These are the 38 updates to saas in version 2025.2, sorted with newest version first:
This supersedes 2025.2-2025-08-13-10-46-12 and earlier versions.
Fixed problem occurring when an editable ID field on a Data Container popup on a Web UI Multi Edit Data Container component was poorly styled and did not support tabbing, which created an unusable ID field.
Support is added for the below context help placement for Read Only fields on Web UI Node Editor.
This supersedes 2025.2-2025-08-12-11-48-33 and earlier versions.
If there are two basket collections for a current user with uppercase and lowercase user ID, Web UI will now prioritize using the lowercase one if both are empty. Otherwise, the system will use the one that is not empty.
This supersedes 2025.2-2025-08-07-12-35-39 and earlier versions.
When an OIEP is triggered, it checks if the AuthFunction exists. This check returned an unhandled NullPointerException. This has been fixed.
General security improvements have been implemented to enhance the overall security of our systems.
Fixed an unexpected exception that occurred in the STEP Workbench when editing unique key values.
This supersedes 2025.2-2025-08-07-12-14-42 and earlier versions.
Fixed a problem that, in very rare cases (race condition), could result in missing BLOB values (values longer than 4000 characters). The garbage collector could mistake a BLOB value for garbage (and thus remove it) if it was written at the very same time as the garbage collector ran for that specific BLOB value.
Cassandra-based STEP systems could, in rare cases, become unresponsive due to the internal Cassandra Garbage Collector holding a snapshot open for a long time. This fix improves Cassandra tokenscanner to enable Cassandra GC to release the In-Memory snapshot more frequently to avoid using many resources in the STEP system.
General security improvements have been implemented to enhance the overall security of our systems.
For more information on this fix, see issue-796457 and issue-796462.
This supersedes 2025.2-2025-07-31-21-01-02 and earlier versions.
Updated the documentation URL on the About Page to match documentation URL on the STEP main page.
When exporting data in BMEcat format, if a Price mapping was configured to use a transformation, the export could incorrectly include ARTICLE_PRICE elements with no corresponding PRICE_AMOUNT element. The PRICE_AMOUNT element is a mandatory element for ARTICLE_PRICE elements, so the ARTICLE_PRICE should only be included if there is a PRICE_AMOUNT. For this solution, the following configuration property needs to be added: FixSpecific.ISSUE794374=true
Fixed a problem in which the Log tab with attribute log lines with line breaks was unavailable in the STEP Workbench due to an unexpected exception.
We bUI: When generating a PDF proof from a Proof View screen that is configured to 'Generate proof in background', if the generation of the proof failed, the background process was incorrectly marked as 'Succeeded'. This has now been changed so that if the proof generation fails, the background process also fails.
STEP Workbench: When using the Advanced Search panel and having a search resulting in a large result list, showing the result list could take unnecessarily long time in the workbench client. This is only a problem for the systems where the limit on the number of shown results has been increased to more than the default of 100. This has now been fixed.
General security improvements have been implemented to enhance the overall security of our systems.
To fix an issue that occurred when creating assets without content using the Asset Importer, optional skipping of System Properties for empty assets has been added.
Fixed an unexpected ConcurrentModification exception when purging assets from the Recycle Bin.
Created a customer-specific solution to correct an issue with a finished process that would not auto-delete from the Web UI Warning Notification Panel.
For more information on this fix, see issue-796457 and issue-796462.
This supersedes 2025.2-2025-07-29-17-51-18 and earlier versions.
Fixed a problem, when alerts, shown by using the Run Business Action on Web UI Node Editor, were cleaned up when evaluating validation business conditions on the same screen.
When using the mapper to import either assets, classifications, or entities, the import functionality will now allow removal of entity to product, asset to product, and classification to product cross-references, if the reference type has been configured to remove un-mapped reference types in the import configuration.
We bUI: When generating a PDF proof from a Proof View screen that is configured to 'Generate proof in background', if the generation of the proof failed, the background process was incorrectly marked as 'Succeeded'. This has now been changed so that if the proof generation fails, the background process also fails.
Added a null check to avoid an error if a configuration in the Web UI Node List has an attribute configured that has been deleted after the configuration has been done.
Fixed an issue that occurred when users received an unexpected error stating the external XSLT service had not processed the task yet and had ended the attempt. To address the issue, a private config property was introduced that increases the waiting time of the external XSLT microservice, which is set by default to 2500 milliseconds instead of hard-coded at the 2500 limit. This property cannot be larger than 10000 (10 seconds).
