merge: sync orchestration before queue 00001KV5W3PJ3

This commit is contained in:
Keisuke Hirata 2026-06-18 22:11:00 +09:00
commit 63d864e1f0
No known key found for this signature in database
12 changed files with 1590 additions and 82 deletions

View File

@ -1,8 +1,8 @@
---
title: 'Plugin: execute Plugin Tool with minimal WASM runtime'
state: 'inprogress'
state: 'done'
created_at: '2026-06-15T14:48:59Z'
updated_at: '2026-06-17T09:50:53Z'
updated_at: '2026-06-18T12:39:30Z'
assignee: null
readiness: 'implementation_ready'
risk_flags: ['plugin', 'wasm', 'tool-runtime', 'sandbox', 'capability-boundary', 'cancellation']

View File

@ -128,3 +128,185 @@ State handling:
- Ticket remains `inprogress` because acceptance and worktree side effects already happened; it is not silently returned to `queued`.
---
<!-- event: implementation_report author: yoi-orchestrator at: 2026-06-18T12:05:59Z -->
## Implementation report
Implementation resumed:
Previous `SpawnPod` failure was recovered after Orchestrator runtime refresh. Reused existing clean child worktree and successfully spawned Coder Pod.
Worktree:
- `/home/hare/Projects/yoi/.worktree/00001KV5W3PHW-plugin-wasm-tool-runtime`
- branch: `impl/00001KV5W3PHW-plugin-wasm-tool-runtime`
Coder Pod:
- `yoi-coder-00001KV5W3PHW`
No duplicate `queued -> inprogress` transition was performed; this resumes the already accepted in-progress work.
---
<!-- event: implementation_report author: yoi-coder-00001KV5W3PHW at: 2026-06-18T12:31:01Z -->
## Implementation report
Implementation report from Coder Pod `yoi-coder-00001KV5W3PHW`:
Commit:
- `10d12148 feat: run plugin tools through wasm runtime`
Changed files:
- `Cargo.lock`
- `crates/manifest/src/plugin.rs`
- `crates/pod/Cargo.toml`
- `crates/pod/src/feature/plugin.rs`
- `package.nix`
Implemented behavior:
- Added minimal sandboxed Plugin Tool WASM runtime using `wasmi`.
- Enabled registered Plugin Tool calls now execute through normal `Tool::execute` / `ToolOutput` / `ToolResult` paths.
- Added manifest-side helper to load resolved runtime module from pinned package:
- validates runtime kind/ABI;
- re-reads package from resolved path;
- verifies deterministic package digest before selecting runtime entry;
- bounds package/module size and reports safe diagnostics.
- Added `yoi-plugin-wasm-1` host ABI:
- required export: `yoi_tool_call()`;
- required guest export: `memory`;
- host imports under `yoi:tool` only:
- `tool_name_len() -> i32`
- `tool_name_read(ptr, len) -> i32`
- `input_len() -> i32`
- `input_read(ptr, len) -> i32`
- `output_write(ptr, len) -> i32`
- No WASI, filesystem, network, environment, or arbitrary host imports are provided; unsupported imports fail closed.
Bounds / failure behavior:
- Input JSON bound: `64 KiB`.
- Output JSON bound: `64 KiB`.
- Output summary bound: `1 KiB`.
- Linear memory bound: `2 MiB`.
- Table bound: `256`.
- Fuel bound: `5_000_000`.
- Wall timeout: `1s`.
- Diagnostics/errors are bounded and sanitized before surfacing as ordinary Tool errors.
- Malformed input JSON, malformed output JSON, output schema mismatch, oversized output, missing runtime/package/module, unsupported imports, load failure, and non-terminating execution fail closed.
Tests added/updated cover:
- Sample Plugin package Tool executes through WASM runtime.
- Tool input JSON reaches the module.
- Output JSON returns through ordinary Tool result path.
- Malformed input JSON.
- Malformed output JSON.
- Output schema mismatch.
- Oversize output.
- Non-terminating/high-fuel execution fail closed.
- Missing runtime/package/module diagnostic is safe and bounded.
- No ambient WASI/fs/network/env import access.
- Existing Plugin Tool registration behavior.
Validation reported by coder:
- Passed: `cargo fmt --check`
- Passed: `cargo check -p pod`
- Passed: `cargo test -p manifest plugin::`
- Passed: `cargo test -p pod feature::plugin`
- Passed: `git diff --check`
- Passed: `nix build .#yoi`
- Passed: `nix build .#yoi --no-link`
Residual risks / notes:
- ABI is intentionally minimal and not a general Plugin capability model.
- Permission grants, richer host APIs, WASI-like APIs, per-plugin runtime configuration knobs, and schema-driven rich output contracts remain follow-up work.
- Cancellation has no explicit `ToolExecutionContext` cancellation token today; this implementation provides safe timeout/fuel-bounded execution and maps timeout/join failure into ordinary Tool errors.
---
<!-- event: review author: yoi-reviewer-00001KV5W3PHW at: 2026-06-18T12:37:14Z status: approve -->
## Review: approve
Review result: approve
Evidence checked:
- Child worktree のみ確認。
- `/home/hare/Projects/yoi/.worktree/00001KV5W3PHW-plugin-wasm-tool-runtime`
- implementation commit: `10d12148`
- base: `d32fb3bc`
- 変更箇所を確認。
- `crates/pod/src/feature/plugin.rs`
- `crates/manifest/src/plugin.rs`
- `crates/pod/Cargo.toml`
- `Cargo.lock`
- `package.nix`
- Read-only validation:
- Passed: `git diff --check d32fb3bc..10d12148`
Acceptance review:
- WASM runtime は `wasmi` を使い、WASI setup はなく、host import surface は `yoi_plugin_tool_v1` の最小 input/output 関数だけに限定されている。
- `wasi_snapshot_preview1.fd_write``env.getenv` など ambient-style imports は拒否される。
- Plugin Tools は ordinary `Feature` tool contribution / `ToolExecutor` path で `ToolOutput` / `ToolError` を返し、hidden context injection path は見当たらない。
- module/input/output/summary/error/memory/table/stack/fuel などの bounds が実装されている。
- malformed output、invalid UTF-8、schema mismatch、oversize output は `ToolError` として fail closed する。
- runtime module selection は resolved plugin record、resolved package path、recorded digest、runtime kind/ABI、entry path に基づき、archive digest を再検証してから module bytes を使用している。
- non-termination は fuel exhaustion と runtime timeout wrapper で fail closed する。現行 `ToolExecutionContext` に cancellation token がないため、明示的 cancellation-token integration がない点は、この Ticket の範囲では blocker ではないと判断した。
- `wasmi``default-features = false``std` / `extra-checks` のみ、`wat` は dev-only。WASI/network/process capability library の追加は見当たらない。
Validation not run by reviewer:
- `cargo test` / `cargo check` / `nix build` は read-only review scope では build artifacts を書くため未実行。Coder reported validation を evidence とした。
Risk assessment:
- residual risk は低〜中程度。
- sandbox boundary は Ticket scope に対して妥当。No WASI / no fs / no network / no env imports、bounded IO、bounded diagnostics、deterministic digest check がある。
- `tokio::time::timeout``spawn_blocking` の実行中 task を強制 abort しないが、fuel により guest execution は有限化される。より強い cooperative cancellation semantics は Tool execution API/runtime cancellation model の follow-up として扱うべき。
- Permission grants / richer host API authority は正しく未実装の follow-up に残っている。
Conclusion:
- Approved. No blocker remains.
---
<!-- event: implementation_report author: yoi-orchestrator at: 2026-06-18T12:39:30Z -->
## Implementation report
Integration report:
Merged implementation branch into orchestration branch:
- branch: `impl/00001KV5W3PHW-plugin-wasm-tool-runtime`
- implementation commit: `10d12148 feat: run plugin tools through wasm runtime`
- merge commit: `05cd788c merge: plugin wasm tool runtime`
Reviewer result:
- `approve` recorded from `yoi-reviewer-00001KV5W3PHW`.
Validation in Orchestrator worktree `/home/hare/Projects/yoi/.worktree/orchestration`:
- Passed: `cargo fmt --check`
- Passed: `cargo check -p pod`
- Passed: `cargo test -p manifest plugin::`
- Passed: `cargo test -p pod feature::plugin`
- 21 passed
- Passed: `git diff --check`
- Passed: `nix build .#yoi --no-link`
Outcome:
- Enabled Plugin Tool calls now execute through minimal `wasmi` runtime with no WASI/fs/network/env host imports.
- Runtime module selection is deterministic from resolved package path/digest/runtime config and rechecks package digest.
- Tool input/output/errors are bounded and returned through ordinary Tool result/error path.
- Missing/malformed/oversize/schema-mismatch/non-terminating execution fails closed with safe Tool errors.
- Permission grants, richer host APIs, WASI-like APIs, and stronger cancellation-token integration remain follow-up scope.
Cleanup planned:
- Stop related coder/reviewer Pods.
- Remove only child implementation worktree/branch for this Ticket.
---
<!-- event: state_changed author: yoi-orchestrator at: 2026-06-18T12:39:30Z from: inprogress to: done reason: merged_validated field: state -->
## State changed
Reviewer approved, implementation branch merged into the orchestration branch, and focused plus packaging validation passed in the Orchestrator worktree. Marking Ticket done in the orchestration branch.
---

