mirror of
https://github.com/software-mansion/argent.git
synced 2026-09-14 19:27:14 +08:00
fix/docs-high-severity
7 Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
 Hubert Gancarczyk
|
afd722924f |
feat(flow): support concurrent flow recordings (#574)
## Problem
Only one flow could be recorded at a time, anywhere on the machine.
Recording state was three module globals in `flow-utils.ts` — harmless
in a per-client process, but the tool-server is a **host-wide singleton
per install bundle**. Every MCP client, every subagent sharing a
parent's connection, every `argent` CLI call and every project on the
machine attaches to the same process, the same `Registry`, and those
same three globals. HTTP requests are served concurrently with no
serialization.
Three failures followed, all reproduced against the pre-change code:
1. **A second recording clobbered the first.** `startRecordingSession`
just overwrote, so the first agent's next `flow-add-step` silently
appended into the *other* agent's flow file — in a different project.
2. **Any agent could finish any agent's recording.**
`flow-finish-recording`'s schema was literally `z.object({})`.
3. **Replay stomped recording.** `resolveFlowFilePath` called
`setActiveProjectRoot`, so a `flow-execute` in project B rebound the
global root mid-recording of project A. Already worked around, locally
only, in `flow-add-step`.
## Approach
Recording state is now a `Map` keyed by the resolved flow file path,
`<project_root>/.argent/flows/<name>.yaml` — the identity of the
artifact being built, so "same key" means "same output file" and two
sessions on one key are a genuine collision rather than an accident of
scoping. `flow-add-step`, `flow-add-echo` and `flow-finish-recording`
each take `name` + `project_root`, so every call is self-contained;
required rather than optional-with-fallback, since a fallback would
succeed while an agent is alone and fail only once a second recording
exists.
`resolveFlowFilePath` is pure, which fixes defect 3 and retires
`FLOW_PROJECT_ROOT_REQUIRED` (the state it guarded can no longer exist).
Rejected alternatives — keying by device udid (flows are deliberately
device-portable, and `flow-add-echo`/`flow-finish-recording` have no
device at all), an opaque `recordingId` (an agent that loses the token
cannot recover), and per-caller identity from the transport
(`ToolContext` carries no caller identity, and subagents share their
parent's MCP connection, so it would fail the exact scenario being
fixed).
### Locking, and why readers needed more than a lock
Every mutation of one flow file runs under a lock keyed by that path —
appends, `flow-start-recording`'s truncate-and-register, and
`flow-finish-recording`'s read-summarize-clear. Keyed by file rather
than owned by the session, because a restart *replaces* the session: a
session-owned lock cannot exclude the operation that supersedes it. Per
file, not global, so two recordings never queue behind each other. An
append whose session was finished, restarted or evicted meanwhile fails
with `FLOW_NO_ACTIVE_RECORDING` rather than writing into a different
take.
The lock serializes writers only, and no reader of a flow YAML can join
it — `flow-execute`'s own load, its `run:` fragment load,
`flow-read-prerequisite`, `flow-add-step`'s sibling-fragment check, and
the `argent` CLI reading from another process entirely, where an
in-process lock cannot reach. Both writers used a plain `fs.writeFile`,
which opens `O_TRUNC`, so a reader could land between the truncate and
the write. That case is silent: `parseFlow("")` returns `{ steps: [] }`
with no error and `summarize` derives `ok` from "no failures", so a
`flow-execute` racing an append reports a **top-level PASS having
replayed zero steps**. Both writes now go through a sibling temp file
and a rename, so a reader sees either the whole old file or the whole
new one — and unlike the lock, that also holds cross-process.
### `stop-all-simulator-servers` gets a `devices` scope
Every agent is told to call this at session end, and unscoped it walked
the whole registry — so agent A wrapping up tore down agent B's devtools
mid-recording, degrading B's flow to brittle coordinate taps. It now
takes an optional `devices` scope and reports `unmatched` ids, so a
mistyped id doesn't read as a clean machine.
Which namespaces count as "owned by a device" turned out to be the
load-bearing part, because nothing cascades to most of them and
`unmatched` turns any omission into an actively wrong diagnosis — a
correct device id reported as a typo. The set now covers `AXService`
(the in-sim ax daemon, spawned `--timeout 3600`; an iOS session that
only ran boot/launch/describe owns this and nothing else),
`ScreenRecordingSession` (an ffmpeg child plus the touch-visualizer
overlay it enabled on the device), `NativeProfilerSession` (an xctrace
child, or an on-device perfetto process and its trace file) and
`JsRuntimeDebugger` (a bound loopback server, the CDP socket to Metro, a
log handle). The debugger URNs interpose the Metro port —
`<ns>:<port>:<deviceId>` — so the matcher understands both shapes,
consuming only the first colon so a wireless adb serial after the port
still compares whole.
`stop-simulator-server` shared none of this and had drifted: it looked
URNs up with an exact, case-sensitive `services.get()`, so a lower-cased
UDID silently no-op'd there while the scoped stop-all reaped it. Both
tools now share one matcher. Their namespace *sets* stay deliberately
different, documented where they are defined — `stop-simulator-server`
is also the documented recovery for a wedged transport, and widening it
to devtools/AX would make a routine retry drop the native-devtools
connection another agent's recording depends on, which is the hazard the
`devices` scope exists to prevent.
## Verification
**End-to-end, three devices, one shared tool-server** (iPhone 16 Pro,
iPhone 17 Pro, Pixel 3a): three interleaved recordings, plus an
unrelated `flow-execute` in a fourth project while all three were live.
Both iOS devices held their own live native-devtools concurrently (two
per-UDID sockets, `connected: true`). Each YAML contained exactly its
own steps in order, taps captured as selectors rather than coordinates,
and neither picked up another's or the third root's path. All three
replayed green — also through the **released 0.17.0 CLI** against the
modified server, which checks wire compatibility. A device-scoped
teardown of sim 1 left sim 2 and the emulator running.
