Guidelines for Creating a Custom Action

Last modified: November 19, 2024

Introduction

For details on custom actions, please see Custom Actions in the ATS 1 Reference Guide.

These are the definitions of Mendix actions and core actions:

Using Mendix Actions

Always try to use a Mendix action first. This table explains why you should use a Mendix action:

Reason Description
Automatic context A Mendix action will automatically adjust to the Mendix Platform, ensuring that it will work correctly.
Visibility checks A Mendix action only searches for elements that are visible.
Browser support A Mendix action always works in the supported browser.
Mendix support A Mendix action always works on the latest version of Mendix.
Waiting for background processes A Mendix action waits until the Mendix app is finished rendering. This aspect is also covered by the Mendix Wait action.

However, there are reasons to not use Mendix actions. For example, sometimes a custom action can get too complex using a Mendix action, because of the input parameters that are required. For another example, when using a Mendix action, you have six input parameters:

But when you use a core action, you have four input parameters:

The difference is because of the required Widget Name input parameter in a Mendix action. When you use a core action, this is not always required. So if you don not need the widget name, a core action is a better solution. This is a valid reason to use the core actions; otherwise, the custom action is not user-friendly.

Using Visual/Optical Components

Only use items that are visually present on the page, like the text inside a certain widget. Sometimes it seems easier to use an attribute or a certain class from a widget, but doing this right the first time will save you time in the future. Invest the time and use the visual components instead of taking the easy way out.

This is a visual component:

The User Approach

When building a custom action, always think about what the user is doing. In addition, keep in mind the things you do without performing an event (like reading a data grid), as these are important when building Search Context Actions.

Some examples of questions you should ask yourself are:

  • How do I find the widget using visual components?
  • How do I trigger an event? (click, enter text, etc.)

Another example is with a checkbox. Do not set it using the JavaScript check function, but click it, because that’s what a user would do. If you want to set the checkbox based on its current state, use the precondition and the click action. By doing this, you invest time in a more complex solution to save time later.

Building It Generically

When building an action for an unsupported widget, keep in mind that the action should work on the widget on every occasion. When a widget has two ways for entering information, use the most obvious one. For example, when using jQuery to find child nodes, make sure you use the generic aspects of the child node to specify it. Some class names are application-specific, and you cannot use them. You don’t want to have to rebuild this when you need it for another application.

This also applies to optional components in the class name. Some widgets have optional components that you can add. For example, when using a selector to find a child node, avoid using classes like glyphicon. These classes are most likely optional. Therefore, if the application does not use the glyphicon option, your action will not work.

Try to find the generic aspect of the child node. In most cases, the mx-name- is unique; therefore, use the mx-name- of the widget to find the widget for the Mendix actions. When the mx-name- is available, use it!

This is the mx-name in the debugger:

Background Processes

When using core actions to trigger an event, always use a Mendix Wait action after triggering the event. The Mendix Wait action covers the Mendix app rendering process. The action waits until the application has finished its background processes, like communicating with the database. If the Mendix app no longer renders, then the action passes. This only works on Mendix applications, so when you need to cover a bit of HTML for logging in, the Mendix Wait action will not work.

This is the Mendix Wait action:

For more information, please see Mendix Wait.

Action Parameters

When defining action parameter names, always use the ATS naming conventions.

In addition, include the optional input parameters in the action parameters for an unsupported widget action.

In almost every situation, include the Search Context input parameter. When an AUT has a lot of snippets, mx-names are not unique. Use Search Context to specify the element in which ATS should look.

These are some of the most used action parameters:

Action parameter Description Required, Optional, or Both Data Type
Widget Name The mx-name of the widget. Required String
Value The value that needs to be set or found inside a widget. Both String
Search Context Provides the element in which ATS will look for the widget. Optional Web Element
Index The index of the widget/item to find. Optional Integer
Column Name The name of the column. Both String
Row Name The name of the row. Both String
Widget Returns the widget. Required Web Element

These are the three most used action parameters:

Last but not least, give a description to the action parameters, especially when it requires specific information. You can also give a generic example of what the user should enter, making it more user-friendly.

Describing the Steps and Defining the Output

When you describe your test steps in ATS, delivering support and updating the action is a lot easier. This will refresh your memory, and it will make it easier to understand for other people.

In addition, describe the output of each test step to make it easier to understand the custom action. Another person should be able to understand what the action does by reading the description and output of the steps.

This is an example:

Data Type Awareness

Be aware of data types. All the action parameters require a certain data type, which means it will only accept or suggest that data type. For example, an input parameter that requires an integer will not accept 3.8, because it only accepts whole numbers.

Set each action parameter to a certain parameter. A Search Context action most likely returns a web element. An unsupported widget action that gets the value of a label returns a string, and so on.

These are the different data types in ATS:

Data Type Description Examples
Text A piece of text that can contain letters, numbers, spaces, and other characters. “ATS123”, “Helloworld!”
Number A whole number. 123, 4423
Boolean A truth value. true, false
Date/Time ⚠ The Date/Time type is deprecated and should not be used. N/A
Page Element Represents a DOM element. (See the image below this table.)
Any Lets the user choose the data type. Choose String for “Helloworld!”
Drop-Down One of the values of the given drop-down menu. Red, Green, Blue; Todo, Running, Done
Float ⚠ The Float type is deprecated and should not be used. N/A
Currency ⚠ The Currency type is deprecated and should not be used. N/A

The JavaScript actions have three different versions, based on the data types:

They all produce outputs based on their data type. This means that you cannot return a string inside the Execute JavaScript web element action.

Good luck building custom actions! Please post any questions about custom actions in the Mendix Community under the category Testing.