request_parser.c 6.91 KB
#include <stdlib.h>
#include <string.h>
#include <stdlib.h>
#include <stdio.h>
#include <unistd.h>
#include <ctype.h>
#include <sys/types.h>

#include "class.h"
#include "http/request_parser.h"
#include "interface/class.h"
#include "interface/stream_reader.h"
#include "http/request.h"
#include "http/request_queue.h"

static
void
httpRequestParserParse(HttpRequestParser);

static
void
ctor(void * _this, va_list * params)
{
	HttpRequestParser this = _this;

	this->request_queue = new(HttpRequestQueue);

	this->buffer = malloc(HTTP_REQUEST_PARSER_READ_CHUNK);
	this->buffer[0] = 0;
}

static
void
dtor(void * _this)
{
	HttpRequestParser this = _this;

	free(this->buffer);
	delete(&(this->request_queue));
} 

static
void
_clone(void * _this, void * _base)
{
	HttpRequestParser this = _this;
	HttpRequestParser base = _base;
	size_t            chunks;

	/**
	 * every parser has its own queue...
	 */
	this->request_queue = new(HttpRequestQueue);
	this->buffer_used   = base->buffer_used;

	chunks = this->buffer_used / HTTP_REQUEST_PARSER_READ_CHUNK;
	chunks++;

	this->buffer = malloc(chunks * HTTP_REQUEST_PARSER_READ_CHUNK);
	memcpy(this->buffer, base->buffer, this->buffer_used);
}

static
size_t
get_data(void * _this, int fd)
{
	HttpRequestParser this = _this;
	size_t            remaining, chunks;
	char              buffer[1024];

	size_t size = read(fd, buffer, 1024);

	if (0 < size) {
		remaining = this->buffer_used % HTTP_REQUEST_PARSER_READ_CHUNK;
		chunks    = this->buffer_used / HTTP_REQUEST_PARSER_READ_CHUNK;

		/**
		 * because a division always rounds down
		 * chunks holds exactly the currently allocated chunks if
		 * remaining equals 0 but there is no space left.
		 * Else chunks holds the actually allocated amount of chunks
		 * minus 1.
		 * For this reason chunks always has to be increased by 1.
		 */
		chunks++;

		if (size >= remaining) {
			this->buffer =
				realloc(this->buffer, chunks * HTTP_REQUEST_PARSER_READ_CHUNK);
		}

		memcpy(this->buffer + this->buffer_used, buffer, size);
		this->buffer_used += size;
		this->buffer[this->buffer_used] = 0;

		httpRequestParserParse(this);
	}

	return size;
}

INIT_IFACE(Class, ctor, dtor, _clone);
INIT_IFACE(StreamReader, get_data);
CREATE_CLASS(HttpRequestParser, NULL, IFACE(Class), IFACE(StreamReader));

static
inline
char *
httpRequestLineGet(char ** data)
{
	char * line_end = strstr(*data, "\r\n");
	char * ret      = *data;

	if (NULL == line_end) {
		return NULL;
	}

	*line_end = 0;
	*data = line_end + 2;

	return ret;
}

static
inline
void
httpRequestSkip(char ** data)
{
	for (; 0 != **data && ! isalpha(**data); (*data)++);
}

/**
 * SDBM hashing algorithm:
 *
 * this algorithm was created for sdbm (a public-domain reimplementation of
 * ndbm) database library. it was found to do well in scrambling bits,
 * causing better distribution of the keys and fewer splits. it also happens
 * to be a good general hashing function with good distribution. the actual
 * function is hash(i) = hash(i - 1) * 65599 + str[i]; what is included below
 * is the faster version used in gawk. [there is even a faster, duff-device
 * version] the magic constant 65599 was picked out of thin air while
 * experimenting with different constants, and turns out to be a prime. this
 * is one of the algorithms used in berkeley db (see sleepycat) and elsewhere. 
 */
static
inline
unsigned long
sdbm(unsigned char * str)
{
	unsigned long hash = 0;
	int c;

