The web in a box - a next generation web framework for the Perl programming language

IRC log for #mojo, 2014-03-13

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

All times shown according to UTC.

Time Nick Message
00:03 maxhq joined #mojo
00:05 punter marcus, your OAuth2 plugin for Mojo doesn't pass the tests. The fix is in a pull request in github, please apply. Thanx
00:11 punter Sorry for saying "please apply", it slipped. I meant "Could you please apply?" :-)
00:11 punter thx
00:11 good_news_everyone joined #mojo
00:11 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/Y2E5LA
00:11 good_news_everyone mojo/master faf0c70 Sebastian Riedel: make sure content type is not affected by variants
00:11 good_news_everyone left #mojo
00:17 good_news_everyone joined #mojo
00:17 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/rFb5GQ
00:17 good_news_everyone mojo/master 11e493b Sebastian Riedel: added hook example for choosing a template variant
00:17 good_news_everyone left #mojo
00:19 sri it's funny how rails folks are starting to complain about how rack is holding back ruby on the web
00:20 * sri bets there will be a custom async/heavily threaded web server trend soonish
00:24 sri although ruby is kinda trapped, i don't see an async eco system in their future... even eventmachine is kinda niche
00:25 sri perhaps a heavy investment in proper threads
00:27 btyler joined #mojo
00:32 jack joined #mojo
01:01 dnbe joined #mojo
01:02 dnbe This may be a dumb question, but I'm finding no real references on this, but can someone speak towards Mojo's input sanitization?
01:02 dnbe Does it do any by default (e.g. stuff coming in via a form it creates using the helpers?)
01:06 d4rkie joined #mojo
01:06 laouji joined #mojo
01:21 jberger__ joined #mojo
01:27 jb360 joined #mojo
01:36 thowe joined #mojo
01:36 disputin joined #mojo
01:37 thowe Is there an "app" method for Mojolicious?  I can't seem to find one documented and yet that seems to be what's used on line 13 here?  https://github.com/tempire/MojoExample/blob/master/lib/MojoFull.pm
01:37 punter What are some possible reasons for getting a 404 on a new route on a new site?
01:37 punter I'm getting a 404, and can't figure out the reason
01:39 thowe ah, it's the Controller....
01:42 bbery punter: kind of new to mojo but I'd spell check match your: get '/path/' => sub call to the render('index'  . . . .  to the @@ index.html.ep  . . . line
01:43 punter I spell checked the path
01:43 punter I'll try to render text instead of json
01:44 punter nothing changed
01:45 bbery Is is safe to assume that Mojo is using Ruby erb style variable parsing in templates? as <%= title %> looks like it
02:03 punter found what caused my problem (the class with the route was named API.pm instead of Api.pm)
02:03 punter left #mojo
02:08 klapperl_ joined #mojo
02:18 hummeleBop joined #mojo
02:27 abra_ joined #mojo
03:25 bd joined #mojo
04:07 sujithm joined #mojo
04:23 laouji joined #mojo
04:55 cleverfox joined #mojo
04:56 mikegrb joined #mojo
05:05 mikegrb joined #mojo
05:05 fhelmber_ joined #mojo
05:08 hesperaux_ joined #mojo
05:46 hesperaux__ joined #mojo
05:54 laouji joined #mojo
06:04 sujithm joined #mojo
07:01 Eke- joined #mojo
07:22 sujithm joined #mojo
07:29 ver joined #mojo
07:30 Vandal joined #mojo
07:42 D4RK-PH0ENiX joined #mojo
07:47 avkhozov joined #mojo
08:03 eseyman joined #mojo
08:15 punter joined #mojo
08:18 punter thanks, marcus . (I hope you'll be uploading soon)
08:18 cpan_mojo Mojolicious-Plugin-OAuth2 1.1 by Marcus Ramberg - http://metacpan.org/release/MRAMBERG/Mojolicious-Plugin-OAuth2-1.1
08:18 punter :-)
08:19 marcus punter: was just waiting for a rebase
08:36 trone joined #mojo
08:38 punter joined #mojo
08:49 rawler joined #mojo
08:55 basiliscos joined #mojo
09:08 rawler joined #mojo
09:19 * sri wonders if respond_to should support variants too
09:20 sri ->respond_to('html+tablet' => {...}, ...);
09:21 denisboyun joined #mojo
09:25 * batman doesn't get what variants solve
09:26 batman why do you want to render another markup to tablet than desktop?
