mirror of
https://github.com/stablyai/orca.git
synced 2026-09-25 00:02:35 +00:00
NeilandOrca
6b6a3b666a
Three journeys attempted; none promoted, and the reasons are recorded in the ledger rather than rounded up. MaxSessions=1 against real OpenSSH, with the cap read back from `sshd -T` rather than assumed, and remote pids read on the container two independent ways that must agree, each carrying its kernel start time. Two disjoint mutations discriminate — one reddens only the reconnect clause, the other only the two restart clauses. But the disconnect clause is a forward guard: four separate guard removals left it green, so nothing shipped is load-bearing for it. Lazy discovery samples sshd's own accept log and live session census across a 22s window with the in-use host as a positive control. No mutation reddens its third clause alone — the real cross-host lease scoping is load-bearing, but removing it breaks the sibling host during setup, so the failure carries no clause information. The paired-runtime skew spec pairs two real processes at different versions and refuses to run rather than degrade into a same-version pairing that would look green and prove nothing. No production code changes. Co-authored-by: Orca <help@stably.ai>
65 lines
2.2 KiB
TypeScript
65 lines
2.2 KiB
TypeScript
/**
|
|
* Asks a remote pane's own shell for its pid, through the production write path.
|
|
*
|
|
* Why not read it from app state: a pane id, a leaf id and a PTY id can all
|
|
* match while a different process runs underneath. The shell answers on the
|
|
* container's filesystem, so the pid is observed where the process actually
|
|
* lives — and a pane that no longer reaches its shell cannot answer at all.
|
|
*/
|
|
import type { Page } from '@stablyai/playwright-test'
|
|
import { expect } from '@stablyai/playwright-test'
|
|
import { sendToTerminal } from './terminal'
|
|
import {
|
|
execDockerSshRelayTargetCommand,
|
|
shellQuote,
|
|
type DockerSshRelayTarget
|
|
} from './docker-ssh-relay-target'
|
|
|
|
function readRemoteFile(target: DockerSshRelayTarget, remotePath: string): string | null {
|
|
try {
|
|
return execDockerSshRelayTargetCommand(target, `cat ${shellQuote(remotePath)}`)
|
|
} catch {
|
|
return null
|
|
}
|
|
}
|
|
|
|
/**
|
|
* `probePath` must be unique per probe: a stale file from an earlier phase would
|
|
* answer for a shell that is no longer there.
|
|
*/
|
|
export async function readRemoteShellPid(
|
|
page: Page,
|
|
target: DockerSshRelayTarget,
|
|
args: { ptyId: string; probePath: string; timeoutMs?: number }
|
|
): Promise<number> {
|
|
if (readRemoteFile(target, args.probePath) !== null) {
|
|
throw new Error(`Remote shell pid probe path already exists: ${args.probePath}`)
|
|
}
|
|
// Write-then-rename: the redirect creates the file before printf fills it, and
|
|
// a poll that caught it empty would read pid 0.
|
|
await sendToTerminal(
|
|
page,
|
|
args.ptyId,
|
|
`printf '%s\\n' "$$" > ${args.probePath}.part && mv ${args.probePath}.part ${args.probePath}\r`
|
|
)
|
|
let pid: number | null = null
|
|
await expect
|
|
.poll(
|
|
() => {
|
|
const raw = readRemoteFile(target, args.probePath)
|
|
const parsed = raw === null ? Number.NaN : Number(raw.trim())
|
|
pid = Number.isInteger(parsed) && parsed > 0 ? parsed : null
|
|
return pid
|
|
},
|
|
{
|
|
timeout: args.timeoutMs ?? 60_000,
|
|
message: `Remote pane ${args.ptyId} never reported its shell pid to ${args.probePath}`
|
|
}
|
|
)
|
|
.not.toBeNull()
|
|
if (pid === null) {
|
|
throw new Error(`Remote pane ${args.ptyId} reported no shell pid`)
|
|
}
|
|
return pid
|
|
}
|