View File

@ -1,8 +1,8 @@
---
title: 'Panel startup latency E2E を一覧データ描画完了基準に修正する'
state: 'inprogress'
state: 'done'
created_at: '2026-06-15T16:44:06Z'
updated_at: '2026-06-17T09:50:53Z'
updated_at: '2026-06-18T12:25:14Z'
assignee: null
readiness: 'implementation_ready'
risk_flags: ['panel', 'e2e', 'startup-latency', 'readiness-metric', 'ticket-list-rendering']

View File

@ -124,3 +124,161 @@ State handling:
- Ticket remains `inprogress` because acceptance and worktree side effects already happened; it is not silently returned to `queued`.
---
<!-- event: implementation_report author: yoi-orchestrator at: 2026-06-18T12:05:59Z -->
## Implementation report
Implementation resumed:
Previous `SpawnPod` failure was recovered after Orchestrator runtime refresh. Reused existing clean child worktree and successfully spawned Coder Pod.
Worktree:
- `/home/hare/Projects/yoi/.worktree/00001KV62PF32-panel-rows-ready-e2e`
- branch: `impl/00001KV62PF32-panel-rows-ready-e2e`
Coder Pod:
- `yoi-coder-00001KV62PF32`
No duplicate `queued -> inprogress` transition was performed; this resumes the already accepted in-progress work.
---
<!-- event: implementation_report author: yoi-coder-00001KV62PF32 at: 2026-06-18T12:19:45Z -->
## Implementation report
Implementation report from Coder Pod `yoi-coder-00001KV62PF32`:
Commit:
- `fffdfd27 test: assert panel rows-ready fixture data`
Changed files:
- `tests/e2e/src/lib.rs`
- Added concrete `ExpectedPanelTicketRow` matcher.
- Exposed fixture ready/planning Ticket ids/titles/states.
- Added `wait_for_first_visible_frame`, `wait_for_fixture_ticket_rows_ready`, and delayed-row negative assertion helpers.
- `tests/e2e/tests/panel.rs`
- Replaced main startup readiness assertion from `rows.len() >= 2` / “full ready” to concrete ready fixture Ticket row readiness.
- Added matcher regression test proving absent/wrong fixture data is rejected.
- Added delayed reload assertion that first frame can arrive while fixture rows are still not rendered.
- `crates/tui/src/multi_pod.rs`
- Added comments clarifying `panel_ready` is first-visible-frame only, while `rows_rendered` snapshots are the data-backed readiness surface.
Behavior / evidence:
- Main startup readiness now distinguishes:
- first frame: legacy `panel_ready`, can be loading/empty;
- rows ready: concrete `rows_rendered` containing expected fixture Ticket id/title/state/kind;
- full/background completion: still observable via background task events, but no longer the startup readiness proxy.
- Startup readiness assertion is no longer satisfied by first frame alone or by weak `rows.len() >= N` checks.
- Delayed reload test asserts first visible frame can arrive while expected fixture Ticket row remains absent for the delayed window.
Validation reported by coder:
- Passed: `cargo test -p yoi-e2e --features e2e --no-run`
- Passed: `cargo test -p yoi-e2e --features e2e panel -- --nocapture`
- Passed: `cargo check -p yoi-e2e -p yoi -p tui`
- Passed: `cargo fmt --check`
- Passed: `git diff --check`
- Not run: `nix build .#yoi` — no runtime/package/dependency changes.
Measured values from focused post-commit reruns:
- `panel_fixture_ticket_row_ready_has_startup_budget`
- first visible frame: `20.403589ms` / budget `1.5s`
- fixture rows ready: `120.605891ms` / budget `5s`
- `panel_first_visible_render_arrives_before_background_reload`
- first visible frame: `20.311085ms` / budget `1.5s`
- reload held; test asserts ready fixture Ticket row remains absent for delayed window before exit.
Repository status:
- Child implementation worktree clean at `fffdfd27`.
Residual gaps:
- None known.
---
<!-- event: review author: yoi-reviewer-00001KV62PF32 at: 2026-06-18T12:24:00Z status: approve -->
## Review: approve
Review result: approve
Evidence checked:
- 対象 worktree / branch / commit が期待通りで clean であることを確認。
- branch: `impl/00001KV62PF32-panel-rows-ready-e2e`
- HEAD: `fffdfd2721fed5171d4dd9780f893b9bb323ab8a`
- base: `d32fb3bc3c82f1cb7e4a29d4fdf6e908ba495c7f`
- Diff scope は以下3ファイルに限定されていることを確認。
- `crates/tui/src/multi_pod.rs`
- `tests/e2e/src/lib.rs`
- `tests/e2e/tests/panel.rs`
Acceptance / intent review:
- Main startup readiness は `panel_fixture_ticket_row_ready_has_startup_budget` で concrete fixture Ticket row readiness を待つ形になっている。
- `rows.len() >= N` ではなく、fixture Ticket row の kind / id / title / status-state を matcher で確認している。
- `panel_ready` / `wait_for_first_visible_frame` は first-visible-frame only としてコメント・helper 名で区別されている。
- `wait_for_fixture_ticket_rows_ready` が startup rows-ready signal として定義されている。
- `panel_fixture_ticket_row_matcher_rejects_absent_fixture_data` により absent/wrong fixture data を reject する挙動が確認されている。
- delayed reload test は first frame が先に出る一方、held reload 中は expected fixture Ticket row が描画されないことを確認している。
- 既存 Panel mouse / row selection E2E は残っており、action dispatch なし・mouse capture behavior の coverage は維持されている。
- fixture isolation / no-provider / no-network 前提は維持されている。
Validation performed by reviewer:
- Passed: `cargo fmt --check`
- Passed: `git diff --check d32fb3bc..HEAD`
Validation not run by reviewer:
- `cargo test` / `cargo check` は read-only review scope では build/artifact 書き込みを伴うため未実行。Coder reported validation を evidence として確認した。
Risk assessment:
- residual risk は低い。変更は test/harness focused で Ticket boundary に沿っている。
- timing test は環境依存性が残るが、readiness predicate は first frame/loading paint ではなく concrete fixture row rendering を測るよう修正されている。
Conclusion:
- Approved. No blocker remains.
---
<!-- event: implementation_report author: yoi-orchestrator at: 2026-06-18T12:25:14Z -->
## Implementation report
Integration report:
Merged implementation branch into orchestration branch:
- branch: `impl/00001KV62PF32-panel-rows-ready-e2e`
- implementation commit: `fffdfd27 test: assert panel rows-ready fixture data`
- merge commit: `226eca7a merge: panel rows-ready e2e`
Reviewer result:
- `approve` recorded from `yoi-reviewer-00001KV62PF32`.
Validation in Orchestrator worktree `/home/hare/Projects/yoi/.worktree/orchestration`:
- Passed: `cargo test -p yoi-e2e --features e2e --no-run`
- Passed: `cargo test -p yoi-e2e --features e2e panel -- --nocapture`
- panel E2E: 6 passed
- first visible frame examples: ~20.34ms / budget 1.5s
- fixture rows ready example: ~120.54ms / budget 5s
- Passed: `cargo check -p yoi-e2e -p yoi -p tui`
- Passed: `cargo fmt --check`
- Passed: `git diff --check`
Outcome:
- Startup readiness E2E now uses concrete fixture Ticket row render readiness instead of first visible/loading frame or weak row-count checks.
- `panel_ready` remains first-visible-frame only; `rows_rendered` fixture row matching is the data-backed readiness surface.
- Delayed reload coverage verifies first frame can arrive before rows-ready and expected fixture Ticket row remains absent while reload is held.
- Existing Panel mouse/row selection E2E remains covered.
Cleanup planned:
- Stop related coder/reviewer Pods.
- Remove only child implementation worktree/branch for this Ticket.
---
<!-- event: state_changed author: yoi-orchestrator at: 2026-06-18T12:25:14Z from: inprogress to: done reason: merged_validated field: state -->
## State changed
Reviewer approved, implementation branch merged into the orchestration branch, and E2E-focused validation passed in the Orchestrator worktree. Marking Ticket done in the orchestration branch.
---

