Salesforce Connector Release Notes - Mule 4

Salesforce Connector now supports Salesforce API v55.0.

Added support for Salesforce API v54.0.

Added the Cache events in memory parameter for the Replay Topic Listener and Replay Channel Listener sources.

Added support for Salesforce API v53.0.

Added the Query Xml Stream operation.

The connector threw a NumberFormatException when using an empty string in proxy-related system properties. (SE-21968)

The Update operation did not update picklist values. (SE-21243)

Fixed reported security vulnerabilities by updating the following libraries: (CONN-8620)

The On Deleted Object source was not generating events. (CPLAT-1351)

The Create Batch operation did not serialize null values properly. (SE-21393)

Added support for Salesforce API v52.0.

Salesforce Connector was not able to parse the CANNOT_EXECUTE_FLOW_TRIGGER status code returned from the batch job result. (SE-19814)

Optimized the connector to improve its overall stability and performance.

There was a NullPointerException while resolving metadata for an account that didn’t have a Change Data Capture event channel set. (CONN-8146)

The Query Result Stream operation returned only the first set of results when reading the result of a batch query that contained multiple response IDs. Use the two new operations Query Result List and Query Result Stream By Id instead. The Query Result Stream operation is now deprecated. (SE-20508)

The On New Object source returned only a partial amount of records when the queried item had a large number of records. (SE-20729)

When using the Salesforce date mapping, the mapping failed and the "Cannot coerce Object" error message occurred. (COCN-74)

Synchronization issue that caused data to not persist in Object Store for the Replay Streaming Channel and Replay Topic sources. (SE-20296)

The object types of the sources were not filtered correctly to contain only queryable objects. (CCONN-768)

The connector did not interpret API errors for the Query operation correctly. (SE-20227)

Query operations that used the special character $ in fields failed. (SE-20281)

Made the following library updates to fix reported security vulnerabilities: (CONN-7951)

The connector reconnection mechanism for sources sometimes created the Bayeux client multiple times, which caused the connector to receive platform events twice. (SE-19315)

Some internal resources were not implemented correctly, which might have resulted in higher memory use than in previous connector versions. (SE-18711)

Added the optional Assignment Rule Id field to the Create Job Bulk Api V2 operation.(SE-19721)

Commons Codec commons-codec:commons-codec:1.14 to commons-codec:commons-codec:1.15

Guava com.google.guava:guava:19.0-jre to com.google.guava:guava:30.1-jre

Junit junit:junit:4.12.0 to junit:junit:4.13.1

Added support for Salesforce API v51.0.

The connection now supports setting a login request timeout value that forces an exception to be raised if the API request to obtain a session takes too long.

The Replay Streaming Channel source went into a resubscription loop after it encountered a Buffering capacity exceeded error. The connector now uses a streaming transformer for the events being processed by the source. As a result, the field used to specify the size of the event queue was deprecated. (SE-18873)

Parameter values for the Search operation did not resolve properly when the search query contained newline. (SE-19254)

The connector did not support a timeout configuration for login requests. You can now set a timeout for login requests by using the new Login Request Timeout field. (SE-19535)

Upgraded XMLBeans org.apache.xmlbeans:xmlbeans:2.5.0 to org.apache.xmlbeans:xmlbeans:3.1.0. (CONN-7634)

Parameter values for the Search operation did not resolve properly when newline was used in the search query. (SE-19254)

Upgraded the Object Store Connector dependency version to 1.1.6 to avoid duplicate object store errors in a cluster environment. (SE-18227)

Logging for sources now identifies the target Salesforce instance. (SE-19174)

The Query Result Stream operation failed with an Out Of Memory Error. (SE-19258)

Parameter values for the Search operation were not resolved properly. (SE-19254)

The List metadata and Read metadata operations were not correctly managing the Settings type. (SE-18716)

The OAuth 2.0 connection was not properly managing token expiration. (CONN-7346)

Upgraded Apache CXF org.apache.cxf:cxf-rt-wsdl:3.3.4 to org.apache.cxf:cxf-rt-wsdl:3.4.1.

