From 7066ae01978799fd638acfe82e6b1267ce1b30a8 Mon Sep 17 00:00:00 2001 From: Volodymyr Herashchenko Date: Mon, 16 Nov 2026 12:15:22 +0200 Subject: [PATCH] feat(docs): add roadmap --- docs/roadmap/README.md | 54 ++++++++++++++++++++++++++++++++++++++++++ docs/roadmap/WSC.md | 51 +++++++++++++++++++++++++++++++++++++++ 2 files changed, 105 insertions(+) create mode 100644 docs/roadmap/README.md create mode 100644 docs/roadmap/WSC.md diff --git a/docs/roadmap/README.md b/docs/roadmap/README.md new file mode 100644 index 0000000..0b21957 --- /dev/null +++ b/docs/roadmap/README.md @@ -0,0 +1,54 @@ +# Wallet-Session Roadmap + +This table has a list of all the features we'd like. Each top-level item links to a markdown file +which has a detailed description of the feature and its sub-feature. + +Before working on a nontrivial feature, please make sure that it has a document in this tree. + +## Taking Work + +While working on a feature, try to keep the `Status` and `PRs` column in this table up to date: + +* When a feature is first suggested, throw it in the table with status `Needs design`. +* Once it has a design document, change this to `Proposed`. +* Once it is in progress, change it to `In progress` and then `Completed`. + +To take on a feature, first make sure that it has an entry in the table and appears in a +linked design document. Each design document should have an owner listed. (If there is no +owner, please volunteer or ask somebody to take this on.) **The owner's responsibility is +to respond to complaints when things are stalled or blocked.** The owner doesn't necessarily +need write or review code (unless this is blocking progress) or even maintain the design +document; they are just the central point-of-contact for the feature. + +Once the feature has a document, go ahead and just open PRs against the repo as you normally +would. (If there are controversial high-level design considerations in your PR, you may need +to first PR to update the design document, but ideally this won't happen too much because +we don't want process friction when implementing stuff we've agreed to do.) + +## Current Priorities + +This section gives an idea of what we are working on. We should endeavor to update at least once +per quarter. + +Our highest priorities right now (Q4 2026) are: + +* the outstanding client review findings `WSC-1` +* the generalized client API `WSC-2` +* the high-level async API `WSC-5`, designed alongside `WSC-2` + +WASM and UniFFI support follow once the generalized API is stable. + +# All Features + +Feel free to add any feature to this list, no matter how futuristic or +low-priority. Its purpose is to define the DAG of dependencies so that we can +understand what is blocking the high-priority items. + +| ID | Name | Description | Dependencies | Status | PRs | +|--------|----------------------|------------------------------------------------------------------------------------------------------|---------------|--------------|-----| +| WSC | [Wallet Session Client](WSC.md) | Client crate as a library to communicate with the Relay | - | - | - | +| WSC-1 | Review findings | Address all client review findings | - | In progress | #14 | +| WSC-2 | Generalized core | Expose high-level client functions as a library API so application protocols can be built on top | WSC-1 | Needs design | - | +| WSC-3 | WASM support | Verify wasm-bindgen exposure, dependency support and a `wasm32-unknown-unknown` build | WSC-2, WSC-5 | Needs design | - | +| WSC-4 | Bindings support | Verify UniFFI bindings for the client, on the same API surface as WASM | WSC-2, WSC-5 | Needs design | - | +| WSC-5 | Async API | High-level async API spec, applied wherever it fits best without prescribing implementation details | WSC-1 | Needs design | - | diff --git a/docs/roadmap/WSC.md b/docs/roadmap/WSC.md new file mode 100644 index 0000000..7b827cf --- /dev/null +++ b/docs/roadmap/WSC.md @@ -0,0 +1,51 @@ +# Wallet Session Client + +**Owner:** TBD + +## Goal + +Finish the client in [`crates/client`](../../crates/client/) as a production-ready +library for the [relay protocol](../spec.md), then make its high-level API +available to application protocols and language bindings. + +## 1. Review Findings (WSC-1) + +Address the outstanding findings from security scanning. + +## 2. Generalized Core (WSC-2) + +Expose high-level client functions as a library API so application protocols can +be built on top. Transport, pairing, encryption, replay protection and +reliability remain internal to the client; a protocol supplies its messages and +handlers. + +This needs a design: the API shape, extension points and error model are not yet +defined. + +## 3. WASM Support (WSC-3) + +Verify that the client can be exposed as a `wasm-bindgen` module. Audit all +dependencies, confirm the crate compiles for `wasm32-unknown-unknown`, and +document any required platform abstractions. + +## 4. Bindings Support (WSC-4) + +Provide the same kind of verification for UniFFI bindings. The binding surface +should follow the generalized core and async API rather than exposing internals. + +## 5. Async API (WSC-5) + +Write a high-level specification for the client's async API: how callers send, +receive, await events and cancel operations, and where the async boundary fits +best for native, WASM and UniFFI consumers. This is a design-level specification, +not an implementation description. + +## Acceptance Criteria + +* All review findings are resolved or explicitly deferred. +* An application protocol can be implemented on the generalized API without + copying transport, encryption, replay or reliability code. +* The client compiles for `wasm32-unknown-unknown` and can be loaded as a + `wasm-bindgen` module. +* UniFFI bindings build and expose the same high-level API surface. +* The async API is documented at a level sufficient to implement a binding.