Project Status

Opt-in 0.2 alpha

The 0.2 alpha release run publishes eight packages at 0.2.0-alpha.0: @vooya/vite, @vooya/vue, @vooya/react, @vooya/build-core, @vooya/provider-rust, @vooya/preset, @vooya/rspack, and @vooya/webpack. @vooya/core and @vooya/compiler remain at 0.1.0-beta.0; Solid and Svelte adapters remain at 0.1.0-beta.2. Use this exact mixed version graph; there is no Solid/Svelte alpha package in this release. Octane remains private and deferred.

The original publication workflow failed after publishing; independent registry recovery and GitHub prerelease reconciliation are complete. See the release record for evidence and limits.

The default quickstart continues to use beta. Alpha is an explicit opt-in for the managed toolchain, bounded Vite 8/Vite+ and Rust bundler paths, and SSR-safe client-mounted islands. See alpha installation. Publication does not establish every framework or SSR combination, and a source fixture is not by itself a registry-consumer test.

The first Vooya beta, 0.1.0-beta.0, was published for all ten public packages. The current npm beta channel has @vooya/vite, @vooya/react, @vooya/solid, and @vooya/svelte at 0.1.0-beta.2, and @vooya/vue at 0.1.0-beta.1. The other five packages (@vooya/core, @vooya/compiler, @vooya/build-core, @vooya/rspack, and @vooya/webpack) remain at 0.1.0-beta.0. Use each package's npm beta tag. It remains a prerelease, not a stable compiler or a production compatibility promise. The main branch may contain unreleased work.

Vue and React are supported first-party adapters. Solid and Svelte are shipped experimental adapters, not unfinished package placeholders: their Rust-file components, events, Store actions, reactive snapshots, and declarations have browser evidence. Their tested scope is narrower; see the framework capability matrix.

The beta.0 release workflow passed the full release gate, all four framework browser fixtures, and clean Vue/React consumers using both packed artifacts and exact npm versions. Solid/Svelte do not yet have that separate clean npm-consumer acceptance step.

The beta.1 release run passed the release gate, exact registry-consumer verification, and GitHub Release publication for the updated package set. This is release evidence for those versions; it does not expand the framework compatibility matrix.

The current versions include two fixes found while integrating Vooya Lab: Vite preserves explicit ?raw Rust source imports, and the four adapters accept generated Store interfaces without an arbitrary string index signature. The Store type fix is available since Vue beta.1 and React/Solid/Svelte beta.2; the raw import fix requires Vite beta.2.

The original ten beta packages share one release workflow (alpha adds @vooya/preset and @vooya/provider-rust, for twelve public packages):

• @vooya/compiler
• @vooya/core
• @vooya/build-core
• @vooya/vite
• @vooya/vue
• @vooya/react
• @vooya/solid
• @vooya/svelte
• @vooya/rspack
• @vooya/webpack

Package versions may differ. Build packages pin internal dependencies exactly; install the documented package tags instead of forcing every package to the same version. Release checks verify the complete dependency graph.

Changesets 3.0.3 manages independent package versions and changelogs, with no fixed or linked version groups. Published source and dependency changes carry package-scoped release notes. The Release workflow prepares a release PR on the selected release branch after an explicit prepare dispatch. Publication requires a separate publish dispatch for the reviewed version commit, followed by the full gate, exact registry checks and per-package GitHub Releases. Ordinary source and version merges do not publish automatically. The receipt records the tested commit and verified package set. This automation does not promote a prerelease to stable or expand the compatibility claims below. See the release guide.

Published beta capabilities

• Compile Rust-file components and stores into application-level WASM.
• Generate typed mount, prop update, event, dispose, and ABI bindings.
• Import one .rs file as a Vue 3, React 19, Solid 1.9, or Svelte 5 component through Vite.
• Generate mirrored TypeScript declarations under .vooya/types from component contracts.
• Compile optional PostCSS-based scoped styles.
• Map extracted Rust diagnostics back to .rs source lines.
• Configure registry, Git, and application-relative path dependencies.
• Recover from Rust build errors and coalesce rapid development saves.
• Validate Vue Counter, React Counter, TaskList, and 100,000-row DataGrid flows in real browsers.
• Validate loop-created Rust listeners, cloned event dispatch, and repeated mount/unmount behavior in both Vue and React browser fixtures.
• Ship vooya doctor for coherent Cargo-selected Rust target, CLI-version, and rustup-path diagnostics.
• Demonstrate a Vue-hosted 150,000-point Rust/WASM Canvas scatter plot.
• Use Vooya Lab as a self-hosting and evidence program; Lab findings that affect product contracts return here as focused issues and fixes.
• Build packed npm artifacts from a project outside the repository checkout.
• Verify a test-only precompiled Vue WASM consumer in a clean Vite project without Cargo, Rust, a Rust target, wasm-bindgen, or the Vite plugin.
• Extract versioned Rust-file schema records from WASM, validate file groups, and generate central .d.rs.ts declarations for Vue, React, Solid, and Svelte consumers.
• Build and exercise a real Rust-file component and store through Vite 7 + Vue 3.5, including scoped CSS, the public vooya authoring dependency, generated WASM bindings, and Chromium interaction.
• Build and exercise Rust-file components and instance-scoped stores through Vite 7 + React 19, including StrictMode mount cleanup, store actions, snapshot subscriptions, and atomic component prop patches.
• Build and exercise a Rust-file Component and instance-scoped Store through Vite 7 + Solid 1.9, including a callback event, generated Accessor declaration, Store action, and Accessor-driven DOM update. Adapter unit tests cover owner cleanup and a Store that resolves after its owner is gone.
• Build and exercise a Rust-file Component and instance-scoped Store through Svelte 5 + Vite 7, including callback delivery, Store action, Component prop update, generated Readable declarations, and Chromium without runtime errors. The same fixture unmounts a child and verifies exactly one Component handle disposal plus one generated Store disposal.
• Verify Rust-file Vite development rebuilds, failed-build recovery, and successful full reloads without preserving component state.
• Expose the instance-scoped Store contract to Vue, React, Solid, and Svelte through generated useName() entries with aligned names and fields; retain each framework's native Ref, snapshot, Accessor, or Readable state container and keep useVooyaStore as an advanced adapter-level helper.
• Generate a Vite virtual module for .rs stores, including independent factory creation, snapshot reads, subscriptions, actions, and disposal.
• Reconcile the generated Rust root and copied module tree when .rs files are added, removed, or renamed.
• Exercise the DOM-only rsx! runtime in Chromium, including signal bindings, owned disposal, conditional branches, and keyed list identity.

