feat(adapter-utils): add observable sandbox sync progress (#8395)

## Thinking Path

> - Paperclip is the control plane for running AI-agent companies, so
long-running remote work needs to stay observable to human operators.
> - Cloud / sandbox agents are an active roadmap area, and their
workspace sync path is part of the runtime substrate every remote coding
run depends on.
> - In the sandbox and SSH execution-target flows, Paperclip logged that
sync had started, then often went silent for the full transfer window.
> - That made large remote syncs feel stalled and also hid a real
performance problem in the command-managed sandbox upload path.
> - The first part of this pull request threads a throttled
progress-reporting surface through the adapter execution-target stack so
sync and restore work can emit meaningful updates.
> - The second part fixes the command-managed sandbox transport itself:
it removes the old serial 32KB append bottleneck, but also falls back
away from the single-stream path when a provider-backed sandbox runner
cannot surface mid-flight stdin progress.
> - The result is that sandbox and SSH transfers are both faster and
more observable, including the live Daytona-style sandbox case that
previously only emitted `0%` and `100%`.

## Linked Issues or Issue Description

No public GitHub issue exists for this bug, so it is described inline
below following the bug report template.

### What happened

- Remote sandbox and SSH workspace syncs could spend a long time
transferring data while only logging a start line (`Syncing workspace
and runtime assets to sandbox environment`) and, at best, a terminal
line.
- In the command-managed sandbox path, the original upload
implementation also paid a large performance cost by appending base64
data in many small sequential remote writes (thousands of serial 32KB
round-trips on a large workspace).
- After the initial transport rewrite, live provider-backed sandbox runs
still only emitted `0%` and `100%` because the single-stream stdin RPC
buffered progress until completion.

### Expected behavior

- Long-running sandbox and SSH syncs should periodically report how much
of the transfer is complete (a percentage and/or MB transferred) so an
operator can tell the run is healthy and making progress rather than
stuck.
- The main sandbox upload path should not be artificially slow.
- A transfer that fails partway should leave an explicit failure marker
in the log rather than a dangling intermediate percentage.

### Steps to reproduce

1. Run an agent against a sandbox (command-managed) or SSH
(remote-managed) execution target with a non-trivial workspace.
2. Watch the run log during the workspace/runtime asset sync phase.
3. Observe that the log shows the sync start line and then stays silent
for the full transfer (live provider-backed sandbox runs only show `0%`
then `100%`).

### Paperclip version or commit

- Branch `PAPA-825-provide-status-updates-when-syncing-sandboxes` off
`master`.

### Deployment mode

- Self-hosted / local instance using sandbox (command-managed) and SSH
(remote-managed) execution targets, including provider-backed sandbox
runners.

## What Changed

- Added shared throttled runtime progress reporting and threaded
`onProgress` through the adapter execution-target surface and adapter
`execute.ts` entrypoints.
- Added sync and restore progress reporting for the command-managed
sandbox path and the SSH/remote-managed path, including git
import/export progress where totals are known.
- Reworked command-managed sandbox transfer behavior so uploads use the
faster single-stream path when appropriate, but fall back to chunked
progress-emitting writes when the runner cannot expose mid-stream stdin
progress.
- Marked provider-backed environment sandbox runners as not supporting
single-stream stdin progress so live sandbox runs emit meaningful
intermediate updates instead of only `0%` and `100%`.
- Emit an explicit terminal failure marker (`failed at NN% (x/y MB)`)
when an SSH/tar transfer rejects, so a failed sync no longer leaves a
dangling intermediate percentage in the log.
- Run the SSH sync/restore size estimate (local directory walk / remote
`du` probe) concurrently with the transfer instead of awaiting it before
opening the pipe, so progress instrumentation no longer adds startup
latency proportional to workspace file count.
- Added and extended focused regression coverage for runtime progress
throttling and the new failure marker, command-managed sandbox
transfers, sandbox orchestration, SSH transfer progress, and environment
execution-target wiring.

## Verification

- `pnpm exec vitest run
packages/adapter-utils/src/runtime-progress.test.ts
packages/adapter-utils/src/ssh-fixture.test.ts
packages/adapter-utils/src/command-managed-runtime.test.ts
packages/adapter-utils/src/sandbox-managed-runtime.test.ts`
- `pnpm exec vitest run
packages/adapter-utils/src/command-managed-runtime.test.ts
server/src/__tests__/environment-execution-target.test.ts`
- `npx tsc --noEmit` for `packages/adapter-utils`

## Risks

- The provider-backed sandbox fallback now prefers chunked
command-managed writes when progress hooks are active, so
small-to-medium uploads may trade some raw throughput for observable
intermediate progress on runtimes that cannot surface true mid-stream
stdin progress.
- Progress percentages on tar-based transfers still depend on estimates
in some cases, so operators may briefly see MB-only lines before the
estimate resolves, then near-final clamping before the terminal `100%`
line.
- This PR changes shared execution-target behavior used by multiple
adapters, so regressions would most likely appear in remote runtime
setup/teardown flows rather than in a single adapter.

## Model Used

- Initial implementation: OpenAI GPT-5.4 via Codex local agent
(`codex_local`), high reasoning mode.
- Observability follow-ups (failure marker, concurrent size estimate,
added tests): Claude Opus 4.8 via Claude Code (`claude_local`).

## Checklist

- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [ ] If this change affects the UI, I have included before/after
screenshots
- [ ] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [ ] I will address all Greptile and reviewer comments before
requesting merge

