| 1 | import { open, link, rename, unlink, lstat, realpath, opendir } from 'node:fs/promises'; |
| 2 | import { constants } from 'node:fs'; |
| 3 | import { DatabaseSync } from 'node:sqlite'; |
| 4 | import { dirname, basename, resolve } from 'node:path'; |
| 5 | import { randomUUID } from 'node:crypto'; |
| 6 | import { setTimeout as delay } from 'node:timers/promises'; |
| 7 | import { PET_BIN_MS, validatePetBucket, encodePetJSONL, decodePetJSONL } from '../../dist/core/pet-telemetry.js'; |
| 8 | |
| 9 | async function syncDirectory(path) { |
| 10 | let directory; |
| 11 | try { directory = await open(path, 'r'); await directory.sync(); } |
| 12 | catch (error) { |
| 13 | // Node cannot open/sync directory handles on Windows. File data is still |
| 14 | // synced before its atomic replacement on that platform. |
| 15 | if (process.platform !== 'win32' || !['EPERM', 'EISDIR', 'EINVAL', 'ENOTSUP'].includes(error.code)) throw error; |
| 16 | } finally { await directory?.close(); } |
| 17 | } |
| 18 | |
| 19 | // SQLite's OS lock is released even after process death. This empty sidecar |
| 20 | // contains no events or recorder state; keep its pathname so later processes |
| 21 | // coordinate on the same inode. No PID files or stale-lock deletion are needed. |
| 22 | async function lockRecorder(path) { |
| 23 | const name = `${path}.writer-lock`; |
| 24 | try { const created = await open(name, 'wx', 0o600); await created.close(); } |
| 25 | catch (error) { if (error.code !== 'EEXIST') throw error; } |
| 26 | const identity = await lstat(name, { bigint: true }); |
| 27 | if (!identity.isFile() || identity.nlink !== 1n || identity.size !== 0n) |
| 28 | throw new Error('Invalid pet recorder lock; existing files were preserved.'); |
| 29 | let database; |
| 30 | const check = async () => { |
| 31 | const current = await lstat(name, { bigint: true }); |
| 32 | if (!current.isFile() || current.nlink !== 1n || current.size !== 0n |
| 33 | || current.dev !== identity.dev || current.ino !== identity.ino) |
| 34 | throw new Error('The pet recorder lock was replaced; existing files were preserved.'); |
| 35 | }; |
| 36 | try { |
| 37 | database = new DatabaseSync(name); |
| 38 | database.exec('PRAGMA busy_timeout = 0; BEGIN EXCLUSIVE'); |
| 39 | await check(); |
| 40 | return { check, close: () => { database.close(); } }; |
| 41 | } catch (error) { |
| 42 | database?.close(); |
| 43 | if (error.errcode === 5 || error.errcode === 6) throw new Error('Another pet recorder is using this output.'); |
| 44 | throw error; |
| 45 | } |
| 46 | } |
| 47 | |
| 48 | /** Replaces the CLI's unbounded append-only output. Each complete segment is |
| 49 | * replayable on its own; the same live pathname always holds the newest one. */ |
| 50 | export async function createPetRecorder(path, { maxBuckets = 216_000, maxBytes = 64 * 1024 * 1024, report = () => {}, resume = false } = {}) { |
| 51 | if (!Number.isSafeInteger(maxBuckets) || maxBuckets < 1 || maxBuckets > 216_000 |
| 52 | || !Number.isSafeInteger(maxBytes) || maxBytes < 1 || maxBytes > 64 * 1024 * 1024) |
| 53 | throw new Error('Invalid pet recording segment limit.'); |
| 54 | path = resolve(await realpath(dirname(resolve(path))), basename(path)); |
| 55 | let lock = await lockRecorder(path), output, sequence = 0, bytes = 0, segment = 0, busy = false, restart = false, expectedMtime; |
| 56 | try { |
| 57 | try { output = await open(path, 'wx', 0o600); } |
| 58 | catch (error) { |
| 59 | if (!resume || error.code !== 'EEXIST') throw error; |
