Package com.smartgwt.client.util

Class AutoTest

java.lang.Object
com.smartgwt.client.util.AutoTest
public class AutoTest extends Object
Standalone class providing a general interface for integration with Automated Testing Tools

For an overview of Automated Testing in Smart GWT, see the documentation here.

AutoTestLocators

The AutoTest class is responsible for creating and resolving autoTestLocators - identifier strings which can reliably identify an interactive element in a Smart GWT application.

These locators are designed to work across browsers, Smart GWT versions, and even major changes to your application's UI. They do not rely on the DOM structure, and are intended to be able to identify the target element reliably without requiring pervasive explicitly specified component IDs.

Note that AutoTest locators are the only supported way to identify Smart GWT generated DOM elements when recording tests for automated testing tools. Using XPath DOM locators to identify elements is not a supported approach. The DOM structure and DOM element attributes generated by Smart GWT components are not guaranteed to be identical for functionally equivalent elements across different Smart GWT builds, different browsers, and for dynamically generated components, could even change across page reloads.

The primary APIs for working with AutoTest locators are as follows:

  • getLocator() to retrieve a locator for an element
  • getElement() to resolve a stored locator back to an element
Developers should also be aware of the installLocatorShortcut() method which allows you to retreive locators via a simple key combo + mouseDown event on the target element.

See the AutoTestLocator overview documentation for details on how locators are structured.

See the Reliable AutoTestLocators overview documentation for best practices for generating high quality locators.

AutoTest locator variables

AutoTest locators may be written to include dynamic variables to be resolved at runtime. See the following APIs for details:
  • setVariable(): Set up a named variable for later use
  • storeLocatorResult(): Resolve a locator against the current environment and store it as a variable
  • getVariable(): Retrieves the value of a previously defined variable

Locator testRoot

For tests that target a standard widget hierarchy such as a screen, the testRoot may be used to reliably identify the widget hierarchy in question wherever it is rendered inside a larger application.

Waiting for asynchronous operations

The AutoTest class provides utilities to wait for asynchronous actions, including both system delays (such as redraws that fire on system idle, for example), and asynchronous RPC operations.
  • waitForElement(): this method will wait for asynchronous operations to complete before resolving an AutoTestLocator to an element and firing a callback function. This allows recorded tests to reference elements that are blocked or unavailable until some asynchronous action(s) complete.
  • waitForSystemDone(): this method will wait for any outstanding asynchronous system operations to complete and then fire the specified callback.
  • isSystemDone(): this method may be called to check whether there are currently any outstanding asynchronous system operations.
Note that Smart GWT's support for integrating with high level tools such as Cypress and Selenium automatically makes use of these APIs.
See Also:

  • Constructor Details

    • AutoTest

      public AutoTest()

  • Method Details

    • getFullPathLocator

      public static String getFullPathLocator()
      Convenience method to generate a AutoTestLocator with no search segments and an explicit path from the locator root to the target.

      This method sets getLocator() with the following configuration settings:

      • useSearchSegments:false
      • useMinimalFallbackAttributes:false
      • useCompactFallbackSyntax:false
      Returns:
      full-path locator string to identify the target. See AutoTestLocator

    • getFullPathLocator

      public static String getFullPathLocator(Element target)
      See Also:

    • getFullPathLocator

      public static String getFullPathLocator(Element target, boolean checkForNativeHandling)
      See Also:

    • getFullPathLocator

      public static String getFullPathLocator(Element target, boolean checkForNativeHandling, String[] coords)
      Convenience method to generate a AutoTestLocator with no search segments and an explicit path from the locator root to the target.

      This method sets getLocator() with the following configuration settings:

      • useSearchSegments:false
      • useMinimalFallbackAttributes:false
      • useCompactFallbackSyntax:false
      Parameters:
      target - Target within the page. This may be a DOM element, a Canvas or a FormItem. If null, a locator for the last mouse event target DOM element will be generated.
      checkForNativeHandling - If this parameter is passed in, check whether the target element responds to native browser events directly rather than going through the Smart GWT widget/event handling model. If we detect this case, return null rather than a live locator. This allows us to differentiate between (for example) an event on a Canvas handle, and an event occurring directly on a simple <a href=...> tag written inside a Canvas handle.
      coords - X, Y page position
      Returns:
      full-path locator string to identify the target. See AutoTestLocator

    • getLocatorWithIndicators

      public static QualityIndicatedLocator getLocatorWithIndicators(Element DOMElement)
      Returns the QualityIndicatedLocator associated with some DOM element in a Smart GWT application page. If coords, representing the page position, is passed in, the locator may be generated with a specific trailing "target area" identifer that will map back to the appropriate, potentially different, physical coordinates, even if the widget is moved. The coords argument will only have an effect in cases where the mouse position over the target could potentially change behavior.
      Parameters:
      DOMElement - DOM element within in the page. If null the locator for the last mouse event target will be generated
      Returns:
      Locator details allowing the AutoTest subsystem to find an equivalent DOM element on subsequent page loads.

