Camelia, the Perl 6 bug

IRC log for #mojo, 2011-02-08

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

All times shown according to UTC.

Time Nick Message
00:10 GitHub72 joined #mojo
00:10 GitHub72 mojo: master Sebastian Riedel * 0801b43 (1 files in 1 dirs): added new rule - http://bit.ly/fLYVde
00:10 GitHub72 left #mojo
00:13 perlrocks Twitter: "Added a new rule to formalize requirements for the #mojolicious installation process. http://t.co/NbmPNFM #perl" --kraih http://twitter.com/kraih/status/34766670362841088
00:15 GitHub10 joined #mojo
00:15 GitHub10 mojo: master Sebastian Riedel * 665c066 (1 files in 1 dirs): fixed typo - http://bit.ly/gT4Uj3
00:15 GitHub10 left #mojo
01:23 GitHub164 joined #mojo
01:23 GitHub164 mojo: master Sebastian Riedel * 1515d2a (1 files in 1 dirs): cleanup - http://bit.ly/hADE66
01:23 GitHub164 left #mojo
01:32 MojoGuest381 joined #mojo
01:32 MojoGuest381 From: http://news.ycombinator.com/item?id=1277067 (199 hits)
01:33 MojoGuest381 left #mojo
01:38 perlrocks Twitter: "After the hook was decided to dispatch mojolicious to want something before the method executes the controller" (ja) --shiba_yu36 http://twitter.com/shiba_yu​36/status/34788140971462656
01:38 sri too bad, outcasts is awful
01:57 perlrocks Twitter: "I just installed Mojolicious (the Web in Box) with Perl Package Manager Index (PPM) from activestate's new PPM index  http://t.co/NCELEXE" --pythondj http://twitter.com/pythond​j/status/34792888135061504
02:10 metaperl1 left #mojo
02:13 sherr left #mojo
02:16 bellair1 joined #mojo
02:26 bellair1 left #mojo
02:32 esskar left #mojo
02:32 esskar joined #mojo
02:33 crab hi
02:36 tempire hooray for one minute!
02:36 tempire do I get credit for that?  I want an ascii representation of my avatar displayed for every mojolicious installed.
02:37 tempire :D
02:37 tempire that should cover it
02:38 tempire or B-)
02:38 crab what are you talking about?
02:38 tempire http://t.co/NbmPNFM
02:40 stephen left #mojo
02:41 MojoGuest987 joined #mojo
02:41 MojoGuest987 From: http://news.ycombinator.com/item?id=1277067 (201 hits)
02:41 MojoGuest477 joined #mojo
02:41 MojoGuest477 From: http://news.ycombinator.com/item?id=1277067 (201 hits)
02:41 MojoGuest987 left #mojo
02:42 MojoGuest294 joined #mojo
02:42 MojoGuest294 From: http://news.ycombinator.com/item?id=1277067 (202 hits)
02:43 MojoGuest477 left #mojo
02:44 MojoGuest294 left #mojo
02:46 MojoGuest926 joined #mojo
02:46 MojoGuest926 From: http://news.ycombinator.com/item?id=1277067 (203 hits)
02:47 MojoGuest926 left #mojo
03:10 crab my README.foo naming scheme doesn't work very well with github
03:16 ryan left #mojo
03:35 crab how can i set ->secure(1) on my session cookie?
03:52 perlrocks Twitter: "Mojolicious:: Lite to use SpeedyCGI runs in a separate file if the template. __DATA__ Be the first to write error when running the template below. __DATA__ Not the first time the following loaded?" (ja) --clicktx http://twitter.com/clicktx​/status/34821844439670785
04:04 tholen42 joined #mojo
04:06 tl left #mojo
04:42 metaperl joined #mojo
05:13 perlrocks Twitter: "MojoX::Auth::Simple 0.04.03: Perl extension for login authentication for Mojolicious http://bit.ly/hBSIgk" --lordnaastik http://twitter.com/lordnaast​ik/status/34842299691180032
05:13 perlrocks Twitter: "Mojolicious::Plugin::TweetButton 0.0003: TweetButton Helper Plugin http://bit.ly/e2sX0Z" --lordnaastik http://twitter.com/lordnaast​ik/status/34842300513255424
05:29 dj joined #mojo
05:44 dj left #mojo
05:47 perlrocks Twitter: "Mojolicious::Plugin::TweetButton 0.0003: TweetButton Helper Plugin http://bit.ly/eRgp9D" --J03_PY http://twitter.com/J03_PY/status/34850783723786241
05:47 perlrocks Twitter: "► Mojolicious::Plugin::TweetButton 0.0003: TweetButton Helper Plugin http://bit.ly/eRgp9D linux" --YNK33 http://twitter.com/YNK33/status/34850783837032449
05:51 sherr joined #mojo
05:55 perlrocks Twitter: "AppandExt: Mojolicious:: Plugin:: TweetButton 0.0003 http://goo.gl/fb/mSQRI" (ja) --S64Pontal http://twitter.com/S64Ponta​l/status/34852923175342080
06:21 kaare joined #mojo
06:21 tholen joined #mojo
06:24 ysyrota joined #mojo
06:29 ltriant left #mojo
06:34 Foxcool joined #mojo
06:37 tholen left #mojo
07:19 koban joined #mojo
07:24 su-bzero sri: $cache->set(foo => 'bar', expires => '+60m') is available?