09:26 batman s/markup/document/
09:36 spoot joined #mojo
09:37 sri http://pastie.org/8913698 # possible patch
09:41 sujithm joined #mojo
09:41 mattastrophe joined #mojo
09:44 * batman thinks it's bad ux design to serve different pages
09:45 jkg I think it's reasonable to offer a different UI on different devices. even if tablet screens are getting closer to desktop screens, tapping is not at all like mouse-clicking.
09:46 jkg I mean in general, I'm not mojo-hardcore enough to have a strong opinion on whether variants are the best way to do that.
09:48 batman different ui != different page
09:48 jkg sure, it's not the only way.
09:48 batman i'm -always- serving the same html to an iphone as a hi-res desktop
09:49 meshl joined #mojo
09:49 batman but it's a bad way. serving different pages often make users choose what to include and not.
09:49 batman hiding information becuase it's a small screen is just evil.
09:49 batman it's a really bad user experience
09:49 jkg is it, always?
09:49 batman yes.
09:49 batman always.
09:51 batman think abou it: what do you want to hide? if ti's something you could exclude on a small screen, it should probably be excluded on any screen
09:51 jkg there are a number of sites I visit which have separate mobile versions with _totally_ different page structures to the "full" desktop sites, with, yes, less information, and yet I find the mobile versions more useful.
09:51 batman because it's simply not important enough
09:51 sri yea, because serving IE workarounds to your mobile users makes so much sense
09:51 Jonneh joined #mojo
09:52 sri they certainly won't care about a little overhead
09:52 batman <!--[if gt IE 9]> ftw
09:52 batman or lt even ;)
09:53 sri because latency is a non-issue for mobile
09:53 jkg batman++ for making me think about this more and sri++ for variants, I think :)
09:53 sri actually... ignore me... i'm not interested in the discussion
09:54 * batman need to get some lunch
09:54 batman and i hope i never need to use variants. that will make me very sad.
09:54 crab if it's something you could exclude on a small screen, it should probably be excluded on any screen ← a bit extremist, don't you think?
09:55 batman crab: no.
09:55 jkg crab: I'd have said absolutist. but yeah.
09:55 crab what's the reason for larger screens then?
09:55 batman most people don't read what is written on a web page, so that means you should exclude as much as possible
09:55 batman you should always design for the smallest screen and then make it good on the desktop.
09:56 batman ^ imo
09:56 crab how? by adding unimportant rubbish?
09:56 batman i'm always talking like an absolutist. that's just the way i talk. sorry about that.
09:56 batman crab: yes. yes! YES!!!
09:56 batman no. of course not.
09:57 crab you can either be a batist or a manist. make your choice!
09:57 batman no idea what that means
09:57 batman anyhow. i need to get some lunch.
09:58 rawler hi.. is it possible to get a full url from "url_for"?
09:58 crab ->to_abs
09:59 rawler crab: ahh, thanks! :)
09:59 rawler *doh* I've seen it mentioned a few times but confused it with absolute paths.. :S
10:00 batman rawler: remember that ->to_abs will also include username:password when using basic auth
10:00 rawler ouch! :S but not a problem for my case.. :)
10:00 rawler (perhaps something that should be documented at http://mojolicio.us/perldoc/Mojo/URL#to_abs ?)
10:01 sri batman: that is false
10:03 sri anyone else think variants should be removed again?
10:04 sri i'll throw away the respond_to patch for now
10:18 batman rawler: sri is right. i was thinking about the request url.
10:25 rawler oki.. works for me either way.. :)
10:29 dpetrov_ joined #mojo
10:40 al joined #mojo
11:00 d4rkie joined #mojo
11:10 denis_boyun joined #mojo
11:13 Akron joined #mojo
11:22 mjb152 joined #mojo
11:27 mjb152 hi, I;m having problems with mysql and "server gone away" messages.  I added a "has" entry, and then setup as helper inside the startup, but I still get the problem.  http://pastebin.com/iFmWsJAY
11:27 mjb152 is it because I have other helpers which refer to the DB also in the startup section ?
11:28 Akron mjb152: DBIx::Connector probably?
11:28 Akron That should reconnect whenever a connection is lost.
11:29 mjb152 I'm using standard DBI, but when I tried to use DBIx::Connector it complained about not having "prepare" , so it meant having to rewrite all the helpers.
