Camelia, the Perl 6 bug

IRC log for #mojo, 2012-07-26

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

All times shown according to UTC.

Time Nick Message
00:13 zhutingting joined #mojo
00:43 jontaylor joined #mojo
00:43 tqm tempire: here you go... work in progress but works for me https://github.com/tqm/Mojol​iocious-Plugin-CanonicalHost
00:45 tqm yay, my first public GitHub repo - be gentle <-- that's what she said ;)
00:46 tempire so if the request doesn't match the specified domain, it redirects somewhere?
00:47 tqm yes, if you run app on localhost:3000 under daemon and enable it... it will redirect to host
00:47 tqm basically the host part has to match, case insensitive
00:47 tqm port is ignored but you can rewrite it as well
00:47 tqm needs better docs, some examples etc - will write those soon
00:49 tempire I'm confused
00:49 tqm with debug=>1 it will show on daemon console/log what it did
00:49 tqm ok, what can I explain?
00:52 tqm one reason for writing this plugin was that I wanted to deploy app using hypnotoad and make sure it doesn't use 'www.' at the URL when clients call it - redirect from www.domain.tld to domain.tld, maintaining the rest of the URL
00:52 tqm trivial to do if you have reverse proxy in front of app but haven't seen this done in Mojolicious yet
00:53 tempire so if it doesn't match the specified host, it redirects to the specified host
00:53 tqm so to make sure app runs only on https on mydomain.tld I set (scheme=>'https', host=>'mydomain.tld', port=>443)
00:53 tqm basically yes
00:54 tempire ok
00:54 tqm leaving only one hostname/url under which the app resides... helpful for SEO and few other bits
00:56 tqm I was testing a web app once, where developers made proper authentication+authorization system that worked great, but it assumed the hostname is their production hostname (hardcoded values!), it was enough I added hosts file entry (<IP> failapp) and I bypassed the whole lot
00:56 tempire marcus, sri: so was the tt problem determined?
00:56 tqm google and duplicate content was the least of the worry at this point :)
00:57 tempire hmm, tqm: http://search.cpan.org/~alexbyk/Moj​olicious-Plugin-RedirectHost-0.03/
00:58 tqm hehehe, nice... shame docs in Russian and that cpan kills cyrylic encoding :(
00:59 tqm oh well, at least I got some more practice writing plugins - time now wasted, rather well spent
00:59 tempire latest dev release fixed it http://search.cpan.org/~alexbyk/Mojol​icious-Plugin-RedirectHost-0.03_2/lib​/Mojolicious/Plugin/RedirectHost.pm
00:59 tempire not wasted.  now you've crossed the line into github territory
01:00 tqm I had account for years, never used it thiugh, never needed to
01:02 tqm the good news... my code is not polluting CPAN :D
01:03 tqm I'll see you tomorrow, ermmm... later today :D good night
01:04 tempire come to think of it, you could do it with a header condition as well
01:04 tempire get '/' => (host => qr/^((?!domain.com).)*$/) => sub {
01:06 jnap joined #mojo
01:06 laouji joined #mojo
01:14 scott joined #mojo
01:32 jnap_ joined #mojo
01:43 xaka joined #mojo
02:03 xaka1 joined #mojo
02:04 d4rkie joined #mojo
02:07 scott__ joined #mojo
02:12 zhutingting1 joined #mojo
02:36 Psyche^ joined #mojo
02:42 xaka joined #mojo
02:55 noganex_ joined #mojo
03:07 zhutingting joined #mojo
03:44 asarch joined #mojo
03:51 tempire ok
03:51 tempire proposal
03:51 tempire mojo get http://www.reddit.com/r/perl.json dump
03:51 tempire data dumpers it
03:51 tempire if Data::Printer is installed, it pretty prints it with amazing colors
03:52 tempire it takes 3 additional lines of code
03:52 tempire that is, without perl trickery
03:56 GitHub52 joined #mojo
03:56 GitHub52 [mojo] tempire created json-pretty-print (+1 new commit): http://git.io/hhyltA
03:56 GitHub52 [mojo/json-pretty-print] pretty prints json if dump keyword provided - with colors if Data::Printer is installed - Glen Hinkle
03:56 GitHub52 left #mojo
03:56 tempire not much to it
04:07 lukep joined #mojo
04:26 sri tempire: -1 from me
04:26 tempire because
04:27 sri that DDP use line stuff makes me nervous
04:27 tempire because it assumes Data::Printer won't break?
