Replace register/replace with publish so new graphs can ship without rewriting old ones, and keep in-flight work on the version it started with.
Co-authored-by: Cursor <cursoragent@cursor.com>
Lower the compiler baseline to 17, rename listDefinitions to queryDefinitions, and align README/usage/roadmap with planned vs out-of-scope work.
Co-authored-by: Cursor <cursoragent@cursor.com>
Give hosts an append-only timeline, post-commit OrdoEventListener hooks, and durable ACTION results without blocking the approval flow.
Co-authored-by: Cursor <cursoragent@cursor.com>
- Added query method to InMemoryApprovalTaskRepository for filtering and pagination of approval tasks.
- Enhanced InMemoryProcessInstanceRepository with query method for filtering and pagination of process instances.
- Introduced Page and PageRequest classes for handling paginated results.
- Updated JdbcApprovalTaskRepository and JdbcProcessInstanceRepository to support querying with pagination.
- Added tests for querying, filtering, and pagination in InMemory and JDBC repositories.
- Implemented dynamic WHERE clause building for SQL queries in PageSupport.
Map the existing graph model from files so hosts can register flows without Java literals. Existing ids still insert-only on restart.
Co-authored-by: Cursor <cursoragent@cursor.com>
Require ProcessDefinition transitions so the engine can branch or end by conditionKey, persist them in JDBC, and pin repository text files to LF.
Co-authored-by: Cursor <cursoragent@cursor.com>
- ApprovalStep now has candidates + ApprovalPolicy (ANY/ALL) instead of a single assignee; ApprovalStep.single(...) kept as a convenience factory for the single-approver case
- AssigneeResolver resolves per-candidate; TaskStatus gains SKIPPED for auto-skipped sibling tasks
- DefaultOrdoEngine creates one task per candidate and advances per-step according to policy (ANY: first approval wins and skips the rest, only rejects once everyone rejects; ALL: fail-fast on first rejection, only advances once everyone approves)
- Added ApprovalTaskRepository.findByInstanceIdAndStepId and JDBC/in-memory implementations
- Added V2 migration: ordo_approval_step.policy column + ordo_step_candidate table
- Updated ApprovalStepMapper/ApprovalTaskMapper/JdbcProcessDefinitionRepository for the new schema, handling null action for SKIPPED tasks
- Migrated all call sites (tests, ordo-example, rhizome demo) to ApprovalStep.single(...) and the 3-arg AssigneeResolver
- Added ANY/ALL policy test coverage; fixed JdbcTestSupport to apply the V2 migration and strip SQL comments before splitting on ';'
- OrdoJdbcAutoConfiguration: use string-based @AutoConfigureAfter(name = ...)
to reference both Spring Boot 3.x and 4.x DataSource/Flyway autoconfiguration
classes, since Boot 4 relocated these to dedicated modules.
- flatten-maven-plugin: switch flattenMode from resolveCiFriendliesOnly to
defaults so inter-module dependency versions are fully resolved (not just
the project's own <version>) when installed, fixing dependency resolution
for external consumers.
- Align sibling-module dependency versions to use \ consistently.
- Ignore generated .flattened-pom.xml files.
Lightweight linear approval engine (v0.1):
- ordo-api: domain model, OrdoEngine port, repository SPI with
conditional updates (insertIfAbsent, completeIfPending) and a
TransactionExecutor port for atomic multi-step writes
- ordo-core: DefaultOrdoEngine running register/start/approve/reject
inside a transaction boundary; in-memory engine and repositories
- ordo-storage-jdbc: thread-bound JDBC transactions, normalized V1
schema migration, Jackson-based ProcessContext JSON codec
- tests: unit tests plus H2 integration tests; PostgreSQL integration
tests run against a local instance via ordo.test.pg.* properties and
skip when unreachable
Co-Authored-By: Claude Code <noreply@anthropic.com>
