What’s New in Mule 4.3

The 4.3 version of the Mule runtime engine provides a huge performance boost to your applications, while continuing to expedite application development on Anypoint Platform and to offer a resilient runtime for use cases that are vital to the functioning your organization.

MuleSoft recommends that you migrate your applications to Mule 4.3 to benefit from all the improvements that this new version provides.

Threading Model

Mule 4.3 handles threads differently than Mule 4.2 and 4.1.

In earlier versions, Mule owned and managed three different thread pools, each dedicated to a different Processing Type (CPU_LIGHT, IO, CPU_INTENSIVE). Tasks were submitted to these pools using the Proactor design pattern.

Starting with Mule 4.3, those thread pools have been combined into one unique thread pool, even though the proactor pattern is still applied.

See Execution Engine for complete details about the new threading model and upgrade instructions.

DataWeave Updates

The Mule 4.3.0 release includes DataWeave 2.3.0 features.

The update Syntax

This feature adds an easy way to update single values in nested data structures without requiring you to understand functional recursion.

The following array of objects serves as the input payload to DataWeave source.

Input:

[
   {
       "ID" : 4128506,
       "name": "Ken",
       "age": 30
   },
   {
       "ID" : 1823940,
       "name": "Tomo",
       "age": 70
   },
   {
       "ID": 9086582,
       "name": "Kajika",
       "age": 10
   }
]

The following DataWeave script uses the map function to iterate over the elements in the input array of objects. It uses the new update operator to modify two of the key-value pairs in each object: it concatenates (++) the string " (Christian)" to the value of a "name" key, if the value is "Tomo", and it increases the value of all "age" keys by 1.

Source:

%dw 2.0
output application/json
---
payload map ((user) ->
   user update {
       case name at .name if(name == "Tomo") -> name ++ " (Christian)"
       case age at .age -> age + 1
   }
)

The JSON output shows that the string " (Christian)" is appended to the value "Tomo" and that the "age" values are incremented by 1. Note that the values of the "ID" keys remain the same, as do the "name" key values that do not match "Tomo".

Output:

[
 {
   "ID": 4128506,
   "name": "Ken",
   "age": 31
 },
 {
   "ID": 1823940,
   "name": "Tomo (Christian)",
   "age": 71
 },
 {
   "ID": 9086582,
   "name": "Kajika",
   "age": 11
 }
]

For complete documentation, see Update Operator.

Streaming Features and Enhancements

XML Streaming:
The DataWeave reader for the XML format now supports streaming. You need to specify the collectionPath and streaming reader properties to activate streaming.

When streaming, the XML parser can start processing content without having all the XML content, which is particularly helpful when dealing with large content.

See XML Format for details.
Improved JSON Streaming:
The JSON reader now supports streaming the entire JSON document, instead of arrays only.

See JSON Format for details.

Literal Types

This feature adds DataWeave support for literal types. A literal type is a type that represents exactly one value, which means, for example, that DataWeave can represent the String value "hello" with the type hello. This feature makes function overloading very dynamic and when combined with Union types, it enables DataWeave to represent an enumeration of literal values as a type, for example:

type Weekdays = "Monday" | "Tuesday" | "Wednesday" | "Thursday" | "Friday"
type Days = Weekdays | "Saturday" | "Sunday"

fun greeting(day : "Monday") = "Best of luck in this new work week!"
fun greeting(day : "Friday") = "YAY, it's Friday! Have an awesome day!"
fun greeting(day: Weekdays) = "Have a nice day at work!"
fun greeting(day: Days) = "We hope you have a great day!"

The greeting function provides:

a general message ("We hope you have a great day!") on weekends (Saturdays and Sundays only).
a work- specific message ("Have a nice day at work!") on Tuesdays through Thursdays.
customized messages for the start and end of the week (Mondays and Fridays).

Note that DataWeave uses the first function it finds that accepts the arguments of the function call, based on the order in which the functions are declared in the script.

For complete documentation, see Literal Types.

Period and DateTime Consistency

This feature adds an easy way to operate with periods, transform the result into different units, and decompose the period into its different parts.

The following example shows how to deconstruct a Duration value. It defines a period variable in the header that subtracts one DateTime value from another. The script decomposes the result of the operation by selecting the hours, minutes, and secs values from the resulting period value.

