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

IRC log for #mojo, 2014-03-04

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

All times shown according to UTC.

Time Nick Message
00:03 good_news_everyone joined #mojo
00:03 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/a4SGFw
00:03 good_news_everyone mojo/master d73ea17 Sebastian Riedel: test invalid bareword
00:03 good_news_everyone left #mojo
00:12 sri if we deprecate j() what should we put into ojo for oneliners?
00:23 drift joined #mojo
00:24 jack joined #mojo
00:26 aggrolite joined #mojo
00:33 davido_ dj and ej
00:43 marty joined #mojo
00:47 drift joined #mojo
01:00 punter joined #mojo
01:05 bluescreen_ joined #mojo
01:08 d4rkie joined #mojo
01:16 mattastrophe joined #mojo
01:21 jberger__ joined #mojo
01:33 ka2u joined #mojo
01:39 sri i've also considered new b() methods
01:40 sri b($bytes)->to_perl... b->from_perl({foo => 'bar'});
01:40 sri ;p
01:40 sri functions with multiple chars don't fit into the theme
01:41 sri b->to_json({foo => 'bar'}); b($bytes)->from_json;
01:49 good_news_everyone joined #mojo
01:49 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/iflHQQ
01:49 good_news_everyone mojo/master 3ea1eb3 Sebastian Riedel: fixed null decoding error and improved Mojo::JSON tests
01:49 good_news_everyone left #mojo
02:01 laouji joined #mojo
02:01 marty joined #mojo
02:05 sri of course the same could have been done with x()
02:06 sri b('<p>test</p>')->to_dom->find(...
02:06 sri :)
02:10 jnbek is there any caching built in to the Mojo::UserAgent ?
02:10 sri no
02:10 jnbek so say a remote call fails, you can load the cached copy
02:10 jnbek ok
02:10 jnbek any plans to? thoughts against it? etc
02:11 sri no. really fucking complicated.
02:12 jnbek I suppose using Storable would be easy enough to just store the $tx object if it succeeds, update the Storable key, else load the old one.... hmmm.... ( something I could do )
02:12 jnbek seem to have missed a critical . in my last post >_<
02:12 sri oh, you mean shitty caching... i thought "correct" caching
02:13 jnbek yea, just cheap "OMG!! The METAR site is down/loaded and I didn't get my updated data..." caching
02:13 jnbek :)
02:14 jnbek but proper caching would be nice to explore too....
02:14 jnbek hmm
02:17 jnbek ok thanks :) that'll keep me busy for a while
02:17 jnbek thanks for the excellent work on Mojolicious sir
02:17 sri :)
02:22 aggrolite joined #mojo
02:32 klapperl_ joined #mojo
02:37 bayashi joined #mojo
03:35 aggrolite joined #mojo
04:10 damaya joined #mojo
04:10 damaya bpmedley: I have that system I was telling you about for testing Mojo with wrk.
04:11 damaya bpmedley: the other system I was using is completely unreliable, has hardware faults, and a lot of other issues. This one does not.
04:11 damaya It has 128 32GB DIMMS
04:11 damaya And 4 SPARC T5-8 procs
04:11 damaya brb
04:27 jberger__ Wow, I think I've finally figured it out
04:27 laouji joined #mojo
04:28 jberger__ Mlehmann gets more civil the more valid his point is
04:29 jberger__ Thus the wildly irrational screaming and ad- hominem attacks over hash ordering now make perfect sense
04:30 damaya jberger, do you know the test bpmedley was doing?
04:30 damaya I'd like to run it just to see what I can get from this system
04:30 damaya oddly I cannot seem to find it on his github page
04:33 damaya found it... hmm, not going to use nginx
04:34 jberger__ https://github.com/jberger/FrameworkBenchmarks
04:34 jberger__ All you really need is the main script
04:35 jberger__ And the db setup files, which they wrote in an inexplicably non functional way
04:35 damaya root@192:~# prtconf | grep "Memory"
04:35 damaya Memory size: 4193024 Megabytes...