The connector stopped working intermittently when using a dynamic configuration. (SE-18473)

The 'Type' parameter values were not populated for multiple operations. (SE-18912)

Added a new advanced parameter, Throw Exception When Organization Daily Limit Exceeded, which enables the connector to throw an exception if the limit is exceeded at deployment time.

Enhanced logging messages for the Streaming sources.

The Query operation was not working properly when a dynamic query with more than nine parameters was used. (SE-18478)

The Streaming sources were not throwing an exception when the daily API limit was exceeded, which is fixed with the addition of a new advanced parameter Throw Exception When Organization Daily Limit Exceeded. (SE-17568)

Added support for Salesforce API v50.0.

The Create job bulk api v 2 and Create query job bulk api v 2 operations now support the Sforce-Call-Options header to specify a default namespace.

The Create job bulk api v 2 operation now supports the hardDelete value in the Operation field.

The Batch result stream operation now supports specifying an output MIME type based on the content type provided when creating the batch.

The result of the Batch result stream operation could not be processed successfully. The new, optional Output Mime Type field fixes this issue by enabling you to specify the content type provided when creating the batch. (SE-17872)

Updating a Date field value decreased the date by one day. (SE-18095)

Processing events using the Replay Channel listener or the Replay Topic listener created many objects in memory that were not properly cleared, which caused applications to crash. (SE-18141)

The Describe SObject operation returned a null field type for some fields. (SE-17817)

Warning messages were displayed in the logs for an Invalid Session ID found in SessionHeader error. (SE-17784)

Test connection was successful, even though there were handshake issues for the streaming sources. (SE-17680)

Mutual TLS did not work for the streaming sources. (SE-17991)

The Create operation did not return the duplicated record ID. (SE-17697)

The Replay topic listener and Replay channel listener sources threw Object-Store-related exceptions when the Resume from the Last Replay Id feature was not enabled. (CONN-6984)

Resume from the Last Replay Id feature of the Replay sources did not work in a runtime cluster. (SE-17312)

Update Metadata operation failed when using a custom field. (SE-17355)

The Retrieve Job Unprocessed Results BulkV2 operation is added.

The Retrieve Job Failed Results BulkV2 and Retrieve Job Successful Results BulkV2 operations now return the original fields of the records

Logging improvements over all operations and connection types.

Subscribe Streaming Channel and Replay Streaming Channel sources now display all types of events available in the configured organisation:

The Query Result Stream operation was returning a ClientInputError because of a mismatched content type. The optional Content Type field can now be used with this operation for jobs that have different content types. (SE-17310)

Upgraded Commons Codec commons-io:commons-io:2.5 to commons-io:commons-io:2.7.

Added support for Salesforce API v49.0.

Added the Merge Multiple operation to enable merging up to 200 records in a single SOAP call.

Upgraded the following dependencies to fix vulnerability issues:

Creating objects with fields that accept array values failed. (SE-16444)

DataSense resolution failed for objects that have relationships with other objects. (SE-16373)

Reconnection did not work for the Invoke APEX SOAP Method operation. (SE-16419)

The Update Metadata operation failed for the Custom Metadata Type. (SE-16074)

Added support for multiple keys in the JKS file specified for the SAML and JWT authentication mechanisms. Use the new Certificate Alias field to specify which certificate should be used. (CONN-6019, CONN-6020)

Added support for custom headers for the Invoke APEX SOAP operation. (SE-15884)

Bulk API V2 operations now return the response headers and status as message attributes.

[403::Unknown client] messages showed up in the logs. The logging level for this message is now debug. (SE-16152)

For the Convert Lead operation, the ownerId field could not be specified. (SE-16169)

The Get Query Job Results-Bulk Api v2 operation threw an OutOfMemory error. A new Max Records Per Page field was added to control the number of records on each page retrieved from the API. (SE-16298)

The Replay Channel listener input source detected an ObjectDoesNotExistException and received platform events multiple times. (SE-15814)

Metadata operations failed for DocumentFolder, EmailFolder, ReportFolder, and DashboardFolder. (SE-16057)

Salesforce Connector failed when used with FTP Connector. (CONN-6077)

Upgraded the following dependency to fix vulnerability issues:

Added support for API v48.0.