| 60 | const original = await lstat(path, { bigint: true }); |
| 61 | if (!original.isFile() || original.size > 64n * 1024n * 1024n) |
| 62 | throw new Error('The previous pet recording is not a bounded regular file; it was preserved.'); |
| 63 | output = await open(path, constants.O_RDWR | constants.O_APPEND | constants.O_NOFOLLOW | constants.O_NONBLOCK); |
| 64 | const held = await output.stat({ bigint: true }); |
| 65 | if (held.dev !== original.dev || held.ino !== original.ino || held.size !== original.size) |
| 66 | throw new Error('The previous pet recording changed while opening; it was preserved.'); |
| 67 | // Read at most the size already checked, including a single growth byte. |
| 68 | const contents = Buffer.alloc(Number(held.size) + 1); |
| 69 | let length = 0; |
| 70 | while (length < contents.length) { |
| 71 | const { bytesRead } = await output.read(contents, length, contents.length - length, length); |
| 72 | if (!bytesRead) break; |
| 73 | length += bytesRead; |
| 74 | } |
| 75 | if (length !== Number(held.size)) throw new Error('The previous pet recording changed while reading; it was preserved.'); |
| 76 | const text = new TextDecoder('utf-8', { fatal: true }).decode(contents.subarray(0, length)); |
| 77 | if (text && !text.endsWith('\n')) throw new Error('The previous pet recording has an incomplete final row; it was preserved.'); |
| 78 | decodePetJSONL(text); |
| 79 | const unchanged = await output.stat({ bigint: true }); |
| 80 | if (unchanged.size !== original.size || unchanged.mtimeNs !== original.mtimeNs) |
| 81 | throw new Error('The previous pet recording changed while reading; it was preserved.'); |
| 82 | bytes = length; restart = true; expectedMtime = original.mtimeNs; |
| 83 | // Continue archive numbering without collecting a growing directory list. |
| 84 | const prefix = `${basename(path)}.segment-`; |
| 85 | for await (const entry of await opendir(dirname(path))) { |
| 86 | if (!entry.name.startsWith(prefix)) continue; |
| 87 | const suffix = entry.name.slice(prefix.length); |
| 88 | if (!/^[0-9]{6,}\.jsonl$/.test(suffix)) continue; |
| 89 | const number = Number(suffix.slice(0, -6)); |
| 90 | if (!Number.isSafeInteger(number) || number >= Number.MAX_SAFE_INTEGER) |
| 91 | throw new Error('Pet archive numbering is exhausted; existing files were preserved.'); |
| 92 | segment = Math.max(segment, number); |
| 93 | } |
| 94 | } |
| 95 | expectedMtime ??= (await output.stat({ bigint: true })).mtimeNs; |
| 96 | } catch (error) { try { await output?.close(); } finally { lock.close(); } throw error; } |
| 97 | return { |
| 98 | async append(bucket) { |
| 99 | if (!output || busy) throw new Error('Pet recorder is closed or already writing.'); |
| 100 | validatePetBucket(bucket); |
| 101 | busy = true; |
| 102 | try { |
| 103 | await lock.check(); |
| 104 | const held = await output.stat({ bigint: true }), current = await lstat(path, { bigint: true }); |
| 105 | if (!current.isFile() || held.dev !== current.dev || held.ino !== current.ino || held.size !== BigInt(bytes) || held.mtimeNs !== expectedMtime) |
| 106 | throw new Error('The live pet recording was changed or replaced externally; existing files were preserved.'); |
| 107 | const encode = seq => encodePetJSONL([{ ...bucket, sequence: seq, simTimeMs: seq * PET_BIN_MS }]); |
| 108 | let row = encode(sequence), size = Buffer.byteLength(row); |
| 109 | if (restart || sequence >= maxBuckets || bytes + size > maxBytes) { |