---------

Co-authored-by: Paperclip <noreply@paperclip.ing>
This commit is contained in:
Devin Foley
2026-06-20 13:03:42 -07:00
committed by GitHub
parent 5cace19aca
commit 07e98d2b2c
19 changed files with 1377 additions and 110 deletions
@@ -1,15 +1,80 @@
import { mkdir, mkdtemp, readFile, rm, writeFile } from "node:fs/promises";
import os from "node:os";
import path from "node:path";
import { execFile as execFileCallback } from "node:child_process";
import { execFile as execFileCallback, spawn } from "node:child_process";
import { promisify } from "node:util";
import { afterEach, describe, expect, it } from "vitest";
import { prepareCommandManagedRuntime } from "./command-managed-runtime.js";
import {
createCommandManagedRuntimeClient,
prepareCommandManagedRuntime,
type CommandManagedRuntimeRunner,
} from "./command-managed-runtime.js";
import type { RunProcessResult } from "./server-utils.js";
const execFile = promisify(execFileCallback);
interface SpawnRunnerHandle {
runner: CommandManagedRuntimeRunner;
calls: Array<{ command: string; args?: string[]; cwd?: string; stdin?: string }>;
}
// A runner that actually executes the shell scripts (piping stdin through a real
// pipe so multi-MB payloads work) and replays stdout through onLog in several
// chunks so the streaming readFile byte-counter is exercised.
function makeSpawnRunner(options: { supportsSingleStreamStdinProgress?: boolean } = {}): SpawnRunnerHandle {
const calls: Array<{ command: string; args?: string[]; cwd?: string; stdin?: string }> = [];
const runner: CommandManagedRuntimeRunner = {
supportsSingleStreamStdinProgress: options.supportsSingleStreamStdinProgress,
execute: async (input) =>
await new Promise<RunProcessResult>((resolve) => {
calls.push({ command: input.command, args: input.args, cwd: input.cwd, stdin: input.stdin });
const startedAt = new Date().toISOString();
const command =
input.command === "sh" ? "/bin/sh" : input.command === "bash" ? "/bin/bash" : input.command;
const child = spawn(command, input.args ?? [], {
cwd: input.cwd,
env: { ...process.env, ...input.env },
});
let stdout = "";
let stderr = "";
child.stdout.on("data", (chunk) => {
stdout += chunk.toString("utf8");
});
child.stderr.on("data", (chunk) => {
stderr += chunk.toString("utf8");
});
child.on("error", () => {
resolve({ exitCode: 127, signal: null, timedOut: false, stdout, stderr, pid: null, startedAt });
});
child.on("close", async (code) => {
if (input.onLog && stdout.length > 0) {
const chunkSize = Math.max(1, Math.ceil(stdout.length / 4));
for (let offset = 0; offset < stdout.length; offset += chunkSize) {
await input.onLog("stdout", stdout.slice(offset, offset + chunkSize));
}
}
resolve({
exitCode: code ?? 0,
signal: null,
timedOut: false,
stdout,
stderr,
pid: child.pid ?? null,
startedAt,
});
});
if (input.stdin != null) child.stdin.write(input.stdin);
child.stdin.end();
}),
};
return { runner, calls };
}
function toArrayBuffer(buffer: Buffer): ArrayBuffer {
return Uint8Array.from(buffer).buffer;
}
describe("command managed runtime", () => {
const cleanupDirs: string[] = [];
@@ -117,7 +182,9 @@ describe("command managed runtime", () => {
await expect(readFile(path.join(remoteWorkspaceDir, "README.md"), "utf8")).resolves.toBe("local workspace\n");
await expect(readFile(path.join(remoteWorkspaceDir, ".paperclip-runtime", "state.json"), "utf8")).rejects
.toMatchObject({ code: "ENOENT" });
expect(calls.every((call) => call.stdin == null)).toBe(true);
// The single-stream upload pipes the tarball through exactly one stdin-backed
// process (the speed fix); nothing else streams stdin.
expect(calls.filter((call) => call.stdin != null).length).toBe(1);
await mkdir(path.join(remoteWorkspaceDir, ".paperclip-runtime"), { recursive: true });
await writeFile(path.join(remoteWorkspaceDir, "README.md"), "remote workspace\n", "utf8");
@@ -129,7 +196,9 @@ describe("command managed runtime", () => {
.toBe("{\"keep\":true}\n");
await expect(readFile(path.join(localWorkspaceDir, ".paperclip-runtime", "remote-state.json"), "utf8")).rejects
.toMatchObject({ code: "ENOENT" });
expect(calls.every((call) => call.stdin == null)).toBe(true);
// Restore streams the download through `base64`/onLog (no stdin), so the only
// stdin-backed call remains the single upload from prepare.
expect(calls.filter((call) => call.stdin != null).length).toBe(1);
});
it("runs setup commands from a stable root cwd when staging into a nested remote workspace dir", async () => {
@@ -143,64 +212,7 @@ describe("command managed runtime", () => {
await mkdir(remoteBaseDir, { recursive: true });
await writeFile(path.join(localWorkspaceDir, "README.md"), "local workspace\n", "utf8");
const calls: Array<{
command: string;
args?: string[];
cwd?: string;
env?: Record<string, string>;
stdin?: string;
timeoutMs?: number;
}> = [];
const runner = {
execute: async (input: {
command: string;
args?: string[];
cwd?: string;
env?: Record<string, string>;
stdin?: string;
timeoutMs?: number;
}): Promise<RunProcessResult> => {
calls.push({ ...input });
const startedAt = new Date().toISOString();
try {
const result = await execFile(input.command === "sh" ? "/bin/sh" : input.command, input.args ?? [], {
cwd: input.cwd,
env: {
...process.env,
...input.env,
},
maxBuffer: 32 * 1024 * 1024,
timeout: input.timeoutMs,
});
return {
exitCode: 0,
signal: null,
timedOut: false,
stdout: result.stdout,
stderr: result.stderr,
pid: null,
startedAt,
};
} catch (error) {
const err = error as NodeJS.ErrnoException & {
stdout?: string;
stderr?: string;
code?: string | number | null;
signal?: NodeJS.Signals | null;
killed?: boolean;
};
return {
exitCode: typeof err.code === "number" ? err.code : null,
signal: err.signal ?? null,
timedOut: Boolean(err.killed && input.timeoutMs),
stdout: err.stdout ?? "",
stderr: err.stderr ?? "",
pid: null,
startedAt,
};
}
},
};
const { runner, calls } = makeSpawnRunner();
await prepareCommandManagedRuntime({
runner,
@@ -217,4 +229,99 @@ describe("command managed runtime", () => {
expect(calls.every((call) => call.cwd === "/")).toBe(true);
await expect(readFile(path.join(remoteWorkspaceDir, "README.md"), "utf8")).resolves.toBe("local workspace\n");
});
it("uploads a multi-MB payload in a single process and preserves exact bytes", async () => {
const rootDir = await mkdtemp(path.join(os.tmpdir(), "paperclip-command-write-"));
cleanupDirs.push(rootDir);
const remotePath = path.join(rootDir, "nested", "payload.bin");
// ~3 MB of every byte value so the test catches any non-binary-safe handling.
const payload = Buffer.alloc(3 * 1024 * 1024);
for (let i = 0; i < payload.length; i++) payload[i] = i % 256;
const { runner, calls } = makeSpawnRunner({ supportsSingleStreamStdinProgress: true });
const client = createCommandManagedRuntimeClient({ runner, commandCwd: "/", timeoutMs: 30_000 });
const progress: Array<{ done: number; total: number | null }> = [];
await client.writeFile(remotePath, toArrayBuffer(payload), {
onProgress: (done, total) => {
progress.push({ done, total });
},
});
// Exactly one upload process: O(1) round-trips regardless of payload size.
expect(calls.length).toBe(1);
expect(calls[0].stdin).toBeTypeOf("string");
const written = await readFile(remotePath);
expect(written.equals(payload)).toBe(true);
// Progress is monotonically non-decreasing and reaches the total.
expect(progress.length).toBeGreaterThan(0);
for (let i = 1; i < progress.length; i++) {
expect(progress[i].done).toBeGreaterThanOrEqual(progress[i - 1].done);
}
expect(progress.at(-1)).toEqual({ done: payload.length, total: payload.length });
});
it("falls back to chunked upload progress when the runner cannot report mid-stream stdin progress", async () => {
const rootDir = await mkdtemp(path.join(os.tmpdir(), "paperclip-command-write-fallback-"));
cleanupDirs.push(rootDir);
const remotePath = path.join(rootDir, "nested", "payload.bin");
const payload = Buffer.alloc(12 * 1024 * 1024);
for (let i = 0; i < payload.length; i++) payload[i] = i % 256;
const { runner, calls } = makeSpawnRunner({ supportsSingleStreamStdinProgress: false });
const client = createCommandManagedRuntimeClient({ runner, commandCwd: "/", timeoutMs: 30_000 });
const progress: Array<{ done: number; total: number | null }> = [];
await client.writeFile(remotePath, toArrayBuffer(payload), {
onProgress: (done, total) => {
progress.push({ done, total });
},
});
const written = await readFile(remotePath);
expect(written.equals(payload)).toBe(true);
// Provider-backed sandbox runners cannot surface mid-flight progress for a
// single stdin RPC, so we intentionally use several large append commands.
expect(calls.length).toBeGreaterThan(2);
expect(progress.length).toBeGreaterThan(2);
for (let i = 1; i < progress.length; i++) {
expect(progress[i].done).toBeGreaterThanOrEqual(progress[i - 1].done);
}
expect(progress.at(-1)).toEqual({ done: payload.length, total: payload.length });
});
it("streams a download through onLog and reports monotonic byte progress to the total", async () => {
const rootDir = await mkdtemp(path.join(os.tmpdir(), "paperclip-command-read-"));
cleanupDirs.push(rootDir);
const remotePath = path.join(rootDir, "download.bin");
const payload = Buffer.alloc(512 * 1024);
for (let i = 0; i < payload.length; i++) payload[i] = (i * 7) % 256;
await writeFile(remotePath, payload);
const { runner } = makeSpawnRunner();
const client = createCommandManagedRuntimeClient({ runner, commandCwd: "/", timeoutMs: 30_000 });
const progress: Array<{ done: number; total: number | null }> = [];
const bytes = await client.readFile(remotePath, {
onProgress: (done, total) => {
progress.push({ done, total });
},
});
expect(Buffer.from(bytes as ArrayBuffer).equals(payload)).toBe(true);
// The runner replays stdout in several chunks, so progress arrives in steps.
expect(progress.length).toBeGreaterThan(1);
for (let i = 1; i < progress.length; i++) {
expect(progress[i].done).toBeGreaterThanOrEqual(progress[i - 1].done);
}
expect(progress.every((entry) => entry.total === payload.length)).toBe(true);
expect(progress.at(-1)?.done).toBe(payload.length);
});
});
@@ -8,8 +8,16 @@ import {
} from "./sandbox-managed-runtime.js";
import { preferredShellForSandbox, shellCommandArgs } from "./sandbox-shell.js";
import type { RunProcessResult } from "./server-utils.js";
import type { RuntimeProgressSink } from "./runtime-progress.js";
export interface CommandManagedRuntimeRunner {
/**
* True only when `execute({ stdin })` can surface useful in-flight progress
* for a single stdin-backed command. Provider-backed sandbox runners usually
* complete the entire RPC before returning, so they should leave this false