129
Cargo.lock generated
View File

@ -1633,6 +1633,12 @@ version = "0.2.186"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "68ab91017fe16c622486840e4c83c9a37afeff978bd239b5293d61ece587de66"
[[package]]
name = "libm"
version = "0.2.16"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "b6d2cec3eae94f9f509c767b45932f1ada8350c4bdb85af2fcab4a3c14807981"
[[package]]
name = "libredox"
version = "0.1.16"
@ -2360,6 +2366,8 @@ dependencies = [
"tools",
"tracing",
"uuid",
"wasmi",
"wat",
"workflow",
]
@ -3342,6 +3350,12 @@ dependencies = [
"windows-sys 0.61.2",
]
[[package]]
name = "spin"
version = "0.9.8"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "6980e8d7511241f8acf4aebddbb1ff938df5eebe98691418c4468d0b72a96a67"
[[package]]
name = "stable_deref_trait"
version = "1.2.1"
@ -3354,6 +3368,16 @@ version = "1.1.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "a2eb9349b6444b326872e140eb1cf5e7c522154d69e7a0ffb0fb81c06b37543f"
[[package]]
name = "string-interner"
version = "0.19.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "23de088478b31c349c9ba67816fa55d9355232d63c3afea8bf513e31f0f1d2c0"
dependencies = [
"hashbrown 0.15.5",
"serde",
]
[[package]]
name = "string_cache"
version = "0.8.9"
@ -4229,7 +4253,17 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "990065f2fe63003fe337b932cfb5e3b80e0b4d0f5ff650e6985b1048f62c8319"
dependencies = [
"leb128fmt",
"wasmparser",
"wasmparser 0.244.0",
]
[[package]]
name = "wasm-encoder"
version = "0.246.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "61fb705ce81adde29d2a8e99d87995e39a6e927358c91398f374474746070ef7"
dependencies = [
"leb128fmt",
"wasmparser 0.246.2",
]
[[package]]
@ -4240,8 +4274,8 @@ checksum = "bb0e353e6a2fbdc176932bbaab493762eb1255a7900fe0fea1a2f96c296cc909"
dependencies = [
"anyhow",
"indexmap",
"wasm-encoder",
"wasmparser",
"wasm-encoder 0.244.0",
"wasmparser 0.244.0",
]
[[package]]
@ -4257,6 +4291,56 @@ dependencies = [
"web-sys",
]
[[package]]
name = "wasmi"
version = "0.51.5"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "bb321403ce594274827657a908e13d1d9918aa02257b8bf8391949d9764023ff"
dependencies = [
"spin",
"wasmi_collections",
"wasmi_core",
"wasmi_ir",
"wasmparser 0.228.0",
]
[[package]]
name = "wasmi_collections"
version = "0.51.5"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "e9b8e98e45a2a534489f8225e765cbf1cb9a3078072605e58158910cf4749172"
dependencies = [
"string-interner",
]
[[package]]
name = "wasmi_core"
version = "0.51.5"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "c25f375c0cdf14810eab07f532f61f14d4966f09c747a55067fdf3196e8512e6"
dependencies = [
"libm",
]
[[package]]
name = "wasmi_ir"
version = "0.51.5"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "624e2a68a4293ecb8f564260b68394b29cf3b3edba6bce35532889a2cb33c3d9"
dependencies = [
"wasmi_core",
]
[[package]]
name = "wasmparser"
version = "0.228.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "4abf1132c1fdf747d56bbc1bb52152400c70f336870f968b85e89ea422198ae3"
dependencies = [
"bitflags 2.11.0",
"indexmap",
]
[[package]]
name = "wasmparser"
version = "0.244.0"
@ -4269,6 +4353,39 @@ dependencies = [
"semver",
]
[[package]]
name = "wasmparser"
version = "0.246.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "71cde4757396defafd25417cfb36aa3161027d06d865b0c24baaae229aac005d"
dependencies = [
"bitflags 2.11.0",
"indexmap",
"semver",
]
[[package]]
name = "wast"
version = "246.0.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "fe3fe8e3bf88ad96d031b4181ddbd64634b17cb0d06dfc3de589ef43591a9a62"
dependencies = [
"bumpalo",
"leb128fmt",
"memchr",
"unicode-width",
"wasm-encoder 0.246.2",
]
[[package]]
name = "wat"
version = "1.246.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "4bd7fda1199b94fff395c2d19a153f05dbe7807630316fa9673367666fd2ad8c"
dependencies = [
"wast",
]
[[package]]
name = "web-sys"
version = "0.3.94"
@ -4730,9 +4847,9 @@ dependencies = [
"serde",
"serde_derive",
"serde_json",
"wasm-encoder",
"wasm-encoder 0.244.0",
"wasm-metadata",
"wasmparser",
"wasmparser 0.244.0",
"wit-parser",
]
@ -4751,7 +4868,7 @@ dependencies = [
"serde_derive",
"serde_json",
"unicode-xid",
"wasmparser",
"wasmparser 0.244.0",
]
[[package]]

View File

@ -637,6 +637,146 @@ pub fn resolve_plugin_config_for_startup(
snapshot
}
/// Load the recorded WASM runtime module for a resolved plugin package.
///
/// Restore and execution paths use this helper instead of reading arbitrary
/// package paths directly so module selection remains tied to the resolved
/// package identity, runtime manifest entry, and deterministic package digest.
pub fn read_resolved_plugin_runtime_module(
record: &ResolvedPluginRecord,
limits: &PluginDiscoveryLimits,
) -> Result<Vec<u8>, PluginDiagnostic> {
let runtime = record.manifest.runtime.as_ref().ok_or_else(|| {
PluginDiagnostic::new(
PluginDiagnosticKind::Missing,
PluginDiagnosticPhase::Manifest,
"resolved plugin package does not declare a WASM runtime",
)
.with_source(record.source)
.with_identity(&record.identity)
.with_package(&record.package_label)
.with_digest(&record.digest)
})?;
if runtime.kind != "wasm" {
return Err(PluginDiagnostic::new(
PluginDiagnosticKind::Api,
PluginDiagnosticPhase::Manifest,
"plugin runtime kind is unsupported",
)
.with_source(record.source)
.with_identity(&record.identity)
.with_package(&record.package_label)
.with_digest(&record.digest));
}
if runtime.abi.as_deref() != Some("yoi-plugin-wasm-1") {
return Err(PluginDiagnostic::new(
PluginDiagnosticKind::Api,
PluginDiagnosticPhase::Manifest,
"plugin WASM ABI is unsupported",
)
.with_source(record.source)
.with_identity(&record.identity)
.with_package(&record.package_label)
.with_digest(&record.digest));
}
let metadata = fs::metadata(&record.package_path).map_err(|error| {
PluginDiagnostic::new(
PluginDiagnosticKind::Io,
PluginDiagnosticPhase::Discovery,
format!(
"resolved plugin package metadata could not be read: {}",
safe_io_error(&error)
),
)
.with_source(record.source)
.with_identity(&record.identity)
.with_package(&record.package_label)
.with_digest(&record.digest)
})?;
if !metadata.is_file() {
return Err(PluginDiagnostic::new(
PluginDiagnosticKind::Malformed,
PluginDiagnosticPhase::Discovery,
"resolved plugin package is not a regular file",
)
.with_source(record.source)
.with_identity(&record.identity)
.with_package(&record.package_label)
.with_digest(&record.digest));
}
if metadata.len() > limits.max_package_size_bytes {
return Err(PluginDiagnostic::new(
PluginDiagnosticKind::Bounds,
PluginDiagnosticPhase::Discovery,
"resolved plugin package exceeds the configured package size bound",
)
.with_source(record.source)
.with_identity(&record.identity)
.with_package(&record.package_label)
.with_digest(&record.digest));
}
let bytes = fs::read(&record.package_path).map_err(|error| {
PluginDiagnostic::new(
PluginDiagnosticKind::Io,
PluginDiagnosticPhase::Discovery,
format!(
"resolved plugin package content could not be read: {}",
safe_io_error(&error)
),
)
.with_source(record.source)
.with_identity(&record.identity)
.with_package(&record.package_label)
.with_digest(&record.digest)
})?;
let archive = parse_stored_zip(&bytes, &record.package_label, record.source, limits)?;
let actual_digest = deterministic_digest(&archive.files);
if !digest_matches(&record.digest, &actual_digest) {
return Err(PluginDiagnostic::new(
PluginDiagnosticKind::Digest,
PluginDiagnosticPhase::Resolution,
"resolved plugin package digest does not match current package content",
)
.with_source(record.source)
.with_identity(&record.identity)
.with_package(&record.package_label)
.with_digest(actual_digest));
}
validate_manifest_path(
&runtime.entry,
&archive,
&record.package_label,
record.source,
&record.manifest.id,
)?;
let normalized = normalize_archive_path(&runtime.entry).ok_or_else(|| {
PluginDiagnostic::new(
PluginDiagnosticKind::Traversal,
PluginDiagnosticPhase::Manifest,
"plugin manifest references a path outside the package root",
)
.with_source(record.source)
.with_identity(&record.identity)
.with_package(&record.package_label)
.with_digest(&record.digest)
})?;
archive.files.get(&normalized).cloned().ok_or_else(|| {
PluginDiagnostic::new(
PluginDiagnosticKind::Missing,
PluginDiagnosticPhase::Manifest,
"plugin runtime module entry is missing from the package",
)
.with_source(record.source)
.with_identity(&record.identity)
.with_package(&record.package_label)
.with_digest(&record.digest)
})
}
#[derive(Clone, Debug)]
struct PluginStore {
source: PluginSourceKind,

View File

@ -35,11 +35,13 @@ workflow-crate = { package = "workflow", path = "../workflow" }
uuid = { workspace = true, features = ["v7"] }
session-metrics = { workspace = true }
arc-swap = "1.9.1"
wasmi = { version = "0.51.1", default-features = false, features = ["std", "extra-checks"] }
[dev-dependencies]
dotenv = "0.15.0"
futures = { workspace = true }
tempfile = { workspace = true }
wat = "1.241.2"
[build-dependencies]
toml = { workspace = true }

