A blog on Computer Science, Security, Programming, and more...

Programmer Competency Matrix

Written by Matt

Here's a fun page I came across for gauging the skill level of a programmer: Programmer Competency Matrix

Feels good to be a strong Level 3 in systems programming, algorithms and data structures. Where do you stand?


Redirect STDOUT of Child to Parent Process in C

Written by Matt

Here's some C code for how to do this on Linux... mostly posting it for my own reference. I've commented the code extensively. (compile with -std=c99 or -std=gnu99 if using GCC)

#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>
#include <sys/types.h>
#include <sys/wait.h>

#define BUFSIZE 256

void child_process(int pipe_pair[2]){

    // closes the child's STDOUT descriptor and replaces it with
    // the write pipe linked to the parent's read pipe
    if (dup2(pipe_pair[1], STDOUT_FILENO) == -1){ 
      // if dup2 fails, perror writes to stderr so this reports
      // appropriately either way as the child and parent share stderr
      // parent will know if the child failed since we return 1
      // (ideally return errno so that calling processes know the error code)

    // duplicated by dup2 above, no longer needed
    // close read end as we will never read from stdout

    // printf writes to stdout by default
    // we could also use fprintf(stdout, ...)
    printf("Hello, parent!\n"); 
    // make sure the write buffer is flushed before we exit

    // close to make sure read() returns 0 in the parent

    // child exits

void parent_process(int pipe_pair[2], pid_t cpid){

    // cstatus will store the return of the child process
    // buf will hold the child's writes to stdout --
    // {0} initializes the array elements to 0x00
    int cstatus; 
    char buf[BUFSIZE] = {0}; 

    close(pipe_pair[1]); // we won't write to stdout

    // read until closed, or error (0 or -1, respectively)
    for (int n = 0; (n = read(pipe_pair[0], buf, BUFSIZE)) > 0;){ 
      printf("Received %d bytes from child process: ", n); 
      // (needed otherwise write() may output before 
      // printf since stdio output to stdout is line buffered)
      // writes just what we read so no need to reset buf
      write(STDOUT_FILENO, buf, n); 

    // close read pipe

    // waits for child process with pid 'cpid' to
    // return and stores the exit code in cstatus
    waitpid(cpid, &cstatus, 0); 

    printf("Child exit status was: %d\n", cstatus);

    // terminate parent


int main(int argc, char **argv){

  // cpid stores the process id of the child process
  // stdout_pipe array = pipe descriptor pair -- 
  // [0] is the read end, [1] is the write end
  pid_t cpid; 
  int stdout_pipe[2]; 

  // call that creates the two unidirectional pipe streams 
  // and stores the descriptors in the array
  if (pipe(stdout_pipe) == -1){ 

  // fork happens here, cpid will have the child's
  // process id or -1 if the call fails
  cpid = fork(); 
  if (cpid == -1){

  // child (fork returns 0 in the child and the child ID for the parent)
  if (cpid == 0) 
  // else when cpid is not 0 or -1 we're in the parent
    parent_process(stdout_pipe, cpid);

  // we shouldn't get here, but return int from main for correctness
  return 0;

This is a cheap and easy way to do interprocess communication, but also useful if you have code that outputs to stdout / takes from stdin and you want to integrate it into your program without rewriting all the output code to use a different write mechanism, or you just don't want to maintain two separate codebases. Simply fork the process and replace STDIN/STDOUT/STDERR in the child before executing the code you need to execute.

Topic: Programming tags: linux, snippet, c, code, system

Getting Raw HTTP Request in PHP

Written by Matt

In case you ever want to get a raw HTTP request for logging, analysis or filtering reasons, the code is simply:

function http_raw_request(){
  $http_request = ""; 

  foreach (getallheaders() as $key => $value){
    $http_request .= $key . ': ' . $value . "\n";

  $http_request .= "\n";
  $http_request .= file_get_contents("php://input");

  return $http_request;


If, like me, you use php-fastcgi and lack the function getallheaders, it can be implemented as followslink:

if (!function_exists('getallheaders')){ 
  function getallheaders(){ 
    $headers = ''; 
    foreach ($_SERVER as $name => $value){ 
      if (substr($name, 0, 5) == 'HTTP_'){ 
        $headers[str_replace(' ', '-', ucwords(strtolower(str_replace('_', ' ', substr($name, 5)))))] = $value; 
    return $headers; 

Example capture on a comment for this blog:

 Host: heapspray.net                                                               +
 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:15.0) Gecko/20100101 Firefox/15.0.1+
 Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8           +
 Accept-Language: en-us,en;q=0.5                                                   +
 Accept-Encoding: gzip, deflate                                                    +
 Connection: keep-alive                                                            +
 Referer: http://heapspray.net/post/added-comments/                                +
 Content-Type: application/x-www-form-urlencoded                                   +
 Content-Length: 38                                                                +

Very useful for anything from site statistics to debugging and filtering spam (most spambots have very different request headers than most browsers).