| 110 | row = encode(0); size = Buffer.byteLength(row); |
| 111 | if (size > maxBytes) throw new Error('Pet bucket exceeds the recording segment byte limit.'); |
| 112 | const temporary = resolve(dirname(path), `.${basename(path)}.next-${randomUUID()}`); |
| 113 | const archive = `${path}.segment-${String(segment + 1).padStart(6, '0')}.jsonl`; |
| 114 | let next, installed = false, created = false; |
| 115 | try { |
| 116 | next = await open(temporary, 'wx', 0o600); created = true; |
| 117 | await next.writeFile(row); await next.sync(); |
| 118 | const nextIdentity = await next.stat({ bigint: true }); |
| 119 | await next.close(); next = undefined; |
| 120 | await output.sync(); |
| 121 | // link is exclusive: a collision never replaces someone else's |
| 122 | // archive. Persist this name before replacing the live pathname. |
| 123 | await link(path, archive); await syncDirectory(dirname(path)); |
| 124 | // Windows can reject replacement while either writer handle is |
| 125 | // open. Both files are synced and the old file is archived first. |
| 126 | await output.close(); output = undefined; |
| 127 | for (let attempt = 0; ; attempt++) { |
| 128 | await lock.check(); |
| 129 | const destination = await lstat(path, { bigint: true }); |
| 130 | if (!destination.isFile() || destination.dev !== held.dev || destination.ino !== held.ino |
| 131 | || destination.size !== held.size || destination.mtimeNs !== held.mtimeNs) |
| 132 | throw new Error('The live pet recording changed while rotating; existing files were preserved.'); |
| 133 | try { await rename(temporary, path); break; } |
| 134 | catch (error) { |
| 135 | // A reader or file scanner can briefly deny replacement on |
| 136 | // Windows. Retry for under two seconds, checking identity each |
| 137 | // time; persistent denial still stops without deleting history. |
| 138 | if (process.platform !== 'win32' || !['EPERM', 'EBUSY'].includes(error.code) || attempt >= 20) throw error; |
| 139 | await delay(Math.min(100, (attempt + 1) * 25)); |
| 140 | } |
| 141 | } |
| 142 | installed = true; |
| 143 | // The first row is already published. Account for it before any |
| 144 | // fallible cleanup/report so a later append sees the actual file. |
| 145 | sequence = 1; bytes = size; segment++; restart = false; |
| 146 | output = await open(path, constants.O_WRONLY | constants.O_APPEND); |
| 147 | const reopened = await output.stat({ bigint: true }); |
| 148 | if (reopened.dev !== nextIdentity.dev || reopened.ino !== nextIdentity.ino) |
| 149 | throw new Error('The live pet recording was replaced externally after rotation.'); |
| 150 | expectedMtime = reopened.mtimeNs; |
| 151 | await syncDirectory(dirname(path)); |
| 152 | report(`Archived pet recording: ${archive}`); |
| 153 | } finally { |
| 154 | await next?.close(); |
| 155 | if (created && !installed) await unlink(temporary); |
| 156 | } |
| 157 | return; |
| 158 | } else { |
| 159 | await output.writeFile(row); |
| 160 | expectedMtime = (await output.stat({ bigint: true })).mtimeNs; |
| 161 | } |
| 162 | bytes += size; sequence++; |
| 163 | } finally { busy = false; } |
| 164 | }, |
| 165 | async close() { |
| 166 | if (busy) throw new Error('Wait for the pet recorder write before closing.'); |
| 167 | const current = output, heldLock = lock; output = undefined; lock = undefined; |
| 168 | try { if (current) { try { await current.sync(); } finally { await current.close(); } } } |
| 169 | finally { heldLock?.close(); } |
| 170 | }, |
| 171 | }; |
| 172 | } |
| 173 |