07:26 rupertthehacker joined #mojo
07:33 Christian joined #mojo
07:34 Christian morning
07:43 crab s-bzero: what are you talking about?
07:45 perlrocks Twitter: "url_for helper returned with Mojo:: URL is because each clone and disposable or not enter. URL query string containing Mojolicious generate - Getting Started with Perl code samples http://bit.ly/gJe8wX" (ja) --5_1_5 http://twitter.com/5_1_5/status/34880376891510784
07:46 marcus hmm
07:46 marcus again, .ja PUZZLEMENT
07:47 marcus do you really need to use clone to change the query parameters on the Mojo::URL object returned by url_for helper?
07:48 su-bzero crab: About expiration for cache.
07:48 su-bzero crab: For example, I want to cache some value for an hour.
07:48 su-bzero Not forever.
07:48 su-bzero So I need to control time.
07:48 marcus oooh. time traveler
07:48 su-bzero It's as a wish for adding expires option for ->set
07:49 su-bzero ;)) Time-machine needed.
07:49 marcus su-bzero: or Hiro Nakamura
07:49 su-bzero Who is it? ;)
07:50 marcus http://heroeswiki.com/Hiro_Nakamura
07:51 crab su-bzero: where is this cache? i didn't know mojo had a cache
07:52 su-bzero crab: https://github.com/kraih/mojo/commit/60​5cd25d83864076ef2cf3c6d9a12b05f80cb675
07:52 marcus that is a route cache
07:52 marcus why would you ever need to expire it?
07:53 marcus oh, I see. su-bzero wants to subvert it for his own nefarious purposes
07:53 crab ah, very new.
07:53 perlrocks Twitter: "1.02 Mojolicious Does not yet ended. One or ー Mojo:: Client implemented like lazy loading." (ja) --5_1_5 http://twitter.com/5_1_5/status/34882500924809216
07:54 crab does anyone use, say, SMS-based confirmation for certain actions? e.g. having to enter a one-time password sent by sms before you can change your password or something.
07:54 su-bzero marcus: Yes. I will use it as key-value database. ;)))
07:54 breaker313 joined #mojo
07:55 su-bzero crab: Need a provider?
07:55 crab no, i have a provider. i'm thinking more about the semantics of the operation.
07:56 marcus OH, PEOPLE OF NIPPON-KOKU, HOW I WISH I COULD COMMUNICATE WITH YOU.
07:56 su-bzero marcus: Why do we need a cache for routes?
07:56 crab suppose i have a bridged setup like this: $confirmed = $r->bridge->to('auth#has_token'); $confirmed->route('/foo')->via​('post')->to('blah#operation')
07:56 marcus su-bzero: to save cpu cycles. it is green.
07:57 marcus su-bzero: sri says it saves about 100rps for the 11th route defined in your system.
07:57 crab should the bridge allow the request if a valid token is submitted? or should e.g. i store the POST parameters on the server and ask for the token to be submitted, and then <do some magic to recreate the originally-saved request>?
07:58 su-bzero ;)) I thought that it could be used for caching some data in Mojo without Redis or MongoDB. ;)
07:58 crab in other words, does the token authenticate the user or does it authorize a request?
07:58 marcus damnit, you guys have totally distracted me from coding on Net-ActiveCollab
07:58 crab what is that?
07:58 AmeliePoulain joined #mojo
07:58 su-bzero Sorry. Need to go too. ;) See you later. But write something about Mojo::Cache... I will read later. ;)
07:59 crab activecollab looks scary
08:00 marcus crab: we use it for project management and time reporting
08:01 marcus so I need a perl module so I can integrate it into our IRC bot :-p
08:01 charsbr marcus: do you need translation from or to japanese? :)
08:01 crab or, in other-other words, do i need the token to say "let the guy who holds this token POST to blah#operation" or do i need the token to say "process the earlier-submitted request"?
08:01 crab hmm. when i put it like that, i'm afraid (1) is by far the easier thing to implement.
08:01 marcus charsbr: sometimes google translate leaves me unsatisfied. What does http://twitter.com/5_1_5/status/34882500924809216 mean for instance? :)
08:02 dj joined #mojo
08:03 crab /msg charsbr now's your chance to give him an even more incomprehensible explanation!
08:03 crab dammit
08:04 marcus charsbr: it's nice to see so many .jp people tweet about mojolicious tho.
08:04 charsbr marcus: apparently I need to wait more for Mojolicious 1.02. and it looks like delayed load is implemented in Mojo::Client
08:04 * marcus feeds crab to the alligators
08:04 crab is sri a cult figure in japan? do people buy little action figures of him?