Re-run against the final build:
- Two interleaved recordings across two project roots on one device,
with a **separate process** polling one of the flow files as fast as it
could for the duration. Zero empty or truncated observations; each YAML
held exactly its own steps; both replayed green (6/6 and 5/5). Against
the pre-change in-place write the same reader observes zero-step reads
interleaved among the real ones — the silent-PASS input.
- A scoped teardown of the iPhone reaped `AXService` and
`NativeDevtools` for that UDID only, reported no `unmatched`, and left
the second sim and the emulator running. Lower-casing the UDID matched;
a genuinely bogus id was still reported.
- With a screen recording live on the Pixel 3a, a scoped teardown reaped
`ScreenRecordingSession` and the **ffmpeg child actually exited** (1 →
0). Before the fix it survived to the 180 s cap while the tool called
the correct serial a typo.
The pre-change repros were re-run afterwards to confirm observed
behavior changed, not just the tests.
`npx vitest run` in `packages/tool-server` (295 files, 3154 tests) and
`packages/argent-cli` (280), `tsc --build`, `tsc --noEmit` on both test
projects, eslint and prettier — all green.
## Review notes
Several rounds of review found real defects in this change; each was
reproduced before fixing and re-verified after. Worth knowing when
reading the diff, since the tests exist to pin them:
- `flow-start-recording` truncated the `.yaml` outside any lock, so a
step from the take being discarded landed in the freshly reset file and
was reported as success (26-29 of 200 runs).
- `flow-finish-recording` never took the lock; its `await fs.readFile`
was a yield during which a concurrent append committed, leaving the
reported summary disagreeing with disk (11-16 of 200).
- A failed finish destroyed the recording. Hand-editing the `.yaml`
mid-recording is a documented workflow, so `parseFlow` can legitimately
throw — and the error told the agent to call `flow-start-recording`,
which truncates the very take it would recover.
- `unmatched` reported a device this session had just stopped, because
`disposeService` returns a node to IDLE and keeps it.
- A bare IP claimed every wireless-adb device at that address, since an
adb serial is itself `ip:port`.
- The eviction backstop stamped with `Date.now()`; ties at millisecond
resolution let it drop the session that was just used.
- `requireRecordingSession` ended with "Call flow-start-recording first"
— reached for a key that was finished, superseded or evicted as well as
one never started, and on those branches the file on disk is fully
populated while that call truncates it and reports no `restarted`.
- The atomic write regressed twice on its own terms: its scratch name
was derived from the flow file's basename, which has no length cap, so a
long flow name that appended fine before failed `ENAMETOOLONG`; and the
cleanup guard began after the temp file was created, leaking a scratch
file into a committed directory on any write that failed after opening.
- Two tests passed against implementations they claimed to reject: the
`stop-simulator-server` narrowness guard used a udid that classifies as
*android*, where the iOS-only services it guards can never appear, so
widening the iOS branch kept it green; and the client-mode finish
asserted only the *shape* of `savedTo`, never its content, though that
directive is the only thing that lands the file in client mode.
## Notes for reviewers
- Client/remote persist mode is the least-exercised path, and now has
its own concurrency tests (overlapping appends, and a restart
superseding an in-flight one) — it previously had none. The
`project_root` key is byte-stable across the file boundary (`kind:
"probe"` passes it through unchanged), which is what keeps a remote
recording addressable from `flow-add-step`, which declares no file
input.
- Two spellings of one flow file still mint two sessions — and two
independent locks — over one file, which bypasses every guarantee here.
Two ways in: the root spelled two ways (a symlink, or a case variant on
APFS), and the likelier one, the *name* cased two ways on a
case-insensitive volume. Neither is normalized away because the correct
normalization is the filesystem's: case-folding the key would wrongly
merge two distinct flows on ext4, and resolving symlinks is impossible
in client mode, where the root does not exist on this host. Documented
at `getFlowPath`.
- No recording tool can detect that another agent took its key.
`flow-add-step`, `flow-add-echo` and `flow-finish-recording` all
re-resolve the key on every call, so after a takeover they act on
whichever take is now live and report success: an append lands in the
other agent's file, and a finish finishes and clears it. The only case
that fails loudly is the narrow one the liveness check covers - an
append already IN FLIGHT when the restart lands. There is no caller
identity to key ownership on (see the rejected alternatives above), so
this is spelled out for the agent under "Pick a name unique to your
task" in `argent-create-flow` rather than mechanized.
- Skew hazard in one direction only: a *new* client against an *older*
tool-server. Zod strips the undeclared `name`/`project_root`, and the
old server falls back to its module globals — so concurrent recordings
silently collapse onto one session. The other direction (released 0.17.0
CLI against this server) was verified end-to-end.
- `packages/argent-private/docs/reference.md` still describes
`flow-add-step` as appending to "the current recording", and names two
tool ids that do not exist (`flow-insert-echo`/`flow-run` are file
names; the ids are `flow-add-echo`/`flow-execute`). It is a submodule,
so the fix cannot ride on this branch.
- A restart reported `discardedSteps` from the superseded session's
in-memory flow while it truncates the *file*. Hand-editing the `.yaml`
mid-recording is documented, so the two diverge: four steps wiped, one
reported. It now counts the file, and reports no number at all when that
file cannot be read or parsed - 0 is the answer a genuinely empty take
gives. Client mode still counts from memory (this host cannot see the
client's file), and the description now says which mode it is promising.
- `ChromiumJsRuntimeDebugger` cascades from `ChromiumCdp` and was
therefore already being torn down - silently, while `NetworkInspector`
and `ReactProfilerSession` cascade the same way and were named.
`stopped` is documented as the services that were live and got shut
down, so it is listed now too.
- The client-mode case named for a superseded in-flight append restarted
*between* calls, so the next append re-resolved the key and succeeded -
the guard was never reached, and neutering it left the whole file green.
It now parks an append in its live sub-tool call across the restart.
- `stop-tools.test.ts`' mock handed the tool the live service map where
the real `getSnapshot` copies, so a cascade rewrote the state the sweep
was still reading and the answer depended on map insertion order. The
mock copies and recurses like `_teardown` now, and the cascade case runs
under both orders.