04:28 sri because it uses Data:Printer
04:28 sri i'm not very keen on that
04:29 tempire but it's soooo pretty
04:29 sri but the whole feature has not convinced me
04:29 sri it seems so patchwork and not fit in
04:30 tempire it seems natural to me.  I always want mojo get to handle my json for me, color coding seems a natural extension given that it can still work with Data::Dumper
04:31 tempire all command line tools seem patchwork to me
04:31 sri that's not a good reason to make them more patchy
04:32 sri if anything it makes me want to block all changes until that is fixed ;p
04:32 tempire what about just with Data::Dumper?
04:35 sri it's still not well thought out
04:35 sri even Data::Dumper is configured wrong
04:36 sri why would you even eval use Data::Dumper?
04:36 tempire so as to not include it if it wasn't needed
04:36 tempire seemed like something you would do
04:36 sri what's the point?
04:37 tempire 'cause the dump should still work even if ::Printer isn't installed
04:37 sri don't think i would ever use a string eval if it wasn't necessary :)
04:47 sri disclaimer: i did not have breakfast yet, so i'm not very open minded towards new ideas ;p
04:50 sri but yea, where do we draw the line for get command features?
04:51 sri new dependencies for something that's not even one of the primary features seems silly
04:55 tempire I can understand that
04:55 tempire but then what's the point of mojo get?
04:56 sri tech demonstration
04:56 sri it's one of those parts i'd rather see forked than extended
04:56 tempire nonsense.  it's entirely too useful to just be a demo
04:57 sri think i figured out the inconsistency part that bugged me
04:57 sri you handled dump before json pointers
04:58 tempire it's not thought out, I'll give you that.  The goal was to have minimum impact on the existing stuff.
05:02 sri if you believe in it fight for it :)
05:06 sri mojo get http://search.twitter.com/search.json?q=perl /results/1 dump
05:06 sri that would be kinda neat
05:06 sri although there are many implementation problems
05:09 sri like if the result is not json you end up with a "unknown command" error from the dom code
05:15 laouji joined #mojo
05:16 sri Data::Printer installs stuff like Class::MOP, Mac::SystemDirectory and IPC::Run3 Oo
05:20 sri perl -Mojo -MDDP -E 'p([lala => 'foo'])'
05:20 sri ouch!
05:29 sri tempire: why do we reencode to unreadable json if a selector returns a structure?
05:34 sri tempire: what about this? http://hastebin.com/raw/yofaruxawe
05:36 Britzel joined #mojo
05:37 sri oh well, afraid i don't see a clean solution
05:39 sri actually i'm not even sure why it can't just be a separate tool
05:39 sri mojo get http://search.twitter.com/search.json?q=perl /results/1 | colorify
05:44 rem_lex|pivo joined #mojo
05:47 sri time for breakfast \o/
05:48 rem_lex joined #mojo
05:50 arpadszasz joined #mojo
05:57 spleenjack joined #mojo
06:01 sri tempire: oops, sorry for complaining about implementation details, completely missed that it was just a branch
06:02 sri tempire++ # using branches
06:02 tempire what, you thought I did it in master?
06:03 * sri nods
06:03 tempire that would be obnoxious
06:03 Foxcool joined #mojo
06:12 xaka1 joined #mojo
06:21 tempire the separate cli tool makes sense.  I was thinking it would be a quick addition.  But a separate tool is one line in an executable file.
06:23 ovnimancer joined #mojo
06:29 sri why doesn't Data::Printer have such a tool?
