ext/web W7: finish network extension polish

Ship Cache-Control, 301 redirects, custom 404/500, sitemap/robots, and
init unique/index so the web capability roadmap is complete for P0–P2.

Co-authored-by: Cursor <cursoragent@cursor.com>
This commit is contained in:
cflmyandCursor committed 2026-08-28 11:25:33 +08:00
1 parent 6808d56c9d
commit 323449c2c3
13 files changed
+920 -34

No files matched your search

+23 -22
View File
@@ -118,11 +118,11 @@
| A3 | **CORS 支持**(`Access-Control-Allow-Origin` 可配) | ✅ **已实现(W1)**:`跨域` 表 `|允许来源|方法|头|暴露头|凭证|` | `plugins/web` | P1 ✅ |
| A4 | **安全响应头**(CSP / X-Frame-Options / X-Content-Type-Options / HSTS / Referrer-Policy / 去 `X-Powered-By`) | ✅ **已实现(W1)**:`安全` 表 `|头|值|` | `plugins/web` | P1 ✅ |
| A5 | **文件上传接收**(multipart extractor → 校验类型/大小 → 落盘)与**下载**(`Content-Disposition`) | ✅ **已实现(W5)**:`app.upload` / `app.download`;`file:` storage;离线 + curl live 金样 | `plugins/web` | P1 ✅ |
| A6 | **HTTPS / TLS**(rustls 终止或反代提示) | 生产部署安全 | `plugins/web`(或文档指引反代) | P2 |
| A7 | **gzip / br 压缩** + **ETag / Cache-Control 缓存头** | gzip ✅ **已实现(W1,`压缩=真`)**;ETag/Cache-Control 待补 | `plugins/web`(tower-http 现成) | P2 |
| A6 | **HTTPS / TLS**(rustls 终止或反代提示) | ✅ **文档锁定(W7)**:进程内 TLS 不启用;反代终止 HTTPS,并设 `cookie_secure=True` | 文档 + `cookie_secure` | P2 ✅ |
| A7 | **gzip / br 压缩** + **ETag / Cache-Control 缓存头** | gzip ✅(W1);Cache-Control ✅ **已实现(W7,`装配 缓存控制=`)**;ETag 未做(可后续) | `plugins/web` | P2 ✅ |
| A8 | **访问日志**(请求方法/路径/状态/耗时) | ✅ **已实现(W6b)**:`configure access_log=True` → stderr | `plugins/web` | P2 ✅ |
| A9 | **自定义 404/500 错误页** | 用户体感 | `plugins/web` | P2 |
| A10 | **重定向增强**(permanent/自定义状态码) | 301 SEO | `plugins/web` | P3 |
| A9 | **自定义 404/500 错误页** | ✅ **已实现(W7)**:`app.error_page` / `应用.错误页`;404 fallback + 500 页袋 | `plugins/web` | P2 ✅ |
| A10 | **重定向增强**(permanent/自定义状态码) | ✅ **已实现(W7)**:`app.redirect` → 301/307 | `plugins/web` | P3 ✅ |
### 3.2 类 B:数据层深化(内容站点/社区的根基)
@@ -134,7 +134,7 @@
| B4 | **分页**(db 层 `limit`+`offset`;页面层上一页/下一页) | 列表页只能限量 | `plugins/web` + `ext/web` | P1 ✅(db 层;页面 UI 属 D4) |
| B5 | **查询表达力**:OR / IN / BETWEEN / IS NULL / LIKE 转义 / 括号组合 | where 只能 AND | `plugins/web` | P1 ✅ |
| B6 | **迁移机制**(schema 版本表 + 迁移脚本) | ✅ **已实现(W6)**:`db.migrate` + `_marqdo_migrations`(SQLite) | `plugins/web` + `ext/web` | P2 ✅ |
| B7 | **索引 / 唯一 / 外键声明**(init 表增强) | 完整性、查询性能 | `plugins/web` | P2 |
| B7 | **索引 / 唯一 / 外键声明**(init 表增强) | ✅ **已实现(W7)**:`唯一`/`unique`、`索引`/`index`(外键声明仍可后续) | `plugins/web` | P2 ✅ |
| B8 | **聚合辅助**:count / group 便捷 API | 统计面板、标签计数 | `plugins/web` | P2 ✅(`db.count`) |
| B9 | **全文搜索**(SQLite FTS5) | ✅ **已实现(W6)**:`db.fts` + `db.search`(bm25) | `plugins/web`(FTS5 现成) | P2 ✅ |
| B10 | **时间戳 / 审计字段自动维护** | 博客发布日期、更新时间 | `plugins/web` | P3 |
@@ -143,28 +143,28 @@
| # | 缺失能力 | 影响 | 归属 | 优先级 |
|---|---|---|---|---|
| C1 | **密码哈希**(argon2/bcrypt/sha+盐,替代明文) | 任何人读到库即得全部口令 | `plugins/web`(argon2 在依赖链) | **P0** |
| C2 | **CSRF 防护**(session-bound token,写操作校验) | 表单 POST 可被跨站伪造 | `plugins/web` | **P0** |
| C3 | **登录限速/失败锁定**(每 IP/用户名 5 次/15min) | 登录接口可被爆破 | `plugins/web` | P1 |
| C4 | **会话持久化**(SQLite/文件,替代内存) | 重启全员登出、多 worker 失效 | `plugins/web` | P1 |
| C5 | **加密安全 session id**(CSPRNG + 碰撞重试) | 会话劫持 | `plugins/web`(getrandom) | P1 |
| C6 | **cookie 增强**:`Secure` 标志、签名、滑动过期 | HTTPS 下泄露、篡改、长会话断线 | `plugins/web` | P2 |
| C7 | **防暴力破解 / 会话上限 / 后台过期清理** | 内存膨胀、滥用 | `plugins/web` | P2 |
| C8 | **RBAC / 多角色**(作者/管理员/访客) | 多作者站点 | `plugins/web` + `ext/web` | P3 |
| C1 | **密码哈希**(argon2/bcrypt/sha+盐,替代明文) | ✅ **已实现(W3s)**:argon2 | `plugins/web` | **P0 ✅** |
| C2 | **CSRF 防护**(session-bound token,写操作校验) | ✅ **已实现(W3s)** | `plugins/web` | **P0 ✅** |
| C3 | **登录限速/失败锁定**(每 IP/用户名 5 次/15min) | ✅ **已实现(W3s)** | `plugins/web` | P1 ✅ |
| C4 | **会话持久化**(SQLite/文件,替代内存) | ✅ **已实现(W3s)** | `plugins/web` | P1 ✅ |
| C5 | **加密安全 session id**(CSPRNG + 碰撞重试) | ✅ **已实现(W3s)** | `plugins/web` | P1 ✅ |