11:31 Akron DBIx::Connector makes prepared statements necessary? I didn't know that.
11:32 mjb152 no, sorry. in all my code I prepare the sql statements and then execute them.  When I changed to DBIx::COnnector and setup a helper for the db, all the dbi calls failed.
11:32 mjb152 think I need to find some examples on using DBIx::Connector.
11:35 mjb152 joined #mojo
11:35 Akron Ah - I see. https://github.com/kraih/mojo/wiki/Hypnotoad-prefork-web-server
11:36 Akron Oh - sorry, that's no example.
11:37 Akron mjb152: Here e.g. http://blogs.perl.org/users/joel_berger/2013/10/some-code-ports-to-mojolicious-just-for-fun.html
11:37 mjb152 here's a great example,  that returns the handle,   but then I guess its a case of changing all my code to stop using DBI and use DBIx::Connector
11:37 mjb152 https://github.com/benvanstaveren/Mojolicious-Plugin-Database/issues/4#issuecomment-37270628
11:39 mjb152 all my helpers use $dbh  , preapre some sql, eval it, and return the data
11:40 Akron Ah. Well - I don't know the Database Plugin, this might ensure the connection by default. The problem with connections and why you should rely on DBIx::Connector is more complex with preforking.
11:41 mjb152 yes I can see thats the way to go.   Think I might be a bit confused the returned connection and handle, so I'd better read up some more.
11:41 Akron I guess with DBIx::Connector you simply wrap this in a run or in a transaction. Not too complicated to change that, I guess.
11:41 Akron Good luck!
11:41 purl You'll need it.
11:41 mjb152 :-)
11:41 neilhwatson joined #mojo
11:56 sri hehe, connection timeouts are annoying... which is why i made Mango take care of everything
12:01 Akron Especially annoying in rfc2324.
12:05 mire__ joined #mojo
12:06 mjb152 ok that was much easier than I thought.  I used DBIx::Connector as per that example,  and setup a helper in the startup to return the connector.   Then just before all my normal DBI stuff,  I used a dbh to refer to the connector, and all works well with no other changes.  my $dbh  = $conn->dbh;
12:06 mjb152 I thought I was stuck with using "run".
12:07 sri send a patch for Mojolicious::Plugin::Database
12:17 chorny joined #mojo
12:39 Jonneh joined #mojo
12:48 sujithm joined #mojo
13:25 spoot joined #mojo
13:32 pernicious mjb152: Is this for MySQL?
13:32 pernicious ah, I see it is
13:32 pernicious If you like the DBI way of things, might want to look at Mojar::Mysql::Connector
13:33 pernicious passing it "-dbh => 1" gives you a 'dbh' helper that takes care of timeouts
13:34 pernicious so $self->dbh gives you a connection guaranteed to be working (assuming the server is working, obviously)
13:58 abhishekisnot1 joined #mojo
14:01 bluescreen joined #mojo
14:05 btyler joined #mojo
14:08 hesco joined #mojo
14:16 hesco left #mojo
14:16 hesco joined #mojo
14:23 sujithm joined #mojo
14:26 gryphon joined #mojo
14:26 punter joined #mojo
14:26 sh4 joined #mojo
14:29 jrbeilke joined #mojo
14:46 btyler http://krondo.com/blog/?p=1209 # been playing with twisted, and this seemed like a particularly clear introduction to the async model (particularly the 3 task diagrams); maybe a mojo book writer will find some inspiration :)
14:58 mjb152 joined #mojo
15:00 maxhq joined #mojo
15:08 rem_lex joined #mojo
15:12 mire__ joined #mojo
15:15 bluescreen_ joined #mojo
15:16 sujithm joined #mojo
15:16 bluescreen_ joined #mojo
15:20 batman is there some hooks that is not run in apps that is mounted inside another app?
15:21 Jonneh_ joined #mojo
15:28 ua_1 joined #mojo
15:30 bluescreen__ joined #mojo
15:31 bluescreen joined #mojo
15:37 malikai "Any solution that involves perl and redis will eventually lead to the rewriting of JMS" - coworker
15:37 * malikai lulz
15:38 dotan_ joined #mojo
15:39 malikai i really need to find a way out of the java asylum
15:39 malikai actually, it's a java asylum ward of the RHEL prison complex..