06:30 sri detect json/yaml on stdin and pretty print it :o
06:31 sri mojo get /foo.json | ddp
06:34 sri cpanm App::ddp
06:34 AmeliePoulain joined #mojo
06:35 * sri pokes garu
06:36 tempire indeed. I was surprised it didn't have one.
06:46 Vandal joined #mojo
07:17 mire joined #mojo
07:20 pau4o joined #mojo
07:25 taryk joined #mojo
07:26 mattastrophe joined #mojo
07:58 venkatk_ joined #mojo
08:08 mattp_ joined #mojo
08:11 cosmincx joined #mojo
08:12 geira joined #mojo
08:15 davido joined #mojo
08:47 tm joined #mojo
08:47 tm morning
08:48 kthakore mornin
08:50 davido Yes, it is. ;)
08:55 sri i can confirm this
08:56 omega afternoon here
09:03 spleenjack joined #mojo
09:05 marcus wow, I thought oauth2 was all done, and it seems they've managed to cock it up in the last strech :-(
09:06 marcus ( http://hueniverse.com/2012/07/o​auth-2-0-and-the-road-to-hell/ )
09:07 marcus damn you, enterprise!
09:08 sri marcus: wtf
09:12 kthakore Hello, I am deploying an android app that talks to a mojolicious server. I am using a token/apikey after user logs in on the phone to auth each request.
09:13 kthakore What other tips should I be aware for security?
09:13 marcus kthakore: use https?
09:13 kthakore should do that ... good point
09:13 kthakore should we roll our own certs?
09:13 davido :) Don't store your passwords in the clear, do salt them randomly per passphrase.
09:14 kthakore davido: I am using SHA digest for our passwords
09:14 marcus kthakore: you could do that, buying a cert would be fine too. if you plan to do a ios client, it doesn't support self-signed certs.
09:16 kthakore marcus: ah ok. How do I set up HTTPS with mojo?
09:17 marcus kthakore: http://mojolicio.us/perldoc​/Mojo/Server/Daemon#listen
09:17 kthakore davido: if I do a random salt per passphrase how I keep track of the random salt when it comes to authenticating the password?
09:18 marcus (assuming you meant, how do I use a custom certificate, using https is just a matter of listening on https)
09:19 kthakore marcus: yeah that is what I meant (custom cert)
09:20 davido kthakore Store the salt along with the digest.  It's ok to do so.
09:20 kthakore ok
09:20 davido two fields is the simplest implementation.
09:20 omega meh if oauth2 is a no-go, what other options are there?
09:21 kthakore two fields?
09:21 davido one for salt, one for digest.
09:21 sri damn, oauth 2.0 looks soooo bad
09:21 kthakore ah right
09:21 davido and salt to as many bits as your passphrase digest for maximum entropy.
09:21 kthakore ok
09:21 davido use a cryptographically solid random number generator.
09:22 kthakore I am looking at Crypt::Passwd::XS
09:22 davido or just go with Authen::Passphrase::SaltedSHA512
09:23 kthakore cool
09:23 kthakore thanks
09:25 davido Let me know if you find it inadequate in some way.  If you need Roles Based Access Control, I put Class::User::DBIx on CPAN after specifically designing it to be easy to use with Mojolicous::Plugin::Authentication and Mojolicious::Plugin::Authorization.  It uses Authen::Passphrase::SaltedSHA512 as its validation tentacle.
09:26 kthakore I am using that right now
09:26 sri omega: you can always do what facebook does and use draft 12, from two years ago
09:27 davido Anyway, I think I'm getting off-topic for here.
09:28 sri davido: doesn't look off-topic to me
09:28 davido Thanks.
09:28 nic tempire: mojo get ... dump will be very useful, but will that include colour codes?
09:29 nic I wouldn't expect 'dump' to prettyprint to that extent
09:29 davido Authen::Passphrase::SaltedSHA512 just specializes Authen::Passphrase by setting up some sane defaults.... and I did a lot of research into what random number generation module would be a good choice for generating good salts.
09:29 nic could the selector be changed to something like 'pretty' (and in the future someone could extend it to more than json)?
