Do you fire ajax requests through the MVC framework of choice, or directly to the CFC?
I'm leaning towards bypassing the MVC, since I need no 'View' from the ajax request.
What are the pro's of routing ajax calls through MVC framework, like Coldbox?
update: found this page http://ortus.svnrepository.com/coldbox/trac.cgi/wiki/cbAjaxHints but I am still trying the wrap my mind around what benefits it brings over the complexity it introduces...
Henry, I make my Ajax requests to proxy objects of my model. Typically, I am outside of a 'framework' when doing so. That being said, it may be (very) necessary to utilize your framework, such as working within a set security model.
I can't really see any benefit of bypassing the MVC framework - in combination, those three elements are your application.
Your ajax elements are really part of the view. As Luca says, the view outputs the results of the model and controller.
Look at it this way - if you made an iPhone-friendly web interface (that is, a new View), would you bypass the model and controller?
Luis Majano, creator of ColdBox said:
These are the two schools of ajax
interaction henry.
I prefer the proxy approach because it
adds the following:
Debugging
Tracing in the debugger
AOP interception points
Security
Setting availability
The proxy will relay to the event model, so I can use local interception
points, local AOP, plugins, etc.
In other words, it can be a highly
monitored call instead of a simple
service cfc call, which you can still
do.
I, for one, love to have my execution
profiler running (part of the coldbox
debugger), so I can see when ajax
requests come in and when they come
out. I can see the data requested and
the data sent back. I don't have to
look in log files, or try to imagine
results or problems. It really helps
out in debugging.
However, it would be a developer
choice in which way you decide to go.
My personal preference is to always
use my proxy to event delegation
because it gives me much more
flexibility, debugging and peace of
mind.
The purpose of the "view" in MVC frameworks is to show the data after the "model" and "controller" have generated it. If you don't need the "view", then what's the point of using such a design pattern?
I agree with Luca. It also bypasses any kind of sanitization and filtering logic you have in your MC stack. It basically negates any kind of query processing that you may or may not have in place.
Yeah, I wouldn't bypass your framework, figure out what's causing you grief and hunt down the offending pieces, adding logic to exclude common components such as headers or footers, and looking for methods injecting whitespace that while fine for html is annoying or down right problematic when parsing json.
Adding output="false" especially in your application.cfc and it's methods would be the first thing I cleaned up.
I am a strong believer in NEVER directly accessing the CFC's directly, I find it creates long term problems when a major refactor might want to consolidate or eliminate components, the direct accesses potentially make this harder than it should be, especially if a third party is hitting your ajax from another domain(e.g. flash remoting).
+1 to Steve's answer.
Related
this question could be a duplicate but I have a specific use case.
The app is a single page Django app. The user can choose from several options in a drop-down box. Depending on the choice, a few input boxes will need to be rendered. As an example, if the user chooses to order a pizza, the options in another drop-down menu should be related to the toppings whereas an order for drinks should provide options related to the type or brands. The resulting number of input boxes could be 5 with one option or 10 with a different option.
I can think of rendering the page using JS or by using Python in the back-end. Because this app will be a commercial one, I do not need SEO as users need to log into the app first. I also want to minimize the amount of code that a competitor can 're-use' from my work.
Will client-side rendering open up security problems? If not, is client-side the better way to go?
This question is more of a theoretical/opinion-based nature than technical, but let me provide some answers.
Will client-side rendering open up security problems?
Generally, web application security is a server-side concern, not client-side. You can do things like input validation on the client-side, but the minimum practice for security is to sanitize, validate, and authenticate all request data anyway, so the client-side checks are more of for convenience and improved user experience than security. I'm not saying that there are no such things as client-side security concerns, but it's something I don't think is generally a cause of worry. Client-side rendering specifically and especially doesn't sound like something to be careful about: regardless of what your client-side code does, whatever <form> and <input> markup it generates, your server-side code should always handle the submitted data as if it could be malicious.