    • getLocatorWithIndicators

      public static QualityIndicatedLocator getLocatorWithIndicators(Element DOMElement, boolean checkForNativeHandling)
      See Also:

    • getLocatorWithIndicators

      public static QualityIndicatedLocator getLocatorWithIndicators(Element DOMElement, boolean checkForNativeHandling, String[] coords)
      Returns the QualityIndicatedLocator associated with some DOM element in a Smart GWT application page. If coords, representing the page position, is passed in, the locator may be generated with a specific trailing "target area" identifer that will map back to the appropriate, potentially different, physical coordinates, even if the widget is moved. The coords argument will only have an effect in cases where the mouse position over the target could potentially change behavior.
      Parameters:
      DOMElement - DOM element within in the page. If null the locator for the last mouse event target will be generated
      checkForNativeHandling - If this parameter is passed in, check whether the target element responds to native browser events directly rather than going through the Smart GWT widget/event handling model. If we detect this case, return null rather than a live locator. This allows us to differentiate between (for example) an event on a Canvas handle, and an event occurring directly on a simple <a href=...> tag written inside a Canvas handle.
      coords - X, Y page position
      Returns:
      Locator details allowing the AutoTest subsystem to find an equivalent DOM element on subsequent page loads.

    • getMinimalLocator

      public static String getMinimalLocator()
      Convenience method to generate a AutoTestLocator that is as compact as possible.

      This method sets getLocator() with the following configuration settings:

      • useSearchSegments:true
      • searchSegmentsIncludeHidden:false
      • useMinimalFallbackAttributes:true
      • useCompactFallbackSyntax:true
      Returns:
      minimal locator string to identify the target. See AutoTestLocator

    • getMinimalLocator

      public static String getMinimalLocator(Element target)
      See Also:

    • getMinimalLocator

      public static String getMinimalLocator(Element target, boolean checkForNativeHandling)
      See Also:

    • getMinimalLocator

      public static String getMinimalLocator(Element target, boolean checkForNativeHandling, String[] coords)
      Convenience method to generate a AutoTestLocator that is as compact as possible.

      This method sets getLocator() with the following configuration settings:

      • useSearchSegments:true
      • searchSegmentsIncludeHidden:false
      • useMinimalFallbackAttributes:true
      • useCompactFallbackSyntax:true
      Parameters:
      target - Target within the page. This may be a DOM element, a Canvas or a FormItem. If null, a locator for the last mouse event target DOM element will be generated.
      checkForNativeHandling - If this parameter is passed in, check whether the target element responds to native browser events directly rather than going through the Smart GWT widget/event handling model. If we detect this case, return null rather than a live locator. This allows us to differentiate between (for example) an event on a Canvas handle, and an event occurring directly on a simple <a href=...> tag written inside a Canvas handle.
      coords - X, Y page position
      Returns:
      minimal locator string to identify the target. See AutoTestLocator

    • installLocatorShortcut

      public static void installLocatorShortcut()
      This method allows developers to retrieve locators for elements on the page via a key-combo plus mouseDown on the element in question.
      It may be invoked from a bookmarklet stored in the browser, giving developers a one-click way to retrieve locators for any Smart GWT application

      When installLocatorShortcut() is invoked, it will register a Page-level mouseDown handler which, if the Shift+Ctrl or Shift+Meta key-combo are being held down will display the locator for the element under the mouse in a text-box and also copy it to clipboard.

      As with the isc.showConsole() method, developers may wish to create a bookmark in their browser to quickly enable this functionality on any Smart GWT application, without any changes to the application code:

      1. Create a new bookmark in your browser.
      2. Enter url "javascript:isc.AutoTest.installLocatorShortcut()".
      3. Label the bookmark as "Locator Shortcut"
      4. Consider adding this to the Bookmarks Toolbar. This allows you to enable the feature with a single click from any Smart GWT application.