09:29 sri nic: dump will prolly not become a feature
09:30 nic 'dump' would be v useful to me, but I'd like it to be pure content, no difficult embedded codes
09:30 sri a much better solution would be for Data::Printer to get a command line tool you can pipe to
09:30 omega sri: hehe :p
09:30 nic ah, yes
09:30 sri mojo get /foo.js | ddp
09:30 nic \o/
09:31 sri *+on
09:31 davido And with Class::User::DBI, the intent was to provide a user login, that could be constrained by passphrase as well as ip-whitelisting (if so desired), and also to provide both roles based access control, and domain (or locality) based access control.
09:32 davido So far I haven't needed to modify any of the UI as I've put it in place behind Mojolicous::Plugin::Authentication and ::Authorization.
09:33 davido Unfortunately it's all a little heavy on dependencies. ;)
09:42 daxim joined #mojo
10:05 andrefs joined #mojo
10:13 mire joined #mojo
10:23 wolf__ joined #mojo
10:41 cosmincx joined #mojo
10:46 fhelmber_ joined #mojo
10:59 tm kthakore: first use SSL certs - you can get them for free these days
11:00 tm kthakore: salt and hash passwords... or even hash several times - I haven't seen yet rainbow tables that are geared for cracking even 2 rounds of SHA1, not to mention SHA256 and above
11:01 tm marcus: awesome post on OAuth 2.0 - wanted to implement it in my app, now having second thoughts
11:02 marcus tm: more depressing than awesome, imo
11:03 tm of course, awesome in the respect of someone standing up and showing the ugly truth
11:03 tm even more because of the guy's role in OAuth community
11:04 tm so what other alternatives do we have? some providers already moved to 2.0 and will be killing 1.x
11:04 robhammond joined #mojo
11:04 marcus oauth2 sucks too.
11:04 taryk joined #mojo
11:05 tm google deprecated 1.0 on 20 April... 2.0 it is if you want to use google https://developers.google.com/accounts/docs/OAuth
11:05 marcus someone should fork draft 12.
11:05 * tm just doesn't want the burden of keeping user credentials in any form
11:06 marcus I guess draft 12 is what I support in the mojo plugin
11:07 tm hmm for now I just use Yubikeys all the way as primary access method - no other credentials, just list of allowed YKs
11:08 fhelmbe__ joined #mojo
11:30 mattp_ joined #mojo
11:32 diegok joined #mojo
11:33 MisterHatt joined #mojo
11:35 sandeepdb joined #mojo
11:41 cosmincx joined #mojo
11:41 marcus http://neilb.org/reviews/http-requeste​rs.html?utm_source=feedburner&amp;utm_​medium=feed&amp;utm_campaign=Feed%3A+P​erlWeekly+%28Perl+Weekly+newsletter%29
11:42 crab validation... tentacle?
11:42 * crab shivers
11:42 bc547 maybe not completely related to mojo, but does anyone know if it is possible to somehow specify a db connection or query timeout with DBIx::Connector? Right now, the entire mojo app stalls if the db is not reachable
11:42 crab i use Crypt::Eksblowfish::Bcrypt to store passwords.
11:49 MisterHatt joined #mojo
11:55 Lucas1 joined #mojo
11:58 mattastrophe joined #mojo
12:15 * marcus replies to the user agent post
12:24 GitHub77 joined #mojo
12:24 GitHub77 [mojo] kraih pushed 1 new commit to master: http://git.io/1Rq5SA
12:24 GitHub77 [mojo/master] use more smartmatch - Sebastian Riedel
12:24 GitHub77 left #mojo
12:26 sri not sure i really like that use case for smartmatch yet, but automatic undef handling for eq comparison is kinda neat
12:27 marcus it makes it prettier tho
12:27 sri performance also seems equal
12:27 marcus less sure about !! as a replacement for scalar
12:27 sri pretty much at least
12:28 sri !! is not a replacement for scalar
12:28 sri i just like to explicitly boolify stuff when the return value was previously random garbage :)
12:29 marcus https://github.com/kraih/mojo/commit/fb9a1​e2c794269a201acd56f380c160a938e564c#L3L44 ?
