should also reconsider incresing the default timeout, to make changing it per connection less common
17:40 tempire it's 15 now?
17:40 sri and 20 for user agent
17:41 tempire the only requests I've heard were the people who wanted essentially unlimited so they didn't have to think about the process.
17:41 sri websockets and long-polling are common users of the timeout api
17:42 tempire that's what I did for moi as an easy fix.  Think I should think about it more and do some sort of keepalive thing.
17:44 tempire has anyone made an asset pipeline yet?
17:44 sri not that i'm aware of
17:45 * tempire reads http://guides.rubyonrails.org/asset_pipeline.html
17:46 * sri likes this one more https://github.com/rstacruz/sinatra-assetpack
17:48 sri this is the one behind the rails asset pipeline https://github.com/sstephenson/sprockets
17:49 tempire not sure what benefit an asset pipeline brings to images
17:49 sri embedding images in css
17:50 sri base64 encoded
17:50 sri for less roundtrips
17:50 tempire sinatra pipeline makes more sense.
17:53 sri reducing roundtrips is the most important function of an asset pipeline imo
17:54 tempire as in cacheing?
17:54 tempire I presume
17:54 sri caching only works after the first visit
17:54 tempire oh, many  many files.
17:54 sri yes, i mean http roundtrips
17:55 tempire yeah, that's what I'm doing here.  I'm going to have to implement something like this anyway, might as well make it a tool.
17:55 tempire maybe a mojo command to see the results.
17:55 sri inlining images can be huge
17:56 sri don't think i would make it a command
17:56 sri just a plugin that checks on startup
17:56 sri compares mtimes and stuff
17:56 tempire not as a primary thing, but as a transparency thing letting you know what's happening.
17:56 sri could do that with debug logging too
17:56 sri just showing changes
17:57 sri but i suppose there are many many options to make it fun :)
17:58 tempire hmm.  you know what would be really neat.  something like ./myapp routes that shows templates and assets used.
17:58 sri hmm
17:59 sri not sure about that
17:59 tempire would be killer in a large application.
18:00 sri well, you can just take the paths from app->renderer->paths and app->statis->paths and list files
18:00 tempire though I guess the debug log states all that.
18:01 * tempire notes for future reference
18:01 sri it would always be incomplete though
18:01 sri since you can't list all DATA files and templates
18:01 sri they might get lazy loaded later on
18:04 tempire have you used the sinatra asset thing?
18:05 tempire and does sinatra make the separation between templates and layouts just like mojolicious?
18:07 GitHub145 joined #mojo
18:07 GitHub145 [mojo] kraih pushed 1 new commit to master: http://git.io/SqT0lw
18:07 GitHub145 [mojo/master] static_class and template_class are also reserved stash values - Sebastian Riedel
18:07 GitHub145 left #mojo
18:08 sri tempire: kinda
18:08 sri tempire: http://www.sinatrarb.com/intro.html # all you need to know
18:09 pfig joined #mojo
18:10 pfig_ joined #mojo
18:12 sri which reminds me of something i wanted to add :)
18:14 sri support for the OPTIONS verb
18:29 GitHub56 joined #mojo
18:29 GitHub56 [mojo] kraih pushed 1 new commit to master: http://git.io/p3WEQA
18:29 GitHub56 [mojo/master] added OPTIONS support - Sebastian Riedel
18:29 GitHub56 left #mojo
18:33 sri objections? :)
18:37 Psyche^ joined #mojo
18:38 kanishka joined #mojo
18:41 sri wonder if there's a number higher than 50+ for retweets https://twitter.com/#!/oper​a/status/180422389371383808
18:48 sri no idea why people assume i'm mad, the guy managing the twitter account obviously didn't know what the failraptor is
18:50 inokenty joined #mojo
19:04 batman joined #mojo
19:06 Psyche^ joined #mojo
19:09 xaka sinatra has send_file! really want this stuff in mojo
19:21 tempire so my issue is that I only want the necessary assets loaded with different pages.  doesn't look like sinatra-assetpack handles that complexity.
19:25 tempire who's assuming  you're mad?
19:48 mire_ joined #mojo
19:52 sri some folks on twitter
19:52 sri hmm... no gsoc for perl this year :o
19:52 tempire not enough submissions?
19:53 sri no idea, just heard the perl foundation didn't get accepted
19:57 sri tempire: http://boingboing.net/2012/03/16/r​iaa-prez-twirls-mustache-in-a.html # so much for SOPA
19:59 sri i don't quite get what the ISPs gain from it
20:00 tempire I really need to run for office
20:01 tempire gotta make a lot of money first so I can fund my own campaign
20:01 sri THE FREE MARKET WILL SOLVE IT!
20:02 tempire it would if they would let it
20:02 sri once companies get that powerful can there be a free market anymore?
20:03 tempire the problem isn't the corporations, it's the govt+corporation collusion
20:03 tempire if the govt keeps their hands out of it, there's no benefit to corporate lobbying
20:03 tempire but they get involved and try to "fix" it, screwing up the whole balanced system
20:04 sri hmm
20:05 tempire "I believe in freedom except for the parts that I can't control"
20:05 tempire if people really understood what the word tolerance meant, maybe they wouldn't be so ridiculously foolish
20:06 tempire Tolerance: the amount of pressure something can withstand before it breaks.
20:06 tempire discomfort is not the enemy.  it's the teacher.