* and let the caller choose a chunked upload path when progress is requested.
*/
supportsSingleStreamStdinProgress?: boolean;
execute(input: {
command: string;
args?: string[];
@@ -40,7 +48,14 @@ function mergeRuntimeExcludes(entries: string[] | undefined): string[] {
return [...new Set([".paperclip-runtime", ...(entries ?? [])])];
}
const REMOTE_WRITE_BASE64_CHUNK_SIZE = 32 * 1024;
// Largest base64 body we hand to the runner as a single stdin string. Normal
// multi-MB workspace/asset tarballs stay well under this and upload in one
// round-trip; anything larger uses the bounded chunked-append fallback so a
// runaway stdin string can't blow the runner/provider RPC limits.
const REMOTE_WRITE_SINGLE_STREAM_MAX_BASE64_BYTES = 96 * 1024 * 1024;
// Fallback chunk size (base64 bytes). Kept a multiple of 4 so each chunk is a
// self-contained base64 unit that decodes cleanly on its own.
const REMOTE_WRITE_FALLBACK_BASE64_CHUNK_SIZE = 4 * 1024 * 1024;
function toBuffer(bytes: Buffer | Uint8Array | ArrayBuffer): Buffer {
if (Buffer.isBuffer(bytes)) return bytes;
@@ -62,13 +77,21 @@ export function createCommandManagedRuntimeClient(input: {
shellCommand?: "bash" | "sh" | null;
}): SandboxManagedRuntimeClient {
const shellCommand = preferredShellForSandbox(input.shellCommand);
const runShell = async (script: string, opts: { stdin?: string; timeoutMs?: number } = {}) => {
const runShell = async (
script: string,
opts: {
stdin?: string;
timeoutMs?: number;
onLog?: (stream: "stdout" | "stderr", chunk: string) => Promise<void>;
} = {},
) => {
const result = await input.runner.execute({
command: shellCommand,
args: shellCommandArgs(script),
cwd: input.commandCwd,
stdin: opts.stdin,
timeoutMs: opts.timeoutMs ?? input.timeoutMs,
onLog: opts.onLog,
});
requireSuccessfulResult(result, script);
return result;
@@ -78,25 +101,101 @@ export function createCommandManagedRuntimeClient(input: {
makeDir: async (remotePath) => {
await runShell(`mkdir -p ${shellQuote(remotePath)}`);
},
writeFile: async (remotePath, bytes) => {
const body = toBuffer(bytes).toString("base64");
writeFile: async (remotePath, bytes, options) => {
const buffer = toBuffer(bytes);
const total = buffer.byteLength;
const body = buffer.toString("base64");
const remoteDir = path.posix.dirname(remotePath);
const remoteTempPath = `${remotePath}.paperclip-upload.b64`;
const remoteTempPath = `${remotePath}.paperclip-upload`;
const canUseSingleStreamProgressPath =
!options?.onProgress || input.runner.supportsSingleStreamStdinProgress === true;
await runShell(
`mkdir -p ${shellQuote(remoteDir)} && rm -f ${shellQuote(remoteTempPath)} && : > ${shellQuote(remoteTempPath)}`,
);
for (let offset = 0; offset < body.length; offset += REMOTE_WRITE_BASE64_CHUNK_SIZE) {
const chunk = body.slice(offset, offset + REMOTE_WRITE_BASE64_CHUNK_SIZE);
await runShell(`printf '%s' ${shellQuote(chunk)} >> ${shellQuote(remoteTempPath)}`);
// Primary path: a single round-trip. Stream the entire base64 body to one
// `base64 -d` process via stdin, decode straight into a temp file, then
// atomically rename into place. This replaces the previous loop that did
// one `printf >> tmpfile` shell round-trip per 32 KB — thousands of serial
// processes for a large workspace — with exactly one process.
if (
body.length <= REMOTE_WRITE_SINGLE_STREAM_MAX_BASE64_BYTES &&
canUseSingleStreamProgressPath
) {
await options?.onProgress?.(0, total);
await runShell(
`mkdir -p ${shellQuote(remoteDir)} && ` +
`base64 -d > ${shellQuote(remoteTempPath)} && ` +
`mv -f ${shellQuote(remoteTempPath)} ${shellQuote(remotePath)}`,
{ stdin: body },
);
await options?.onProgress?.(total, total);
return;
}
// Bounded fallback for payloads too large to hand the runner as one stdin
// string: append the base64 body to a remote temp file in large chunks
// (orders of magnitude fewer round-trips than the old 32 KB loop), decoding
// each self-contained chunk on arrival and emitting progress per write,
// then atomically rename into place.
await runShell(
`base64 -d < ${shellQuote(remoteTempPath)} > ${shellQuote(remotePath)} && rm -f ${shellQuote(remoteTempPath)}`,
`mkdir -p ${shellQuote(remoteDir)} && ` +
`rm -f ${shellQuote(remoteTempPath)} && : > ${shellQuote(remoteTempPath)}`,
);
for (let offset = 0; offset < body.length; offset += REMOTE_WRITE_FALLBACK_BASE64_CHUNK_SIZE) {
const chunk = body.slice(offset, offset + REMOTE_WRITE_FALLBACK_BASE64_CHUNK_SIZE);
await runShell(`base64 -d >> ${shellQuote(remoteTempPath)}`, { stdin: chunk });
const decodedSoFar = Math.min(total, Math.floor(((offset + chunk.length) * 3) / 4));
await options?.onProgress?.(decodedSoFar, total);
}
await runShell(`mv -f ${shellQuote(remoteTempPath)} ${shellQuote(remotePath)}`);
await options?.onProgress?.(total, total);
},
readFile: async (remotePath) => {
const result = await runShell(`base64 < ${shellQuote(remotePath)}`);
return Buffer.from(result.stdout.replace(/\s+/g, ""), "base64");
readFile: async (remotePath, options) => {
// Decoded file size up front so download progress can be a percentage.
// Only paid when a progress hook is attached.
let totalBytes: number | null = null;
if (options?.onProgress) {
const sizeResult = await runShell(`wc -c < ${shellQuote(remotePath)}`);
const parsed = Number.parseInt(sizeResult.stdout.trim(), 10);
totalBytes = Number.isFinite(parsed) && parsed >= 0 ? parsed : null;
}
// Stream the remote `base64` stdout through a byte counter, decoding each
// 4-char-aligned slice incrementally rather than buffering the whole
// response as one string. Falls back to the buffered result when the
// runner does not surface incremental stdout.
const decodedChunks: Buffer[] = [];
let b64Remainder = "";
let decodedSoFar = 0;
let streamed = false;
const result = await runShell(`base64 < ${shellQuote(remotePath)}`, {
onLog: async (stream, chunk) => {
if (stream !== "stdout") return;
streamed = true;
const data = b64Remainder + chunk.replace(/\s+/g, "");
const alignedLen = data.length - (data.length % 4);
if (alignedLen > 0) {
const decoded = Buffer.from(data.slice(0, alignedLen), "base64");
decodedChunks.push(decoded);
decodedSoFar += decoded.byteLength;
}
b64Remainder = data.slice(alignedLen);
if (options?.onProgress) {
const done = totalBytes != null ? Math.min(totalBytes, decodedSoFar) : decodedSoFar;
await options.onProgress(done, totalBytes);
}
},
});
let out: Buffer;
if (streamed) {
if (b64Remainder.length > 0) {
decodedChunks.push(Buffer.from(b64Remainder, "base64"));
}
out = Buffer.concat(decodedChunks);
} else {
out = Buffer.from(result.stdout.replace(/\s+/g, ""), "base64");
}
await options?.onProgress?.(out.byteLength, totalBytes ?? out.byteLength);
return out;
},
listFiles: async (remotePath) => {
const result = await runShell(
@@ -146,6 +245,9 @@ export async function prepareCommandManagedRuntime(input: {
installCommand?: string | null;
/** When provided alongside `installCommand`, skip the install if `command -v <detectCommand>` succeeds. */
detectCommand?: string | null;
// Upload progress sink. Forwarded to prepareSandboxManagedRuntime; the child
// task wires it into the byte-counting writeFile/readFile transport.
onProgress?: RuntimeProgressSink;
}): Promise<PreparedSandboxManagedRuntime> {
const timeoutMs = input.spec.timeoutMs && input.spec.timeoutMs > 0 ? input.spec.timeoutMs : 300_000;
const workspaceRemoteDir = input.workspaceRemoteDir ?? input.spec.remoteCwd;
@@ -193,6 +295,7 @@ export async function prepareCommandManagedRuntime(input: {
workspaceExclude: mergeRuntimeExcludes(input.workspaceExclude),
preserveAbsentOnRestore: input.preserveAbsentOnRestore,
assets: input.assets,
onProgress: input.onProgress,
});
}
}
@@ -227,5 +330,6 @@ export async function prepareCommandManagedRuntime(input: {
workspaceExclude: mergeRuntimeExcludes(input.workspaceExclude),
preserveAbsentOnRestore: input.preserveAbsentOnRestore,
assets: input.assets,
onProgress: input.onProgress,
});
}
+11 -1
View File
@@ -28,6 +28,9 @@ import {
} from "./server-utils.js";
import { sanitizeRemoteExecutionEnv } from "./remote-execution-env.js";
import { preferredShellForSandbox, shellCommandArgs } from "./sandbox-shell.js";
import type { RuntimeProgressSink } from "./runtime-progress.js";
export type { RuntimeProgressSink } from "./runtime-progress.js";
export interface AdapterLocalExecutionTarget {
kind: "local";
@@ -70,7 +73,7 @@ export interface PreparedAdapterExecutionTargetRuntime {
workspaceRemoteDir: string | null;
runtimeRootDir: string | null;
assetDirs: Record<string, string>;
restoreWorkspace(): Promise<void>;
restoreWorkspace(onProgress?: RuntimeProgressSink): Promise<void>;
}
export interface AdapterExecutionTargetProcessOptions {
@@ -930,6 +933,11 @@ export async function prepareAdapterExecutionTargetRuntime(input: {
installCommand?: string | null;
/** When provided alongside `installCommand`, skip the install if the binary is already on PATH. */
detectCommand?: string | null;
// Optional progress sink for the workspace/asset upload. The returned
// `restoreWorkspace(onProgress?)` accepts its own sink for teardown. Both are
// forwarded down to the transport so the sandbox/SSH children can attach byte
// counters without further changes here.
onProgress?: RuntimeProgressSink;
}): Promise<PreparedAdapterExecutionTargetRuntime> {
const target = input.target ?? { kind: "local" as const };
if (target.kind === "local") {
@@ -950,6 +958,7 @@ export async function prepareAdapterExecutionTargetRuntime(input: {
workspaceLocalDir: input.workspaceLocalDir,
workspaceRemoteDir: input.workspaceRemoteDir,
assets: input.assets,
onProgress: input.onProgress,
});
return {
target,
@@ -980,6 +989,7 @@ export async function prepareAdapterExecutionTargetRuntime(input: {
assets: input.assets,
installCommand: input.installCommand,
detectCommand: input.detectCommand,
onProgress: input.onProgress,
});
return {
target,
+9
View File
@@ -61,6 +61,15 @@ export {
redactCommandText,
} from "./command-redaction.js";
export { buildSandboxNpmInstallCommand } from "./sandbox-install-command.js";
export { createRuntimeProgressReporter } from "./runtime-progress.js";
export type {
RuntimeProgressSink,
RuntimeProgressPhase,
RuntimeProgressDirection,
RuntimeProgressTarget,
RuntimeProgressReporter,
RuntimeProgressReporterOptions,
} from "./runtime-progress.js";
export { inferOpenAiCompatibleBiller } from "./billing.js";
// Keep the root adapter-utils entry browser-safe because the UI imports it.
// The sandbox callback bridge stays available via its dedicated subpath export.
@@ -6,6 +6,7 @@ import {
syncDirectoryToSsh,
} from "./ssh.js";
import { captureDirectorySnapshot } from "./workspace-restore-merge.js";
