Release History / Changelog

This is a high-level overview of the additions, removals, and changes for releases of wasmcloud and related tooling. This is not a complete list. For all of the specific details, check out the release notes for each release in the appropriate repository.

Release 0.57

Added
  • Added WASMCLOUD_CTL_TOPIC_PREFIX environment variable for specifying a prefix other than wasmbus.ctl in #452
Changed
  • wasmCloud hosts releases are now built with mix releases instead of Distillery, changing the commands necessary to run the host in #451. Largely this will require any scripts running wasmCloud to change foreground to start, and start to daemon.

Release 0.56

Added
  • Added Policy enforcement over NATS to allow users to specify custom topics to additionally restrict invocations and starting actors or providers in #442
Changed
  • Fixed console warnings for duplicate dropdown IDs in #445
  • Fixed a helm naming issue for structured logging in #443

Release 0.55

Added
  • Added OCI_REGISTRY environment variable for specifying a registry to use credentials for in #423
  • Add IPv6 support for NATS connections in #408
  • OpenTelemetry Tracing Support in #400, #403, #406, #407, #411
Changed
  • Log messages now are sent to stderr instead of stdout in #418
  • Actor messages are now round-robined between available actors using internal dispatch instead of a queue subscribe in #408, #415, #420, #422

Release 0.54

Added
  • Structured logging support in #368 which will format all host and actor logs as JSON for easily parsed logs. Human-readable text logs are still the default option.
Changed
  • HostData struct passed to providers on stdin now includes a structured_logging_enabled field to notify providers of the request for structured logs in #368

Release 0.53

Added
  • NATS message “bubble wrap” that prevents host processes from exiting when there’s a temporary loss of NATS connectivity in #363
  • Invocation chunking support, any invocation that exceeds the NATS max size limit will be automatically chunked and delivered with actors and providers running at least wasmbus-rpc 0.8.3 #354
Changed
  • Identified and fixed a memory spike with starting capability providers, each capability provider will now require only ~10Mb free memory to start rather than ~500Mb in #347
  • Capability providers that exit, gracefully or due to panic, will now be properly cleaned up and removed from the host inventory in #360
  • Actors that exit, gracefully or due to panic, because of a Wasmex issue will now safely exit #366
  • RPC failures that are caused by a timeout will now indicate the cause of the issue in the host log in #362

Release 0.52

Added
  • Ability to configure the dashboard listen host address at runtime in #317
  • Support for human-friendly host names in #319
  • Events for invocation success and failure in #320
  • Kubernetes applier support in #321
  • Experimental support for Bindle as an alternative to provider archives in #323
Changed
  • Now performs first host heartbeat and actor/provider health checks immediately, changed in #328

Release 0.51

Added
  • Emit claims for actors and providers in their respective start events (ActorStarted, ProviderStarted)
  • Human friendly relative times for claims expiration and not-before-dates
  • Actors can now be “hot watched” in the web dashboard which will automatically reload an actor upon changes. This dramatically reduces the developer feedback loop for actor development
  • Image reference is now included on various events for further information
  • Release tarballs from host_core, which is a packaged version of a host without the wasmCloud dashboard and UI assets
Changed
  • Updated a warning message in the dashboard logs to indicate unhandled events are ignored, rather than unsupported
  • Acknowledge actor and provider start commands in the control interface once they pass validation which avoids waiting for OCI downloads to occur before returning from the command. This comes with the caveat that when you receive the response from the control interface for these commands, it’s likely that the actor and provider have not fully started yet. The preferred way to assert this is to monitor the event stream for the respective Started events

Release 0.50

Added
  • The wasmcloud host has been ported to Elixir (OTP) to provide greater flexibility, scalability and reliability. Actor and Capability providers won’t see any elixir code, but this change is expected to make it easier to develop several features on the roadmap.
  • A new dashboard, built with Phoenix, shows the status of the lattice, and enables starting and stopping actors and capability providers.
  • A new host implementation in JavaScript that runs in the browser or on nodejs. A WebAssembly actor can be deployed without changes to any host - on linux, macos, windows, or in a browser.
Changed
  • Capability providers communicate with the wasmcloud host over a secure protocol via NATS, rather than being loaded as a library (as they were with host versions <= 0.18). This change allows us to support capability providers written in languages other than Rust, and is more flexible
  • The Interface Definition Language (IDL) has changed from WIDL to Smithy, and a new code generator generates code from smithy libraries.

