v0.8.0

Added

• Node API to subscribe to peer connection events #625
• Support private network secured by pre-shared key #635

Changed

• Update libp2p to version 0.53.2 and apply API changes #631
• Move relay registration into main network service event loop #632
• Adjust level of replication and materialization logs #639

Fixed

• Handle connection ids greater than 9 in Peer impl of Human trait #634
• Check if blob file exists before deleting it from fs #636
• Inconsistent blob storage warning was wrongly shown #638
• Safely handle missing document when retrieving document view from store #637

v0.7.4

Added

• Introduce PeerAddress struct to help resolve String to internal address types #621
• Re-dial all configured known peers on schedule #622

Changed

• Increase max_pending_connections_in and max_pending_connections_out #628

Fixed

• Re-materialize blobs which were only partially written to disc due to node crash #618
• Include all logs for target schemas during replication #620
• Re-apply unhandled operations during startup of materializer service #623

v0.7.3

Fixed

• Handle null values returned from empty child relation queries #614

v0.7.2

Changed

• Accept domain name and ip addresses for peers in configuration file and cli #612
• Allow setting config file path via environment variables #611

v0.7.1

Added

• Export ConfigFile which can be de/serialized to and from a config file #607

Fixed

• Fix bug where known schemas are not replicated between nodes #603.

v0.7.0

Added

• Introduce a migrate API on the Node to allow publishing data programmatically #598

Changed

• Update p2panda-rs to v0.8.1 #599

v0.6.0

Added

• Serve static files from blobs directory #480 🥞
• Add method to store for pruning document views #491
• Introduce BlobStore #484
• Task for automatic garbage collection of unused documents and views #500 🥞
• Blobs directory configuration #549
• Integrate Bytes operation value #554
• Implement dependency replication for blob_v1 and blob_piece_v1 documents #514
• Remove deleted/unused blobs from the file system #571

Changed

• HTTP routes to serve files with correct content type headers #544
• Build a byte buffer over paginated pieces when assembling blobs #547
• Stream blob data in chunks to files to not occupy too much memory #551
• Remove unused methods from EntryStore #560
• Updates for new hash serialization in p2panda-rs #569
• Use libp2p 0.5.3 #570
• Optimize test data generation methods #572
• Use SocketAddr in network config instead of MultiAddr #576
• Update p2panda-rs to 0.8.0 #585
• Update libp2p to 0.52.4 #596

Fixed

• Make sure temporary directory does not run out of scope #557
• Deduplicate generated schema field by key in proptests #558
• Do not panic when blob does not have all pieces yet #563
• Fix blob tasks being triggered too often #578
• Fix schema tasks being triggered too often #581
• Fix blobs getting corrupted when written to the file system #587
• Fix pagination bug when only one field is selected and sorted at the same time #593
• Fix SQLite in-memory database overwrite by giving them each a random name #595

v0.5.0

Added

• Dial peers discovered via mDNS #331
• Simplify network CLI options and configuration #322
• Introduce autonat and relay network protocols #314
• Introduce identify and rendezvous network protocols / behaviours #304
• Introduce libp2p networking service and configuration #282
• Create and validate abstract queries #302
• Support paginated, ordered and filtered collection queries #308
• SQL query for collections #311
• Add custom validation to GraphQL scalar types #318
• Introduce property tests for GraphQL query API with proptest #338
• Introduce initial libp2p network behaviour for replication protocol #365
• Replication protocol session manager #363
• Replication message de- / serialization #375
• Naive protocol replication #380
• Integrate replication manager with networking stack #387 🥞
• Reverse lookup for pinned relations in dependency task #434
• Persist and maintain index of operation's position in document #438
• Introduce dialer behaviour with retry logic #444
• Introduce peer sampling to the replication service #463
• Only replicate and materialize configured "supported schema" #569
• Parse supported schema ids from config.toml #473
• Fix relayed connections, add DCUtR Holepunching and reduce CLI args #502
• Announce supported schema ids in network before replication #515
• Allow & block lists, direct dial known peers, connect to multiple relays #542

Changed

• Migrate CLI from structopt to clap #289
• Rework test runner and test node population patterns and refactor test_utils #277
• Implement API changes to p2panda-rs storage traits, new and breaking db migration #268
• Move all test utils into one module #275
• Use new version of async-graphql for dynamic schema generation #287
• Restructure graphql module #307
• Removed replication service for now, preparing for new replication protocol #296
• Bring back e2e GraphQL API tests #342
• Move GraphQL types into separate modules #343
• Set default order for root queries to document id #352
• Remove property tests again because of concurrency bug #347
• Incrementally update documents in materializer #280
• Decouple p2panda's authentication data types from libp2p's #408
• Make TaskInput an enum and other minor clean ups in materialiser #429
• Use libp2p 0.52.0 #425
• Check for duplicate entries arriving to Ingest before consuming #439
• Replicate entries in their topologically sorted document order #442
• Remove "quick commit" from materialization service #450
• Reduce WARN level logging in network and replication services #467
• mDNS and AutoNAT disabled by default #475
• By default, nodes support any schema #487
• Rework networking service #502
• Deduplicate peer connections when initiating replication sessions #525
• Improve consistency and documentation of configuration API #528
• Improve log level config and user interface #539