Source:

%dw 2.0
output application/json
var period = (|2010-12-10T12:10:12| - |2010-12-10T10:02:10|)
---
{
   hours:  period.hours,
   minutes:  period.minutes,
   secs:  period.secs,
}

The JSON output shows the resulting number of hours, minutes, and secs values selected from period.

Output:

{
  "hours": 2,
  "minutes": 8,
  "secs": 2
}

For complete documentation, see Period.

Custom MIME Types

This feature adds a way to define the output MIME type separately from the output data format. It also provides a simpler way of defining output and input formats:

%dw 2.0
output json
---
{
	Hello: “there”
}

For documentation, see Setting MIME Types.

Performance Improvements

DataWeave improves the performance of memory usage, functions (like groupBy), and the internal execution engine in this release.

Elimination of common subexpressions:
Common subexpression elimination is included to optimize scripts. Now, DataWeave optimizes the scripts by defining internal variables to represent subexpressions that are repeated, thereby replacing the repeated executions of that subexpression with a single execution.
Source
```
%dw 2.0
output json
---
{
   code: payload.message.code,
   message : payload.message.value,
   user : payload.message.users[0].name,
   contact: payload.message.users[0].email
}
```
Internally, the script in the example above is treated as shown the following example, where variables are defined for each repeated expression so that they execute only once.
Internal Representation
```
%dw 2.0
output json

var fakeVariable1 = payload.message
var fakeVariable2 = fakeVariable1.users[0]
---
{
   code: fakeVariable1.code,
   message : fakeVariable1.value,
   user : fakeVariable2.name,
   contact: fakeVariable2.email
}
```
Memory management improvements:
Memory management is now centralized. Several memory management strategies are included, such as pooling direct memory or using heap memory exclusively.

DataWeave Modules

DataWeave introduces the Types (dw::core::Types) module to support type introspection. This new module includes a number of functions and new types. See Types (dw::core::Types) for details.

DataWeave also introduces a number of new functions in existing modules:

Core (dw::Core) module: entriesOf, keysOf, namesOf, valuesOf.
Arrays (dw::core::Arrays) module: firstWith
Objects (dw::core::Objects) module: everyEntry, someEntry, and takeWhile; dropWhile also includes a new function that iterates over and ignores items in an array until a condition becomes true.
Strings (dw::core::Strings) module: withMaxSize enables you to specify a maximum size for a given value. A value that exceeds the size limit will be truncated. If the size is no greater than the limit, the value remains the same.

Deprecations:
The entrySet, keySet, nameSet, valueSet functions in the Objects module are deprecated in this release and replaced with the functions entriesOf, keysOf, namesOf, valuesOf in the Core module.

Intersection Type

DataWeave introduces the Intersection type. The Intersection type appends Object types.

The syntax for intersection type is:

TypeA & TypeB & ...

For documentation, see Type System

Improvements to Parsing Error Messages

Parsing error messages now reflect more accurately missing tokens, incomplete definitions, and other common issues. Tips and examples of how to solve the issues are also included.

New DataWeave Reader and Writer Properties

DataWeave introduces a number of new and enhanced reader and writer properties for supported DataWeave formats.

Binary (application/octet-stream) format: encoding
DataWeave (application/dw) format: onlyData writer property and all the reader properties (externalResources, javaModule, and onlyData)
JSON (application/json) format: writeAttributes
XML (application/xml) format: escapeCR writer property and several reader properties (collectionPath, maxAttributeSize, optimizeFor, supportDtd, streaming)
YAML (application/yaml) format: maxEntityCount reader property

For descriptions, see the links to these formats in Supported Data Formats.

Core Components

The following Core components have been improved, adding new functionality and capabilities.

Until Successful component:
Max Retries (maxRetries) and Milliseconds Between Retries (millisBetweenRetries) now support a number or an expression that resolves to a number.

For complete documentation, see Until Successful Scope.
Batch Aggregator component adds the preserveMimeTypes attribute:
Aggregated records are passed into the aggregator as an array containing each record’s payload. However, the MIME types associated with those payloads are not preserved by default. You can preserve record’s MIME types by specifying the preserveMimeTypes attribute in a batch aggregator. This change makes it easier to handle records from a payload that is in a structured text format, such as JSON or XML.

For complete documentation, see Refining Batch Steps Processing.

API Gateway

API Gateway now includes a disablePolicies annotation. Setting it to true enables you to block the application of policies to the requestor.

Also, the release introduces several enhancements to resiliency, including a better backoff mechanism and improvements of the policy deployer mechanism.