import type { RuntimeProgressSink } from "./runtime-progress.js";
export interface RemoteManagedRuntimeAsset {
key: string;
@@ -20,7 +21,7 @@ export interface PreparedRemoteManagedRuntime {
workspaceRemoteDir: string;
runtimeRootDir: string;
assetDirs: Record<string, string>;
restoreWorkspace(): Promise<void>;
restoreWorkspace(onProgress?: RuntimeProgressSink): Promise<void>;
}
function asObject(value: unknown): Record<string, unknown> {
@@ -69,6 +70,9 @@ export async function prepareRemoteManagedRuntime(input: {
workspaceLocalDir: string;
workspaceRemoteDir?: string;
assets?: RemoteManagedRuntimeAsset[];
// Upload progress sink. Threaded for the byte-counting transport rewrite; the
// child task wires it into the workspace/asset transfers.
onProgress?: RuntimeProgressSink;
}): Promise<PreparedRemoteManagedRuntime> {
const baseWorkspaceRemoteDir = input.workspaceRemoteDir ?? input.spec.remoteCwd;
const workspaceRemoteDir = path.posix.join(
@@ -84,6 +88,7 @@ export async function prepareRemoteManagedRuntime(input: {
spec: input.spec,
localDir: input.workspaceLocalDir,
remoteDir: workspaceRemoteDir,
onProgress: input.onProgress,
});
const restoreExclude = preparedWorkspace.gitBacked ? [".git", ".paperclip-runtime"] : [".paperclip-runtime"];
const baselineSnapshot = await captureDirectorySnapshot(input.workspaceLocalDir, {
@@ -101,6 +106,8 @@ export async function prepareRemoteManagedRuntime(input: {
remoteDir,
followSymlinks: asset.followSymlinks,
exclude: asset.exclude,
onProgress: input.onProgress,
progressLabel: asset.key,
});
}
} catch (error) {
@@ -110,6 +117,7 @@ export async function prepareRemoteManagedRuntime(input: {
remoteDir: workspaceRemoteDir,
baselineSnapshot,
restoreGitHistory: preparedWorkspace.gitBacked,
onProgress: input.onProgress,
});
throw error;
}
@@ -120,13 +128,14 @@ export async function prepareRemoteManagedRuntime(input: {
workspaceRemoteDir,
runtimeRootDir,
assetDirs,
restoreWorkspace: async () => {
restoreWorkspace: async (onProgress?: RuntimeProgressSink) => {
await restoreWorkspaceFromSshExecution({
spec: input.spec,
localDir: input.workspaceLocalDir,
remoteDir: workspaceRemoteDir,
baselineSnapshot,
restoreGitHistory: preparedWorkspace.gitBacked,
onProgress,
});
},
};
@@ -0,0 +1,281 @@
import { describe, expect, it } from "vitest";
import { createRuntimeProgressReporter } from "./runtime-progress.js";
const MB = 1024 * 1024;
function makeClock(start = 0) {
let value = start;
return {
now: () => value,
advance: (ms: number) => {
value += ms;
},
};
}
describe("createRuntimeProgressReporter", () => {
it("formats the message with phase, label, direction, target, percent and MB", async () => {
const lines: string[] = [];
const reporter = createRuntimeProgressReporter({
sink: (line) => {
lines.push(line);
},
phase: "Syncing",
label: "workspace",
direction: "to",
target: "sandbox",
});
await reporter.report(12.6 * MB, 31.4 * MB);
expect(lines).toEqual(["[paperclip] Syncing workspace to sandbox: 40% (12.6/31.4 MB)\n"]);
});
it("omits the label when none is provided (e.g. git history)", async () => {
const lines: string[] = [];
const reporter = createRuntimeProgressReporter({
sink: (line) => {
lines.push(line);
},
phase: "Importing git history",
direction: "to",
target: "ssh",
});
await reporter.report(4 * MB, 4 * MB);
expect(lines).toEqual(["[paperclip] Importing git history to ssh: 100% (4.0/4.0 MB)\n"]);
});
it("suppresses intermediate emits that neither cross a step nor exceed the interval", async () => {
const clock = makeClock();
const lines: string[] = [];
const reporter = createRuntimeProgressReporter({
sink: (line) => {
lines.push(line);
},
phase: "Syncing",
label: "workspace",
direction: "to",
target: "sandbox",
now: clock.now,
});
// First report always emits (step 0 crossed).
await reporter.report(1 * MB, 100 * MB); // 1%
// Still within the same 10% step and under 2s -> suppressed.
await reporter.report(2 * MB, 100 * MB); // 2%
await reporter.report(5 * MB, 100 * MB); // 5%
expect(lines).toHaveLength(1);
expect(lines[0]).toBe("[paperclip] Syncing workspace to sandbox: 1% (1.0/100.0 MB)\n");
});
it("emits when the percentage crosses a 10% step", async () => {
const clock = makeClock();
const lines: string[] = [];
const reporter = createRuntimeProgressReporter({
sink: (line) => {
lines.push(line);
},
phase: "Syncing",
label: "workspace",
direction: "to",
target: "sandbox",
now: clock.now,
});
await reporter.report(1 * MB, 100 * MB); // 1% -> emit (step 0)
await reporter.report(15 * MB, 100 * MB); // 15% -> crosses into step 1 -> emit
expect(lines).toHaveLength(2);
expect(lines[1]).toBe("[paperclip] Syncing workspace to sandbox: 15% (15.0/100.0 MB)\n");
});
it("emits on the time threshold even without a step crossing", async () => {
const clock = makeClock();
const lines: string[] = [];
const reporter = createRuntimeProgressReporter({
sink: (line) => {
lines.push(line);
},
phase: "Syncing",
label: "workspace",
direction: "to",
target: "sandbox",
now: clock.now,
});
await reporter.report(1 * MB, 100 * MB); // emit
await reporter.report(2 * MB, 100 * MB); // suppressed (same step, no time elapsed)
clock.advance(2000);
await reporter.report(3 * MB, 100 * MB); // 3% same step, but 2s elapsed -> emit
expect(lines).toHaveLength(2);
expect(lines[1]).toBe("[paperclip] Syncing workspace to sandbox: 3% (3.0/100.0 MB)\n");
});
it("always emits the terminal 100% line via report reaching the total", async () => {
const clock = makeClock();
const lines: string[] = [];
const reporter = createRuntimeProgressReporter({
sink: (line) => {
lines.push(line);
},
phase: "Restoring",
label: "workspace",
direction: "from",
target: "sandbox",
now: clock.now,
});
await reporter.report(1 * MB, 100 * MB); // emit
await reporter.report(100 * MB, 100 * MB); // terminal -> always emit
expect(lines[lines.length - 1]).toBe(
"[paperclip] Restoring workspace from sandbox: 100% (100.0/100.0 MB)\n",
);
});
it("complete() emits the terminal 100% line even when intermediate emits were throttled", async () => {
const clock = makeClock();
const lines: string[] = [];
const reporter = createRuntimeProgressReporter({
sink: (line) => {
lines.push(line);
},
phase: "Syncing",
label: "workspace",
direction: "to",
target: "sandbox",
now: clock.now,
});
await reporter.report(1 * MB, 100 * MB); // emit
await reporter.report(5 * MB, 100 * MB); // suppressed
await reporter.complete();
expect(lines[lines.length - 1]).toBe(
"[paperclip] Syncing workspace to sandbox: 100% (100.0/100.0 MB)\n",
);
});
it("complete() is idempotent and does not double-emit after a terminal report", async () => {
const lines: string[] = [];
const reporter = createRuntimeProgressReporter({
sink: (line) => {
lines.push(line);
},
phase: "Syncing",
label: "workspace",
direction: "to",
target: "sandbox",
});
await reporter.report(100 * MB, 100 * MB); // terminal
await reporter.complete();
await reporter.complete();
expect(lines).toHaveLength(1);
});
it("reports MB-only (no percent) when the total is unknown, plus a completion line", async () => {
const clock = makeClock();
const lines: string[] = [];
const reporter = createRuntimeProgressReporter({
sink: (line) => {
lines.push(line);
},
phase: "Restoring",
label: "workspace",
direction: "from",
target: "ssh",
now: clock.now,
});
await reporter.report(2 * MB, null); // first emit
await reporter.report(4 * MB, null); // suppressed (no time elapsed)
clock.advance(2000);
await reporter.report(8 * MB, null); // time elapsed -> emit
await reporter.complete();
expect(lines).toEqual([
"[paperclip] Restoring workspace from ssh: 2.0 MB\n",
"[paperclip] Restoring workspace from ssh: 8.0 MB\n",
"[paperclip] Restoring workspace from ssh: 8.0 MB\n",
]);
});
it("fail() emits a terminal failure marker with the last percent instead of a dangling line", async () => {
const clock = makeClock();
const lines: string[] = [];
const reporter = createRuntimeProgressReporter({
sink: (line) => {
lines.push(line);
},
phase: "Syncing",
label: "workspace",
direction: "to",
target: "ssh",
now: clock.now,
});
await reporter.report(40 * MB, 100 * MB); // emit at 40%
await reporter.fail();
expect(lines).toEqual([
"[paperclip] Syncing workspace to ssh: 40% (40.0/100.0 MB)\n",
"[paperclip] Syncing workspace to ssh: failed at 40% (40.0/100.0 MB)\n",
]);
});
it("fail() falls back to an MB marker when the total is unknown", async () => {
const lines: string[] = [];
const reporter = createRuntimeProgressReporter({
sink: (line) => {
lines.push(line);
},
phase: "Restoring",
direction: "from",
target: "ssh",
});
await reporter.report(3 * MB, null);
await reporter.fail();
expect(lines.at(-1)).toBe("[paperclip] Restoring from ssh: failed after 3.0 MB\n");
});
it("fail() is suppressed after a terminal completion and complete() after a failure", async () => {
const lines: string[] = [];
const reporter = createRuntimeProgressReporter({
sink: (line) => {
lines.push(line);
},
phase: "Syncing",
label: "workspace",
direction: "to",
target: "ssh",
});
await reporter.report(100 * MB, 100 * MB); // terminal complete
await reporter.fail(); // suppressed — already completed
expect(lines).toHaveLength(1);
const failLines: string[] = [];
const failed = createRuntimeProgressReporter({
sink: (line) => {
failLines.push(line);
},
phase: "Syncing",
label: "workspace",
direction: "to",
target: "ssh",
});
await failed.report(20 * MB, 100 * MB);
await failed.fail();
await failed.complete(); // suppressed — already failed
expect(failLines).toHaveLength(2);
expect(failLines.at(-1)).toContain("failed at 20%");
});
});
@@ -0,0 +1,152 @@
// Shared, throttled progress reporting for execution-target sync/restore.
//
// Transports (sandbox / SSH) own the byte counting and call `report()` as bytes
// move; orchestrators own the per-phase label and direction. The reporter
// throttles emits so a long transfer doesn't flood the log: a line is emitted
// only when the percentage crosses a step boundary (default every 10%) or once
// at least `minIntervalMs` has elapsed since the last emit. The terminal
// completion line is always emitted via `complete()` (or when `report()` reaches
// the known total).
/** A sink for fully-formatted progress lines (newline included). */
export type RuntimeProgressSink = (line: string) => void | Promise<void>;
export type RuntimeProgressPhase =
| "Syncing"
| "Restoring"
| "Importing git history"
| "Exporting git history";
export type RuntimeProgressDirection = "to" | "from";
export type RuntimeProgressTarget = "sandbox" | "ssh";
export interface RuntimeProgressReporterOptions {
sink: RuntimeProgressSink;
phase: RuntimeProgressPhase;
/** Optional per-phase label, e.g. "workspace" or an asset key. */