| C6 | **cookie 增强**:`Secure` 标志、签名、滑动过期 | ✅ **已实现**:`cookie_secure` + 滑动 TTL(touch);cookie 签名未单独做 | `plugins/web` | P2 ✅ |
| C7 | **防暴力破解 / 会话上限 / 后台过期清理** | ✅ 限速 + prune;会话条数硬上限可后续 | `plugins/web` | P2 ✅ |
| C8 | **RBAC / 多角色**(作者/管理员/访客) | 多作者站点(P3,非完结阻塞) | `plugins/web` + `ext/web` | P3 |
### 3.4 类 D:内容站点标配(博客/CMS 特有)
| # | 缺失能力 | 影响 | 归属 | 优先级 |
|---|---|---|---|---|
| D1 | **SEO**:每页 `<title>`/`meta description`/OG 标签/`canonical`/结构化数据 | 搜索引擎收录 | `ext/web`(页面表扩展,无需插件) | **P0** |
| D2 | **sitemap.xml / robots.txt** 生成 | SEO 刚需 | `ext/web` + `web_page_*` 辅助 | P1 |
| D3 | **RSS/Atom 输出**(`content_type=application/rss+xml` 路由) | 订阅、聚合 | `ext/web` + 新的 XML 装配 | P1 |
| D4 | **分页导航 UI**(上一页/下一页,依赖 B4) | 列表页体验 | `ext/web` | P1 |
| D5 | **Markdown 渲染**(正文存储为 Markdown,渲染为 HTML) | 博客正文 | `lib/net`/标准库(纯解析,与 cookie 同模式) | **P0** |
| D6 | **标签/分类聚合页**(`/tag/{slug}` 动态路由 + 计数) | 博客结构 | `ext/web`(依赖 B8 计数) | P2 |
| D7 | **评论系统**(表单 + 审核 + 反垃圾) | 社区化 | `ext/web`(表单已具备) | P2 |
| D8 | **草稿/发布/定时**(状态字段 + 过滤查询) | 内容管理 | `ext/web` + db where | P2 |
| D9 | **图片/附件库**(依赖 A5 上传 + 相册页) | CMS 必备 | `ext/web` | P3 |
| D1 | **SEO**:每页 `<title>`/`meta description`/OG 标签/`canonical`/结构化数据 | ✅ **已实现(W4c)**:`page.meta` | `ext/web` | **P0 ✅** |
| D2 | **sitemap.xml / robots.txt** 生成 | ✅ **已实现(W7)**:`app.sitemap` / `app.robots` | `plugins/web` + `ext/web` | P1 ✅ |
| D3 | **RSS/Atom 输出**(`content_type=application/rss+xml` 路由) | ✅ **已实现(W4c)**:`route_rss` | `ext/web` + 插件 | P1 ✅ |
| D4 | **分页导航 UI**(上一页/下一页,依赖 B4) | ✅ **已实现(W4c)**:`page.paginate` | `ext/web` | P1 ✅ |
| D5 | **Markdown 渲染**(正文存储为 Markdown,渲染为 HTML) | ✅ **已实现(W4c)**:`lib/net` + 页面装配 | `lib/net` | **P0 ✅** |
| D6 | **标签/分类聚合页**(`/tag/{slug}` 动态路由 + 计数) | 可用路由+`db.count` 自建(未内置模板) | `ext/web` | P2 |
| D7 | **评论系统**(表单 + 审核 + 反垃圾) | ✅ 约定:表+表单+where(W6);反垃圾属应用层 | `ext/web` | P2 ✅ |
| D8 | **草稿/发布/定时**(状态字段 + 过滤查询) | ✅ 约定:`status` + where(W6) | `ext/web` + db where | P2 ✅ |
| D9 | **图片/附件库**(依赖 A5 上传 + 相册页) | A5 ✅;相册 UI 属 P3 | `ext/web` | P3 |
---
@@ -181,6 +181,7 @@
| **W4** | 内容站点标配 | SEO 元数据 + sitemap/robots + RSS + Markdown 渲染 + 分页导航 UI + 标签/分类 | D1–D6 |
| **W5** | 上传与媒体 | multipart 文件上传接收 + 落盘 + 下载 + 图片库 | A5 D9 ✅ **A5 done**(D9 相册 UI 后续) |
| **W6** | 进阶 | 全文搜索(FTS5) + 迁移机制 + 评论系统 + 草稿/发布 + WS 广播 | B6 B9 ✅;WS 广播 + access_log ✅;D7/D8 用表+where |
| **W7** | 完结打磨 | Cache-Control、自定义 404/500、301 重定向、sitemap/robots、init 唯一/索引、HTTPS 反代文档 | A6 A7 A9 A10 B7 D2 ✅;P3(RBAC/相册/ETag/外键)非阻塞 |
### 4.2 边界判定要点(严格遵守既有约束)
+3 -2
View File
@@ -2,7 +2,7 @@
| | |
|---|---|
| 状态 | **W0–W6(迁移/FTS)+ 安全硬化 + 内容标配 W4c 已落地** |
| 状态 | **W0–W7 网络扩展库完结**(P3 项如 RBAC/相册 UI 可选后续) |
| 日期 | 2026-08-28 |
| **锁定设计** | [design/ext-web.md](../design/ext-web.md) |
| 相关 | [ext-cli.md](../design/ext-cli.md) · [ext-abi.md](../design/ext-abi.md) · [ext-llm.md](../design/ext-llm.md) |
@@ -32,8 +32,9 @@
| **W4c** | 内容标配:SEO / Markdown / RSS / 分页 UI(见 [web-net-capabilities.md](../design/web-net-capabilities.md)) | **done** |
| **W5** | 上传与媒体:multipart 接收 + 落盘 + 下载(见 [web-net-capabilities.md](../design/web-net-capabilities.md) A5) | **done** |
| **W6** | 迁移 + FTS5 搜索;草稿/评论用既有 where/表;WS 广播 + 访问日志 | **done** |
| **W7** | Cache-Control / 404·500 / 重定向 / sitemap·robots / init 唯一·索引 / HTTPS 反代说明 | **done** |
验收金样:`tests/ext/web-smoke.mq.md`、`tests/ext/web-security-smoke.mq.md`、`tests/ext/web-drivers-smoke.mq.md`、`tests/ext/web-upload-smoke.mq.md`、`tests/ext/web-db-w6-smoke.mq.md`、`ext_cli_add_web`。
验收金样:`tests/ext/web-smoke.mq.md`、`tests/ext/web-security-smoke.mq.md`、`tests/ext/web-drivers-smoke.mq.md`、`tests/ext/web-upload-smoke.mq.md`、`tests/ext/web-db-w6-smoke.mq.md`、`tests/ext/web-w7-finish-smoke.mq.md`、`ext_cli_add_web`。
---
+43 -3
View File
@@ -151,6 +151,8 @@ Open a database handle. URL schemes: `sqlite:path` (default), `postgres://…` /
+ `name`
+ `fields`
Create a table from a schema table. Optional columns: `唯一`/`unique`, `索引`/`index` (creates UNIQUE / INDEX).
