mirror of
https://github.com/stablyai/orca.git
synced 2026-09-22 16:02:32 +00:00
* fix(windows): drop EDR-flagged -ExecutionPolicy Bypass from encoded PowerShell
MDE flags `-ExecutionPolicy Bypass` paired with base64 `-EncodedCommand` as a
behavioural signal. Measured on Windows 11: neither `-Command` nor
`-EncodedCommand` is execution-policy gated (both run under an explicit
`-ExecutionPolicy Restricted` and `AllSigned`; only `-File` fails), so the
switch was a pure no-op on every one of these command lines.
Removes the switch from all four sites that spelled it, and de-encodes the one
site whose payload never passes through a re-parsing shell:
- ssh-remote-powershell: one chokepoint for ~40 remote-Windows call sites.
Base64 kept — the remote sshd DefaultShell re-parses this string.
- setup-agent-sequencing / windows-cmd-runner-delayed-launch: base64 kept —
these strings are typed into a terminal pane.
- windows-interactive-login-spawn: base64 kept — `cmd.exe /c start` re-parses,
and the cmd-safe-token guard rejects the `&` and `"` in the raw relay script.
- windows-mobile-firewall local runner: `-EncodedCommand` -> `-Command`, since
execFile reaches CreateProcess with no shell in between.
The setup startup gate keeps execution-policy relief in-payload (process scope),
because it evals a user-authored startup command that may invoke a `.ps1`, and a
`.ps1` IS gated. Caught by the real-process suite; mirrors the agent-hooks
launcher's trade.
The elevated firewall child deliberately stays encoded: `Start-Process
-ArgumentList` joins its array into one ShellExecuteEx string without quoting
and PowerShell re-splits on whitespace, measured to collapse `C:\My App\...`
to `C:\My App\...` — a firewall rule for the wrong program.
* test(ssh): enforce the no-script-file invariant remote payloads rely on
Dropping `-ExecutionPolicy Bypass` from `powerShellCommand` is a no-op only
while no remote payload loads a PowerShell script file — execution policy has
never gated anything else. That invariant held by inspection and was guarded by
nothing, so a future payload that dot-sourced, used `-File`, or imported a
`.psm1` would break only on a remote host with a Restricted/AllSigned
LocalMachine policy and no GPO: a failure on someone else's machine.
States the invariant at the wrapper, and adds a ratchet that scans every module
importing it for `.ps1`/`.psm1`, `Import-Module`, `-File`, and dot-sourcing.
The scan discovers importers itself (13 today) so new ones are covered, and
asserts it found some, so an emptied list cannot pass vacuously.
Mutation-checked: injecting each construct into a real importer fails the
matching case and names the file. The first dot-source pattern passed a
`;`-prefixed sample but missed `powerShellCommand(". '$x'")` — the likelier
shape — so the pattern now accepts a string-literal start and the self-test
samples carry their surrounding quotes.
* test(ssh): close two blind spots in the remote-payload ratchet
Both found by independent mutation testing of the ratchet itself, and both let
a real violation pass while the guard reported green.
`-File` was matched case-sensitively, so `-file $scriptVar` slipped through —
PowerShell switches are case-insensitive, and with a variable path the `.ps1`
pattern does not cover for it, so that shape escaped both nets. The naive fix
is wrong: bare /-File\b/i matches `--credential-file`, `--log-file` and
`--body-file`, which occur in three of these importers. Anchoring to a token
boundary catches the lowercase, odd-spacing and argv-element forms with zero
offenders across all 14.
Comment stripping paired a `/*` appearing inside a string (a glob such as
'src/*.ts') with any later comment close and deleted everything between, hiding
violations in the gap. Anchoring the block strip to line start, as the `//`
strip already was, fixes it — verified by injecting an `Import-Module` after a
glob string: the unanchored form misses it, the anchored form catches it.
Extends the same case-insensitivity to `.ps1`/`.psm1` and `Import-Module`,
which had the identical flaw (`import-module`, `DEPLOY.PS1` are legitimate
spellings); measured to add no false positive.
Each construct now carries the fixtures it must catch AND the near-misses it
must not, so a future tightening cannot quietly trade one for the other — the
negative fixtures are what would have caught the naive `-File` fix. Non-vacuity
bound tightened to >10 against 14 importers.
* docs(ssh): state what the remote-payload ratchet cannot see
The scan matches source text, so a script file reached only through a variable
(`& $scriptPath`) never appears in source and no pattern can catch it. The
ratchet narrows the hole; the invariant note on `powerShellCommand` covers the
remainder.
Recorded because a guard that reads as complete coverage when it is not is
worse than one that states its edge: the next author trusts it further than it
deserves, and should learn this limit from the test rather than an incident.
* test(ssh): scan remote payloads with the shared source walk
The ratchet had its own tree walk and comment stripper. The walk skipped
neither node_modules/dist/.git nor dot-directories and excluded tests by
`.test.ts` alone, so its importer count -- the guard's own goalpost -- could
be wrong about what it scanned. The stripper was anchored to line start to
dodge a `/*` inside a glob string, which silently skipped trailing comments;
`stripComments` tracks quote state and handles both.
Importer set re-derived against the shared walk: 15, floor unchanged at 10.
