Camelia, the Perl 6 bug

IRC log for #mojo, 2013-05-28

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

All times shown according to UTC.

Time Nick Message
00:16 sri the whole keep CGI.pm discussion is kinda funny
00:16 sri no new standard will emerge, ever
00:17 sri the web has changed, and keeps changing quickly
00:27 jberger I think (as I said in my blog post) that having certain web-inspired modules in the core is probably a good idea
00:27 jberger param parsing and entities etc
00:27 jberger things that all frameworks would benefit from
00:28 jberger I think its funny that one of the arguments I'm seeing more than expected is that newbies benefit from the presence of CGI.pm??!?
00:28 sri fast percent encoding in core would have been nice
00:29 sri entities i dunno, they keep changing
00:29 jberger really?
00:29 jberger but you get the idea, webby utility modules, not a "framework"
00:29 sri they are defined in the html5 spec now, which is being changed all the time (html5.1 now)
00:30 jberger is there a subset which is reasonably stable?
00:31 sri sure, but still, browsers are iterating so quickly, you could never keep up with anything html5 in core
00:31 sri getting stuff like ipv6 finally into core is much bigger imo
00:32 sri (more basic building blocks for web frameworks)
00:33 jberger ipv6 in the core seems like a no-brainer
00:33 sri base64 being in core is great, punycode would also be nice
00:34 jberger yeah, building blocks, exactly
00:34 sri param parsing is already a bit high level
00:36 sri re entities... fast escaping for &, <, >, " and ' would be nice, that's something everybody needs
00:38 sri then again, some things already in core are still pretty weak, such as monotonic clock support
00:39 sri or Sub::Name functionality not being in Scalar::Util
00:46 jberger I think what is clear to most people is that you should not encourage new projects with CGI.pm and that is what you do by having it in the core
00:46 sri i actually have a counter argument ;p
00:46 sri even though i agree...
00:46 jberger and what's worse, when newcomers see it as "the way" in Perl, versus Sinatra or Tornado or something, they don't stay with Perl
00:47 sri cgi.rb is in ruby core!
00:47 sri perl community is different though, much more fragmented
00:50 jberger TIMTOWTDI is good and all, but can't we encourage something from this decade? and if not from the previous one? why go back two decades??? its like encouraging writing new projects in COBOL
00:50 jberger (and yes I saw the hosted COBOL thingy :-) )
00:50 shmuel joined #mojo
00:53 sri to be fair, newer tech is already encouraged http://perldoc.perl.org/perlfaq9.html
00:55 jberger that is, if they make it to perlfaq9
00:56 jberger most find a PERL tutorial online
00:56 jberger if those users started a PERL CGI tutorial and went to use it and it wasn't there, they might investigate whether is was a good idea to continue
01:03 egopro joined #mojo
01:32 btyler joined #mojo
01:33 hesperaux__ joined #mojo
01:47 d4rkie joined #mojo
01:56 Meiermann joined #mojo
02:25 basiliscos joined #mojo
02:38 inokenty joined #mojo
02:45 cooper joined #mojo
03:32 fildon_ joined #mojo
03:50 preflex_ joined #mojo
03:55 asarch joined #mojo
04:29 zivester joined #mojo
05:21 sh4 joined #mojo
05:27 jakir joined #mojo
05:32 marcus ASS THROUGH NUTS
05:32 marcus http://www.indiegogo.com/pro​jects/rob-ford-crackstarter
05:49 Britzel_ joined #mojo
05:59 davido joined #mojo
06:12 suy joined #mojo
06:16 sivoais joined #mojo
06:34 dotan joined #mojo
06:44 dod joined #mojo
06:50 yakudza joined #mojo
07:02 mire_ joined #mojo
07:06 dod1 joined #mojo
07:10 bc547 joined #mojo
07:14 dpetrov_ joined #mojo
07:30 nicolaas joined #mojo
07:50 jzawodn joined #mojo
07:56 Vandal joined #mojo
08:07 cosmincx joined #mojo
08:15 dod joined #mojo
08:32 njlg joined #mojo
08:38 nic joined #mojo
08:51 fhelmber_ joined #mojo
08:53 dod joined #mojo
09:23 dod joined #mojo
09:25 reneeb joined #mojo
09:28 JiYu joined #mojo
09:31 Vandal joined #mojo
09:36 maxhq joined #mojo
10:12 nelio joined #mojo
10:16 rem_lex|pivo joined #mojo
10:16 Vandal is there a way to chain a bridges? something like: my $bridge = $r->bridge('/user')->to('user#auth​')->to('foo#base')->to('bar#baz'); ?
