**** BEGIN LOGGING AT Thu Nov 13 03:00:00 2014
Nov 13 14:33:26 * DougReeder waves hello
Nov 13 14:34:00 * GodGinrai waves back
Nov 13 14:34:16 <DougReeder> For a 2-way binding, is there any event sent or hook for when the “to” end changes the value?
Nov 13 14:34:54 <GodGinrai> there should be
Nov 13 14:35:06 <GodGinrai> What do you need it for?
Nov 13 14:35:19 <Semanticfire_2> hi guys
Nov 13 14:35:24 <DougReeder> I have the bingding {from: ".person.favorite", to: ".$.header.favorite", oneWay: false },
Nov 13 14:35:25 <GodGinrai> morning Semanticfire_2
Nov 13 14:35:49 <DougReeder> … but when the “to” end changes it, personChanged is not called.
Nov 13 14:36:07 <GodGinrai> DougReeder: that's not a bindings problem
Nov 13 14:36:26 <DougReeder> Should I use some favoriteChanged() method?
Nov 13 14:36:35 <GodGinrai> DougReeder: yes.
Nov 13 14:38:15 <DougReeder> on PersonModel?
Nov 13 14:39:15 <GodGinrai> DougReeder: I guess?  I don't know your Kind structure.
Nov 13 14:39:43 <GodGinrai> Alternatively, you could also attach an observer to person.favorite
Nov 13 14:40:18 <GodGinrai> depending on what you are needing to do when person.favorite changes
Nov 13 14:41:02 <DougReeder> Since person is constantly being set, would I need to reestablish the observer every time?
Nov 13 14:41:18 <Semanticfire_2> GodGinrai, its late afternoon here ;)
Nov 13 14:41:37 <GodGinrai> Semanticfire_2: http://www.total-knowledge.com/~ilya/mips/ugt.html
Nov 13 14:42:45 <GodGinrai> DougReeder: I believe they are like bindings and will just work whenever something is actually where they are pointing. (IIRC, Bindings just use observers to detect the changes in the first place)
Nov 13 14:53:26 <DougReeder> I’m not certain this is optimal, but this does what I asked: this.person.addObserver("favorite", function (oldFavorite, newFavorite, propChanged) {
Nov 13 14:53:50 <DougReeder> … which prsumably needs to be called every time person changes
Nov 13 14:55:16 <Semanticfire_2> GodGinrai, nice, didn't know that one :)
Nov 13 15:07:06 <GodGinrai> Semanticfire_2: yw ;)
Nov 13 15:08:42 <GodGinrai> btw, is there a way in JS to include a JSON file programmatically?  Or do I need to use XHR to load it?
Nov 13 15:14:06 <Semanticfire_2> what do you mean ?
Nov 13 15:15:52 <GodGinrai> Basically, I need to load some JSON that will be in the same directory structure as the code that's running.  (same server, in other words)
Nov 13 15:16:05 <GodGinrai> But I need to be able to choose when to load it
Nov 13 15:16:11 <GodGinrai> and what particular files to load
Nov 13 15:19:36 <Semanticfire_2> ah like that
Nov 13 15:20:19 <Semanticfire_2> well if you need to get it from the server there is no other option
Nov 13 15:20:26 <Semanticfire_2> since the browser needs to fetch it as well
Nov 13 15:21:06 <Semanticfire_2> if the set of choices you want to choose from is static you might work with a variable and load all the scripts at once
Nov 13 15:21:18 <Semanticfire_2> but its not json file then
Nov 13 15:21:41 <Semanticfire_2> whats your issues with xhr ?
Nov 13 15:21:50 <DougReeder> The above works, but the observer needs to be added every time person changes.  Currently, I’m trying { name: "header", kind: "ContactHeader", favoriteChanged: function (oldFavorite, newFavorite, header) {
Nov 13 15:22:28 <DougReeder> … that is, observe the other end. The header object doesn’t change, so the observer can be added once.
Nov 13 15:22:46 <GodGinrai> DougReeder: if you just need something to key off of whenever it changes, why not use a transform function in the binding?
Nov 13 15:23:32 <DougReeder> Also a possibility, but that would be a side effect, and it seems wrong for transforms to have side effects.
Nov 13 15:23:39 <GodGinrai> Semanticfire_2: My issue with XHR is that it is a lot of setup for what other languages accomplish with include(file) or require(file)
Nov 13 15:24:29 * DougReeder shrugs “But that might be less smell than other techniques.
Nov 13 15:25:06 <GodGinrai> DougReeder: Sure, technically a transform function should only transform the data, but it seems like the cleanest way to solve your problem (aside from writing in a changed method for your Person Model)
Nov 13 15:25:08 <Semanticfire_2> GodGinrai, but other languages don't get it from a remote filesystem :)
Nov 13 15:25:48 <DougReeder> And the transform method knows which way the data is flowing.
Nov 13 15:26:02 <GodGinrai> Semanticfire_2: Sure, but this is the same remote filesystem I already got the current files from.  HTML can say it easy enough with <script> blocks, why can't JS do the same?
Nov 13 15:29:18 <DougReeder> GodGinrai , you could add a <script> tag to the page when you want to load, with a type of text/json or whatever.
Nov 13 15:30:02 <DougReeder> … or application/json if you follow RFC4627
Nov 13 15:30:21 <GodGinrai> DougReeder: That's basically as much work as using XHR
Nov 13 15:30:36 <DougReeder> The browser won’t do anything with it, but you can grab the innerHTML and parse it yourself.