04:35 sri oh benchmarks
04:35 purl benchmarks are like college lab reports....often fudged to achieve the desired results
04:35 damaya Anyone know how to get CPU info in Solaris?
04:36 sri if you didn't optimize something in mojolicious as a result i hate you!
04:36 jberger__ https://github.com/jberger/FrameworkBenchmarks/blob/mojolicious/mojolicious/app.pl
04:36 damaya jberger, are you saying I need to set up a DB on here?
04:36 jberger__ For anything useful yes
04:36 damaya Getting Perl + Mojolicious is not a problem, a DB may be unfortunately.
04:36 damaya dammi
04:37 damaya t
04:37 jberger__ But I never thought it would be worth it on your hardware anyway
04:37 jberger__ It's about optimizing away the latency
04:37 damaya Hm, seriously?
04:37 purl is that for real?
04:38 damaya I wish I had known that.
04:38 jberger__ Sure once you do that it should just scale with hardware power as expected
04:38 damaya You probably said it and I didn't even see it.
04:38 sri how many cores do you got?
04:39 damaya Alright, well I think then that it is easiest not to do it, because setting up a DB on here will be a pain. It doesn't have an internet connection, no pkgutil, and my Solaris knowledge is pretty crap.
04:39 damaya I don't even know how to check that sri
04:39 damaya cat: cannot open /proc/cpuinfo: No such file or directory
04:40 sri type psrinfo
04:40 damaya http://pastebin.com/QweFrt9S
04:41 sri that's a lot of cores
04:41 damaya It seems that it has 128 cores.
04:41 damaya 128 cores and 4TB memory.
04:41 moritz wtf.
04:42 sri if you want to test anything just for fun try https://github.com/kraih/mojo/blob/master/examples/fast.pl and https://github.com/kraih/mojo/blob/master/examples/hello.pl
04:42 sri those are my profiling apps
04:42 jberger__ sri, I was more proud that I got a full Nonblocking and readable benchmark implementation in in under 100 lines plus the template
04:43 sri perl -Ilib examples/fast.pl prefork -m production -w 256 -l http://*:8080
04:44 sri jberger: why non-blocking in a benchmark?
04:44 jberger__ It's not my benchmark
04:44 jberger__ Just my implementation
04:44 sri you get more rps out of it by wasting memory with more workers and blocking mongodb operations ;)
04:45 sri those benchmarks seem to ignore memory usage, so use as much as you can!
04:46 jberger__ Eh as you say benchmarks are crap, I want to show off what mojo can do
04:46 jberger__ The cool kids want non blocking, show off mojo vs node
04:47 sri i don't think it works if you're slower
04:48 jberger__ http://www.techempower.com/benchmarks/#section=environment&amp;hw=i7&amp;test=json
04:48 jberger__ Can't you have both?
04:48 sri game is rigged against you from the start if you want to compare yourself to node.js with a full framework
04:48 jberger__ Hypnotoad it out
04:49 sri in my benchmarks non-blocking hypnotoad apps have been really really slow
04:49 sri you can cope with latency, but that's not a bottleneck in micro benchmarks
04:50 sri heck, mongodb keeps your full benchmark dataset in memory!
04:51 sri you need long queries or real-time push stuff to come out on top
04:51 jberger__ Well feel free to comment on the pr if you think three old version was better:  https://github.com/TechEmpower/FrameworkBenchmarks/pull/797
04:52 jberger__ s/three/the/
04:52 sri i don't care *that* much
04:53 sri no clue what was there before
04:53 sri but if you made sure your version is faster, i trust you
04:53 jberger__ It was bad
04:53 jberger__ It was just the dancer app
04:56 sri when i participate in benchmarks i do cheat too though ;p
04:57 sri what's the point if you don't optimize the hell out of what you got
04:57 jberger__ https://github.com/TechEmpower/FrameworkBenchmarks/blob/master/mojolicious/app.pl
04:57 jberger__ That was the old one
04:57 jberger__ Well feel free to comment
04:58 jberger__ I'm all for taking advantage where possible
04:58 sri looks really shitty
04:58 sri but mysql is a good choice for benchmarks :o
04:58 jberger__ Mine or the old one :-P
04:59 sri nice double encoding there
04:59 jberger__ Btw they do make a distinction between real - world and stripped down
04:59 damaya What I find interesting is that I ran these benchmarks (not all, but many), and found much different results. Any idea why that may be?