10:31 kitt_vl joined #mojo
10:55 ver joined #mojo
11:03 nicolaas Vandal maybe using "redirect_to" in your chainable methods ?
11:03 Vandal what?
11:04 nicolaas Vandal for example, in your User controller "auth" method, you can do stuff and then call $self->redirect_to('/Foo/base');
11:05 nicolaas and in Foo base method, you can do the same to /Bar/baz
11:05 Vandal nicolaas, and all stuff from auth method disapeard
11:05 nicolaas and the baz method do the "$self->render"
11:06 Vandal you kidding, right?
11:06 nicolaas no
11:07 nicolaas here is my bridge method : http://pastebin.com/NvWWNQEu
11:08 nicolaas after that, I define routes from the bridge (instead of "routes" object)
11:08 nicolaas like this : $b->route('/:controller/:acti​on')->via('GET','POST')->to( ... )
11:08 nicolaas you can use "redirect_to" to chain methods
11:08 nicolaas or "render(partial=>1)"
11:09 nicolaas if you want to compose your render content
11:10 nicolaas the content to render, sorry. In this case, User auth & Foo base make a partial render
11:10 Vandal you saying nonsense
11:10 nicolaas well, I might have misunderstood your problem then
11:12 nic nicolaas++
11:14 Vandal http://pastebin.com/cccsfyqA
11:14 kwa nicolaas: Doesn't redirect_to actually perform a 302 redirect? I'm not sure you want to do that to chain methods.
11:15 Vandal kwa, exactly
11:15 nicolaas kwa you're probably right, I didn't check actually.
11:16 nicolaas if Vandal wants to print the content generated from the code of several methods, I think partial rendering is the best way, but maybe he doesn't want that at all .. in this case, sorry I didn't get the right idea
11:17 Vandal nicolaas, here is wat I want: http://pastebin.com/cccsfyqA
11:17 Vandal I Don't want to Repeat mY self
11:18 Vandal so I thought that there is may be esy way to chain bridges
11:18 Vandal *easy
11:18 nicolaas Vandal yes but what do the user#auth, foo#base and bar#baz methods do ? do they build something to render ?
11:18 kwa Vandal: Might be better explaining why you want to do that. Then I could at least offer up an alternative.
11:19 Vandal nicolaas, they filling up the Stash with data
11:19 Vandal kwa, what do you mean?
11:21 kwa Well, depending on what the additional chained methods do, you might want to refactor the code into plugins which can be called from a single bridge to build up the stash.
11:23 nicolaas Vandal I think I guess what you mean, but I don't know if there is another way to do it better than writing all your calls to user#auth, foo#base & bar#baz in a single bridge callback
11:24 nicolaas like this : http://pastebin.com/bjnkGWWf
11:24 Vandal kwa, I hate Plugins, I already have controllers
11:24 nicolaas (must add the bar baz method though)
11:26 Vandal nicolaas, it is even more code than as is now
11:26 nicolaas Vandal yes but in one bridge. I don't know if there is a better solution, I'm interested in that answer too now
11:27 kwa Vandal: Fair enough, but as I'm not sure what you're trying to achieve, I cannot offer an alternative. I can't personally see the benefit of chaining multiple routes just to build up the stash. Sounds like you might want to be doing something else.
11:30 nicolaas I think a bridge only returns 1 or 0, so you can't "chain" them like you asked.
