pack.c

php-4.4.7学习linux时下载的源代码

	}

	efree(argv);
	efree(formatcodes);
	efree(formatargs);
	output[outputpos] = '\0';
	RETVAL_STRINGL(output, outputpos, 1);
	efree(output);
}
/* }}} */

/* {{{ php_unpack
 */
static long php_unpack(char *data, int size, int issigned, int *map)
{
	long result;
	char *cresult = (char *) &result;
	int i;

	result = issigned ? -1 : 0;

	for (i = 0; i < size; i++) {
		cresult[map[i]] = *data++;
	}

	return result;
}
/* }}} */

/* unpack() is based on Perl's unpack(), but is modified a bit from there.
 * Rather than depending on error-prone ordered lists or syntactically
 * unpleasant pass-by-reference, we return an object with named paramters 
 * (like *_fetch_object()). Syntax is "f[repeat]name/...", where "f" is the
 * formatter char (like pack()), "[repeat]" is the optional repeater argument,
 * and "name" is the name of the variable to use.
 * Example: "c2chars/nints" will return an object with fields
 * chars1, chars2, and ints.
 * Numeric pack types will return numbers, a and A will return strings,
 * f and d will return doubles.
 * Implemented formats are A, a, h, H, c, C, s, S, i, I, l, L, n, N, f, d, x, X, @.
 */
/* {{{ proto array unpack(string format, string input)
   Unpack binary string into named array elements according to format argument */
PHP_FUNCTION(unpack)
{
	zval **formatarg;
	zval **inputarg;
	char *format;
	char *input;
	int formatlen;
	int inputpos, inputlen;
	int i;

	if (ZEND_NUM_ARGS() != 2 || 
        zend_get_parameters_ex(2, &formatarg, &inputarg) == FAILURE) {
		WRONG_PARAM_COUNT;
	}

	convert_to_string_ex(formatarg);
	convert_to_string_ex(inputarg);

	format = Z_STRVAL_PP(formatarg);
	formatlen = Z_STRLEN_PP(formatarg);
	input = Z_STRVAL_PP(inputarg);
	inputlen = Z_STRLEN_PP(inputarg);
	inputpos = 0;

	array_init(return_value);

	while (formatlen-- > 0) {
		char type = *(format++);
		char c;
		int arg = 1, argb;
		char *name;
		int namelen;
		int size=0;

		/* Handle format arguments if any */
		if (formatlen > 0) {
			c = *format;

			if (c >= '0' && c <= '9') {
				arg = atoi(format);

				while (formatlen > 0 && *format >= '0' && *format <= '9') {
					format++;
					formatlen--;
				}
			} else if (c == '*') {
				arg = -1;
				format++;
				formatlen--;
			}
		}

		/* Get of new value in array */
		name = format;
		argb = arg;

		while (formatlen > 0 && *format != '/') {
			formatlen--;
			format++;
		}

		namelen = format - name;

		if (namelen > 200)
			namelen = 200;

		switch ((int) type) {
			/* Never use any input */
			case 'X': 
				size = -1;
				break;

			case '@':
				size = 0;
				break;

			case 'a': 
			case 'A':
				size = arg;
				arg = 1;
				break;

			case 'h': 
			case 'H': 
				size = (arg > 0) ? (arg + (arg % 2)) / 2 : arg;
				arg = 1;
				break;

			/* Use 1 byte of input */
			case 'c': 
			case 'C':
			case 'x':
				size = 1;
				break;

			/* Use 2 bytes of input */
			case 's': 
			case 'S': 
			case 'n': 
			case 'v':
				size = 2;
				break;

			/* Use sizeof(int) bytes of input */
			case 'i': 
			case 'I':
				size = sizeof(int);
				break;

			/* Use 4 bytes of input */
			case 'l': 
			case 'L': 
			case 'N': 
			case 'V':
				size = 4;
				break;

			/* Use sizeof(float) bytes of input */
			case 'f':
				size = sizeof(float);
				break;

			/* Use sizeof(double) bytes of input */
			case 'd':
				size = sizeof(double);
				break;
		}