View File

@ -1,17 +1,23 @@
//! Plugin package contributions for model-visible Tool schemas.
//!
//! This module registers *enabled* plugin package tool surface definitions as
//! unavailable Tool stubs. It deliberately does not execute plugin code or grant
//! plugin permissions; the runtime/WASM executor belongs to a later boundary.
//! This module registers *enabled* plugin package tool surface definitions and
//! executes Tool calls through the minimal sandboxed `yoi-plugin-wasm-1` WASM
//! ABI. It deliberately does not grant filesystem, network, environment, hook,
//! service, ingress, or richer host API authority; those remain follow-up
//! boundaries.
use std::collections::HashSet;
use std::sync::Arc;
use std::time::Duration;
use async_trait::async_trait;
use llm_worker::tool::{
Tool, ToolDefinition, ToolError, ToolExecutionContext, ToolMeta, ToolOrigin, ToolOutput,
};
use manifest::plugin::{PluginConfig, PluginSurface, ResolvedPluginRecord};
use manifest::plugin::{
PluginConfig, PluginDiscoveryLimits, PluginSurface, ResolvedPluginRecord,
read_resolved_plugin_runtime_module,
};
use serde_json::Value;
use super::{
@ -115,7 +121,8 @@ impl FeatureModule for PluginToolFeature {
})?;
context.tools().register(ToolContribution::new(
tool.name.clone(),
plugin_runtime_missing_definition(
plugin_wasm_tool_definition(
self.record.clone(),
tool.name.clone(),
tool.description.clone(),
tool.input_schema.clone(),
@ -127,7 +134,18 @@ impl FeatureModule for PluginToolFeature {
}
}
fn plugin_runtime_missing_definition(
const PLUGIN_WASM_HOST_MODULE: &str = "yoi:tool";
const PLUGIN_WASM_ENTRYPOINT: &str = "yoi_tool_call";
const PLUGIN_WASM_MAX_INPUT_BYTES: usize = 64 * 1024;
const PLUGIN_WASM_MAX_OUTPUT_BYTES: usize = 64 * 1024;
const PLUGIN_WASM_MAX_SUMMARY_BYTES: usize = 1024;
const PLUGIN_WASM_FUEL: u64 = 5_000_000;
const PLUGIN_WASM_TIMEOUT: Duration = Duration::from_secs(1);
const PLUGIN_WASM_MEMORY_BYTES: usize = 2 * 1024 * 1024;
const PLUGIN_WASM_TABLE_ELEMENTS: usize = 256;
fn plugin_wasm_tool_definition(
record: ResolvedPluginRecord,
name: String,
description: String,
input_schema: Value,
@ -139,7 +157,8 @@ fn plugin_runtime_missing_definition(
.description(description.clone())
.input_schema(input_schema.clone())
.origin(origin.clone()),
Arc::new(PluginRuntimeMissingTool {
Arc::new(PluginWasmTool {
record: record.clone(),
name: name.clone(),
origin: origin.clone(),
}) as Arc<dyn Tool>,
@ -147,29 +166,390 @@ fn plugin_runtime_missing_definition(
})
}
struct PluginRuntimeMissingTool {
struct PluginWasmTool {
record: ResolvedPluginRecord,
name: String,
origin: ToolOrigin,
}
#[async_trait]
impl Tool for PluginRuntimeMissingTool {
impl Tool for PluginWasmTool {
async fn execute(
&self,
_input_json: &str,
input_json: &str,
_ctx: ToolExecutionContext,
) -> Result<ToolOutput, ToolError> {
Err(ToolError::ExecutionFailed(format!(
"plugin tool runtime missing/unavailable for `{}` from `{}` (digest {}, package {} api {})",
self.name,
self.origin.plugin_ref,
self.origin.digest,
self.origin.package_version,
self.origin.package_api_version
)))
if input_json.len() > PLUGIN_WASM_MAX_INPUT_BYTES {
return Err(ToolError::InvalidArgument(format!(
"plugin tool `{}` input exceeds {} bytes",
self.name, PLUGIN_WASM_MAX_INPUT_BYTES
)));
}
serde_json::from_str::<Value>(input_json).map_err(|error| {
ToolError::InvalidArgument(format!(
"plugin tool `{}` input is not valid JSON: {}",
self.name,
bounded_message(error.to_string())
))
})?;
let record = self.record.clone();
let name = self.name.clone();
let plugin_ref = self.origin.plugin_ref.clone();
let digest = self.origin.digest.clone();
let input = input_json.as_bytes().to_vec();
let execution =
tokio::task::spawn_blocking(move || run_plugin_wasm_tool(record, name, input));
match tokio::time::timeout(PLUGIN_WASM_TIMEOUT, execution).await {
Ok(Ok(Ok(output))) => Ok(output),
Ok(Ok(Err(error))) => Err(ToolError::ExecutionFailed(format!(
"plugin WASM tool `{}` from `{}` (digest {}) failed closed: {}",
self.name,
plugin_ref,
digest,
error.bounded_message()
))),
Ok(Err(error)) => Err(ToolError::ExecutionFailed(format!(
"plugin WASM tool `{}` from `{}` (digest {}) cancelled/failed to join: {}",
self.name,
plugin_ref,
digest,
bounded_message(error.to_string())
))),
Err(_) => Err(ToolError::ExecutionFailed(format!(
"plugin WASM tool `{}` from `{}` (digest {}) timed out after {:?}",
self.name, plugin_ref, digest, PLUGIN_WASM_TIMEOUT
))),
}
}
}
#[derive(Debug)]
enum PluginWasmError {
Package(String),
Module(String),
Execution(String),
Output(String),
}
impl PluginWasmError {
fn bounded_message(&self) -> String {
match self {
Self::Package(message) => {
bounded_message(format!("package/module load error: {message}"))
}
Self::Module(message) => bounded_message(format!("WASM module error: {message}")),
Self::Execution(message) => bounded_message(format!("WASM execution error: {message}")),
Self::Output(message) => bounded_message(format!("WASM output error: {message}")),
}
}
}
#[derive(Debug)]
struct PluginWasmHostState {
tool_name: Vec<u8>,
input: Vec<u8>,
output: Vec<u8>,
output_error: Option<String>,
store_limits: wasmi::StoreLimits,
}
fn run_plugin_wasm_tool(
record: ResolvedPluginRecord,
tool_name: String,
input: Vec<u8>,
) -> Result<ToolOutput, PluginWasmError> {
let limits = PluginDiscoveryLimits::default();
let module_bytes = read_resolved_plugin_runtime_module(&record, &limits)
.map_err(|diagnostic| PluginWasmError::Package(diagnostic.message))?;
if module_bytes.len() > limits.max_file_size_bytes as usize {
return Err(PluginWasmError::Package(format!(
"WASM runtime module exceeds {} bytes",
limits.max_file_size_bytes
)));
}
let mut config = wasmi::Config::default();
config.consume_fuel(true);
config.set_max_recursion_depth(64);
config.set_max_stack_height(8 * 1024 * 1024);
let engine = wasmi::Engine::new(&config);
let module = wasmi::Module::new(&engine, &module_bytes[..])
.map_err(|error| PluginWasmError::Module(error.to_string()))?;
validate_wasm_imports(&module)?;
let store_limits = wasmi::StoreLimitsBuilder::new()
.memory_size(PLUGIN_WASM_MEMORY_BYTES)
.table_elements(PLUGIN_WASM_TABLE_ELEMENTS)
.instances(1)
.tables(1)
.memories(1)
.trap_on_grow_failure(true)
.build();
let mut store = wasmi::Store::new(
&engine,
PluginWasmHostState {
tool_name: tool_name.into_bytes(),
input,
output: Vec::new(),
output_error: None,
store_limits,
},
);
store.limiter(|state| &mut state.store_limits);
store
.set_fuel(PLUGIN_WASM_FUEL)
.map_err(|error| PluginWasmError::Execution(error.to_string()))?;
let mut linker = wasmi::Linker::<PluginWasmHostState>::new(&engine);
define_plugin_wasm_host_imports(&mut linker)?;
let instance = linker
.instantiate_and_start(&mut store, &module)
.map_err(|error| PluginWasmError::Execution(error.to_string()))?;
let entry = instance
.get_typed_func::<(), ()>(&store, PLUGIN_WASM_ENTRYPOINT)
.map_err(|error| PluginWasmError::Module(error.to_string()))?;
entry
.call(&mut store, ())
.map_err(|error| PluginWasmError::Execution(error.to_string()))?;
if let Some(error) = store.data().output_error.clone() {
return Err(PluginWasmError::Output(error));
}
decode_plugin_wasm_output(&store.data().output)
}
fn validate_wasm_imports(module: &wasmi::Module) -> Result<(), PluginWasmError> {
for import in module.imports() {
if import.module() != PLUGIN_WASM_HOST_MODULE {
return Err(PluginWasmError::Module(format!(
"unsupported import module `{}`; only `{}` is available",
import.module(),
PLUGIN_WASM_HOST_MODULE
)));
}
match import.name() {
"tool_name_len" | "tool_name_read" | "input_len" | "input_read" | "output_write" => {}
other => {
return Err(PluginWasmError::Module(format!(
"unsupported host import `{}`; no filesystem, network, environment, or WASI imports are available",
other
)));
}
}
}
Ok(())
}
fn define_plugin_wasm_host_imports(
linker: &mut wasmi::Linker<PluginWasmHostState>,
) -> Result<(), PluginWasmError> {
linker
.func_wrap(
PLUGIN_WASM_HOST_MODULE,
"tool_name_len",