To address an error that occurred when attempting to bulk-assign tasks to users in the Web UI, changes have been made to ensure the correct node is used for calculating valid assignees for a task in the Web UI Node List.
It has been seen that the STEP background process (BGP) system stops scheduling new BGPs due to internal duplicates of values used for controlling the state of the BGP (running, waiting, succeeded,etc.). This fix makes the BGP system able to survive cases where LOVs used by BGPs have duplicates. For instance, the status attribute containing the state running, waiting, succeeded, etc. Also ensured specifically for the status attribute that creating duplicates is not possible anymore. This also includes removing any duplicates already created.
Introduced a customer-specific solution which will clean up the business condition alerts on Node Details and Initiate Item screens in the Web UI when the business condition does not return messages anymore (the errors have been fixed).
In the Web UI Advanced Search component, it was not possible to search for values containing multiple consecutive spaces. This has been fixed.
A fix was made so it is possible to delete BGPs of type 'Refreshing Node Collection.'
A problem occurred when values could not be correctly identified as alphanumeric. In these instances, the system would fall back on simply comparing them as numeric, which caused all values to tie, which caused the system to then fall back on sorting by the ID column instead. This fix includes a check in the default case to see if the values are numeric. If not, a string comparison is performed instead.
Saving a page in Web UI with multiple changes could, in some cases, fail with an exception. This has been fixed.
Fixed a problem that occurred with importing new publications when importing from Excel.
Progress bar is incremented in InDesign Queue when input is more than 90 nodes. This has been fixed.
Previously, in an edge case where a user group had been deleted after being assigned as a default assignee in a workflow, users would encounter a NullPointerException if the workflow remained unchanged and valid object types were automatically initiated. This issue made it difficult to diagnose the problem as users would only discover the cause by attempting to edit the workflow. This has been improved: now, the null pointer is handled gracefully, and a clear, immediate error message is presented to the user, eliminating the need for additional investigation.
This supersedes 2025.2-2025-07-24-13-41-07 and earlier versions.
Fixed a NullPointerException occurring in the AutoCare PCdb import.
There was a problem searching in the Web UI for values, including multiple consecutive spaces. A previous fix did not address the problem sufficiently, thus this new fix reverts the old fix and adds the better and more stable fix.
If a system was configured with the configuration property Calculated.TemplateRoot that was set to a directory that did not exist, an error would occur when editing a Calculated Attribute function. This has been fixed.
In the Web UI Advanced Search component, it was not possible to search for values containing multiple consecutive spaces. This has been fixed.
Asset Importer will now ignore the contents of the macOS resource folder when importing assets from a zip file.
Added detailed debug logging to List Value Generator.
Fixed a problem in which the value generator was inaccessible with invalid values in the STEP Workbench.
The CSV import for List Value Generators did not handle BOM characters and as a result unwanted BOM characters got imported within the values. The fix handles BOM characters correctly so they do not end up imported within the values.
The order was not honored. A proper map structure (linked hash map) to honor the order from configuration has been added to fix this issue.
A fix was made so it is possible to delete BGPs of type 'Refreshing Node Collection.'
Now, an ID label in workbench will not be translated in Japanese and just say 'ID.'
This supersedes 2025.2-2025-07-23-10-59-44 and earlier versions.
Fixed problem with business function results caching, which lead to missing export values that were related to business function mapping.
Now, Web UI configuration will reload Matching Algorithms cache every 30 minutes in addition to doing a Web UI configuration last edit date check to make sure that latest changes in the configuration are reflected.
Fixed the REFERENCE_FEATURE_SYSTEM information to be in line with BMECat 1.2 specifications.
Addresses a specific edge case encountered during Asynchronous Translation on a customer system, where two translation files were generated—one of which was empty. This resulted in a premature end of file error when importing the response from Lionbridge. The fix introduces an additional sanitization step that filters out empty files from the list prior to processing. This covers files that are 'Truly empty' meaning that they have a size of 0Kb, 'Effectively empty' files still pass as they do have a root defined within.
There was a problem searching in the Web UI for values, including multiple consecutive spaces. A previous fix did not address the problem sufficiently, thus this new fix reverts the old fix and adds the better and more stable fix.
A fix was made for for exporting asset references on the child nodes in the streaming mapper.
When saving a ZIP file during an SFTP delivery, the correct extension (".zip") is now used in the file name rather than the exported file's file extension type.
In the Web UI Advanced Search component, it was not possible to search for values containing multiple consecutive spaces. This has been fixed.