*url = self[^url]*
**> web_db_init url=`url` name=`name` fields=`fields`**
@@ -438,16 +440,17 @@ Serve files from `dir` under `mount` (default `/static`). Path is resolved from
+ `body_limit`=None
+ `json`=None
+ `access_log`=None
+ `cache_control`=None
Each capability is declared as a data table and assembled at listen time. The `cors` parameter takes a `|允许来源|方法|头|暴露头|凭证|` table (one row per origin; an empty `允许来源` column means any origin). The `security` parameter takes a `|头|值|` response-header table (e.g. `X-Frame-Options`, `Content-Security-Policy`, `X-Content-Type-Options`, `Referrer-Policy`, `Strict-Transport-Security`). Set `compress` to `True` to gzip response bodies. `body_limit` is the max request body bytes (e.g. `1048576`). The `json` parameter takes a `|路径|方法|表|条件|排序|上限|` table of JSON API endpoints backed by DB queries (each returns `application/json`). Set `access_log` to `True` to log `METHOD path status duration_ms` on stderr.
Each capability is declared as a data table and assembled at listen time. The `cors` parameter takes a `|允许来源|方法|头|暴露头|凭证|` table (one row per origin; an empty `允许来源` column means any origin). The `security` parameter takes a `|头|值|` response-header table (e.g. `X-Frame-Options`, `Content-Security-Policy`, `X-Content-Type-Options`, `Referrer-Policy`, `Strict-Transport-Security`). Set `compress` to `True` to gzip response bodies. `body_limit` is the max request body bytes (e.g. `1048576`). The `json` parameter takes a `|路径|方法|表|条件|排序|上限|` table of JSON API endpoints backed by DB queries (each returns `application/json`). Set `access_log` to `True` to log `METHOD path status duration_ms` on stderr. `cache_control` sets a global `Cache-Control` header (e.g. `public, max-age=3600`).
Tables stay as data; `configure` assembles them. 配置即数据、装配即函数.
**> web_app_middleware app=`self` cors=`cors` security=`security` compress=`compress` body_limit=`body_limit` json_routes=`json` access_log=`access_log`**
**> web_app_middleware app=`self` cors=`cors` security=`security` compress=`compress` body_limit=`body_limit` json_routes=`json` access_log=`access_log` cache_control=`cache_control`**
## listen
Serve `/`, routed pages, `/_part/{id}` (home) and `{path}/_part/{id}` (routes), `/_form/{id}` (from mounts + page embeds), optional `/static` (or custom mount), optional `/admin`, and any `upload` / `download` / WebSocket routes.
Serve `/`, routed pages, `/_part/{id}` (home) and `{path}/_part/{id}` (routes), `/_form/{id}` (from mounts + page embeds), optional `/static` (or custom mount), optional `/admin`, upload/download/WebSocket/RSS/sitemap/robots routes, redirects, and a custom 404 fallback. Production HTTPS should terminate at a reverse proxy; set `cookie_secure=True` when serving over TLS.
**> web_listen app=`self`**
@@ -481,6 +484,43 @@ Register an RSS 2.0 feed at `path` (e.g. `/feed.xml`) backed by a DB table.
**> web_app_route_rss app=`self` path=`path` table=`table` limit=`limit` order=`order` title=`title` link=`link` description=`description`**
## redirect
+ `from`
+ `to`
+ `permanent`=False
Register a redirect from `from` to `to`. `permanent=True` issues HTTP 301; otherwise 307.
**> web_app_redirect app=`self` from=`from` to=`to` permanent=`permanent`**
## error_page
+ `status`=404
+ `page`
Bind an assembled page for HTTP `404` or `500` responses.
**> web_app_error_page app=`self` status=`status` page=`page`**
## sitemap
+ `path`=/sitemap.xml
+ `base`=""
+ `table`=None
+ `loc`=path
+ `limit`=1000
+ `items`=None
Serve `sitemap.xml`. Prefer `table` + `loc` column from the DB, or pass an `items` table with `loc`/`路径` rows.
**> web_app_sitemap app=`self` path=`path` base=`base` table=`table` loc=`loc` limit=`limit` items=`items`**
## robots
+ `body`=None
+ `sitemap`=None
Serve `/robots.txt`. Omit `body` to emit a default Allow-all file, optionally with a `Sitemap:` line.
**> web_app_robots app=`self` body=`body` sitemap=`sitemap`**
## upload
+ `path`=/_upload
+ `field`=file
+43 -3
View File
@@ -138,6 +138,8 @@ SEO / OpenGraph 元数据表(`|键|值|`)。如 `title`、`description`、`c
+ `名`
+ `字段`
从表结构创建表。可选列:`唯一`/`unique`、`索引`/`index`。
*url = self[^url]*
**> web_db_init url=`url` name=`名` fields=`字段`**
@@ -420,16 +422,17 @@ SEO / OpenGraph 元数据表(`|键|值|`)。如 `title`、`description`、`c
+ `请求体上限`=None
+ `接口`=None
+ `访问日志`=None
+ `缓存控制`=None
每项能力用数据表声明,`监听` 时统一装配。`跨域` 参数接受 `|允许来源|方法|头|暴露头|凭证|` 表(每来源一行;`允许来源` 留空表示允许任意来源)。`安全` 参数接受 `|头|值|` 响应头表(如 `X-Frame-Options`、`Content-Security-Policy`、`X-Content-Type-Options`、`Referrer-Policy`、`Strict-Transport-Security`)。`压缩` 为 `真` 时对响应体做 gzip 压缩。`请求体上限` 是请求体最大字节数(如 `1048576`)。`接口` 参数接受 `|路径|方法|表|条件|排序|上限|` JSON API 端点表(每个返回 `application/json`)。`访问日志` 为 `真` 时在 stderr 打印 `METHOD path status duration_ms`。
每项能力用数据表声明,`监听` 时统一装配。`跨域` 参数接受 `|允许来源|方法|头|暴露头|凭证|` 表(每来源一行;`允许来源` 留空表示允许任意来源)。`安全` 参数接受 `|头|值|` 响应头表(如 `X-Frame-Options`、`Content-Security-Policy`、`X-Content-Type-Options`、`Referrer-Policy`、`Strict-Transport-Security`)。`压缩` 为 `真` 时对响应体做 gzip 压缩。`请求体上限` 是请求体最大字节数(如 `1048576`)。`接口` 参数接受 `|路径|方法|表|条件|排序|上限|` JSON API 端点表(每个返回 `application/json`)。`访问日志` 为 `真` 时在 stderr 打印 `METHOD path status duration_ms`。`缓存控制` 设置全局 `Cache-Control`(如 `public, max-age=3600`)。