08:04 marcus oh wait, sri ate those :/ we need some new creatures from the crypt.
08:05 marcus crab: you mean like cthulhu?
08:05 marcus or godzilla..
08:05 purl somebody said godzilla was la.znet.com/~callgirl or http://www.winternet.com/~mikelr/flame67.html or a giant lizard who enjoys destroying Tokyo
08:06 marcus charsbr: thanks for the translation :) Now I just have to figure out what delayed load means.
08:06 marcus charsbr: he might mean async loading, which is already supported in mojo.
08:07 marcus oh, maybe he's talking about the on_start stuff.
08:08 dj left #mojo
08:08 koban left #mojo
08:08 charsbr maybe he's talking about a mail he sent last month "[Mojolicious] Speed-up of Mojolicious/Mojo on CGI"
08:11 crab i begin to get a horrible sinking feeling that there's no wasy way to set ->secure(1) on my session cookies
08:12 crab s/wa/ea/
08:12 su-bzero1 joined #mojo
08:12 btn left #mojo
08:12 btn joined #mojo
08:12 charsbr crab: sri is not a cult figure yet in japan, but several people including me promoted mojolicious, and I'll give a talk later this month, so several people started learning
08:12 btn left #mojo
08:13 charsbr but if you create a cool figure (or doll), several otakus would definitely buy it :)
08:14 marcus charsbr: ah. lazy loading of ->client to speed up cgi. Now it all makes sense. Thanks :)
08:15 charsbr :)
08:16 su-bzero_ left #mojo
08:16 marcus https://github.com/kraih/mojo/commit/25​6327da4ee6d01bf0948d5c232c4703c24a8039
08:18 charsbr nod
08:18 marcus sri++ # Bending under pressure
08:18 marcus =)
08:19 marcus (it is better than breaking)
08:21 metaperl left #mojo
08:24 metaperl joined #mojo
08:28 extreme joined #mojo
08:33 otaviof joined #mojo
08:43 GabrielVieira joined #mojo
08:43 metaperl left #mojo
08:48 GabrielVieira2 left #mojo
08:53 marcus hmm
08:53 marcus having some trouble with my proxy plugin on trunk
08:53 marcus even tho it calls ->rendered
08:55 marcus oh wait. I need render_later, don't I..
08:55 btn joined #mojo
08:57 perlrocks Twitter: "Cool! RT: @kraih: After reviewing the test suite i've decided to try lazy loading Mojo::Client, this can speed up #mojolicious cgi apps 30%." --sharifulin http://twitter.com/shariful​in/status/34898527276634112
09:08 breaker313_ joined #mojo
09:09 breaker313 left #mojo
09:16 Foxcool left #mojo
09:29 metaperl joined #mojo
09:36 marcus sri: shouldn't Waiting for delayed response, forgot to render or resume? go away if I call render_later?
09:37 spleenjack joined #mojo
10:05 koban joined #mojo
10:31 breaker313 joined #mojo
10:32 breaker313_ left #mojo
10:55 crab hm.
10:55 crab in the general case, there is no reliable way to turn a POST request body into an html form which, when submitted, will produce substantially the same request, right?
10:58 robn joined #mojo
10:59 robn left #mojo
11:25 fmerges joined #mojo
11:25 metaperl left #mojo
11:38 perlrocks Twitter: "Mojolicious::LiteとHTML::F​illInForm::Liteでフォームのフィルイン - たつみの殴り書き http://htn.to/fg5vgk" --__gfx__ http://twitter.com/__gfx__​/status/34939046157361152
12:04 crab damn and blast. i can't find any good way to set my session cookie to be https-only.
12:06 crab oh. maybe create SecureSessions inheriting from Sessions and override sessions in my app, then say sub store { my $c = $self->SUPER::store(@_); $c->secure(1); } or something
12:14 perlrocks Twitter: "Mojolicious:: Lite and HTML:: FillInForm:: Lite fill in a form - scrawled in http://bit.ly/hh1fr1 Tatsumi" (ja) --wPerlnews http://twitter.com/wPerlnew​s/status/34948072794492928
12:27 crab wtf.
12:50 marcus mm
12:59 crab well, it's a nice theory, but it doesn't actually seem to work.
13:01 crab after i do $c->signed_cookie(....), should $c->res->cookie('blah')->secure(1) work?
13:08 sri good morning sunshines
13:09 sri marcus: you do not have to clone a url to change the query
13:09 crab ah, it works, but it doesn't change the response because the first cookie has already been ->headers->add'ed
13:10 stephan48 good morning sri
13:10 crab it feels very wrong to have to remove a Set-Cookie from the headers and then add the edited cookie back into res->cookies.
13:12 marcus sri: thought as much. I would assume url_for returned a fresh url object every time
13:13 sri hmm, a picture of godzilla associated with my name seems appropriate
13:13 crab sri: any advice? (i want to make my session cookie https-only)
13:14 sri crab: don't think you can yet
13:15 sri marcus: message will never go away
13:15 purl Message for will stored.