Fixed the export delivery plugin to rename the final file correctly when using legacy storage folder options. Also, we improved text in the exception when this fails.
Fixed a problem occurring where it was not possible to import an empty embedded number attribute value in STEP Workbench.
Fixed the issue where child objects were exported instead of the selected ones when using eCatalog.
A property that dictates whether or not an event should be generated before or after the object has been updated for a customer-specific use case.
This supersedes 2025.2-2025-07-17-09-17-24 and earlier versions.
Channel IDs from PDX that are longer than 100 characters are now hashed using UUID.nameUUIDFromBytes(...). The resulting 36-character ID is stored in the PDX Channels LOV. Example: id=f2c8c1a0-6c3e-3b3e-8a7e-4f5c8e4f6f3e, value={"name":"Skin Care", "parentChannelID":"Arts_PROD_Sample_Channel-KBduCjzT"} The original full Channel ID is preserved in the PDS Status Data Container.
This supersedes 2025.2-2025-07-16-07-25-54 and earlier versions.
Now, Web UI configuration will reload Matching Algorithms cache every 30 minutes in addition to doing a Web UI configuration last edit date check to make sure that latest changes in the configuration are reflected.
The problem with sorting was related to a pagination problem on some node fields, for example, Approval Status. They do not have lower case representation indexed in Elasticsearch, which led to a pagination problems as the system tried to use it. A fix has been prepared for Approval Status and other static fields that have enumerated values. The search, after (pagination) mechanism, takes lower case values into account. Otherwise, it uses the original value.
Now multi-value attribute values with units will not provoke unexpected exceptions in the Globally Configured Unfolding Data Container component in the Web UI.
Increased time for which the GC Overhead Sensor can be executed before causing the warning about slow execution from 3 to 20 seconds - due to the slow GC logs reading.
Now, the Hide Equal functionality works correctly on the Multi Workspace Screen when applied to attribute values that do not exist in the approved workspace.
To fix an environment outage, the amount of undo / redo operations on a Mass Creation Screen in the Web UI will be reduced to 100 in order to save server memory.
This supersedes 2025.2-2025-07-15-11-38-58 and earlier versions.
When trying to export Excel Smartsheet, it will no longer fail and throw a NullPointerException if the exported product has not been initiated in a workflow.
Added detailed logging at level FINEST when node is not found in workspace during export.
Backported latest improvements to global representation caching to 11.1-MP6.
Fixed a NullPointerException that occurred during the Change Package impact analysis when handling the FunctionDependency tag of a local business rule of a workflow.
To fix an error thrown when attempting a bulk update, a special attribute holding the search URL is allowed to exceed 1M limit via a property. The limit is in place to protect against running out of memory, but given that relatively few values will be held on the heap for this attribute, allowing it presents no problem.
Fixed a NullPointerException occurring in the Globally Configured Multi Edit Data Container in the Web UI's design mode.
Now, an assignee's type (group or user) will be correctly determined by the Submit from Grid action when being opened in a Task List with status flag filtering.
Fixed a ClassCastException error thrown by the outbound integration endpoint (OIEP) Poller with GDSN Data Pool configured.
Exceptions are prevented when doing drag and drop in workbench from a product to a plan.
The fix ensures that spaces in the name of a downloaded file are preserved and not substituted with plus ('+') characters.
The Import InDesign Package action now uses the shared directory for the InDesign documents to avoid InDesign failing to open / download the document if more servers in a cluster are sharing the task.
When using a “STEPGROUP AG” tag in an InDesign product template for which the product does not have values for any of its attributes in the specified attribute group, the text from the start STEPGROUP tag to the end STEPGROUP tag is deleted.
Fixed a NullPointer Exception (NPE) occurring when retrieving an asset from the recycle bin.
This supersedes 2025.2-2025-07-11-11-27-48 and earlier versions.
This solution fixes the preview of images from Image Viewer in workbench from being cropped on the right and bottom. Instead of referencing the entirety of the Image Viewer, including the Title Border, it now references the draw area exclusively. Images are therefore presented and scaled and centered according to the draw area only.
When using the Web UI component Matching On Save with the Run Business Action button that calls the Address Standardization Business Action, evaluation of match codes were calculated on saved data before calling Business Action. Now, it is fixed and calculation of match codes is also called after the Business Action. To enable this functionality, which is not standard, contact Stibo Systems Support.
Fixed problem so that changes in the JavaScript 'Applies If' condition are now saved when being edited via the 'Pen' icon in STEP Workbench.
Exceptions are prevented when doing drag and drop in workbench from a product to a plan.