* fix(setup): report a failed execution-policy relief instead of swallowing it
The in-payload Set-ExecutionPolicy carried -ErrorAction SilentlyContinue and
an empty catch, so any failure vanished. A Windows PowerShell 5.1 install with
duplicate extended type data fails every cmdlet in Microsoft.PowerShell.Security
-- autoload, not policy -- and the user then saw only their own .ps1 being
refused, with no trace that the relief had been attempted or why.
-ErrorAction Stop is what routes a non-terminating failure into the catch at
all; the catch reports the FullyQualifiedErrorId to stderr and deliberately
does not rethrow, so a broken policy cmdlet cannot take down the startup this
gate exists to run. Success path is unchanged and stays stderr-clean.
Verified by execution on a clean child environment: success -> policy=Bypass,
stderr empty; shadowed failing cmdlet -> diagnostic on stderr and the gate
still continues; the old empty catch -> silent.
---------
Co-authored-by: Orca Worker <orca-worker@localhost>
668 lines
25 KiB
TypeScript
668 lines
25 KiB
TypeScript
import { spawn } from 'node:child_process'
|
|
import { chmodSync, mkdirSync, mkdtempSync, readFileSync, rmSync, writeFileSync } from 'node:fs'
|
|
import { join } from 'node:path'
|
|
import { tmpdir } from 'node:os'
|
|
|
|
import { afterEach, describe, expect, it, vi } from 'vitest'
|
|
|
|
import { getDefaultRepoHookSettings } from './constants'
|
|
import {
|
|
createSequencedSetupAgentCommands,
|
|
createSetupAgentSequenceNonce,
|
|
SETUP_COMPLETE_MESSAGE,
|
|
getSetupAgentSequenceShellForTests,
|
|
resolveSetupAgentSequenceLaunchCommand,
|
|
SETUP_AGENT_SEQUENCE_STARTUP_COMMAND_ENV,
|
|
SETUP_AGENT_SEQUENCE_STARTUP_SCRIPT_ENV
|
|
} from './setup-agent-sequencing'
|
|
import {
|
|
DEFAULT_SETUP_AGENT_STARTUP_POLICY,
|
|
shouldWaitForSetupBeforeAgentStartup
|
|
} from './setup-agent-startup-policy'
|
|
|
|
const TEMP_DIRS: string[] = []
|
|
const WINDOWS_PROCESS_TEST_TIMEOUT_MS = 30_000
|
|
|
|
afterEach(() => {
|
|
for (const dir of TEMP_DIRS.splice(0)) {
|
|
rmSync(dir, { recursive: true, force: true })
|
|
}
|
|
})
|
|
|
|
describe('createSequencedSetupAgentCommands', () => {
|
|
it('defaults agent startup to immediate unless the wait policy is explicit', () => {
|
|
expect(DEFAULT_SETUP_AGENT_STARTUP_POLICY).toBe('start-immediately')
|
|
expect(getDefaultRepoHookSettings().setupAgentStartupPolicy).toBe('start-immediately')
|
|
expect(shouldWaitForSetupBeforeAgentStartup(undefined)).toBe(false)
|
|
expect(shouldWaitForSetupBeforeAgentStartup('start-immediately')).toBe(false)
|
|
expect(shouldWaitForSetupBeforeAgentStartup('wait-for-setup')).toBe(true)
|
|
})
|
|
|
|
it('uses the original sequenced startup command as the launch hint when present', () => {
|
|
expect(
|
|
resolveSetupAgentSequenceLaunchCommand(
|
|
{ [SETUP_AGENT_SEQUENCE_STARTUP_COMMAND_ENV]: 'omp --resume' },
|
|
'powershell wait-wrapper'
|
|
)
|
|
).toBe('omp --resume')
|
|
expect(
|
|
resolveSetupAgentSequenceLaunchCommand(
|
|
{ [SETUP_AGENT_SEQUENCE_STARTUP_COMMAND_ENV]: ' ' },
|
|
'powershell wait-wrapper'
|
|
)
|
|
).toBe('powershell wait-wrapper')
|
|
})
|
|
|
|
it('wraps POSIX setup and startup commands with a matching nonce marker', () => {
|
|
const result = createSequencedSetupAgentCommands({
|
|
runnerScriptPath: '/repo/.git/orca/setup-runner.sh',
|
|
startupCommand: "codex 'fix bug'",
|
|
platform: 'posix',
|
|
nonce: 'nonce-123',
|
|
waitTimeoutSeconds: 9
|
|
})
|
|
|
|
expect(result.setupCommand).toMatch(/^bash -lc /)
|
|
expect(result.setupCommand).toContain('bash /repo/.git/orca/setup-runner.sh')
|
|
expect(result.setupCommand).toContain('printf')
|
|
expect(result.setupCommand).toContain('nonce-123 "$status"')
|
|
expect(result.setupCommand).toContain(
|
|
'mv -f /repo/.git/orca/setup-runner.sh.nonce-123.done.tmp'
|
|
)
|
|
const startupScript = result.startupEnv?.[SETUP_AGENT_SEQUENCE_STARTUP_SCRIPT_ENV]
|
|
expect(result.startupCommand).toBe(
|
|
`bash -lc 'eval "$${SETUP_AGENT_SEQUENCE_STARTUP_SCRIPT_ENV}"'`
|
|
)
|
|
expect(startupScript).toContain('deadline=$((SECONDS + 9))')
|
|
expect(startupScript).not.toContain('date +%s')
|
|
expect(startupScript).toContain('Waiting for setup to finish before starting agent...')