- Prose the code contradicted, each verified against source before
rewriting: Vega owns `JsRuntimeDebugger`/`NetworkInspector` once the
debugger has run (and can match anything at all through an ERROR node);
"nothing here cascades" was denied twice in its own file; an iOS
boot/launch/describe session owns `NativeDevtools` too; `:tcp` is
mintable but unreached in production; the tool-server is a singleton
*per install bundle*, not per machine; `argent flow list` does enumerate
`.argent/flows`; the stop-tools comments narrated a "before" that only
existed on this branch.
Found while verifying, out of scope, worth their own issues: a `tool:
flow-execute` self-reference recurses past both cycle guards (`run:` is
guarded, a nested `flow-execute` reseeds the stack); react-profiler
artifacts collide host-wide on a second-resolution filename;
`stop-metro` is port-scoped and kills another agent's bundler;
`debugger-reload-metro`'s HTTP fallback broadcasts to every attached
client. A tool's `capability` is enforced only by the HTTP layer, not by
`registry.invokeTool`, so a call that reaches the registry another way -
`flow-add-step` takes `command` as a bare string - can resolve a service
the tool declares it does not support, leaving an ERROR node behind
(e.g. `SimulatorServer:<vega serial>`).
---------
Co-authored-by: Hubert Gancarczyk <claude-hubert.gancarczyk@swmansion.com>
|
||
|
Ignacy Łątka
|
a5c6730d05 |
feat(ios-profiler): attributable memory leaks via malloc_stack_logging (#351)
## What Opt-in `malloc_stack_logging` flag on `native-profiler-start` that makes iOS memory leaks **attributable** — they come back with a real responsible frame + library instead of `<Call stack limit reached>`. ## Why `native-profiler-start` attaches to the already-running app. iOS records malloc allocation backtraces only when `MallocStackLogging` is set **at process launch**, so an attached app has none: Instruments' Leaks scanner finds the leaked blocks but can't attribute them and emits the placeholder `<Call stack limit reached>` (that string is Instruments', not ours). ## How When `malloc_stack_logging: true`, the profiler cold-launches the app **under** xctrace instead of attaching: ``` xctrace record --template <Argent.tracetemplate> --device <udid> \ --env MallocStackLogging=1 --output <…> --no-prompt \ --launch -- <App.app> ``` - `.app` path resolved via `simctl get_app_container`; the running instance is terminated first for a clean cold start. - The existing Leaks export / parser / render path is reused **unchanged** — no parser changes. - Default behaviour is untouched: attach to the running app (no relaunch, no overhead). The flag is opt-in because it restarts the app (loses state) and adds allocator overhead — for leak-attribution passes, not CPU/hang work. - The report relabels unattributable leaks with a hint to re-run with `malloc_stack_logging: true`, instead of surfacing the raw placeholder. ## Evidence (real iPhone 16 sim, same scroll workload) Before (attach) — every leak row: ``` responsible-frame="<Call stack limit reached>" responsible-library="" ``` After (`--env MallocStackLogging=1 --launch`) — same Leaks export: | leaked-object | responsible-frame | library | |---|---|---| | Malloc 1008 B ×3 | `itanium_demangle::OutputBuffer::grow(...)` | libc++abi | | Malloc 48 B ×1 | `hermes::vm::JSTypedArrayBase::createBuffer(...)` | hermes | ## Tests - `test/ios-instruments/malloc-stack-logging.test.ts` — the malloc launch argv (`--env MallocStackLogging=1`, `--launch -- <app>` last, with and without `--notify-tracing-started`, `--no-prompt`) vs default `--attach`; terminate-first / get_app_container ordering; failed-start behavior; per-test `ARGENT_IOS_CAPTURE` scrubbing. - `test/ios-instruments/leak-attribution-render.test.ts` — unattributable vs attributed leak rendering, capture-mode-aware hints. - `test/ios-instruments/load-freshness.test.ts` — per-capture residue clearing on `profiler-load`, live/recovery-pending load refusals. - `test/ios-instruments/stop-recovery.test.ts` — capture-mode pairing at stop (clean and recovery paths). - `test/profiler-combined-leaks.test.ts`, `test/profiler-leak-stacks.test.ts`, `test/ios-leak-pipeline.test.ts` — shared caveat line, evidence-first hints, attribution-normalized grouping. - Full `test/ios-instruments/` suite green (9 files / 70 tests); full tool-server suite green; `tsc --noEmit` + `tsc -p tsconfig.test.json` clean. ## Notes E2E'd live twice during review (real iPhone 16 Pro / iPhone 17 Pro sims on Xcode 26.6): guard refusals fire before the app is touched, the `--device` cold launch's deadlock on degraded Xcode is confirmed (so the guard is warranted), and the malloc argv/terminate/relaunch behavior checked out on-device. Docs updated: `IOS_PROFILER_REFERENCE.md` and the `argent-native-profiler` skill. ## Review follow-ups (post-approval hardening) Later commits address the remaining review threads and several subsequent multi-lens review passes: - **Capture-mode is a real signal, end to end.** The session records the mode of the in-flight recording at start (`recordingMallocStackLogging`), pairs it with `exportedFiles` at stop (`mallocStackLogging`), and freezes it into `parsedData` at analyze/load — so the analyze report, the combined report, and the `leak_stacks` drill-down all name the capture mode of the data they actually render instead of inferring it (or mislabeling it after a newer capture starts). One shared `renderUnattributedLeaksNote()` keeps the wording identical across reports, and attribution evidence outranks the flag: a resolved frame proves the target ran under malloc stack logging however it was launched. - **`profiler-load` restores a clean slate.** Loading clears all per-capture residue (capture mode, `cpuFilterPid`, `wallClockStartMs`, `traceFile`) so a loaded trace is never analyzed with a previous live capture's PID filter, freshness anchor, or report path — and refuses to load while a recording is in flight or a crashed capture awaits `native-profiler-stop`'s partial-trace recovery, which the load would otherwise orphan. - **Failed starts are non-destructive.