-
Fixed an issue where after an automation was deployed to an account, data within the automation was still associated with the source account instead of the target account. The most noticeable side effect was that dynamic field mappings would not appear when actions in the deployed automation were edited.
-
Fixed an issue introduced in the previous release where temporary file URL generation within batch jobs became broken due to the system initializing an automation uuid value so that trigger/action assemblies can be better tested when run on the server due to some modules needing a non-empty automation uuid.
-
The “CSV to XML” subassembly has been deprecated and a new version has been added that emits columns as XML nodes rather than attributes.
-
Fixed an isolated rendering issue with the automation editor caused by a certain arrangement of conditionals, where the conditionals would not be rendered in the expected layout.
-
Fixed the File - CSV module so it works as expected instead of returning a 500 error.
-
Fixed the “Transform Date Time” “modify and reformat date time” action so that when a timezone offset is specified, the offset is computed taking into account the input date and time value, which in turn takes into account any future or past timezone changes due to Daylight Saving. Previously the offset was being computed using the server’s current date and time. So if the input date resulted in a different timezone offset due to Daylight Saving, the resulting value was incorrect.
-
Improved the rendering of the automation editor such that when automation diagrams are loaded the “Loading…” spinner stays onscreen until the diagram is fully rendered (avoiding the screen being empty for a second or two for very large diagrams) and the diagram no longer jumps around due to collapsed groups when initially shown.
-
The Text Value module in the assembly editor now defaults the editor syntax to “Text”, rather than “XML”. If the text starts with “<” the syntax is defaulted to “XML”. If the text starts with “<jsp” the syntax is defaulted to “JSP”.
-
A change has been made to the automation processing engine in the way that it requests Java to perform garbage collection. Previously the system would ask Java to perform garbage collection only when all automations were idle, before every 100th execution of an automation. The theory was that it was best to not to request garbage collection while automations were processing, to avoid any performance penalty. Unfortunately this caused a problem b/c a single long-running automation would prevent garbage collection from occurring, which caused memory usage to grow without bounds until the long-running automation was finished. A contributing factor was that Java was not automatically executing its garbage collector, seemingly b/c of the system’s continuous processing. Now garbage collection is requested before every 50th execution of an automation, even if other automations are being executed. In addition, the automation execution engine now requests Java garbage collection after every 50th data row is processed. These changes should help to reduce peak memory usage and avoid memory exhaustion.
-
The automation processing engine now writes automation history data into Lucene upon every 50th processed trigger data row even if each data row experienced an error. Previously data was written into Lucene on every 50th trigger data row only if not error occurred. This change reduces memory consumption and improves the performance of storing automation history data for the scenario when an automation processes lots of trigger data rows and many rows encounter errors.
-
The automation editor’s Connections screen now always allows the Default account to be deleted. Previously the Default account could only be deleted when it was the last connection. If the app has a second account connected, that connection will then become the default account and any automations that reference the second connection are turned off. If the app has more than 2 accounts connected, a selection dialog will appear allowing one of the named accounts to be selected as the new Default account. If a selection is made, all automations referencing the selected account will be turned off.
-
Fixed an issue with the automation editor’s top-level automation history screen that would sometimes cause the total number of tasks to be displayed as “null”.
-
The Admin Console’s “Memory Usage” screen now has a button that allows a Java heap snapshot of the server to be saved. The button is only visible to the root admin account.
-
The Automation Editor’s field mappings now display just the first 20 characters of the trigger/action name, rather than the item’s step number (which could become very long for complex action logic). Mousing over the trigger/action name will display the item’s step number and full name in a tooltip.
-
Fixed an issue where sometimes when connecting to an OAuth app the previously connected app’s authorization dialog would erroneously appear instead of the selected app’s authorization dialog.
-
The Transform Number’s “evaluate math formula” action no longer adds zero to the computed result, which could cause a zero being appended to the final value rather than the originally intended conversion to a numeric result.