11:31 nic I want my toaster to toast my bread without me having to slice it
11:31 nic I hate slicing bread
11:31 Vandal kwa, I have controllers User, Category, Pages, Cart, Item and to render page I need to fill stash with data about User, list of Categories, list of Pages, what in Cart and how many, list of Items and for of all this I need methods in respected controller
11:32 nicolaas Vandal that's why you have to use the cb method of the bridge, and put it all in it
11:32 kwa nicolaas: If a bridge returns a positive value, the chain continues, if 0 or undef it will break the chain and attempt to render.
11:32 nicolaas kwa oh okay, my bad
11:34 kwa Vandal: Any you want to fill the stash with this information for a specific route, or every request?
11:34 Vandal kwa, for a specific route
11:35 Vandal some routes need Pages some dont
11:35 Vandal some routes need Items some dont
11:35 Vandal etc
11:52 kwa Vandal, gotcha.
11:52 kwa In that case, I think you might want nested routes: http://mojolicio.us/perldoc/Mojoli​cious/Guides/Routing#Nested_routes
11:54 Vandal kwa, nope, I have different Controllers and it dosent make less code
12:00 kwa Vandal, I'll have a think about it. I can see where you're coming from, though I generally build the stash in the auth bridge once I've authenticated them. Then in nested routes I load additional information from the DB which isn't relevant to the majority of requests.
12:00 inokenty left #mojo
12:10 kwa Ahaa interesting! Vandal, you can used nested bridges to do what you want.
12:10 kwa Vandal: https://github.com/kraih/mojo/blob/​master/t/mojolicious/routes.t#L141
12:17 KindTwo joined #mojo
12:19 SmokeMachine joined #mojo
12:20 libsysguy joined #mojo
12:33 mire joined #mojo
12:33 nicolaas joined #mojo
12:36 Vandal kwa, this is a little better, thank you :)
12:36 kwa No probs. :)
12:46 libsysguy are there any good tutorials on sql database models?
12:47 libsysguy I see in the growing guide there is an example of using a perl module as a model
12:50 kwa libsysguy, do you use DBI (or its derivatives) or DBIx::Class?
12:52 kwa libsysguy, check out the wiki in any case: https://github.com/kraih/mojo/wiki
12:54 mak joined #mojo
13:10 SmokeMachine joined #mojo
13:30 moltar joined #mojo
13:31 zivester joined #mojo
13:39 kej joined #mojo
13:40 inokenty joined #mojo
13:43 btyler joined #mojo
13:57 libsysguy thanks kwa
13:58 libsysguy kea I was using DBIx::Class in my catalyst project
13:59 libsysguy but I was going to see if I could work on a mojo derivative of it as well
14:00 Vandal mojo derivative of what?
14:00 libsysguy just a project I'm working on
14:00 libsysguy I was going to try and develop it in parallel with catalyst and mojo
14:01 d4rkie joined #mojo
14:15 kwa libsysguy: If you're using DBIx::Class then check tempire's awesome example of using DBIx::Class and Mojolicious -> https://github.com/tempire/MojoExample
14:19 libsysguy ah ha, yeah that is the one I was going off of
14:19 dpetrov_ joined #mojo
14:20 libsysguy so all you have to do to load the model is load that helper from startup?
14:20 libsysguy thats pretty nice
14:24 sh4 joined #mojo
14:31 dod joined #mojo
14:36 doby joined #mojo
14:40 libsysguy can I ask what voodoo makes has, around, under, etc work in mojo?
14:41 libsysguy it looks like Moose, but I can't find a use statement for it
14:43 abra joined #mojo
14:44 bluescreen joined #mojo
14:53 sri http://mojolicio.us/perldoc/Mojo/Base
14:56 libsysguy ah okay
14:56 libsysguy I thought it was Mojo::Base but I wasn't looking at the functions :p
14:58 libsysguy I'm surprised it wasn't just called Mo :p
15:01 sri it has a long history, started out completely different
15:01 libsysguy oh? is it on a current track of stability or is it still in flux?
