tierra's C++ vs dataw0lf's Python speedhash challenge!

glibchash.c

// Kill me if I ever have to use glibc ever again, almost as bad as Perl.

#define _GNU_SOURCE

#include <stdio.h>
#include <unistd.h>
#include <string.h>
#include <sys/stat.h>
#include <sys/types.h>
#include <search.h>

int main (int argc, char* argv[])
{
	ENTRY e, *ep;
	FILE *hash_fd, *list_fd;
	// This will need to be increased to fit the list sadly.
	char* wordlist[100000];
	char* hashfile = argv[1];
	char* wordfile = argv[2];
	char* word = NULL;
	size_t wordsize = 0;
	int bytes_read = 0, i;
	unsigned int count = 0, match_count = 0;

	hash_fd = fopen(hashfile, "r");
	if(hash_fd == NULL) {
		fprintf(stderr, "Error opening hash file.\n");
		return 1;
	}

	while((bytes_read = getline(&word, &wordsize, hash_fd)) != -1) {
		wordlist[count] = strtok(word, "\n");
		count++;
		word = NULL;
		wordsize = 0;
	}
	close(fileno(hash_fd));

	hcreate(count);
	for(i = 0; i < count; i++) {
		e.key = wordlist[i];
		e.data = (char*)i;
		ep = hsearch(e, ENTER);
		if (ep == NULL) {
			fprintf(stderr, "Hash entry failed.\n");
			return 1;
		}
	}

	list_fd = fopen(wordfile, "r");
	if(list_fd == NULL) {
		fprintf(stderr, "Error opening word file.\n");
		return 1;
	}

	while((bytes_read = getline(&word, &wordsize, list_fd)) != -1) {
		e.key = strtok(word, "\n");
		if(hsearch(e, FIND))
			match_count++;			
		word = NULL;
		wordsize = 0;
	}

	close(fileno(list_fd));
	printf("Matching Lines: %d\n", match_count);
	return 0;
}

stlhash.cpp

#define _GNU_SOURCE

#include <stdio.h>
#include <unistd.h>
#include <sys/stat.h>
#include <sys/types.h>
#include <string>
#include <map>

using namespace std;

int main (int argc, char* argv[])
{
	map<string, bool> tehmap;
	FILE *hash_fd, *list_fd;
	char* hashfile = argv[1];
	char* wordfile = argv[2];
	char* word = NULL;
	string stlword;
	size_t wordsize = 0;
	int bytes_read = 0, i;
	unsigned int match_count = 0;

	printf("Initialized variables, opening list of words to add to hash table...\n");
	
	hash_fd = fopen(hashfile, "r");
	if(hash_fd == NULL) {
		fprintf(stderr, "Error opening hash file.\n");
		return 1;
	}

	printf("Adding words from file to hash table...\n");

	while((bytes_read = getline(&word, &wordsize, hash_fd)) != -1) {
		stlword = strtok(word, "\n");
		tehmap[stlword] = true;
		word = NULL;
		wordsize = 0;
	}
	close(fileno(hash_fd));

	printf("Opening list of words to check against hash table...\n");

	list_fd = fopen(wordfile, "r");
	if(list_fd == NULL) {
		fprintf(stderr, "Error opening word file.\n");
		return 1;
	}

	printf("Checking words against entries in the hash table...\n");
	
	while((bytes_read = getline(&word, &wordsize, list_fd)) != -1) {
		stlword = strtok(word, "\n");
		if(tehmap.find(stlword) != tehmap.end())
			match_count++;
		word = NULL;
		wordsize = 0;
	}

	close(fileno(list_fd));
	printf("Matching Lines: %d\n", match_count);
	return 0;
}