      To uninstall the locator shortcut, call uninstallLocatorShortcut()

    • isCanvasDone

      public static Boolean isCanvasDone(Canvas canvas)
      Returns whether the canvas is in a consistent state with no pending operations. Returns null if the supplied argument is not a valid canvas. Otherwise, returns true or false according as the conditions below are all satisfied:
      • page has finished loading
      • canvas is drawn
      • canvas isn't dirty
      • canvas has no queued overflow operations
      • canvas is not animating
      Parameters:
      canvas - canvas to test
      Returns:
      whether canvas is 'done' as described above

    • isElementClickable

      public static Boolean isElementClickable(Canvas canvas)
      Returns whether the instance is ready to be clicked on by a Selenium test. Returns null if the supplied instance is not a valid canvas or form item. Otherwise, returns true or false according as the conditions below are all satisfied:
      • page has finished loading
      • no network operations are outstanding (configurable, see implicitNetworkWait)
      • canvas is visible, enabled, and not masked,
      • canvas satisfies isCanvasDone()
      • if canvas is a TileGrid, it satisfies isTileGridDone()
      • if canvas is a TileLayout, it satisfies isTileLayoutDone()
      • if canvas is a ListGrid or body of a ListGrid, it satisfies isGridDone()
      • if canvas is a DynamicForm, it satisfies isFormDone()
      Note that for a form item in a DynamicForm, the DynamicForm must satisfy the third condition above, while the container widget of the element must satisfy the remaining conditions.
      Parameters:
      canvas - instance to test
      Returns:
      whether instance is 'clickable' as described above

    • isElementClickable

      public static Boolean isElementClickable(FormItem item)
      Returns whether the instance is ready to be clicked on by a Selenium test. Returns null if the supplied instance is not a valid canvas or form item. Otherwise, returns true or false according as the conditions below are all satisfied:
      • page has finished loading
      • no network operations are outstanding (configurable, see implicitNetworkWait)
      • canvas is visible, enabled, and not masked,
      • canvas satisfies isCanvasDone()
      • if canvas is a TileGrid, it satisfies isTileGridDone()
      • if canvas is a TileLayout, it satisfies isTileLayoutDone()
      • if canvas is a ListGrid or body of a ListGrid, it satisfies isGridDone()
      • if canvas is a DynamicForm, it satisfies isFormDone()
      Note that for a form item in a DynamicForm, the DynamicForm must satisfy the third condition above, while the container widget of the element must satisfy the remaining conditions.
      Parameters:
      item - instance to test
      Returns:
      whether instance is 'clickable' as described above

    • isElementReadyForKeyPresses

      public static Boolean isElementReadyForKeyPresses(Canvas canvas)
      Given a DOM element, returns whether the associated Smart GWT Canvas or FormItem is ready to receive keyPress events from a Selenium test. Returns null if the locator is not valid or doesn't represent a valid Canvas or FormItem. Otherwise, returns true or false according as the conditions below are all satisfied:
      Parameters:
      canvas - DOM element to test (element obtained from canvas, form item, or Smart GWT locator if provided)
      Returns:
      whether element is 'ready' as described above

    • isElementReadyForKeyPresses

      public static Boolean isElementReadyForKeyPresses(FormItem item)
      Given a DOM element, returns whether the associated Smart GWT Canvas or FormItem is ready to receive keyPress events from a Selenium test. Returns null if the locator is not valid or doesn't represent a valid Canvas or FormItem. Otherwise, returns true or false according as the conditions below are all satisfied:
      Parameters:
      item - DOM element to test (element obtained from canvas, form item, or Smart GWT locator if provided)
      Returns:
      whether element is 'ready' as described above

    • isFormDone

      public static Boolean isFormDone(Canvas canvas)
      Returns whether the DynamicForm is in a consistent state with no pending operations. Returns null if the supplied argument is not a valid DynamicForm. Otherwise, returns true or false according as the conditions below are all satisfied:
      • page has finished loading
      • form has no pending delayed "set values" or "set values focus" operations
      • all contained items satisfy isItemDone()
      Parameters:
      canvas - canvas to test
      Returns:
      whether canvas is 'done' as described above

    • isGridDone

      public static Boolean isGridDone(Canvas canvas)
      Returns whether the ListGrid supplied or containing the supplied canvas is in a consistent state with no pending operations. Returns null if the supplied canvas is not a ListGrid or contained in a ListGrid. Otherwise, returns true or false according as the conditions below are all satisfied:
      • page has finished loading
      • no pending scroll operations
      • no pending filter editor operations
      • no unsaved edits to the grid records
      • no asynchronous regrouping operations are in progress
      • no outstanding fetch/filter operations are present for the ResultSet
      • no outstanding sort operations are present that will update the ListGrid
      Parameters:
      canvas - canvas to test
      Returns:
      whether canvas is 'done' as described above