15:02 sri it's much older than all the Moose alternatives that have popped up recently
15:02 libsysguy I really like the functionality it provides
15:02 libsysguy namely Moo :p
15:02 sri depends on your definition of stability
15:03 sri recent additions have been the inclusion of "use utf8" for example
15:03 sri has for example didn't change in years though
15:04 libsysguy under is new?
15:04 sri under is not in Mojo::Base
15:04 libsysguy disregard that
15:04 libsysguy lol
15:22 mire joined #mojo
15:36 kwa sri, I know you said it was by design, but I read the docs all weekend and the rudimentary test I posted on pastbin re: respond_to really looks like a bug (http://pastebin.com/CzqZGNYt). If it's not could you please explain why?
15:38 sri kwa: try making it work differently and watch tests fail :)
15:40 kwa sri, so it's not a bug, or it's a right pain in the bum to get to play nice? :/ :)
15:41 kwa I've had to apply the following hack in my bridge to get respond_to to do what I want: $self->stash(format => $self->match->stack->[1]->{format}) if $self->match->stack->[1]->{format};
15:41 kwa Is it fine inspecting the stack like that?
15:46 good_news_everyone joined #mojo
15:46 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/ke33iw
15:46 good_news_everyone mojo/master f7afcd5 Sebastian Riedel: reuse HMAC functionality from Digest::SHA
15:46 good_news_everyone left #mojo
15:47 sri missed that when removing HMAC-MD5 in 4.0 :)
15:47 MeierM joined #mojo
15:47 kbenson joined #mojo
15:49 Akron joined #mojo
15:52 libsysguy routing question, whats the best way to do it?
15:52 libsysguy do I have to define all my routes in startup?
15:52 kwa Technically no, but generally, yes.
15:52 libsysguy that could get...large
15:52 Akron The only perl project in GSoC is "Port the test suite of wget from Perl to Python, because the C-Programmers of wget are confused by Perl". wtf?!
15:53 good_news_everyone joined #mojo
15:53 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/mSDzzw
15:53 good_news_everyone mojo/master 5d3f2fd Sebastian Riedel: removed a few unnecessary hmac_sha1_sum tests
15:53 good_news_everyone left #mojo
15:53 kwa libsysguy: Read http://mojolicio.us/perldoc​/Mojolicious/Guides/Routing - using placeholders you can generally do what you want in only a few routes.
15:53 sri kwa: not a bug, if you can make it more dwim, more power to you... but i doubt it
15:53 sri kwa: yes, inspecting the frames like that is fine
15:55 sri kwa: don't forget that many stash values are special and used by the renderer for example, if you merge them too early things explode
15:55 kwa sri: Fair enough. It just seemed weird that if I passed format=txt|json via the query string it did what I wanted.
15:56 sri and even non-special ones might affect helpers in existing apps if they get merged differently
15:56 kwa (As opposed to just adding .json|.txt for respond_to.)
15:56 kwa Ahh gotcha.
15:57 sri well, format is even more special
15:57 libsysguy It looks like routing is very different in Mojo::Lite as opposed to Full
15:57 tempire libsysguy: same routing, different syntax
15:57 KindOne joined #mojo
15:57 libsysguy heh
15:58 libsysguy so Mojo::Lite employs the Mojo::ForTheLazy module :p
15:59 sri kwa: but i doubt special casing format would work well... it can still mess up render/respond_to calls if you're expecting the current behavior
16:00 sri since only an endpoint can pick up format it might be an option to do it, but there would very likely be breakage
16:01 sri kwa: think of it as the current route picking up stash values right before your action runs, not in advance for all of them
16:02 kwa sri: I need to make sure respond_to returns what the client is expecting if I render an error. The same bridge supports json and html requests, I wasn't getting json on errors when testing. I'm really forced to apply the hack in the bridge in this case.
16:03 sri like i said, i'm open for proposals, but it won't be easy
16:04 kwa One thing I've noticed after all the reading, is that mojolicious likes to exit the dispatch chain. Whereas before the digging through the docs I wanted to detach via plugins like I would do in Catalyst. Both different beasts.