15:41 * malikai gets back to looking busy for the guards
15:44 gryphon joined #mojo
15:48 meshl joined #mojo
15:48 batman malikai: haha :)
16:32 d4rkie joined #mojo
16:33 arthas joined #mojo
16:51 spoot joined #mojo
16:52 hesco joined #mojo
16:53 punter When defining routes, what's the difference between $r->route and $r->all ?
16:54 Andy2 joined #mojo
17:08 sri hmmmm... i like self-hiding menu bars... think i might add one to mojolicious
17:10 crab what is a self-hiding menu bar?
17:10 maxhq joined #mojo
17:10 crab and how do you find it?
17:10 sri you'll see in a few minutes :)
17:14 crab not if it's properly hidden!
17:15 pernicious Is it live yet?  I genuinely can't see it.  Amazing!
17:18 good_news_everyone joined #mojo
17:18 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/j8SjJA
17:18 good_news_everyone mojo/master 5e47444 Sebastian Riedel: improved built-in templates with unobtrusive menu bar
17:18 good_news_everyone left #mojo
17:18 sri live on http://mojolicio.us
17:19 sri very noticeable when browsing the docs http://mojolicio.us/perldoc
17:19 sri it's rather tricky to make it work with the frontpage hack :S
17:20 pernicious works nicely in the docs
17:20 sri also works for the 404 and 500 pages
17:21 pernicious Does it mean you have to go back to the start of the doc if you want the menu bar?
17:21 sri scroll up
17:22 crab oh i see. it auto-hides if you scroll down.
17:22 sri it reappears if you scroll back a few pixels
17:22 crab kinda like the mobile chrome's location bar on my phone.
17:22 pernicious ah, I get it
17:23 sri disappears after about 250 pixels scrolling down, and reappears as soon as you scroll back up, and disappears again after a few pixels down
17:23 pernicious We might need a dancing dog to bark those instructions to the user</management-input>
17:24 crab of course, the dog has to be a user-selectable breed, including cats</cultural-review>
17:24 pernicious :D
17:24 sri only on the frontpage it sometimes jumps a little because of the menubar offset
17:24 sri maybe you see it when you scroll very very fast
17:25 sri i think it might be a bug
17:25 sri (in chrome)
17:25 genio works fine for me in chrome.  no oddities
17:26 sri http://mojolicio.us/ # scroll up and down very fast
17:27 genio sri: yea, but that's only noticable if you actually grab the scroll bar and shove your mouse up and down
17:27 jrbeilke seems to happen right when I hit the threshold for fixing the menubar to the top of the window
17:27 genio with the scroll wheel it works pretty darn quickly
17:27 sri the menu bar gets pulled into the middle and performs the transition animation to the top... it's quite odd... since the class responsible for it has already been removed
17:28 sri actually it's the opposite of what i said :)
17:29 sri there might be a lag that results in the transition class to be applied too early
17:30 sri although... even like that i think it's worth it
17:30 * sri loves unobtrusive menu bars
17:34 sri wonder how it looks in ie10
17:35 genio same as chrome for me
17:35 sri \o/
17:35 genio ie 10 standards mode
17:37 denisboyun joined #mojo
17:38 sri http://pastie.org/8916429 # with this patch is triggers all the time if you scroll slowly past the menu bar
17:41 sri haha, think i got a reliable solution \o/
17:42 good_news_everyone joined #mojo
17:42 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/p_W71g
17:42 good_news_everyone mojo/master c10c550 Sebastian Riedel: less jumpy menu bar
17:42 good_news_everyone left #mojo
17:42 sri http://mojolicio.us/ # i can't trigger it anymore
17:44 sujithm joined #mojo
17:50 mire__ joined #mojo
17:57 good_news_everyone joined #mojo
17:57 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/KpEebQ
17:57 good_news_everyone mojo/master d937a41 Sebastian Riedel: mention how to customize templates directories in rendering guide
17:57 good_news_everyone left #mojo
18:03 sri the menu bar works quite well with mobile chrome
18:20 malikai heh.. it even works with resizing
18:21 malikai er/zooming
18:23 cleverfox joined #mojo
18:24 malikai http://i.imgur.com/jxBZG.jpg
18:40 punter marcus, I believe you fine Mojolicious OAuth2 module has certain omissions. Mind if I code them a bit to make it more complete? (I need it more complete for my project)
18:41 punter I'll tell you soon which ones.