|
|
expect(startupScript).toContain('[ "$seen" = nonce-123 ]')
|
|
expect(startupScript).toContain(
|
|
'rm -f /repo/.git/orca/setup-runner.sh.nonce-123.done /repo/.git/orca/setup-runner.sh.nonce-123.done.tmp'
|
|
)
|
|
expect(startupScript).toContain('exec codex')
|
|
expect(startupScript).toContain('fix bug')
|
|
expect(result.startupEnv).toEqual(
|
|
expect.objectContaining({
|
|
[SETUP_AGENT_SEQUENCE_STARTUP_COMMAND_ENV]: "codex 'fix bug'",
|
|
[SETUP_AGENT_SEQUENCE_STARTUP_SCRIPT_ENV]: startupScript
|
|
})
|
|
)
|
|
})
|
|
|
|
it('announces success so the pane stops showing the waiting line', () => {
|
|
const commands = createSequencedSetupAgentCommands({
|
|
runnerScriptPath: '/repo/.git/orca/setup-runner.sh',
|
|
startupCommand: 'codex',
|
|
platform: 'posix',
|
|
nonce: 'nonce-1'
|
|
})
|
|
const script = commands.startupEnv?.ORCA_SEQUENCED_STARTUP_SCRIPT ?? ''
|
|
// Why ordering: `eval`/`exec` never returns, so a later message never renders.
|
|
expect(script.indexOf(SETUP_COMPLETE_MESSAGE)).toBeGreaterThan(-1)
|
|
expect(script.indexOf(SETUP_COMPLETE_MESSAGE)).toBeLessThan(
|
|
script.indexOf('eval "$ORCA_SEQUENCED_STARTUP_COMMAND"')
|
|
)
|
|
})
|
|
|
|
it('announces success on the native Windows gate too', () => {
|
|
const commands = createSequencedSetupAgentCommands({
|
|
runnerScriptPath: 'C:\\repo\\.git\\orca\\setup-runner.cmd',
|
|
platform: 'windows',
|
|
startupCommand: 'codex',
|
|
nonce: 'nonce-2'
|
|
})
|
|
const decoded = Buffer.from(
|
|
commands.startupCommand.split('-EncodedCommand ')[1] ?? '',
|
|
'base64'
|
|
).toString('utf16le')
|
|
expect(decoded).toContain(SETUP_COMPLETE_MESSAGE)
|
|
expect(decoded.indexOf(SETUP_COMPLETE_MESSAGE)).toBeLessThan(
|
|
decoded.indexOf('Invoke-Expression $startup')
|
|
)
|
|
})
|
|
|
|
it('leaves the failure and timeout messages as the only other outcomes', () => {
|
|
const script =
|
|
createSequencedSetupAgentCommands({
|
|
runnerScriptPath: '/repo/.git/orca/setup-runner.sh',
|
|
startupCommand: 'codex',
|
|
platform: 'posix',
|
|
nonce: 'nonce-3'
|
|
}).startupEnv?.ORCA_SEQUENCED_STARTUP_SCRIPT ?? ''
|
|
// Silence on success is what made a healthy worktree look stuck.
|
|
expect(script).toContain(SETUP_COMPLETE_MESSAGE)
|
|
expect(script).toContain('Setup failed; skipping agent startup.')
|
|
expect(script).toContain('Timed out waiting for setup before starting agent.')