16:04 kwa ...exit it gracefully.
16:05 sri detach in catalyst is a hack really, i don't like using exceptions for flow control
16:05 kwa sri: It's fair enough, I understand the reasons a bit more now, thanks. It just wasn't immediately obvious when I was trying to get the bridge to DWIM.
16:06 sri i also don't like features that wouldn't work async
16:06 kwa Yeah, just adding a return before my API calls satifies the flow control. Just a change in mindset.
16:06 sri detach in a timer callback for example would be an event loop killer
16:10 Vandal joined #mojo
16:11 kwa True. Cheers for your time.
16:12 * kwa runs home
16:27 bluescreen_ joined #mojo
16:41 libsysguy is it possible/recommended to add routes in the controller?
16:42 libsysguy apologies if I
16:42 libsysguy I'm asking too many questions
16:49 jberger libsysguy, you might be interested in this: http://blog.kraih.com/post/43198653667​/mojolicious-hack-of-the-day-mojolyst
16:50 libsysguy very helpful
16:50 libsysguy jberger++
16:50 libsysguy hey…jberger I think you're the one that got me interested in looking at mojo in my reddit post
16:51 jberger oh hi
16:51 libsysguy :)
16:51 jberger I'm trying to be more proactive in promoting Mojolicious, that has included me monitoring reddit for places to do so :-)
16:52 libsysguy well all I have ever known is catalyst
16:52 jberger I can imagine its quite a step
16:52 libsysguy and I'm not allowed to use it at $work
16:52 libsysguy so mojo was the next logical alternative…I am really digging the async stuff
16:52 jberger not allowed to use which ?
16:53 libsysguy and yes, there has been much headdesking this weekend
16:53 libsysguy not allowed to use Catalyst
16:53 jberger too heavy?
16:53 libsysguy the overhead of Moose is too large
16:53 jberger understood
16:53 libsysguy precisely
16:53 jberger yeah, Mojo::Base is good for most things
16:53 libsysguy I've just gotten spoiled with syntactic sugar
16:53 jberger I've been using Moo+Type::Tiny lately
16:53 jberger for bigger things
16:54 jberger async is the good stuff though
16:54 jberger its worth the pain once you get through it
16:54 libsysguy well here at $work all our projects are written in straight CGI
16:54 libsysguy wheee
16:54 jberger :-)
16:55 libsysguy well I keep telling myself there was a steep learning curve for Catalyst as well
16:55 libsysguy I should expect no less from Mojo, but screencasts++
16:55 jberger have you read Mojolicious::Lite and Mojolicious::Guides::Growing/
16:55 jberger ?
16:56 jberger those two really make it rather clear, especially after letting it sink in for a while
16:56 libsysguy I have to some degree, I've been more focused on the growing guide
16:56 jberger you really should read Lite too
16:56 jberger its kinda the "tutorial"
16:56 jberger growing assumes you have read it already
16:56 libsysguy I think if I would have read the lite guide routing would have been less painful
16:57 jberger we recommend you read Lite and the Guides in this order: http://mojolicio.us/perldoc#TUTORIAL
16:58 libsysguy all caps for dummies that want to skip through it eh?
16:58 libsysguy :p
16:58 jberger READ ALL THE THINGS~!!!
16:58 jberger :-)
17:02 rem_lex| joined #mojo
17:03 basiliscos joined #mojo
17:05 libsysguy that is one majorly nice thing about Mojo is the documentation
17:06 tempire ^ sri
17:06 tempire how times have changed!
17:06 libsysguy heh
17:08 libsysguy is anybody presenting any Mojo related stuff at YAPC this year?
17:09 basiliscos1 joined #mojo
17:10 * jberger would have loved to, but can't make it to yapc this year :-(
17:11 jberger tempire are you going?
