We have a legacy PHP/Zend application built with dojo/dijit and dojox/mobile. We want to start to remodel the application using DOJO2. Much of the existing UI is already in the form of dijit AJAX pulls against server code.
We've been through the dojo2 tutorials, as well as become fond of webstorm.
As folks do such migrations, is it most often done as a new, dojo2 application that makes pulls against legacy (existing) server code, or is it more common to point webstorm's 'dist' directory to the existing application's javascript folder and create 'stub' server pages that just include the new DOJO2 code.
Or does it not really matter?
We're just looking to see what is the most common path forward with DOJO2 while minimizing the mucking that goes on with Zend's module routers, and don't want to start down a path that has some unknown 'gotchas'.
Dojo 2 follows a reactive architecture and has a different build path to legacy Dojo. You may find it easier to re-write each php page's TS/JS code at a time to use Dojo 2. I'm afraid I am not familiar with Zend so cannot provide specific advice for this.
Related
First off, I'm brand new to Dojo.
I'm integrating it into our existing web app.
We initially only need the Calendar widget functionality.
I'm looking to keep the number and size of files as small as possible.
I don't believe downloading just the base code file will be sufficent?
http://dojotoolkit.org/download/
Additionally, the Dojo toolkit download is a huge zip (Even if I was to only use compressed files)
Am I left with downloading the toolkit and manually removing everything I don't need?
Is there no custom download builder like jquery ui?
Well, the dojo library is much larger than jquery ui and I don't know of an equivalent to the download builder. If you are just interested in using dojo for a single widget, you might consider exploring a different library.
To use dojox/Calendar, you are still going to need the many dependencies it has on other dojo modules. You can do this manually, but it will be tedious.
One thing you can do is run dojo's build system to package dojox/Calendar and all of its dependencies into a single file. This isn't a trivial task and requires a good understanding of dojo's AMD loader and package system.
If you want to go down this route, I would clone the dojo-boilerplate project on github. It contains everything you need to do this out of the box. Then follow the build system tutorial to understand how you set this up. From there you can have your app depend on dojox/Calendar to produce the file you include on your page to consume it.
I suggest that you put the whole thing (yes, it's a lot of tiny files) to your server.
Dojo 1.9 is written so that when users visit, their computers will only download the individual pieces on an as-needed basis. This is possible because every piece (AMD modules) is explicit about what it needs.
Once you have something that works, you can choose speed-up loading times by using the build system. Basically, this involves going: "If the user wants this thing, they'll probably want all this other stuff, so create a big minified lump and give it to them whenever they start asking." Best of all, it doesn't have to be perfect: If you miss including something, the users browser will still request it a la carte.
At work we're using the Dojo Boilerplate starting application which helps give some initial organization to the build process.
I am converting a good old ASP.Net website to a single page application using Ember.js in a ASP.NET Web API project.
All the devs of my team and myself are pretty new to javascript. We spent the last 2 weeks learning the basis and comparing SPA frameworks. I apologize in advance if my question sounds stupid :)
All the Ember tutorials I have found so far included all Handlebars templates into one single file. I assumed it would be pretty obvious to split them into separates files (*.hbs) when the time would come, but it's not. I might be totally missing something here, but I found about 4 ways to get my templates back when I need them. I'd like to know which method you would recommend:
Concatenate and then inject all the template files when the app loads. I could write some C# code on the server-side that concatenates all the templates files into a single one when the app loads (i.e. each time a visitor enter the app). It seems odd to me, in terms of processing, but also because the generated HTML file will be pretty heavy.
Load each template dynamically via Ajax when I need it. Pretty much what is done here. I kinda like this solution even though I haven't tried it yet. It makes sense to me to get asynchronously a template when I need it instead of loading the entire app on the first load.
Use the Bundling mechanism of Asp.Net MVC. I found stuff like csharp-ember-handlebars to precompile the templates on the server-side and return them as a single javascript file. It works-ish but I feel like the precompiled file will become pretty heavy as I add new templates.
Use Grunt with the plugin grunt-ember-handlebars to precompile the templates. I'm not familiar with Grunt but if I understand well all the devs working on the project will have to install Node.js + Grunt + learn how to use a command prompt + remember to run the command before each commit (if they modified a template). This is not obvious for the web designers. And adding grunt to the build actions will require the entire dev team (working on other projects) to have grunt on their machine (not acceptable).
I need to find a simple and elegant solution to address this issue. My project is in a solution with 35 other projects and I cannot add too much complexity to the build, neither depend on unstable libraries. Maybe I have been too optimistic when I thought I could use Ember for my project. Any suggestion would be welcome!
Your #3 is the most ideal (and common) way that I've seen applications handle templates. With a compiled and minified template file you really don't have to worry to much about performance problems in regards to adding new templates, especially if you take advantage of caching.
One benefit to having the templates compiled and available off-the-bat is that users only need to Download Your Resources Once™, as apposed to downloading resources for each subsequent page load. This leads to a fantastic user experience.