label?: string;
direction: RuntimeProgressDirection;
target: RuntimeProgressTarget;
/** Emit when the percentage crosses this step. Default 10. */
stepPercent?: number;
/** Emit when at least this many ms have elapsed since the last emit. Default 2000. */
minIntervalMs?: number;
/** Injectable clock for deterministic tests. Default `Date.now`. */
now?: () => number;
}
export interface RuntimeProgressReporter {
/**
* Report progress. Throttled: only emits on a step crossing or after
* `minIntervalMs`. When `totalBytes` is known and `doneBytes` reaches it, the
* terminal 100% line is emitted and the reporter is marked complete.
*/
report(doneBytes: number, totalBytes: number | null): Promise<void>;
/**
* Emit the terminal completion line if it hasn't been emitted yet. Idempotent.
*/
complete(doneBytes?: number, totalBytes?: number | null): Promise<void>;
/**
* Emit a terminal failure line if no terminal line has been emitted yet, so a
* failed transfer leaves an explicit marker instead of a dangling percentage.
* Idempotent and mutually exclusive with `complete()`.
*/
fail(doneBytes?: number, totalBytes?: number | null): Promise<void>;
}
const BYTES_PER_MB = 1024 * 1024;
function formatMb(bytes: number): string {
return (Math.max(0, bytes) / BYTES_PER_MB).toFixed(1);
}
function clampPercent(value: number): number {
if (!Number.isFinite(value)) return 0;
return Math.min(100, Math.max(0, Math.round(value)));
}
export function createRuntimeProgressReporter(
options: RuntimeProgressReporterOptions,
): RuntimeProgressReporter {
const stepPercent = options.stepPercent && options.stepPercent > 0 ? options.stepPercent : 10;
const minIntervalMs =
options.minIntervalMs && options.minIntervalMs > 0 ? options.minIntervalMs : 2000;
const now = options.now ?? Date.now;
const prefix = `[paperclip] ${options.phase}${options.label ? ` ${options.label}` : ""} ${options.direction} ${options.target}`;
let lastEmitAt: number | null = null;
let lastStep = -1;
let lastDoneBytes = 0;
let lastTotalBytes: number | null = null;
let completed = false;
function buildLine(doneBytes: number, totalBytes: number | null): string {
if (totalBytes != null && totalBytes > 0) {
const pct = clampPercent((doneBytes / totalBytes) * 100);
return `${prefix}: ${pct}% (${formatMb(doneBytes)}/${formatMb(totalBytes)} MB)\n`;
}
return `${prefix}: ${formatMb(doneBytes)} MB\n`;
}
function buildFailLine(doneBytes: number, totalBytes: number | null): string {
if (totalBytes != null && totalBytes > 0) {
const pct = clampPercent((doneBytes / totalBytes) * 100);
return `${prefix}: failed at ${pct}% (${formatMb(doneBytes)}/${formatMb(totalBytes)} MB)\n`;
}
return `${prefix}: failed after ${formatMb(doneBytes)} MB\n`;
}
async function emit(doneBytes: number, totalBytes: number | null): Promise<void> {
lastEmitAt = now();
if (totalBytes != null && totalBytes > 0) {
lastStep = Math.floor(((doneBytes / totalBytes) * 100) / stepPercent);
}
await options.sink(buildLine(doneBytes, totalBytes));
}
return {
async report(doneBytes, totalBytes) {
lastDoneBytes = doneBytes;
lastTotalBytes = totalBytes;
if (completed) return;
const elapsedOk = lastEmitAt == null || now() - lastEmitAt >= minIntervalMs;
if (totalBytes != null && totalBytes > 0) {
const terminal = doneBytes >= totalBytes;
const step = Math.floor(((doneBytes / totalBytes) * 100) / stepPercent);
const stepOk = step > lastStep;
if (terminal || stepOk || elapsedOk) {
await emit(doneBytes, totalBytes);
}
if (terminal) completed = true;
return;
}
// Unknown total: no step boundaries, throttle purely on elapsed time.
if (elapsedOk) {
await emit(doneBytes, totalBytes);
}
},
async complete(doneBytes, totalBytes) {
if (completed) return;
completed = true;
const total = totalBytes !== undefined ? totalBytes : lastTotalBytes;
const done =
doneBytes !== undefined
? doneBytes
: total != null && total > 0
? total
: lastDoneBytes;
await options.sink(buildLine(done, total));
},
async fail(doneBytes, totalBytes) {
if (completed) return;
completed = true;
const total = totalBytes !== undefined ? totalBytes : lastTotalBytes;
const done = doneBytes !== undefined ? doneBytes : lastDoneBytes;
await options.sink(buildFailLine(done, total));
},
};
}
@@ -197,6 +197,89 @@ describe("sandbox managed runtime", () => {
await expect(readFile(path.join(remoteWorkspaceDir, "src", "main.ts"), "utf8")).resolves.toBe("x\n");
});
it("emits throttled, labeled upload and restore progress with direction and percentages", async () => {
const rootDir = await mkdtemp(path.join(os.tmpdir(), "paperclip-sandbox-progress-"));
cleanupDirs.push(rootDir);
const localWorkspaceDir = path.join(rootDir, "local-workspace");
const remoteWorkspaceDir = path.join(rootDir, "remote-workspace");
const localAssetsDir = path.join(rootDir, "local-assets");
await mkdir(localWorkspaceDir, { recursive: true });
await mkdir(localAssetsDir, { recursive: true });
await writeFile(path.join(localWorkspaceDir, "README.md"), "workspace\n", "utf8");
await writeFile(path.join(localAssetsDir, "skill.md"), "skill\n", "utf8");
// Drive byte progress in 100 fine (1%) increments so the throttle has many
// chances to emit; the reporter must collapse them to ~one line per 10% step.
const driveProgress = async (
total: number,
onProgress: ((done: number, total: number | null) => void | Promise<void>) | undefined,
) => {
if (!onProgress) return;
for (let i = 1; i <= 100; i++) {
await onProgress(Math.floor((total * i) / 100), total);
}
};
const client: SandboxManagedRuntimeClient = {
makeDir: async (remotePath) => {
await mkdir(remotePath, { recursive: true });
},
writeFile: async (remotePath, bytes, options) => {
await mkdir(path.dirname(remotePath), { recursive: true });
const buffer = Buffer.from(bytes);
await writeFile(remotePath, buffer);
await driveProgress(buffer.byteLength, options?.onProgress);
},
readFile: async (remotePath, options) => {
const buffer = await readFile(remotePath);
await driveProgress(buffer.byteLength, options?.onProgress);
return buffer;
},
listFiles: async () => [],
remove: async (remotePath) => {
await rm(remotePath, { recursive: true, force: true });
},
run: async (command) => {
await execFile("sh", ["-c", command], { maxBuffer: 32 * 1024 * 1024 });
},
};
const lines: string[] = [];
const prepared = await prepareSandboxManagedRuntime({
spec: {
transport: "sandbox",
provider: "test",
sandboxId: "sandbox-1",
remoteCwd: remoteWorkspaceDir,
timeoutMs: 30_000,
apiKey: null,
},
adapterKey: "test-adapter",
client,
workspaceLocalDir: localWorkspaceDir,
assets: [{ key: "skills", localDir: localAssetsDir }],
onProgress: (line) => {
lines.push(line);
},
});
const uploadWorkspaceLines = lines.filter((line) => line.includes("Syncing workspace to sandbox"));
const uploadAssetLines = lines.filter((line) => line.includes("Syncing skills to sandbox"));
expect(uploadWorkspaceLines.length).toBeGreaterThan(0);
expect(uploadAssetLines.length).toBeGreaterThan(0);
// 100 reported increments must be throttled to at most ~one line per 10% step.
expect(uploadWorkspaceLines.length).toBeLessThanOrEqual(11);
// Reaches 100% and shows the MB breakdown.
expect(uploadWorkspaceLines.some((line) => line.includes("100%"))).toBe(true);
expect(uploadWorkspaceLines.every((line) => /\(\d+\.\d\/\d+\.\d MB\)/.test(line))).toBe(true);
await prepared.restoreWorkspace();
const restoreLines = lines.filter((line) => line.includes("Restoring workspace from sandbox"));
expect(restoreLines.length).toBeGreaterThan(0);
expect(restoreLines.length).toBeLessThanOrEqual(11);
expect(restoreLines.some((line) => line.includes("100%"))).toBe(true);
});
it("creates a valid empty workspace tarball when the local workspace is empty", async () => {
const rootDir = await mkdtemp(path.join(os.tmpdir(), "paperclip-sandbox-empty-"));
cleanupDirs.push(rootDir);
@@ -4,6 +4,7 @@ import os from "node:os";
import path from "node:path";
import { promisify } from "node:util";
import { captureDirectorySnapshot, mergeDirectoryWithBaseline } from "./workspace-restore-merge.js";
import { createRuntimeProgressReporter, type RuntimeProgressSink } from "./runtime-progress.js";
const execFile = promisify(execFileCallback);
@@ -23,10 +24,23 @@ export interface SandboxManagedRuntimeAsset {
exclude?: string[];
}
/**
* Per-call byte-level progress hook. `transferredBytes`/`totalBytes` are decoded
* file bytes (not the base64 wire size). `totalBytes` is null when the size is
* not known up front. The transport is the source of truth for byte counts; the
* orchestrator owns the phase label and direction.
*/
export interface SandboxTransferProgressOptions {
onProgress?: (transferredBytes: number, totalBytes: number | null) => void | Promise<void>;
}
export interface SandboxManagedRuntimeClient {
makeDir(remotePath: string): Promise<void>;
writeFile(remotePath: string, bytes: ArrayBuffer): Promise<void>;
readFile(remotePath: string): Promise<Buffer | Uint8Array | ArrayBuffer>;
writeFile(remotePath: string, bytes: ArrayBuffer, options?: SandboxTransferProgressOptions): Promise<void>;
readFile(
remotePath: string,
options?: SandboxTransferProgressOptions,
): Promise<Buffer | Uint8Array | ArrayBuffer>;
listFiles(remotePath: string): Promise<string[]>;
remove(remotePath: string): Promise<void>;
run(command: string, options: { timeoutMs: number }): Promise<void>;
@@ -38,7 +52,7 @@ export interface PreparedSandboxManagedRuntime {
workspaceRemoteDir: string;
runtimeRootDir: string;
assetDirs: Record<string, string>;
restoreWorkspace(): Promise<void>;
restoreWorkspace(onProgress?: RuntimeProgressSink): Promise<void>;
}
function asObject(value: unknown): Record<string, unknown> {
@@ -261,6 +275,36 @@ function tarExcludeFlags(exclude: string[] | undefined): string {
return ["._*", ...(exclude ?? [])].map((entry) => `--exclude ${shellQuote(entry)}`).join(" ");
}
// Bridge a single byte-level transfer to the throttled progress reporter. The
// transport reports decoded bytes via `options.onProgress`; the reporter turns
// them into a throttled, fully-formatted log line. `finish()` emits the terminal
// completion line (idempotent) once the transfer returns.
function makeTransferProgress(
sink: RuntimeProgressSink | undefined,
phase: "Syncing" | "Restoring",
direction: "to" | "from",
label: string,
): { options: SandboxTransferProgressOptions | undefined; finish: () => Promise<void> } {
if (!sink) return { options: undefined, finish: async () => {} };
const reporter = createRuntimeProgressReporter({
sink,
phase,
direction,
target: "sandbox",
label,
});
return {
options: {
onProgress: async (transferredBytes, totalBytes) => {
await reporter.report(transferredBytes, totalBytes);
},
},