Fixed an issue in outbound integration endpoints (OIEPs) where product selection changes were not properly reflected in the endpoint's XML configuration. This fix ensures all product selections made in the UI are correctly persisted in the endpoint's configuration. In order to enable the fix, the following configuration property needs to be added and set to 'true': FixSpecific.ISSUE704765=true.
To fix the issue, it is now possible to select assets in the Test Business Rule dialog.
This supersedes 2025.2-2025-07-09-10-20-05 and earlier versions.
Now, Web UI configuration will reload Matching Algorithms cache every 30 minutes in addition to doing a Web UI configuration last edit date check to make sure that latest changes in the configuration are reflected.
Fixed an issue in Match and Merge Importer's new implementation where groups of records for matching and merging were duplicated, and because of that, the amount of records presented in the execution report's message was higher than amount of records provided in the import file.
This supersedes 2025.2-2025-07-08-14-33-45 and earlier versions.
When a Photoshop EPS file with a large physical dimension specified in the bounding box was converted, the system crashed. This is now solved.
Now, a healthcheck 'History entries pointing to non existing previous versions' will not hold the epoch lock, preventing LFSC processes from starting.
Fixed an unexpected exception in CharactersBetweenTransformation which was caused by an end character being placed before a start character.
Added logging to expose possible error. Further investigation is ongoing.
Fixed a problem that occurred in the mapDB implementation when mapDB files created by the exporter could be left in temporary Linux storage even after the process completed.
This solution fixes the problem with non-exported MIME elements in BMEcat when multi-level references are used in the mapping to get the MIME_SOURCE value. Introduced a temporary config property: FixSpecific.ISSUE778345 - when set to false it uses the values only from the first level references instead of the last level configured in mappings.
Added detailed logging at level FINEST when node is not found in workspace during export.
Now, folders will be cleaned up correctly after AsyncJobRunner finishes and is scheduled for cleanup.
Added configurability for buffer size and internal timeout for the SFTP client.
This supersedes 2025.2-2025-07-06-04-21-09 and earlier versions.
An error occurred because a translation partner expected specific translation file-naming conventions in the translation file they received from the customer. The naming discrepancies caused translation files to be rejected in a specific use case. To fix the issue, the translation file name now defaults to the expected naming convention (translation.xml) should it not find translation files following the translation-BGP_(\\d+)-(.*).xml file-naming pattern.
Now, folders will be cleaned up correctly after AsyncJobRunner finishes and is scheduled for cleanup.
This supersedes 2025.2-2025-07-03-17-53-22 and earlier versions.
Now, an LOV ID sorting order will be correctly respected in an Add Data Container popup on a Web UI Multi Edit Data Container component.
Fixed a problem that occurred when multi-valued Lists of Values (LOVs) with IDs were not saved correctly by the Globally Configured Unfolding Data Container component.
To fix an error that prevented products from being saved, a property was introduced that allows multi-value duplicates in the Web UI Node Editor. The property defaults to false but when toggled will introduce the desired behavior.
An existing HOTFIX-8251 was backported to 2024.3. Fixed an unexpected exception related to navigating to an item after submitting it from the current state.
Reduced memory usage by not showing data containers and headers for Globally Configured Multi Edit Data Containers in Web UI Design Mode.
This supersedes 2025.2-2025-07-01-14-49-59 and earlier versions.
Nodes that were directly defined in the privilege definition were not displayed in the Elasticsearch search screen. This is now fixed and all nodes to which the user has been granted privileges are visible.
After the fix, approval status in Elasticsearch presents proper values in each context.
This supersedes 2025.2-2025-06-30-13-52-07 and earlier versions.
Fixed a bug introduced in pipeline calculate-resolution introduced in a recent commit for a customer fix.
Dual visible soft values in the health check (Softvalues for Dual Visibility) also works when the values are stored under edgeid instead of nodeid.
Assets used on multiple pages could not be purged from the Recycle Bin. This has been fixed.
Now the import preview will be able to read more than 1000 characters from the mapped column cell, which will fix the incorrect exceptions about tag validation.
Fixed a problem in which attributes without values or names were being exported in BMEcat 1.2 exports. To address the issue, a new configuration property has been introduced that, when enabled, does not add article feature elements when feature values or names are empty.
After upgrading to STEP 2025.2, starting BGPs could produce some severe log entries in step.0.log (especially when starting Event Processor BGPs) that are not really severe. These severe log entries have now been removed.
When exporting in generic JSON format, nested references are now correctly handled.