13:15 sri ops
13:16 sri message marcus DUDE!
13:16 purl Message for marcus stored.
13:16 marcus message sri SWEET!
13:16 purl Message for sri stored.
13:17 sri message marcus DUDE
13:17 purl Message for marcus stored.
13:17 marcus message sri SWEET
13:17 purl Message for sri stored.
13:17 * sri pokes purl
13:17 * purl pees
13:17 sri Oo
13:17 marcus purl: stupid bot.
13:17 purl stoopid hyooman
13:18 crab sri: so what i've done is to override Mojolicious::Sessions::store in a derived module and call $self->SUPER::store(@_); i guess since i'm going that far, i might as well also res->headers->remove("Set-Cookie"); and re-add the cookie.
13:19 crab that would be very bad if there were multiple cookies, but i don't have those.
13:19 sri i guess secure as a core feature would be sensible
13:21 crab yes. very much so.
13:21 crab if you can give me some idea of how you want it, i'm willing to send a patch, two-space indent and all.
13:22 crab (of course, i personally think we should just set secure on the session cookie unconditionally, but i'm sure that will annoy people and break things.)
13:22 crab leaking authenticators over plaintext is bad karma.
13:25 Foxcool joined #mojo
13:54 GitHub125 joined #mojo
13:54 GitHub125 mojo: master Sebastian Riedel * 30c8e6b (3 files in 3 dirs): improved css3 support in Mojo::DOM - http://bit.ly/hqh0Ma
13:54 GitHub125 left #mojo
13:58 fhelmber_ joined #mojo
13:59 sri crab: breaking all apps out there is not an option
13:59 sri i guess app->session->secure(1) would work
13:59 sri or actually app->sessions->secure(1)
14:03 rshhh left #mojo
14:04 crab just has secure => 0; in Sessions.pm plus a line or two in store? works for me, coming up.
14:08 rshhh joined #mojo
14:22 su-bzero is now known as su-bzero[A]
14:24 su-bzero1 left #mojo
14:30 Foxcool left #mojo
14:31 breaker313 left #mojo
14:31 breaker313_ joined #mojo
14:32 su-bzero joined #mojo
14:38 su-bzero[A] is now known as su-bzero`
14:39 crab sri: i sent a pull request. i've never done that before, so i hope i didn't screw it up somehow.
14:41 Foxcool joined #mojo
14:49 sri crab: test missing
14:50 MojoGuest39 joined #mojo
14:50 MojoGuest39 From: http://www.networkedblogs.com/33Knp (59 hits)
14:50 MojoGuest39 left #mojo
14:50 sri crab: you might also have to add secure support to the cookiejar for a test case
14:52 kaare_ joined #mojo
14:54 crab if i add "proper" support to the cookiejar, then all my tests will break again. oh well, too bad.
14:54 sri ?
14:54 crab can you give me a brief sketch of where/how you want the test?
14:56 sri one of the tls tests i guess
14:56 crab sri: i'll just have to enable https for testing. (i don't know how, but there must be a way. will look later.)
14:56 sri and of course the cookiejar test first
14:58 sri https://github.com/kraih/mojo/blob/master/t​/mojolicious/websocket_tls_proxy_lite_app.t # this one is a bit over the top, but you get the idea
14:59 sri https://github.com/kraih/mojo/​blob/master/t/mojo/cookiejar.t # this first though
15:00 sri su-bzero: i have no plans to add expiration to the cache
15:01 sri the cache helper does expiration though
15:01 sri or actually the memorize helper
15:01 su-bzero ;(
15:02 sri it's an in memory cache, you can do it in like 10 lines of code yourself
15:03 sri http://mojolicio.us/perldoc?Mojolic​ious/Plugin/DefaultHelpers#memorize
15:04 AmeliePoulain left #mojo
15:07 alnewkirk left #mojo
15:08 damog joined #mojo
15:16 sri interesting, couchone and membase merged :o
15:17 sri "Technologically, we’ll be joining the products together to create a high volume, low latency, elastic clustered Couchbase server system. A Couch that’s Fast, Simple and Elastic."
15:18 Foxcool left #mojo
15:18 crab sri: adding support to cookiejar basically means that ->find('http://blah') will not find a cookie that is marked secure but otherwise matches, right?
15:19 crab or is there more to it?
15:19 sri right
15:19 sri you better check the specs though
15:19 sri in the end you'll be the one to blame if something goes wrong ;p
15:20 crab 'When it sends a "secure" cookie back to a server, the user agent SHOULD use no less than the same level of security as was used when it received the cookie from the server.'
15:21 crab what a vague wording.
15:21 crab If absent, the user agent MAY send the cookie over an
15:21 crab insecure channel.
15:21 sri welcome to my world!
15:22 crab sri: as far as i can see, the change i described above is sensible. i'll implement that.
