🎯 Use case
Reads or normalizes session / transcript history. On the API surface it exposes getPastedTextRefNumLines, formatPastedTextRef, formatImageRef, parseReferences, and expandPastedTextRefs (and more) — mainly functions, hooks, or classes. Dependencies touch Node filesystem and Node path helpers. It composes internal code from bootstrap and utils (relative imports).
Generated from folder role, exports, dependency roots, and inline comments — not hand-reviewed for every path.
🧠 Inline summary
import { appendFile, writeFile } from 'fs/promises' import { join } from 'path' import { getProjectRoot, getSessionId } from './bootstrap/state.js' import { registerCleanup } from './utils/cleanupRegistry.js' import type { HistoryEntry, PastedContent } from './utils/config.js'
📤 Exports (heuristic)
getPastedTextRefNumLinesformatPastedTextRefformatImageRefparseReferencesexpandPastedTextRefsTimestampedHistoryEntryaddToHistoryclearPendingHistoryEntriesremoveLastFromHistory
📚 External import roots
Package roots from from "…" (relative paths omitted).
fspath
🖥️ Source preview
import { appendFile, writeFile } from 'fs/promises'
import { join } from 'path'
import { getProjectRoot, getSessionId } from './bootstrap/state.js'
import { registerCleanup } from './utils/cleanupRegistry.js'
import type { HistoryEntry, PastedContent } from './utils/config.js'
import { logForDebugging } from './utils/debug.js'
import { getClaudeConfigHomeDir, isEnvTruthy } from './utils/envUtils.js'
import { getErrnoCode } from './utils/errors.js'
import { readLinesReverse } from './utils/fsOperations.js'
import { lock } from './utils/lockfile.js'
import {
hashPastedText,
retrievePastedText,
storePastedText,
} from './utils/pasteStore.js'
import { sleep } from './utils/sleep.js'
import { jsonParse, jsonStringify } from './utils/slowOperations.js'
const MAX_HISTORY_ITEMS = 100
const MAX_PASTED_CONTENT_LENGTH = 1024
/**
* Stored paste content - either inline content or a hash reference to paste store.
*/
type StoredPastedContent = {
id: number
type: 'text' | 'image'
content?: string // Inline content for small pastes
contentHash?: string // Hash reference for large pastes stored externally
mediaType?: string
filename?: string
}
/**
* Claude Code parses history for pasted content references to match back to
* pasted content. The references look like:
* Text: [Pasted text #1 +10 lines]
* Image: [Image #2]
* The numbers are expected to be unique within a single prompt but not across
* prompts. We choose numeric, auto-incrementing IDs as they are more
* user-friendly than other ID options.
*/
// Note: The original text paste implementation would consider input like
// "line1\nline2\nline3" to have +2 lines, not 3 lines. We preserve that
// behavior here.
export function getPastedTextRefNumLines(text: string): number {
return (text.match(/\r\n|\r|\n/g) || []).length
}
export function formatPastedTextRef(id: number, numLines: number): string {
if (numLines === 0) {
return `[Pasted text #${id}]`
}
return `[Pasted text #${id} +${numLines} lines]`
}
export function formatImageRef(id: number): string {
return `[Image #${id}]`
}
export function parseReferences(
input: string,
): Array<{ id: number; match: string; index: number }> {
const referencePattern =
/\[(Pasted text|Image|\.\.\.Truncated text) #(\d+)(?: \+\d+ lines)?(\.)*\]/g
const matches = [...input.matchAll(referencePattern)]
return matches
.map(match => ({
id: parseInt(match[2] || '0'),
match: match[0],
index: match.index,
}))
.filter(match => match.id > 0)
}
/**
* Replace [Pasted text #N] placeholders in input with their actual content.
* Image refs are left alone — they become content blocks, not inlined text.
*/
export function expandPastedTextRefs(
input: string,
pastedContents: Record<number, PastedContent>,
): string {
const refs = parseReferences(input)
let expanded = input
// Splice at the original match offsets so placeholder-like strings inside
// pasted content are never confused for real refs. Reverse order keeps
// earlier offsets valid after later replacements.
for (let i = refs.length - 1; i >= 0; i--) {
const ref = refs[i]!
const content = pastedContents[ref.id]
if (content?.type !== 'text') continue
expanded =
expanded.slice(0, ref.index) +
content.content +
expanded.slice(ref.index + ref.match.length)
}
return expanded
}
function deserializeLogEntry(line: string): LogEntry {
return jsonParse(line) as LogEntry
}
async function* makeLogEntryReader(): AsyncGenerator<LogEntry> {
const currentSession = getSessionId()
// Start with entries that have yet to be flushed to disk
for (let i = pendingEntries.length - 1; i >= 0; i--) {
yield pendingEntries[i]!