Is client-side the better way to go?
There are so many more factors to consider in order to answer this, so it's largely a matter of opinion. But since you're asking about Django, then you might want to reduce overall development friction by maximizing Django's features and design—and Django, in my view, is largely a static markup-first framework, meaning minimal use (at first, at least) of client-side JavaScript. Django Forms and Class-Based Views (CBV), for example, work well together to allow rapid development of non-single-page applications.
Your specific use case of an initial drop-down choice determining the main form to be presented could be developed very rapidly in the traditional Django way by giving up your single-page-application requirement, and just providing some initial menu page that will lead to the different views and forms (pizza vs. drinks, etc.), the latter of which you could build rapidly with the help of CBVs. (By the way, your specific use case doesn't seem too unique, actually. It's just the fundamental issue of complexity for which we have programming concepts such as polymorphism and inheritance in object-oriented programming—hence the appropriateness of CBVs.)
I know that single-page applications are nice, and is the fashionable thing nowadays, but I think people underestimate the speed of old-fashioned HTML applications. And by speed I mean not just the user's client-side experience (HTML pages load rather rapidly with HTTP2 and CDNs and all the other modern Web infrastructure tech these days), but also development time.
Besides, you can always just add single-page-like experiences in a progressive manner. Django is particularly suited to an agile-style development strategy where you'd build initial functionality rapidly without much client-side JS, and then just add rich client-side experiences (using React or Vue or something similar) where it will add the most value for users.
I also want to minimize the amount of code that a competitor can 're-use' from my work.
I don't know the full context, but generally I wouldn't worry about this. If you won't do much client-side rendering, then there won't be much client-side code to ‘steal’. But even if you do, unless you specifically write your client-side code in a way that maximizes reusability (either for yourself or for others), I think coders tend anyway to write highly-coupled code, which is to say, your client-side code will tend to be highly dependent on your server-side code's specifics, which means poor reusability. Your competitors could copy your client-side code all they want, but the cost of making it work with their own back-end will be so high that it wouldn't be worth it, they'll just want to write their own.
I'm new to web development, but have recently been getting going with it pretty fast, using Django (which I'm falling in love with). However, while Django is easy to get going with and pick up quite fast, I'm afraid there are concepts I don't know much about, particularly REST and RESTful web services. I hear those terms thrown around a lot, and I'm assuming are important to modern web apps, and I want to know basically what they mean, when I should use them, and how I should use them (package, plugin, etc.).
My web app consists of the following functionality:
Discussion Board which I've implemented so far only using the model layer
Messaging which I've implemented so far only using the model layer
Payments (not yet implemented)
Calendar (not yet implemented)
And that's about it for now. When should I be thinking about REST within these functionalities?
You could get really in depth with the subject, but when I think of it, I think of the URLs your site will be providing. That, for me at least, is the simple way to think of a RESTful service. I also think it's a good way to get to grips with Django & it's generic views.
Taken your example there of a calendar. A RESTful approach to a calendar app, with django's generic views, might implement URLs like;
# ListView
/calendar
# DetailView for a given item in the calendar
/calendar/<id>
# UpdateView for an item
/calendar/<id>/update
# DeleteView for an item
/calendar/<id>/delete
Beyond that, REST requires you consider the HTTP methods in use, so you should then define what methods your URLs will accept in order to better control the interaction. Furthermore, if you were enforcing the HTTP method on each action, you could write a more generic view which didn't expose the /update or /delete URLs. But I'd consider that a more advanced approach & you may wish to start from a more explicit design.
Once you start to write a consistent structure for your apps then you can easily make generic functions, and expand.
There is a whole load of things you could read on this subject, depending on where you see it going.
If you're thinking of building something that can provide an API then there's already a Django framework for this; http://www.django-rest-framework.org/
But if you're getting started & just want to know more about the concepts, Wikipedia is always a good place to look, and finally this looks like a great example for this subject using Django which should hopefully help you on your way.