15:27 otaviof left #mojo
15:30 koban left #mojo
15:32 mattastrophe left #mojo
15:37 crab sri: if you have a bunch of assignments with the '='s aligned and i add a new variable that's longer than the old ones, do you want me to reindent the whole thing?
15:37 sri ?
15:37 crab my $name   = $cookie->name;
15:37 crab my $secure = $cookie->secure || 0;
15:37 crab my $httponly = $cookie->httponly || 0;
15:37 crab was kicked by sri: crab
15:38 tholen joined #mojo
15:42 Shaeto joined #mojo
15:44 Shaeto hi guys, who can me help me to run fastnotes demo. latest 1.02 mojolicious, latest fastnotes from git, "Tue Feb  8 18:43:32 2011 debug Mojolicious::Controller:481 [4115]: Resource "->create_form" not found."
15:45 Shaeto thats all i did manage to get
15:45 sri Shaeto: did you contact the author?
15:45 Shaeto not yet, mb it is very stupid configuration issue
15:46 sri i doubt many people here know the app, best to contact the author first
15:47 Shaeto left #mojo
15:47 Shaeto_ joined #mojo
15:47 Shaeto_ sorry disconnected
15:47 Shaeto_ is now known as Shaeto
15:47 Shaeto so, you suggest to pm mojo author ?
15:48 sri i am the mojo author ;p
15:48 sri i suggest you contact the author of the application
15:48 Shaeto wow :) glad to see you
15:51 Christian nive evening to you all
15:51 sri i can see the problem though, it is using a bug
15:52 sri the /signup route is invalid
15:52 sri the users controller needs a create_form sub
15:53 sri it worked with 1.01, but was a bug
15:54 Christian left #mojo
15:54 Shaeto and auths controller need create_form sub also ?
15:55 Shaeto or only users
15:55 vel joined #mojo
15:55 sri auths too it seems
15:56 Shaeto understand, i investigated Controller.pm and Routes.pm and now understand why google now lies about create_form automatic behaviour :) at least in 1.02
15:56 sri where does google lie?
15:57 Shaeto in some blogs about example application
15:57 sri it has never been documented or unit tested :(
15:57 sri argh
15:58 sri thats really bad
15:59 yko that behavior was useed too much in many applications and examples
15:59 sri the problem is malicious users can exploit it
16:00 sri because of placeholder routes
16:00 yko i understand
16:00 yko i'd wrote 'almost all' even, but that may be my personal feeling. That just explains why so much noize arround that
16:00 sri no matter what i do here it's wrong :(
16:00 sri loss/loss situation
16:01 sri i can keep broken examples working or fix a security issue
16:01 yko ye, sux, just nevermind or create little blog note about that...
16:01 Shaeto agree, small note will help us :)
16:02 sri can't do that
16:02 sri it will get used against me
16:02 yko sri: at the moment there's no way to rollback: all active users already upgraded, so i'ts better to keep security fix
16:02 sri i'm tempted to keep the security issue actually
16:02 yko yes, it will :) it will be used against you anyway :p
16:03 rhaen HEEEELP! I can't see a solution! What is the best way to display an image inside a template?
16:03 marcus left #mojo
16:03 yko Oo
16:03 rhaen I have a model which generates a GIF image. How do I pass it to the template?
16:03 sri breaking backwards compatibility is the worst case atm marketing wise
16:03 rhaen in a Mojolicious::Lite app?
16:03 yko ooh.. :)
16:04 yko rhaen: google for html embedded images
16:05 yko like b($image)->base64_encode or so
16:05 rhaen yeah. ok - got it thanks
16:06 rhaen I was thinking in that direction, however - I thought there might be an easier way
16:06 Shaeto well, stub create_form sub helps and example works now
16:07 yko rhaen: just generate url-key or cookie and mforce browser to make one more request is much easy way
16:11 sri damn, i don't know what to do :(
16:13 yko I think part of problem is that most users still are using Mojolicious from master, not from github
16:13 sri how is that related?
16:13 sri the problem is people made blog posts about something that is not a feature and used it in example apps
16:15 crab joined #mojo
16:15 yko You commit => users (all) upgrade Mojolicious and (sometime) gets problems (even in production, yes)
16:15 sri you can possibly crash small servers with this
16:15 sri yko: that's not the issue here
16:16 sri crab: no pasting into the channel please
16:16 yko And than evil maintainers of some applications came here and argue and rant in their blogs :)
16:17 sri crab: there's a perltidyrc included in the mojolicious distribution you can use
16:24 marcus joined #mojo
16:25 sri well, i've reported the bug in the application https://github.com/koorchik/F​astNotes-Proto/issues/issue/1
16:26 yko sri++
16:26 sri which blog posts mentioned it?
16:26 sri maybe i can add a comment there
16:29 Shaeto btw i tested ALL example applications today - almost all use this "behaviour" and should be changed
16:30 sri oh my
16:30 spleenjack left #mojo
16:31 mattastrophe joined #mojo
16:31 spleenjack joined #mojo
16:32 sri Shaeto: got links?