Added support for subscribing to these newly-added platform events: FOStatusChangedEvent, OrderSummaryCreatedEvent, and OrderSumStatusChangedEvent.

Added support for using the new RecordVisibilityContext clause when using the WITH filteringExpression in a query.

The Get Job State Bulk Api V2 operation threw a SALESFORCE:UNKOWN exception when there were errors processing the job. The operation now returns the actual job state with all of the details provided by the API. (SE-15419)

Null payload values sent to the Invoke Apex REST method were disregarded. A new configuration parameter called Include null values enables you to choose whether to send request parameters that have null values. (SE-15357)

The Get Query Job Results Bulk Api V2 operation did not handle pagination correctly. (SE-15726)

The OAuth JWT connection did not support custom domains for the audience URL. A new, optional field called audienceUrl was added to support this scenario. (SE-14071)

Refactored input sources to process messages correctly and fixed the interaction with the object store. (CONN-5477, SE-15212, SE-15066)

Create operations did not support Long values for the object fields. (SE-14890)

Fixed the APEX invocation issues related to GET requests. (SE-15097 and SE-15039)

The operations for querying job results did not parse the CSV response properly. (SE-14560)

Upgraded the following dependencies to fix vulnerability issues:

Calling the managed package Apex REST API did not work. (SE-14774)

Enabling a reconnection strategy for the connector resulted in different behavior when using different runtimes. (SE-14268)

The scheduler responsible for keeping the connection alive shut down for cached OAuth connections, even though the keep alive feature was enabled. (SE-14647)

The scheduler responsible for keeping the connection alive shut down for cached Basic Authentication connections, even though the keep alive feature was enabled. (SE-14682)

Upgraded some

Removed deprecated connection types

Made the following changes to some operations:

Redesigned the Replay ID functionality

Changed the Mule TLS Context usage

Improved the error types

Removed the dependency on the Dozer library

Removed the keep alive feature as a configurable option because it is no longer needed

Updating a Date field value decreased the date by one day. (SE-18095)

The Create Single operation did not return the duplicated record ID. (SE-17697)

The Delete Job Bulk V2 operation returned NULL when successful. (CONN-6618)

Reconnection was not working for the Invoke APEX SOAP Method operation. (SE-16419)

The Update Metadata operation was failing for the Custom Metadata Type. (SE-16074)

Logs were flooded with warning messages when the connection expired. (SE-15855)

The Replay Channel listener input source detected an ObjectDoesNotExistException and received platform events multiple times. (SE-15814)

Added support for API v48.0.

Added support for subscribing to these newly-added platform events: FOStatusChangedEvent, OrderSummaryCreatedEvent, and OrderSumStatusChangedEvent.

Added support for using the new RecordVisibilityContext clause when using the WITH filteringExpression in a query.

Reconnection failed when using the Basic Username Password authentication type. (SE-15694 and SE-15713)

The Get Bulk Job State v2 operation threw a SALESFORCE:UNKOWN exception when there were errors processing the job. The operation now returns the actual job state with all of the details provided by the API. (CONN-5721)

The connector did not honor the read timeout value set in the connection provider for the Invoke APEX Soap Method operation. (SE-15727)

The Get Query Job Results Bulk Api V2 operation did not handle pagination correctly. (CONN-5931)

Mutual authentication failed when a custom login domain was used. (SE-15552)

The issue that caused the Invoke Apex Rest Method operation to throw an INVALID_SESSION_ID exception even though the reconnection strategy was enabled is fixed. (SE-15429)

The Apache CXF dependency has been upgraded to version 3.3.5 to fix a security vulnerability issue. (SE-15353)

The connector threw an INVALID_TYPE_ON_FIELD_IN_RECORD error when trying to create an entity that contained fields of type NUMBER or CURRENCY. (SE-14890)

The OAuth JWT connection did not support custom domains for the audience URL. A new, optional field called audienceUrl was added to support this scenario. (SE-14071)

The Invoke Apex Rest method failed. (SE-15039)

Platform events did not display in the Streaming channel list of the Subscribe channel input source. (SE-14953)

Salesforce Replay Channel operation failed to process headers from HTTP Response. (SE-14958) Changes to fix SE-13603 were reverted to fix this regression issue.