|
|
})
|
|
|
|
it('keeps the POSIX terminal submission below the canonical input floor', () => {
|
|
const result = createSequencedSetupAgentCommands({
|
|
runnerScriptPath: `/repo/${'nested-worktree/'.repeat(100)}setup-runner.sh`,
|
|
startupCommand: 'codex',
|
|
platform: 'posix',
|
|
nonce: 'long-path'
|
|
})
|
|
|
|
expect(result.startupCommand.length).toBeLessThan(256)
|
|
expect(result.startupCommand).not.toContain('nested-worktree')
|
|
expect(result.startupEnv?.[SETUP_AGENT_SEQUENCE_STARTUP_SCRIPT_ENV]).toContain(
|
|
'nested-worktree'
|
|
)
|
|
})
|
|
|
|
it('uses launch-specific marker paths for overlapping setup gates', () => {
|
|
const first = createSequencedSetupAgentCommands({
|
|
runnerScriptPath: '/repo/.git/orca/setup-runner.sh',
|
|
startupCommand: 'claude',
|
|
platform: 'posix',
|
|
nonce: 'first-launch'
|
|
})
|
|
const second = createSequencedSetupAgentCommands({
|
|
runnerScriptPath: '/repo/.git/orca/setup-runner.sh',
|
|
startupCommand: 'codex',
|
|
platform: 'posix',
|
|
nonce: 'second-launch'
|
|
})
|
|
|
|
expect(first.setupCommand).toContain('/repo/.git/orca/setup-runner.sh.first-launch.done')
|
|
expect(first.startupEnv?.[SETUP_AGENT_SEQUENCE_STARTUP_SCRIPT_ENV]).toContain(
|
|
'/repo/.git/orca/setup-runner.sh.first-launch.done'
|
|
)
|
|
expect(second.setupCommand).toContain('/repo/.git/orca/setup-runner.sh.second-launch.done')
|
|
expect(second.startupEnv?.[SETUP_AGENT_SEQUENCE_STARTUP_SCRIPT_ENV]).toContain(
|
|
'/repo/.git/orca/setup-runner.sh.second-launch.done'
|
|
)
|
|
expect(first.setupCommand).not.toContain('/repo/.git/orca/setup-runner.sh.second-launch.done')
|
|
expect(second.setupCommand).not.toContain('/repo/.git/orca/setup-runner.sh.first-launch.done')
|
|
})
|
|
|
|
it('keeps simple POSIX startup commands eligible for exec when quoted text has separators', () => {
|
|
const result = createSequencedSetupAgentCommands({
|
|
runnerScriptPath: '/repo/.git/orca/setup-runner.sh',
|
|
startupCommand: "codex 'fix this; then test'",
|
|
platform: 'posix',
|
|
nonce: 'nonce-quoted',
|
|
waitTimeoutSeconds: 9
|
|
})
|
|
|
|
const startupScript = result.startupEnv?.[SETUP_AGENT_SEQUENCE_STARTUP_SCRIPT_ENV]
|
|
expect(startupScript).toContain("exec codex 'fix this; then test'")
|
|
expect(startupScript).not.toContain('eval codex')
|
|
})
|
|
|
|
it('preserves POSIX inline environment assignment startup commands', () => {
|
|
const result = createSequencedSetupAgentCommands({
|
|
runnerScriptPath: '/repo/.git/orca/setup-runner.sh',
|
|
startupCommand: 'FOO=bar claude',
|
|
platform: 'posix',
|
|
nonce: 'nonce-env',
|
|
waitTimeoutSeconds: 9
|
|
})
|
|
|
|
const startupScript = result.startupEnv?.[SETUP_AGENT_SEQUENCE_STARTUP_SCRIPT_ENV]
|
|
expect(startupScript).toContain('FOO=bar claude')
|
|
expect(startupScript).toContain('exit "$?"')
|
|
expect(startupScript).not.toContain('exec FOO=bar claude')
|
|
})
|
|
|
|
it('uses the converted Linux marker path for WSL UNC runners on Windows', () => {
|
|
const result = createSequencedSetupAgentCommands({
|
|
runnerScriptPath:
|
|
'\\\\wsl.localhost\\Ubuntu\\home\\jin\\repo\\.git\\worktrees\\feature\\orca\\setup-runner.sh',
|
|
startupCommand: 'claude',
|
|
platform: 'windows',
|
|
nonce: 'nonce-wsl'
|
|
})
|
|
|
|
expect(getSetupAgentSequenceShellForTests(resultPathWsl(), 'windows')).toBe('posix')
|
|
expect(result.setupCommand).toContain(
|
|
'bash /home/jin/repo/.git/worktrees/feature/orca/setup-runner.sh'
|
|
)
|
|
expect(result.setupCommand).toContain(
|
|
'/home/jin/repo/.git/worktrees/feature/orca/setup-runner.sh.nonce-wsl.done'
|
|
)
|
|
expect(result.setupCommand).not.toContain('wsl.localhost')
|
|
})
|
|
|
|
it('keeps remote POSIX runners in bash even from a Windows client', () => {
|
|
const result = createSequencedSetupAgentCommands({
|
|
runnerScriptPath: '/remote/repo/.git/worktrees/feature/orca/setup-runner.sh',
|
|
startupCommand: 'claude',
|
|
platform: 'windows',
|
|
nonce: 'nonce-remote'
|
|
})
|
|
|
|
expect(result.setupCommand).toContain(
|
|
'bash /remote/repo/.git/worktrees/feature/orca/setup-runner.sh'
|
|
)
|
|
expect(result.startupEnv?.[SETUP_AGENT_SEQUENCE_STARTUP_SCRIPT_ENV]).toContain(
|
|
'[ "$seen" = nonce-remote ]'
|
|
)
|
|
})
|
|
|
|
it('preserves WSL shell metadata when sequencing native Windows runners', () => {
|
|
const result = createSequencedSetupAgentCommands({
|
|
runnerScriptPath: 'C:\\repo\\.git\\orca\\setup-runner.sh',
|
|
startupCommand: 'claude',
|
|
platform: 'windows',
|
|