|caller: wasmi::Caller<'_, PluginWasmHostState>| -> i32 {
caller.data().tool_name.len() as i32
},
)
.map_err(|error| PluginWasmError::Module(error.to_string()))?;
linker
.func_wrap(
PLUGIN_WASM_HOST_MODULE,
"input_len",
|caller: wasmi::Caller<'_, PluginWasmHostState>| -> i32 {
caller.data().input.len() as i32
},
)
.map_err(|error| PluginWasmError::Module(error.to_string()))?;
linker
.func_wrap(
PLUGIN_WASM_HOST_MODULE,
"tool_name_read",
|mut caller: wasmi::Caller<'_, PluginWasmHostState>, ptr: i32, len: i32| -> i32 {
write_host_bytes_to_guest(&mut caller, ptr, len, HostBuffer::ToolName)
},
)
.map_err(|error| PluginWasmError::Module(error.to_string()))?;
linker
.func_wrap(
PLUGIN_WASM_HOST_MODULE,
"input_read",
|mut caller: wasmi::Caller<'_, PluginWasmHostState>, ptr: i32, len: i32| -> i32 {
write_host_bytes_to_guest(&mut caller, ptr, len, HostBuffer::Input)
},
)
.map_err(|error| PluginWasmError::Module(error.to_string()))?;
linker
.func_wrap(
PLUGIN_WASM_HOST_MODULE,
"output_write",
|mut caller: wasmi::Caller<'_, PluginWasmHostState>, ptr: i32, len: i32| -> i32 {
read_guest_output(&mut caller, ptr, len)
},
)
.map_err(|error| PluginWasmError::Module(error.to_string()))?;
Ok(())
}
#[derive(Clone, Copy, Debug)]
enum HostBuffer {
ToolName,
Input,
}
fn write_host_bytes_to_guest(
caller: &mut wasmi::Caller<'_, PluginWasmHostState>,
ptr: i32,
len: i32,
buffer: HostBuffer,
) -> i32 {
if ptr < 0 || len < 0 {
return -1;
}
let bytes = match buffer {
HostBuffer::ToolName => caller.data().tool_name.clone(),
HostBuffer::Input => caller.data().input.clone(),
};
if len as usize != bytes.len() {
return -1;
}
let Some(memory) = caller
.get_export("memory")
.and_then(|export| export.into_memory())
else {
return -1;
};
match memory.write(caller, ptr as usize, &bytes) {
Ok(()) => bytes.len() as i32,
Err(_) => -1,
}
}
fn read_guest_output(
caller: &mut wasmi::Caller<'_, PluginWasmHostState>,
ptr: i32,
len: i32,
) -> i32 {
if ptr < 0 || len < 0 {
caller.data_mut().output_error = Some("guest output pointer/length is invalid".into());
return -1;
}
let len = len as usize;
if len > PLUGIN_WASM_MAX_OUTPUT_BYTES {
caller.data_mut().output_error = Some(format!(
"guest output exceeds {} bytes",
PLUGIN_WASM_MAX_OUTPUT_BYTES
));
return -1;
}
let Some(memory) = caller
.get_export("memory")
.and_then(|export| export.into_memory())
else {
caller.data_mut().output_error = Some("guest did not export linear memory".into());
return -1;
};
let mut output = vec![0; len];
if memory.read(&*caller, ptr as usize, &mut output).is_err() {
caller.data_mut().output_error = Some("guest output memory range is invalid".into());
return -1;
}
caller.data_mut().output = output;
len as i32
}
fn decode_plugin_wasm_output(bytes: &[u8]) -> Result<ToolOutput, PluginWasmError> {
if bytes.is_empty() {
return Err(PluginWasmError::Output(
"guest did not call output_write".into(),
));
}
if bytes.len() > PLUGIN_WASM_MAX_OUTPUT_BYTES {
return Err(PluginWasmError::Output(format!(
"guest output exceeds {} bytes",
PLUGIN_WASM_MAX_OUTPUT_BYTES
)));
}
let text = std::str::from_utf8(bytes)
.map_err(|error| PluginWasmError::Output(format!("guest output is not UTF-8: {error}")))?;
let value: Value = serde_json::from_str(text).map_err(|error| {
PluginWasmError::Output(format!("guest output is not valid JSON: {error}"))
})?;
let Value::Object(map) = value else {
return Err(PluginWasmError::Output(
"guest output JSON must be an object".into(),
));
};
for key in map.keys() {
if key != "summary" && key != "content" {
return Err(PluginWasmError::Output(format!(
"guest output contains unsupported key `{key}`"
)));
}
}
let summary = match map.get("summary") {
Some(Value::String(summary)) if !summary.is_empty() => summary.clone(),
Some(Value::String(_)) => {
return Err(PluginWasmError::Output(
"guest output summary must not be empty".into(),
));
}
Some(_) => {
return Err(PluginWasmError::Output(
"guest output summary must be a string".into(),
));
}
None => {
return Err(PluginWasmError::Output(
"guest output must include a summary string".into(),
));
}
};
if summary.len() > PLUGIN_WASM_MAX_SUMMARY_BYTES {
return Err(PluginWasmError::Output(format!(
"guest output summary exceeds {} bytes",
PLUGIN_WASM_MAX_SUMMARY_BYTES
)));
}
let content = match map.get("content") {
Some(Value::String(content)) => {
if content.len() > PLUGIN_WASM_MAX_OUTPUT_BYTES {
return Err(PluginWasmError::Output(format!(
"guest output content exceeds {} bytes",
PLUGIN_WASM_MAX_OUTPUT_BYTES
)));
}
Some(content.clone())
}
Some(Value::Null) | None => None,
Some(_) => {
return Err(PluginWasmError::Output(
"guest output content must be a string or null".into(),
));
}
};
Ok(ToolOutput { summary, content })
}
fn bounded_message(message: impl Into<String>) -> String {
let message = message.into();
let mut sanitized = String::with_capacity(message.len().min(512));
for ch in message.chars() {
if ch.is_control() && ch != '\n' && ch != '\t' {
sanitized.push(' ');
} else {
sanitized.push(ch);
}
if sanitized.len() >= 512 {
sanitized.truncate(512);
sanitized.push('…');
break;
}
}
sanitized
}
fn validate_declared_tool_names(record: &ResolvedPluginRecord) -> Result<(), FeatureInstallError> {
let mut seen = HashSet::new();
for tool in &record.manifest.tools {
@ -371,8 +751,14 @@ fn is_supported_schema_keyword(key: &str) -> bool {
#[cfg(test)]
mod tests {
use super::*;
use manifest::plugin::{PluginPackageManifest, SourceQualifiedPluginId};
use manifest::plugin::{
PluginDiscoveryOptions, PluginEnablementConfig, PluginPackageManifest,
PluginRuntimeManifest, SourceQualifiedPluginId, resolve_plugin_config_for_startup,
};
use serde_json::json;
use std::fs;
use std::path::Path;
use tempfile::TempDir;
fn tool(name: &str) -> manifest::plugin::PluginToolManifest {
manifest::plugin::PluginToolManifest {
@ -645,27 +1031,349 @@ mod tests {
}
#[tokio::test]
async fn registered_tool_executes_as_runtime_missing_error() {
let mut pending = Vec::new();
let mut hooks = crate::hook::HookRegistryBuilder::new();
let report = super::super::FeatureRegistryBuilder::default()
.with_module(PluginToolFeature::new(record(vec![tool("PluginSearch")])))
.install_into_pending(&mut pending, &mut hooks);
assert!(
report
.reports
.iter()
.all(|feature_report| feature_report.diagnostics.is_empty()),
"{:#?}",
report.reports
);
async fn registered_plugin_tool_executes_wasm_and_returns_normal_tool_result() {
let (_dir, record) = resolved_record_with_wasm(input_reaches_guest_module());
let origin = PluginToolFeature::new(record.clone()).origin();
let tool = PluginWasmTool {
record,
name: "PluginEcho".into(),
origin,
};
let output = tool
.execute(r#"{"x":1}"#, ToolExecutionContext::default())
.await
.unwrap();
assert_eq!(output.summary, "input reached");
assert_eq!(output.content.as_deref(), Some("ordinary tool result path"));
let result = llm_worker::tool::ToolResult::from_output("call-1", output);
assert_eq!(result.summary, "input reached");
assert!(
result
.content
.unwrap()
.contains("ordinary tool result path")
);
}
#[tokio::test]
async fn malformed_input_json_fails_before_wasm_execution() {
let (_dir, record) = resolved_record_with_wasm(input_reaches_guest_module());
let origin = PluginToolFeature::new(record.clone()).origin();
let tool = PluginWasmTool {
record,
name: "PluginEcho".into(),
origin,
};
let error = tool
.execute("not json", ToolExecutionContext::default())
.await
.unwrap_err();
assert!(error.to_string().contains("input is not valid JSON"));
}
#[tokio::test]
async fn malformed_output_fails_closed() {
let (_dir, record) = resolved_record_with_wasm(output_module(b"not json"));
let error =
run_plugin_wasm_tool(record, "PluginEcho".into(), br#"{}"#.to_vec()).unwrap_err();
assert!(error.bounded_message().contains("not valid JSON"));
}
#[tokio::test]
async fn schema_mismatch_output_fails_closed() {
let (_dir, record) = resolved_record_with_wasm(output_module(br#"{"summary":1}"#));
let error =
run_plugin_wasm_tool(record, "PluginEcho".into(), br#"{}"#.to_vec()).unwrap_err();