finish: async () => {
await reporter.complete();
},
};
}
export async function prepareSandboxManagedRuntime(input: {
spec: SandboxRemoteExecutionSpec;
adapterKey: string;
@@ -270,6 +314,9 @@ export async function prepareSandboxManagedRuntime(input: {
workspaceExclude?: string[];
preserveAbsentOnRestore?: string[];
assets?: SandboxManagedRuntimeAsset[];
// Upload progress sink. Threaded for the byte-counting transport rewrite; the
// child task wires it into writeFile/readFile.
onProgress?: RuntimeProgressSink;
}): Promise<PreparedSandboxManagedRuntime> {
const workspaceRemoteDir = input.workspaceRemoteDir ?? input.spec.remoteCwd;
const runtimeRootDir = path.posix.join(workspaceRemoteDir, ".paperclip-runtime", input.adapterKey);
@@ -287,7 +334,13 @@ export async function prepareSandboxManagedRuntime(input: {
const workspaceTarBytes = await fs.readFile(workspaceTarPath);
const remoteWorkspaceTar = path.posix.join(runtimeRootDir, "workspace-upload.tar");
await input.client.makeDir(runtimeRootDir);
await input.client.writeFile(remoteWorkspaceTar, toArrayBuffer(workspaceTarBytes));
const workspaceUpload = makeTransferProgress(input.onProgress, "Syncing", "to", "workspace");
await input.client.writeFile(
remoteWorkspaceTar,
toArrayBuffer(workspaceTarBytes),
workspaceUpload.options,
);
await workspaceUpload.finish();
const preservedNames = new Set([".paperclip-runtime", ...(input.preserveAbsentOnRestore ?? [])]);
const findPreserveArgs = [...preservedNames].map((entry) => `! -name ${shellQuote(entry)}`).join(" ");
await input.client.run(
@@ -311,7 +364,9 @@ export async function prepareSandboxManagedRuntime(input: {
const assetTarBytes = await fs.readFile(assetTarPath);
const remoteAssetDir = path.posix.join(runtimeRootDir, asset.key);
const remoteAssetTar = path.posix.join(runtimeRootDir, `${asset.key}-upload.tar`);
await input.client.writeFile(remoteAssetTar, toArrayBuffer(assetTarBytes));
const assetUpload = makeTransferProgress(input.onProgress, "Syncing", "to", asset.key);
await input.client.writeFile(remoteAssetTar, toArrayBuffer(assetTarBytes), assetUpload.options);
await assetUpload.finish();
await input.client.run(
`sh -c ${shellQuote(
`rm -rf ${shellQuote(remoteAssetDir)} && ` +
@@ -334,7 +389,8 @@ export async function prepareSandboxManagedRuntime(input: {
workspaceRemoteDir,
runtimeRootDir,
assetDirs,
restoreWorkspace: async () => {
restoreWorkspace: async (onProgress?: RuntimeProgressSink) => {
const restoreSink = onProgress ?? input.onProgress;
await withTempDir("paperclip-sandbox-restore-", async (tempDir) => {
const remoteWorkspaceTar = path.posix.join(runtimeRootDir, "workspace-download.tar");
await input.client.run(
@@ -345,7 +401,9 @@ export async function prepareSandboxManagedRuntime(input: {
)}`,
{ timeoutMs: input.spec.timeoutMs },
);
const archiveBytes = await input.client.readFile(remoteWorkspaceTar);
const workspaceRestore = makeTransferProgress(restoreSink, "Restoring", "from", "workspace");
const archiveBytes = await input.client.readFile(remoteWorkspaceTar, workspaceRestore.options);
await workspaceRestore.finish();
await input.client.remove(remoteWorkspaceTar).catch(() => undefined);
const localArchivePath = path.join(tempDir, "workspace.tar");
const extractedDir = path.join(tempDir, "workspace");
@@ -11,6 +11,7 @@ import {
readSshEnvLabFixtureStatus,
restoreWorkspaceFromSshExecution,
runSshCommand,
syncDirectoryFromSsh,
syncDirectoryToSsh,
startSshEnvLabFixture,
stopSshEnvLabFixture,
@@ -54,6 +55,31 @@ async function startSshEnvLabFixtureOrSkip(statePath: string, label: string) {
}
}
interface ParsedProgressLine {
raw: string;
percent: number | null;
doneMb: number | null;
totalMb: number | null;
}
function parseProgressLine(line: string): ParsedProgressLine {
const trimmed = line.trimEnd();
const percentMatch = trimmed.match(/:\s*(\d+)%\s*\(([\d.]+)\/([\d.]+) MB\)$/);
if (percentMatch) {
return {
raw: trimmed,
percent: Number.parseInt(percentMatch[1]!, 10),
doneMb: Number.parseFloat(percentMatch[2]!),
totalMb: Number.parseFloat(percentMatch[3]!),
};
}
const mbMatch = trimmed.match(/:\s*([\d.]+) MB$/);
if (mbMatch) {
return { raw: trimmed, percent: null, doneMb: Number.parseFloat(mbMatch[1]!), totalMb: null };
}
return { raw: trimmed, percent: null, doneMb: null, totalMb: null };
}
describe("ssh env-lab fixture", () => {
const cleanupDirs: string[] = [];
@@ -199,6 +225,142 @@ describe("ssh env-lab fixture", () => {
expect(result.stdout).not.toContain("appledouble-present");
}, SSH_FIXTURE_TEST_TIMEOUT_MS);
it("reports throttled upload progress with a clamped percent and terminal 100% line", async () => {
const rootDir = await mkdtemp(path.join(os.tmpdir(), "paperclip-ssh-fixture-"));
cleanupDirs.push(rootDir);
const statePath = path.join(rootDir, "state.json");
const localDir = path.join(rootDir, "local-overlay");
await mkdir(localDir, { recursive: true });
// Multiple files large enough that tar emits several pipe chunks, so the
// byte counter crosses several step boundaries before the stream closes.
for (let index = 0; index < 4; index += 1) {
await writeFile(path.join(localDir, `blob-${index}.bin`), Buffer.alloc(256 * 1024, index + 1));
}
const started = await startSshEnvLabFixtureOrSkip(statePath, "SSH upload progress test");
if (!started) return;
const config = await buildSshEnvLabFixtureConfig(started);
const remoteDir = path.posix.join(started.workspaceDir, "overlay-progress");
const lines: ParsedProgressLine[] = [];
await syncDirectoryToSsh({
spec: { ...config, remoteCwd: started.workspaceDir },
localDir,
remoteDir,
onProgress: (line) => {
lines.push(parseProgressLine(line));
},
progressLabel: "workspace",
});
expect(lines.length).toBeGreaterThan(0);
for (const line of lines) {
expect(line.raw).toContain("Syncing workspace to ssh");
}
// Monotonically increasing byte counts.
const doneSeries = lines.map((line) => line.doneMb ?? 0);
for (let index = 1; index < doneSeries.length; index += 1) {
expect(doneSeries[index]!).toBeGreaterThanOrEqual(doneSeries[index - 1]!);
}
// Percent clamped to <= 99% on every line emitted before the stream closed.
for (const line of lines.slice(0, -1)) {
if (line.percent != null) expect(line.percent).toBeLessThanOrEqual(99);
}
// Terminal completion line is 100% with matching done/total.
const last = lines.at(-1)!;
expect(last.percent).toBe(100);
expect(last.doneMb).toBe(last.totalMb);
}, SSH_FIXTURE_TEST_TIMEOUT_MS);
it("reports restore progress with a terminal completion line", async () => {
const rootDir = await mkdtemp(path.join(os.tmpdir(), "paperclip-ssh-fixture-"));
cleanupDirs.push(rootDir);
const statePath = path.join(rootDir, "state.json");
const localDir = path.join(rootDir, "local-overlay");
const restoreDir = path.join(rootDir, "restore-target");
await mkdir(localDir, { recursive: true });
await mkdir(restoreDir, { recursive: true });
for (let index = 0; index < 4; index += 1) {
await writeFile(path.join(localDir, `blob-${index}.bin`), Buffer.alloc(256 * 1024, index + 1));
}
const started = await startSshEnvLabFixtureOrSkip(statePath, "SSH restore progress test");
if (!started) return;
const config = await buildSshEnvLabFixtureConfig(started);
const spec = { ...config, remoteCwd: started.workspaceDir } as const;
const remoteDir = path.posix.join(started.workspaceDir, "restore-source");
await syncDirectoryToSsh({ spec, localDir, remoteDir });
const lines: ParsedProgressLine[] = [];
await syncDirectoryFromSsh({
spec,
remoteDir,
localDir: restoreDir,
onProgress: (line) => {
lines.push(parseProgressLine(line));
},
progressLabel: "workspace",
});
expect(lines.length).toBeGreaterThan(0);
for (const line of lines) {
expect(line.raw).toContain("Restoring workspace from ssh");
}
// Terminal completion line: either an exact 100% (probe succeeded) or a
// final MB-received line (probe unavailable). Either is a valid terminal.
const last = lines.at(-1)!;
expect(last.percent === 100 || (last.percent === null && last.doneMb !== null)).toBe(true);
// The restored files round-tripped through the byte-counting transport.
await expect(readFile(path.join(restoreDir, "blob-0.bin"))).resolves.toEqual(
Buffer.alloc(256 * 1024, 1),
);
}, SSH_FIXTURE_TEST_TIMEOUT_MS);
it("reports exact git-history import percentage from the known bundle size", async () => {
const rootDir = await mkdtemp(path.join(os.tmpdir(), "paperclip-ssh-fixture-"));
cleanupDirs.push(rootDir);
const statePath = path.join(rootDir, "state.json");
const localRepo = path.join(rootDir, "local-workspace");
await mkdir(localRepo, { recursive: true });
await git(localRepo, ["init"]);
await git(localRepo, ["checkout", "-b", "main"]);
await git(localRepo, ["config", "user.name", "Paperclip Test"]);
await git(localRepo, ["config", "user.email", "test@paperclip.dev"]);
await writeFile(path.join(localRepo, "tracked.bin"), Buffer.alloc(256 * 1024, 7));
await git(localRepo, ["add", "tracked.bin"]);
await git(localRepo, ["commit", "-m", "initial"]);
const started = await startSshEnvLabFixtureOrSkip(statePath, "SSH git import progress test");
if (!started) return;
const config = await buildSshEnvLabFixtureConfig(started);
const spec = { ...config, remoteCwd: started.workspaceDir } as const;
const lines: ParsedProgressLine[] = [];
await prepareWorkspaceForSshExecution({
spec,
localDir: localRepo,
remoteDir: started.workspaceDir,
onProgress: (line) => {
lines.push(parseProgressLine(line));
},
});
const importLines = lines.filter((line) => line.raw.includes("Importing git history to ssh"));
expect(importLines.length).toBeGreaterThan(0);
// Known bundle size -> exact percentage with no "workspace" label.
for (const line of importLines) {
expect(line.raw).not.toContain("workspace");
expect(line.percent).not.toBeNull();
}
const lastImport = importLines.at(-1)!;
expect(lastImport.percent).toBe(100);
expect(lastImport.doneMb).toBe(lastImport.totalMb);
}, SSH_FIXTURE_TEST_TIMEOUT_MS);
it("can dereference local symlinks while syncing to the remote fixture", async () => {
const rootDir = await mkdtemp(path.join(os.tmpdir(), "paperclip-ssh-fixture-"));
cleanupDirs.push(rootDir);
+295 -16
View File
@@ -4,10 +4,17 @@ import { constants as fsConstants, createReadStream, createWriteStream, promises
import net from "node:net";
import os from "node:os";
import path from "node:path";
import { Transform } from "node:stream";
import type { CommandManagedRuntimeRunner } from "./command-managed-runtime.js";
import type { RunProcessResult } from "./server-utils.js";
import type { DirectorySnapshot } from "./workspace-restore-merge.js";
import { mergeDirectoryWithBaseline } from "./workspace-restore-merge.js";
import {
createRuntimeProgressReporter,