The HAProxy based loadbalancer is used for on-prem STEP systems, including development systems, so it needs to support the multitenant Instrument server. SaaS systems implement this support via the Kubernetes Ingress, which sits in front of the STEP app server and is not related to this change.
The loadbalancer frontend needs to be able to understand the shape of the STEP cluster, but when new server types are created, the frontend does not understand those. This is a change to simply ignore new server types and rely on the already known types.
To fix an issue with a failing OIEP (outbound integration endpoint), a Rhino-xml third-party library was added to re-enable E4X (ECMAScript for XML) XML API in JavaScript.
Added third-party library to Asset Push Sidecar to fix the problem that Asset Push stopped pushing assets.
This supersedes 2025.2-2025-06-27-14-58-45 and earlier versions.
When running the Asset Importer in an IIEP, load the Asset Importer Configration according to the context configured on the endpoint, rather than default context. This makes it so that mapped attributes will get mapped against LOV values in the correct context.
Returning LOV and multivalued attributes in REST API v2 now.
To fix an error that prevented products from being saved, a property was introduced that allows multi-value duplicates in the Web UI Node Editor. The property defaults to false but when toggled will introduce the desired behavior.
Introduced a configuration property that is set to false by default. When the property is set to true, the data profiling MapDB files will be cleaned up using the system GC on both Windows and Linux systems. Previously this approach was used only on Windows.
Fixed a problem that occurred when records with a '0' match score were not merged even if they had a confirmed duplicate relation.
Under certain system configurations, the user-applied sorting of products in a Print on Demand publication was incorrectly ignored in the resulting pagination. This has been fixed.
Fixed a memory leak in the Business Rule Based Message Processor. Previously, the processor would find all the nodes to process and then process all the nodes. Now, the processor will find and handle nodes one by one. The benefit of this change is that there is no memory leak from loading all the nodes (as well as any related nodes and their values) into heap memory. Previously, when a batch contained a collection with many nodes, it would cause the application server to crash with out of memory.
Improved the error message for when you make multiple concurrent uploads of the same file to the same integration endpoint.
{code:title=before}
Operation failed as an exclusive lock could not be obtained for endpoint ID '
To fix an environment outage, the amount of undo / redo operations on a Mass Creation Screen in the Web UI will be reduced to 100 in order to save server memory.
This supersedes 2025.2-2025-06-27-04-20-51 and earlier versions.
Fixed a memory leak in the Business Rule Based Message Processor. Previously, the processor would find all the nodes to process and then process all the nodes. Now, the processor will find and handle nodes one by one. The benefit of this change is that there is no memory leak from loading all the nodes (as well as any related nodes and their values) into heap memory. Previously, when a batch contained a collection with many nodes, it would cause the application server to crash with out of memory.
Added third-party library to Asset Push Sidecar to fix the problem that Asset Push stopped pushing assets.
This supersedes 2025.2-2025-06-25-16-42-19 and earlier versions.
Fix for solving side-effect in new version of com.stibo.util
This supersedes 2025.2-2025-06-20-15-51-44 and earlier versions.
Fixed renaming of files using the $autodetected-extension$ in a template.
Added third-party library to Asset Push Sidecar to fix the problem that Asset Push stopped pushing assets.
This supersedes 2025.2-2025-06-19-18-52-03 and earlier versions.
Added gamma correction for 32-bit TIFF images to make these images look similar to other image viewers when viewed from the workbench image viewer.
Some File-Not-Found errors could occur during Hotfolder handling and copying of files when under higher load, thus stopping the endpoint. This file handling has been made more robust, which should improve this problem. And additional logging has been added in order to better determine where the problem is, if it still occurs.
Fixed renaming of files using the $autodetected-extension$ in a template.
Fixed an unexpected exception on a Task List with Streaming, Sorting, and Filtering enabled.
To fix an issue that occurred where asset content was missing from asset objects, logging has been added in order to follow the progress of the health check for missing BLOBs.
Fixed a problem that, in very rare cases (race condition), could result in missing BLOB values (values longer than 4000 characters). The problem lied in that the Garbage Collector mistakenly could consider a BLOB value as garbage (and thus remove it) if it was written at the very same time as the garbage collector ran for that specific BLOB value.
Added detailed logging at level FINEST when node is not found in workspace during export.
Customers using a multi-server cluster without In-Memory use the Level 2 cache. And, the communication between those caches was accidentally restricted, which could cause the communication to fail. This causes optimistic lock verification errors that seem to be at random places. This has been fixed so that STEP does not impose such restriction itself, but leaves imposing the correct network restrictions to the network setup.