16:33 Shaeto yes ec
16:33 Shaeto sec
16:34 breaker313_ left #mojo
16:41 Shaeto sorry other seems Lite applications with own issues, for example https://github.com/diegok/PhotoBooth , http://github.com/shoorick/mojowka, http://github.com/vti/bootylicious . didn't work also but i am not sure that they affected by "unable to find template" problem. i'll try to investigate and report here
16:41 sri lite apps are not affected
16:42 sri just full mojolicious apps
16:43 marcus left #mojo
16:44 Shaeto understand, okay ill update bug trackers for Lite applications, they SHOULD work to get new people :)
16:45 sri that will be good
16:53 jfuller joined #mojo
16:56 sri ooooh
16:56 sri there might be a workaround
17:03 GitHub159 joined #mojo
17:03 GitHub159 mojo: master Sebastian Riedel * 4877dbf (3 files in 3 dirs): made action-less templates work again for normal Mojolicious applications - http://bit.ly/fkQ22k
17:03 GitHub159 left #mojo
17:03 jwang joined #mojo
17:03 sri ok, please test this
17:04 sri it might have horrible side effects
17:06 ka2u_ left #mojo
17:06 crab should i (copy websocket_tls_proxy_lite_app.t and) create tls_lite_app.t and add my test there?
17:07 sri you can add it to the websocket test
17:07 sri should be small enough
17:07 sri or actually, just add a new test
17:08 crab new test in the same file, you mean? ok.
17:08 sri nono
17:08 sri just add a new file
17:08 crab ok. tls_lite_app.t?
17:08 sri ye
17:09 * sri pokes yko
17:09 * sri pokes Shaeto
17:10 * yko pokes sri back
17:10 sri the last commit should fix the security isse and keep action-less templates working
17:11 Shaeto sec i'll test
17:11 marcus joined #mojo
17:12 sri marcus looks a bit unstable today
17:12 marcus yeah, setting up my drobo
17:13 marcus needed some reboots of the mini that's hosting my znc
17:13 sri that's a storage thingy isn't it?
17:13 marcus yeah
17:13 marcus that you can stuff any kind of drives into
17:14 marcus takes up to 4 drives. uses the biggest one for parity
17:14 crab sigh. i suppose i'll have to find a new quote.
17:15 sri you do
17:15 Shaeto i can confirm fastnotes works with latest changes. not sure about "horrible side effects" :)
17:15 sri and the patch will be rejected until it is a really good one!
17:15 sri Shaeto: \o/
17:16 sri might eat your offspring or so
17:16 marcus awesome side effects
17:16 yko ill take two
17:17 sri might multiply your offspring too
17:17 extreme left #mojo
17:17 yko i still don't have children only because i'm not sure ill get two for price as for one
17:17 marcus cheaper than forking
17:18 yko so such side effect is ok for me :D
17:18 ysyrota left #mojo
17:19 crab can lite apps have a sub startup?
17:19 sri no
17:20 sri they *are* the sub start
17:20 sri *+up
17:20 yko crab: just write the code :)
17:22 vel left #mojo
17:23 crab is $client->get("https://localhost:$port/")->success->body,
17:24 crab what is that "success"?
17:24 crab ah, it returns the response. ok.
17:24 yko $tx->res || undef
17:27 alnewkirk joined #mojo
17:32 minimalist joined #mojo
17:34 Shaeto left #mojo
17:35 fmerges left #mojo
17:36 tholen left #mojo
17:41 crab sri: ok, done
17:47 sri crab: you forgot to squash
17:48 crab oh, you want them squashed? ok.
17:51 sri crab: your use of httponly is wrong too
17:51 sri httponly has no meaning for our cookiejar
17:53 sri in browsers it prevents cookies from being accessible via javascript
17:53 crab oh, damn.
17:54 sri squash in the fix too :)
17:55 marcus can you squash once you've pushed to github?
17:55 marcus or do you have to delete your fork and push again?
17:55 sri i think you can
17:56 sri since it doesn't really change the underlying structure
17:57 marcus http://www.gitready.com/advanced/2009/0​2/10/squashing-commits-with-rebase.html tutorial says it's bad if you've shared with others
17:57 marcus but I suppose it might work
17:57 sri right
17:57 sri it's not nice, but should work
17:57 sri best to just use branches all the time
17:59 crab marcus: i can push -f
17:59 marcus crab: good times
17:59 purl DON'T OPEN IT!  IT'LL CRASH AOL!
18:00 crab sri: resubmitted
18:04 sri looks good
18:10 ilbot2 joined #mojo
18:12 crab so how can i use Test::Mojo and test a http+https app?
18:12 sri you can't atm i guess
18:12 crab because now my whole test suite is br... oh, great.
18:13 yko sri: because of random port generation?
18:13 sri you've opened a huge can of worms, https is even an optional feature
18:14 marcus pandora's https box
18:14 crab if i make it all-https can i do it then?