04:59 jberger__ Yeah I saw that
04:59 damaya I ran them according to the directions on github as well. I guess it could possibly be due to the fact that I was dealing with faulty hardware? Not really too sure.
04:59 sri AND WTF!!!!??? "my $id = int rand 10000 + 1;"
04:59 damaya Try the go benchmark vs mojolicious and you will see what I mean.
05:00 sri so one benchmark may randomly get the same row over and over from the db cache?
05:00 sri WTF!
05:00 * sri runs away screaming
05:01 jberger__ Again not my spec
05:01 jberger__ Abd
05:02 jberger__ And the spec specifically requires that, not just a fake random to the same number
05:02 damaya sri, so examples/fast.pl is running... do I use wrk now to actually test it? And, can wrk be on the same system, or can it be on a different system?
05:07 sri jberger__: question is if this is what we really want http://mojolicio.us/perldoc/Mojo/JSON#j
05:07 dotandimet joined #mojo
05:08 damaya man, sorry all but it's not really worth it if the benchmark is not useful. The system is in an area without internet access, doesn't have make, gcc, etc.
05:09 damaya If there is a use for it then I can take the time to get everything set up.
05:11 sri no use, just fun
05:12 sri we know hypnotoad scales pretty well now, someone did a test with 20 or so cores once
05:13 sri new timers and accept mutex fixed a lot
05:15 ka2u joined #mojo
05:18 jberger__ For the record, I think we should keep j as is
05:18 jberger__ It's still useful in practice
05:19 good_news_everyone joined #mojo
05:19 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/l-5CLw
05:19 good_news_everyone mojo/master 80357a5 Sebastian Riedel: fixed JSON rendering bug
05:19 good_news_everyone left #mojo
05:19 jberger__ We can always deprecate later if it doesn't
05:20 good_news_everyone joined #mojo
05:20 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/NUmEJQ
05:20 good_news_everyone mojo/master ca039a7 Sebastian Riedel: another JSON test
05:20 good_news_everyone left #mojo
05:20 sri $self->render(json => undef) seems ridiculous... but is valid now!
05:22 jberger__ Nice
05:23 jberger__ Hahahahahaha
05:23 purl LOLCON 6 reached.
05:23 jberger__ Patrick Stewart on Colbert as a hick
05:28 sri oh, you could do $c->req->text eq 'null' to make sure... lol
05:28 sri what a ridiculous idiom for a ridiculous feature
05:35 plenum joined #mojo
05:41 good_news_everyone joined #mojo
05:41 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/4wSLKg
05:41 good_news_everyone mojo/master ae504b7 Sebastian Riedel: do not use Exporter as a base class
05:41 good_news_everyone left #mojo
05:45 sujithm joined #mojo
05:46 * sri loves the bsg soundtrack so much
05:49 jberger__ People adopted json and its simplicity over xml
05:50 jberger__ I wonder if they continue to bike shed it like this if people will find something else
05:51 jberger__ json just got harder to use (correctly)
05:51 sri i'm on the edge about it
05:51 sri it did make JSON a little easier to implement
05:52 sri at the cost of everything that depended on strictness
05:52 * jberger__ tries to think on the other side of the coin
05:53 jberger__ It now defines a charset and encoding for text interchange
05:53 sri but yea, it is the epitome of bikeshedding
05:53 sri tim bray just made it different, not better
05:53 * jberger__ prefers his json bike shed blue
05:54 sri the old RFC has been in use since 2006 for crying out loud
05:55 jberger__ If they can do this to json, can we turn strict on by default yet?
05:57 jberger__ Backwards compatibility just took a hit in the legacy internet
05:57 * jberger__ should only fight one battle at a time
06:12 plenum joined #mojo
06:21 sri jberger: made a comment on the pull request
06:24 sri https://eval.in/108854 # "It is *hard* to write secure software in PHP."