Salesforce Replay Channel received messages that were already consumed. (SE-15005) Changes to fix SE-13603 were reverted to fix this regression issue.

Refactored the mechanism interacting with Object Store to fix the issue that occurred when using more than one salesforce:replay-channel in an application in CloudHub, which caused a 429 status code. (SE-13603)

The issue that caused Salesforce Connector 9.8.x to consume more API calls than version 9.7.x is fixed by caching the API response for some requests to avoid multiple calls when they are not needed. (SE-14655)

The issue that caused the Salesforce Query Job results to throw an ArrayIndexOutOfBoundsException because the connector wasn’t parsing the API response properly is fixed. (SE-14560)

The issue that caused the Describe object operation to return fieldType as null for all objects was fixed by mapping the API response properly. (SE-14661)

The issue that caused calling the managed package Apex REST API not to work is fixed. (SE-14774)

The scheduler responsible for keeping the connection alive shut down for cached OAuth connections even though the keep alive feature was enabled. (SE-14647)

The scheduler responsible for keeping the connection alive shut down for cached Basic Authentication connections even though the keep alive feature was enabled. (SE-14682)

The connector could not connect to Salesforce when using Mutual Authentication. (SE-14651)

Enabling a reconnection strategy for the connector resulted in different behavior when using different runtimes. (SE-14268)

Salesforce Connector Retrieve Metadata operation returned API version 37.0 though requesting 47.0. (SE-14393)

Salesforce Connector Upsert operation not adding custom headers to SOAP call. (SE-14384)

Because the Replay Topic and Replay Channel sources use the same object store, they have the same limitations as the object store component. For example, depending on the object store subscription and the application throughput, you might receive related 429 (conflict) errors, such as this one:

Upsert metadata was failing for picklist values. (SE-14043)

Fixed security vulnerabilities issues. (CONN-2878)

Updated File connector dependency for the Salesforce connector to the latest version. (CONN-3078)

Upgraded the Object Store dependency to the latest version. (CONN-3736)

Salesforce disconnects after a few hours. When a cached connection type was used, the reconnection caused a running source to fail. In version 9.7.10, a fix for this issue was provided but there was a NullPointerException occurring when the keep alive feature parameters were empty. (SE-12904)

Salesforce Connector relied on an outdated version of the Java module that caused compatibility issues with other modules. The Java module was removed from the Salesforce Connector dependencies. (SE-13470)

Calls to the Salesforce Create Job Bulk v2 operation with a large volume of data caused java.lang.OutOfMemoryError: Java heap space error. This issue has been fixed by changing the HTTP client to the one in the Mule SDK. To enable this functionality the mule.http.requestStreaming.enable property has to be set as a VM argument.

Failed to upsert metadata. (SE-13472)

Create operation with a parent relationship failed. (SE-12755)

Salesforce disconnects after a few hours. When a cached connection type was used, the reconnection caused a running source to fail. (SE-12904)

When using the Replay Topic or Replay Channel operations, if you have multiple workers deployed in CloudHub, duplicate messages occur. You can work around this by modifying your app to screen out duplicate messages, or you can provide an app as a receiver, and use a shared queue to pass messages to process between workers in your app. (SE-9992)

Fixed issue with connecting to metadata operations. (CONN-3069)

The Invoke Apex REST API operation calls to the Salesforce connector did not create new connections for each request in Mule 4. (SE-12635)

The Get Updated Objects operation did not return updated records. This was caused by the Salesforce API ignoring the seconds value for the API’s startDate and endDate fields. See the Arguments section in the Salesforce getUpdated operation. This was fixed so that when two or more subsequent calls occur in less than a minute, each call after the first returns an empty array to avoid duplicates. Messages are not lost because the time intervals are calculated based on the first time this operation is called and because the API disregards the seconds value. (SE-12253)

OAuth v2.0 failed to get the correct session and threw INVALID_SESSION_ID after being idle. (SE-10988)

Upgraded the Object Store dependency to the latest version. The Salesforce connector now uses Object Store connector version 1.1.3 internally. (CONN-2519)

Exposed the Salesforce session key from the Salesforce connector. This is now fixed and the session key is no longer visible in the logs. (CONN-2511)

