The JML Continuum: Writing a C/C++ App run from apache

A few weeks back I started wondering on ways to improve the performance of certain areas of my site (as I do periodically) and while I came across examples of PHP compilers in their many forms, I started wondering if I should just go whole hog and write some parts of the application right in C/C++. Sure, everyone is jumping on nodejs, but I've worked in both C and C++ languages before and while it has been sometime, I thought it might be an interesting exercise to try writing some of the back end processes in either one. After all, I'm running mysql as my db, and there are mysql c headers and api, so it should be fairly straight forward. I found quickly that some of this documentation was lacking. Sure I can create a C/C++ executable that'll give me info from my DB on demand, but getting it back to the requester was being a pain. Digging around and experimenting led me to finding a few things that I'd like to share here.

int MIMEHeader() { 
     cout << "Content-type: text/html" << endl << endl << endl; 
     cout << "<!DOCTYPE html PUBLIC \"-//W3C//DTD XHTML 1.0 Strict//EN\" \"http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd\">"<<endl; 
     cout <<"<HTML xlmns='http://www.w3.org/1999/xhtml' xml:lang='en' lang='en'>" <<endl;
}

int main() {
     MIMEHeader();
     cout<<">head<>/head><body>"<<endl;        
     cout<<"Hello World!"<<endl;    
     cout<<"</body>"<<endl;
     return TRUE;
}

First off, yes, you really do need three "endl"'s on the first line. One finishes the current line and the next to tell apache you really mean that this is html and it should serve it. The next two sets up the doctype and get things ready for you to output your info. Calling this function from our main function causes things to be ready for output. In the example here, we tack on the proverbial "Hello World" example.
The next step here is to get apache to proxy the output of our program to the end user's web browser. You'll need this block in your sections affected.

<Directory "/path/your/website/x">
     AllowOverride None
     Options ExecCGI
     Order allow,deny
     Allow from all
</Directory>
ScriptAlias /x /path/to/your/website/x

In the example, I used /x as my subdir of compiled code. So, resolving it would be "http://example.com/x/programName".

At this point you're good to go, and boy is it FAST. In my non-exhaustive testing, a simple App involving a MySQL query takes less than 1ms more than the actual MySQL query. Obviously, your results may vary. Here's the example C++ I wrote for a start before I started adding more fancy api type features to it. I compiled it on CentOS 5.4 with stock CentOS distributed MySQL 5.1:

// Compile this using:
// g++ -o mysqltest mysqltest.cpp -L/usr/lib64/mysql -lmysqlclient -I/usr/include/mysql && ./mysqltest
//

#include <iostream>
#include <mysql/mysql.h>

using namespace std;
MYSQL *connection, mysql;
MYSQL_RES *result;
MYSQL_ROW row;
int query_state;

int MIMEHeader() {
 cout << "Content-type: text/html"<<endl<<endl<<endl<<"<HTML><BODY><PRE>";
}

int Footer() { 
 cout << "</PRE></BODY></HTML>";
}

int main() {
 char * env;
 MIMEHeader();
 mysql_init(&mysql);
 connection = mysql_real_connect(&mysql,"localhost","myUs3r","myPa55word","myBookmarks",0,0,0);
 if (connection == NULL) { 
  cout << mysql_error(&mysql) << endl;
  return 1;
 } 
 query_state=mysql_query(connection,"select * from links where id like '%blogger.com';");
 if (query_state!=0) {
  cout << mysql_error(&mysql) << endl;
                return 1;
 }
 result = mysql_store_result(connection);
 cout << "<table>"<<endl;
 while  (( row = mysql_fetch_row(result)) != NULL ) {
  cout << "<tr><td>" << row[0] <<"</td><td>"<< row[1] <<"</td><td>"<< row[2] <<"</td></tr>"<<endl;
 }
 cout << "</table>"<<endl;
 Footer();
 mysql_free_result(result);
 mysql_close(connection);
 return 0;
}

2013-11-22

Writing a C/C++ App run from apache

No comments:

Post a Comment