claude-code-session-manager 0.11.1 → 0.12.0

package/src/main/search.cjs

@@ -0,0 +1,337 @@
+/**
+ * Search IPC — fuzzy file path search + content grep.
+ *
+ * - Prefers ripgrep (`rg`) when present on PATH for both fast file enumeration
+ *   (`rg --files`) and content search (`rg --json`). Falls back to a Node
+ *   fs-based recursive walk capped at depth 10 when rg is missing.
+ * - All cwds are routed through `config.cjs`'s `validatePath` so the renderer
+ *   can never search outside the home directory.
+ * - 10s wall-clock ceiling per call — search runs that overflow are killed
+ *   and return partial results.
+ * - Ignored dirs (both modes): .git / node_modules / dist / build / .next /
+ *   .cache / coverage / .turbo.
+ *
+ * Designed to back two renderer modals: QuickOpen (Cmd+P) and GlobalSearch
+ * (Cmd+Shift+F).
+ */
+
+const { ipcMain } = require('electron');
+const { spawn } = require('node:child_process');
+const fsp = require('node:fs/promises');
+const path = require('node:path');
+const os = require('node:os');
+const { z } = require('zod');
+
+const { validatePath } = require('./config.cjs');
+
+const TIMEOUT_MS = 10_000;
+const MAX_FILES = 5000;
+const MAX_TEXT_MATCHES = 500;
+const MAX_WALK_DEPTH = 10;
+const IGNORED_DIRS = new Set([
+  '.git', 'node_modules', 'dist', 'build', '.next',
+  '.cache', 'coverage', '.turbo', '.parcel-cache', '.pnpm-store',
+]);
+
+// ── ripgrep detection ─────────────────────────────────────────────────────
+let _rgAvailable = null; // null = unknown, true/false = cached
+function probeRipgrep() {
+  if (_rgAvailable !== null) return _rgAvailable;
+  return new Promise((resolve) => {
+    let resolved = false;
+    const done = (val) => { if (!resolved) { resolved = true; _rgAvailable = val; resolve(val); } };
+    try {
+      const proc = spawn('rg', ['--version'], { stdio: ['ignore', 'pipe', 'pipe'] });
+      proc.on('error', () => done(false));
+      proc.on('exit', (code) => done(code === 0));
+      // Hard timeout: probe shouldn't take more than 1s.
+      setTimeout(() => { try { proc.kill('SIGKILL'); } catch { /* noop */ } done(false); }, 1000);
+    } catch {
+      done(false);
+    }
+  });
+}
+
+// ── helpers ───────────────────────────────────────────────────────────────
+function expandHome(p) {
+  if (!p) return p;
+  if (p === '~') return os.homedir();
+  if (p.startsWith('~/')) return path.join(os.homedir(), p.slice(2));
+  return p;
+}
+
+/**
+ * Spawn rg with argv (no shell), capture stdout/stderr, enforce TIMEOUT_MS.
+ * Resolves to { ok, stdout, stderr, exitCode, timedOut }.
+ */
+function runRg(args, cwd) {
+  return new Promise((resolve) => {
+    let stdout = '';
+    let stderr = '';
+    let timedOut = false;
+    let proc;
+    try {
+      proc = spawn('rg', args, { cwd, stdio: ['ignore', 'pipe', 'pipe'] });
+    } catch (e) {
+      resolve({ ok: false, stdout: '', stderr: e.message, exitCode: -1, timedOut: false });
+      return;
+    }
+    const timer = setTimeout(() => {
+      timedOut = true;
+      try { proc.kill('SIGTERM'); } catch { /* noop */ }
+      setTimeout(() => { try { proc.kill('SIGKILL'); } catch { /* noop */ } }, 500);
+    }, TIMEOUT_MS);
+
+    proc.stdout.on('data', (chunk) => {
+      stdout += chunk.toString('utf8');
+      // Soft cap on buffer growth — if rg yields more than 10 MB of stdout
+      // something is wrong with the query; just kill it.
+      if (stdout.length > 10 * 1024 * 1024) {
+        try { proc.kill('SIGTERM'); } catch { /* noop */ }
+      }
+    });
+    proc.stderr.on('data', (chunk) => { stderr += chunk.toString('utf8'); });
+    proc.on('error', (e) => {
+      clearTimeout(timer);
+      resolve({ ok: false, stdout, stderr: e.message, exitCode: -1, timedOut });
+    });
+    proc.on('exit', (code) => {
+      clearTimeout(timer);
+      // rg exits 1 when zero matches found — that's still "ok" for us.
+      resolve({ ok: code === 0 || code === 1, stdout, stderr, exitCode: code ?? -1, timedOut });
+    });
+  });
+}
+
+/**
+ * Recursive fs walk capped at MAX_WALK_DEPTH and ignoring IGNORED_DIRS.
+ * Returns absolute paths. Bails out early at MAX_FILES.
+ */
+async function fsWalk(rootAbs) {
+  const out = [];
+  const startedAt = Date.now();
+  async function walk(dir, depth) {
+    if (depth > MAX_WALK_DEPTH) return;
+    if (out.length >= MAX_FILES) return;
+    if (Date.now() - startedAt > TIMEOUT_MS) return;
+    let entries;
+    try {
+      entries = await fsp.readdir(dir, { withFileTypes: true });
+    } catch {
+      return;
+    }
+    for (const entry of entries) {
+      if (out.length >= MAX_FILES) return;
+      if (entry.name.startsWith('.') && entry.name !== '.env') {
+        // Skip dotfiles by default — matches rg's defaults closely enough
+        // for our use case. .env is a common exception users may want.
+        if (entry.isDirectory()) continue;
+      }
+      if (entry.isDirectory()) {
+        if (IGNORED_DIRS.has(entry.name)) continue;
+        await walk(path.join(dir, entry.name), depth + 1);
+      } else if (entry.isFile()) {
+        out.push(path.join(dir, entry.name));
+      }
+    }
+  }
+  await walk(rootAbs, 0);
+  return out;
+}
+
+/**
+ * Naive content grep for the rg-fallback path. Reads each file up to 1 MB,
+ * splits on newlines, returns the first MAX_TEXT_MATCHES occurrences. Skips
+ * binary files heuristically (NUL byte in first 4 KiB).
+ */
+async function fsGrep(rootAbs, needle, opts) {
+  const caseSensitive = !!opts.caseSensitive;
+  const limit = Math.max(1, Math.min(MAX_TEXT_MATCHES, opts.limit ?? MAX_TEXT_MATCHES));
+  const haystack = caseSensitive ? needle : needle.toLowerCase();
+  const startedAt = Date.now();
+  const files = await fsWalk(rootAbs);
+  const matches = [];
+  for (const file of files) {
+    if (matches.length >= limit) break;
+    if (Date.now() - startedAt > TIMEOUT_MS) break;
+    let buf;
+    try {
+      const st = await fsp.stat(file);
+      if (st.size > 1024 * 1024) continue;
+      buf = await fsp.readFile(file);
+    } catch { continue; }
+    // Binary sniff.
+    const sniff = buf.slice(0, 4096);
+    if (sniff.includes(0)) continue;
+    const text = buf.toString('utf8');
+    const lines = text.split('\n');
+    for (let i = 0; i < lines.length; i++) {
+      if (matches.length >= limit) break;
+      const line = lines[i];
+      const target = caseSensitive ? line : line.toLowerCase();
+      const idx = target.indexOf(haystack);
+      if (idx !== -1) {
+        matches.push({
+          path: file,
+          line: i + 1,
+          column: idx + 1,
+          text: line.length > 400 ? line.slice(0, 400) + '…' : line,
+        });
+      }
+    }
+  }
+  return matches;
+}
+
+// ── public API: search:files ──────────────────────────────────────────────
+async function searchFiles({ cwd, query, opts }) {
+  const limit = Math.max(1, Math.min(MAX_FILES, opts?.limit ?? MAX_FILES));
+  let resolved;
+  try { resolved = validatePath(expandHome(cwd)); }
+  catch (e) { return { ok: false, files: [], error: e.message, usedRipgrep: false }; }
+
+  const usedRipgrep = await probeRipgrep();
+  let absPaths = [];
+  if (usedRipgrep) {
+    // `rg --files` enumerates the tree honoring .gitignore. We add explicit
+    // --glob '!dir' exclusions on top in case the repo isn't a git repo.
+    const args = [
+      '--files', '--hidden', '--follow',
+      '--max-filesize', '5M',
+    ];
+    for (const dir of IGNORED_DIRS) {
+      args.push('--glob', `!**/${dir}/**`);
+    }
+    const result = await runRg(args, resolved);
+    if (result.timedOut) {
+      return { ok: false, files: [], error: 'search timeout (10s)', usedRipgrep: true };
+    }
+    if (!result.ok) {
+      return { ok: false, files: [], error: result.stderr || `rg exit ${result.exitCode}`, usedRipgrep: true };
+    }
+    absPaths = result.stdout.split('\n')
+      .filter((l) => l.length > 0)
+      .map((rel) => path.join(resolved, rel));
+  } else {
+    absPaths = await fsWalk(resolved);
+  }
+
+  // Fuzzy filter / sort happens in renderer. Main returns the candidate set.
+  // When query is provided we do a coarse case-insensitive substring filter
+  // to cut payload size — renderer's fuzzy ranker handles the precise scoring.
+  if (query && query.length > 0) {
+    const q = query.toLowerCase();
+    absPaths = absPaths.filter((p) => p.toLowerCase().includes(q));
+  }
+  absPaths = absPaths.slice(0, limit);
+
+  // Shape each entry like files:list so QuickOpen can reuse FileNode types.
+  const files = absPaths.map((abs) => ({
+    name: path.basename(abs),
+    path: abs,
+    isDirectory: false,
+    isFile: true,
+  }));
+  return { ok: true, files, error: null, usedRipgrep };
+}
+
+// ── public API: search:text ───────────────────────────────────────────────
+async function searchText({ cwd, query, opts }) {
+  const limit = Math.max(1, Math.min(MAX_TEXT_MATCHES, opts?.limit ?? MAX_TEXT_MATCHES));
+  const caseSensitive = !!opts?.caseSensitive;
+  if (!query || query.length === 0) {
+    return { ok: true, matches: [], error: null, usedRipgrep: false };
+  }
+  let resolved;
+  try { resolved = validatePath(expandHome(cwd)); }
+  catch (e) { return { ok: false, matches: [], error: e.message, usedRipgrep: false }; }
+
+  const usedRipgrep = await probeRipgrep();
+  if (!usedRipgrep) {
+    const matches = await fsGrep(resolved, query, { caseSensitive, limit });
+    return { ok: true, matches, error: null, usedRipgrep: false };
+  }
+
+  // rg --json gives us { type, data: { ... } } lines. We use --fixed-strings
+  // so users don't accidentally trigger regex on `.` or `(`. Renderer can
+  // toggle regex later if we ever surface that option.
+  const args = [
+    '--json', '--fixed-strings', '--hidden', '--follow',
+    '--max-filesize', '5M',
+    '--max-count', String(limit),
+  ];
+  if (!caseSensitive) args.push('--ignore-case');
+  for (const dir of IGNORED_DIRS) {
+    args.push('--glob', `!**/${dir}/**`);
+  }
+  args.push('--', query);
+
+  const result = await runRg(args, resolved);
+  if (result.timedOut) {
+    return { ok: false, matches: [], error: 'search timeout (10s)', usedRipgrep: true };
+  }
+  if (!result.ok) {
+    return { ok: false, matches: [], error: result.stderr || `rg exit ${result.exitCode}`, usedRipgrep: true };
+  }
+
+  const matches = [];
+  const lines = result.stdout.split('\n');
+  for (const raw of lines) {
+    if (matches.length >= limit) break;
+    if (!raw) continue;
+    let evt;
+    try { evt = JSON.parse(raw); } catch { continue; }
+    if (evt.type !== 'match') continue;
+    const data = evt.data;
+    if (!data) continue;
+    const rel = data.path?.text;
+    if (!rel) continue;
+    const abs = path.isAbsolute(rel) ? rel : path.join(resolved, rel);
+    const lineNum = data.line_number ?? 0;
+    let text = data.lines?.text ?? '';
+    if (text.endsWith('\n')) text = text.slice(0, -1);
+    if (text.length > 400) text = text.slice(0, 400) + '…';
+    const column = (data.submatches && data.submatches[0]?.start != null)
+      ? data.submatches[0].start + 1
+      : 1;
+    matches.push({ path: abs, line: lineNum, column, text });
+  }
+  return { ok: true, matches, error: null, usedRipgrep: true };
+}
+
+// ── IPC registration ──────────────────────────────────────────────────────
+const searchFilesSchema = z.object({
+  cwd: z.string().min(1).max(4096),
+  query: z.string().max(1024).optional().default(''),
+  opts: z.object({
+    limit: z.number().int().min(1).max(MAX_FILES).optional(),
+  }).optional(),
+});
+
+const searchTextSchema = z.object({
+  cwd: z.string().min(1).max(4096),
+  query: z.string().min(1).max(1024),
+  opts: z.object({
+    limit: z.number().int().min(1).max(MAX_TEXT_MATCHES).optional(),
+    caseSensitive: z.boolean().optional(),
+  }).optional(),
+});
+
+function registerSearchHandlers() {
+  ipcMain.handle('search:files', (_e, payload) => {
+    const parsed = searchFilesSchema.parse(payload);
+    return searchFiles(parsed);
+  });
+  ipcMain.handle('search:text', (_e, payload) => {
+    const parsed = searchTextSchema.parse(payload);
+    return searchText(parsed);
+  });
+}
+
+module.exports = {
+  registerSearchHandlers,
+  // exported for tests
+  searchFiles,
+  searchText,
+  probeRipgrep,
+};

package/src/preload/api.d.ts

@@ -417,6 +417,47 @@ export interface FilesRenameResult { ok: boolean; newPath?: string; error: strin
 export interface FilesDeleteResult { ok: boolean; error: string | null }
 export interface FilesShellResult { ok: boolean; error?: string }
 
+export interface SearchFileEntry {
+  name: string;
+  path: string;
+  isDirectory: false;
+  isFile: true;
+}
+export interface SearchFilesResult {
+  ok: boolean;
+  files: SearchFileEntry[];
+  error: string | null;
+  usedRipgrep: boolean;
+}
+export interface SearchTextMatch {
+  path: string;
+  line: number;
+  column: number;
+  text: string;
+}
+export interface SearchTextResult {
+  ok: boolean;
+  matches: SearchTextMatch[];
+  error: string | null;
+  usedRipgrep: boolean;
+}
+
+export interface RepoLanguageStats { files: number; lines: number }
+export interface RepoTopDirectory { path: string; fileCount: number; totalLines: number }
+export interface RepoGitStatusSummary { uncommitted: number; branch: string | null }
+export interface RepoAnalyzeResult {
+  ok: true;
+  cwd: string;
+  totalFiles: number;
+  totalLines: number;
+  languageBreakdown: Record<string, RepoLanguageStats>;
+  topDirectories: RepoTopDirectory[];
+  gitStatus: RepoGitStatusSummary;
+  truncated: boolean;
+  durationMs: number;
+}
+export interface RepoAnalyzeError { ok: false; error: string }
+
 export interface WatcherInfo {
   watcherId: string;
   tabId: string;
@@ -788,6 +829,13 @@ export interface SessionManagerAPI {
     openExternal: (path: string) => Promise<FilesShellResult>;
     showInFinder: (path: string) => Promise<FilesShellResult>;
   };
+  search: {
+    files: (cwd: string, query?: string, opts?: { limit?: number }) => Promise<SearchFilesResult>;
+    text: (cwd: string, query: string, opts?: { limit?: number; caseSensitive?: boolean }) => Promise<SearchTextResult>;
+  };
+  repo: {
+    analyze: (cwd: string) => Promise<RepoAnalyzeResult | RepoAnalyzeError>;
+  };
   history: {
     aggregate: (req?: HistoryAggregateRequest) => Promise<HistoryAggregateResult>;
     listConversations: () => Promise<ListConversationsResult>;

package/src/preload/index.cjs

@@ -163,6 +163,13 @@ contextBridge.exposeInMainWorld('api', {
     openExternal: (path) => ipcRenderer.invoke('files:open-external', { path }),
     showInFinder: (path) => ipcRenderer.invoke('files:show-in-finder', { path }),
   },
+  search: {
+    files: (cwd, query, opts) => ipcRenderer.invoke('search:files', { cwd, query, opts }),
+    text: (cwd, query, opts) => ipcRenderer.invoke('search:text', { cwd, query, opts }),
+  },
+  repo: {
+    analyze: (cwd) => ipcRenderer.invoke('repo:analyze', { cwd }),
+  },
   schedule: {
     state: () => ipcRenderer.invoke('schedule:state'),
     setConfig: (partial) => ipcRenderer.invoke('schedule:set-config', partial),