17:11 tempire I am not
17:11 tempire I enjoyed giving the talk
17:11 tempire but it wasn't quite my crowd of people
17:12 jberger hey! I was in that crowd :-P
17:12 mrphilov joined #mojo
17:12 tempire it's true. I went to your talk.
17:12 jberger I remember
17:12 libsysguy I don't think tempire appreciated your heckling jberger
17:12 tempire I remember people pretending they understood what you were talking about.
17:12 jberger at least you were at the XS one
17:13 jberger people didn't follow THAT ONE??!
17:13 jberger oh man, then the really must not have gotten anything from my sciencey talk!
17:13 tempire It was the sciencey talk that I went to
17:13 tempire as I recall
17:13 jberger oh
17:13 jberger yeah, I wouldn't be surprised if no one understood a word of that
17:14 tempire someone else gave a statistics talk
17:14 * jberger has a memory that likes to play tricks on him
17:14 sri tempire: :)
17:14 jberger yeah, Maggie gave a talk on PDL-Stats
17:14 tempire was she an asian girl from shutterstock?
17:14 jberger yeah
17:15 sri jberger: i love the idea of type checks... but they are soooooooooooo expensive :/
17:15 jberger http://perl4science.github.io/videos/
17:15 tempire people tried to pretend they understood that one as well :)
17:15 jberger sri: I agree absolutely
17:16 tempire I'll go again if I have an amazing talk to give.
17:16 tempire But not just to consume
17:17 jberger sri: I remember the hit that Mojo::Base took when I tried to add even just BUILD/BUILDARGS support
17:17 jberger tempire, I think we need to do better promoting the async/non-blocking features of mojo
17:17 libsysguy ^^
17:17 jberger I get asked rather often about the differences between Dancer and Mojo
17:17 tempire I'm in the midst of writing a fully non-blocking app right now
17:18 tempire when I finish, I might put something together
17:18 libsysguy this may just be me, but I love full app examples
17:18 libsysguy foss++
17:18 jberger tempire's talks are gorgeous!
17:18 jberger libsysguy, they are nice, but hard, because you essentially have to show several files
17:19 jberger makes it harder to follow in examples
17:19 jberger but I personally use Full for everything
17:19 jberger https://github.com/jberger/Gali​leo/blob/master/lib/Galileo.pm
17:20 libsysguy I see some advantages of using Lite, I just can't stand the thought of template data in my files
17:21 sri jberger: yea, async is hard to explain though
17:21 jberger libsysguy, agreed, however, sometimes its nice to have template data in the same files as plugins
17:21 libsysguy I could see it for plugins
17:22 rihegher joined #mojo
17:22 libsysguy I like your reasoning for jQuery usage
17:23 jberger ?
17:23 jberger "because everything uses jQuery" hehe
17:23 libsysguy "http://jquery.com/ - because everything uses jQuery"
17:23 jberger took me a moment
17:23 jberger :-0
17:23 libsysguy yeah, its funny because there has been some heated debate as to what js library to use here at $work
17:24 libsysguy YUI vs jQuery holy war
17:24 preaction Angular! (which uses jquery)
17:24 libsysguy YUI is winning :-(
17:24 abra joined #mojo
17:24 jberger preaction, is doing some cool stuff with angular
17:24 * jberger show learn more about it
17:25 jberger s/show/should/
17:25 * tempire uses backbone
17:25 preaction last i used YUI was v2, and it was great for its time, but the UI set leaves a lot to be desired for skinning
17:26 libsysguy agreed, they are on v3 now and imo it hasn't gotten much better
17:26 preaction which leaves just the core libraries, which are comparable. but Angular uses jquery, and Angular is some next-level stuff
17:27 preaction really, the choices you should evaluate are: Ember, Backbone, and Angular. I think Knockout is another one
17:28 libsysguy I thought it was just between backbone and angular
17:28 preaction Ember is the evolution of Sproutcore, which was some really pretty stuff
17:32 preaction huh. just did the first 5 minutes of the knockout tutorial. some really cool stuff there
17:33 phillipadsmith joined #mojo
17:43 libsysguy it seems like there is some crossover between stash and render
17:43 libsysguy would that be an incorrect assumption?
