AskaEth
f82c76df27
CI / Backend (build / vet / test / cross-compile) (push) Has been cancelled
CI / Frontend (tsc + vite build) (push) Has been cancelled
feat(documents): PDF 文本层抽取(S6 第一段;扫描件仍诚实拒绝)
- 依赖:github.com/ledongthuc/pdf(**BSD-3-Clause**,非 MIT —— 子代理核了 LICENSE 并纠正了任务书的写法;零传递依赖;无 cgo/无外部进程;体积 strip 后 +285.5 KiB / +0.74%,在 detached worktree 里量取,产出写到仓库外)。被否候选:pdfcpu 无文本抽取、rsc.io/pdf 停更、pdfium/poppler/mupdf 需 cgo 撞红线。
- 实现:魔数 %PDF- 是唯一入口判定(名字/声明指向 PDF 而字节无魔数 ⇒ ErrUnsupported,绝不退回纯文本抽);逐页 GetPlainText(页间仅去首尾空白);ExtractorVersion v2→v3(所有 sidecar 一次性重抽);页数上限 5000 + 字节安全上限复用;新增哨兵 ErrPDFEncrypted / ErrPDFDamaged。
- 诚实边界:无文本层(扫描件)⇒ ErrNoTextLayer 且文案含「图片型/无文本层 PDF,需要 OCR(当前未做)」;加密/损坏各自可判、绝不 panic、不回空正文假装成功。
- 白名单与成对裁决:.pdf 纳入(22→23),且 ContentAllowed 做双向裁决(.pdf 只收嗅探为 application/pdf;PDF 字节只收在 .pdf 名下);读侧真 PDF 如实下发 application/pdf,仍 attachment + nosniff(不可内联)。
- 1 处 panic 兜底已如实登记:该库对「xref 指向文件外」畸形件会在 NumPage() 直接 panic(已用仓库外程序复现),单二进制常驻后端不能因 587 字节上传件崩掉 ⇒ defer recover 折成 ErrPDFDamaged;损坏类主用例仍走库的真出错路径。
- 主会话独立验收:隔离 worktree(B1 的 14 文件)build/vet 绿、全树 59 包 0 FAIL、PDF 用例 13 PASS、双平台交叉编译绿;变异独立复现(魔数裁决短路 ⇒ 魔数/成对裁决用例红)。
- ⚠️ 契约漂移(已排 C2):Docs 仍写「22 项白名单 / 嗅探仅放行 text/plain」,与本切片的 23 项 + 成对裁决不一致 ⇒ 需升 v0.61 登记 .pdf 与新的边界。
2026-10-04 19:46:36 +08:00
..
2026-10-04 12:30:16 +08:00
2026-10-04 12:30:16 +08:00
2026-10-04 14:21:41 +08:00
2026-10-04 13:42:37 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:28:25 +08:00
2026-10-04 14:00:25 +08:00
2026-10-03 22:28:25 +08:00
2026-10-04 12:43:04 +08:00
2026-10-04 12:43:04 +08:00
2026-10-04 12:43:04 +08:00
2026-10-04 12:43:04 +08:00
2026-10-04 18:11:43 +08:00
2026-10-04 13:06:54 +08:00
2026-10-04 18:11:43 +08:00
2026-10-04 10:46:51 +08:00
2026-10-04 13:07:52 +08:00
2026-10-04 14:21:41 +08:00
2026-10-04 14:21:41 +08:00
2026-10-03 22:28:25 +08:00
2026-10-04 14:21:41 +08:00
2026-10-04 12:30:16 +08:00
2026-10-04 14:21:41 +08:00
2026-10-03 22:28:25 +08:00
2026-10-04 19:44:53 +08:00
2026-10-04 19:44:53 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:28:25 +08:00
2026-10-04 13:42:37 +08:00
2026-10-04 18:57:58 +08:00
2026-10-04 18:11:43 +08:00
2026-10-03 22:28:25 +08:00
2026-10-04 13:07:52 +08:00
2026-10-03 22:28:25 +08:00
2026-10-04 12:07:51 +08:00
2026-10-04 13:07:52 +08:00
2026-10-03 22:28:25 +08:00
2026-10-04 14:21:41 +08:00
2026-10-03 22:28:25 +08:00
2026-10-04 18:11:43 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:07:56 +08:00
2026-10-04 17:59:28 +08:00
2026-10-04 17:59:28 +08:00
2026-10-04 19:46:36 +08:00
2026-10-04 14:37:37 +08:00
2026-10-04 17:25:35 +08:00
2026-10-03 22:28:25 +08:00
2026-10-04 19:44:53 +08:00
2026-10-04 19:46:36 +08:00
2026-10-04 17:07:20 +08:00
2026-10-04 19:46:36 +08:00
2026-10-04 14:37:37 +08:00
2026-10-03 22:28:25 +08:00
2026-10-04 10:37:22 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:28:25 +08:00
2026-10-04 17:42:16 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:07:56 +08:00
2026-10-04 14:13:27 +08:00
2026-10-04 14:13:27 +08:00
2026-10-04 18:57:58 +08:00
2026-10-04 14:21:52 +08:00
2026-10-04 14:21:52 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:07:56 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:28:25 +08:00
2026-10-04 13:58:03 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:28:25 +08:00
2026-10-04 13:58:03 +08:00
2026-10-04 13:58:03 +08:00
2026-10-04 13:58:03 +08:00
2026-10-04 13:58:03 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:28:25 +08:00
2026-10-04 14:21:41 +08:00
2026-10-04 16:27:35 +08:00
2026-10-04 17:25:35 +08:00
2026-10-04 19:44:53 +08:00
2026-10-04 19:44:53 +08:00
2026-10-03 22:28:25 +08:00
2026-10-04 17:42:16 +08:00
2026-10-04 17:42:16 +08:00
2026-10-04 17:42:16 +08:00
2026-10-04 14:53:40 +08:00
2026-10-04 14:21:52 +08:00
2026-10-03 22:28:25 +08:00
2026-10-04 17:59:28 +08:00
2026-10-04 17:59:28 +08:00
2026-10-04 17:42:16 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:07:56 +08:00
2026-10-04 17:59:28 +08:00
2026-10-04 14:53:40 +08:00
2026-10-04 17:59:28 +08:00
2026-10-04 17:25:35 +08:00
2026-10-04 17:42:16 +08:00
2026-10-04 14:17:58 +08:00
2026-10-04 17:59:28 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:10:33 +08:00
2026-10-04 17:59:28 +08:00
2026-10-04 17:59:28 +08:00
2026-10-04 12:07:46 +08:00
2026-10-04 14:53:40 +08:00
2026-10-04 12:10:21 +08:00
2026-10-03 22:07:56 +08:00
2026-10-03 22:10:33 +08:00
2026-10-04 17:59:28 +08:00
2026-10-03 22:07:56 +08:00
2026-10-03 22:28:25 +08:00
2026-10-03 22:07:56 +08:00
2026-10-04 17:25:35 +08:00
2026-10-04 17:25:35 +08:00
2026-10-03 22:10:33 +08:00
2026-10-04 14:25:13 +08:00
2026-10-04 17:59:28 +08:00
2026-10-04 11:26:35 +08:00