Skip to content

RELEASE_NOTES.md

Latest commit

 

History

History
38 lines (23 loc) · 3.7 KB

RELEASE_NOTES.md

File metadata and controls

38 lines (23 loc) · 3.7 KB

Payroll Engine — Release Notes

v1.0.0 - Apr 2026

Highlights

  • Wage type caches — WageTypeCycleCache and WageTypeConsolidatedCycleCache bulk-load cycle and consolidated wage-type results once per employee when wage types are tagged via the corresponding payroll cluster sets, reducing database round-trips during payrun scripts
  • PasswordAvailable — user API and client models expose whether a password is set without transferring the hash
  • Exchange import — payrun jobs that end with Abort/Cancel and CompletedJobStatus=Abort are treated as an expected test outcome during import, not as a hard import failure

Backend

Database Update required — run the Update-Model scripts (SQL Server and MySQL) from schema 0.9.7 → 1.0.0 before starting the backend. The nine Payroll columns ClusterSetCase, ClusterSetCaseField, ClusterSetCollector, ClusterSetCollectorRetro, ClusterSetWageType, ClusterSetWageTypeRetro, ClusterSetCaseValue, ClusterSetWageTypePeriod, and ClusterSetWageTypeLookup are replaced by a single ClusterSet JSON column; existing data is migrated in the scripts. The persistence layer enforces minimum schema version 1.0.0 — the backend will not start against a pre-1.0.0 database.

New Features

  • WageTypeCycleCache — for wage types tagged via Payroll.ClusterSet.ClusterSetWageTypeCycle, GetWageTypeResults data for the cycle/YTD path is loaded in bulk once per employee at payrun employee start; the cache is reset and reloaded before retro re-evaluation
  • WageTypeConsolidatedCycleCache — for wage types tagged for the consolidated cache, GetConsolidatedWageTypeResults is bulk-loaded once per employee; calls with noRetro=true bypass the cache by design
  • PasswordAvailable in the user API response — derived from the stored hash in memory; the password hash remains [JsonIgnore] on the model

Bug Fixes

  • ClusterSet persistence — pre-serialize ClusterSet as PascalCase JSON before write to avoid sql_variant conversion issues; register JsonObjectTypeHandler<PayrollClusterSets> for correct Dapper deserialization on read
  • Payrun jobs — CompletedJobStatus=Abort is no longer applied to successfully completed jobs; abort remains reserved for processor failures and expected-abort test scenarios
  • PayrunProcessorSettings — XML documentation for MaxParallelPersist default corrected (validated by load tests)

Breaking Change

  • Database schema 1.0.0 — Payroll cluster-set columns consolidated into one ClusterSet JSON column; breaking change (see Database Update above)

Libraries

  • Client.Core — IPayroll / Payroll: individual ClusterSet* string properties replaced by a single ClusterSet property of type PayrollClusterSets breaking change; PasswordAvailable on IUser / User; ExchangeImport honors intentional abort status for test expectations; project references use $(Version) for PE dependencies; release workflow no longer patches versions with sed
  • Client.Scripting — fixed double-encoding of strings in PayrunFunction; GitHub Actions pages.yml patches PE dependency versions before build

Breaking Change

  • JSON schema (PayrollEngine.Exchange.schema.json) — the former top-level Payroll properties clusterSetCase, clusterSetCaseField, clusterSetCollector, clusterSetCollectorRetro, clusterSetWageType, clusterSetWageTypeRetro, clusterSetCaseValue, and clusterSetWageTypePeriod are removed; the same references must appear under clusterSet (PayrollClusterSets). Existing payroll JSON files used for import/export must be migrated — breaking change