12:29 sri yea, it returned a random number for true
12:29 marcus yeah
12:29 marcus the number of subscribers, rather?
12:29 sri in some other places it even was random objects
12:30 sri marcus: that's arbitrary
12:32 sri too bad smartmatch doesn't take care of undef for regex comparison
12:34 marcus sri: did you find my reply to the user agent comparison acceptable?
12:34 sri marcus: not seen it
12:35 marcus sri: http://neilb.org/reviews/http-requesters.ht​ml?utm_source=feedburner&amp;utm_medium=fee​d&amp;utm_campaign=Feed%3A+PerlWeekly+%28Pe​rl+Weekly+newsletter%29#comment-598437964
12:35 * marcus & # trampoline
12:35 sri Oo
12:36 sri marcus++ # good answer, most people making the LWP comparison completely miss the async aspect
12:39 sri marcus: i don't understand the final note bit though
12:40 sri if you mean the POST form encoding thingy he wants, then that is already a separate module with Mojo::Parameters
12:42 sri say Mojo::Parameters->new(username => $user, password => $pword);
12:45 marcus sri:  some of the code to make a full body is inside the form method and _multipart method of transactor tho.
12:45 sri marcus: he did not say he wants multipart forms
12:46 marcus sri: I suppose not.
12:47 sri in fact, i doubt he actually knows what he wants
12:47 marcus Added a note.
12:48 sri his research was a little sloppy
12:49 sri and why do i feel dirty after using smartmatch? :S
12:54 marcus Guess you're squeamish :)
12:55 Netfeed isn't most of those light tests usually, "test the thing that my module is good at and see everyone elses suck"?
12:55 jnap joined #mojo
12:56 sri i don't think he has a horse in the race
12:56 Netfeed s/my module/the one i like/ then
12:56 marcus seems he has two
12:56 marcus HTTP-Lite and HTTP-Client
12:56 marcus http://search.cpan.org/~neilb/
12:57 marcus interesting, he's actually reviewing his own modules as well
12:57 sri ah, i've not really payed any attention, since finding the fastest http client is rather uninteresting
12:57 marcus apparently he got comaint for it.
12:58 marcus seems he took over maint for both of those.
12:58 marcus I'd assume the fastest one is the c based one :)
12:58 sri except for curl::simple xD
12:59 Netfeed ease of use > speed in most cases imho
13:01 sri http is also rarely about raw performance, but coping with latency
13:03 tm that's where async comes into play... mariage of WWW::Mechanize with AnyEvent prooved it many times - now I would use Mojo::UserAgent to avoid all that cruft
13:10 tm guys, where would you host your app if some functions can get cpu heavy (think imagemagick) - any cloud or VPS or go dedicated/host yourself?
13:11 * tm is trying to pick some brains :D
13:15 sri dedicated
13:15 sri for that stuff i go with hetzner boxes
13:18 tm yeah, that's what I thought... mojo app is flying on mid-range vps no problem but image processing worker can take >5min to do the job, while basic stuff is just ~10sec - they oversell quite a lot in this case; to compare old laptop (2-core intel 1.6GHz) does the same image processing in 3sec
13:18 tm thx for vote for hetzner, never used them but will check for sure
13:20 gryphon joined #mojo
13:21 CromeDome joined #mojo
13:22 rwstaune` joined #mojo
13:25 d4rkie joined #mojo
13:28 preflex_ joined #mojo
13:29 Lucas1 joined #mojo
13:41 robhammond joined #mojo
13:45 inokenty joined #mojo
13:48 fhelmber_ joined #mojo
13:49 GabrielVieira joined #mojo
13:52 * sri wonders if there's a more pretty way to do ($foo || '') =~ /whatever/;
13:57 jnap joined #mojo
13:57 bc547 sri: like (defined($foo) && ($foo=~/whatever/)) ?
13:59 sri not a bad start, but that's a lot longer and doesn't work so well for ($self->foo || '') =~ /whatever/;
14:01 bc547 true
14:08 tm not sure if it makes much sense... ($Self->foo ||= '') =~ /whatever/;
14:09 sri wat?
