Capture Incoming HTTP Requests to the Web Server

Introduction

Few days ago, I had a problem in my CGI application where I must capture the data inside the HTTP POST request coming to the application. The post data are XML data that I must parse and process. In this tip/trick, I will share the idea and the solution I found.

Background

There are few things at least you must know, types of HTTP requests and the HTTP request format, so I suggest you check this detailed article on Wikipedia on the HTTP protocol. I also suggest you review some CGI concepts especially the environment variables.

Using the Code

The idea behind the solution is the CGI environment variables, especially the CONTENT_LENGTH variable. This variable indicates the number of bytes (size) of the data passed within a POST HTTP request.

Any C/C++ application has standard input and output, so when you run the application as CGI, the standard input/output will be through server. So to read the post data, you read the value of CONTENT_LENGTH and then read from stdin. Here is how it is done:

char* len_ = getenv("CONTENT_LENGTH");
if(len_)
{
    long int len = strtol(len_, NULL, 10);
    cout << "LENGTH = " << len << endl;
    char* postdata = (char*)malloc(len + 1);
    if (!postdata) { exit(EXIT_FAILURE); }
    fgets(postdata, len + 1, stdin);
    cout << "DATA = " << postdata << endl;
    free(postdata);
}
else
{
    // No CONTENT_LENGTH, therefore no post data available.
}

Points of Interest

Always check for the value of CONTENT_LENGTH, if no data is available within the request, it becomes null. And always allocate extra location for the null terminator.