Nov 13 15:32:04 <DougReeder> Why can’t you load it with the code? Is it a large amout of data?
Nov 13 15:33:00 <Semanticfire_2> yeah I'm failing to see that as well :)
Nov 13 15:33:05 <GodGinrai> DougReeder: They are config files that tell the app what to do.  The config file that needs to be loaded changes depending on what we need.
Nov 13 15:33:41 <DougReeder> Is the data small enough you could load all of them and only use the one you want?
Nov 13 15:33:56 <DougReeder> Does the data change after the code has loaded?
Nov 13 15:34:53 <GodGinrai> DougReeder: While that could possibly work when deployed to production, (where one single "app" gets deployed) the load times during testing would be crap and waste our time
Nov 13 15:35:27 <Semanticfire_2> that what I suggested as wel, put the JSON in variables and in one file and select a variable
Nov 13 15:35:38 <DougReeder> If you want to load data that can’t be known until after the code is running, XHR is the way to go. Many libs can take card of the boilerplate.
Nov 13 15:36:31 <GodGinrai> Semanticfire_2: can't do that.  The config files need to be separate because they get deployed depending on the "app" that we are deploying.  This is an engine that reads these config files to work as an app for specific companies
Nov 13 15:37:53 <Semanticfire_2> so there is a specific app for a company ? or a generic app which load a specific set of config files ?
Nov 13 15:38:31 <DougReeder> At the network level, either your config data is one or part of the HTTP requests during page load, or an HTTP request after page load.
Nov 13 15:39:09 <DougReeder> …and the xfer is what slows you down (usually)
Nov 13 15:39:45 <GodGinrai> Semanticfire_2: Generic engine that uses the config files to build the interface described by them
Nov 13 15:40:38 <Semanticfire_2> then xhr is the way to go
Nov 13 15:41:08 <GodGinrai> damn :\
Nov 13 15:42:04 <Semanticfire_2> why is that so bad ?
Nov 13 15:42:25 <Semanticfire_2> is the config data large ?
Nov 13 15:42:55 <DougReeder> GodGinrai , can you use your app build process to copy a specific config file into a standard location?
Nov 13 15:44:53 <GodGinrai> DougReeder: for testing, we have an interface which allows us to load any of the configs on the fly.  So once again, that doesn't work for testing
Nov 13 15:45:51 <GodGinrai> Semanticfire_2: It's not about how large the data is.  It's about having to populate objects and write handlers just to include a damn file. <.<
Nov 13 15:47:35 <Semanticfire_2> you chose the wrong platform :)
Nov 13 15:48:16 <Semanticfire_2> there is plenty of jquery abstractions for it
Nov 13 15:49:34 <Semanticfire_2>  $.get( "<http://jsonstuff>", function( data ) { do stuff with jsondata });
Nov 13 15:50:35 <Semanticfire_2> and to prevent loading it everytime you might as well push it into local storage, saves you a server roundtrip
Nov 13 15:50:48 <GodGinrai> Semanticfire_2: nothing I can do about it.  I tried pushing Enyo and no one cares
Nov 13 15:52:20 <DougReeder> require() is synchronous.  It doesn’t seem to me that a synchronous XHR request is much more complicated than require().
Nov 13 15:53:54 <GodGinrai> DougReeder: We have to support old versions of IE, so synchronous XHR requests are out the door.  Besides, that's still more code than a single line for require
Nov 13 15:54:57 <DougReeder> Well, with a requirement to support old IE, simplicity is no longer a possibility.
Nov 13 15:55:31 <DougReeder> Simple, Fast, old IE: choose one.
Nov 13 15:56:17 <GodGinrai> <.<
Nov 13 15:57:52 <DougReeder> Wrap the XHR in your own require() function.
Nov 13 15:59:12 <GodGinrai> Yea... -_-
Nov 13 15:59:19 <DougReeder> In old IE, just spin until the async XHR returns.
Nov 13 15:59:21 <chall> you can always shim it just for old IE
Nov 13 16:00:17 <chall> as an aside, how old are you talking?
Nov 13 16:00:26 <chall> IE 6?
Nov 13 16:00:50 <GodGinrai> no, not 6 at least
Nov 13 16:00:53 <GodGinrai> maybe 7?
Nov 13 16:00:57 <GodGinrai> at least 8
Nov 13 16:01:04 <chall> that doesn't seem too bad, really
Nov 13 16:01:22 <chall> I don't recall IE 8 being that hard to support
Nov 13 16:01:36 <GodGinrai> yea, it's not TOO bad
Nov 13 16:01:52 <GodGinrai> but doesn't support sync xhr iirc
Nov 13 16:02:30 <chall> according to the Google search I just made, it should
Nov 13 16:03:28 <chall> the browser might freeze while it's doing it's thing, but it should work
Nov 13 16:03:34 <chall> its*
Nov 13 16:04:34 <chall> I'm not sure why you'd want to do it sync anyways, but, then, I'm a fan of evented programming
Nov 13 16:11:22 <GodGinrai> chall: Because I'm not trying to do anything asynchronous here.  I just want to require data
Nov 13 16:11:39 <GodGinrai> chall: This is all because javascript doesn't have a simple include() or require()
Nov 13 16:47:46 <dmanderson> Morning Enyos
Nov 13 16:50:48 <GodGinrai> morning dmanderson
Nov 13 16:50:53 <Semanticfire_2> ey dmanderson
**** BEGIN LOGGING AT Fri Nov 14 01:15:17 2014
**** ENDING LOGGING AT Fri Nov 14 03:00:00 2014