I read Scott Gu's article about built-in support for bundling and minification in ASP .NET 4.5.
However there's no mention of embedded resources, which is a pity.
In the past I've been using a Codeplex project called Client Dependency Framework which supported embedded resources.
Seems like a pretty major omission to me. Is support planned?
I'm pretty sure you could write your own transformer to handle this.
Create a class that implements System.Web.Optimization.IBundleTransform.
Then in the Process method get the contents of the embedded resource. This shouldn't be too difficult. This blog post might be helpful.
Then add the transform to the bundle.
e.g.
var bundle = new Bundle("~/Test").Include("~/Content/Site.css");
bundle.Transforms.Add(new EmbeddedResourceTransformer());
Note that I am using the nuget package from System.Web.Optimization, not Microsoft.Web.Optimization (I have no idea why there are two different namespace implementations, and whether the syntax would be the same in both).
I also can't vouch for the performance of doing it this way as opposed to the file system.
Hope that helps!
Just a few comments on above answers since I don't have enough rep. to comment directly...
The answer from Hainesy suggests using a BundleTransform. I believe this is too late in the process to include an embedded resource. The BundleTransform is helpful for converting things inside the css or javascript after the contents are pulled from the original file and before they are put into the bundled file. For example, if you need to modify image URL's in CSS to point to local relative url for dev and to a CDN url for production.
The link from user960567 explains how to use embedded resources, but there's a catch. That process will only work for something like a common control used from another project.
e.g. If you create a textbox that needs CSS and JS then it allows you to create a HTML helper in the common project that will add the textbox and the script tags that pull in the embedded resource into the page. It does not allow you to pull the embedded resource from the common project into a bundle in another project. This will basically create a separate script or style tag for each embedded resource which may not be what you want (at least it's not what I was looking for.)
I've written a detailed article about how you can use the bundle and minification technology to wrap up your external resources here.
I am rather new to Sitecore and would like to know a bit more about the regular approach to a new project. I'm therefore willing to listen and try out some of the experienced Sitecore developers solutions. I have alot of questions, i won't ask them all. I am just very curious to the approach other people have.
What would be the best approach to start a Sitecore project?
How would you set your project up?
What will be your approach looking at the recycling of code in future projects?
In short: What experiences do YOU have (if you have worked with or are working on Sitecore projects) and how would you recommend other people to work with Sitecore.
Right now we are busy on building Sitecore blocks that we can just recycle in other projects but i know for sure there are 1001 handy tips and tricks out there. I hope we have some Sitecore pro's # stackoverflow that could help a bit.
Here is some general setup info, based on how we do things.
Subversion
This is not Sitecore specific but we set up our repository like this
branches - This is used for working on big updates to the site that may take a while. Say for example I wanted to update how all of the sidebars on the site worked, and this was going to take a few weeks to complete. What we do is create a new branch, and set up another sitecore instance for this dev branch and do what we need to do. When it is complete we merge it back into the trunk for testing and deployment.
tags - This is used for keeping a copy of code that will never be merged back into the trunk (that is the difference between this and branches), so for example when we deploy an update to a site we can create a tag of said code so we can go back to it if necessary.
trunk - The active code, anything checked in here should always be deployable.
The Trunk
This is where we are actively developing/fixing bugs, depending on which part of the project that we are on. We set it up something like this (as an example the project is called TheProject)
We keep our solution file at the root of this folder, this will reference the various libraries in the src folder as well as the web project in the website folder.
docs - A place to put documentation about the site. I strongly suggest that as you complete features/sections you write up a little guide about any special knowledge needed for it to work. So say I am working on a featured content box on a landing page. This box will automatically pull some content unless it is explicitly overridden. What I do when I complete something like this is I write a guide for the customer, using a lot of screenshots. I send the guide to the customer as well as put it in the docs folder. This both helps the customer train their staff, as well as helps new developers come up to speed with how things are done.
lib - This is where we keep any DLLs we are going to need to reference in our projects.
test - A place to put unit tests.
src - This is where we keep our project specific library code. So in here we would have a folder called TheProject.Library, and in there would be the visual studio project for said
web/Website - This is where we have Sitecore installed and is the root of the site. In here we have a project called something along the lines of TheProject.Web. In the project we add all of the general stuff like the web.config/layouts folder and so on.
General Sitecore Code Library
One the best things you can do is from the start setup a general Sitecore library that can be added onto over time. Then when you write any code for a project that is not only applicable to the project, you can add it there. It may seem obvious, but this will really help in the long term. You will end up with much more solid code, see link text .
So when we are done with all this we have something like this as a solution/project structure
TheProject (The solution)
TheProject.Library
TheProject.Web
MyCompany.SitecoreLibrary (our general sitecore library)
Tools
This is another general thing, but I find it can really help speed up Sitecore development. If you find yourself doing something over and over in Sitecore, using API write a tool to do it for you. This not only helps with solving whatever problem you are tackling, but also helps to get you more familiar with the API.