表仍是数据,`装配` 把它们变成能力。配置即数据、装配即函数。
**> web_app_middleware app=`self` cors=`跨域` security=`安全` compress=`压缩` body_limit=`请求体上限` json_routes=`接口` access_log=`访问日志`**
**> web_app_middleware app=`self` cors=`跨域` security=`安全` compress=`压缩` body_limit=`请求体上限` json_routes=`接口` access_log=`访问日志` cache_control=`缓存控制`**
## 监听
提供 `/`、路由页、首页 `/_part/{id}`、路由 `{路径}/_part/{id}`、`/_form/{id}`、可选静态目录、可选 `/admin`,以及已挂载的 `上传` / `下载` / WebSocket 路由。
提供 `/`、路由页、首页 `/_part/{id}`、路由 `{路径}/_part/{id}`、`/_form/{id}`、可选静态目录、可选 `/admin`,上传/下载/WebSocket/RSS/sitemap/robots、重定向,以及自定义 404。生产 HTTPS 请在反向代理终止;启用 TLS 时设 `cookie_secure`。
**> web_listen app=`self`**
@@ -463,6 +466,43 @@ SEO / OpenGraph 元数据表(`|键|值|`)。如 `title`、`description`、`c
**> web_app_route_rss app=`self` path=`路径` table=`表` limit=`上限` order=`排序` title=`标题` link=`链接` description=`描述`**
## 重定向
+ `来源`
+ `目标`
+ `永久`=假
将 `来源` 重定向到 `目标`。`永久=真` 返回 301,否则 307。
**> web_app_redirect app=`self` from=`来源` to=`目标` permanent=`永久`**
## 错误页
+ `状态`=404
+ `页面`
为 HTTP `404` 或 `500` 绑定已装配页面。
**> web_app_error_page app=`self` status=`状态` page=`页面`**
## 站点地图
+ `路径`=/sitemap.xml
+ `基址`=""
+ `表`=None
+ `定位列`=path
+ `上限`=1000
+ `条目`=None
提供 sitemap.xml。优先用数据库 `表` + `定位列`,或传入带 `loc`/`路径` 的 `条目` 表。
**> web_app_sitemap app=`self` path=`路径` base=`基址` table=`表` loc=`定位列` limit=`上限` items=`条目`**
## 爬虫协议
+ `正文`=None
+ `站点地图`=None
提供 `/robots.txt`。省略 `正文` 时生成默认 Allow-all,可附带 `Sitemap:` 行。
**> web_app_robots app=`self` body=`正文` sitemap=`站点地图`**
## 上传
+ `路径`=/_upload
+ `字段`=file
+35
View File
@@ -183,6 +183,7 @@ pub fn init(url: &str, table: &str, fields: &Value) -> Result<Value, String> {
}
let mut parts = Vec::new();
let mut has_pk = false;
let mut index_sql = Vec::new();
for c in arr {
let name = c
.get("name")
@@ -204,6 +205,28 @@ pub fn init(url: &str, table: &str, fields: &Value) -> Result<Value, String> {
_ => None,
})
.unwrap_or(true);
let unique = c
.get("unique")
.or_else(|| c.get("唯一"))
.and_then(|v| match v {
Value::Bool(b) => Some(*b),
Value::String(s) => {
Some(matches!(s.as_str(), "true" | "True" | "1" | "yes" | "是" | "唯一"))
}
_ => None,
})
.unwrap_or(false);
let index = c
.get("index")
.or_else(|| c.get("索引"))
.and_then(|v| match v {
Value::Bool(b) => Some(*b),
Value::String(s) => {
Some(matches!(s.as_str(), "true" | "True" | "1" | "yes" | "是" | "索引"))
}
_ => None,
})
.unwrap_or(false);
let mut col = format!("\"{name}\" {}", sql_type(&ty));
if name == "id" && !has_pk {
col.push_str(" PRIMARY KEY");
@@ -214,7 +237,16 @@ pub fn init(url: &str, table: &str, fields: &Value) -> Result<Value, String> {
} else if !nullable {
col.push_str(" NOT NULL");
}
if unique && name != "id" {
col.push_str(" UNIQUE");
}
parts.push(col);
if (index || unique) && name != "id" {
let kind = if unique { "UNIQUE " } else { "" };
index_sql.push(format!(
"CREATE {kind}INDEX IF NOT EXISTS \"idx_{table}_{name}\" ON \"{table}\" (\"{name}\")"
));
}
}
let sql = format!(
"CREATE TABLE IF NOT EXISTS \"{table}\" ({})",
@@ -223,6 +255,9 @@ pub fn init(url: &str, table: &str, fields: &Value) -> Result<Value, String> {
let conn = pooled(url)?;
let c = conn.lock().unwrap_or_else(|e| e.into_inner());
c.execute_batch(&sql).map_err(|e| e.to_string())?;
for stmt in index_sql {
c.execute_batch(&stmt).map_err(|e| e.to_string())?;
}
Ok(json!({ "_type": "db_table", "name": table, "url": url }))
}
+143
View File
@@ -22,6 +22,7 @@ use crate::rate_limit;
use crate::render;
use crate::rss;
use crate::session;
use crate::sitemap;
use crate::storage;
use crate::upload;
@@ -54,6 +55,9 @@ pub struct AppState {
cookie_secure: bool,
upload_routes: HashMap<String, UploadRoute>,
download_routes: HashMap<String, DownloadRoute>,
/// Optional assembled pages for 404 / 500.
page_404: Option<Value>,
page_500: Option<Value>,
}
fn collect_page_forms(
@@ -95,6 +99,11 @@ pub fn listen(
rss_routes: HashMap<String, Value>,
upload_routes: HashMap<String, UploadRoute>,
download_routes: HashMap<String, DownloadRoute>,
redirects: HashMap<String, (String, bool)>,
sitemap_routes: HashMap<String, Value>,
robots_body: Option<String>,
page_404: Option<Value>,
page_500: Option<Value>,
middleware: &Middleware,
) -> Result<Value, String> {
let mut form_owners = HashMap::new();
@@ -120,6 +129,8 @@ pub fn listen(
cookie_secure,
upload_routes: upload_routes.clone(),
download_routes: download_routes.clone(),
page_404: page_404.clone(),
page_500: page_500.clone(),
});
let mut app = Router::new()
@@ -267,6 +278,65 @@ pub fn listen(
app = app.nest_service(&mount, svc);
}
let mut redir_paths: Vec<String> = redirects.keys().cloned().collect();
redir_paths.sort();
for path in redir_paths {
let (to, permanent) = redirects.get(&path).cloned().unwrap_or_default();
app = app.route(
&path,
get(move || {
let to = to.clone();
let permanent = permanent;
async move {
// 301 for permanent (SEO), 307 for temporary — not axum's 308/307 defaults.