Fixed an issue where both imports and change package installs were unjustifiably failing when they included a link type which added an attribute group, in cases where the link type had data containers as a valid source.
Fixed problem, when existing files were not overwritten completely by new SFTP delivery plugin - instead the data was written inside those files. Now the existing files will be completely overwritten by SFTP delivery plugins.
We avoid the issue by limiting the number of cell indices for deep linking saved in the url
Updated dependency versions for libraries used by XSLT local transformation.
Now, an OK button will be disabled on the Add Data Container Popup, which contains only LOV-mandatory attributes, if the attribute values are not provided by user.
Fixed a memory leak caused by accumulating unused product URLs during the import.
Fixed problem, when depending LOV with similar ID on one LOV value with Value from another LOV value was causing invalid validation error in WebUI Node Editor screen when performing LOV cross validation business condition.
Approving object that is already fully approved now does nothing. Previously it could do an update in the revision of the object which could cause optimistic lock exception if doing this currently.
We fixed the way character tags are parsed in the case of multi-valued Lists of Values (LOVs) where values also contain other tags.
HOTFIX-8332: The $autodetected-extension$ filename macro now also works for Office documents. HOTFIX-8333: Custom image conversions can now be used to generate thumbnails of Office documents.
Fixed an Import Manager mapping problem when using an Intermediate Variable with a multi-value Constant to map into a Data Container where the value was not separated correctly from the multi-value Constant.
Fixed an issue that caused the average processing time of Approve Recycle Bin jobs to be more than 40 minutes if even one unapproved object was in the Recycle Bin. To fix the issue, a new health check was added and the node in recycle-bin missing a front revision was revised.
Using Attribute Group Attributes data selector in a table, if "Exclude if empty" was selected and a child product was overriding the sort value with an empty value, the empty value on the child was incorrectly being ignored and the parent value would be used instead. This resulted in values appearing in the table when they should have been excluded.
The loadbalancer frontend needs to be able to understand the shape of the STEP cluster, but when new server types are created, the frontend does not understand those. This is a change to simply ignore new server types and rely on the already known types.
This supersedes 2025.2-2025-06-19-15-23-50 and earlier versions.
The problem with sorting was related to a pagination problem on some node fields, for example, Approval Status. They do not have lower case representation indexed in Elasticsearch, which led to a pagination problems as the system tried to use it. A fix has been prepared for Approval Status and other static fields that have enumerated values. The search, after (pagination) mechanism, takes lower case values into account. Otherwise, it uses the original value.
This supersedes 2025.2-2025-06-18-14-30-35 and earlier versions.
An error occurred because a translation partner expected specific translation file-naming conventions in the translation file they received from the customer. The naming discrepancies caused translation files to be rejected in a specific use case. To fix the issue, the translation file name now defaults to the expected naming convention (translation.xml) should it not find translation files following the translation-BGP_(\\d+)-(.*).xml file-naming pattern.
Addresses a specific edge case encountered during Asynchronous Translation on a customer system, where two translation files were generated—one of which was empty. This resulted in a premature end of file error when importing the response from Lionbridge. The fix introduces an additional sanitization step that filters out empty files from the list prior to processing. This covers files that are 'Truly empty' meaning that they have a size of 0Kb, 'Effectively empty' files still pass as they do have a root defined within.
Under certain system configurations, creation of InDesign packages could result in missing files. The Create InDesign Package background process can now be configured to always use a shared directory.
General security improvements have been implemented to enhance the overall security of our systems.
To fix an issue that occurred when creating assets without content using the Asset Importer, optional skipping of System Properties for empty assets has been added.
Improved handling of an edge case when customer-defined behavior conflicts with recommended practices: https://doc.stibosystems.com/doc/version/latest/web/content/busrules/conditions/business_condition_evaluate_javascript.html 'Although it is technically possible to write business conditions to change data, doing so can result in unpredictable errors and should therefore not be attempted.' This specific issue was observed in the Datasheet OIEP when attempting to revive a deleted asset. In this case, a customer-defined business rule triggered during the revival process and attempted to modify the asset, resulting in an OptimisticLockException. This update improves resilience by gracefully catching the exception, allowing the process to continue without termination.
Decreased 'Severe' severity for log messages related to extension API errors to 'Finest.'
Fixed the issue, which was not updated during ECLASS import.
This supersedes 2025.2-2025-06-16-13-48-57 and earlier versions.
Fixed the incorrect status when images were not updated to upstream systems via Asset Push notify / sidecars.