** Per-capture descriptors — and the recovery flags that gate `native-profiler-stop`'s partial-trace export — are stamped/reset only on a successful start, so a failed attempt leaves the previous capture's still-loaded exports fully described and a pending abnormal-end recovery intact. - **Guards keep session state coherent.** `profiler-load`, `native-profiler-analyze`, and `profiler-combined-report` all refuse while a recording is in flight or a crashed capture awaits recovery, via one shared `isCaptureInFlight`/`inFlightGuardMessage` helper whose message distinguishes recording vs 10-min-cap timeout vs unexpected exit (so the stated cause matches what `native-profiler-stop` reports next). Proceeding would otherwise wedge the session (load) or render one capture's frozen data under another's live fields (analyze/combined). - **The combined report anchors on a frozen clock.** The recording's start time is frozen into `parsedData` at analyze, and the iOS combined report anchors its hangs to that frozen value rather than the live session field — so a capture started after analyze can never shift a real hang↔commit correlation into "unmatched". Android keeps its live anchor (it re-derives hangs from the live trace). The combined-report guard's recovery advice routes through `native-profiler-analyze`, the only step that rebuilds `parsedData`. - **Leak grouping is attribution-normalized.** `aggregateLeaks` keys on `(objectType, frame-normalized-the-way-isLeakAttributed-matches)`, so mixed unattributed spellings (`Unknown`, the sentinel, whitespace variants) can't split one group per type. - **The malloc launch target is unambiguous.** Since the cold launch terminates the resolved app, `app_process` resolution prefers an exact `CFBundleExecutable` match and, when builds still tie (dev + prod sharing an executable/display name), refuses and points at the globally-unique `CFBundleIdentifier` — which it also accepts as an `app_process` — instead of killing whichever app the plist enumerates first. - **Report tables escape `|`.** Demangled C++ frames (`operator|` included) are real headline content under malloc mode; every free-form cell across the leak, CPU-hotspot, function caller/callee, and thread-breakdown tables — symbol *and* thread/queue-name columns — is escaped so a stray pipe can't shift the row's columns. - **Android start is non-destructive too.** `startNativeProfilerAndroid` now starts perfetto before mutating session state, mirroring iOS, so a failed start no longer burns a prior capped capture's pending partial-trace recovery. |
||
|
filip131311
|
c8f6b4baf7 |
feat(ios-profiler): pluggable capture strategy with --all-processes fallback for the xctrace --device deadlock (#380)
## What
Encapsulates the iOS native CPU capture behind an `IosCaptureStrategy`
interface with **two interchangeable implementations**, selected per
environment. This is a drop-in refactor: the default path is
byte-identical to today's behaviour; a new fallback path is used only on
broken Xcode versions.
## Why
On **Xcode 26.4–27.0**, `xctrace record --device <sim>` **deadlocks at
the recording-start handshake** (a host-side xctrace regression: the
in-sim `DTServiceHub` waits forever for a `recording-started` reply the
regressed xctrace never sends). Capture never starts, so there's nothing
to recover — `native-profiler-start` just hangs until the cap. Older
xctrace (≤ 26.3) works fine, which is how we know the bug is in
xctrace's host side, not the simulator.
The only hang-free, Apple-entitled path that still profiles a simulator
app is to record the **host** with `--all-processes` (no `--device`).
Simulator apps run as ordinary host processes on the shared host kernel,
so host-wide kperf sampling captures the target app's threads too — at
the same ~1 kHz PET rate, with fully symbolicated user callstacks — and
the host path finalizes cleanly.
## How
- **`utils/ios-profiler/capture-strategy/`** (new): `IosCaptureStrategy`
interface + two impls:
- **`device`** (default): `xctrace --device <sim> --attach <pid|name>` —
record args are identical to the current code, so healthy environments
are unaffected. No post-filter (already scoped).
- **`all-processes`** (fallback): `xctrace --all-processes` (host-wide),
then filter the export to the app PID.
- **`select.ts`**: picks `device` vs `all-processes` from the active
Xcode version (26.4–27.0 ⇒ fallback), with an
`ARGENT_IOS_CAPTURE=device|all-processes` env override to force either
direction. So the new path runs **only on degraded environments** unless
explicitly overridden.
- **`platforms/ios.ts`**: builds the xctrace argv via
`strategy.buildRecordArgs(...)`, stores `strategy.cpuFilterPid(target)`
on the session, gates the cold-start retry on `strategy.attachesByName`,
and guards that the all-processes path has a running app PID to scope
by.
- **Pipeline**: `runIosProfilerPipeline(files, { cpuFilterPid })` →
`parseCpuFile/parseCpuXml(xml, targetPid)`. When a filter PID is set,
`time-profile` rows are kept only for that PID using the export's
`<process>`/`<pid>` id/ref table, preserving the existing two-pass
row↔sample alignment. No filtering when unset (device path), so existing
behaviour is unchanged.
Lifecycle (startup readiness, stop/finalize, export) and all analyses
are shared between strategies.
## Fidelity
`--all-processes` data fully supports the analyses the tool-server
provides — **hotspots, native call trees, per-thread breakdown, app
call-chains, hangs** (verified against the export schema and Argent's
parser). The only thing absent is **kernel callstacks**
(`cp-kernel-callstack` is empty for simulator processes) — **no Argent
analysis consumes them, and the `--device` path lacks them for
simulators too, so there is no regression.**
## Caveats / notes
- All-processes capture is system-wide; samples are filtered to the app
PID post-export. Trace files are larger and export takes a few extra
seconds.
- The target app must be on-CPU during capture to be sampled (true of
any sampling profiler).
- The Xcode-version range in `select.ts` is the currently-known-broken
set; revisit when Apple ships a fix (or just rely on the env override).
## Testing
- New `capture-strategy` module typechecks clean in isolation;
full-package `tsc` shows no errors in any changed file.