Not enough information was provided for errors. Exceptions had the wrong error type and errors for an incorrect or missing message were fixed. (CONN-1183)

Salesforce connector autoreplays after restarting the application in CloudHub. (SE-9992)

Salesforce connector OAuth v2.0 fails to get the right session and throws INVALID_SESSION_ID after being idle. This issue still appears for the sources and operations that use the Bulk API and the Apex API. (SE-10988)

Salesforce connector replayId feature gap between 3.x and 4.x. (SE-11258)

Salesforce connector autoreplays after restarting the application in CloudHub. (SE-9992)

Salesforce connector OAuth v2.0 fails to get the right session and throws INVALID_SESSION_ID after being idle. This issue still appears for the sources and operations that use the Bulk API and the Apex API. (SE-10988)

Salesforce connector Get Updated Objects is not returning updated records. (SE-12253)

OAuth v2.0 handles the refresh_token value for Core API and Metadata API operations. (SE-11869)

Salesforce connector replayId feature gap between 3.x and 4.x. (SE-11258)

Salesforce connector autoreplays after restarting the application in CloudHub. (SE-9992)

Salesforce connector OAuth v2.0 fails to get the right session and throws INVALID_SESSION_ID after being idle. This issue still appears for the sources and operations that use the Bulk API and the Apex API. (SE-10988)

Salesforce connector replayId feature gap between 3.x and 4.x. (SE-11258)

Salesforce connector autoreplays after restarting the application in CloudHub. (SE-9992)

Salesforce connector OAuth v2.0 fails to get the right session and throws INVALID_SESSION_ID after being idle. This issue still appears for the sources and operations that use the Bulk API and the Apex API. (SE-10988)

Metadata propagation error. (SE-11757)

Batch aggregator executed an Upsert operation and did not return the right number of failed records if an error occurred. (SE-10501)

NullPointerException occurred when resolving metadata. (CONN-1938)

OAuth v2.0 failed to get the right session and threw INVALID_SESSION_ID after being idle. This was fixed partially for core operations (Create, Retrieve, Upsert, Update, Merge, Delete, Convert Lead, Find Duplicates, and others). (SE-10988)

Vulnerability reported for Salesforce connector 9.2.2 - Part of the vulnerabilities reported were fixed by upgrading the dependencies that the connector uses. (SE-9824)

Salesforce connector replayId feature gap between 3.x and 4.x. (SE-11258)

Salesforce connector autoreplays after restarting the application in CloudHub. (SE-9992)

Salesforce connector OAuth v2.0 fails to get right session and throws INVALID_SESSION_ID after being idle. This issue still appears for the sources and operations that use the Bulk API and the Apex API. (SE-10988)

Metadata issues for sources - The DataSense keys and output were wrong in the SubscribeTopic, SubscribeChannel, ReplayTopic, and ReplayChannel sources. With this version, the metadata has been fixed. (CONN-1918)

Duplicate values with objectType case - Issues in the input and output DataSense for different objects where some of the fields were duplicated. (CONN-451)

Vulnerability reported for Salesforce connector 9.2.2 - Part of the vulnerabilities reported were fixed by upgrading the dependencies the connector uses. (SE-9824)

Batch aggregator with Salesforce Connector executing an upsert operation does not return the right number of failed records when something went wrong. (SE-10501)

Salesforce connector replayId feature gap between 3.x and 4.x. (SE-11258)

Salesforce connector autoreplays after restarting the application in CloudHub. (SE-9992)

Salesforce Subscribe topic connector is disconnected - We have upgraded the CometD library org.cometd.java:cometd-java-client:3.0.10 to a newer version org.cometd.java:cometd-java-client:3.1.5 per a Salesforce recommendation, otherwise there might be connectivity issues. Also we have fixed an issue related to "Resume from last replay Id" feature of the Replay Topic and Replay Channel sources where if there were no events for 24 hours and a connection issue appeared, the reconnection stopped working. (SE-11365)

Salesforce connector memory leak - There was an issue in cached connection providers where some objects could not be cleared properly and this was leading to a memory leak. (CONN-1625)