An error occurred because a translation partner expected specific translation file-naming conventions in the translation file they received from the customer. The naming discrepancies caused translation files to be rejected in a specific use case. To fix the issue, the translation file name now defaults to the expected naming convention (translation.xml) should it not find translation files following the translation-BGP_(\\d+)-(.*).xml file-naming pattern.
Addresses a specific edge case encountered during Asynchronous Translation on a customer system, where two translation files were generated—one of which was empty. This resulted in a premature end of file error when importing the response from Lionbridge. The fix introduces an additional sanitization step that filters out empty files from the list prior to processing. This covers files that are 'Truly empty' meaning that they have a size of 0Kb, 'Effectively empty' files still pass as they do have a root defined within.
The problem with sorting was related to a pagination problem on some node fields, for example, Approval Status. They do not have lower case representation indexed in Elasticsearch, which led to a pagination problems as the system tried to use it. A fix has been prepared for Approval Status and other static fields that have enumerated values. The search, after (pagination) mechanism, takes lower case values into account. Otherwise, it uses the original value.
Now, an LOV ID sorting order will be correctly respected in an Add Data Container popup on a Web UI Multi Edit Data Container component.
Fixed a faulty sort comparator in Tree Navigator.
Fixed an unexpected validation exception on a Web UI Node Editor screen caused by evaluation of a Business Condition for LOV attributes, where one value had the same ID as another LOV value's value.
Now, an OK button will be disabled on the Add Data Container Popup, which contains only LOV-mandatory attributes, if the attribute values are not provided by user.
Fixed unexpected error occurring when double clicking on a Workflow Pie Chart on a Web UI home page.
General security improvements have been implemented to enhance the overall security of our systems.
Now, the Web UI compressed component store size will be limited to avoid OutOfMemory problems.
Fixed a problem that occurred when the Web UI Node List did not clean the selected item after it was deleted or removed from the list as a result of a Run Business Action execution.
This supersedes 2025.2-2025-06-12-17-52-29 and earlier versions.
Improved performance of list value generator.
Fixed a problem in Auto Classification Rules where the SkipIfRulesDoNotAllow property was not working as intended.
General security improvements have been implemented to enhance the overall security of our systems.
This supersedes 2025.2-2025-06-10-11-54-12 and earlier versions.
When Match and Merge Importer uses anew implementation and there is a Standardized Business Action configured that adds or removes a reference type on an incoming node, such changes were not reflected in the result record. Now, this is fixed and changes made through a Standardized Business Action are visible on the result record.
Fix for upgrade library in component Blob storage to final version.
Fix for upgrade library in Blob storage component to latest version.
Fixed an issue in the Match and Merge Importer V2 implementation where importing the same entity twice resulted in default source information being added after the second import attempt. Now, the importer correctly preserves or updates source information without adding defaults on repeated imports.
General security improvements have been implemented to enhance the overall security of our systems.
Now, an HTML function Web UI component will not be displayed in a design mode in order to allow users to correct it if it causes an infinite reload.
Now, an Object Type Search descriptor in the Web UI will automatically transform an 'AND' object type search into an 'OR' object type search (as one node cannot have multiple object types).
This supersedes 2025.2-2025-06-04-12-01-23 and earlier versions.
When a user first logs into the workbench after a STEP system has been upgraded, a dialog is presented that offers to open the update notes for the new release (2025.2). This update notes link was broken and is now fixed.
Improved performance of list value generator.
Now, on a Web UI Multi Context Edit screen, when updating a cell, another cell will not be updated when clicking on another cell to exit the edit mode.
Information published in 2025.2 documentation.
Fixed a problem where empty inherited references were not shown in GraphQLv2 response.
Fix for upgrade library in InMemory component to latest version.
Fixed unexpected error occurring when double clicking on a Workflow Pie Chart on a Web UI home page.
Improved the Log tab for Event Processors in workbench: Now, log messages from Event Processors and the event queue are sorted and deduplicated. Log messages for all nodes will be sorted in workbench.
Backport of ISSUE-93079
General security improvements have been implemented to enhance the overall security of our systems.
General security improvements have been implemented to enhance the overall security of our systems.
This supersedes 2025.2-2025-06-02-18-51-44 and earlier versions.
General security improvements have been implemented to enhance the overall security of our systems.
Added a config-dependent solution to make a guardrail protecting against large values more flexible.
Fixed mapping problem occurring when mapping a variable with a constant data source, and then choosing a Data Container mapping for this variable, in the import wizard in STEP Workbench.