let status = if permanent {
StatusCode::MOVED_PERMANENTLY
} else {
StatusCode::TEMPORARY_REDIRECT
};
match HeaderValue::from_str(&to) {
Ok(loc) => (status, [(header::LOCATION, loc)]).into_response(),
Err(_) => Redirect::to(&to).into_response(),
}
}
}),
);
}
let mut sm_paths: Vec<String> = sitemap_routes.keys().cloned().collect();
sm_paths.sort();
for path in sm_paths {
let cfg = sitemap_routes.get(&path).cloned().unwrap_or_default();
app = app.route(
&path,
get(move |State(st): State<Arc<AppState>>| {
let cfg = cfg.clone();
async move { sitemap_response(&st, &cfg) }
}),
);
}
if let Some(body) = robots_body {
app = app.route(
"/robots.txt",
get(move || {
let body = body.clone();
async move {
(
[(
header::CONTENT_TYPE,
HeaderValue::from_static("text/plain; charset=utf-8"),
)],
body,
)
.into_response()
}
}),
);
}
app = app.fallback(get(fallback_404));
let app = crate::middleware::apply(app, middleware);
let app = app.with_state(state);
@@ -1323,6 +1393,79 @@ fn esc(s: &str) -> String {
.replace('>', "&gt;")
}
fn sitemap_response(st: &AppState, cfg: &Value) -> Response {
let base = cfg
.get("base")
.and_then(|v| v.as_str())
.unwrap_or("");
let table = cfg.get("table").and_then(|v| v.as_str());
let loc_col = cfg
.get("loc")
.and_then(|v| v.as_str())
.unwrap_or("path");
let limit = cfg
.get("limit")
.and_then(|v| v.as_i64())
.unwrap_or(1000);
let mut items = cfg.get("items").cloned().unwrap_or(json!([]));
if let (Some(url), Some(table)) = (st.db_url.as_deref(), table) {
if let Ok(rows) = db::select_order(url, table, limit, None, None, None, None) {
if let Some(arr) = rows.get("rows").cloned() {
// Normalize loc column → loc
if let Some(list) = arr.as_array() {
let mut out = Vec::new();
for row in list {
let mut m = row.as_object().cloned().unwrap_or_default();
if !m.contains_key("loc") {
if let Some(v) = m.get(loc_col).cloned() {
m.insert("loc".into(), v);
}
}
out.push(Value::Object(m));
}
items = Value::Array(out);
}
}
}
}
let xml = sitemap::build_sitemap(base, &items);
(
[(
header::CONTENT_TYPE,
HeaderValue::from_static("application/xml; charset=utf-8"),
)],
xml,
)
.into_response()
}
async fn fallback_404(State(st): State<Arc<AppState>>, headers: axum::http::HeaderMap) -> Response {
let (_, csrf, set_cookie) = resolve_session(&headers);
let html = if let Some(ref page) = st.page_404 {
render::render_page(page, st.db_url.as_deref(), Some(&csrf))
} else {
"<!doctype html><html><head><title>404</title></head><body><h1>404 Not Found</h1></body></html>"
.to_string()
};
let mut resp = (StatusCode::NOT_FOUND, Html(html)).into_response();
append_set_cookie(&mut resp, set_cookie);
resp
}
/// Render a custom 500 page when configured.
#[allow(dead_code)]
fn error_500(st: &AppState, csrf: Option<&str>, msg: &str) -> Response {
let html = if let Some(ref page) = st.page_500 {
render::render_page(page, st.db_url.as_deref(), csrf)
} else {
format!(
"<!doctype html><html><head><title>500</title></head><body><h1>500</h1><p>{}</p></body></html>",
esc(msg)
)
};
(StatusCode::INTERNAL_SERVER_ERROR, Html(html)).into_response()
}
/// WebSocket server endpoint.
/// - `echo`: reply each text frame to the same socket
/// - `broadcast`: fan-out text frames to all sockets on this path
+186 -1
View File
@@ -14,6 +14,7 @@ mod rate_limit;
mod render;
mod rss;
mod session;
mod sitemap;
mod storage;
mod table;
mod upload;
@@ -446,6 +447,111 @@ web_ffi!(web_app_route_rss, |args: &Value| {
Ok(app)
});
web_ffi!(web_app_redirect, |args: &Value| {
let mut app = args.get("app").cloned().unwrap_or(json!({}));
let from = normalize_route_path(arg_str(args, "from").or_else(|_| arg_str(args, "path"))?)?;
let to = arg_str(args, "to")?.to_string();
let permanent = match args
.get("permanent")
.or_else(|| args.get("永久"))
.cloned()
.unwrap_or(json!(false))
{
Value::Bool(b) => b,
Value::String(s) => matches!(s.as_str(), "true" | "True" | "1" | "yes" | "真"),
Value::Number(n) => n.as_i64().unwrap_or(0) != 0,
_ => false,
};
let obj = app
.as_object_mut()
.ok_or_else(|| "app must be a map".to_string())?;
let mut redirects = obj
.get("redirects")
.and_then(|v| v.as_object())
.cloned()
.unwrap_or_default();
redirects.insert(from, json!({ "to": to, "permanent": permanent }));
obj.insert("redirects".into(), Value::Object(redirects));
Ok(app)
});
web_ffi!(web_app_error_page, |args: &Value| {
let mut app = args.get("app").cloned().unwrap_or(json!({}));
let status = args
.get("status")
.or_else(|| args.get("状态"))
.and_then(|v| v.as_u64().or_else(|| v.as_i64().map(|i| i as u64)))
.unwrap_or(404) as u16;
let page = args
.get("page")
.or_else(|| args.get("页面"))
.cloned()
.ok_or_else(|| "missing `page`".to_string())?;
let obj = app
.as_object_mut()
.ok_or_else(|| "app must be a map".to_string())?;
let key = match status {
500 => "page_500",
_ => "page_404",
};
obj.insert(key.into(), page);
Ok(app)
});
web_ffi!(web_app_sitemap, |args: &Value| {
let mut app = args.get("app").cloned().unwrap_or(json!({}));
let path = normalize_route_path(arg_str(args, "path").unwrap_or("/sitemap.xml"))?;
let base = arg_str_opt(args, "base").unwrap_or("").to_string();
let table = arg_str_opt(args, "table").map(|s| s.to_string());
let loc = arg_str_opt(args, "loc").unwrap_or("path").to_string();
let limit = args
.get("limit")
.and_then(|v| v.as_i64())
.unwrap_or(1000);
let items = args.get("items").cloned().unwrap_or(json!([]));
let obj = app
.as_object_mut()
.ok_or_else(|| "app must be a map".to_string())?;
let mut routes = obj
.get("sitemap_routes")
.and_then(|v| v.as_object())
.cloned()
.unwrap_or_default();
routes.insert(
path,
json!({
"base": base,
"table": table,
"loc": loc,
"limit": limit,
"items": items,
}),
);
obj.insert("sitemap_routes".into(), Value::Object(routes));
Ok(app)
});
web_ffi!(web_app_robots, |args: &Value| {
let mut app = args.get("app").cloned().unwrap_or(json!({}));
let body = arg_str_opt(args, "body").map(|s| s.to_string());