Vulnerability reported for Salesforce connector 9.2.2 - A part of the vulnerabilities reported were fixed by upgrading some of the dependencies the connector uses. (SE-9824)

Batch aggregator with Salesforce Connector executing an upsert operation does not return the right number of failed records when something went wrong. (SE-10501)

Salesforce connector replayId feature gap between 3.x and 4.x. (SE-11258)

Salesforce connector autoreplays after restarting the application in CloudHub. (SE-9992)

Salesforce invokes Apex metadata request for each invoke. Now the connector does not fetch all the metadata to validate the connection, it only performs a Get Server Timstamp operation which should tell the app if reconnection is necessary or not. (SE-10925)

Vulnerability reported for Salesforce connector 9.2.2. A part of the vulnerabilities reported were fixed by upgrading some of the dependencies the connector uses. (SE-9824)

Batch aggregator with Salesforce Connector executing an upsert operation does not return the right number of failed records when something went wrong. (SE-10501)

Salesforce connector replayId feature gap between 3.x and 4.x. (SE-11258)

Salesforce connector autoreplays after restarting the application in CloudHub. (SE-9992)

Salesforce definition for a relationship (__r usage) threw error. Fixed an issue regarding object relationships in Salesforce. There was a bug causing troubles when creating objects with custom fields that had relationships to other objects. (SE-10599)

Salesforce Connector return wrong metadata when include a subquery with a list. Fixed output metadata issues for the query operation when there was a subquery included to retrieve related objects. (SE-10255)

Subscribe topic operation throws Invalid SObject Type: null error. Fixed metadata issues for the Subscribe Topic source. There was an exception thrown when trying to retrieve metadata for different object types. (CONN-1660)

Salesforce Metadata create order issue. There was an issue with reference fields in input metadata, that now is fixed. (SE-10375)

Salesforce connector create job bulk v 2 operation content type field is not valid. The only valid content type for this operation is CSV. We could not remove the parameter or the other content types, but the field was moved to the Advanced tab and a proper Summary was added for it. The operation is also fail fast (fails without making an API call) if a type other than CSV is specified. (SE-11043)

Removed the (Deprecated) keyword in the name of the OAuth 2.0 Connection which was added by mistake in a previous release. The connection was not changed and can be used the same way as up until now.

Salesforce invokes Apex metadata request for each invoke. (SE-10925)

Batch aggregator with Salesforce Connector executing an upsert operation does not return the right number of failed records when something went wrong. (SE-10501)

Salesforce connector replayId feature gap between 3.x and 4.x. (SE-11258)

Salesforce connector autoreplays after restarting the application in CloudHub. (SE-9992)

Added new versions of connection providers for JWT and SAML, because the existing ones do not match the Salesforce API behavior (one session with the same credentials at a time).

Improved connection performance when executing operations by making better use of the Mule SDK’s thread pools.

Fixed an issue where the connector was unable to handle empty body requests to Apex services.

Fixed an issue where for the Retrieve operation when the specified ID was invalid the connector was trying to reconnect because of a wrong exception that was thrown.

Metadata for Apex classes didn’t work if an instance had more than 200 records. Fixed by using the search service to query the instance for Apex classes instead of a simple query.

Null values were set as blank by DataWeave when obtained from the Salesforce Connector. Fixed by changing the returning JSONObject.NULL fields to null.

Improved the API call that validates the connection to improve connector performance.

Invalid Session ID and Reconnection Strategy works only after multiple invalid requests when this connector performs a concurrent query. Fixed by implementing a new CachedOauthUserPassConnectionProvider with keep-alive functionality.

GetSessionId was not returning the correct ID after the session was invalidated, because it was not throwing the correct exception. This was fixed by adding a call to the API, that returns a ConnectionException if the session ID is invalid

The operation "Create Content Version" was failing if the payload was of binary data type. This was fixed by checking the data type before casting to String, so no exception is thrown

Apex method call recognized an output field as a date if the return string has a similar format as Date. Methods in Apex come with a return type parameter which is now included in the condition which converts a string to a date type.

Invoking Apex REST was not reconnecting when a session expired. Fixed by properly handling exceptions thrown by the Apex API.

Fixed problem that occurred when attempting to create a new Content Version that had data as a Base64 format picture, it is not required to decode the Base64 string since the decoding is handled by the connector.

