Asynchronous API Programming

Overview

Whilst M-Files COM API calls are typically synchronous (i.e. the application doesn’t continue until the API call completes), the M-Files Vault object exposes a property named Async which can be used to make asynchronous calls. Using asynchronous calls - rather than synchronous calls, which wait for the process to finish before continuing - ensures that the user interface doesn’t appear to “lock”, and the user experience remains fluid.

Using asynchronous calls are currently optional when targeting the M-Files Desktop client. It is mandatory when targeting M-Files Classic Web.

Altering synchronous to asynchronous calls

The following is a standard synchronous call to CheckOut. Once the call completes, the return value of the call (the checked-out ObjectVersion) is assigned to the objectVersion variable and the code continues.

// Check out the object (objver) with synchronous call.
var objectVersion = vault.ObjectOperations.CheckOut( objver.ObjID );

// Process checked out object.

Altering code to use an asynchronous approach involves two steps:

  1. Alter the call to use Vault.Async rather than Vault.
  2. Provide one or more callback functions to handle when the asynchronous method call completes.

Using an asynchronous approach, a callback function is passed after the standard arguments. At some point in the future when the call to CheckOut completes, the callback function is called and the return value is available as objectVersion.

vault.Async.ObjectOperations.CheckOut( objver.ObjID,      
	function ( objectVersion ) {
		// Process checked out object.
	} );

It is important to note that the call to CheckOut completes almost immediately. If later lines of code require the return value of the call then they must be moved inside the callback function.

Multiple callback functions

Each asynchronous call can be provded with up to three callback functions, in this order:

  1. A callback function for a successful call.
  2. A callback function for a failed call.
  3. A callback function for cleanup. This callback function is always executed, regardless of whether the call was successful or not.

The successful callback

The success callback is executed if the API call completed successfully. It always receives a single parameter which is the return value of the call. For example: GetProperties defines two parameters (the ObjVer to get the properties of, and the whether to update from the server or not) and returns a PropertyValues object, so the parameter would be a single object of type PropertyValues:

vault.Async.ObjectPropertyOperations.GetProperties(objVer, 
	false, // UpdateFromServer = false
	function(propertyValues){
		// Handle the properties.
	})

If only a success callback is provided (i.e. failed callback is not provided) then exceptions will not be visible to the user.

The failed callback

The failed callback is executed if the API call does not complete (i.e. it throws an exception). Only one of the success and failed callbacks will be executed, depending on whether the call succeeds or fails. The failed callback always receives three parameters:

  1. The short error message.
  2. The long error message.
  3. An object detailing the exception.
vault.Async.ObjectPropertyOperations.GetProperties(objVer, 
	false, // UpdateFromServer = false
	function(propertyValues){
		// Handle the properties.
	}),
	function(shorterror, longerror, errorobj){
		// Handle the exception
	})

Reporting exceptions

To report an exception to the user, MFiles.ReportException can be used. This is passed the third parameter of the failed callback, which is the object which details the exception:

vault.Async.ObjectPropertyOperations.GetProperties(objVer, 
	false, // UpdateFromServer = false
	function(propertyValues){
		// Handle the properties.
	}),
	function(shorterror, longerror, errorobj){
		// Show the exception to the user.
		MFiles.ReportException(errorobj);
	})

The cleanup callback

The cleanup callback is always executed - if provided - regardless of whether the call completes or excepts. It is always executed after the successful callback or the failed callback and receives no parameters.

vault.Async.ObjectPropertyOperations.GetProperties(objVer, 
	false, // UpdateFromServer = false
	function(propertyValues){
		// Handle the properties.
	}),
	function(shorterror, longerror, errorobj){
		// Show the exception to the user.
		MFiles.ReportException(errorobj);
	}),
	function(){
		// Clean up any resources required.
	}

An important note on supported object types

Asynchronous calls clone each input parameter and the return value, therefore all object types being passed in and out of the API call must be cloneable. Cloneable object types can be checked using the M-Files COM API reference to see whether each object has a Clone method.

For example: PropertyValues has a Clone method, so methods using it can be called asynchronously. However, SourceObjectFiles does not have a Clone method, so methods using it cannot be called asynchronously.

Note that M-Files Classic Web has specific wrappers around some of these uncloneable objects; SourceObjectFiles can be used asynchronously in M-Files Classic Web, but cannot be used in M-Files Desktop client. To work around this, check the current platform and call methods such as CreateNewObject synchronously on M-Files Desktop and asynchronously on M-Files Classic Web.