type RuntimeProgressDirection,
type RuntimeProgressPhase,
type RuntimeProgressSink,
} from "./runtime-progress.js";
export interface SshConnectionConfig {
host: string;
@@ -409,6 +416,152 @@ function tarSpawnEnv(): NodeJS.ProcessEnv {
};
}
// Converts a tar `--exclude` pattern into a regexp for the local-size estimate.
// We only need approximate fidelity here (the estimate feeds a clamped percent),
// so we support the literal names and `*`/`?` globs used in practice.
function tarPatternToRegExp(pattern: string): RegExp {
const escaped = pattern
.replace(/[.+^${}()|[\]\\]/g, "\\$&")
.replace(/\*/g, "[^/]*")
.replace(/\?/g, "[^/]");
return new RegExp(`^${escaped}$`);
}
// Walks `localDir` summing regular-file sizes, mirroring tar's `--exclude`
// handling (plus the implicit `._*`) and `followSymlinks` so the to-ssh upload
// can report an estimated total before tar finishes producing the stream.
async function estimateLocalDirSize(input: {
localDir: string;
exclude?: string[];
followSymlinks?: boolean;
}): Promise<number> {
const regexes = ["._*", ...(input.exclude ?? [])].map(tarPatternToRegExp);
const isExcluded = (relPath: string, base: string) =>
regexes.some((regex) => regex.test(relPath) || regex.test(base));
let total = 0;
const walk = async (dir: string, relative: string): Promise<void> => {
const entries = await fs.readdir(dir, { withFileTypes: true }).catch(() => []);
for (const entry of entries) {
const entryRelative = relative ? `${relative}/${entry.name}` : entry.name;
if (isExcluded(entryRelative, entry.name)) continue;
const full = path.join(dir, entry.name);
const stats = await (input.followSymlinks ? fs.stat(full) : fs.lstat(full)).catch(() => null);
if (!stats) continue;
if (stats.isDirectory()) {
await walk(full, entryRelative);
} else if (stats.isFile()) {
total += stats.size;
}
}
};
await walk(input.localDir, "");
return total;
}
// Best-effort remote size probe for the from-ssh restore. `du -sk` is POSIX and
// available on the BSD/Linux remotes we target; it over-counts (block-rounded,
// includes excluded dirs) which keeps the reported percent safely below 100
// until the stream actually closes. Returns null when unavailable so the caller
// falls back to MB-received mode.
async function probeRemoteDirSize(input: {
spec: SshConnectionConfig;
remoteDir: string;
}): Promise<number | null> {
try {
const result = await runSshScript(
input.spec,
`du -sk ${shellQuote(input.remoteDir)} 2>/dev/null | cut -f1`,
{ timeoutMs: 15_000, maxBuffer: 16 * 1024 },
);
const kilobytes = Number.parseInt(result.stdout.trim(), 10);
return Number.isFinite(kilobytes) && kilobytes > 0 ? kilobytes * 1024 : null;
} catch {
return null;
}
}
interface TransferProgress {
// Backpressure-respecting counter to splice into a transport pipe.
counter: Transform;
// Last cumulative byte count observed by the counter.
transferred: () => number;
// Emit the terminal completion line. Idempotent.
finish: () => Promise<void>;
// Emit a terminal failure marker instead of a completion line. Idempotent.
fail: () => Promise<void>;
}
// Wraps a throttled progress reporter behind a counting Transform so transports
// can `source.pipe(progress.counter).pipe(dest)`. When `totalBytes` is a known
// exact size (e.g. a git bundle) the reporter emits an exact percentage. When it
// is an estimate (tar upload / remote probe) we clamp the reported bytes to 99%
// of the estimate so an inaccurate total never shows a premature 100%; `finish`
// then emits the terminal 100% (or, in MB-only mode, the final MB) line.
//
// `totalBytes` may be a promise so an expensive size estimate (a local dir walk
// or a remote `du` probe) runs concurrently with the transfer instead of
// blocking the pipe from opening. Until it resolves the counter reports bytes in
// MB-only mode, then adopts the percentage once the total is known; `finish`
// awaits the estimate so the terminal 100% line is still guaranteed.
function createTransferProgress(input: {
onProgress: RuntimeProgressSink;
phase: RuntimeProgressPhase;
direction: RuntimeProgressDirection;
label?: string;
totalBytes: number | null | Promise<number | null>;
estimated: boolean;
}): TransferProgress {
const reporter = createRuntimeProgressReporter({
sink: input.onProgress,
phase: input.phase,
direction: input.direction,
label: input.label,
target: "ssh",
});
let total: number | null = null;
let cap: number | null = null;
const applyTotal = (value: number | null) => {
total = value != null && value > 0 ? value : null;
cap = total != null && input.estimated ? Math.floor(total * 0.99) : null;
};
const totalReady: Promise<void> =
input.totalBytes != null && typeof (input.totalBytes as Promise<number | null>).then === "function"
? (input.totalBytes as Promise<number | null>).then(applyTotal, () => applyTotal(null))
: (applyTotal(input.totalBytes as number | null), Promise.resolve());
let transferred = 0;
let chain: Promise<void> = Promise.resolve();
const enqueue = (work: () => Promise<void>) => {
chain = chain.then(work).catch(() => undefined);
};
const counter = new Transform({
transform(chunk: Buffer, _encoding, callback) {
transferred += chunk.length;
const reported = cap != null ? Math.min(transferred, cap) : transferred;
const totalSnapshot = total;
enqueue(() => reporter.report(reported, totalSnapshot));
callback(null, chunk);
},
});
return {
counter,
transferred: () => transferred,
finish: async () => {
await chain.catch(() => undefined);
await totalReady.catch(() => undefined);
await reporter.complete(total != null ? total : transferred, total).catch(() => undefined);
},
fail: async () => {
await chain.catch(() => undefined);
await reporter.fail(transferred, total).catch(() => undefined);
},
};
}
async function runSshScript(
config: SshConnectionConfig,
script: string,
@@ -493,6 +646,7 @@ async function streamLocalFileToSsh(input: {
spec: SshConnectionConfig;
localFile: string;
remoteScript: string;
progress?: TransferProgress;
}): Promise<void> {
const auth = await createSshAuthArgs(input.spec);
const sshArgs = [
@@ -525,7 +679,12 @@ async function streamLocalFileToSsh(input: {
});
source.on("error", fail);
ssh.on("error", fail);
source.pipe(ssh.stdin ?? null);
if (input.progress) {
input.progress.counter.on("error", fail);
source.pipe(input.progress.counter).pipe(ssh.stdin ?? null);
} else {
source.pipe(ssh.stdin ?? null);
}
ssh.on("close", (code) => {
if (settled) return;
settled = true;
@@ -542,6 +701,7 @@ async function streamSshToLocalFile(input: {
spec: SshConnectionConfig;
remoteScript: string;
localFile: string;
progress?: TransferProgress;
}): Promise<void> {
const auth = await createSshAuthArgs(input.spec);
const sshArgs = [
@@ -569,7 +729,12 @@ async function streamSshToLocalFile(input: {
reject(error);
};
ssh.stdout?.pipe(sink);
if (input.progress) {
input.progress.counter.on("error", fail);
ssh.stdout?.pipe(input.progress.counter).pipe(sink);
} else {
ssh.stdout?.pipe(sink);
}
ssh.stderr?.on("data", (chunk) => {
sshStderr += String(chunk);
});
@@ -594,6 +759,7 @@ async function importGitWorkspaceToSsh(input: {
localDir: string;
remoteDir: string;
snapshot: LocalGitWorkspaceSnapshot;
onProgress?: RuntimeProgressSink;
}): Promise<void> {
const bundleDir = await fs.mkdtemp(path.join(os.tmpdir(), "paperclip-ssh-bundle-"));
const bundlePath = path.join(bundleDir, "workspace.bundle");
@@ -628,11 +794,31 @@ async function importGitWorkspaceToSsh(input: {
`git -C ${shellQuote(input.remoteDir)} update-ref -d ${shellQuote(tempRef)} >/dev/null 2>&1 || true`,
].join("\n");
await streamLocalFileToSsh({
spec: input.spec,
localFile: bundlePath,
remoteScript: remoteSetupScript,
});
// The git bundle is a real local file of known size, so report an exact
// percentage. No `workspace` label: the "Importing git history" phase is
// already self-describing in the log line.
const progress = input.onProgress
? createTransferProgress({
onProgress: input.onProgress,
phase: "Importing git history",
direction: "to",
totalBytes: (await fs.stat(bundlePath)).size,
estimated: false,
})
: null;
try {
await streamLocalFileToSsh({
spec: input.spec,
localFile: bundlePath,
remoteScript: remoteSetupScript,
progress: progress ?? undefined,
});
await progress?.finish();
} catch (error) {
await progress?.fail();
throw error;
}
} finally {
await runLocalGit(input.localDir, ["update-ref", "-d", tempRef], {
timeout: 10_000,
@@ -648,6 +834,7 @@ async function exportGitWorkspaceFromSsh(input: {
localDir: string;
importedRef?: string;
resetLocalWorkspace?: boolean;
onProgress?: RuntimeProgressSink;
}): Promise<string> {
const bundleDir = await fs.mkdtemp(path.join(os.tmpdir(), "paperclip-ssh-bundle-"));
const bundlePath = path.join(bundleDir, "workspace.bundle");
@@ -665,11 +852,30 @@ async function exportGitWorkspaceFromSsh(input: {
'cat "$tmp_bundle"',
].join("\n");
await streamSshToLocalFile({
spec: input.spec,
remoteScript: exportScript,
localFile: bundlePath,
});
// The remote bundle size isn't known before streaming, so report bytes
// received (MB mode) with a terminal completion line.
const progress = input.onProgress
? createTransferProgress({
onProgress: input.onProgress,
phase: "Exporting git history",
direction: "from",
totalBytes: null,
estimated: false,
})
: null;
try {
await streamSshToLocalFile({
spec: input.spec,
remoteScript: exportScript,
localFile: bundlePath,
progress: progress ?? undefined,
});
await progress?.finish();
} catch (error) {
await progress?.fail();
throw error;
}
await runLocalGit(input.localDir, ["fetch", "--force", bundlePath, `refs/paperclip/ssh-sync/export:${importedRef}`], {
timeout: 60_000,
@@ -1049,6 +1255,8 @@ export async function syncDirectoryToSsh(input: {
remoteDir: string;
exclude?: string[];
followSymlinks?: boolean;
onProgress?: RuntimeProgressSink;
progressLabel?: string;
}): Promise<void> {
const auth = await createSshAuthArgs(input.spec);
const sshArgs = [
@@ -1059,7 +1267,27 @@ export async function syncDirectoryToSsh(input: {
`sh -c ${shellQuote(`mkdir -p ${shellQuote(input.remoteDir)} && tar -xf - -C ${shellQuote(input.remoteDir)}`)}`,
];
await new Promise<void>((resolve, reject) => {
// tar's archive size isn't known until tar finishes, so estimate it from the
// local file sizes and clamp the reported percent to 99% until the pipe closes.
// The estimate walk runs concurrently with the transfer so it never delays the
// pipe from opening on large workspaces.
const progress = input.onProgress
? createTransferProgress({
onProgress: input.onProgress,
phase: "Syncing",
direction: "to",
label: input.progressLabel,
totalBytes: estimateLocalDirSize({
localDir: input.localDir,
exclude: input.exclude,