- Verified the `device` strategy emits the exact current `record` argv,
and `--all-processes` host capture finalizes hang-free with per-sample
symbolicated user stacks filterable to one PID.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
|
||
|
filip131311
|
c01f41cd5f |
feat(remote): input-side file boundary — flows, screenshot-diff, profiler reports, workspace tools (#319)
## Why PR #278 fixed the **output** half of the remote file boundary: files a tool *produces* travel to the client as `ArtifactHandle`s. But seven tools still break when the tool-server runs on a different machine than the agent, in one (or both) of two ways: - **Output-side leftovers** — `native-profiler-analyze` and `screenshot-diff` still returned raw host paths (and embedded them in report/summary prose), so the agent can't open them remotely. - **Input-side gap** — nothing moved a file the *agent* owns onto the tool-server. Any tool taking a client-local path as a parameter (`screenshot-diff` baselines, the three flow tools' `project_root`, `react-profiler-component-source`, `gather-workspace-data`) silently read/wrote the wrong machine's disk. Flows were fully broken in remote mode: the YAML landed on the server's disk under the agent's path. ## What ### Output side (apply the existing pattern) - **native-profiler-analyze** — `reportFile` is now an `ArtifactHandle` (mirrors `react-profiler-analyze`'s `fileArtifact()`); the inline report now says *"use the Read tool on the `reportFile` path in this result"* instead of embedding the server path. - **screenshot-diff** — `diffPath` / `contextDiffPath` are now `ArtifactHandle`s; the MCP adapter renders the context-diff image from materialized bytes (legacy string-path results from older servers still render co-located). The summary references the result fields instead of embedding host paths. ### Input side (new, symmetric mechanism) A declarative **file-input boundary**, mirroring the artifact gate in the opposite direction: - A tool declares `fileInputs: [{ target, path: "${param}…", kind }]` on its `ToolDefinition`; the declaration ships to the client via `GET /tools`. - Before sending a call, the client (`@argent/tools-client`, shared by CLI and MCP) interpolates each template from the args and replaces the target arg with a wrapper: `{ __argentFileInput, path, size, mtimeMs, content? }`. File bytes are base64-inlined **only when routed to an external tool-server** (`argent link` / `ARGENT_TOOLS_URL`) — unlinked local calls never pay for encoding. - The server resolves wrappers **before zod validation**, so tools always execute against a plain local path: - path matches on the server's own filesystem (existence for dirs, size+mtime for files) → used **in place**, zero copies — the exact dual of the artifact materializer's co-located gate; - otherwise `kind: "file"` content is materialized into a temp file (with size-integrity check, 32 MB cap); - `kind: "directory"` (trees that can't ride in a call) → 422 with actionable remote-mode guidance; - `kind: "probe"` → never fails; the tool learns host-presence via `ctx.fileInputs` and adapts. - The reverse channel: a tool whose output belongs in the **agent's project** returns a `__argentClientFile` directive (path + content); the client writes it and rewrites the directive to the path. Writes are constrained to `**/.argent/flows/<safe-name>.yaml` so a tool-server can't direct writes anywhere else on the client machine. ### Tool migrations | Tool | Boundary use | |---|---| | `flow-start-recording` | `probe` on `project_root`: present → host persistence exactly as before; absent (remote) → recording kept in server memory, YAML returned as a client-write directive | | `flow-add-step` / `flow-add-echo` / `flow-finish-recording` | follow the session's persist mode; every mutation returns `savedTo` (path locally, directive remotely) so the agent-side file is always current | | `flow-execute` / `flow-read-prerequisite` | `file` input derived from `${project_root}/.argent/flows/${name}.yaml` into an internal `flow_file` param — read in place locally, uploaded when remote | | `screenshot-diff` | `baselinePath`/`currentPath` as `file` inputs; `outputDir` now **optional** (temp default) and probe-ignored when absent on the host | | `react-profiler-component-source`, `gather-workspace-data` | `directory` gate — unchanged locally; clear error remotely instead of a silent empty AST index / all-nulls snapshot | ### Version skew Both halves degrade to today's behavior: an old client sends plain strings (the resolver passes them through, tools take their legacy paths), and an old server doesn't advertise `fileInputs` (the new client never wraps). Wrappers on undeclared params fail the tool's own schema validation, so nothing can be smuggled through free-form args. ## Tests **Unit** (all green: registry 68, tool-server 992, tools-client 107, argent-mcp 66, argent-cli 117): - `tool-server/test/file-inputs.test.ts` — in-place gate, stat-mismatch → upload, truncated-upload rejection, directory gate, probe semantics, undeclared-target smuggling. - `argent-tools-client/test/file-inputs.test.ts` — wrapping (stat-only vs inlined content), multi-param template derivation, explicit-override respect, directive writes incl. path-validation refusals (relative, traversal, outside `.argent/flows`, bad charset/extension). - `tool-server/test/flows/flow-remote-recording.test.ts` — remote recording stays in memory + returns directives, nothing written on the host, replay/read from a boundary-resolved `flow_file`. - Updated existing screenshot-diff / native-profiler tests to the artifact-handle contract. **Local / co-located (unlink scenario)** — isolated tool-server + CLI on an iPhone 17 Pro simulator: full flow record→finish→read-prerequisite→execute loop (YAML written via the same path as before, in-place gate hit, no byte copies); screenshot-diff with and without `outputDir`; wire-level checks of the unlinked client shape (wrapper without content → in-place read) and the legacy plain-string shape; missing-file → 422 with a precise message. **Remote (Lima Linux VM as the agent machine, tool-server on the host Mac, `--host` bind + bearer auth)** — via both the CLI and the real MCP stdio server: - flow recording from the VM: YAML lands **in the VM's project**, `savedTo` rewritten to the VM path, *no* agent-path directories created on the host; replay from the VM uploads the YAML and drives the host simulator. - `screenshot` materializes into the VM temp cache; `screenshot-diff` round-trips both directions (VM baseline uploaded, live capture on host, diff artifacts downloaded back; MCP returns `image,text` content with the context diff inlined). - `gather-workspace-data` / `react-profiler-component-source` with VM-only paths → the new actionable 422. ## Notes - `native-profiler-analyze` was not exercised live (xctrace export is unusable on this machine's iOS 26.4 simulators — known, unrelated); its change is the same one-line `fileArtifact` pattern `react-profiler-analyze` already ships, covered by unit tests. - `react-profiler-analyze`'s own `project_root` (AST resolve stage) is left as-is deliberately: it degrades gracefully (report still renders) and gating it would break currently-working remote analysis. Tracked as follow-up alongside nested invocations (`run-sequence`/`flow-execute` steps bypass the HTTP boundary by design). 🤖 Generated with [Claude Code](https://claude.com/claude-code) --------- Co-authored-by: Cursor <cursoragent@cursor.com> Co-authored-by: Claude Fable 5 <noreply@anthropic.com> |
||