To understand the idea of a resource, take a look at this answer
I'd like to create a highly maintainable code, with unit tests.
I have read the best practices so directories are organized as per best practice, with just one bundle named AppBundle, and I'm using annotations.
I have an issue with business logic. I've read that I should not put business logic inside controllers. As the vast majority of my code was not meant to be reused I put the logic inside the controllers. Yesterday I read here that "logic should live in the controller unless you’re going to unit test it or until you need to re-use it" and this made me faint: am I not going to unit test if I put my logic inside controllers?
I understand that "controllers should be as lean as possible, with few lines of code to glue stuff together". But what about forms? A form often has a logic like "display the form, if it's valid do something and display another page". Now if I'm going to put the form creation and logic inside a service (or a model?) I'd have to put the page render command inside that, so basically the controller would be 1 line, with all the rest inside the service, and the actual decision of which page to display would be done inside the service, not the controller itself...so what's the point of a controller, just the routing?
I really need to understand this before proceeding (i've already developed 3 months on this and I'd have to rework a lot, but better now than never)...
Thank you!
EDIT: some additional considerations to address some of the comments below.
I have a clear understanding of how forms work in Symfony, with the creation of the form and it's management with the "isValid()" in the same place.
Let's imagine the following controller which performs the following operations:
get the current customer from security context
queries the DB to get the field $user->getIsBillable()
if the user is not billable
queries the DB to find if someone else is paying for him
else //the user is billable
create and manage form1
if the user is not billable but there is someone who is paying for him
check if the license is active (I have to call an external API)
if the license is not ok
redirect to a page to update credit card info
if the user is not billable
create and manage form2
else
create and manage form3
render the license management page with all the necessary forms created
What would be supposed to do? Put all this stuff into a service which would return the forms (or null if a form is not created)? Consider that in the "is valid" of some of the forms I render the same page with simply updated info. Or should I create a service foreach of the forms creation and keep the logic in the controller?
You should keep your logic outside the controllers in order for your code to be more reusable and to reduce code duplication.
The best practice is to have your business logic registered as a Symfony service instead.
The controller should be as lean as possible as it only handles the view and accessing your services to run business logic functionality.
Using services will make your code more reusable and it will be easier to write unit tests for each part of your code.
Forms are handled in a different way in Symfony2 and you should read more about it: http://symfony.com/doc/current/book/forms.html
Like most php frameworks, symfony2 has its problems.
Turns out symfony2 best practices are not always best practices. But in most cases symfony allows for loose coupling and proper dependency injection, it just doesn't tell you in the "book".
For example a very important topic: http://symfony.com/doc/current/cookbook/controller/service.html
I usually have my form types as services, my controllers as services and other services for the business logic. My controller actions are ideally around 3-5 lines of code. Sometimes when handling a file upload they will be slightly longer though.
You can unit test your business logic services, but for your controllers and forms you will need some sort of integration testing, or use heavy mocking.
I understand that "Selenium" is a good tool for testing such things, I haven't used it yet.
One thing that I would like add to other answers provided there to you is that Unit Test is different from Functional Test.
Unit Test
You need Unit Test when you want to test individual "unit" of code (such as method in a class)
Functional Test
You need Functional Test when you want to test a "suite" of functionalities (like many methods interactions, db interactions, web services and so on)
Recap
Ideally you want to write Unit Test(s) for a class that contains heavy business logic whereas you need Functional Test(s) for testing whole controller (so a page is render with all class interactions and db ones, the page is render, the page is posted and so on)
The best approach, to me - and it's only an opinion of a devel that tried as today to understand testing from a newbie point of view - is to mix each other: with Unit Test (and isolation, of course!) I'm sure that (simplifying) every input that I provide to my method will return the correct output, while functional test will help me to test the whole thing and interactions.