Added log information to integration endpoints to enable troubleshooting of errors
Fixed an IllegalStateException error that occurred when using Outbound Integration Endpoints (OIEPs) with the aspect 'File named by asset ID' or 'File named by asset Name' in the output template mapping.
Empty assets were not migrated from Oracle to Cassandra. A workaround has been made to solve this problem.
Fixed an issue that occurred when users received an unexpected error stating the external XSLT service had not processed the task yet and had ended the attempt. To address the issue, a private config property was introduced that increases the waiting time of the external XSLT microservice, which is set by default to 2500 milliseconds instead of hard-coded at the 2500 limit. This property cannot be larger than 10000 (10 seconds).
Oracle Data Migration: Fixed the problem that empty asset content was not migrated.
Fixed a problem where PDF metadata could become mixed up or corrupted when importing multiple files at the same time. This fix ensures that each file's information remains intact even when the system processes several PDFs simultaneously.
When a Photoshop EPS file with a large physical dimension specified in the bounding box was converted, the system crashed. This is now solved.
When a user first logs into the workbench after a STEP system has been upgraded, a dialog is presented that offers to open the update notes for the new release (2025.2). This update notes link was broken and is now fixed.
Improved logging for low-memory detector when using OneQueue BGP scheduler was added after an investigation of background processes not starting due to heap usage.
Fixed a problem when exporting the product parent with an empty URL selection when exporting leaf products ended up exporting only the parent and excluding the product children.
Fixed a problem with missing 8BIM profiles when "Keep Profiles" was set in Image Conversion Configuration.
Excessive caching could cause STEP to crash due to running out of memory. It is now possible to set a limit on the number of objects that are cached.
The OrphanDataContainers health-check had an error in its fix code that could cause a severe data inconsistency in STEP if the fix mechanism of that health-check is used. This has been fixed.
General security improvements have been implemented to enhance the overall security of our systems.
Improvements to new feature in 2025.2 with having BGP files in Blob Storage.
Fixed an issue that caused the average processing time of Approve Recycle Bin jobs to be more than 40 minutes if even one unapproved object was in the Recycle Bin. To fix the issue, a new health check was added and the node in recycle-bin missing a front revision was revised.
Suppressed unnecessary logging from OIEP Business Processor, which could have caused the out-of-memory (OOM) problem.
Now, the data container key will use a default unit if it was not specified by the "withAttributeValue" JavaScript Business Rule API, which will make sure that a new data container is not created if there is already a data container object with the same attribute value and default unit.
The issue solves a problem of empty node id and url while reindexing data.
This supersedes 2025.2-2025-05-28-19-58-16 and earlier versions.
To improve Asset Push functionality, Asset Push now runs on Java 21.
We have implemented a fix that ensures
Fixed unhandled exception when adding a new Auto Classification Rule to the table with applied filtering in STEP Workbench.
Improvements to new feature in 2025.2 with having BGP files in Blob Storage.
General security improvements have been implemented to enhance the overall security of our systems.
The issue solves a problem of empty node id and url while reindexing data.
This supersedes 2025.2-2025-05-26-06-27-16 and earlier versions.
Implemented a change to the Inbound Business Rule Message processor ensuring that business rules marked as "run as privileged" will run as such.
General security improvements have been implemented to enhance the overall security of our systems.
Fixed a problem occurring when adding a data container on a Web UI Node Editor via the Globally Configured Multi Edit Data Container component caused an 'Unsaved changes...' popup to appear when trying to submit a current item via Submit Action.
General security improvements have been implemented to enhance the overall security of our systems.
Fixed a problem when users with restricted privileges coud not access a Web UI if they did not have access to Web UI logo attributes.
Improved Web UI Pie Chart legend styling so that it now wraps to the next line instead of shifting and overlapping the chart when the legend values are too long.
Improved the Log tab for Event Processors in workbench: Now, log messages from Event Processors and the event queue are sorted and deduplicated. Log messages for all nodes will be sorted in workbench.
Fixed a problem where it was possible to save a data container without a key value on a Web UI Node Editor via the Globally Configured Data Container component.
General security improvements have been implemented to enhance the overall security of our systems.
General security improvements have been implemented to enhance the overall security of our systems.
This supersedes 2025.2-2025-05-23-14-56-37 and earlier versions.
Added extra check when matching users by key value.
This supersedes 2025.2-2025-05-22-11-39-33 and earlier versions.
Implemented a change to the Inbound Business Rule Message processor ensuring that business rules marked as "run as privileged" will run as such.