#include <sys/mman.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <unistd.h>
#include <libxml/HTMLparser.h>
#include <string.h>
int main (int argc, char ** argv) {
	xmlInitParser();
	htmlDocPtr xmldoc;
	char * txtdoc;
	struct stat s;
	int fd = open(argv[1], O_RDONLY);
	htmlParserCtxtPtr c;
	stat(argv[1], &s);
	txtdoc = mmap(NULL, s.st_size, PROT_READ, MAP_PRIVATE, fd, 0);
	xmlInitParser();

	c = htmlNewParserCtxt();
	xmldoc = htmlCtxtReadMemory(c, txtdoc, strlen(txtdoc), "", NULL, HTML_PARSE_RECOVER);
	/* by the way: why/how/when does libxml2 use networking when HTML_PARSE_NOT is not specified? */
	htmlFreeParserCtxt(c);
	xmlFreeDoc(xmldoc);
	
	close(fd);
	munmap(txtdoc, s.st_size);
	xmlCleanupParser();
	return 0;
}