Answer
So, returning to your question - that seems wider that at first look could seems - a good way to procede is to keep all business logic outside controllers (even if you're sure [and to me you'll never be sure enaugh] that code you put in will be never reused elsewhere) by defining services (as said in other answers). Then you will do some unit test(s) onto your services and after that you will probably need some Functional Test(s) to check that all "the glue keep things together correctly".
I don't know if I'm missing something (it's likely). Just to be more complete, I will attach my older question on Symfony2 (I was studying it and MVC pattern at the time)
Structuring your code to be clean and maintainable it is a very hard stuff, I suggest you to start studying the SOLID principles than having a superficial understanding of MVC:
Single Responsability Principle - SRP
Open/Closed Principle - OCP
Liskov Substitution - LSP
Interface Segregation - ISP
Dependency Inversion Principle - DIP
I would like to use ember-data in a project I am building, but the API I am working with does not follow REST conventions.
For example, all the HTTP requests are POST and the naming conventions of the endpoints are unique to the actions they perform. e.g. /api/thing/retrieve would require me to post a JSON object with some parameters, and would return a 'thing' to me.
Do I use the Restful adapter and reopen the class and redefine the various find functions?
OR
Do I have to make a new adapter?
OR
Should I just abandon ember-data and and use ajax in my models(or maybe even controllers)?
I'm not sure how I would handle any of those options. Any guidance would be appreciated.
The only information which I have seen on this subject has been an article by the Discourse folks linked below.
http://eviltrout.com/2013/03/23/ember-without-data.html
I personally have toyed around with the reopenClass method in the article, and would probably drop it into a mixin or something to that effect if I had a consistent but non-REST API which I was calling regularly.
I would say that, if your API is consistent (reliable) then you should create/extend the DS.Adapter (not DS.RESTAdapter) to implement to your specification.
All the hooks are there, you will just end up defining it once which all models can use.
I would also read through the Basic Adapter code - (https://github.com/emberjs/data/blob/master/packages/ember-data/lib/adapters/basic_adapter.js) it might be a better staring point for you then DS.Adapter.
If your API is not reliable, then you are probably better off with just using the $.ajax() calls as necessary. But, in my opinion, that does not scale well.
A link worth reading when looking at Basic Adapter: http://emberjs.com/blog/2013/03/22/stabilizing-ember-data.html
One last note, building an ORM or even a something more simple then an ORM is not a trivial task, that for me, makes using ember-data worth the effort, and yes sometimes pain.
Currently I'm working with ASP.NET 2.0, which may explain why I'm not as up on this as I might be. However, I don't see a full solution in my Googling of ASP.NET MVC, etc.
Here's my background thinking.
Firstly, data-bound templates are really useful. I'm currently dealing with lots of legacy code whereby people are building up controls programmatically, both on the client and the server, and it's a huge pain.
Secondly, sometimes you want controls to be data-bound on the client, sometimes on the server. The most obvious case for databinding on the server is where you're trying to account for people turning off javascript. But issues of speed, caching, bandwidth etc. all play their part as to deciding where to bind.
Now, on the server I can write UserControls with databinding points. And on the client I can write templates and bind them with JQuery (I'm currently using the microtemplating engine by John Resig as amended by Rick Strahl). But ideally there should be a way to write a template once and let the plumbing make it available for both server and client-side data binding. I guess that XML/XSLT would be one approach to this, but a horrible one. So what else is there? It needn't be an ASP.NET 2.0 solution; I'd just like to think that somewhere there is a fix.
HAML
You can create "datasource" objects that are independent of our databound controls / templates.
To use them with your databound control, instead of attaching them declaratively, e.g.:
<asp:gridview ...datasource="myDataSource"...>
you can attach them with code:
(some event)
me.Gridview1.datasource = "myXMLDataSource"
---or---
me.Gridview1.datasource = "mySQLDataSource"
If you set the datasources up ahead of time (either in the .aspx or in the code-behind is OK), then in this way, you can switch datasources based on some event, or logic, when you want to, without having to re-code / re-publish anything.