Fixed

• Correct use of sqlx transactions #285
• Fix race-condition of mutably shared static schema store during testing #269
• Introduce flag to requeue tasks in worker queue, fixes race-condition in materialization logic #286
• Update breaking API calls for new p2panda-rs 0.7.0 version #293
• Handle decoding and parsing empty pinned relation lists #336
• Make value on QueryRow an option #339
• Fix race conditions in SQLite database test runner #345
• Remove duplicate Cursor implementation, query code clean up #346
• Fix SQL ordering to correctly assemble resulting rows to documents #347
• Fix parsing lists arguments twice in GraphQL #354
• Sort paginated query field rows by document view id #354
• Fix missing field when filtering owner #359
• Bind untrusted user filter arguments in SQL query #358
• Reintroduce property tests for GraphQL query api #362
• Fix cursor pagination over ordered queries #412
• Fix issue where filtered queries fail on updated fields #409
• Fix insertion of view before document is materialized #413
• Handle duplicate document view insertions in reduce task #410
• Fix race condition when check for existing view ids was too early #420
• Use fork of asynchronous-codec to temporarily fix CBOR decoding bug #440
• Fix composing of circuit relay address #451
• Correct selection of comparison field during pagination #471
• Don't check for affected_rows on task deletion #461
• Do not critically fail when view does not exist due to race condition #460
• Do nothing on log insertion conflict #468
• Don't update or announce an update in schema provider if a schema with this id exists already #472
• Do nothing on document_view insertion conflicts #474
• Only over-write http_port when cli arg is passed #489
• Move deserialization into PeerMessage to distinct variants correctly #538

Open Sauce

• CI: Temporary workaround for Rust compiler bug #417
• CI: Update actions #455

v0.4.0

Added

• GraphQL replication service gets and verifies new entries and inserts them into the db #137
• Dynamically generated GraphQL types and query fields for accessing materialised documents #141 🥞
• validation and domain modules used for publish and next args API #204 🥞
• Schema task and schema provider that update when new schema views are materialised #166 🥞
• Service ready signal #218
• Validate operations against their claimed schema #245

Changed

• Refactor scalars and replication API, replace graphql-client with gql_client #184
• Give error types of worker a string for better debugging #194
• Bump p2panda-rs which now supports log id's starting from 0 #207
• Removed unused field entry_hash from operation data model #221
• Detach test helpers from test storage provider implementation #237
• Remove Scalar suffix from scalar types in GraphQL schema #231
• Implement new API for untagged operations #245
• Use DocumentStore trait from p2panda_rs #249
• Increase broadcast channel sizes #257
• Use transaction in insert_document() #259
• Use Human display strings in info logging #251
• Implement p2panda-rs API changes for PublicKey and previous() #262

Fixed

• Don't return errors from SchemaStore when a schema could not be constructed #192
• Filter out deleted documents in get_documents_by_schema SQL query #193
• Resolve implicit __typename field on dynamically generated GraphQL objects #236
• Allow Content-Type header #236
• Do not forget to register DocumentIdScalar #252
• Fix pagination during replication #199
• Load DocumentId to be able to resolve it in meta field query #258

Mushroom Time

• Write a lib-level doc string #263

v0.3.0

Added

• Introduce GraphQL endpoint #81
• Generic task queue with worker pool #82
• Service manager #90
• Service error handling, refactor runtime #92
• Refactor module structure, propagate errors in worker to service manager #97
• Restructure storage modules and remove JSON RPC #101
• Implement new methods required for replication defined by EntryStore trait #102
• Implement SQL OperationStore #103
• GraphQL client API with endpoint for retrieving next entry arguments #119
• GraphQL endpoint for publishing entries #123
• Implement SQL DocumentStore #118
• Implement SQL SchemaStore #130
• Reduce and dependency tasks #144
• GraphQL endpoints for replication #100
• Inform materialization service about new operations #161
• e2e publish entry tests #167
• Reschedule pending tasks on startup #168
• Debug logging in reduce task #175

Changed

• Move to tokio async runtime #75
• Implement SQL storage using p2panda_rs storage provider traits #80
• Improve Signal efficiency in ServiceManager #95
• EntryStore improvements #123
• Improvements for log and entry table layout #124
• Update StorageProvider API after p2panda-rs changes #129
• Move SqlStorage into shared Context #135
• Refactor tests to use fixtures exported from p2panda-rs #147
• Use DocumentViewId for previous operations #147
• Use VerifiedOperation #158
• Refactor test_db helper method #176
• Update publishEntry params and nextEntryArgs response fields #181

Fixed

• Fix high CPU usage of idle workers #136
• Improve CI, track test coverage #139
• Fix compatibility with PostgreSQL, change sqlx runtime to tokio #170
• Use UPSERT for inserting or updating documents #173
• Don't critically fail reduce task when document missing #177