}
// Read from global history file (shared across all projects)
const historyPath = join(getClaudeConfigHomeDir(), 'history.jsonl')
try {
for await (const line of readLinesReverse(historyPath)) {
try {
const entry = deserializeLogEntry(line)
// removeLastFromHistory slow path: entry was flushed before removal,
// so filter here so both getHistory (Up-arrow) and makeHistoryReader
// (ctrl+r search) skip it consistently.
if (
entry.sessionId === currentSession &&
skippedTimestamps.has(entry.timestamp)
) {
continue
}
yield entry
} catch (error) {
// Not a critical error - just skip malformed lines
logForDebugging(`Failed to parse history line: ${error}`)
}
}
} catch (e: unknown) {
const code = getErrnoCode(e)
if (code === 'ENOENT') {
return
}
throw e
}
}
export async function* makeHistoryReader(): AsyncGenerator<HistoryEntry> {
for await (const entry of makeLogEntryReader()) {
yield await logEntryToHistoryEntry(entry)
}
}
export type TimestampedHistoryEntry = {
display: string
timestamp: number
resolve: () => Promise<HistoryEntry>
}
/**
* Current-project history for the ctrl+r picker: deduped by display text,
* newest first, with timestamps. Paste contents are resolved lazily via
* `resolve()` — the picker only reads display+timestamp for the list.
*/
export async function* getTimestampedHistory(): AsyncGenerator<TimestampedHistoryEntry> {
const currentProject = getProjectRoot()
const seen = new Set<string>()
for await (const entry of makeLogEntryReader()) {
if (!entry || typeof entry.project !== 'string') continue
if (entry.project !== currentProject) continue
if (seen.has(entry.display)) continue
seen.add(entry.display)
yield {
display: entry.display,
timestamp: entry.timestamp,
resolve: () => logEntryToHistoryEntry(entry),
}
if (seen.size >= MAX_HISTORY_ITEMS) return
}
}
/**
* Get history entries for the current project, with current session's entries first.
*
* Entries from the current session are yielded before entries from other sessions,
* so concurrent sessions don't interleave their up-arrow history. Within each group,
* order is newest-first. Scans the same MAX_HISTORY_ITEMS window as before —
* entries are reordered within that window, not beyond it.
*/
export async function* getHistory(): AsyncGenerator<HistoryEntry> {
const currentProject = getProjectRoot()
const currentSession = getSessionId()
const otherSessionEntries: LogEntry[] = []
let yielded = 0
for await (const entry of makeLogEntryReader()) {
// Skip malformed entries (corrupted file, old format, or invalid JSON structure)
if (!entry || typeof entry.project !== 'string') continue
if (entry.project !== currentProject) continue
if (entry.sessionId === currentSession) {
yield await logEntryToHistoryEntry(entry)
yielded++
} else {
otherSessionEntries.push(entry)
}
// Same MAX_HISTORY_ITEMS window as before — just reordered within it.
if (yielded + otherSessionEntries.length >= MAX_HISTORY_ITEMS) break
}
for (const entry of otherSessionEntries) {
if (yielded >= MAX_HISTORY_ITEMS) return
yield await logEntryToHistoryEntry(entry)
yielded++
}
}
type LogEntry = {
display: string
pastedContents: Record<number, StoredPastedContent>
timestamp: number
project: string
sessionId?: string
}
/**
* Resolve stored paste content to full PastedContent by fetching from paste store if needed.
*/
async function resolveStoredPastedContent(
stored: StoredPastedContent,
): Promise<PastedContent | null> {
// If we have inline content, use it directly
if (stored.content) {
return {
id: stored.id,
type: stored.type,
content: stored.content,
mediaType: stored.mediaType,
filename: stored.filename,
}
}
// If we have a hash reference, fetch from paste store
if (stored.contentHash) {
const content = await retrievePastedText(stored.contentHash)
if (content) {
return {
id: stored.id,
type: stored.type,
content,
mediaType: stored.mediaType,
filename: stored.filename,
}
}
}
// Content not available
return null
}
/**
* Convert LogEntry to HistoryEntry by resolving paste store references.