assert!(error.bounded_message().contains("summary must be a string"));
}
#[tokio::test]
async fn oversize_output_fails_closed() {
let (_dir, record) = resolved_record_with_wasm(oversize_output_module());
let error =
run_plugin_wasm_tool(record, "PluginEcho".into(), br#"{}"#.to_vec()).unwrap_err();
assert!(error.bounded_message().contains("exceeds"));
}
#[tokio::test]
async fn nonterminating_execution_fails_closed_with_fuel_boundary() {
let (_dir, record) = resolved_record_with_wasm(nonterminating_module());
let error =
run_plugin_wasm_tool(record, "PluginEcho".into(), br#"{}"#.to_vec()).unwrap_err();
let message = error.bounded_message();
assert!(
message.contains("Execution")
|| message.contains("fuel")
|| message.contains("execution"),
"{message}"
);
}
#[tokio::test]
async fn missing_runtime_module_returns_safe_bounded_tool_error() {
let record = record_with_missing_package_runtime();
let origin = PluginToolFeature::new(record.clone()).origin();
let tool = PluginWasmTool {
record,
name: "PluginSearch".into(),
origin,
};
let (_, tool) = pending[0]();
let error = tool
.execute("{}", ToolExecutionContext::default())
.await
.unwrap_err();
assert!(error.to_string().contains("runtime missing/unavailable"));
assert!(error.to_string().contains("project:example"));
let message = error.to_string();
assert!(message.contains("failed closed"));
assert!(message.contains("metadata could not be read"));
assert!(message.len() < 900);
assert!(message.contains("project:example"));
}
#[tokio::test]
async fn ambient_wasi_fs_network_env_imports_are_unavailable() {
let (_dir, record) = resolved_record_with_wasm(wasi_import_module());
let error =
run_plugin_wasm_tool(record, "PluginEcho".into(), br#"{}"#.to_vec()).unwrap_err();
let message = error.bounded_message();
assert!(message.contains("unsupported import module"), "{message}");
assert!(message.contains("wasi_snapshot_preview1"), "{message}");
}
fn record_with_missing_package_runtime() -> ResolvedPluginRecord {
let mut record = record(vec![tool("PluginSearch")]);
record.manifest.runtime = Some(PluginRuntimeManifest {
kind: "wasm".into(),
entry: "plugin.wasm".into(),
abi: Some("yoi-plugin-wasm-1".into()),
});
record
}
fn resolved_record_with_wasm(wasm: Vec<u8>) -> (TempDir, ResolvedPluginRecord) {
let dir = TempDir::new().unwrap();
let package_dir = dir.path().join(".yoi/plugins");
fs::create_dir_all(&package_dir).unwrap();
let package_path = package_dir.join("example.yoi-plugin");
write_plugin_package(&package_path, &wasm);
let config = PluginConfig {
enabled: vec![PluginEnablementConfig {
id: "project:example".parse().unwrap(),
surfaces: vec![PluginSurface::Tool],
..PluginEnablementConfig::default()
}],
resolved: Vec::new(),
diagnostics: Vec::new(),
};
let options = PluginDiscoveryOptions::new(dir.path());
let resolved = resolve_plugin_config_for_startup(&config, &options);
assert!(
resolved.diagnostics.is_empty(),
"{:#?}",
resolved.diagnostics
);
assert_eq!(resolved.resolved.len(), 1);
(dir, resolved.resolved[0].clone())
}
fn write_plugin_package(path: &Path, wasm: &[u8]) {
let manifest = br#"schema_version = 1
id = "example"
name = "Example"
version = "1.0.0"
description = "Example plugin"
surfaces = ["tool"]
[runtime]
kind = "wasm"
entry = "plugin.wasm"
abi = "yoi-plugin-wasm-1"
[[tools]]
name = "PluginEcho"
description = "Echo plugin tool"
input_schema = { type = "object", additionalProperties = true }
"#;
write_stored_zip(
path,
&[("plugin.toml", manifest.as_slice()), ("plugin.wasm", wasm)],
);
}
fn input_reaches_guest_module() -> Vec<u8> {
let ok = br#"{"summary":"input reached","content":"ordinary tool result path"}"#;
let bad = br#"{"summary":"input missing"}"#;
let wat = format!(
r#"(module
(import "yoi:tool" "input_len" (func $input_len (result i32)))
(import "yoi:tool" "input_read" (func $input_read (param i32 i32) (result i32)))
(import "yoi:tool" "output_write" (func $output_write (param i32 i32) (result i32)))
(memory (export "memory") 1)
(data (i32.const 0) "{}")
(data (i32.const 128) "{}")
(func (export "yoi_tool_call")
(local $len i32)
(local.set $len (call $input_len))
(if (i32.eq (local.get $len) (i32.const 7))
(then
(drop (call $input_read (i32.const 512) (local.get $len)))
(if (i32.eq (i32.load8_u (i32.const 517)) (i32.const 49))
(then (drop (call $output_write (i32.const 0) (i32.const {}))))
(else (drop (call $output_write (i32.const 128) (i32.const {}))))
)
)
(else (drop (call $output_write (i32.const 128) (i32.const {}))))
)
)
)"#,
wat_bytes(ok),
wat_bytes(bad),
ok.len(),
bad.len(),
bad.len()
);
wat::parse_str(wat).unwrap()
}
fn output_module(output: &[u8]) -> Vec<u8> {
let wat = format!(
r#"(module
(import "yoi:tool" "output_write" (func $output_write (param i32 i32) (result i32)))
(memory (export "memory") 1)
(data (i32.const 0) "{}")
(func (export "yoi_tool_call")
(drop (call $output_write (i32.const 0) (i32.const {})))
)
)"#,
wat_bytes(output),
output.len()
);
wat::parse_str(wat).unwrap()
}
fn oversize_output_module() -> Vec<u8> {
let wat = format!(
r#"(module
(import "yoi:tool" "output_write" (func $output_write (param i32 i32) (result i32)))
(memory (export "memory") 2)
(func (export "yoi_tool_call")
(drop (call $output_write (i32.const 0) (i32.const {})))
)
)"#,
PLUGIN_WASM_MAX_OUTPUT_BYTES + 1
);
wat::parse_str(wat).unwrap()
}
fn nonterminating_module() -> Vec<u8> {
wat::parse_str(
r#"(module
(memory (export "memory") 1)
(func (export "yoi_tool_call")
(local $remaining i32)
(local.set $remaining (i32.const 100000000))
(loop $again
(local.set $remaining (i32.sub (local.get $remaining) (i32.const 1)))
(br_if $again (local.get $remaining))
)
)
)"#,
)
.unwrap()
}
fn wasi_import_module() -> Vec<u8> {
wat::parse_str(
r#"(module
(import "wasi_snapshot_preview1" "fd_write" (func $fd_write))
(memory (export "memory") 1)
(func (export "yoi_tool_call"))
)"#,
)
.unwrap()
}
fn wat_bytes(bytes: &[u8]) -> String {
bytes
.iter()
.map(|byte| format!(r#"\{:02x}"#, byte))
.collect()
}
fn write_stored_zip(path: &Path, files: &[(&str, &[u8])]) {
let mut out = Vec::new();
let mut central = Vec::new();
for (name, data) in files {
let offset = out.len() as u32;
let name_bytes = name.as_bytes();
let crc = crc32(data);
write_u32(&mut out, 0x0403_4b50);
write_u16(&mut out, 20);
write_u16(&mut out, 0);
write_u16(&mut out, 0);
write_u16(&mut out, 0);
write_u16(&mut out, 0);
write_u32(&mut out, crc);
write_u32(&mut out, data.len() as u32);
write_u32(&mut out, data.len() as u32);
write_u16(&mut out, name_bytes.len() as u16);
write_u16(&mut out, 0);
out.extend_from_slice(name_bytes);
out.extend_from_slice(data);
write_u32(&mut central, 0x0201_4b50);
write_u16(&mut central, 20);
write_u16(&mut central, 20);
write_u16(&mut central, 0);
write_u16(&mut central, 0);
write_u16(&mut central, 0);
write_u16(&mut central, 0);
write_u32(&mut central, crc);
write_u32(&mut central, data.len() as u32);
write_u32(&mut central, data.len() as u32);
write_u16(&mut central, name_bytes.len() as u16);
write_u16(&mut central, 0);
write_u16(&mut central, 0);
write_u16(&mut central, 0);
write_u16(&mut central, 0);
write_u32(&mut central, 0);
write_u32(&mut central, offset);
central.extend_from_slice(name_bytes);
}
let central_offset = out.len() as u32;
let central_size = central.len() as u32;
out.extend_from_slice(&central);
write_u32(&mut out, 0x0605_4b50);
write_u16(&mut out, 0);
write_u16(&mut out, 0);
write_u16(&mut out, files.len() as u16);
write_u16(&mut out, files.len() as u16);
write_u32(&mut out, central_size);
write_u32(&mut out, central_offset);
write_u16(&mut out, 0);
fs::write(path, out).unwrap();
}
fn write_u16(out: &mut Vec<u8>, value: u16) {
out.extend_from_slice(&value.to_le_bytes());
}
fn write_u32(out: &mut Vec<u8>, value: u32) {
out.extend_from_slice(&value.to_le_bytes());
}
fn crc32(data: &[u8]) -> u32 {
let mut crc = 0xffff_ffffu32;
for &byte in data {
crc ^= byte as u32;
for _ in 0..8 {
let mask = if crc & 1 == 1 { 0xedb8_8320 } else { 0 };
crc = (crc >> 1) ^ mask;
}
}
!crc
}
}