shell: { family: 'posix', executable: 'wsl.exe' },
|
|
nonce: 'nonce-wsl-shell'
|
|
})
|
|
|
|
expect(result.setupCommand).toContain('bash /mnt/c/repo/.git/orca/setup-runner.sh')
|
|
expect(result.setupCommand).toContain(
|
|
'/mnt/c/repo/.git/orca/setup-runner.sh.nonce-wsl-shell.done'
|
|
)
|
|
expect(result.startupEnv?.[SETUP_AGENT_SEQUENCE_STARTUP_SCRIPT_ENV]).toContain(
|
|
'/mnt/c/repo/.git/orca/setup-runner.sh.nonce-wsl-shell.done'
|
|
)
|
|
})
|
|
|
|
it('wraps native Windows runners in a cmd-pinned setup and startup gate', () => {
|
|
const result = createSequencedSetupAgentCommands({
|
|
runnerScriptPath: 'C:\\repo\\.git\\orca\\setup-runner.cmd',
|
|
startupCommand: "codex --model gpt-5 'fix !PATH! & test'",
|
|
platform: 'windows',
|
|
nonce: 'nonce-win',
|
|
waitTimeoutSeconds: 3
|
|
})
|
|
const setupPowerShell = decodePowerShellScript(result.setupCommand)
|
|
const startupPowerShell = decodePowerShellScript(result.startupCommand)
|
|
|
|
expect(result.setupCommand).toContain(
|
|
'powershell.exe -NoProfile -NonInteractive -EncodedCommand'
|
|
)
|
|
expect(setupPowerShell).toContain("$runner = 'C:\\repo\\.git\\orca\\setup-runner.cmd'")
|
|
expect(setupPowerShell).toContain('$nonce + ":" + $setupStatus')
|
|
expect(result.startupCommand.match(/powershell\.exe/g)).toHaveLength(1)
|
|
expect(result.startupCommand).toContain(
|
|
'powershell.exe -NoProfile -NonInteractive -EncodedCommand'
|
|
)
|
|
expect(startupPowerShell).toContain('AddSeconds(3)')
|
|
expect(startupPowerShell).toContain('Missing setup marker path.')
|
|
expect(startupPowerShell).toContain('Timed out waiting for setup before starting agent.')
|
|
expect(startupPowerShell).toContain('Setup failed; skipping agent startup.')
|
|
expect(startupPowerShell).toContain(
|
|
'Remove-Item -LiteralPath $marker, $tmp -Force -ErrorAction SilentlyContinue'
|
|
)
|
|
expect(result.startupCommand).not.toContain('%ERRORLEVEL%')
|
|
expect(startupPowerShell).toContain('Invoke-Expression')
|
|
expect(result.startupCommand).not.toContain('fix !PATH! & test')
|
|
expect(result.startupEnv).toEqual({
|
|
[SETUP_AGENT_SEQUENCE_STARTUP_COMMAND_ENV]: "codex --model gpt-5 'fix !PATH! & test'"
|
|
})
|
|
// Why: `-EncodedCommand` is not execution-policy gated — only `-File` is — so the switch
|
|
// was a no-op, and base64 beside `-ExecutionPolicy Bypass` is a heavily EDR-flagged shape.
|
|
// The base64 itself must stay: these strings are typed into a terminal pane.
|
|
expect(result.setupCommand).not.toMatch(/-ExecutionPolicy/i)
|
|
expect(result.startupCommand).not.toMatch(/-ExecutionPolicy/i)
|
|
// Why: dropping the switch alone would break a user startup command that invokes a
|
|
// `.ps1` — a `.ps1` IS policy gated even though `-EncodedCommand` is not. The relief
|
|
// moves into the payload, where it is not part of the flagged command-line shape.
|
|
expect(startupPowerShell).toContain(
|
|
'Set-ExecutionPolicy -Scope Process -ExecutionPolicy Bypass -Force -ErrorAction Stop'
|
|
)
|
|
// Why `-ErrorAction Stop` and a reporting catch: autoload can fail for reasons that are
|
|
// not about policy at all (a 5.1 install with duplicate extended type data fails every
|
|
// cmdlet in Microsoft.PowerShell.Security), and the old SilentlyContinue plus `catch {}`
|
|
// hid that -- the user saw only their own script being refused. The catch must report and
|
|
// must NOT rethrow, or a broken policy cmdlet would take the whole startup with it.
|
|
expect(startupPowerShell).not.toContain('catch {}')
|
|
expect(startupPowerShell).toMatch(/catch \{ \[Console\]::Error\.WriteLine\(/)
|
|
expect(startupPowerShell).toContain('$_.FullyQualifiedErrorId')
|
|
expect(startupPowerShell).not.toMatch(/catch \{[^}]*throw/)
|
|
// Why: the autoloaded module's progress record would otherwise corrupt this gate's stderr.
|
|
expect(startupPowerShell).toContain("$ProgressPreference = 'SilentlyContinue'")
|
|
expect(startupPowerShell).toContain('$ProgressPreference = $orcaProgress')
|
|
// The setup gate only ever launches a .cmd/.bat runner, so it needs no relief.
|
|
expect(setupPowerShell).not.toMatch(/Set-ExecutionPolicy/i)
|
|
})
|
|
|
|
it('launches a batch runner through the cmd launcher inside a Git Bash gate', () => {
|
|
// Regression (#6896): a Git Bash terminal with a batch setup script still gets a .cmd
|
|
// runner, and the gate must not hand that runner to bash. The gate itself stays POSIX
|
|
// because the Git Bash pane types it and quoted the startup command for bash.