*/
async function logEntryToHistoryEntry(entry: LogEntry): Promise<HistoryEntry> {
const pastedContents: Record<number, PastedContent> = {}
for (const [id, stored] of Object.entries(entry.pastedContents || {})) {
const resolved = await resolveStoredPastedContent(stored)
if (resolved) {
pastedContents[Number(id)] = resolved
}
}
return {
display: entry.display,
pastedContents,
}
}
let pendingEntries: LogEntry[] = []
let isWriting = false
let currentFlushPromise: Promise<void> | null = null
let cleanupRegistered = false
let lastAddedEntry: LogEntry | null = null
// Timestamps of entries already flushed to disk that should be skipped when
// reading. Used by removeLastFromHistory when the entry has raced past the
// pending buffer. Session-scoped (module state resets on process restart).
const skippedTimestamps = new Set<number>()
// Core flush logic - writes pending entries to disk
async function immediateFlushHistory(): Promise<void> {
if (pendingEntries.length === 0) {
return
}
let release
try {
const historyPath = join(getClaudeConfigHomeDir(), 'history.jsonl')
// Ensure the file exists before acquiring lock (append mode creates if missing)
await writeFile(historyPath, '', {
encoding: 'utf8',
mode: 0o600,
flag: 'a',
})
release = await lock(historyPath, {
stale: 10000,
retries: {
retries: 3,
minTimeout: 50,
},
})
const jsonLines = pendingEntries.map(entry => jsonStringify(entry) + '\n')
pendingEntries = []
await appendFile(historyPath, jsonLines.join(''), { mode: 0o600 })
} catch (error) {
logForDebugging(`Failed to write prompt history: ${error}`)
} finally {
if (release) {
await release()
}
}
}
async function flushPromptHistory(retries: number): Promise<void> {
if (isWriting || pendingEntries.length === 0) {
return
}
// Stop trying to flush history until the next user prompt
if (retries > 5) {
return
}
isWriting = true
try {
await immediateFlushHistory()
} finally {
isWriting = false
if (pendingEntries.length > 0) {
// Avoid trying again in a hot loop
await sleep(500)
void flushPromptHistory(retries + 1)
}
}
}
async function addToPromptHistory(
command: HistoryEntry | string,
): Promise<void> {
const entry =
typeof command === 'string'
? { display: command, pastedContents: {} }
: command
const storedPastedContents: Record<number, StoredPastedContent> = {}
if (entry.pastedContents) {
for (const [id, content] of Object.entries(entry.pastedContents)) {
// Filter out images (they're stored separately in image-cache)
if (content.type === 'image') {
continue
}
// For small text content, store inline
if (content.content.length <= MAX_PASTED_CONTENT_LENGTH) {
storedPastedContents[Number(id)] = {
id: content.id,
type: content.type,
content: content.content,
mediaType: content.mediaType,
filename: content.filename,
}
} else {
// For large text content, compute hash synchronously and store reference
// The actual disk write happens async (fire-and-forget)
const hash = hashPastedText(content.content)
storedPastedContents[Number(id)] = {
id: content.id,
type: content.type,
contentHash: hash,
mediaType: content.mediaType,
filename: content.filename,
}
// Fire-and-forget disk write - don't block history entry creation
void storePastedText(hash, content.content)
}
}
}
const logEntry: LogEntry = {
...entry,
pastedContents: storedPastedContents,
timestamp: Date.now(),
project: getProjectRoot(),
sessionId: getSessionId(),
}
pendingEntries.push(logEntry)
lastAddedEntry = logEntry
currentFlushPromise = flushPromptHistory(0)
void currentFlushPromise
}
export function addToHistory(command: HistoryEntry | string): void {
// Skip history when running in a tmux session spawned by Claude Code's Tungsten tool.
// This prevents verification/test sessions from polluting the user's real command history.
if (isEnvTruthy(process.env.CLAUDE_CODE_SKIP_PROMPT_HISTORY)) {
return
}
// Register cleanup on first use
if (!cleanupRegistered) {
cleanupRegistered = true
registerCleanup(async () => {
// If there's an in-progress flush, wait for it
if (currentFlushPromise) {
await currentFlushPromise
}
// If there are still pending entries after the flush completed, do one final flush
if (pendingEntries.length > 0) {
await immediateFlushHistory()
}
})
}
void addToPromptHistory(command)
}
export function clearPendingHistoryEntries(): void {
pendingEntries = []
lastAddedEntry = null
skippedTimestamps.clear()
}
/**
* Undo the most recent addToHistory call. Used by auto-restore-on-interrupt:
* when Esc rewinds the conversation before any response arrives, the submit is
* semantically undone — the history entry should be too, otherwise Up-arrow
* shows the restored text twice (once from the input box, once from disk).
*
* Fast path pops from the pending buffer. If the async flush already won the
* race (TTFT is typically >> disk write latency), the entry's timestamp is
* added to a skip-set consulted by getHistory. One-shot: clears the tracked
* entry so a second call is a no-op.
*/
export function removeLastFromHistory(): void {
if (!lastAddedEntry) return
const entry = lastAddedEntry
lastAddedEntry = null
const idx = pendingEntries.lastIndexOf(entry)
if (idx !== -1) {
pendingEntries.splice(idx, 1)
} else {
skippedTimestamps.add(entry.timestamp)
}
}