I am developing a widget, and am using extJS framework (along with stomp,orbited servers). It requires 5 different Javascript files. To avoid impacting load time for users, I am trying to load them asynchronously.
Through some callback stacks, I have managed to asynchronously load them in order they are required.
However, the widget does not work at all in the asynchronous case. I then copied the "modified" resultant HTML DOM from Chrome's developer tools, and created a static HTML file.
The static file works. I am printing few alerts when different JS files are loaded in both cases. The order of these alerts is the same.
How do I detect and fix the error. Stumped after lot of debugging.
(Link removed as they were on production server. Managed to detect the issue and find a workaround - please see answer below - keeping for anyone's reference).
The problem is that Ext.onReady is called as soon as document is ready, which is BEFORE any other scripts are loaded asynchronously.
Thus, the initialization code was not getting called. I manage to work it around by calling Ext.each("dummy args", initialization_function, this) in the last javascript that is loaded. This made things work for me.
Related
EDIT: Is there anywhere to get an un-minified version of the Here-api to use when debugging? It's impossible for me to figure out what 'v' is and why it may be undefined.
We're using the HERE API both from our website where it works flawlessly and our old RDP C++ application which runs a similar webpage in an embedded IE window. It should be using a stripped version of IE11 I believe.
We recently upgraded to the new HERE API after routing stopped working in the old one, and it worked for a while but a while ago it suddenly didn't. And no one can recall making any changes that could affect this.
I have narrowed it down to a single line of code where it crashes. (platform is already defined in the scope through our geo-service script, the same one being used for the web that works)
var defaultLayers = platform.createDefaultLayers();
This is an initialization of the map layers that is required for the maps to work, but we simply can't perform this action through this embedded browser window even though we run almost identical code on the web.
We receive two error messages of:
'v' is undefined
With a reference to some dynamically generated eval code.
This is the only lead I've managed to dig up, it's not much but I'm hoping someone else has encountered a similar issue and can point me in the right direction what to look for.
I found the issue, it was totally self inflicted...
When we implemented the solution there was an issue in the core js file from Here which made it not work in our servers due to some path being wrong. To fix this we had changed the path so it worked and then hosted our own version of the core file.
This worked great, until Here put up a new minor release which is automatically distributed through the same content link as before. This meant the minified files were no longer in sync with the variable names, thus causing v to never be defined where it should, since in our file it was probably named something else.
It was only by chance I noticed that the core js was side-loaded like that, I was looking in the completely opposite direction the entire time and didn't even consider the loading might've been tinkered with.
We have quite a large MVC4 application and we would like to have Selenium go through every page and make sure it loads - some sort of smoke test.
I can use reflection to go through the assembly, find all controllers and all actions, check if actions are not post, come up with parameters for actions that require parameters.
Then I'll feed this list to Selenium and check that everything I need on the pages is done appropriately.
But before I start playing with reflection, I'd like to check if this has already been done, so I don't reinvent the bicycle. I have googled for such thing, but could not find anything.
p.s. Writing the reflection code is not an issue. Selenium is covered as well. Just checking if this has already been done.
The AttributeRouting project has a route debugger in place, which does work even if you don't use attribute routing inside your project.
You can see the class that handles displaying the routes over on Github but I'm not sure it will display the routing information when the project isn't run locally. You may need to adapt that code so you can access it safely from your Selenium instance (and make it machine readable using JSON or something).
This is my first attempt at working with Angular.js after using Backbone.js for some projects. I used Yeoman to get the initial angular scaffolding setup as well as to manage various dependencies etc.
The first issue I've run into is with the angular $resource service I setup to hit an external API which returns a simple array of names. In using the Chrome web inspector I see that the data is returned as expected but never renders. I suspect this is due to the call by asynchronous and when the page renders there is no data to show. Shortly after the page renders the data is returned from the API but the page never re-renders again.
The Angular docs state pretty clearly in the tutorial that this should happen automagically via the angular data binding but so far it's never worked for me.
I put the code I'm experimenting with now up in a repo on GitHub at https://github.com/chad-skidmore/angular-test
I've not found any documented issues so far regarding this issue after spending the last couple days hacking around to try and get it to work. If I pass an array to the controller it works fine of course which is one reason I think that there is something hosed in the data binding for async external resource calls.
Any pointers or help would be appreciated.
You can also get to the same code on plnkr: http://plnkr.co/edit/BAHhsT2l1HKnpjiKLZMK
Thanks,
chad
To be more specific, when a cshtml is needed, what happens? I would need to build an application with only ashx and a template engine, without the MVC stuffs, simple rendering, but i would'not like to loose the capability to change the cshtml files on the fly. So I have somehow to manage their recompilation, possibly without recycling the application server loading and unloading AppDomains.
On the first request I can build the page and load it in a sort of cache (like RazorEngine on codeplex), then reuse it. When the page changes i should change the page and that's it.
But as far as I understood a new assembly is built for every page, so if there are frequent changes (that is likely to happens in my environment) tons of assemblies will get loaded.
The question is,
How it works for the standard MVC ?
CodeDom is used, or directly IL that is subsequently added without creating new assemblies?
One thing that comes me in mind now is that after building the assembly, it can be decompiled and then the IL loaded directly on the app, that would make sense
Ask for further clarification if it's unclear! (and any suggestion to this Stackoverflow newbie is welcome)
Ok,
First the cshtml files are loaded then the files are compild into their own assemblies with the standard CSharp/VbCodeCompiler, no code dom is present as far as i understood...
Once you load an actionscript page from scratch (in my case loading XML data from a file, initializing various other elements in a fairly time consuming way) if you navigate away from that page and then return to it, (via the browser 'back' key for example) is their a way to forego the previous initialization and just immediately bring up the previous Actionscript page in its fully initialized state.
Flash Player will always start up your SWF file from scratch. In most cases, navigating back will load the SWF and things like XML files from the browser's cache, and you'll be up and running faster than if you were to go there the first time. However, you're responsible for adding logic to remember the most recent state and restore it when the SWF is loaded again.
If you're using Flex, the History Manager might be useful. For AS3 without a framework, you might check out SWFAddress instead. Both of those use the browser's URL and history features to track the application state, which can allow you to navigate through the SWF as if it contains "pages" and even bookmark specific locations in the SWF. Alternatively, or maybe even simultaneously, you might also consider using Flash Player's cookie-like flash.net.SharedObject functionality to remember the the state. You lose the more granular browser navigation, but you might be able to remember more about a specific single state.