|
|
const result = createSequencedSetupAgentCommands({
|
|
runnerScriptPath: 'C:\\repo\\.git\\orca\\setup-runner.cmd',
|
|
startupCommand: "claude 'fix the user'\\''s login'",
|
|
platform: 'windows',
|
|
shell: { family: 'posix' },
|
|
nonce: 'nonce-gitbash-cmd'
|
|
})
|
|
|
|
expect(result.setupCommand).toContain(
|
|
'powershell.exe -NoProfile -NonInteractive -EncodedCommand'
|
|
)
|
|
expect(result.setupCommand).not.toMatch(/bash\s+\S*setup-runner/)
|
|
expect(decodePowerShellScript(result.setupCommand)).toContain(
|
|
"$runner = 'C:\\repo\\.git\\orca\\setup-runner.cmd'"
|
|
)
|
|
// Why: PowerShell's `Invoke-Expression` cannot parse the POSIX `'\''` escaping a Git Bash
|
|
// pane produces, so the gate that evaluates the startup command must be bash.
|
|
expect(result.setupCommand).toMatch(/^bash -lc /)
|
|
expect(result.startupCommand).toMatch(/^bash -lc /)
|
|
expect(result.startupCommand).not.toContain('Invoke-Expression')
|
|
expect(result.startupEnv?.[SETUP_AGENT_SEQUENCE_STARTUP_SCRIPT_ENV]).toContain(
|
|
'eval "$ORCA_SEQUENCED_STARTUP_COMMAND"'
|
|
)
|
|
// Why: bash writes and reads the marker here, so it needs the /c/... form of the path.
|
|
expect(result.setupCommand).toContain(
|
|
'/c/repo/.git/orca/setup-runner.cmd.nonce-gitbash-cmd.done'
|
|
)
|
|
expect(result.startupEnv?.[SETUP_AGENT_SEQUENCE_STARTUP_SCRIPT_ENV]).toContain(
|
|
'/c/repo/.git/orca/setup-runner.cmd.nonce-gitbash-cmd.done'
|
|
)
|
|
})
|
|
|
|
it.skipIf(process.platform !== 'win32')(
|
|
'executes the native Windows setup-to-agent sequence through cmd.exe',
|
|
async () => {
|
|
const tempDir = join(makeTempDir(), 'path with spaces')
|
|
mkdirSync(tempDir)
|
|
const runnerScriptPath = join(tempDir, 'setup runner.cmd')
|
|
const startupScriptPath = join(tempDir, 'agent-startup.cmd')
|
|
const logPath = join(tempDir, 'sequence.log')
|
|
|
|
writeFileSync(
|
|
runnerScriptPath,
|
|
['@echo off', `>> "${logPath}" echo setup-done`, 'exit /b 0'].join('\r\n'),
|
|
'utf8'
|
|
)
|
|
writeFileSync(
|
|
startupScriptPath,
|
|
['@echo off', `>> "${logPath}" echo agent-start`, 'exit /b 0'].join('\r\n'),
|
|
'utf8'
|
|
)
|
|
|
|
const commands = createSequencedSetupAgentCommands({
|
|
runnerScriptPath,
|
|
startupCommand: `cmd.exe /d /c "${startupScriptPath}"`,
|
|
platform: 'windows',
|
|
nonce: 'windows-sequence',
|
|
waitTimeoutSeconds: 2
|
|
})
|
|
|
|
const setupExit = await waitForExit(
|
|
spawnWindowsCommand(tempDir, 'run-setup.cmd', commands.setupCommand)
|
|
)
|
|
expect(setupExit.code).toBe(0)
|
|
expect(readIfExists(`${runnerScriptPath}.windows-sequence.done`)).toBe(
|
|
'windows-sequence:0\r\n'
|
|
)
|
|
|
|
const startupExit = await waitForExit(
|
|
spawnWindowsCommand(
|
|
tempDir,
|
|
'run-startup.cmd',
|
|
commands.startupCommand,
|
|
commands.startupEnv
|
|
)
|
|
)
|
|
|
|
expect(startupExit.code).toBe(0)
|
|
expect(startupExit.stderr).toContain('Waiting for setup to finish before starting agent...')
