Permalink
Fetching contributors…
Cannot retrieve contributors at this time
117 lines (96 sloc) 5.44 KB

xTuple Extension Tutorial FAQ

How to fork and clone xtuple-extensions

Log in to github.com (you will need a user account, which is free to set up), go to https://github.com/xtuple/xtuple-extensions, and click the "fork" button in the upper-right.

Then, navigate on the command line of your development environment to the directory that contains the xtuple repository. Type

git clone git@github.com:yourusername/xtuple-extensions.git

(with your own username, of course.)

Congratulations! The xtuple-extensions directory is now created. It is important that this repository sits alongside the core xtuple repository.

Why create a new schema?

We do this to avoid namespace collisions, and generally to keep every extension code neatly cordoned off into its own area. It's easy enough to do and our xm (ORM) layer hides all of these details from the client. For the purposes of this tutorial all our tables will be in the ic schema, whereas the ORM-backed views will live in xm.

Why not use native postgres functions to create tables?

Our xt.create_table and xt.add_column functions ensure that you can run the same script over and over and not have to make separate install and update packages. The system will only add things that weren't there before.

Where should I put ORM definitions?

By convention, we put new orms in the orm/models directory, and extensions to existing orms in the orm/ext directory. Unlike with the sql scripts, you don't need to have a master file like the manifest.js that references them all. The core build tool will find all the files in these directories and load them in the appropriate order based on the dependency chain.

What is wrapping the sample client code?

At the beginning of all of our client-side files you'll see a jshint declaration, and the entirety of the code is wrapped in an anonymous function which is executed immediately. If possible, we also "use strict". These are all good practices and you should follow them when writing in our style. Moreover, you'll also see that for the extension client-side files (except for core.js and the package.js files), the code is wrapped in another, named function that is not executed immediately. This allows us to load the code of the extension and actually execute the code later on in our setup process.

For the sake of concision the code examples in this tutorial will ignore the jsdoc and anonymous function wrappers (your code will work without them), but you will see them in the actual implementation of this sample extension.

You might find yourself copying and pasting the tops and bottoms of client-side files, so as to avoid writing the jshint, "use strict", and wrapper functions. This is fine to do, but make sure that you rename the XT.extensions functions. There must only be one XT.extensions.iceCream.initModels function, for example.

What are the differences between the base model classes?

We extend XM.IceCreamFlavor off of XM.Document, which is itself extended off of XM.Model, which itself extends Backbone-relational and Backbone models. XM.Document has some features on top of XM.Model, such as having a user-defined key. This is the most commonly extended base object in backbone-x.

For the larger business objects, we'll typically define one or two lighterweight models (and ORMs) to be used in lists or as nested objects within other business objects. So, along with the "editable" model XM.Quote, we have XM.QuoteListItem and XM.QuoteRelation, which have fewer fields and are not editable. These lightweight models are extended from XM.Info. In the case of IceCreamFlavor our editable model is so light that we don't need to define any others.

How do I update the strings file?

The /path/to/xtuple-extensions/source/xtuple-ice-cream/client/en/strings.js file provides the English translations for all the visible text in your app. If you're diligent about always entering your visible text in the format "_myWord".loc(), then you'll notice the underscores onscreen, which will remind you to add the English translation into the strings.js file.

By the end of your tutorial your strings.js file will look like this.

Why do we need a new table to extend contact?

In a perfect world, we would just go into the cntct table and add a column. This is not an option. We're writing a humble extension here! We have no authority to make changes to core tables in the public schema.

Adding a new table is next-easiest approach. The good news is that when we complete this plumbing in the database, the Contact business object will appear in the application as if this field were in it from the beginning.

What is the XM collection cache?

There are many collections of data that we want to be immediately accessible to the client, such as a list of the supported currencies, or incident priorities, or anything else that we'll want to use to fuel a picker dropdown. When we declare that we want a collection to be included in our cache, then it will be fetched when we start the app, and will be accessible as XM.termsTypes, for example. These caches are automatically kept up-to-date if you go to the setup area and change their models.