18:45 genio (house hunting)--
19:03 julian_ joined #mojo
19:12 jesteves joined #mojo
19:13 jesteves Hello everyone!  First time around :-)
19:15 go|dfish jesteves: Welcome
19:15 jesteves Hopefully someone can help with some situation I'm facing...
19:16 jesteves I wpuld define it as a "caching issue"...
19:16 crab never mind defining it. describe it instead.
19:17 crab then we'll beat it with a stick.
19:17 crab it may or may not help you, but hey, it'll sure make us feel better.
19:20 jesteves In my app I need to handle connections for several "virtual hosts", and I'm finding that some routes and templates that should be defined only one of them gets "cached" for *all* of them...
19:24 jesteves Argh, sorry guys, didn't pay attention of the hour.  I'll try to make myself clear in a moment :-S
19:43 jesteves Back again...  I guess my question is "if you would need to have some templates or routes defined based on which URL is used to access the app, what would you recommend to do?"
19:47 jesteves In my scenario I cannot separate my "virtual hosts" in different instances of my app, running in different ports.  I need to handle all of them with just one copy of the app, so I need to identify for which "virtual host" is the connection for and handle it accordingly.
19:52 disputin joined #mojo
19:53 jesteves I'm already have a working solution ($app has 'sites') for most of it.  "Shared templates" or "shared routes" work just fine.  The problem is on templates or routes that have to be defined in only one "host" and not in the others...
19:55 jesteves Since $app->renderer and $app->routes are singletons,  once one "host" has defined a template or route for it, it gets defined for all of them...
19:57 crab http://mojolicio.us/perldoc/Mojolicious/Guides/Routing#Conditions
19:57 crab does that help?
19:58 crab since you're trying to describe things in an indirect and high-level way, i still am not sure i understand your problem.
19:59 crab but if it's what i think it is, you can write a host condition and add over(Host => whatever) to selected routes
20:01 batman http://home.thorsen.pm/?_size=iphone-5 # new plugin on it's way. https://github.com/jhthorsen/mojolicious-plugin-responsinator
20:01 batman feedback is more than welcome
20:04 good_news_everyone joined #mojo
20:04 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/uix7kw
20:04 good_news_everyone mojo/master 68fb6e9 Sebastian Riedel: slightly faster scrolling
20:04 good_news_everyone left #mojo
20:07 jesteves Ok, I'll give it a thought with conditions.  But, even then, just to be certain: if in the handler for one condition I set something in $app, that setting will apply to whatever condition, isn't it?
20:15 crab your question makes no sense
20:17 crab when you say "set something in $app", what do you mean? that might help to understand what "apply to whatever condition" means. maybe.
20:18 sri can confirm, am confused
20:23 jesteves joined #mojo
20:32 jrbeilke what happens when a routing condition returns undef? will it continue checking other routes? I'm working on a store app where all the urls are in the form /:id.html and I need to check the db whether it's a vendor/category/product and route accordingly
20:33 beyondcreed joined #mojo
20:36 mattastrophe joined #mojo
20:42 X joined #mojo
20:42 X left #mojo
20:58 punter What's the difference between $r->all and $r->route ?
21:21 cleverfox joined #mojo
21:23 marty jrbeilke:  the first matching route is used.  No others are tried after that.
21:25 good_news_everyone joined #mojo
21:25 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/NguRyw
21:25 good_news_everyone mojo/master 37dc0cc Sebastian Riedel: mention that conditions need to return a true value
21:25 good_news_everyone left #mojo
21:28 marty punter:  out of curiosity, where did you find $r->all in the documentation.   I recall seeing that but I cannot find it.
21:29 good_news_everyone joined #mojo
21:29 good_news_everyone [mojo] kraih tagged v4.89 at 24ac6f7: http://git.io/eLktEg
21:29 good_news_everyone left #mojo
21:31 cpan_mojo Mojolicious 4.89 by Sebastian Riedel - http://metacpan.org/release/SRI/Mojolicious-4.89
21:34 denis_boyun joined #mojo
21:36 * sri is a sucker for litttle design details like the new menu bar :D
21:40 punter marty, I don't remember.
21:41 punter probably "any" is what I meant, marty
21:42 punter So new question: (Anyone) What's the difference between $r->any and $r->route ?