|
|
expect(readFileSync(logPath, 'utf8')).toBe('setup-done\r\nagent-start\r\n')
|
|
},
|
|
WINDOWS_PROCESS_TEST_TIMEOUT_MS
|
|
)
|
|
|
|
it.skipIf(process.platform === 'win32')(
|
|
'ignores stale markers until the matching setup run finishes, even when startup launches first',
|
|
async () => {
|
|
const tempDir = makeTempDir()
|
|
const runnerScriptPath = join(tempDir, 'setup-runner.sh')
|
|
const startupScriptPath = join(tempDir, 'startup.sh')
|
|
const logPath = join(tempDir, 'sequence.log')
|
|
const markerPath = `${runnerScriptPath}.fresh-sequence.done`
|
|
|
|
writeExecutable(
|
|
runnerScriptPath,
|
|
[
|
|
'#!/bin/sh',
|
|
`printf 'setup-start\\n' >> ${quoteSh(logPath)}`,
|
|
'sleep 1',
|
|
`printf 'setup-done\\n' >> ${quoteSh(logPath)}`
|
|
].join('\n')
|
|
)
|
|
writeExecutable(
|
|
startupScriptPath,
|
|
['#!/bin/sh', `printf 'agent-start\\n' >> ${quoteSh(logPath)}`].join('\n')
|
|
)
|
|
writeFileSync(markerPath, 'stale:0\n', 'utf8')
|
|
|
|
const commands = createSequencedSetupAgentCommands({
|
|
runnerScriptPath,
|
|
startupCommand: `bash ${quoteSh(startupScriptPath)}`,
|
|
platform: 'posix',
|
|
nonce: 'fresh-sequence',
|
|
waitTimeoutSeconds: 5
|
|
})
|
|
|
|
const startupExitPromise = waitForExit(
|
|
spawn('bash', ['-lc', commands.startupCommand], {
|
|
stdio: 'pipe',
|
|
env: { ...process.env, ...commands.startupEnv }
|
|
})
|
|
)
|
|
await sleep(250)
|
|
expect(readIfExists(logPath)).toBe('')
|
|
expect(readFileSync(markerPath, 'utf8')).toBe('stale:0\n')
|
|
|
|
const setupExit = await waitForExit(
|
|
spawn('bash', ['-lc', commands.setupCommand], { stdio: 'pipe' })
|
|
)
|
|
expect(setupExit.code).toBe(0)
|
|
|
|
const startupExit = await startupExitPromise
|
|
expect(startupExit.code).toBe(0)
|
|
|
|
expect(readFileSync(logPath, 'utf8')).toBe('setup-start\nsetup-done\nagent-start\n')
|
|
expect(readIfExists(markerPath)).toBe('')
|
|
expect(readIfExists(`${markerPath}.tmp`)).toBe('')
|
|
}
|
|
)
|
|
|
|
it.skipIf(process.platform === 'win32')(
|
|
'runs compound POSIX startup cleanup commands after setup succeeds',
|
|
async () => {
|
|
const tempDir = makeTempDir()
|
|
const runnerScriptPath = join(tempDir, 'setup-runner.sh')
|
|
const logPath = join(tempDir, 'sequence.log')
|
|
|
|
writeExecutable(
|
|
runnerScriptPath,
|
|
['#!/bin/sh', `printf 'setup-done\\n' >> ${quoteSh(logPath)}`].join('\n')
|
|
)
|
|
|
|
const commands = createSequencedSetupAgentCommands({
|
|
runnerScriptPath,
|
|
startupCommand: `printf 'agent-start\\n' >> ${quoteSh(logPath)}; printf 'cleanup\\n' >> ${quoteSh(logPath)}`,
|
|
platform: 'posix',
|
|
nonce: 'compound-sequence',
|
|
waitTimeoutSeconds: 5
|
|
})
|
|
|
|
const setupExitPromise = waitForExit(
|
|
spawn('bash', ['-lc', commands.setupCommand], { stdio: 'pipe' })
|
|
)
|
|
const startupExit = await waitForExit(
|
|
spawn('bash', ['-lc', commands.startupCommand], {
|
|
stdio: 'pipe',
|
|
env: { ...process.env, ...commands.startupEnv }
|
|
})
|
|
)
|
|
const setupExit = await setupExitPromise
|
|
|
|
expect(setupExit.code).toBe(0)
|
|
expect(startupExit.code).toBe(0)
|
|
expect(readFileSync(logPath, 'utf8')).toBe('setup-done\nagent-start\ncleanup\n')
|
|
const startupScript = commands.startupEnv?.[SETUP_AGENT_SEQUENCE_STARTUP_SCRIPT_ENV]
|
|
expect(startupScript).toContain('eval')
|
|
expect(startupScript).not.toContain('exec printf')
|
|
}
|
|
)
|
|
|
|
it.skipIf(process.platform === 'win32')(
|
|
'prefers the env-provided startup command after setup succeeds',
|
|
async () => {
|
|
const tempDir = makeTempDir()
|
|
const runnerScriptPath = join(tempDir, 'setup-runner.sh')
|
|
const startupScriptPath = join(tempDir, 'startup.sh')
|
|
const logPath = join(tempDir, 'sequence.log')
|
|
|
|
writeExecutable(
|
|
runnerScriptPath,
|
|
['#!/bin/sh', `printf 'setup-done\\n' >> ${quoteSh(logPath)}`].join('\n')
|
|
)
|
|
writeExecutable(
|
|
startupScriptPath,
|
|
[
|
|
'#!/bin/sh',
|
|
'if [ "$FOO" = "bar" ]; then',
|
|
` printf 'env-start\\n' >> ${quoteSh(logPath)}`,
|
|
'fi'
|
|
].join('\n')
|
|
)
|
|
|
|
const commands = createSequencedSetupAgentCommands({
|
|
runnerScriptPath,
|
|
