Help with splitting const char?

[corey]

Hi, I am having some trouble with getting this working. I am working on some base code that I have. And I want to split the buffer and replace some things. I have tried countless other things, none have really worked.

I do not want to use boost or any thing like that, I just need a short piece of code to help me. Please do not ask me to use boost, I do not want to.

Here is what I am trying to achieve:

I have used Detours 2.1 and some online resources to get a working hook for Send. I am capturing the HTTP headers and I want to find and edit the Accept-Encoding line.

From:

GET / HTTP/1.1
[removed to shorten the post]
Accept-Encoding: gzip,deflate

To:

GET / HTTP/1.1
[removed to shorten the post]
Accept-Encoding: identity

Here is all that I have so far in my hooked Send:

int WINAPI Mine_Send( SOCKET s, const char *buf, int len, int flags )
{
	//MessageBox(NULL,buf,NULL,NULL);
	if (strncmp(buf,"POST",lstrlen("POST"))==0)
	{
		//this was just a test
	}
		ofstream sendlog;
		sendlog.open ("C:\\Users\\[User]\\Desktop\\buf.txt",ios::app);
		sendlog <<"Socket: " << s <<endl << "Buffer: " << buf <<endl;
		sendlog.close();
    return Real_Send( s, buf, len, flags ); 
}

I would really appreciate it if someone could help me out. I am still learning and would like to take the time out to mention that a short snippet would be the best for me.

Thanks.

[hamsterman]

You mean that you get the "From" as the argument in buf and you want to pass the edited version to Real_Send ?
I suggest that you use strings. A combination of find(), substr() and operator + should be sufficient. see http://www.cplusplus.com/reference/string/string/ for how to use them. What code you need depends on exact requirements. If the input will always end with what you posted, it could be as simple as

std::string input( buf );
std::string output = input.substr(0, input.length()-12) + "identity";

Then pass output.c_str() and output.length() to Real_Send.

[corey]

No, the order of the header options will be different most of the time. Do you happen to know how to get each line (so many in between each \n)? I could just use that and then use strncmp to check if it starts with "Accept-Encoding:". If it does, I can just replace the line. Thanks.

[hamsterman]

If you use c functions, you'll need to deal with reallocation yourself..

Here's a bit of code to process every line:

bool newline = true;
for( const char * ptr = buf; *ptr != 0; ptr++ ){
   if(newline){
      //do whatever you like
      newline = false;
   }
   if( *ptr == '\n' ) newline = true;
}

Note that if since buf is const char*, you'll need to make a copy. If you don't know how much exactly the length will change, it's a bit tricky..

Will there be only one line like this?
If yes, try

std::string input = buf;
int pos = input.find("Accept-Encoding:");
int end_of_that_line = input.find('\n', pos);
std::string output = input.substr(0, pos) + "Accept-Encoding: identity" + input.substr(end_of_that_line);

Do test this before using it though, as I haven't.

Also, can it be that "Accept-Encoding:" would occur somewhere in the middle of a line and would not need to be replaced? If yes, check that pos == 0 || input[pos-1] == '\n'

[corey]

Thanks, I'll try this out when I get home. It looks perfect for my scenario.