View File

@ -151,9 +151,13 @@ pub(crate) async fn run(
#[cfg(feature = "e2e-test")]
{
if !emitted_panel_ready {
// `panel_ready` is a first-visible-frame signal only. E2E tests that need
// list/data readiness must wait for a concrete `rows_rendered` fixture row.
crate::e2e_observer::emit("panel", "panel_ready", serde_json::json!({}));
emitted_panel_ready = true;
}
// Emit every drawn row snapshot separately so tests can assert data-backed row
// readiness without conflating it with the first frame.
app.emit_rows_rendered();
}

View File

@ -40,7 +40,7 @@ rustPlatform.buildRustPackage rec {
filter = sourceFilter;
};
cargoHash = "sha256-Y1siH1oDe9It7ntx83DJO5fzV9LtC7+qq9V6RPlRxUY=";
cargoHash = "sha256-ud+3INcXnT5W26Bz0K4QXUqoqw3p/ER9c4F2Fhq3YuQ=";
depsExtraArgs = {
# Older fetchCargoVendor utilities used crates.io's API download endpoint,

View File

@ -267,12 +267,78 @@ pub struct RenderedPanelRow {
pub rect: PanelRect,
}
#[derive(Debug, Clone, PartialEq, Eq)]
pub struct ExpectedPanelTicketRow {
pub id: String,
pub title: String,
pub status: String,
}
impl ExpectedPanelTicketRow {
pub fn new(id: impl Into<String>, title: impl Into<String>, status: impl Into<String>) -> Self {
Self {
id: id.into(),
title: title.into(),
status: status.into(),
}
}
pub fn matches(&self, row: &RenderedPanelRow) -> bool {
row.key.kind == "ticket"
&& row.key.id == self.id
&& row.title == self.title
&& row.status.as_deref() == Some(self.status.as_str())
}
fn description(&self) -> String {
format!(
"ticket row id={} title={:?} status={}",
self.id, self.title, self.status
)
}
}
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct RowsRendered {
pub selected: Option<PanelRowKey>,
pub rows: Vec<RenderedPanelRow>,
}
impl RowsRendered {
pub fn fixture_ticket_row(
&self,
expected: &ExpectedPanelTicketRow,
) -> Option<&RenderedPanelRow> {
self.rows.iter().find(|row| expected.matches(row))
}
pub fn has_fixture_ticket_row(&self, expected: &ExpectedPanelTicketRow) -> bool {
self.fixture_ticket_row(expected).is_some()
}
}
fn rows_rendered_event_has_fixture_ticket(
event: &HarnessEvent,
expected: &ExpectedPanelTicketRow,
) -> bool {
serde_json::from_value::<RowsRendered>(event.data.clone())
.map(|rows| rows.has_fixture_ticket_row(expected))
.unwrap_or(false)
}
fn describe_rows(rows: &RowsRendered) -> String {
rows.rows
.iter()
.map(|row| {
format!(
"{}:{} title={:?} status={:?}",
row.key.kind, row.key.id, row.title, row.status
)
})
.collect::<Vec<_>>()
.join(", ")
}
#[derive(Debug, Clone)]
pub enum KeyPress {
CtrlC,
@ -445,6 +511,69 @@ impl PanelHarness {
}
}
/// Waits for the legacy `panel_ready` observer event, which means the first
/// panel frame was painted. It intentionally does not mean workspace rows or
/// Ticket data are ready.
pub fn wait_for_first_visible_frame(&mut self, timeout: Duration) -> Result<HarnessEvent> {
self.wait_for(
"first visible panel frame (panel_ready, before rows readiness)",
timeout,
|event| event.event == "panel_ready",
)
}
/// Waits until a concrete fixture Ticket row has rendered. This is the
/// startup rows-ready signal; it validates id + title + state rather than
/// using only the number of rendered rows.
pub fn wait_for_fixture_ticket_rows_ready(
&mut self,
expected: &ExpectedPanelTicketRow,
timeout: Duration,
) -> Result<RowsRendered> {
let description = expected.description();
let event = self.wait_for(
format!("fixture Ticket rows ready ({description})"),
timeout,
|event| {
event.event == "rows_rendered"
&& rows_rendered_event_has_fixture_ticket(event, expected)
},
)?;
serde_json::from_value(event.data).map_err(HarnessError::from)
}
pub fn assert_fixture_ticket_row_not_rendered(
&mut self,
expected: &ExpectedPanelTicketRow,
duration: Duration,
) -> Result<()> {
let start = Instant::now();
while start.elapsed() < duration {
if let Some(rows) = self
.events()?
.iter()
.filter(|event| event.event == "rows_rendered")
.filter_map(|event| serde_json::from_value::<RowsRendered>(event.data.clone()).ok())
.find(|rows| rows.has_fixture_ticket_row(expected))
{
self.flush_output_artifact()?;
return Err(HarnessError::Protocol(format!(
"fixture Ticket row rendered before data-backed rows readiness was expected: {}; rows: {}",
expected.description(),
describe_rows(&rows)
)));
}
if let Some(status) = self.child.try_wait()? {
self.flush_output_artifact()?;
return Err(HarnessError::Protocol(format!(
"process exited with {status} while asserting fixture Ticket rows stayed delayed"
)));
}
thread::sleep(Duration::from_millis(20));
}
Ok(())
}
pub fn wait_for_rows(&mut self, min_rows: usize) -> Result<RowsRendered> {
let event = self.wait_for("rows_rendered", DEFAULT_WAIT, |event| {
event.event == "rows_rendered"
@ -781,6 +910,9 @@ pub struct FixtureCleanupReport {
pub report_path: PathBuf,
}
pub const READY_FIXTURE_TICKET_TITLE: &str = "Ready E2E Ticket";
pub const PLANNING_FIXTURE_TICKET_TITLE: &str = "Planning E2E Ticket";
#[derive(Debug)]
pub struct FixtureWorkspace {
temp_root: Option<TempDir>,
@ -792,6 +924,8 @@ pub struct FixtureWorkspace {
pub xdg_config_home: PathBuf,
pub xdg_runtime_dir: PathBuf,
pub artifacts_dir: PathBuf,
pub ready_ticket_id: String,
pub planning_ticket_id: String,
}
impl FixtureWorkspace {
@ -832,7 +966,7 @@ impl FixtureWorkspace {
fs::create_dir_all(dir)?;
}
let fixture = Self {
let mut fixture = Self {
temp_root: Some(temp_root),
root,
workspace,
@ -842,6 +976,8 @@ impl FixtureWorkspace {
xdg_config_home,
xdg_runtime_dir,
artifacts_dir,
ready_ticket_id: String::new(),
planning_ticket_id: String::new(),
};
fixture.write_fixture_metadata("created", None)?;
@ -867,7 +1003,7 @@ impl FixtureWorkspace {
&fixture.xdg_config_home,
&fixture.xdg_runtime_dir,
&fixture.artifacts_dir,
"Ready E2E Ticket",
READY_FIXTURE_TICKET_TITLE,
)?;
run_yoi(
binary,
@ -880,7 +1016,7 @@ impl FixtureWorkspace {
&fixture.artifacts_dir,
&["ticket", "state", &first, "ready"],
)?;
let _second = create_ticket(
let second = create_ticket(
binary,
&fixture.workspace,
&fixture.home,
@ -889,12 +1025,22 @@ impl FixtureWorkspace {
&fixture.xdg_config_home,
&fixture.xdg_runtime_dir,
&fixture.artifacts_dir,
"Planning E2E Ticket",
PLANNING_FIXTURE_TICKET_TITLE,
)?;
fixture.ready_ticket_id = first;
fixture.planning_ticket_id = second;
fixture.write_fixture_metadata("ready", None)?;
Ok(fixture)
}
pub fn ready_fixture_ticket_row(&self) -> ExpectedPanelTicketRow {
ExpectedPanelTicketRow::new(
self.ready_ticket_id.clone(),
READY_FIXTURE_TICKET_TITLE,
"ready",
)
}
pub fn panel_config(&self, binary: PathBuf) -> PanelHarnessConfig {
PanelHarnessConfig {
binary,
@ -999,6 +1145,18 @@ impl FixtureWorkspace {
"xdg_config_home": &self.xdg_config_home,
"xdg_runtime_dir": &self.xdg_runtime_dir,
"artifacts_dir": &self.artifacts_dir,
"tickets": {
"ready": {
"id": &self.ready_ticket_id,
"title": READY_FIXTURE_TICKET_TITLE,
"state": "ready"
},
"planning": {
"id": &self.planning_ticket_id,
"title": PLANNING_FIXTURE_TICKET_TITLE,
"state": "planning"
}
},
"env_runtime_policy": {
"tested_yoi_uses_env_clear": true,
"host_runtime_inherited": false,