    • isItemDone

      public static Boolean isItemDone(FormItem item)
      Returns whether the FormItem is in a consistent state with no pending operations. Returns null if the supplied argument is not a valid FormItem. Otherwise, returns true or false according as the conditions below are all satisfied:
      • page has finished loading
      • if the container widget of the item isn't the parent DynamicForm, then the container widget must satisfy isCanvasDone() (or isGridDone(), etc., as appropriate)
      • the item cannot have any fetches in progress for missing display/value field values
      • picklists (SelectItem or ComboBoxItem) cannot have any pending row fetches
      • any contained FormItems must satisfy isItemDone() themselves
      • any contained Canvas must satisfy isCanvasDone() (or isGridDone(), etc., as appropriate)
      Parameters:
      item - canvas to test
      Returns:
      whether item is 'done' as described above

    • isSystemDone

      public static boolean isSystemDone()
      Returns a boolean to indicate whether the current application is fully loaded with no pending operations, and is ready for user interaction. This includes the following checks: In addition to this the options parameter allows developers to check additional criteria as documented under SystemDoneConfig.

      Note: waitForSystemDone() will poll this method repeatedly to check for all pending actions being complete. The SystemWaitConfig parameter of that method allow the user to speciofy which actions to wait for. By default waitForSystemDone() will wait for all actions to complete, including queued redraws, timers, network requests, as well as page load and comopnent level isComponentDone checks.

      Depending on your application configuration, it is possible that this method will not be able to determine whether it is truly in a busy state. Some examples include:

      • If SystemDoneConfig.includeTimers is true and the application has a constantly running background timer to perform some ongoing polling action or similar, this method may return false even though the application is ready for interaction
      • Similarly if SystemDoneConfig.includeRedraws is true, and the application is using messaging, or some other mechanism to periodically refresh component(s) such that they are repeatedly marked as dirty, this method may return false even though the application is ready for interaction
      • If this application makes use of some third-party library to perform asynchronous actions, this method may return true even when the application is waiting on a response from this third-party tool.
      Returns:
      whether loaded page is 'done' as described above

    • isSystemDone

      public static boolean isSystemDone(SystemDoneConfig options)
      Returns a boolean to indicate whether the current application is fully loaded with no pending operations, and is ready for user interaction. This includes the following checks: In addition to this the options parameter allows developers to check additional criteria as documented under SystemDoneConfig.

      Note: waitForSystemDone() will poll this method repeatedly to check for all pending actions being complete. The SystemWaitConfig parameter of that method allow the user to speciofy which actions to wait for. By default waitForSystemDone() will wait for all actions to complete, including queued redraws, timers, network requests, as well as page load and comopnent level isComponentDone checks.

      Depending on your application configuration, it is possible that this method will not be able to determine whether it is truly in a busy state. Some examples include:

      • If SystemDoneConfig.includeTimers is true and the application has a constantly running background timer to perform some ongoing polling action or similar, this method may return false even though the application is ready for interaction
      • Similarly if SystemDoneConfig.includeRedraws is true, and the application is using messaging, or some other mechanism to periodically refresh component(s) such that they are repeatedly marked as dirty, this method may return false even though the application is ready for interaction
      • If this application makes use of some third-party library to perform asynchronous actions, this method may return true even when the application is waiting on a response from this third-party tool.
      Parameters:
      options - This parameter determines what specific outstanding actions are required to be completed before the system is considered "done" with current processing. If passed as a boolean rather than a SystemDoneConfig object, this determines whether the system should wait for pending redraws.
      Returns:
      whether loaded page is 'done' as described above

    • isSystemDone

      public static boolean isSystemDone(boolean options)
      Returns a boolean to indicate whether the current application is fully loaded with no pending operations, and is ready for user interaction. This includes the following checks: In addition to this the options parameter allows developers to check additional criteria as documented under SystemDoneConfig.

      Note: waitForSystemDone() will poll this method repeatedly to check for all pending actions being complete. The SystemWaitConfig parameter of that method allow the user to speciofy which actions to wait for. By default waitForSystemDone() will wait for all actions to complete, including queued redraws, timers, network requests, as well as page load and comopnent level isComponentDone checks.