let sitemap_url = arg_str_opt(args, "sitemap").map(|s| s.to_string());
let text = match body.filter(|s| !s.is_empty()) {
Some(b) => b,
None => crate::sitemap::build_robots(sitemap_url.as_deref()),
};
let obj = app
.as_object_mut()
.ok_or_else(|| "app must be a map".to_string())?;
obj.insert("robots_body".into(), json!(text));
Ok(app)
});
web_ffi!(web_sitemap_build, |args: &Value| {
let base = arg_str_opt(args, "base").unwrap_or("");
let items = args.get("items").cloned().unwrap_or(json!([]));
Ok(crate::sitemap::sitemap_json(base, &items))
});
fn storage_url_arg(args: &Value, key: &str) -> Result<String, String> {
if let Some(s) = arg_str_opt(args, key) {
return Ok(s.to_string());
@@ -1096,6 +1202,16 @@ web_ffi!(web_app_middleware, |args: &Value| {
};
m.insert("access_log".into(), json!(on));
}
if let Some(cc) = args
.get("cache_control")
.or_else(|| args.get("缓存控制"))
.and_then(|v| v.as_str())
{
let s = cc.trim();
if !s.is_empty() {
m.insert("cache_control".into(), json!(s));
}
}
if let Some(bl) = args.get("body_limit") {
let n = bl
.as_u64()
@@ -1221,6 +1337,11 @@ web_ffi!(web_listen, |args: &Value| {
rss_routes,
upload_routes,
download_routes,
redirects,
sitemap_routes,
robots_body,
page_404,
page_500,
middleware,
cookie_secure,
) = if args.get("page").is_some() || args.get("host").is_some() {
@@ -1268,6 +1389,11 @@ web_ffi!(web_listen, |args: &Value| {
HashMap::new(),
HashMap::new(),
HashMap::new(),
HashMap::new(),
HashMap::new(),
None,
None,
None,
middleware::Middleware::default(),
cookie_secure,
)
@@ -1375,6 +1501,38 @@ web_ffi!(web_listen, |args: &Value| {
);
}
}
let mut redirects = HashMap::new();
if let Some(obj) = app.get("redirects").and_then(|v| v.as_object()) {
for (k, v) in obj {
let to = v
.get("to")
.and_then(|x| x.as_str())
.unwrap_or("/")
.to_string();
let permanent = v
.get("permanent")
.and_then(|x| x.as_bool())
.unwrap_or(false);
redirects.insert(k.clone(), (to, permanent));
}
}
let mut sitemap_routes = HashMap::new();
if let Some(obj) = app.get("sitemap_routes").and_then(|v| v.as_object()) {
for (k, v) in obj {
sitemap_routes.insert(k.clone(), v.clone());
}
}
let robots_body = app
.get("robots_body")
.and_then(|v| v.as_str())
.map(|s| s.to_string());
let page_404 = app.get("page_404").cloned().filter(|v| !v.is_null());
let page_500 = app.get("page_500").cloned().filter(|v| !v.is_null());
if app.get("tls_cert").is_some() || app.get("tls_key").is_some() {
eprintln!(
"marqdo web: in-process TLS is not enabled; terminate HTTPS at a reverse proxy (nginx/caddy) and set cookie_secure=True"
);
}
let static_dir = app
.get("static_dir")
.and_then(|v| v.as_str())
@@ -1432,6 +1590,11 @@ web_ffi!(web_listen, |args: &Value| {
rss_routes,
upload_routes,
download_routes,
redirects,
sitemap_routes,
robots_body,
page_404,
page_500,
middleware,
cookie_secure,
)
@@ -1460,6 +1623,11 @@ web_ffi!(web_listen, |args: &Value| {
rss_routes,
upload_routes,
download_routes,
redirects,
sitemap_routes,
robots_body,
page_404,
page_500,
&middleware,
)
});
@@ -1549,6 +1717,23 @@ pub unsafe extern "C" fn marqdo_plugin_init(host: *const MarqdoHostApi) -> c_int
"app,path,table,limit,order,title,link,description",
web_app_route_rss as PluginFn,
),
(
"web_app_redirect",
"app,from,to,permanent",
web_app_redirect as PluginFn,
),
(
"web_app_error_page",
"app,status,page",
web_app_error_page as PluginFn,
),
(
"web_app_sitemap",
"app,path,base,table,loc,limit,items",
web_app_sitemap as PluginFn,
),
("web_app_robots", "app,body,sitemap", web_app_robots as PluginFn),
("web_sitemap_build", "base,items", web_sitemap_build as PluginFn),
(
"web_app_upload",
"app,path,field,storage,prefix,max_bytes,types",
@@ -1651,7 +1836,7 @@ pub unsafe extern "C" fn marqdo_plugin_init(host: *const MarqdoHostApi) -> c_int
),
(
"web_app_middleware",
"app,cors,security,compress,body_limit,json_routes,access_log",
"app,cors,security,compress,body_limit,json_routes,access_log,cache_control",
web_app_middleware as PluginFn,
),
("web_form_new", "table,action,id", web_form_new as PluginFn),
+19
View File
@@ -43,6 +43,8 @@ pub struct Middleware {
pub body_limit: Option<u64>,
/// Log `METHOD path status duration_ms` to stderr for each request.
pub access_log: bool,
/// Global `Cache-Control` response header (e.g. `public, max-age=3600`).
pub cache_control: Option<String>,
/// JSON API routes: `path -> { method, table, where?, order?, limit? }`.
pub json_routes: Vec<(String, JsonRoute)>,
}
@@ -89,6 +91,11 @@ pub fn parse(app: &Value) -> Middleware {
.get("access_log")
.and_then(|v| v.as_bool())
.unwrap_or(false);
mw.cache_control = obj
.get("cache_control")
.and_then(|v| v.as_str())
.map(|s| s.trim().to_string())
.filter(|s| !s.is_empty());
mw.body_limit = obj
.get("body_limit")
.and_then(|v| v.as_u64().or_else(|| v.as_i64().map(|i| i as u64)))
@@ -226,6 +233,15 @@ pub fn apply(
app = app.layer(CompressionLayer::new());
}
if let Some(ref cc) = mw.cache_control {
if let Ok(v) = HeaderValue::from_str(cc) {
app = app.layer(SetResponseHeaderLayer::overriding(
header::CACHE_CONTROL,
v,
));
}
}
// Request body limit
if let Some(bytes) = mw.body_limit {
app = app.layer(RequestBodyLimitLayer::new(bytes as usize));
@@ -457,6 +473,9 @@ pub fn summary(app: &Value) -> String {
if mw.access_log {
parts.push("access_log".to_string());
}
if mw.cache_control.is_some() {
parts.push("cache_control".to_string());
}
if !mw.json_routes.is_empty() {
parts.push(format!("json:{}", mw.json_routes.len()));
}
+147
View File
@@ -0,0 +1,147 @@
//! sitemap.xml / robots.txt helpers (W7).
use serde_json::{json, Value};
fn cell_str(v: &Value) -> String {
match v {
Value::Null => String::new(),
Value::String(s) => s.clone(),
Value::Bool(b) => b.to_string(),
Value::Number(n) => n.to_string(),
other => other.to_string(),
}
}
fn xml_escape(s: &str) -> String {
s.replace('&', "&amp;")
.replace('<', "&lt;")
.replace('>', "&gt;")
.replace('"', "&quot;")
.replace('\'', "&apos;")
}
/// Build a sitemap.xml document from rows.