06:24 sri hahahahaha
06:24 purl LOLCON 5 reached.
06:29 sri it's a bit ridiculous about which parts of php people make fun of these days
06:35 sri (that just went through my twitter stream with gazillions of retweets...)
06:39 good_news_everyone joined #mojo
06:39 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/3bV_Dw
06:39 good_news_everyone mojo/master dda5509 Sebastian Riedel: improved Mojo::ByteStream to allow more method chaining
06:39 good_news_everyone left #mojo
06:46 dotandimet joined #mojo
06:47 * sri could use an american breakfast http://i.imgur.com/tXLTPbf.jpg
06:55 sri you know... when i google "murican breakfast" i did not expect to stumble over something this brilliant :O http://data2.whicdn.com/images/65090154/large.jpg
07:11 good_news_everyone joined #mojo
07:11 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/MYKvUA
07:11 good_news_everyone mojo/master 9f2e0b0 Sebastian Riedel: test full JSON roundtrip
07:11 good_news_everyone left #mojo
07:13 good_news_everyone joined #mojo
07:13 good_news_everyone [mojo] kraih tagged v4.87 at c98435f: http://git.io/YGSyTw
07:13 good_news_everyone left #mojo
07:16 dotandimet joined #mojo
07:18 cleverfox joined #mojo
07:24 KindOne joined #mojo
07:27 hrupp joined #mojo
07:37 basiliscos joined #mojo
07:42 Vandal joined #mojo
08:02 rawler joined #mojo
08:09 dod joined #mojo
08:12 odc joined #mojo
08:24 trone joined #mojo
08:35 sujithm joined #mojo
08:40 KindOne joined #mojo
08:42 laouji joined #mojo
08:46 judofyr joined #mojo
08:48 dpetrov_ joined #mojo
08:51 KindOne joined #mojo
09:04 laouji joined #mojo
09:04 preaction left #mojo
09:05 dotan sri: that murican breakfast looks like stills from Epic Meal Time.
09:17 laouji joined #mojo
09:33 sujithm joined #mojo
09:36 KindOne joined #mojo
09:36 meshl joined #mojo
09:42 aleksey joined #mojo
09:49 plenum joined #mojo
09:53 laouji_ joined #mojo
10:05 nic joined #mojo
10:12 Jonneh joined #mojo
10:14 ladnaV joined #mojo
10:17 maxhq joined #mojo
10:53 snap_ joined #mojo
11:06 rem_lex| joined #mojo
11:11 sujithm joined #mojo
11:12 denisboyun joined #mojo
11:28 mire__ joined #mojo
11:30 Jonneh is there a mojolicious way of getting at the output stream as a filehandle?
11:31 Jonneh I have this weird thing that's outputting a bunch of data to stdout, but in morbo it's alll coming out in the server output, rather than to the client
11:36 odc joined #mojo
11:48 sujithm joined #mojo
12:09 batman Jonneh: STDOUT is not connected to the client. this is not cgi.
12:10 batman Jonneh: if you want CGI and mojo, you can use https://metacpan.org/pod/Mojolicious::Plugin::CGI
12:10 Jonneh yeah I understand, but is there an equivalent so that I can do non-blocking IO
12:10 batman $c->write("data");
12:10 batman $c = $self in your controller class.
12:10 Jonneh yeah but I need a filehandle because the thing producing the data is blocking but can "stream" to a filehandle
12:11 Jonneh looking at using tie but feels super hacky
12:11 batman watch the filehandle with mojo and $c->write() on new data
12:12 batman it's what i do to read from STDOUT and $c->write() to the client in M::P::CGI
12:12 batman https://metacpan.org/source/JHTHORSEN/Mojolicious-Plugin-CGI-0.07/lib/Mojolicious/Plugin/CGI.pm#L191
12:16 Jonneh specifically I want to stream output from http://perldoc.perl.org/IO/Compress/Zip.html the problem is that the zip call itself blocks so I need to pass it something that ends up getting straight to the output stream going to the client
12:30 bpmedley Jonneh: https://gist.github.com/brianmed/8590672 <-- See how the filehandle is watched?  Do you need a better example?