17:50 chansen joined #mojo
18:01 rihegher left #mojo
18:04 btyler joined #mojo
18:05 dpetrov1 joined #mojo
18:19 good_news_everyone joined #mojo
18:19 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/TzFAgQ
18:19 good_news_everyone mojo/master 790e80e Sebastian Riedel: better HMAC-SHA1 tests
18:19 good_news_everyone left #mojo
18:22 sri libsysguy: nope
18:22 libsysguy yayy I'm starting to get it
18:22 libsysguy :p
18:23 sri it's why stuff like this works :)
18:23 sri perl -Mojo -E 'a({text => "Hello Mojo!"})->start' get /
18:23 libsysguy dat one-liner
18:32 BeDa joined #mojo
18:35 whitebook joined #mojo
18:47 SmokeMachine joined #mojo
18:53 gtodd libsysguy: if you are starting to get it then your understanding will reveal to you how Mojo is able to get itself
18:53 asarch joined #mojo
18:54 alester joined #mojo
18:54 gtodd perl -Mojo -E 'a({text => "I got myself \n"})->start' get /
18:55 libsysguy so basally that starts the mojo embedded server then makes a get request to itself and prints the output
18:58 btyler joined #mojo
19:01 yko so noone of mojo core goes to yapc this year?
19:03 libsysguy i see mst in here, I know he's going, but I don't know if he's core
19:27 bluescreen joined #mojo
19:43 fhelmber_ joined #mojo
20:24 nelio joined #mojo
20:39 mire joined #mojo
21:03 Meiermann joined #mojo
21:09 Tester123 joined #mojo
21:09 sh4 joined #mojo
21:11 Tester123 Cygwin install error: Makefile:1605: recipe for target `test_dynamic' failed - Full log: http://pastebin.com/cTT2L2Qm
21:12 Tester123 Any Ideas on this error?
21:21 sri it's cygwin, sockets are always broken somehow there
21:37 bpmedley joined #mojo
21:52 lukep joined #mojo
22:07 mire joined #mojo
22:11 inokenty joined #mojo
22:43 russum joined #mojo
22:44 sri kwa: re content negotiation in bridges, this patch appears to break nothing http://pastie.org/7977181
22:44 sri no idea if special casing "format" is a good idea though
22:48 sri and just because it doesn't break existing tests doesn't mean apps won't break that depend on the current behavior
22:53 sri jberger, marcus, tempire, crab: maybe you should have a look at the patch too, the tests make the problem pretty clear
22:57 jberger I'm not sure that I have any opinion
22:58 sri think i have an idea for how to explain it in documentation, in case we decide to do it
23:07 hesperaux joined #mojo
23:10 jberger is the issue that the format is stashed earlier?
23:10 jnbek joined #mojo
23:10 sri the other way around
23:10 jberger *possible issue
23:11 sri or do you mean "issue of the patch"
23:11 jberger yes, the issue of the patch
23:11 sri i don't know if there's an issue
23:12 sri doubt it though
23:13 jberger I wouldn't think so either
23:13 jberger by the time most things get to us, the format is stashed anyway
23:13 sri oh wait
23:13 sri i missed a problem this doesn't address
23:14 sri embedded routes behind bridges :)
23:14 sri the format is not available yet at the time of the first dispatch cycle
23:14 * jberger hides head in the sand
23:15 yakudza joined #mojo
23:15 sri umm... embedded apps of course
23:15 sri no way to ever make that work though
23:26 asarch joined #mojo
23:34 sri my explanation sounds something like "Stash values will be merged with each dispatch cycle and if possible a C<format> value will be set before the first."
23:34 sri with a format bridge example below
23:34 sri in the bridge section of the routing guide
23:37 whitebook joined #mojo
23:41 sri http://pastie.org/7977357 # ok, here's the full patch with docs
23:56 sri well, it's still inconsistent... so i won't merge it unless it gets a few votes

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