///
/// Each row may use keys: `loc`/`url`/`路径`, optional `lastmod`/`更新`,
/// `changefreq`/`频率`, `priority`/`优先级`.
pub fn build_sitemap(base: &str, items: &Value) -> String {
let base = base.trim_end_matches('/');
let rows: Vec<Value> = match items {
Value::Array(a) => a
.iter()
.map(|row| match row {
Value::String(s) => json!({ "loc": s }),
Value::Object(_) => row.clone(),
other => json!({ "loc": cell_str(other) }),
})
.collect(),
Value::Object(m) => {
// Columnar GFM: { loc: [...], ... }
let locs = m
.get("loc")
.or_else(|| m.get("路径"))
.or_else(|| m.get("path"))
.or_else(|| m.get("url"));
match locs {
Some(Value::Array(a)) => a
.iter()
.enumerate()
.map(|(i, loc)| {
let mut obj = serde_json::Map::new();
obj.insert("loc".into(), loc.clone());
for (k, v) in m {
if k == "loc" || k == "路径" || k == "path" || k == "url" {
continue;
}
if let Some(Value::Array(col)) = Some(v) {
if let Some(cell) = col.get(i) {
obj.insert(k.clone(), cell.clone());
}
}
}
Value::Object(obj)
})
.collect(),
_ => Vec::new(),
}
}
_ => Vec::new(),
};
let mut out = String::from(
"<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n\
<urlset xmlns=\"http://www.sitemaps.org/schemas/sitemap/0.9\">\n",
);
for row in &rows {
let Some(obj) = row.as_object() else {
continue;
};
let loc_raw = obj
.get("loc")
.or_else(|| obj.get("url"))
.or_else(|| obj.get("路径"))
.or_else(|| obj.get("path"))
.map(cell_str)
.unwrap_or_default();
if loc_raw.is_empty() {
continue;
}
let loc = if loc_raw.starts_with("http://") || loc_raw.starts_with("https://") {
loc_raw
} else if loc_raw.starts_with('/') {
format!("{base}{loc_raw}")
} else {
format!("{base}/{loc_raw}")
};
out.push_str(" <url>\n");
out.push_str(&format!(" <loc>{}</loc>\n", xml_escape(&loc)));
if let Some(lm) = obj
.get("lastmod")
.or_else(|| obj.get("更新"))
.map(cell_str)
.filter(|s| !s.is_empty())
{
out.push_str(&format!(" <lastmod>{}</lastmod>\n", xml_escape(&lm)));
}
if let Some(cf) = obj
.get("changefreq")
.or_else(|| obj.get("频率"))
.map(cell_str)
.filter(|s| !s.is_empty())
{
out.push_str(&format!(
" <changefreq>{}</changefreq>\n",
xml_escape(&cf)
));
}
if let Some(pr) = obj
.get("priority")
.or_else(|| obj.get("优先级"))
.map(cell_str)
.filter(|s| !s.is_empty())
{
out.push_str(&format!(" <priority>{}</priority>\n", xml_escape(&pr)));
}
out.push_str(" </url>\n");
}
out.push_str("</urlset>\n");
out
}
/// Default robots.txt body allowing all and pointing at a sitemap URL.
pub fn build_robots(sitemap_url: Option<&str>) -> String {
let mut out = String::from("User-agent: *\nAllow: /\n");
if let Some(u) = sitemap_url.filter(|s| !s.is_empty()) {
out.push_str(&format!("Sitemap: {u}\n"));
}
out
}
/// ABI helper returning `{ body }`.
#[allow(dead_code)]
pub fn robots_json(sitemap_url: Option<&str>, body: Option<&str>) -> Value {
let text = match body.filter(|s| !s.is_empty()) {
Some(b) => b.to_string(),
None => build_robots(sitemap_url),
};
json!({ "body": text })
}
+30 -3
View File
@@ -66,7 +66,10 @@ fn boolish(v: &Value) -> bool {
}
}
/// Schema table → `[{name,type,nullable}, …]`
const UNIQUE_KEYS: &[&str] = &["唯一", "unique", "uniq"];
const INDEX_KEYS: &[&str] = &["索引", "index", "idx"];
/// Schema table → `[{name,type,nullable,unique,index}, …]`
pub fn as_fields(table: &Value) -> Value {
match table {
Value::Array(rows) => {
@@ -81,7 +84,15 @@ pub fn as_fields(table: &Value) -> Value {
.map(cell_str)
.unwrap_or_else(|| "text".into());
let nullable = pick(m, NULL_KEYS).map(boolish).unwrap_or(true);
out.push(json!({ "name": name, "type": ty, "nullable": nullable }));
let unique = pick(m, UNIQUE_KEYS).map(boolish).unwrap_or(false);
let index = pick(m, INDEX_KEYS).map(boolish).unwrap_or(false);
out.push(json!({
"name": name,
"type": ty,
"nullable": nullable,
"unique": unique,
"index": index,
}));
}
}
Value::Array(out)
@@ -90,6 +101,8 @@ pub fn as_fields(table: &Value) -> Value {
let names = pick(m, FIELD_KEYS).map(as_str_list).unwrap_or_default();
let types = pick(m, TYPE_KEYS).map(as_str_list).unwrap_or_default();
let nulls = pick(m, NULL_KEYS).map(as_str_list).unwrap_or_default();
let uniques = pick(m, UNIQUE_KEYS).map(as_str_list).unwrap_or_default();
let indexes = pick(m, INDEX_KEYS).map(as_str_list).unwrap_or_default();
let mut out = Vec::new();
for (i, name) in names.iter().enumerate() {
if name.is_empty() {
@@ -100,7 +113,21 @@ pub fn as_fields(table: &Value) -> Value {
.get(i)
.map(|s| boolish(&Value::String(s.clone())))
.unwrap_or(true);
out.push(json!({ "name": name, "type": ty, "nullable": nullable }));
let unique = uniques
.get(i)
.map(|s| boolish(&Value::String(s.clone())))
.unwrap_or(false);
let index = indexes
.get(i)
.map(|s| boolish(&Value::String(s.clone())))
.unwrap_or(false);
out.push(json!({
"name": name,
"type": ty,
"nullable": nullable,
"unique": unique,
"index": index,
}));
}
Value::Array(out)
}
+82
View File
@@ -0,0 +1,82 @@
---
title: web W7 finish smoke — unique index, sitemap, cache-control (offline)
import web:ext/web/web.mq.md
import plugin:lib/plugin.mq.md
import sys:lib/sys.mq.md
---
# main
*p = > plugin.native_path name="web"*
1. `p`
> plugin.load path=`p`
2. *
> sys.exit code=1
*store = > web.db url="sqlite:web-fixtures/data/w7-smoke.db"*
> `store`.exec sql="DROP TABLE IF EXISTS posts"
`fields` =
| 字段 | 类型 | 可空 | 唯一 | 索引 |