followSymlinks: input.followSymlinks,
}),
estimated: true,
})
: null;
try {
await new Promise<void>((resolve, reject) => {
const tarArgs = [
...(input.followSymlinks ? ["-h"] : []),
"-C",
@@ -1111,7 +1339,12 @@ export async function syncDirectoryToSsh(input: {
reject(error);
};
tar.stdout?.pipe(ssh.stdin ?? null);
if (progress) {
progress.counter.on("error", fail);
tar.stdout?.pipe(progress.counter).pipe(ssh.stdin ?? null);
} else {
tar.stdout?.pipe(ssh.stdin ?? null);
}
tar.stderr?.on("data", (chunk) => {
tarStderr += String(chunk);
});
@@ -1131,7 +1364,12 @@ export async function syncDirectoryToSsh(input: {
sshExitCode = code;
maybeFinish();
});
}).finally(auth.cleanup);
}).finally(auth.cleanup);
await progress?.finish();
} catch (error) {
await progress?.fail();
throw error;
}
}
export async function syncDirectoryFromSsh(input: {
@@ -1140,6 +1378,8 @@ export async function syncDirectoryFromSsh(input: {
localDir: string;
exclude?: string[];
preserveLocalEntries?: string[];
onProgress?: RuntimeProgressSink;
progressLabel?: string;
}): Promise<void> {
const auth = await createSshAuthArgs(input.spec);
const stagingDir = await fs.mkdtemp(path.join(os.tmpdir(), "paperclip-ssh-sync-back-"));
@@ -1155,6 +1395,21 @@ export async function syncDirectoryFromSsh(input: {
`sh -c ${shellQuote(remoteTarScript)}`,
];
// The remote tar size isn't known locally, so probe the remote directory for
// an estimate (clamped to 99%). The probe runs concurrently with the transfer
// so its round-trip never delays the restore; when it is unavailable we report
// bytes received in MB mode with a terminal completion line.
const progress = input.onProgress
? createTransferProgress({
onProgress: input.onProgress,
phase: "Restoring",
direction: "from",
label: input.progressLabel,
totalBytes: probeRemoteDirSize({ spec: input.spec, remoteDir: input.remoteDir }),
estimated: true,
})
: null;
try {
await new Promise<void>((resolve, reject) => {
const ssh = spawn("ssh", sshArgs, {
@@ -1195,7 +1450,12 @@ export async function syncDirectoryFromSsh(input: {
reject(error);
};
ssh.stdout?.pipe(tar.stdin ?? null);
if (progress) {
progress.counter.on("error", fail);
ssh.stdout?.pipe(progress.counter).pipe(tar.stdin ?? null);
} else {
ssh.stdout?.pipe(tar.stdin ?? null);
}
ssh.stderr?.on("data", (chunk) => {
sshStderr += String(chunk);
});
@@ -1216,9 +1476,13 @@ export async function syncDirectoryFromSsh(input: {
maybeFinish();
});
});
await progress?.finish();
await clearLocalDirectory(input.localDir, input.preserveLocalEntries);
await copyDirectoryContents(stagingDir, input.localDir);
} catch (error) {
await progress?.fail();
throw error;
} finally {
await fs.rm(stagingDir, { recursive: true, force: true }).catch(() => undefined);
await auth.cleanup();
@@ -1229,6 +1493,7 @@ export async function prepareWorkspaceForSshExecution(input: {
spec: SshRemoteExecutionSpec;
localDir: string;
remoteDir?: string;
onProgress?: RuntimeProgressSink;
}): Promise<{ gitBacked: boolean }> {
const remoteDir = input.remoteDir ?? input.spec.remoteCwd;
const gitSnapshot = await readLocalGitWorkspaceSnapshot(input.localDir);
@@ -1239,12 +1504,15 @@ export async function prepareWorkspaceForSshExecution(input: {
localDir: input.localDir,
remoteDir,
snapshot: gitSnapshot,
onProgress: input.onProgress,
});
await syncDirectoryToSsh({
spec: input.spec,
localDir: input.localDir,
remoteDir,
exclude: [".git", ".paperclip-runtime"],
onProgress: input.onProgress,
progressLabel: "workspace",
});
await removeDeletedPathsOnSsh({
spec: input.spec,
@@ -1264,6 +1532,8 @@ export async function prepareWorkspaceForSshExecution(input: {
localDir: input.localDir,
remoteDir,
exclude: [".paperclip-runtime"],
onProgress: input.onProgress,
progressLabel: "workspace",
});
return { gitBacked: false };
}
@@ -1274,6 +1544,7 @@ export async function restoreWorkspaceFromSshExecution(input: {
remoteDir?: string;
baselineSnapshot?: DirectorySnapshot;
restoreGitHistory?: boolean;
onProgress?: RuntimeProgressSink;
}): Promise<void> {
const remoteDir = input.remoteDir ?? input.spec.remoteCwd;
if (input.baselineSnapshot) {
@@ -1289,6 +1560,7 @@ export async function restoreWorkspaceFromSshExecution(input: {
localDir: input.localDir,
importedRef: importedRef ?? undefined,
resetLocalWorkspace: false,
onProgress: input.onProgress,
})
: null;
await syncDirectoryFromSsh({
@@ -1296,6 +1568,8 @@ export async function restoreWorkspaceFromSshExecution(input: {
remoteDir,
localDir: stagingDir,
exclude: input.baselineSnapshot.exclude,
onProgress: input.onProgress,
progressLabel: "workspace",
});
await mergeDirectoryWithBaseline({
baseline: input.baselineSnapshot,
@@ -1330,6 +1604,7 @@ export async function restoreWorkspaceFromSshExecution(input: {
spec: input.spec,
remoteDir,
localDir: input.localDir,
onProgress: input.onProgress,
});
await syncDirectoryFromSsh({
spec: input.spec,
@@ -1337,6 +1612,8 @@ export async function restoreWorkspaceFromSshExecution(input: {
localDir: input.localDir,
exclude: [".git", ".paperclip-runtime"],
preserveLocalEntries: [".git"],
onProgress: input.onProgress,
progressLabel: "workspace",
});
return;
}
@@ -1346,6 +1623,8 @@ export async function restoreWorkspaceFromSshExecution(input: {
remoteDir,
localDir: input.localDir,
exclude: [".paperclip-runtime"],
onProgress: input.onProgress,
progressLabel: "workspace",
});
}
@@ -485,6 +485,7 @@ export async function execute(ctx: AdapterExecutionContext): Promise<AdapterExec
workspaceLocalDir: cwd,
installCommand: SANDBOX_INSTALL_COMMAND,
detectCommand: command,
onProgress: (line) => onLog("stdout", line),
assets: [
{
key: "skills",
@@ -523,7 +524,7 @@ export async function execute(ctx: AdapterExecutionContext): Promise<AdapterExec
executionCwd: effectiveExecutionCwd,
});
const restoreRemoteWorkspace = preparedExecutionTargetRuntime
? () => preparedExecutionTargetRuntime.restoreWorkspace()
? () => preparedExecutionTargetRuntime.restoreWorkspace((line) => onLog("stdout", line))
: null;
const effectivePromptBundleAddDir = executionTargetIsRemote
? preparedExecutionTargetRuntime?.assetDirs.skills ??
@@ -397,6 +397,7 @@ export async function execute(ctx: AdapterExecutionContext): Promise<AdapterExec
workspaceLocalDir: cwd,
installCommand: SANDBOX_INSTALL_COMMAND,
detectCommand: command,
onProgress: (line) => onLog("stdout", line),
assets: [
{
key: "home",
@@ -414,7 +415,7 @@ export async function execute(ctx: AdapterExecutionContext): Promise<AdapterExec
const executionTargetIsSandbox =
runtimeExecutionTarget?.kind === "remote" && runtimeExecutionTarget.transport === "sandbox";
const restoreRemoteWorkspace = preparedExecutionTargetRuntime
? () => preparedExecutionTargetRuntime.restoreWorkspace()
? () => preparedExecutionTargetRuntime.restoreWorkspace((line) => onLog("stdout", line))
: null;
let paperclipBridge: Awaited<ReturnType<typeof startAdapterExecutionTargetPaperclipBridge>> = null;
const remoteCodexHome = executionTargetIsRemote
@@ -364,13 +364,15 @@ export async function execute(ctx: AdapterExecutionContext): Promise<AdapterExec
workspaceLocalDir: cwd,
installCommand: SANDBOX_INSTALL_COMMAND,
detectCommand: command,
onProgress: (line) => onLog("stdout", line),
assets: [{
key: "skills",
localDir: localSkillsDir,
followSymlinks: true,
}],
});
restoreRemoteWorkspace = () => preparedExecutionTargetRuntime.restoreWorkspace();
restoreRemoteWorkspace = () =>
preparedExecutionTargetRuntime.restoreWorkspace((line) => onLog("stdout", line));
effectiveExecutionCwd = preparedExecutionTargetRuntime.workspaceRemoteDir ?? effectiveExecutionCwd;
refreshPaperclipWorkspaceEnvForExecution({
env,
@@ -340,13 +340,15 @@ export async function execute(ctx: AdapterExecutionContext): Promise<AdapterExec
workspaceLocalDir: cwd,
installCommand: SANDBOX_INSTALL_COMMAND,
detectCommand: command,
onProgress: (line) => onLog("stdout", line),
assets: [{
key: "skills",
localDir: localSkillsDir,
followSymlinks: true,
}],
});
restoreRemoteWorkspace = () => preparedExecutionTargetRuntime.restoreWorkspace();
restoreRemoteWorkspace = () =>
preparedExecutionTargetRuntime.restoreWorkspace((line) => onLog("stdout", line));
effectiveExecutionCwd = preparedExecutionTargetRuntime.workspaceRemoteDir ?? effectiveExecutionCwd;
refreshPaperclipWorkspaceEnvForExecution({
env,
@@ -325,8 +325,10 @@ export async function execute(ctx: AdapterExecutionContext): Promise<AdapterExec
timeoutSec,
installCommand: ctx.runtimeCommandSpec?.installCommand ?? null,
detectCommand: ctx.runtimeCommandSpec?.detectCommand ?? command,
onProgress: (line) => onLog("stdout", line),
});
restoreRemoteWorkspace = () => preparedExecutionTargetRuntime.restoreWorkspace();
restoreRemoteWorkspace = () =>
preparedExecutionTargetRuntime.restoreWorkspace((line) => onLog("stdout", line));
effectiveExecutionCwd = preparedExecutionTargetRuntime.workspaceRemoteDir ?? effectiveExecutionCwd;
refreshPaperclipWorkspaceEnvForExecution({
env,
@@ -377,6 +377,7 @@ export async function execute(ctx: AdapterExecutionContext): Promise<AdapterExec
workspaceLocalDir: cwd,
installCommand: SANDBOX_INSTALL_COMMAND,
detectCommand: command,
onProgress: (line) => onLog("stdout", line),
assets: [
{
key: "skills",
@@ -391,7 +392,8 @@ export async function execute(ctx: AdapterExecutionContext): Promise<AdapterExec
: []),
],
});
restoreRemoteWorkspace = () => preparedExecutionTargetRuntime.restoreWorkspace();
restoreRemoteWorkspace = () =>
preparedExecutionTargetRuntime.restoreWorkspace((line) => onLog("stdout", line));
effectiveExecutionCwd = preparedExecutionTargetRuntime.workspaceRemoteDir ?? effectiveExecutionCwd;
refreshPaperclipWorkspaceEnvForExecution({
env: preparedRuntimeConfig.env,
@@ -417,6 +417,7 @@ export async function execute(ctx: AdapterExecutionContext): Promise<AdapterExec
workspaceLocalDir: cwd,
installCommand: SANDBOX_INSTALL_COMMAND,
detectCommand: command,
onProgress: (line) => onLog("stdout", line),
assets: [
{
key: "skills",
@@ -431,7 +432,8 @@ export async function execute(ctx: AdapterExecutionContext): Promise<AdapterExec
: []),
],
});
restoreRemoteWorkspace = () => preparedRemoteRuntime.restoreWorkspace();
restoreRemoteWorkspace = () =>
preparedRemoteRuntime.restoreWorkspace((line) => onLog("stdout", line));
effectiveExecutionCwd = preparedRemoteRuntime.workspaceRemoteDir ?? effectiveExecutionCwd;
refreshPaperclipWorkspaceEnvForExecution({
env,
@@ -75,6 +75,7 @@ export async function resolveEnvironmentExecutionTarget(input: {
timeoutMs,
runner: input.environmentRuntime && input.lease
? {
supportsSingleStreamStdinProgress: false,
execute: async (commandInput) => {
const startedAt = new Date().toISOString();
const result = await input.environmentRuntime!.execute({