12:31 bpmedley batman: Sorry, i missed your example.
12:32 bluescreen joined #mojo
12:36 skullbocks joined #mojo
12:37 bluescreen_ joined #mojo
12:40 bluescreen__ joined #mojo
12:45 batman bpmedley: no worries :) i think your example is easier to read anyway :)
12:45 batman Jonneh: maybe you want https://metacpan.org/pod/Mojo::IOLoop::ForkCall ?
12:52 jberger__ sri thanks for the vote of confidence
12:53 skullbocks joined #mojo
12:57 jberger__ Mmmmm yeah that's pretty much what I have for breakfast every day
13:06 bowtie joined #mojo
13:15 basiliscos joined #mojo
13:16 marcus I usually just kill a whale with a spear and eat it raw for breakfast.
13:17 laouji joined #mojo
13:21 plenum joined #mojo
13:24 mattastrophe joined #mojo
13:34 dotandimet joined #mojo
13:38 cpan_mojo joined #mojo
13:38 jberger__ That's the Norwegian breakfast, as we all know
13:39 jberger__ I need to figure out why cpan_mojo doesn't reconnect well
13:39 jberger__ When it drops off, its still running on my box
13:41 d4rkie joined #mojo
13:46 jberger_ joined #mojo
14:00 d4rkie joined #mojo
14:12 marty omg.  that murican breakfast sammich looks yummy!
14:21 Coke mmmmmm, breakfast sammmmmmich.
14:21 Coke (stupid oatmeal. :P)
14:24 Dandre joined #mojo
14:25 sh4 joined #mojo
14:27 sujithm joined #mojo
14:28 mire__ joined #mojo
14:28 bluescreen joined #mojo
14:29 bowtie_ joined #mojo
14:43 skullbocks joined #mojo
14:44 rdunklau joined #mojo
14:44 rdunklau Hi.
14:44 rdunklau Is there a way to use significant trailing slashes with mojolicious routes ?
14:46 genio haha, that 'murican breakfast does look good!
14:46 zivester joined #mojo
14:58 nic rdunklau: last time I looked, they were significant
14:58 nic (I had to manually make them insignificant)
15:06 rdunklau nic: The documentation explicitly says they are optional (http://mojolicio.us/perldoc/Mojo/Path#trailing_slash )
15:07 rdunklau And I can't figure how to create two different routes for /some_path and /some_path/
15:11 nic In the places where it mattered, I redirected /some_path/ to /some_path (with the controller)
15:12 nic s/with/within/
15:13 nic (but I'm talking about the user req, not route construction)
15:14 gryphon joined #mojo
15:16 rdunklau nic: I'm not sure I follow: you mean you manually redirected stuff with a leading / to an url without one ?
15:17 rdunklau The behaviour I notice is that no redirection takes place, but both /some_path and /some_path/ route to the same destination
15:24 JSI jberger: I tried get headers with "scalar" prefix, and everything is fine! Thanks for your help and sorry again for my false issue report!
15:35 skullbocks joined #mojo
15:37 jberger_ JSI: glad it's working for you
15:37 jberger_ Happy Perling
15:42 bowtie joined #mojo
15:44 nic rdunklau: yes, /some_path and /some_path/ route to the same controller
15:45 nic but I had a controller using relative links, which broke if the user added a trailing '/'
15:45 nic so in that case I redirected the user to the url with no trailing '/'
15:49 rdunklau nic: is there "some way" to change this behaviour ?
15:49 rdunklau It seems like a pretty opinionated choice.
15:49 pernicious slap the users
15:49 rdunklau pernicious: it is not about the users, but about routes.
15:50 rdunklau Right now, I can't declare routes like /user/add and /user/user1/
15:50 rdunklau One would be a route to add a user, the other one to show details about user1
15:50 pernicious It sounds like your problem is something v different
15:50 rdunklau Using mojolicious, my only solution would be to forbid using "add"  as valid parameter for the second route
15:51 rdunklau I meant /user/add and /user/:username/
15:51 sri you have $c->req->url->path->trailing_slash, so you can whatever you like
15:51 arpadszasz_ joined #mojo
15:51 rdunklau sri: I don't understand what you're saying here. Can I use this in a route declaration ?
