2.2.8.5 release notes

 2.2.8.5

  • Issue 2437 - Fixed an issue with variant code completions offered for standalone new expressions.
  • Issue 2440 - Added static final modifiers to Canvas.Test constants. Note that the OST must be regenerated to see the effect of this change.
  • Implemented several improvements and fixes for the Einstein GPT for Developers integration:
    • Added a settings button to the top-right corner of the main Einstein GPT dialog that can be used to configure the following (with more likely to come):
      • Feedback | Include class context information - When checked, the immediate call to the feedback API includes contextual information about the top-level and member declarations of the Apex source file which was active when generation was requested. This is enabled by default but can be disabled if you'd prefer not to share this information with Salesforce.
      • Feedback | Include file information - When set to any value other than None, explicit feedback will include information about the Apex source file which was active when generation was requested. This can be the absolute path of that file, its project- or source root-relative path, or just its filename. This uses the relative path by default but can be disabled by choosing None if you'd prefer not to share this information with Salesforce.
    • Added code completion to the selector for a parent Apex class for generated member declarations.
    • Added validation of system types as potential targets for insertion of generated member declarations.
    • Improved rendering of historical responses such that they include proper Apex syntax highlighting both in the history table and in the tooltip popup.
    • Fixed an issue with event sequencing when iterating on a request that results in creation of a new top-level type or trigger with deploy-on-save and delete propagation enabled.
    • Limited the Generate | Using Einstein GPT action to source files in IC2 modules.
  • Added another recognized FIELD_INTEGRITY_EXCEPTION pattern for falsely-reported errors of valid usages of custom objects/fields in graphQL queries during Tooling API-based LWC deployment. Such failures are automatically retried via the Metadata API.