The clearNullFields config parameter was not used properly and its functionality would have never worked. Fixed the way clearNullFields is handled and now works as intended.

Fixed Salesforce Connector not reconnecting when a Bulk operation session expired. Fixed by changing the exception thrown in this case.

Retrieve record failure Bulk v2 not failing because of column delimiter. Fixed by adding handling for a column delimiter for the retrieve record failure Bulk v2.

Invalid Session ID and Reconnection Strategy works only after multiple invalid requests when the Mule connector performs a concurrent query. Fixed by implementing a new CachedBasicConnectionProvider with keep-alive functionality.

The BasicConnectionProvider was deprecated because a PoolingConnectionProvider does not have sense in the context of Salesforce’s way of handling multiple logins, which returns the same sessionId.

Username Password connection provider was improved by handling invalid connections currently in the pool, so before executing an operation, the connector ensures that the operation uses a valid connection.

Fixed an issue on Query operation where the connector was not able to reconnect after the session became invalid

The Query operation was not handling properly the exceptions thrown by the Salesforce API.

Fixed an issue in the Read Metadata operation where the connector was unable to return picklist values.

Fixed the problem of error message mapping in the Create Bulk v2 operation.

Fixed an issue in the Create Batch Stream operation where the connector was not able to reconnect after the session became invalid.

Fixed an issue in the Query operations where the execution was stopping if the query was invalid. Now a proper error is raised.

Fixed an issue where the connector Attachments were not supporting base64 encoded body.

Fixed an issue on create ContentVersion object that was failing because the Content Data field was not serialized properly.

Fixed an issue for the Batch info list, Batch info, and Job info operations that were going into an infinite loop because they were defaulting to XML content type. Optional parameters were added so the user can specify the content type for the entity so that it will not fail any more.

Get updated objects failed with UNKNOWN error when no records were retrieved. This is no longer the case.

CustomObjectTranslation was not working for the Read Metadata operation. This was fixed.

Update the connector to support Salesforce Spring 2018 (API v42)

Support Bulk API v2

After a time that the connector was not being used, customers got the following error:

Root Exception stack trace:
[UnexpectedErrorFault [ApiFault exceptionCode='INVALID_SESSION_ID'
exceptionMessage='Invalid Session ID found in SessionHeader: Illegal Session. Session not found, missing session hash: jiVw2Gqdz4jdt5QfJJcVLbEHzg4NFOAPHpWosGrWyUA=
This error usually occurs after a session expires or a user logs out.'

The customer couldn’t make APEX call without parameters.

The customer couldn’t make APEX call without parameters with array payload.

When using the Salesforce connector inside a batch job that uses the Batch Aggregator with the Upsert operation, the records were not being marked as failed or successful because the result was not being returned as BulkOperationResult.

Changed Minimum Mule Version to 4.1.1.

Added the isForceResubscribe method in Salesforce Bayeux Client.

Support for Anypoint Studio Proxy configuration as the default.

Trigger operation did not retrieve MetadataKeysId. Solved by adding support for retrieving metadata in case of sources.

Reply topic was losing messages in case the connection has been lost for less than 40 seconds; this is the amount of time that was needed to trigger re-subscribe naturally according to Salesforce documentation. The issue was solved by forcing re-subscription whenever the network connection is lost.

Buffering capacity exceeded message was triggered when the connector consumed more than 10000 messages from a topic. The issue was solved by adding a new attribute for the "Replay streaming channel" operation: Buffer Size In Bytes, it’s an optional parameter with the default value of 26 MB.

Wrong transformation of fields of type Date. Trying to update or insert an Object in Salesforce with fields of type Date caused a failure, this was due to failure in serializing the objects of type Date. The issue was solved by parsing the values received from the Mule flow and converting them into the expected Salesforce accepted format.

Support for new and changed objects in metadata - API v41.

Support for defining any absolute path for the keystore.

Salesforce trigger - To address ad-hoc integrators or citizen integrators with Design Center, you can now start a flow with the Salesforce connector when you create an object. For example, when new data is added to a lead object or an opportunity object, you now have an easier way to initiate a flow.