View File

@ -1,17 +1,62 @@
use std::time::{Duration, Instant};
const FIRST_VISIBLE_RENDER_BUDGET: Duration = Duration::from_millis(1500);
const FULL_READY_BUDGET: Duration = Duration::from_secs(5);
const ROWS_READY_BUDGET: Duration = Duration::from_secs(5);
use yoi_e2e::{
FixtureCleanupReport, FixtureWorkspace, KeyPress, PanelHarness, RenderedPanelRow, yoi_binary,
ExpectedPanelTicketRow, FixtureCleanupReport, FixtureWorkspace, KeyPress, PanelHarness,
PanelRect, PanelRowKey, RenderedPanelRow, RowsRendered, yoi_binary,
};
#[test]
fn panel_fixture_ticket_row_matcher_rejects_absent_fixture_data() {
let expected = ExpectedPanelTicketRow::new("0000000000000", "Ready E2E Ticket", "ready");
let wrong_title = RenderedPanelRow {
key: PanelRowKey {
kind: "ticket".to_string(),
id: "0000000000000".to_string(),
},
title: "Different Ticket".to_string(),
status: Some("ready".to_string()),
action: None,
rect: PanelRect {
x: 0,
y: 0,
width: 10,
height: 1,
},
};
let wrong_kind = RenderedPanelRow {
key: PanelRowKey {
kind: "pod".to_string(),
id: "0000000000000".to_string(),
},
title: "Ready E2E Ticket".to_string(),
status: Some("ready".to_string()),
action: None,
rect: PanelRect {
x: 0,
y: 0,
width: 10,
height: 1,
},
};
assert!(!expected.matches(&wrong_title));
assert!(!expected.matches(&wrong_kind));
let rows = RowsRendered {
selected: None,
rows: vec![wrong_title, wrong_kind],
};
assert!(!rows.has_fixture_ticket_row(&expected));
}
#[test]
fn panel_first_visible_render_arrives_before_background_reload() -> yoi_e2e::Result<()> {
let binary = yoi_binary()?;
let fixture = FixtureWorkspace::new(&binary)?;
assert_fixture_paths_are_isolated(&fixture);
let ready_ticket = fixture.ready_fixture_ticket_row();
let started = Instant::now();
let mut panel =
@ -19,17 +64,15 @@ fn panel_first_visible_render_arrives_before_background_reload() -> yoi_e2e::Res
let remaining = FIRST_VISIBLE_RENDER_BUDGET
.checked_sub(started.elapsed())
.unwrap_or_else(|| Duration::from_millis(0));
panel.wait_for("first visible panel render", remaining, |event| {
event.event == "panel_ready"
})?;
panel.wait_for_first_visible_frame(remaining)?;
let first_visible_elapsed = started.elapsed();
eprintln!(
"panel first visible render: {first_visible_elapsed:?} (budget {FIRST_VISIBLE_RENDER_BUDGET:?}); artifacts at {}",
"panel first visible frame: {first_visible_elapsed:?} (budget {FIRST_VISIBLE_RENDER_BUDGET:?}); artifacts at {}",
panel.artifacts().dir.display()
);
assert!(
first_visible_elapsed <= FIRST_VISIBLE_RENDER_BUDGET,
"first visible render took {first_visible_elapsed:?}, budget {FIRST_VISIBLE_RENDER_BUDGET:?}; artifacts at {}",
"first visible frame took {first_visible_elapsed:?}, budget {FIRST_VISIBLE_RENDER_BUDGET:?}; artifacts at {}",
panel.artifacts().dir.display()
);
@ -42,7 +85,7 @@ fn panel_first_visible_render_arrives_before_background_reload() -> yoi_e2e::Res
events[..ready_index]
.iter()
.all(|event| event.event != "background_task_started"),
"initial render must be emitted before reload/background work starts; artifacts at {}",
"initial frame must be emitted before reload/background work starts; artifacts at {}",
panel.artifacts().dir.display()
);
@ -54,12 +97,13 @@ fn panel_first_visible_render_arrives_before_background_reload() -> yoi_e2e::Res
event.event == "background_task_started"
&& event.data.get("task").and_then(serde_json::Value::as_str) == Some("reload")
})
.expect("held reload should start after first visible render");
.expect("held reload should start after first visible frame");
assert!(
ready_index < reload_started_index,
"first visible render and reload ordering should remain separate; artifacts at {}",
"first visible frame and reload ordering should remain separate; artifacts at {}",
panel.artifacts().dir.display()
);
panel.assert_fixture_ticket_row_not_rendered(&ready_ticket, Duration::from_millis(150))?;
panel.press(KeyPress::CtrlC)?;
let status = panel.expect_exit_within(PanelHarness::default_exit_wait())?;
@ -70,51 +114,46 @@ fn panel_first_visible_render_arrives_before_background_reload() -> yoi_e2e::Res
}
#[test]
fn panel_full_ready_has_separate_startup_budget() -> yoi_e2e::Result<()> {
fn panel_fixture_ticket_row_ready_has_startup_budget() -> yoi_e2e::Result<()> {
let binary = yoi_binary()?;
let fixture = FixtureWorkspace::new(&binary)?;
assert_fixture_paths_are_isolated(&fixture);
let ready_ticket = fixture.ready_fixture_ticket_row();
let started = Instant::now();
let mut panel = PanelHarness::spawn(fixture.panel_config(binary))?;
let first_visible_remaining = FIRST_VISIBLE_RENDER_BUDGET
.checked_sub(started.elapsed())
.unwrap_or_else(|| Duration::from_millis(0));
panel.wait_for(
"first visible panel render",
first_visible_remaining,
|event| event.event == "panel_ready",
)?;
panel.wait_for_first_visible_frame(first_visible_remaining)?;
let first_visible_elapsed = started.elapsed();
eprintln!(
"panel first visible render: {first_visible_elapsed:?} (budget {FIRST_VISIBLE_RENDER_BUDGET:?}); artifacts at {}",
"panel first visible frame: {first_visible_elapsed:?} (budget {FIRST_VISIBLE_RENDER_BUDGET:?}); artifacts at {}",
panel.artifacts().dir.display()
);
assert!(
first_visible_elapsed <= FIRST_VISIBLE_RENDER_BUDGET,
"first visible render took {first_visible_elapsed:?}, budget {FIRST_VISIBLE_RENDER_BUDGET:?}; artifacts at {}",
"first visible frame took {first_visible_elapsed:?}, budget {FIRST_VISIBLE_RENDER_BUDGET:?}; artifacts at {}",
panel.artifacts().dir.display()
);
let full_ready_remaining = FULL_READY_BUDGET
let rows_ready_remaining = ROWS_READY_BUDGET
.checked_sub(started.elapsed())
.unwrap_or_else(|| Duration::from_millis(0));
panel.wait_for("full ready fixture rows", full_ready_remaining, |event| {
event.event == "rows_rendered"
&& event
.data
.get("rows")
.and_then(serde_json::Value::as_array)
.is_some_and(|rows| rows.len() >= 2)
})?;
let full_ready_elapsed = started.elapsed();
let rows = panel.wait_for_fixture_ticket_rows_ready(&ready_ticket, rows_ready_remaining)?;
assert!(
rows.has_fixture_ticket_row(&ready_ticket),
"rows-ready event must contain concrete ready fixture Ticket row; artifacts at {}",
panel.artifacts().dir.display()
);
let rows_ready_elapsed = started.elapsed();
eprintln!(
"panel full ready: {full_ready_elapsed:?} (budget {FULL_READY_BUDGET:?}); artifacts at {}",
"panel fixture rows ready: {rows_ready_elapsed:?} (budget {ROWS_READY_BUDGET:?}); artifacts at {}",
panel.artifacts().dir.display()
);
assert!(
full_ready_elapsed <= FULL_READY_BUDGET,
"full ready took {full_ready_elapsed:?}, budget {FULL_READY_BUDGET:?}; artifacts at {}",
rows_ready_elapsed <= ROWS_READY_BUDGET,
"fixture rows ready took {rows_ready_elapsed:?}, budget {ROWS_READY_BUDGET:?}; artifacts at {}",
panel.artifacts().dir.display()
);