Release 0.18.0

Added
  • Actor and Provider name and revision are now included with the control interface HostInventory query (PR #154). Names and revisions are not displayed yet with wash’s ctl get inventory command.
  • Released wash v0.4.0, the integrated REPL environment no longer requires NATS to run which dramatically simplifies the onboarding experience (wash PR #107)
  • get_actor_identity and get_provider_identity functions on the Host API for retrieving the name, image reference, and revision number of an actor or provider respectively (PR #163)
  • --nsprefix option on the wasmcloud binary which allows you to set a namespace other than “default” with the stock wasmcloud binary (PR #161)
  • Configured a Gitpod environment for a complete pre-configured in-browser development environment for wasmcloud. (PR #162)
Changed
  • Control interface and lattice WIDL files are now stored in the wasmcloud repository instead of capability-providers (PR #154)
  • Host API providers() function now returns a Vec of 3-tuples (public_key, contract_id, link_name) - was previously just public key
Fixed
  • Fixed a bug where the control interface and host API could incorrectly report the OCI image reference of an actor or provider. (PR #159 and PR #163)
  • Fixed a bug where live updating an actor would not update the stored OCI image references (PR #160)

Release 0.17.0

Added
  • Added additional logging and error messages to common operations (PR #143 and PR #148)
  • Added the ability to remove a link on the Host API and control interface (PR #147)
  • Added extensive Rust documentation on the wasmcloud_host crate (PR #151)
  • Added additional queries to the Host API to query the lattice cache. This allows users directly invoking the Host API to retrieve the same information available via the control interface (PR #153)
Fixed
  • Fixed an issue where operations could be called on a host that had not started (PR #142)

Release 0.16.0

The 0.16.0 release includes multiple bug fixes, documentation improvements, and additional features.

Added
  • Added bencher sample for determining wasmcloud performance (PR #100)
Changed
  • Renamed wascc-codec to wasmcloud-provider-core (PR #102)
  • Updated embedded Extras provider to use wasmcloud:extras capability contract (PR #109)
  • Updated actix-rt to 2.1.0 which is based on tokio 1.X. This is a breaking change for users that embed a wasmcloud-host in their Rust program. (PR #114 and PR #135)
  • Changed fields on the Host struct to allow it to be sent between threads (PR #132)
Fixed
  • Fixed a bug where loading a capability provider caused a SIGSEGV on Linux (PR #99)
  • Fixed a bug where the live update operation didn’t allow for differing OCI references (PR #128)

Release 0.15.0

The 0.15.0 release is the single biggest release we’ve had since the project began. It is essentially a bottom-up rewrite of the entire wasmcloud host runtime. In addition to this internal rewrite, we’ve incorporated lessons learned from the past year of development and applied those to the way we build capability providers, define actor interfaces, and support the developer experience. The core theme of this release is the ruthless focus on developer experience.

Added
  • Consolidated multiple CLI tools (nk, wascap, etc) into the new wasmcloud shell wash
  • Added a fully functioning REPL to the wash CLI
  • Provider archives - allows for the bundling of multiple OS/CPU native libraries into a single archive
  • OCI support - You can now deploy actors and capability providers into a host/lattice from an OCI registry
  • First-party capability providers and sample actors are all available in our OCI registry wasmcloud.azurecr.io
  • wasmcloud and wash are now installable via package cloud, homebrew, and other convenient methods
Changed
  • Completely rewrote the wasmcloud-host crate to allow for easier maintenance and adding features
  • The wasmcloud-host crate is now entirely async and based on the actix framework
  • Too many others to list here…