14:10 GitHub51 joined #mojo
14:10 GitHub51 [mojo] kraih pushed 1 new commit to master: http://git.io/6NVuTA
14:10 GitHub51 [mojo/master] small optimizations - Sebastian Riedel
14:10 GitHub51 left #mojo
14:10 tm you do match if $self->foo is defined or do it on empty value, correct?
14:10 tm empty but defined
14:11 sri yes, but i don't see how ||= is better than ||
14:11 tm hmmm it's not as it would change value of $self->foo
14:12 sri it's not necessarily an attribute accessor, and even then it wouldn't be an lvalue
14:12 sri nobody uses lvalue accessors
14:14 sri marcus: is your perl on mountain lion also slightly slower than on lion?
14:14 * sri might have compiled it with clang
14:19 bc547 sri: how about something like (($_=$foo)?/whatever/:'') ?
14:19 sri it should be pretty :)
14:19 bc547 sri: although what you said seems prettier to me :-)
14:19 bc547 :-)
14:20 sri i'd like something along the lines of $foo ~~ 'whatever'
14:30 robhammond joined #mojo
14:47 marcus sri: I'm still using the same perl, and I haven't tested
14:47 marcus (I use perlbrew)
14:49 labrown joined #mojo
14:49 vel joined #mojo
14:54 sri i'm also using perlbrew, but i've rebuilt perl
15:18 jamesw joined #mojo
15:22 xaka joined #mojo
15:27 jnap joined #mojo
15:47 amirite joined #mojo
15:59 mire joined #mojo
16:23 marty_ joined #mojo
16:53 Britzel joined #mojo
16:57 vel joined #mojo
17:12 tempire marcus: I gave neilb quite a bit of information in in blogs.perl post
17:12 tempire http://blogs.perl.org/users/neilb/2012/07/​cpan-modules-for-making-http-requests.html
17:30 jnap joined #mojo
17:30 * sri wonders what features our users would like to see in future versions of mojolicious
17:31 sri we are usually so quick with adding stuff that there's very little left for long term goals :)
17:34 sri perl6 port is a popular long term goal i guess
17:38 xaka joined #mojo
17:43 sri hmm, all my long term goals are spinoff projects :o
17:43 spaceships joined #mojo
17:46 tempire sri: what are they?
17:46 tempire besides mongodb driver
17:48 sri most are still very fuzzy
17:48 mattastrophe joined #mojo
17:50 ZadYree sri, I'm planning to switch from the perl MongoDB module set to the Mojolicious one, will I cry during the migration?
17:51 sri but i've been having some ideas around the topics asset pipeline, message queue, event bus, high level plugins
17:51 tempire ZadYree: in that it's not complete or released, yes.
17:51 sri ZadYree: for all you know my mongodb will totally suck :p
17:52 sri *+driver
17:52 ZadYree aha
17:52 ZadYree I see :P
17:53 tempire I thought about doing an event bus a while back.
17:53 sri at the moment i'm actually just using it to experiment with blocking/non-blocking hybrids some more
17:54 sri it's interesting because the mongodb wire protocol is more versatile than http
17:59 sri showing how a first class model layer for mojolicious could look like might be a worthy goal though
17:59 sri you just need to keep marcus from killing my motivation again :)
18:00 tempire what did marcus do this time
18:00 sri bashing mongodb for the past year of course
18:01 tempire I've been happy for the past year because every project I've had has been able to avoid using a database
18:01 tempire well, directly at least
18:02 Foxcool joined #mojo
18:09 rem_lex|pivo joined #mojo
18:10 sri mongodb driver would be nice for going some more in the meteor.js direction
18:18 whee I would be more entertained with something using redis
18:18 sri but redis is not a database
18:19 whee it stores data :)
18:19 sri it's a global namespace for data structures shared between multiple servers
18:21 sri that said, i love redis and wish it could be used in place of a real database
18:23 fhelmber_ joined #mojo
18:24 whee I keep meaning to look at lamernews to see how he handled everything
18:27 sri tempire: what are your goals? :)
18:27 tempire in terms of?