Resharper
This is more of a general .NET development suggestion, use Resharper(http://www.jetbrains.com/resharper/index.html). I am sort of a a Resharper fan boy, it makes so many things with development easier and quicker. In my mind the biggest advantage though is how easy it makes refactoring code, which is really important to do over time to keep things clean and understandable.
I hope some of this helps.
Gabe
This is, as you said, quite a big question. Here are some of my thoughts:
Developing Environment
First of all when I start a new project I install Sitecore on my developing environment and I make sure everything works. Either during installation or after I place the databases on a separate SQL-server and change the connectionstring accordingly.
I open up Visual Studio and create a solution and include the files needed. I create some kind of HelloWorld rendering and try building the solution so that I can verify that everything is working as it should.
When everything is up and running I create a zip-file of the whole solution, including the data-folder. Now it is time to add this to some kind of version control system, in my case Subversion.
I add the zip-file to subversion and also add all files that I think will be changed during the project, usually I tell subversion to ignore the sitecore folder, this speeds up performance drastically when checking in files.
After I perform a commit-action the other team members of my project can check out the code and start developing (after unzipping the zip-file, off-course)
We all work towards the same database although this goes against Sitecore recommendations, we havent had any problems with this approach however items in GUI created/changed by one developer take some time before it is created/changed for all the others.
We could off-course develop several different projects using the same Sitecore installation but since almost all customers use different versions of Sitecore we have found this approach a bit cumbersome.
Often we set up an automated build-server but this is a whole other issue.
Reusable code and renderings
I would like to say that we create neat packages based on the same codebase that gets reused between projects but unfortunately we are not there yet. Today it is a lot of cut and pasting between solutions.
Uploading code to customer
This is done via sitecore packages, normally with some kind of dynamic selection for what files to include, say all ascx-files in a specific folder changed the last 5 days.
There you have it.
Take a look at this series.
Especially the component architecture part have increased our level of reusability.
When you create your Visual Studio's project in Sitecore's web root folder and you will keep all Sitecore's dlls files inside bin directory, don't forget to add to project's references all these files:
bin\ComponentArt.Web.UI.dll
bin\HtmlAgilityPack.dll
bin\ITHit.WebDAV.Server.dll
bin\Lucene.Net.dll
bin\Mvp.Xml.dll
bin\Newtonsoft.Json.dll
bin\RadEditor.Net2.dll
bin\Sitecore.Kernel.dll
bin\Sitecore.Logging.dll
bin\Sitecore.NVelocity.dll
bin\Sitecore.Zip.dll
Because when you CLEAN your project and you will have reference only Sitecore.Kernel.dll (in most of cases), you will lost most of dlls from bin directory!!
Just say you have a quite large and complicated desktop application written in objective-c/cocoa written properly in MVC. You then wish to replace the V and C so that it is a web application?
Is there anything like Tomcat but for objective c? The closest thing I can think of is somehow convert the Model code into an apache module to load in apache?
Obviously I could embed some sort of HTTP server and write a whole pile of code to manage sessions and requests and responses and so on, but might there be a simpler way?
You can leave the server almost as-is. I'd run it as a daemon in the background.
I'd split the controller part. One part of it resides on the server as a connection between HTTP requests and the daemon.
You could run a self-made http server.
You could build it into a Apache module as you mentioned.
You could access your controller through CGI. This is the easiest option as I see it.
The second part of the controller is written in Javascript in the browser, exchanging data with the server and updating the GUI.
The view part is written completely in javascript.
You could - if you want - leave out the controller on the server and make the model understand HTTP requests.
For the Controller / View part consider using a framework.
jQuery for only a light interface. (Or a similar framework.)
I haven't yet found a framework that leaves all the model stuff to the server. Maybe you can use SproutCore or Cappuccino accordingly.
SproutCore, the full-backed MVC framework that is used by Apple. (Think so at least.)
Cappuccino, another full-backed MVC framework.
Have you looked at Cappuccino? http://cappuccino.org/
I have not used but have heard many good things about it.
There is a new web development platform for Objective-C/Cocoa called Bombax. It sounds like exactly what you're looking for (it is designed to allow you to write entire web applications in Objective-C). Perhaps you could even combine it with Cappuccino. You can check it out at http://www.bombaxtic.com.
The closest you'll probably come to Objective-C web frameworks are either SOPE, or GNUStep, neither of which I have used, but ran across when I was deciding if I wanted to use Rails or something written in Objective-C for my web stuff. There's also ARJDatabase, which is kind of like Core Data, but not source compatible.
The bottom line is, unless you were very careful about how you wrote your code (i.e. wrote it with GNUStep in mind, didn't use anything Apple specific) you're probably going to have to do quite a bit of work anyway. And if you used Core Data at all, there's no way I know of to reuse that outside of a Mac App. And since it's not setup to be used my multiple simultaneous users, you wouldn't want to use that in a web app anyway. Bottom line, I went with Rails.
There is Frothkit - http://code.google.com/p/frothkit/ but it appears not to have been updated for a year.