18:14 sri you can't
18:15 sri Test::Mojo does not know anything about https yet
18:15 sri you would have to make the undelying Mojo::Client code more configurable
18:16 minimalist left #mojo
18:16 sri that goes quite deep though
18:16 crab so i went all the way up shit creek and threw away my own paddle.
18:17 sri pretty much
18:17 crab any chance you'll let me back out the cookiejar changes (except for the bugfix)?
18:18 crab they're somewhat superfluous. i didn't need them to test that ->secure(1) works
18:19 sri too late, already merging
18:20 crab i could submit a separate revert with a comment, then the code would stay around when https/mixed testing is possible...
18:21 sri no going back
18:22 fhelmber_ left #mojo
18:23 crab i guess i can throw away my test suite.
18:24 tholen joined #mojo
18:25 fhelmber_ joined #mojo
18:27 * crab sighs
18:29 sri maybe i can save you, no promises though
18:30 fish__ wtf? syntax error at /usr/local/share/perl/5.10.1/M​ojolicious/Command/Fastcgi.pm line 6, near "has description"
18:31 fish__ all tests are fine
18:32 fhelmber_ left #mojo
18:32 sri there is no syntax error
18:32 sri your installation is broken again
18:32 fish__ sri: well, i just updated it to make sure it is recent
18:34 fish__ sri: http://perlpunks.de/paste/show/4d518c9e.1ed0.3 <- those are the first lines of Fastcgi.pm. looks quite ok i guess. maybe somehow 'has' doesnt get exported?
18:35 sri no point discussing, your installation is broken
18:35 sri the github version does work
18:36 fish__ the tests work too here. maybe there is something left from a older version..
18:39 fish__ just removed /usr/local/share/perl/5.10.1/Mojo* and reinstalled it, this didn't help
18:52 sri crab: yes, looks like i'm saving your tests
18:57 marcus fish__: for some reason, you are using an older Mojo::Base
18:58 marcus perldoc -l Mojo::Base
19:00 sri grr, github bot dead again
19:00 GitHub96 joined #mojo
19:00 GitHub96 mojo: master Abhijit Menon-Sen * 8bb9c2d (3 files in 3 dirs): Allow app->sessions->secure(1) to secure session cookies ...
19:00 GitHub96 mojo: master Sebastian Riedel * 1f7bca4 (7 files in 6 dirs): added experimental support for HTTPS testing
19:00 GitHub96 mojo: master commits 4877dbf...1f7bca4 - http://bit.ly/g1UtwK
19:00 GitHub96 left #mojo
19:00 sri oh
19:00 fish__ oh, i guess its the 'bundled' mojo of vti's bootylicious
19:00 sri just slacking
19:00 fish__ marcus: thanks for the hin
19:00 fish__ t
19:00 sri crab: that should save your day
19:00 marcus fish__: you are welcome
19:01 sri just see the new tls_lite_app.t
19:01 sri $t->client->test_server('https');
19:01 sri that will make test use https internally for now
19:01 sri *+s
19:02 spleenjack left #mojo
19:03 tholen left #mojo
19:04 crab thanks. i was just in the middle of doing something similar in a module derived from Mojo::Client.
19:04 crab and wondering why it wasn't working, which was because i misunderstood when test_server was called.
19:05 sri it's very much experimental though
19:05 sri but i guess you won't mind
19:05 crab indeed.
19:10 crab Use of uninitialized value $address in concatenation (.) or string at /usr/local/share/perl/5.10.1/Mojo/Client.pm line 602.
19:11 crab should be localhost, but where did the value come from before? /me reads diff
19:18 crab in tls_lite_app.t, _tx_info shows $req->url as https://localhost:14473/login?name=sri, but in my tests i just get /nonesuch
19:19 marcus http://showmetheco.de/articles/2011/2/divin​g-into-html5-with-websockets-and-perl.html hmm?
19:19 marcus I thought most browsers had disabled web sockets now
19:19 sri they do
19:19 sri firefox and opera at least
19:20 marcus nice comment on vti's post :D
19:20 sri chrome is experimenting and safari doesn't care
19:21 sri hehe
19:21 crab maybe i need to set the base somewhere.
19:22 sri years might be a bit much, "a year" might be closer
19:22 sri sri-- # overusing words
19:25 sri mainstream is also a bad choice, i doubt they ever will go mainstream
19:25 purl okay, sri.
19:25 sri heh
19:25 * sri pats purl
19:25 * purl stabs
19:25 sri :/
19:27 dotan_ joined #mojo
19:31 crab aha. $self->app is unset when tx_start is called.
19:32 stephen joined #mojo
19:38 crab looks like i have to do: $client = Mojo::Client->new(app => "Wigeon"); $t = Test::Mojo->new(app => "Wigeon", client => $client); $client->test_server('https') and then it works.