Current limits

• Source consumers need Cargo, the WASM target, and wasm-bindgen-cli.
• Vite >=7 <9 is the supported source-authoring bundler range, with Vite 8 as the primary compatibility target. Vite 7 Rust-file Vue, React, Solid, and Svelte production/browser fixtures pass, and the packed Vite 8 fixture covers production, rebuild, HMR, and error recovery. Rspack >=2.1.10 has an experimental first-party adapter with Rsbuild, Rslib, and direct Rspack fixtures. Webpack >=5 has an experimental first-party adapter. Both published non-Vite adapters retain transitional .voo fixture evidence; their .rs integration requires the opt-in 0.2 alpha packages. Webpack 4, Rollup, Turbopack, and other bundlers remain unsupported.
• The newer Vite+ four-framework dev/build matrix runs against current-source tarballs. It does not expand the published beta claim or cover every Vite+ tool.
• Turbopack has no compatibility claim. Its loader API evidence and unverified integration requirements are recorded in Turbopack research; Webpack and Rspack evidence must not be generalized to Turbopack.
• No precompiled component product is currently published; the retained Vue fixture is build-contract evidence, not a user-facing package.
• A non-trivial component still uses some direct web_sys APIs.
• Props, event payloads, action arguments, and store snapshots share the owned ABI v1 mapping for finite numbers, bigint, nullability, vectors, tuples, and string-key maps. User-defined recursive schemas, borrowed values, arbitrary generics, non-string maps, and TypedArray transport remain outside the current boundary.
• Rust Store schema/declarations and generated .rs Store modules are available for Vue, React, Solid, and Svelte. Browser-level interaction is covered by test:rust-vue, test:rust-react, test:rust-solid, and test:rust-svelte; Solid and Svelte claims remain limited to their Vite 7 fixtures and named adapter tests.
• The public vooya Cargo authoring crate is copied into the packed @vooya/core source bundle during the package build, so clean npm consumers can compile .rs files without a checkout of this repository. It is not a separately published Rust crate.
• Reactive dependencies and cleanup are explicit; batch coalesces synchronous signal writes, re-entrant effect cycles are suppressed, and opt-in tracked_effect collects dynamic dependencies. rsx! now compiles keyed for loops and conditional if/else branches with owned cleanup.
• Successful Rust HMR performs a full reload and loses component state; the Rust-file recovery path is covered by test:rust-hmr.
• vooya doctor is a local diagnostic and deterministic PATH candidate selector, not a toolchain installer; it warns when its valid Cargo choice differs from the first Cargo on PATH.
• SSR, hydration, slots, SvelteKit integration, and standalone application rendering are out of the current evidence boundary.
• Prerelease ABI revisions can be breaking.
• The default CI browser evidence is a small Chromium smoke suite plus the bundler integration fixtures. Extended DataGrid, scatter, trace, and Firefox checks remain available through explicit manual commands.
• Vooya Lab cases are evidence and product discovery, not a separate support matrix or an automatic beta gate; see RFC 0011.

Available opt-in 0.2 scope

The optional managed preset, SSR-safe islands and Nuxt fixture, Webpack/Rspack Rust-file integration, expanded Vite+ evidence, and merged Rust provider extraction are available in the opt-in alpha package graph. The release-line review identifies two merged 0.1 maintenance backports; neither has been published. Do not treat the current main branch as a maintenance release candidate.

Octane remains a private workspace experiment, excluded from this 0.2 batch and public release gate. Its independent browser fixture is retained.

Next milestones

Beta.0 passed strict TypeScript checks and Chromium interaction for clean Rust-file Vue and React consumers, first with packed artifacts and then with exact registry versions. Future releases retain these separate acceptance steps. Current source adds separate clean Vite 8 dev/build consumers for all four adapters. Registry browser acceptance remains follow-up work before a broader support claim. Managed toolchain installation through @vooya/preset@0.2.0-alpha.0 addresses the bounded source-authoring setup in #129. It is optional and does not remove host SDK requirements; beta users continue using their system Rust/WASM tools.

1. Extend the existing rsx! conditional/keyed rendering and owned cleanup with evidence from more real component use cases.
2. Design a supported, explicitly named component product on top of the generic precompiled Vue producer.
3. Define state-preserving HMR semantics.
4. Extend the shipped owned struct/unit-enum declarations and scoped-name lookup where real consumers need richer Rust type resolution; recursive and borrowed values still require an ownership design.
5. Expand browser and framework coverage beyond the named fixtures. The existing Vue 3.6/Vapor smoke remains experimental; it does not establish full support.

The benchmark result remains deliberately modest: the first 100,000-row case showed approximate parity with its Vue baseline. See the recorded result rather than assuming WASM is automatically faster.