startupCommand: `printf 'inline-start\\n' >> ${quoteSh(logPath)}`,
|
|
platform: 'posix',
|
|
nonce: 'env-sequence',
|
|
waitTimeoutSeconds: 5
|
|
})
|
|
|
|
const setupExitPromise = waitForExit(
|
|
spawn('bash', ['-lc', commands.setupCommand], { stdio: 'pipe' })
|
|
)
|
|
const startupExit = await waitForExit(
|
|
spawn('bash', ['-lc', commands.startupCommand], {
|
|
stdio: 'pipe',
|
|
env: {
|
|
...process.env,
|
|
...commands.startupEnv,
|
|
[SETUP_AGENT_SEQUENCE_STARTUP_COMMAND_ENV]: `FOO=bar bash ${quoteSh(startupScriptPath)}; printf 'env-cleanup\\n' >> ${quoteSh(logPath)}`
|
|
}
|
|
})
|
|
)
|
|
const setupExit = await setupExitPromise
|
|
|
|
expect(setupExit.code).toBe(0)
|
|
expect(startupExit.code).toBe(0)
|
|
expect(readFileSync(logPath, 'utf8')).toBe('setup-done\nenv-start\nenv-cleanup\n')
|
|
}
|
|
)
|
|
|
|
it.skipIf(process.platform === 'win32')(
|
|
'times out instead of hanging forever when setup never writes a matching marker',
|
|
async () => {
|
|
const tempDir = makeTempDir()
|
|
const runnerScriptPath = join(tempDir, 'setup-runner.sh')
|
|
|
|
writeExecutable(runnerScriptPath, '#!/bin/sh\nexit 0\n')
|
|
|
|
const commands = createSequencedSetupAgentCommands({
|
|
runnerScriptPath,
|
|
startupCommand: 'printf ready',
|
|
platform: 'posix',
|
|
nonce: 'timeout-sequence',
|
|
waitTimeoutSeconds: 1
|
|
})
|
|
|
|
const startupExit = await waitForExit(
|
|
spawn('bash', ['-lc', commands.startupCommand], {
|
|
stdio: 'pipe',
|
|
env: { ...process.env, ...commands.startupEnv }
|
|
})
|
|
)
|
|
|
|
expect(startupExit.code).toBe(124)
|
|
expect(startupExit.stderr).toContain('Timed out waiting for setup before starting agent.')
|
|
}
|
|
)
|
|
})
|
|
|
|
describe('createSetupAgentSequenceNonce', () => {
|
|
it('prefers crypto.randomUUID when available', () => {
|
|
const originalCrypto = globalThis.crypto
|
|
vi.stubGlobal('crypto', { randomUUID: () => 'uuid-1' })
|
|
|
|
expect(createSetupAgentSequenceNonce()).toBe('uuid-1')
|
|
|
|
vi.stubGlobal('crypto', originalCrypto)
|
|
})
|
|
})
|
|
|
|
function resultPathWsl(): string {
|
|
return '\\\\wsl.localhost\\Ubuntu\\home\\jin\\repo\\.git\\worktrees\\feature\\orca\\setup-runner.sh'
|
|
}
|
|
|
|
function makeTempDir(): string {
|
|
const dir = mkdtempSync(join(tmpdir(), 'orca-setup-sequencing-'))
|
|
TEMP_DIRS.push(dir)
|
|
return dir
|
|
}
|
|
|
|
function writeExecutable(path: string, contents: string): void {
|
|
writeFileSync(path, contents, 'utf8')
|
|
chmodSync(path, 0o755)
|
|
}
|
|
|
|
function quoteSh(value: string): string {
|
|
return `'${value.replace(/'/g, `'\\''`)}'`
|
|
}
|
|
|
|
function readIfExists(path: string): string {
|
|
try {
|
|
return readFileSync(path, 'utf8')
|
|
} catch {
|
|
return ''
|
|
}
|
|
}
|
|
|
|
function sleep(ms: number): Promise<void> {
|
|
return new Promise((resolve) => {
|
|
setTimeout(resolve, ms)
|
|
})
|
|
}
|
|
|
|
function spawnWindowsCommand(
|
|
dir: string,
|
|
filename: string,
|
|
command: string,
|
|
env: Record<string, string> = {}
|
|
): ReturnType<typeof spawn> {
|
|
const scriptPath = join(dir, filename)
|
|
// Why: /s strips the quotes Node adds for batch paths containing spaces;
|
|
// argv spawning still exercises cmd.exe's native parser without that loss.
|
|
writeFileSync(scriptPath, `@echo off\r\n${command}\r\nexit /b %ERRORLEVEL%\r\n`, 'utf8')
|
|
return spawn('cmd.exe', ['/d', '/c', scriptPath], {
|
|
stdio: 'pipe',
|
|
env: { ...process.env, ...env }
|
|
})
|
|
}
|
|
|
|
function decodePowerShellScript(command: string): string {
|
|
const encoded = command.match(/-EncodedCommand\s+([A-Za-z0-9+/=]+)/)?.[1]
|
|
if (!encoded) {
|
|
throw new Error('Missing PowerShell encoded command')
|
|
}
|
|
return Buffer.from(encoded, 'base64').toString('utf16le')
|
|
}
|
|
|
|
function waitForExit(
|
|
child: ReturnType<typeof spawn>
|
|
): Promise<{ code: number | null; stderr: string }> {
|
|
return new Promise((resolve, reject) => {
|
|
let stderr = ''
|
|
child.stderr?.on('data', (chunk: Buffer | string) => {
|
|
stderr += chunk.toString()
|
|
})
|
|
child.once('error', reject)
|
|
child.once('close', (code) => {
|
|
resolve({ code, stderr })
|
|
})
|
|
})
|
|
}
|