15:52 sri if you really want to, sure
15:52 pernicious rdunklau: Are you saying you want '/usr/user1' and '/usr/user1/' to do different things?  That sounds like a bad choice
15:52 sri make a condition
15:52 rdunklau pernicious: yes I want
15:53 sri although, making url_for work, might not be possible
15:53 rdunklau What is the rationale behind the choice of "ignoring" trailing slashes ?
15:53 sri it's usually a bad choice
15:54 rdunklau sri: why ?
15:56 rdunklau I like to have generic routes like /entity/<crud_operation> and /entity/<entity_id>/ , I don't see what is so bad about it.
15:58 stephan48 by declaring crud operation first, they are ignored as values for entity_id(cause they match earlier)....
15:59 rdunklau stephan48: yes, if trailing slashes are not considered when matching the url to a route
15:59 pernicious Personally I wouldn't want to support usernames like "add", "create", and "'; drop database;"
15:59 stephan48 pernicious: wheres the problem?
16:00 rdunklau pernicious: I personnally don't care.
16:00 rdunklau And I don't want to write specific validation code for that.
16:01 pernicious stephan48: explained a solution, no code required
16:01 stephan48 pernicious: who doesn't want a user named "'; drop database;"?
16:02 rdunklau pernicious: using this solution, I must ensure that no user can be created with the username "add", because their profile wil not be accessible then.
16:03 pernicious That is much less pain that relying on /xyz and /xyz/ doing very different things
16:03 rdunklau pernicious: why would that be painful ?
16:05 basiliscos joined #mojo
16:07 crab nothing good comes of putting things of a different nature into the same namespace
16:08 stephan48 rdunklau: you don't have that many crudoperations do you?
16:08 aggrolite joined #mojo
16:09 stephan48 i could think of add, edit, delete, fetch... is it really that difficult to block 4 tokens?
16:11 mire__ joined #mojo
16:11 rdunklau stephan48: it is not that difficult, but more difficult than just doing nothing and relying on the routing layer.
16:15 rdunklau But thanks anyway for your responses, I'll put it in a separate namespace.
16:20 zivester joined #mojo
16:26 meshl joined #mojo
16:28 asarch joined #mojo
16:42 arpadszasz_ joined #mojo
16:50 dotan joined #mojo
16:52 sri rdunklau: trailing slash is usually used wrong, gets eaten by deployment servers, results in terrible usability
16:53 sri it's a saner default, you should have to work more to make worse apps
16:53 sri some reading material
16:53 sri http://googlewebmastercentral.blogspot.de/2010/04/to-slash-or-not-to-slash.html
16:53 sri http://programmers.stackexchange.com/questions/186959/trailing-slash-in-restful-api
16:56 bowtie joined #mojo
16:56 sri frameworks like Angular don't even support trailing slash at all
16:57 sri i'm throwing stuff here so i can link to this in the future
16:58 rdunklau sri: Yes, I know about that.
16:59 rdunklau But what deployment server is eating them, if not specifically configured to redirect trailing or non trailing slashes url ?
17:06 btyler joined #mojo
17:07 ka2u joined #mojo
17:08 Jonneh can someone tell me what I'm doing wrong here: https://gist.github.com/popeshoe/9350795
17:09 Jonneh as I understand it, load_chunk should immediately be sent to the client, rather than wait for the whoel thing to finish
17:09 Jonneh which is what happens for me
17:10 batman can i get the error message if $c->render(partial => 1, template => 'foo') fail?
17:13 Jonneh no error, but it does the same thing
17:14 batman Jonneh: sleep(1); does not allow the IOLoop to write anything to the client
17:14 batman you need to do non-blocking for write() and friends to work
17:14 sujithm joined #mojo
17:16 Coke question on the mojo/util.t tests  - html_unescape has the name of the test in a comment rather than in the test description - any reason for that?