      Depending on your application configuration, it is possible that this method will not be able to determine whether it is truly in a busy state. Some examples include:

      • If SystemDoneConfig.includeTimers is true and the application has a constantly running background timer to perform some ongoing polling action or similar, this method may return false even though the application is ready for interaction
      • Similarly if SystemDoneConfig.includeRedraws is true, and the application is using messaging, or some other mechanism to periodically refresh component(s) such that they are repeatedly marked as dirty, this method may return false even though the application is ready for interaction
      • If this application makes use of some third-party library to perform asynchronous actions, this method may return true even when the application is waiting on a response from this third-party tool.
      Parameters:
      options - This parameter determines what specific outstanding actions are required to be completed before the system is considered "done" with current processing. If passed as a boolean rather than a SystemDoneConfig object, this determines whether the system should wait for pending redraws.
      Returns:
      whether loaded page is 'done' as described above

    • isTileGridDone

      public static Boolean isTileGridDone(Canvas canvas)
      Returns whether the TileGrid is in a consistent state with no pending operations. Returns null if the supplied canvas is not a valid TileGrid. Otherwise, returns true or false according as the conditions below are all satisfied:
      • page has finished loading
      • the TileGrid (as a tileLayout) satisfies isTileLayoutDone()
      • the TileGrid has no pending layout animation operations queued
      Parameters:
      canvas - TileGrid to test
      Returns:
      whether TileGrid is 'done' as described above

    • isTileLayoutDone

      public static Boolean isTileLayoutDone(Canvas canvas)
      Returns whether the TileLayout is in a consistent state with no pending operations. Returns null if the supplied canvas is not a valid TileLayout. Otherwise, returns true or false according as the conditions below are all satisfied:
      • page has finished loading
      • the TileLayout (as a canvas) satisfies isCanvasDone()
      • the TileLayout is not currently animating any layout operations
      Parameters:
      canvas - TileLayout to test
      Returns:
      whether TileLayout is 'done' as described above

    • locateCanvasFromDOMElement

      public static Canvas locateCanvasFromDOMElement(Element element)
      Given an element in the DOM, returns the canvas containing this element, or null if the element is not contained in any canvas handle.
      Parameters:
      element - DOM element within in the page
      Returns:
      canvas containing the element, or null if none apply

    • setTestRoot

      public static void setTestRoot(Canvas canvas)
      Sets the implicit root canvas available in scLocators starting "//testRoot[]". Setting this property may enable one to use the same script to test identical widget hierarchies that are rooted under different base widgets.
      Parameters:
      canvas - the implicit root

    • uninstallLocatorShortcut

      public static void uninstallLocatorShortcut()
      Uninstalls the locator shortcut installed by installLocatorShortcut()

    • waitForElement

      public static void waitForElement(String locator, String callback)
      Waits for any in-progress actions to complete and then returns the element for the specified locator.

      By default this method will wait for the system to be done before calling getElement() with the specified locator. The waitStyle attribute of the options parameter may be used to instead wait for the locator to resolve to a clickable element.

      Parameters:
      locator - locator for the element to retrieve. See AutoTestLocator
      callback - callback to fire when the element has been retrieved - takes two parameters:
      • "element" (DOMElement) - the clickable element the locator resolved to (or null)
      • "done" (boolean) - this will be false if the method timed out
      . See Callback

    • waitForElement

      public static void waitForElement(String locator, String callback, ElementWaitConfig options)
      Waits for any in-progress actions to complete and then returns the element for the specified locator.

      By default this method will wait for the system to be done before calling getElement() with the specified locator. The waitStyle attribute of the options parameter may be used to instead wait for the locator to resolve to a clickable element.

      Parameters:
      locator - locator for the element to retrieve. See AutoTestLocator
      callback - callback to fire when the element has been retrieved - takes two parameters:
      • "element" (DOMElement) - the clickable element the locator resolved to (or null)
      • "done" (boolean) - this will be false if the method timed out
      . See Callback
      options - Options to configure the wait

    • waitForSystemDone

      public static void waitForSystemDone(String callback)
      Fires the provided callback when isSystemDone() returns true.

      By default this will wait for the following system actions to complete:

      Note that this behavior may differ from the default settings for isSystemDone()

      The options parameter gives developers finer grained control over what to wait for.

      Parameters:
      callback - Action to fire. This will take a single parameter "done" which will be set to false if the method timed out waiting for isSystemDone() to return true;. See Callback
      See Also:

    • waitForSystemDone

      public static void waitForSystemDone(String callback, SystemWaitConfig options)
      Fires the provided callback when isSystemDone() returns true.

      By default this will wait for the following system actions to complete:

      Note that this behavior may differ from the default settings for isSystemDone()

      The options parameter gives developers finer grained control over what to wait for.

      Parameters:
      callback - Action to fire. This will take a single parameter "done" which will be set to false if the method timed out waiting for isSystemDone() to return true;. See Callback
      options - options to configure the wait

    • getTestRoot

      public static Canvas getTestRoot()
      Gets the implicit root canvas available in scLocators starting "//testRoot[]". Setting this property may enable one to use the same script to test identical widget hierarchies that are rooted under different base widgets.
      Returns:
      the canvas currently set as the test root