how to split binary/post-data/html from a http response?

Question

I'm writing a simple function that split binary/post-data/html from a HTTP response. The HTTP headers are terminated by \r\n\r\n the rest is the message.

I have wrote this:

 #define MAX_BUFFER_SIZE 256
    //... 
    int size = 0;
    int buf_size = MAX_BUFFER_SIZE;
    char * headers = malloc(MAX_BUFFER_SIZE);
    char * newbuf;

    while(httpresponse[size]) {
        if(httpresponse[size]     == '\r' &&
           httpresponse[size + 1] == '\n' &&
           httpresponse[size + 2] == '\r' &&
           httpresponse[size + 3] == '\n') {
            break;
        }

        headers[size] = httpresponse[size];     

        if(size >= buf_size) {
            buf_size += MAX_BUFFER_SIZE;
            newbuf = realloc(headers, buf_size);

           if(NULL ==  newbuf) exit(1);
           headers = newbuf;

         }

           size ++;
        }

        printf("%s\n", headers);

the httpresponse variable, has value-like:

HTTP/1.1 200 OK
Date: Fri, 23 Mar 2012 15:28:17 GMT
Expires: Sat, 23 Mar 2013 15:28:17 GMT
Cache-Control: public, max-age=31536000
Last-Modified: Thu, 14 Apr 2011 15:46:35 GMT
Content-Type: image/jpeg
Content-Length: 12745
X-XSS-Protection: 1; mode=block
Connection: close

���������I1��} �g������'�B�f�p���ohd]sft�����J�������1����瘿ٱ����$3�G�8��4=�E�i����ܼG����H��nbi�"�1��b[Ǘl��++���OPt�W��>�����i�]t�QT�N/,Q�Qz������0�`    N7���M��f��S�Š�x9k��$*

//more binary...

but the above C program, print the following text:

HTTP/1.1 200 OK
Date: Fri, 23 Mar 2012 17:12:09 GMT
Expires: Sat, 23 Mar 2013 17:12:09 GMT
Last-Modified: Thu, 14 Apr 2011 15:46:35 GMT
Content-Type: image/jpeg
Content-Length: 12745
X-XSS-Protection: 1; mode=block
Cache-Control: public, max-age=31536000
Age: 3746
�2�/���ms���|ނ����LQr2K3�v��J.�,�z��^Oy����s(ct���X`iA����I����U�{

instead of:

    HTTP/1.1 200 OK
    Date: Fri, 23 Mar 2012 15:28:17 GMT
    Expires: Sat, 23 Mar 2013 15:28:17 GMT
    Cache-Control: public, max-age=31536000
    Last-Modified: Thu, 14 Apr 2011 15:46:35 GMT
    Content-Type: image/jpeg
    Content-Length: 12745
    X-XSS-Protection: 1; mode=block
    Connection: close

how to fix this? Thanks in advance.

Andrey · Accepted Answer · 2014-08-18 07:29:31Z

1

I use following code:

const QByteArray& data = socket->readAll();
int index = data.indexOf("\r\n\r\n");
QString sHeader;
QString sBody;
if (index < 0)
    sHeader = QString::fromUtf8(data);
else
{
    sHeader = QString::fromUtf8(data.left(index));
    sBody = QString::fromUtf8(data.mid(index + 4));
}

QIssHttpRequestHeader requestHeader(sHeader); // QIssHttpRequestHeader is a copy of QHttpRequestHeader from Qt4

if (requestHeader.method() != "GET")
{
    send501Error(socket);
    return;
}

QUrl url(requestHeader.path());
...

answered Aug 18, 2014 at 7:29

Andrey

99011 silver badges13 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

Murray McDonald · Accepted Answer · 2012-03-23 18:41:54Z

0

I think your code is OK except that malloc doens't initialize the memory it allocates. This means when you BREAK out of your while loop the "headers" string you've captured doesn't have a null byte terminating it.

Two ways to fix - ether use "calloc" which initializes the memory to binary zeroes or insert one line of code before your call to printf:

 headers[size] = '\0'

Cheers

answered Mar 23, 2012 at 18:41

Murray McDonald

6314 silver badges5 bronze badges

3 Comments

Murray McDonald Over a year ago

Hmm -- and httpresponse is a "single-byte wide" string -- how about using the "strstr" function to look for the <cr><lf><cr><lf> instead of using the while loop?

Jack Over a year ago

I have tried it. but if I split the \r\n\r\n with strstr() it returns the binary only. And I want split headers and body(message), for this reason I have used while loop.

Murray McDonald Over a year ago

ahh -- well if you want both pieces why not use a capturing regular expression? /^(.+?)\r\n\r\n(.+)$/s

rene · Accepted Answer · 2012-03-23 19:27:14Z

0

Are you sure about the outputs? you said httpresponse, ends with X-XSS-Protection: 1; mode=block and Connection: close in this case, but your ouptput ends like this: X-XSS-Protection: 1; mode=block andCache-Control: public, max-age=31536000 ??!!?

if you are sure about copying the outputs right, well it seems that the end of your while is incorrect.

answered Mar 23, 2012 at 19:27

rene

1661 gold badge2 silver badges11 bronze badges

Collectives™ on Stack Overflow

how to split binary/post-data/html from a http response?

3 Answers 3

Comments

3 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

Comments

3 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related