17:17 Jonneh batman, bah figured it'd be something like that
17:17 Jonneh thanks anyway
17:18 batman Jonneh: https://metacpan.org/pod/distribution/Mojolicious/lib/Mojolicious/Guides/Cookbook.pod#Timers <--- non-blocking example
17:34 arpadszasz_ joined #mojo
17:35 Jonneh batman: that second example does teh same behavior for me
17:37 Jonneh dunno if i'm being a doofus or missing something, but I should be seeing each number appear after a second right?
17:37 batman probably not if you're using a browser.
17:38 batman but it should be sent over the wire.
17:38 good_news_everyone joined #mojo
17:38 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/YNTTOg
17:38 good_news_everyone mojo/master cf8bd2d Sebastian Riedel: fixed a few typos
17:38 good_news_everyone left #mojo
17:40 batman Jonneh: the browser need a big chunk of data before displaying anything. try writing bigger chunks for each recurring()
17:41 batman and remove any sleep() that you would have
17:46 Jonneh yeah you're right, curl does what i'd expect
17:46 Jonneh but chrome seems to do teh same thing
17:46 Jonneh weird
17:46 Jonneh thanks man
17:48 bowtie joined #mojo
18:02 mudler joined #mojo
18:14 bpmedley_ joined #mojo
18:18 beyondcreed joined #mojo
18:18 punter joined #mojo
18:24 sri rdunklau: anyway, if you think there's a better way the router could handle trailing slash you're welcome to make a proposal
18:52 mattastrophe joined #mojo
19:13 denisboyun joined #mojo
19:30 aggrolite joined #mojo
19:40 davido_ What if j() only encoded if handed a ref, rather than only decoding if handed a string starting with [] or {}?
19:41 sri seems rather pointless
19:42 sri although it reminds me of an actually useful side effect of the new semantics
19:42 davido_ My thought was that I it would be better to not fail to decode valid json, at the expense of being a little less convenient to encode.
19:42 sri $self->render(json => $object_with_to_json_method);
19:42 sri that actually works now
19:43 sri same for $self->send({json => $object_with_to_json_method});
19:43 davido_ Silver linings.
19:43 sri i guess batman might like that for Mandel
19:44 sri davido_: wouldn't that only add true/false?
19:45 davido_ It would permit simple strings to decode with j() without being wrapped in [] or {}.
19:46 sri i don't understand that sentence
19:47 skullbocks joined #mojo
19:49 davido_ Wait, I see your point. :)
19:52 dotandimet joined #mojo
19:55 dotandimet joined #mojo
19:58 sri j() is not mentioned in documentation for anything else than oneliners anymore... so i don't care very much anymore either
19:59 sri i would have moved it completely to ojo if it wasn't used in core in a few spots
20:01 sri official api is encode_json/decode_json with good semantics... some people like j(), and there are a few valid uses... so as far as i'm concerned things can stay the way they are now
20:02 sri new synopsis only shows encouraged use cases now http://mojolicio.us/perldoc/Mojo/JSON#SYNOPSIS
20:10 * sri opened a Mandel issue https://github.com/jhthorsen/mandel/issues/10
20:16 bluescreen_ joined #mojo
20:16 meshl joined #mojo
20:18 bluescreen_ joined #mojo
20:40 lukep joined #mojo
20:41 aggrolite joined #mojo
21:05 pernicious joined #mojo
21:05 beyondcreed joined #mojo
21:08 stephanj joined #mojo
21:17 mattastrophe joined #mojo
21:17 mire__ joined #mojo
21:28 beyondcreed_ joined #mojo
21:50 KindOne joined #mojo
21:54 punter joined #mojo
22:06 bluescreen joined #mojo
22:29 * malicious needs a message queue
22:30 aggrolite joined #mojo
22:34 punter I use a table in MySQL to hold my message queue
22:34 punter I don't use specialized software
22:41 howitdo joined #mojo
23:02 cpan_mojo Mojolicious-Plugin-BootstrapPagination 0.07 by DOKECHIN - http://metacpan.org/release/DOKECHIN/Mojolicious-Plugin-BootstrapPagination-0.07
23:06 mattastrophe joined #mojo
23:51 BenGoldberg joined #mojo

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