20:06 * tempire steps off the soapbox
20:07 * marty builds tempire a bigger soapbox
20:09 marty uhhhg.  perl + fedex = not optimal.  :(
20:09 marty I just want meez rates
20:12 d4rkie joined #mojo
20:16 preflex_ joined #mojo
20:18 augensalat Test::Mojo starts a webserver at a random port listening on all available IP addresses. Is that really necessary. Isn't localhost good enough?
20:19 * augensalat is asking on behalf of a paranoid admin.
20:19 rwstauner joined #mojo
20:20 sri to make us change that you would have to prove that it wouldn't cause portability problems
20:21 augensalat Current tests aren't sufficient?
20:21 sri maybe, that's for you to find out
20:22 sri i'm frightened by the idea of making such changes
20:22 sri based on past fights with cpantersters setups
20:22 augensalat I must admit, that I don't understand the test server completely. It seems to be started in Mojo::UserAgent - which doesn't feel like the right place for that.
20:24 d4rkie joined #mojo
20:26 sri that does not fill me with confidence
20:40 rwstauner joined #mojo
20:51 rwstauner joined #mojo
21:01 tempire A paranoid admin is running the test suite on a production system?
21:01 rwstauner joined #mojo
21:02 * tempire doesn't see why the binded ip would matter
21:05 d4rkie joined #mojo
21:10 sri tempire: would you take responsibility? (as in push a quick re-release if necessary)
21:11 augensalat tempire: No I'm running it on a development server, that is behind a firewall and not directly connected to internet ... paranoid, as I said.
21:11 sri tempire: also, on which interface would you listen?
21:12 tempire sri: no, I was saying I can't see how it would matter that it's binded to 0.0.0.0
21:12 sri oh
21:12 augensalat But this guy noticed processes opening random ports on the dev host.
21:13 tempire hackers!
21:13 purl it has been said that hackers is mostly about teenage punks fighting authority, AFAICT.
21:13 augensalat I wouldn't care either...
21:15 rwstauner joined #mojo
21:44 tempire audacity 2.0!
21:44 tempire unfortunately, STILL UGLY.
21:47 xaka when i do render_static, how do i know that file has been sent? any events?
21:47 vervain augensalat: I think that the admin can address his problem without changes to Mojo.
21:53 augensalat His point is, that he considers a software, that opens random ports as a potential danger. He is not completely wrong, even though it is unlikely, that someone will gain access to that machine through a running Mojo test server.
21:54 vervain And I take his point... it's intention is for development work... which needs relaxed standards.
21:55 vervain And it opens 'user' ports... which any one can do at any time.
21:55 vervain Unless the admin forbids it... which he could do.
21:55 augensalat And it is his job, to care about that.
21:55 vervain Or at least firewall them off.
21:56 vervain I take his concern... I just think it may be an OS / Sysadmin concern... not Mojo. :-)
21:56 vervain Like I say... he can stop it.
21:57 vervain Anyway... good luck... I'm off to fight the spring battle with my blackberry bramble.
21:59 augensalat The test server opens a random port between 10000 and 19999. It's possible, that a $someone has opened a port in that range for whatever reason (a test for ex.) and later forgot to close it.
22:00 vervain I doubt if Mojo tries to open an already used port.
22:00 augensalat But the FW port could be still open
22:00 augensalat shit happens...
22:01 sri no that can't happen
22:01 augensalat sure?
22:01 purl But are you sure you're sure?
22:01 sri if a port is in use it is in use
22:02 tempire pos o tive
22:02 vervain http://web.mit.edu/adorai/www​/seuss-technical-writing.html
22:04 augensalat you can open a port in a FW no matter if an application uses this port
22:04 sri augensalat: look, you're having the entirely wrong argument here, i don't care at all if we limit it to 127.0.0.1, all i want ia a guarantee that it doesn't kill our 100% green stats on cpantesters
22:05 augensalat I understand that. Will try to check this tomorrow.
22:05 vervain Maybe some kind of configuration item?
22:05 sri http://groups.google.com/group/mojolicio​us/browse_thread/thread/816b266038bef343 # lulz?
22:05 vervain So the user can choose to limit but default is to bind to all?
22:06 sri vervain: how would that make anything better?
22:06 sri it sounds like endless new possibilities for stuff to go wrong
22:07 vervain <shrug>  Just throwing crap at the walls.  I personally don't see that there is a problem to solve.
22:09 augensalat We all see it from the developer's POV. Out admin has other priorities. I think he's quite good in his job.
22:10 augensalat s/Out/Our/
22:26 sri hmm
22:27 sri one look at the port generator and i had proof that 127.0.0.1 is actually portable
22:37 sri in fact, not using 127.0.0.1 could now even be considered a bug
22:38 stephan48 in what way portable?
22:39 stephan48 to winschrott?
23:00 sri doh!
23:00 sri i actually found a much bigger flaw
23:04 sri it's the way we generate free ports
23:04 sri there's a short time when other test processes can generate the same port
23:07 sri and i don't think there's a fix
23:22 GitHub54 joined #mojo
23:22 GitHub54 [mojo] kraih pushed 1 new commit to master: http://git.io/taFfDw
23:22 GitHub54 [mojo/master] improved port range and address used for testing - Sebastian Riedel
23:22 GitHub54 left #mojo
23:23 sri not sure the collision flaw is fixable
23:35 brambles joined #mojo
23:54 GitHub24 joined #mojo
23:54 GitHub24 [mojo] kraih pushed 1 new commit to master: http://git.io/_oePpQ
23:54 GitHub24 [mojo/master] fixed typo in port range - Sebastian Riedel
23:54 GitHub24 left #mojo

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