	while ((c = *str++))
		hash = c + (hash << 6) + (hash << 16) - hash;

	return hash;
}

static
inline
int
comp (const void * _a, const void * _b)
{
	const struct HttpRequestHeader * a = _a;
	const struct HttpRequestHeader * b = _b;
	return (a->hash < b->hash)? -1 : (a->hash > b->hash)? 1 : 0;
}

static
void
httpRequestParserParse(HttpRequestParser this)
{
	static HttpRequest request  = NULL;
	static char *      data; // static pointer to unprocessed data
	char *             line;
	int                cont = 1;
	static int         header_idx;

	while(cont) {
		switch(this->state) {
			case HTTP_REQUEST_GARBAGE:
				data = this->buffer; // initialize static pointer
				httpRequestSkip(&data);
				request = new(HttpRequest);

				this->state = HTTP_REQUEST_START;
				break;

			case HTTP_REQUEST_START:
				if (NULL == (line = httpRequestLineGet(&data))) {
					cont = 0;
					break;
				}
				
				{
					char * delim = strchr(line, ' ');

					if (NULL != delim) {
						*delim = 0;
						request->method = malloc(strlen(line) + 1);
						strcpy(request->method, line);
						line = delim + 1;

						for (; *line == ' ' && *line != 0; line++);

						if (0 != *line) {
							delim = strchr(line, ' ');

							if (NULL != delim) {
								*delim = 0;
								request->uri = malloc(strlen(line) + 1);
								strcpy(request->uri, line);
								line = delim + 1;

								for (; *line == ' ' && *line != 0; line++);

								if (0 != *line) {
									request->http_version = malloc(strlen(line) + 1);
									strcpy(request->http_version, line);
								}
							}
						}
					}
				}

				header_idx = 0;
				this->state = HTTP_REQUEST_REQUEST_LINE_DONE;
				break;

			case HTTP_REQUEST_REQUEST_LINE_DONE:
				if (NULL == (line = httpRequestLineGet(&data))) {
					cont = 0;
					break;
				}

				if (0 == strlen(line)) {
					this->state = HTTP_REQUEST_HEADERS_DONE;
					break;
				}

				{
					char * delim = strchr(line, ':');

					*delim = 0;
					(request->header)[header_idx].name = malloc(strlen(line) + 1);
					strcpy((request->header)[header_idx].name, line);
					(request->header)[header_idx].hash = sdbm((unsigned char *)line);

					line = delim + 1;
					for (; *line == ' ' && *line != 0; line++);

					(request->header)[header_idx].value = malloc(strlen(line) + 1);
					strcpy((request->header)[header_idx].value, line);

					header_idx++;
					request->nheader++;
				}

				break;

			case HTTP_REQUEST_HEADERS_DONE:
				/**
				 * @TODO: here comes the body handling
				 */
				qsort(
						request->header,
						request->nheader,
						sizeof(struct HttpRequestHeader),
						comp);

				{
					char * bodylen;

					bodylen = httpRequestHeaderGet(request, "Content-Length");

					if (NULL != bodylen) {
						request->nbody = atoi(bodylen);
						request->body = calloc(1, request->nbody + 1);
						memcpy(request->body, data, request->nbody);
						data += request->nbody;
					}
				}
					
				this->state = HTTP_REQUEST_DONE;
				break;

			case HTTP_REQUEST_DONE:
				/**
				 * enqueue current request
				 */
				this->request_queue->requests[(this->request_queue->nrequests)++] =
					request;

				/**
				 * remove processed stuff from input buffer.
				 */
				memmove(this->buffer,
						data,
						this->buffer_used - (data - this->buffer) + 1);

				this->buffer_used -= data - this->buffer;

				/**
				 * dont continue loop if input buffer is empty
				 */
				if (0 == this->buffer_used) {
					cont = 0;
				}

				/**
				 * prepare for next request
				 */
				this->state = HTTP_REQUEST_GARBAGE;

				break;

			default:
				break;
		}
	}
}

// vim: set ts=4 sw=4: