hob-bot 1.0.3 → 1.1.0

package/server/src/files.ts

@@ -1,6 +1,6 @@
 // /api/files 的取文件规则：路径先解开符号链接，再判断能不能给页面看。
-import { closeSync, openSync, readSync, readdirSync, realpathSync, statSync } from "node:fs";
-import { basename, join, relative, resolve, sep } from "node:path";
+import { closeSync, existsSync, lstatSync, openSync, readSync, readdirSync, realpathSync, statSync } from "node:fs";
+import { basename, isAbsolute, join, relative, resolve, sep } from "node:path";
 import { fileKind, isTextKind, mimeOf } from "../../shared/files.ts";
 import type { FileKind, JobFileEntry, JobFilesView } from "../../shared/types.ts";
 
@@ -11,13 +11,15 @@ const DENY_ROOT_DIRS = new Set(["engine", "lore"]);
 /** 备份文件一律不给（凭据的备份也在里面）。 */
 const DENY_ANY_FILE = /\.bak/i;
 
+/** 任务工作目录里不对页面开放的：密钥目录、git 内部、.env、私钥。 */
+const DENY_WORK_DIRS = /(^|[/\\])\.(ssh|gnupg|aws|git)([/\\]|$)/;
+const DENY_WORK_FILE = /^(\.env.*|id_[a-z0-9]+|.*\.(pem|key))$/i;
+
 /** 预览面板最多拉 1 MB 文本。 */
 export const MAX_TEXT_BYTES = 1024 * 1024;
 
 /** 任务工作目录里最多列 200 个文件。 */
 const MAX_JOB_FILES = 200;
-/** 递归太深的目录不进去（产出文件不该有这么深）。 */
-const MAX_DEPTH = 8;
 const SKIP_DIRS = new Set(["node_modules", ".git"]);
 
 export interface FileInfo {
@@ -68,10 +70,11 @@ function headOf(path: string): Uint8Array {
 }
 
 /**
- * 解析一个路径：realpath 后必须还在数据目录里（符号链接指出去就是 403），
+ * 解析一个路径：realpath 后必须在数据目录或某个任务的工作目录里（符号链接指出去就是 403），
  * 凭据和内部状态文件不开放，不是普通文件当 404。
+ * 包含数据目录的工作目录（家目录、根目录这类）太宽，不算数。
  */
-export function lookupFile(home: string, requested: string): Lookup {
+export function lookupFile(home: string, requested: string, workDirs: string[] = []): Lookup {
   const root = realpathSync(resolve(home));
   let path: string;
   let size: number;
@@ -85,58 +88,75 @@ export function lookupFile(home: string, requested: string): Lookup {
   } catch {
     return { ok: false, status: 404, error: "文件不存在" };
   }
-  if (!path.startsWith(`${root}${sep}`)) return { ok: false, status: 403, error: "只允许读 Hob 自己的文件" };
-  if (denied(root, path)) return { ok: false, status: 403, error: "这个文件不对页面开放" };
+  const inHome = path.startsWith(`${root}${sep}`);
+  const work = inHome ? undefined : workRoots(root, workDirs).find((dir) => path.startsWith(`${dir}${sep}`));
+  if (!inHome && !work) return { ok: false, status: 403, error: "只允许读 Hob 自己的文件" };
+  const relativePath = relative(inHome ? root : work!, path);
+  if ((inHome && denied(root, path)) || DENY_WORK_DIRS.test(relativePath) || DENY_WORK_FILE.test(basename(path)) || DENY_ANY_FILE.test(basename(path))) {
+    return { ok: false, status: 403, error: "这个文件不对页面开放" };
+  }
   const name = basename(path);
   const kind = fileKind(name) ?? (looksText(headOf(path)) ? "text" : "other");
   const text = isTextKind(kind);
   return {
     ok: true,
-    info: { path, name, size, modifiedAt, kind, text, mime: mimeOf(name), ...(/^pages[/\\][0-9a-f-]{36}\.html$/i.test(relative(root, path)) ? { page: true } : {}) },
+    info: { path, name, size, modifiedAt, kind, text, mime: mimeOf(name), ...(/^(?:pages[/\\][0-9a-f-]{36}|knowledge[/\\]pages[/\\][0-9a-f]{64})\.html$/i.test(relative(root, path)) ? { page: true } : {}) },
   };
 }
 
-function walk(dir: string, base: string, depth: number, out: JobFileEntry[]): boolean {
-  if (depth > MAX_DEPTH) return false;
+function workRoots(root: string, dirs: string[]): string[] {
+  const out: string[] = [];
+  for (const dir of new Set(dirs)) {
+    try {
+      const real = realpathSync(resolve(dir));
+      if (!`${root}${sep}`.startsWith(`${real === sep ? "" : real}${sep}`)) out.push(real);
+    } catch {
+      // 目录已经删了：不算
+    }
+  }
+  return out;
+}
+
+function visibleEntries(dir: string) {
+  return readdirSync(dir, { withFileTypes: true }).filter((entry) =>
+    !entry.isSymbolicLink() && (entry.isFile() || (entry.isDirectory() && !SKIP_DIRS.has(entry.name))));
+}
+
+/** 一次只列一层；拒绝越界和符号链接。工作目录还没建出来时当作空的。 */
+export function listJobFiles(workDir: string, subdir = ""): JobFilesView {
+  const parts = subdir.split(/[\\/]/).filter(Boolean);
+  if (isAbsolute(subdir) || parts.some((part) => part === ".." || SKIP_DIRS.has(part))) throw new Error("工作目录路径无效");
+  const base = resolve(workDir);
+  if (!parts.length && !existsSync(base)) return { files: [], total: 0, fileCount: 0, directoryCount: 0 };
+  let dir = base;
   let entries;
   try {
-    entries = readdirSync(dir, { withFileTypes: true });
-  } catch {
-    return false;
+    for (const part of parts) {
+      dir = join(dir, part);
+      const info = lstatSync(dir);
+      if (info.isSymbolicLink() || !info.isDirectory()) throw new Error("不是文件夹");
+    }
+    entries = visibleEntries(dir);
+  } catch (error) {
+    const code = (error as NodeJS.ErrnoException).code;
+    throw new Error(code === "ENOENT" ? "这个文件夹已经不在了" : code === "EACCES" ? "没有权限读取这个文件夹" : "工作目录路径无效");
   }
+  const all: JobFileEntry[] = [];
   for (const entry of entries) {
-    // 符号链接不跟：可能指到目录外面，也可能绕成环
-    if (entry.isSymbolicLink()) continue;
     const path = join(dir, entry.name);
-    if (entry.isDirectory()) {
-      if (SKIP_DIRS.has(entry.name)) continue;
-      if (walk(path, base, depth + 1, out)) return true;
-      continue;
-    }
-    if (!entry.isFile()) continue;
-    let stats;
+    const common = { path, rel: relative(base, path), name: entry.name };
     try {
-      stats = statSync(path);
+      if (entry.isDirectory()) {
+        all.push({ ...common, type: "directory", count: visibleEntries(path).length });
+      } else {
+        const stats = statSync(path);
+        all.push({ ...common, type: "file", size: stats.size, modifiedAt: stats.mtime.toISOString(), kind: fileKind(entry.name) ?? (looksText(headOf(path)) ? "text" : "other") });
+      }
     } catch {
-      continue;
+      // 列的时候刚被删掉或读不了（在跑的任务常有临时文件）：跳过这一项
     }
-    out.push({
-      path,
-      rel: relative(base, path),
-      name: entry.name,
-      size: stats.size,
-      modifiedAt: stats.mtime.toISOString(),
-      kind: fileKind(entry.name) ?? (looksText(headOf(path)) ? "text" : "other"),
-    });
-    if (out.length >= MAX_JOB_FILES + 1) return true;
   }
-  return false;
-}
-
-/** 任务工作目录里的文件：递归、修改时间倒序、最多 200 个，跳过 node_modules 和 .git。 */
-export function listJobFiles(workDir: string): JobFilesView {
-  const all: JobFileEntry[] = [];
-  walk(workDir, workDir, 0, all);
-  all.sort((a, b) => b.modifiedAt.localeCompare(a.modifiedAt));
-  return { files: all.slice(0, MAX_JOB_FILES), hasMore: all.length > MAX_JOB_FILES };
+  all.sort((a, b) => a.type !== b.type ? (a.type === "directory" ? -1 : 1) : a.type === "file" && b.type === "file" ? b.modifiedAt.localeCompare(a.modifiedAt) || a.name.localeCompare(b.name) : a.name.localeCompare(b.name));
+  const directoryCount = all.filter((entry) => entry.type === "directory").length;
+  return { files: all.slice(0, MAX_JOB_FILES), total: all.length, directoryCount, fileCount: all.length - directoryCount };
 }