|
Paweł Fornagiel
|
3ccd89f333 |
feat: add native android profiling (#275)
# Android native profiling via in-process Perfetto WASM Connected with: https://github.com/software-mansion/argent-private/pull/16 ## TL;DR Adds **Android** to the native profiler (previously iOS-only) and does it with a brand-new engine: instead of shipping Perfetto's `trace_processor_shell` native binary and forking it per query, we now run **Perfetto's trace-processor compiled to WebAssembly, in-process** inside the tool-server. - **One ~13 MB `trace_processor.wasm`** runs on every OS/arch — no per-platform binary, no subprocess, fully offline. - The native profiler is refactored from an iOS-only implementation into a **platform-dispatch facade** (`platforms/ios.ts` + `platforms/android.ts`) over a new **`profiler-shared/`** layer, so iOS and Android share the aggregator, types, lifecycle, and time-alignment code. - New **release/pack pipeline** fetches + sha256-verifies the WASM bundle from `argent-private-releases`, and bundles it (plus SQL queries and the Perfetto TraceConfig) into the published `@swmansion/argent` package. End-to-end on a 76 MB trace with 1013 jank rows: **6.3 s** total analyze (the older per-hang-subprocess approach was projected at ~47 minutes and never finished within the tool-call deadline). --- ## TESTING MANUALLY To test manually run the following packing command checked-out on this branch: ``` npm run pack:mcp -- argent-@pFornagiel-native-android-devtools-perfetto ``` This will bundle the package with the new binaries. After this PR is fixed and the binaries are resolved from the main release, the `argent-@pFornagiel-native-android-devtools-perfetto` release can be deleted. ## Why Android needed native CPU/hang/memory profiling to reach parity with iOS. The natural engine is Perfetto's trace-processor, but the obvious ways to ship it are both bad: - **`trace_processor_shell` binary** — per-platform (a mac-arm64 pack shipped a mac-arm64-only tarball), pinned an OS/arch matrix, and cost a fresh ~1.3 s subprocess fork **per query**. With one subprocess forked per detected hang, a large trace projected to ~47 minutes. - **`get.perfetto.dev/trace_processor` Python launcher** — adds a runtime Python 3 dependency, resolves "latest" at first run (defeats version pinning), and makes the first analyze slow via a network fetch. The WASM engine is one cross-platform artifact, in-process and warm-cached: the trace is parsed **once** and reused across the whole analyze run plus every drill-down query. --- ## The three big pieces ### 1. In-process Perfetto WASM engine New package surface in `@argent/native-devtools-android`: | File | Role | | --- | --- | | `src/wasm-trace-processor.ts` (+476) | The whole engine: boot, glue patching, warm-cache, fault containment, chunked parse, cell decoding | | `src/perfetto-engine.d.ts` (+49) | Hand-written types for the version-agnostic RPC decoder (imported by path, so tsc can't see its own `.d.ts`) | | `src/errors.ts` (+68) | `TraceProcessorUnavailableError` (`wasm_load_failed` / `wasm_path_invalid`) | | `src/bundled-meta.ts` (+12) | `PERFETTO_VERSION` stamp, regenerated at pack time | | `src/index.ts` | Re-exports + package-root accessors (`traceProcessorQueriesDir`, `traceConfigPath`) | Public surface: `resolveTraceProcessorAssets()`, `queryWarm<Row>(tracePath, sql)`, `ensureTraceProcessorReady(tracePath)`, `disposeWarmEngine(tracePath)`. Highlights of how it works: - **Boots Google's prebuilt web/worker wasm under Node.** Installs a minimal worker scope (`self`, `WorkerGlobalScope`, `location`), reads the wasm bytes, patches the Emscripten glue (`useMemory64=false`, hand the bytes in directly, expose the bridge), and drives it over a `MessageChannel` + `EngineBase` subclass. Glue patches are applied **by unique string anchors that throw if they don't match** — the canary for glue-format drift on a Perfetto bump. - **Warm-engine cache, one per trace path.** LRU with `MAX_WARM_ENGINES = 3` (bounds resident memory at ~26–76 MB trace + wasm heap each), 5-min idle dispose (timer `unref()`'d), no failed-promise caching. - **Fault containment (`fatal` promise).** `EngineBase.fail()` throws synchronously inside the `MessagePort.onmessage` handler (e.g. an RPC framing error); left unhandled it becomes a Node `uncaughtException` and would **kill the entire tool-server**, wiping every device + profiler session. Each engine exposes a `fatal: Promise<never>`; the handler try/catches, rejects `fatal`, evicts the engine, and every load/query is `Promise.race`'d against it — so a contained fault is a clean rejection and the server stays up. - **Chunked trace parse.** Perfetto's RPC ring buffer rejects any frame > 64 MiB, so the trace is fed in 32 MiB chunks before `notifyEof()`. ### 2. Android native profiler + the shared/dispatch refactor **Capture** (`utils/android-profiler/capture.ts`, `detect-app.ts`): spawns `adb shell perfetto --txt -c - --background-wait -o /data/.../argent-<ts>.pftrace`, piping the TraceConfig on **stdin** (SELinux denies `shell:s0` writes under `/data/misc/perfetto-traces/`, so the config can't be staged as a file), waits for the PID on stdout, then on stop sends `kill -TERM`, polls `/proc/<pid>`, and `adb pull`s the trace. **Pipeline** (`utils/android-profiler/pipeline/`): PerfettoSQL does the parsing and aggregation, so the Android pipeline is intentionally **two files, not iOS's four**. The per-hang annotation pass is **batched** — `runBatchedHangFolds` inlines all hang windows as a `VALUES` table into one SQL script (two derived views unioned in a single terminal SELECT, because the engine returns only the final statement's rows) and runs it once against the warm engine. **The platform-dispatch refactor** — `native-profiler-start/stop/analyze` shrank from monolithic iOS files (~160–330 lines each) into **thin routers** (~60–110 lines) that branch on `api.platform`: ``` native-profiler-start ─┬─ ios? → platforms/ios.ts (xctrace) └─ android → platforms/android.ts (adb perfetto) ``` A new **`utils/profiler-shared/`** layer holds everything both backends use: | File | What was hoisted | | --- | --- | | `types.ts` | Unified `Bottleneck = CpuHotspot \| UiHang \| MemoryLeak \| MemoryRssGrowth`, each now carrying a `platform` field; shared `NativeProfilerAnalyzeResult`; `RECORDING_CAP_MS` | | `aggregate.ts` | Generic `aggregateCpuHotspots(AggregatorInputRow[])` — burst windowing (`BURST_GAP_MS=500`), severity banding (RED >15%, YELLOW 3–15%, <3% dropped). iOS feeds it via a pre-pass; Android feeds it precomputed SQL rows | | `lifecycle.ts` | `shutdownChild` SIGINT→SIGTERM→SIGKILL ladder + `waitForChildExit` (no PID-reuse hazard) | | `thread.ts` | `normalizeThreadName` for both xctrace and Perfetto thread names | | `time-align.ts` | New `buildPerfettoAnchor()` + `windowsOverlap()` alongside the existing iOS/React anchors | | `format.ts` | `formatBytes` (moved out of the combined report) | iOS was refactored to fit this shape: `ios-profiler/lifecycle.ts` and `types.ts` become thin re-exports; `pipeline/02-aggregate.ts` shrank from ~200 to ~60 lines (keeps its iOS-only dominant-function pre-pass, delegates grouping/bursting/ severity to the shared aggregator). **No iOS behavior change** — same outputs, shared internals. **Session blueprint** (`native-profiler-session.ts`): generic `capturePid` / `captureProcess` (was `xctracePid` / `xctraceProcess`), new `platform` and `androidOnDeviceTracePath` fields, and **platform-branched dispose** (iOS SIGKILLs xctrace; Android `kill -KILL`s perfetto, `rm`s the device trace, and disposes the warm WASM engine). **Drill-down tools**: `profiler-load.ts` learns to restore Android `.pftrace` sessions (with an optional metadata sidecar / `app_process` override); `profiler-stack-query.ts` splits into iOS and Android branches. Unlike iOS (which caches `parsedData` in memory), Android **re-queries the `.pftrace`** per drill-down — the warm engine is the cache, so each query costs only SQL execution time. ### 3. Packaging & release pipeline The WASM bundle's three third-party artifacts (`trace_processor.wasm`, `engine_bundle.node.js`, `engine.mjs`) + `LICENSE` are **never committed** (`.gitignore` lists the generated files but not the dir). They're built and sha256-checksummed in `argent-private` CI, published to `argent-private-releases`, and on our side: - **`scripts/download-trace-processor.sh`** (+88, new) — `gh release download` → verify tarball sha256 (**FATAL on mismatch** — the blobs are unsigned) → extract → re-verify via `SHA256SUMS` → assert all four present. - **`scripts/pack-mcp.cjs`** (+60, new) — single pack entry point: downloads simulator-server, native binaries, **and** the trace-processor bundle, then builds and `npm pack`s `@swmansion/argent`. `package.json`'s `pack` / `pack:mcp` now call this script. - **`packages/argent/scripts/bundle-tools.cjs`** (heavily reworked, +574/-...) — copies `assets/trace-processor/`, `assets/queries/`, and `argent.tracecfg.pbtxt` into `packages/argent/assets/`, and regenerates `bundled-meta.ts` to stamp the Perfetto version. - **`publish.yml` / `publish-next.yml`** — run the download step and then **verify**: all four WASM files present, the wasm magic-byte check (`head -c4 … | grep $'\x00asm'`), plus the queries dir, TraceConfig, manifest, and Android helper APK. Asset reorg: `native-devtools-android` now ships `dist/`, `bin/`, and `assets/` (SQL queries, TraceConfig, manifest moved under `assets/`); `.gitignore` swaps the old per-file `packages/argent/manifest.json` entry for the whole `packages/argent/assets/` dir. --- ## Testing 15 new test files under `packages/tool-server/test/android-perfetto/` cover the engine and pipeline, including: - `dispatch.test.ts` — start/stop/analyze route to the right platform backend. - `aggregate-shared.test.ts` — the shared aggregator on both iOS and Android-shaped input. - `hang-folds-batched.test.ts` / `hang-fold.test.ts` — the batched per-hang SQL fold. - `hang-severity.test.ts` — jank-reason → RED/YELLOW classification. - `dispose.test.ts` / `dispose-warm-engine.test.ts` — session dispose kills the daemon, rms the device trace, and tears down the warm engine. - `stop-recovery.test.ts` — `profilingActive` is forced false even if `adb pull` fails (so a retry-start doesn't wedge). - `run-tp.test.ts` — `renderSqlTemplate` token rendering + the "unused substitution" guard. - `manifest-hint.test.ts` — zero-callstack detection emits the manifest hint. - `profiler-load.test.ts` — `.pftrace` session restore with/without the metadata sidecar. Plus `native-profiler-analyze-failure.test.ts` for the engine-unavailable banner, and updated iOS tests for the shared-lifecycle refactor. `@argent/native-devtools-android` gains its own vitest setup (`vitest.config.ts`, `tsconfig.test.json`, `test/errors.test.ts`). --- ## Notes & follow-ups - **Manifest requirement.** Perfetto's `linux.perf` data source needs `/proc/<pid>/mem` access, granted only to a **debuggable** or `<profileable android:shell="true"/>` app. A release build without either silently produces zero callstacks; the pipeline detects this and emits a **manifest hint** instead of an "All clear" report. - **Version pinning lives in the `argent-private` submodule** (`PERFETTO_VERSION` / `PERFETTO_WASM_TAG` / `LYNX_TRACE_PROCESSOR_VERSION`). Bumping is documented in `argent-private/docs/ANDROID_PERFETTO.md`. JS↔engine version skew is structurally impossible (version-agnostic decoder + vendored wasm). - **`disposeWarmEngine` is exported but not wired into session-stop** — warm engines are reclaimed only by the idle timer + LRU cap. Acceptable today (bounded memory); wire it into session dispose if explicit teardown is ever needed. - **Submodule bump.** This PR moves `packages/argent-private` to `22e55f90` — see the companion `argent-private` PR description, including its merge-hygiene note (that branch should pick up latest `argent-private` main before the pointer is finalized). ### Companion docs added in this branch - `utils/android-profiler/ANDROID_PROFILER_REFERENCE.md` — stack, capture, queries, manifest requirement. - `utils/android-profiler/PIPELINE_DESIGN.md` — the "why" (two-file pipeline, shared-aggregator hoist, re-query-don't-cache, batched fold). --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com> |
||
|
Ignacy Łątka
|
2180dd8e6a |
ci: typecheck test files in CI (#202)
We've had lots of type errors in our tests, scripts and other non-core
code despite all CI being green. This PR fixes that.
