为什么卷曲的帖子请求与此普通图书馆邮政请求返回与同一机构的同一地址不同的结果？

Question

我正在使用钢库普通LISP（SBCL），emacs，slime和一个名为 dexador ）的库。

在补充中，我可以做：

CL-USER> (dex:post "https://httpbin.org/post"
          :content "{\"user\": \"user1\", \"pass\":\"abcd\"}")
"{
  \"args\": {}, 
  \"data\": \"{\\\"user\\\": \\\"user1\\\", \\\"pass\\\": \\\"abcd\\\"}\", 
  \"files\": {}, 
  \"form\": {}, 
  \"headers\": {
    \"Accept\": \"*/*\", 
    \"Content-Length\": \"33\", 
    \"Content-Type\": \"text/plain\", 
    \"Host\": \"httpbin.org\", 
    \"User-Agent\": \"Dexador/0.9.15 (SBCL 2.1.9.nixos); Linux; 5.10.94\", 
    \"X-Amzn-Trace-Id\": \"Root=1-62956dc4-1b95d37752a67b8420180f71\"
  }, 
  \"json\": {
    \"pass\": \"abcd\", 
    \"user\": \"user1\"
  }, 
  \"origin\": \"189.2.84.243\", 
  \"url\": \"https://httpbin.org/post\"
}
"

删除逃生字符，请注意JSON输出的JSON键：

  "json": {
    "pass": "abcd", 
    "user": "user1"
  }

如果我尝试在curl上执行此操作，则遵循此教程，这就是我得到的：

$ curl -d "user=user1&pass=abcd" -X POST https://httpbin.org/post

{
  "args": {}, 
  "data": "", 
  "files": {}, 
  "form": {
    "pass": "abcd", 
    "user": "user1"
  }, 
  "headers": {
    "Accept": "*/*", 
    "Content-Length": "20", 
    "Content-Type": "application/x-www-form-urlencoded", 
    "Host": "httpbin.org", 
    "User-Agent": "curl/7.79.1", 
    "X-Amzn-Trace-Id": "Root=1-62956e50-4e880d101ac23b781fe0f9d5"
  }, 
  "json": null, 
  "origin": "189.2.84.243", 
  "url": "https://httpbin.org/post"
}

注意“ json”：null，。

我期望这两个结果都是相同的，因为它们指向相同的地址并具有相同的主体。

为什么它们与众不同？我想念什么吗？

Answer 1

设置NetCat：

nc -l -p 3500

运行您的curl命令：

curl -d "user=user1&pass=abcd" -X POST http://localhost:3500/post

NetCat打印：

POST /post HTTP/1.1
Host: localhost:3500
User-Agent: curl/7.68.0
Accept: */*
Content-Length: 20
Content-Type: application/x-www-form-urlencoded

user=user1&pass=abcd

因此，首先，content-type is application> application> application/x-www -form-urlencoded，因为那是-d的意思是curl。

其次，数据以user = user1＆amp; pass = abcd而发送，而不是JSON。这是因为curl的默认数据模式是application/x-www-form-urlencoded，它不仅将其发送为JSON，因为您希望它会。
实际上，它根本没有实现JSON。

如果要发送JSON，则必须明确执行此操作：

jq --null-input '
    .user |= "user1" | 
    .pass |= "abcd"' \
    | curl \
        -H 'Content-Type: text/plain' \
        -d @- \
        -X POST \
        https://httpbin.org/post

响应：

{
  "args": {}, 
  "data": "{  \"user\": \"user1\",  \"pass\": \"abcd\"}", 
  "files": {}, 
  "form": {}, 
  "headers": {
    "Accept": "*/*", 
    "Content-Length": "36", 
    "Content-Type": "text/plain", 
    "Host": "httpbin.org", 
    "User-Agent": "curl/7.68.0", 
    "X-Amzn-Trace-Id": "Root=1-6295a78d-0d828c3c6d3891174fd22d01"
  }, 
  "json": {
    "pass": "abcd", 
    "user": "user1"
  }, 
  "origin": "71.122.242.250", 
  "url": "https://httpbin.org/post"
}

Answer 2

如果一个人打开 main github存储库，他将能够在用法 part：

(dex:post "https://example.com/login"
          :content '(("name" . "fukamachi") ("password" . "1ispa1ien")))

表单内容定义为a-list或关联列表。关联列表是通用LISP中非常重要的一件事，如果该函数需要一个参数中的A列表，则不应提供字符串：

(dex:post "https://httpbin.org/post" :content "{\"user\": \"user1\", \"pass\": \"abcd\"}")
                                              ^ this is a string not a a-list

正确的函数调用是：

(dex:post "https://httpbin.org/post"
          :content '(("user" . "user1") ("pass" . "abcd")))

返回：

(dex:post "https://httpbin.org/post"
          :content '(("user" . "user1") ("pass" . "abcd")))
#<(SIMPLE-ARRAY CHARACTER (464)) {
  "args": {},
  "data": "",
  "files": {},
  "form": {         ;;
    "pass": "abcd", ;; The data is here
    "user": "user1" ;;
  },                ;;
  "headers": {
    "Content-Length": "20",
    "Content-Type": "application/x-www-form-urlen... {10076E5CCF}>
200