21:42 marty punter:  ahh, well in that case...   http://mojolicio.us/perldoc/Mojolicious/Routes/Route#any
21:42 punter ok
21:43 marty sri++ on kewl menu bar.   Looks nice
21:47 sri \o\
21:47 sri /o/
21:58 jrbeilke thanks marty I'm thinking I'd be better off using a hook to grab the item id from the url and re-map the route based on the item type (ie /:id.html becomes /category/:id.html)
21:59 dvinciguerra joined #mojo
22:01 marty jrbeilke:  A bridge might also be useful in your situation.
22:05 jrbeilke would a bridge be able to modify the route for an incoming request marty?
22:06 trone joined #mojo
22:11 cleverfox joined #mojo
22:13 marty in your bride you could check for what type of :id it is and dispatch accordingly or return 404 if not found.  It may not work for your situation, lots of ways to do it.
22:13 meshl joined #mojo
22:22 meshl joined #mojo
22:24 cleverfox joined #mojo
22:27 punter marcus, Why is it in the OAuth2 plugin that the user is forced to choose async or not async, when this could be determined by IOLoop->is_running?
22:28 punter marcus, my question put better: "Why does the user have to pass async => 1, when that could be determined by IOLoop->is_running?"
22:30 punter marcus, or even better: it should be determined by the presence or absence of the callback or the on_success handler
22:31 punter marcus, Right now you're forcing users to type async => 1 for no reason (I think, please correct me if I'm wrong)
22:34 paladinn joined #mojo
22:48 marcus punter: async is not required if you provide a callback
22:48 punter marcus,  it is required, though, if you provide an on_success handler
22:49 marcus punter: https://github.com/marcusramberg/Mojolicious-Plugin-OAuth2/blob/master/lib/Mojolicious/Plugin/OAuth2.pm#L49
22:50 punter that's not the same callback - $args{'callback'} is one thing, $cb is another
22:50 punter see line 64
22:50 punter 64 only checks $cb
22:51 punter ( marcus )
22:52 punter marcus, from what I understand, the user isn't supposed to enter a "callback => sub {...}" parameter
22:52 punter marcus, only either an on_success one, or a sub at the end of the parameters list
22:53 punter the documentation doesn't say that I'm wrong
22:53 marcus punter: on_success is called even for sync
22:53 marcus punter: so you can't change to rely on that without breaking backwards compat
22:53 punter ok, but what's the point of calling on_success in sync?
22:53 punter o
22:53 punter ok
22:54 punter maybe in v2.0?
22:55 punter Is sri thinking about letting promises be placed where $delay->begin currently live?
22:55 marcus If I changed it in v2.0 I'd probably remove on_success and just return the token in sync.
22:56 marcus I never use any other calling style than the callback at the end
22:56 punter How do you deal with errors then?
22:56 sri no he is not, promises are not on the mojolicious roadmap
22:57 punter won't you have to have an on_error handler in the params, marcus ? So you'll be defining on_error before the callback (which is the last param)?
22:57 sri if you want promises in mojolicious prototype it on the outside and make a proposal for inclusion in core
22:57 marcus punter: Inside the callback. It gets passed without a token
22:57 marcus and with the tx as the last arg
22:57 punter ok
22:58 punter sri, ok, I'll try that.
23:02 punter i guess then that calling with $cb is the proper way to use this module.
23:05 marcus punter: in async mode yes
23:06 punter thanks
23:07 * sri did not know luajit had a 2gb memory limit :o
23:11 mire__ joined #mojo
23:17 bluescreen joined #mojo
23:17 punter thanks, it all makes sense now
23:18 punter There's no improvements to be made in the module
23:25 punter How does one handle errors in Mojo::IOLoop->delay?
23:25 punter I mean, if you have 10 subrefs to execute in the delay, and #5 fails and you decide you don't want to continue with the others
23:25 punter how is this error caught?
23:26 punter And is the RAM that holds all of these subrefs ever freed? (or is there a leak??)
23:27 cstamas joined #mojo
23:33 cleverfox joined #mojo
23:36 marcus punter: just return without calling the rest of the chain. And that should not leak :)
23:37 punter thanks
23:40 mattastrophe joined #mojo
23:53 dvinciguerra joined #mojo
23:55 punter After all this discussion, I think I'm going to use the blocking form of your library, to be able to take advantage of mysql transactions (and detection of any errors in between, in order to rollback)

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