This PR also removes some redundant features like `@ts-check` inline
directives which have been replaced by proper ts config files.
<details>
## AI Summary
Audit of every CI workflow against every package's `package.json` and
`tsconfig` surfaced several gaps where CI did partial coverage. This PR
closes them and fixes every pre-existing type error the new gates
surfaced.
## Gaps closed
| # | Gap | Fix |
|---|---|---|
| 1 | `tsc --build` only covers `src/**`; test files (`test/**`,
`tests/**`) never typecheck. Vitest transforms tests with esbuild —
strips types instead of checking them. | New `tsconfig.test.json` per
test-bearing package (`composite: false`, `noEmit`, `rootDir: "."`,
includes `src/**` + `test/**` (or `tests/**`) + `vitest.config.ts`). New
`typecheck:tests` npm script per package. New CI step `npm run
typecheck:tests --workspaces --if-present`. |
| 2 | Build/publish scripts (`scripts/*.{cjs,mjs}`,
`packages/argent/scripts/*.cjs`, `packages/skills/scripts/install.js`) —
JS, never typechecked. `sync-readme.cjs` runs in `prepack`;
`postinstall.cjs` runs on every install; `bundle-tools.cjs` is
build-critical. A typo only surfaces at runtime. | Root
`tsconfig.scripts.json` with `allowJs` + `checkJs` + `noImplicitReturns`
covers all script paths. Root `typecheck:scripts` npm script. New CI
step. Per-file `// @ts-check` comments removed — tsconfig drives
inclusion. |
| 3 | `argent-cli` had no `typecheck:tests` script. When tests get added
there they'd silently bypass the gate. | Added `tsconfig.test.json` +
`typecheck:tests` script. `--if-present` would have skipped it anyway,
so this is future-proofing. |
| 4 | Local `prettier --check .` walked into the `argent-private`
submodule and reported violations there. CI didn't see this because
`actions/checkout@v4` doesn't init submodules — so local `prettier
--check` did not match CI behavior. | New `.prettierignore` excludes the
submodule plus `dist/`, `node_modules/`, lock files, `tsbuildinfo`.
No-op in CI; aligns local with remote. |
## Test type errors fixed
These were already present. The `npm test` passed because esbuild
stripped the offending types:
- **argent-installer**: widen `scope` literal type for
dead-code-elimination guard; non-null `addAllowlist!`/`removeAllowlist!`
(optional methods on adapter).
- **argent-mcp**: add `.js` extensions for NodeNext module resolution.
- **registry**: broaden `StaticBlueprintResult` blueprint api from `{
id; deps? }` to `Record<string, unknown>` so existing tests can pass
arbitrary api shapes.
- **tool-server**:
- non-null `zodSchema!` on tool defs that always carry one
(`launchAppTool`, `restartAppTool`, etc.)
- cast sentinel `"ignored"` strings to `DeviceInfo` in factory-rejection
tests
- match the 4-arg `dispatchByPlatform<IosServices, AndroidServices,
Params, Result>` signature (tests still passed the old 3-arg shape)
- `assertFlowRunResult` type guard for the `FlowRunResult |
FlowPrerequisiteNotice` union before `.steps` access
- swap broken `typeof import("supertest").default` for static `import
supertest` (supertest is `export = supertest`)
- fill in the 4 fields missing from a `NativeProfilerSessionApi` mock
(`xctraceProcess`, `recordingTimedOut`, `recordingExitedUnexpectedly`,
`lastExitInfo`)
- cast a vitest `Mock` to a plain function at one direct call site where
the `Procedure | Constructable` union loses callability
- fix `update-checker.test.ts` reading `../../package.json` (off-by-one
— went past tool-server; was relying on a Vite resolver quirk)
- replace one cross-package `../../../registry/src/index` import with
`@argent/registry` so registry's private `services` field isn't compared
between src and dist declarations
All fixes are type-level. Full `npm test --workspaces --if-present`
still reports **991 passed across 87 test files**.
</details>
|
||
|
Ignacy Łątka
|
837d088744 |
test: cover #148 (android) (#186)
This PR adds tests for #148. ## Summary Moves the Android emulator branch's test coverage out of #148 so that PR can be reviewed without the test churn. This branch adds the test files back on top of the (now test-stripped) base branch. Tests inseparable from source changes (`describe-tool`, `workspace-reader`, `workspace-reader-integration`) and the `boot-simulator.test.ts` deletion stayed on #148 - base PR's CI passes on its own. ## Test plan - [x] `npx tsc --build` — clean - [x] `npx prettier --check .` — clean - [x] `npm test --workspaces --if-present` — 585/585 tool-server tests pass; all other workspaces green --------- Co-authored-by: Paweł Fornagiel <pawel.fornagiel2@gmail.com> |