|------|------|------|------|------|
| id | int | 否 | | |
| slug | text | 否 | 是 | 是 |
| title | text | 否 | | |
> `store`.init name=posts fields=`fields`
*idx = > `store`.query sql="SELECT name FROM sqlite_master WHERE type='index' AND tbl_name='posts'"*
*nidx = > len value=idx[^rows]*
1. `nidx` >= 1
> print text=unique-ok
2. *
> print text=unique-fail
`items` =
| loc | 更新 |
|-----|------|
| / | 2026-08-28 |
| /about | 2026-08-28 |
*sm = > web_sitemap_build base="https://example.com" items=`items`*
*xml = sm[^xml]*
1. `xml`
> print text=sitemap-ok
2. *
> print text=sitemap-fail
*pg = > web.page title="w7"*
*notfound = > web.page title="Missing"*
*app = > web.app page=`pg` port=18120*
*app = > `app`.configure cache_control="public, max-age=60" access_log=False*
*app = > `app`.redirect from="/old" to="/new" permanent=True*
*app = > `app`.error_page status=404 page=`notfound`*
*app = > `app`.sitemap path="/sitemap.xml" base="https://example.com" items=`items`*
*app = > `app`.robots sitemap="https://example.com/sitemap.xml"*
*rd = app[^redirects]*
*has = rd[^/old]*
1. `has`
> print text=redirect-ok
2. *
> print text=redirect-fail
*rb = app[^robots_body]*
1. `rb`
> print text=robots-ok
2. *
> print text=robots-fail
*mw = app[^middleware]*
*cc = mw[^cache_control]*
1. `cc` == "public, max-age=60"
> print text=cache-control-ok
2. *
> print text=cache-control-fail
*p404 = app[^page_404]*
1. `p404`
> print text=error-page-ok
2. *
> print text=error-page-fail
+31
View File
@@ -0,0 +1,31 @@
---
title: web W7 live server — 404, redirect, sitemap, robots, cache-control
import web:ext/web/web.mq.md
import plugin:lib/plugin.mq.md
import sys:lib/sys.mq.md
---
# main
*p = > plugin.native_path name="web"*
1. `p`
> plugin.load path=`p`
2. *
> sys.exit code=1
`items` =
| loc | note |
|-----|------|
| / | home |
| /about | about |
*pg = > web.page title="w7-live"*
*nf = > web.page title="Not Found Page"*
*app = > web.app page=`pg` port=18121*
*app = > `app`.configure cache_control="public, max-age=120"*
*app = > `app`.redirect from="/legacy" to="/" permanent=True*
*app = > `app`.error_page status=404 page=`nf`*
*app = > `app`.sitemap path="/sitemap.xml" base="http://127.0.0.1:18121" items=`items`*
*app = > `app`.robots sitemap="http://127.0.0.1:18121/sitemap.xml"*
> `app`.listen
+135
View File
@@ -1978,6 +1978,141 @@ access-log-ok",
);
}
#[test]
fn ext_web_w7_finish_smoke() {
let status = Command::new("cargo")
.args(["build", "-p", "marqdo_plugin_web"])
.current_dir(env!("CARGO_MANIFEST_DIR"))
.status()
.expect("build web plugin");
assert!(status.success(), "failed to build marqdo_plugin_web");
assert_out(
"tests/ext/web-w7-finish-smoke.mq.md",
"unique-ok
sitemap-ok
redirect-ok
robots-ok
cache-control-ok
error-page-ok",
);
}
#[test]
fn ext_web_w7_live() {
let status = Command::new("cargo")
.args(["build", "-p", "marqdo_plugin_web"])
.current_dir(env!("CARGO_MANIFEST_DIR"))
.status()
.expect("build web plugin");
assert!(status.success(), "failed to build marqdo_plugin_web");
let script = "tests/ext/web-w7-live-server.mq.md";
let bin = env!("CARGO_BIN_EXE_marqdo");
let mut child = Command::new(bin)
.args(["run", script])
.current_dir(env!("CARGO_MANIFEST_DIR"))
.stdout(Stdio::piped())
.stderr(Stdio::piped())
.spawn()
.expect("spawn w7 live server");
let base = "http://127.0.0.1:18121";
let ready = std::time::Instant::now();
loop {
if ready.elapsed().as_secs() > 10 {
let _ = child.kill();
panic!("w7 live server did not start in time");
}
let out = Command::new("curl")
.args([
"-s",
"-o",
"/dev/null",
"-w",
"%{http_code}",
&format!("{base}/"),
])
.output()
.expect("curl probe");
if String::from_utf8_lossy(&out.stdout) == "200" {
break;
}
std::thread::sleep(std::time::Duration::from_millis(300));
}
// Permanent redirect
let out = Command::new("curl")
.args([
"-s",
"-o",
"/dev/null",
"-w",
"%{http_code}|%{redirect_url}",
&format!("{base}/legacy"),
])
.output()
.expect("curl redirect");
let redir = String::from_utf8_lossy(&out.stdout);
assert!(
redir.starts_with("301|") && redir.contains("/"),
"redirect got {redir}"
);
// Custom 404
let out = Command::new("curl")
.args(["-s", "-w", "\n%{http_code}", &format!("{base}/no-such")])
.output()
.expect("curl 404");
let body = String::from_utf8_lossy(&out.stdout);
assert!(
body.contains("Not Found Page") && body.trim_end().ends_with("404"),
"404 body={body}"
);
// Sitemap
let out = Command::new("curl")
.args(["-s", &format!("{base}/sitemap.xml")])
.output()
.expect("curl sitemap");
let sm = String::from_utf8_lossy(&out.stdout);
assert!(
sm.contains("<urlset") && sm.contains("/about"),
"sitemap={sm}"
);
// robots.txt
let out = Command::new("curl")
.args(["-s", &format!("{base}/robots.txt")])
.output()
.expect("curl robots");
let rb = String::from_utf8_lossy(&out.stdout);
assert!(
rb.contains("User-agent:") && rb.contains("Sitemap:"),
"robots={rb}"
);
// Cache-Control
let out = Command::new("curl")
.args([
"-s",
"-D",
"-",
"-o",
"/dev/null",
&format!("{base}/"),
])
.output()
.expect("curl headers");
let headers = String::from_utf8_lossy(&out.stdout).to_ascii_lowercase();
assert!(
headers.contains("cache-control: public, max-age=120"),
"headers={headers}"
);
let _ = child.kill();
let _ = child.wait();
}
#[test]
fn ext_web_ws_broadcast_live() {
use futures_util::{SinkExt, StreamExt};