19:42 MojoGuest320 joined #mojo
19:42 MojoGuest320 From: http://www.google.de/search?hl=de&amp;cli​ent=safari&amp;rls=en&amp;q=ExtJs+Mojolic​ious&amp;aq=f&amp;aqi=&amp;aql=&amp;oq= (1 hits)
19:42 MojoGuest320 left #mojo
20:02 GitHub77 joined #mojo
20:02 GitHub77 mojo: master Sebastian Riedel * 5bf01c1 (2 files in 2 dirs): new code name :) - http://bit.ly/i0MvCN
20:02 GitHub77 left #mojo
20:02 sri :D
20:04 perlrocks Twitter: "We should have a #mojolicious release with a new code name later this week. :) http://t.co/zvh6H6V #perl" --kraih http://twitter.com/kraih/status/35066485722451968
20:06 crab Eeeexcellent.
20:07 metaperl joined #mojo
20:07 sri http://www.fileformat.info/inf​o/unicode/char/1f3e9/index.htm # hahahahaha
20:07 * sri pokes marcus
20:22 otaviof joined #mojo
20:38 vel joined #mojo
20:40 perlrocks Twitter: "Smiling Cat Face With Heart-Shaped Eyes "@kraih: We should have a #mojolicious release with a new code name later this week. :) #perl"" --mshershnev http://twitter.com/mshershn​ev/status/35075588167311361
20:49 perlrocks Twitter: "“@kraih: @miyagawa Haha, U+1F3E9 totally!” - Mojolicious Love Hotel is going to be great." --marcusramberg http://twitter.com/marcusramb​erg/status/35077778550292480
21:04 ltriant joined #mojo
21:08 dekimsey sri, I've noticed that stderr seems to be disappearing into a black hole when my app is run in fastcgi mode... unfortunately, thats about all I know :/
21:08 sri that's true
21:09 sri forwarding stderr over fastcgi is rather annoying, but patches are welcome
21:09 dekimsey I looked into FCGI to see how they addressed it.... and it made me cry. They used a tied filehandle if I remember correctly
21:09 sri lol
21:10 sri FCGI has that effect
21:11 dekimsey I'll probably look some more, but I don't see my self getting anywhere there. I think at this point just updating the documentation to point out the known issue would be sufficient. Do you have any specific place that should go?
21:11 sri acually i don't consider it an issue
21:11 dekimsey why is that?
21:12 sri none of our deployment environments forward stderr anywhere
21:12 sri if anything it should go to the mojo log file
21:13 dekimsey Hrm, perhaps its another artifact of the environment we have here...
21:13 jfuller sri: Where do you log startup errors?
21:13 dekimsey what happens when you have 'warn'
21:14 dekimsey in your supporting libraries?
21:14 sri depends where stderr goes
21:14 sri atm you are responsible for it
21:16 ltriant left #mojo
21:16 ltriant joined #mojo
21:17 sri i wouldn't mind a discussion about better solutions
21:18 sri possibly on the mailing list
21:18 sri i guess forwarding stderr to the log file by default would be more "out of the box"
21:19 dekimsey I'm fine with it not, I guess it was simply that the stderr was simply disappearing as a result of fastcgi
21:20 dekimsey That's what caused the problem for me. Fixing that one... hell if I know
21:20 sri well, forwarding stderr via fastcgi is a horrible idea anyway
21:21 dekimsey I ended up with a hack for sigwarn that redirects the output to the logger. But to get the logger's caller information to look correct required a little munging.
21:21 dekimsey $SIG{__WARN__} = sub {
21:21 dekimsey unshift @_, $self->app->log;
21:21 dekimsey goto $self->app->log->can('warn');
21:21 dekimsey }
21:21 dekimsey was kicked by sri: dekimsey
21:21 dekimsey joined #mojo
21:21 sri hmm, need to work on the script, still a bit slow with the kicking
21:22 dekimsey lol
21:22 sri dekimsey: no pasting please
21:22 dekimsey aye, sorry about that. I'm not terribly familiar with IRC.
21:23 dekimsey Anyway, setting the sig warn handle deals with the problem. Not sure if its what one would consider The Right Way.
21:24 sri yea
21:24 sri could be useful to have the recipe on the wiki though
21:24 sri *hint hint*
21:24 dekimsey will do.
21:24 sri \o/
21:28 dotan_ left #mojo
22:23 jfuller left #mojo
22:32 vel left #mojo
22:37 rupertthehacker left #mojo
22:40 kaare_ left #mojo
22:49 damog left #mojo
22:49 Akron joined #mojo
22:51 rupertthehacker joined #mojo
22:51 rupertthehacker left #mojo
22:59 auggy joined #mojo
23:01 ltriant left #mojo
23:02 ltriant joined #mojo
23:05 GabrielVieira left #mojo
23:06 GabrielVieira joined #mojo
23:07 MojoGuest223 joined #mojo
23:08 MojoGuest223 left #mojo
23:08 mattastrophe left #mojo
23:08 MojoGuest853 joined #mojo
23:09 MojoGuest853 est
23:10 MojoGuest853 left #mojo
23:48 metaperl left #mojo
23:49 mattastrophe joined #mojo

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