18:28 tempire because the answer is usually cookies
18:28 sri you said you've thought about an event bus too
18:28 * sri supports cookies
18:29 jnap joined #mojo
18:29 sri where do you see mojolicious heading?
18:30 tempire ah, I was working for a mc enterprise at the time, and the management was all about "WE NEED AN ESB IT WILL SOLVE ALL THE PROBLEMS"
18:31 tempire and of course java would do all the things, but even the simplest solutions were quite complicated
18:31 tempire so I started looking into why exactly an esb was the 'best' solution
18:32 sri MOAR BOILERPLATE!
18:32 tempire I think it's interesting to find where rest and esb meet
18:32 tempire this is what I was working towards: http://stackoverflow.com/questions/2383912/m​essage-queue-vs-web-services/5363360#5363360
18:33 tempire that being said, it's clear to me that mojolicious is more than just a web framework, even if that's how it's perceived
18:34 tempire I think people get caught up in what 'guaranteed delivery' means
18:35 tempire I think the best solutions are decentralized
18:36 tempire of all the java esb's, mule seemed the most sane.
18:37 tempire damn, they've corporatized the hell out of their website
18:38 sri you lost me when you mentioned enterprise :)
18:38 * sri likes his event bus minimalistic, zeromq style
18:39 tempire yeah.  though I think it's important to realize they've been doing things for what, 30+ years.  there's some lessons learned in there.
18:40 sri enterprise solutions are always wrong
18:41 tempire I don't think that's true.  it's just a problem of how to get many disparate things communicating in the most efficient manner.
18:41 sri show me a single one that's not hopelessly overengineered
18:42 tempire I'm not talking about the implementations, but the principles behind them
18:42 sri they never solve a technical problem, always management problems and have to ensure job security
18:42 sri what happened to oauth 2.0 is a good example :)
18:43 tempire again, principles, not implementation.
18:44 sri oauth 2.0 is not an implementation
18:45 sri maybe i don't understand what you mean by principles
18:47 tempire in the way I'm using the term, it is.  oauth 2.0 is an implementation of design and flow patterns
18:47 tempire I'm speaking very abstractly.  I don't talk about this stuff much because I haven't worked out good ways to communicate what I'm talking about.
18:47 * tempire digresses
18:48 sri you sound... enterprisey
18:48 tempire I know
18:49 tempire one day you'll see what I'm talking about.  and you will \o/
18:49 * sri nods enterprisey to hide that he didn't understand a single word
18:49 * tempire fails
18:49 tempire :)
18:50 sri i'd say you succeeded, that gibberish will get you through any meeting ;p
19:08 sri oh, making events and non-blocking I/O more approachable is also a good goal for mojolicious
19:08 sri although i have no clue how to get there
19:08 sri it's still such an alien concept for most perl programmers
19:19 ZadYree yup
19:39 Foxcool joined #mojo
19:47 xaka joined #mojo
20:06 trone_ joined #mojo
20:25 andrefs joined #mojo
20:32 vel joined #mojo
20:36 xaka joined #mojo
20:45 andrefs joined #mojo
21:22 amirite joined #mojo
21:36 fhelmber_ joined #mojo
21:47 lukep joined #mojo
21:57 GabrielVieira joined #mojo
22:13 preflex_ joined #mojo
22:14 tm sri, whee: +1 on redis
22:15 tm sorry, catching up with chat from the afternoon when I was afk
22:23 fhelmber_ joined #mojo
22:25 jnap joined #mojo
23:08 jontaylor joined #mojo
23:21 andrefs joined #mojo
23:25 tm hmmm, 3.12 doesn't have fastcgi mode? I guess cgi will know what to do then :D
23:27 tm is the only way now to use psgi?
23:29 tempire tm: separate plugin
23:36 tm so to deploy behind say nginx I can go daemon/morbo/hypnotoad or plack with all the bang?
23:47 GabrielVieira joined #mojo
23:47 tm never mind that ... good night :D
23:55 dvinciguerra joined #mojo

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