		/* Do actual unpacking */
		for (i = 0; i != arg; i++ ) {
			/* Space for name + number, safe as namelen is ensured <= 200 */
			char n[256];

			if (arg != 1 || namelen == 0) {
				/* Need to add element number to name */
				sprintf(n, "%.*s%d", namelen, name, i + 1);
			} else {
				/* Truncate name to next format code or end of string */
				sprintf(n, "%.*s", namelen, name);
			}

			if (size != 0 && size != -1 && INT_MAX - size + 1 < inputpos) {
				php_error_docref(NULL TSRMLS_CC, E_WARNING, "Type %c: integer overflow", type);
				inputpos = 0;
			}

			if ((inputpos + size) <= inputlen) {
				switch ((int) type) {
					case 'a': 
					case 'A': {
						char pad = (type == 'a') ? '\0' : ' ';
						int len = inputlen - inputpos;	/* Remaining string */

						/* If size was given take minimum of len and size */
						if ((size >= 0) && (len > size)) {
							len = size;
						}

						size = len;

						/* Remove padding chars from unpacked data */
						while (--len >= 0) {
							if (input[inputpos + len] != pad)
								break;
						}

						add_assoc_stringl(return_value, n, &input[inputpos], len + 1, 1);
						break;
					}
					
					case 'h': 
					case 'H': {
						int len = (inputlen - inputpos) * 2;	/* Remaining */
						int nibbleshift = (type == 'h') ? 0 : 4;
						int first = 1;
						char *buf;
						int ipos, opos;

						/* If size was given take minimum of len and size */
						if (size >= 0 && len > (size * 2)) {
							len = size * 2;
						} 

						if (argb > 0) {	
							len -= argb % 2;
						}

						buf = emalloc(len + 1);

						for (ipos = opos = 0; opos < len; opos++) {
							char c = (input[inputpos + ipos] >> nibbleshift) & 0xf;

							if (c < 10) {
								c += '0';
							} else {
								c += 'a' - 10;
							}

							buf[opos] = c;
							nibbleshift = (nibbleshift + 4) & 7;

							if (first-- == 0) {
								ipos++;
								first = 1;
							}
						}

						buf[len] = '\0';
						add_assoc_stringl(return_value, n, buf, len, 1);
						efree(buf);
						break;
					}

					case 'c': 
					case 'C': {
						int issigned = (type == 'c') ? (input[inputpos] & 0x80) : 0;
						long v = php_unpack(&input[inputpos], 1, issigned, byte_map);
						add_assoc_long(return_value, n, v);
						break;
					}

					case 's': 
					case 'S': 
					case 'n': 
					case 'v': {
						long v;
						int issigned = 0;
						int *map = machine_endian_short_map;

						if (type == 's') {
							issigned = input[inputpos + (machine_little_endian ? 1 : 0)] & 0x80;
						} else if (type == 'n') {
							map = big_endian_short_map;
						} else if (type == 'v') {
							map = little_endian_short_map;
						}

						v = php_unpack(&input[inputpos], 2, issigned, map);
						add_assoc_long(return_value, n, v);
						break;
					}

					case 'i': 
					case 'I': {
						long v;
						int issigned = 0;

						if (type == 'i') {
							issigned = input[inputpos + (machine_little_endian ? (sizeof(int) - 1) : 0)] & 0x80;
						}

						v = php_unpack(&input[inputpos], sizeof(int), issigned, int_map);
						add_assoc_long(return_value, n, v);
						break;
					}

					case 'l': 
					case 'L': 
					case 'N': 
					case 'V': {
						int issigned = 0;
						int *map = machine_endian_long_map;
						long v;

						if (type == 'l') {
							issigned = input[inputpos + (machine_little_endian ? 3 : 0)] & 0x80;
						} else if (type == 'N') {
							map = big_endian_long_map;
						} else if (type == 'V') {
							map = little_endian_long_map;
						}

						v = php_unpack(&input[inputpos], 4, issigned, map);
						add_assoc_long(return_value, n, v);
						break;
					}

					case 'f': {
						float v;

						memcpy(&v, &input[inputpos], sizeof(float));
						add_assoc_double(return_value, n, (double)v);
						break;
					}

					case 'd': {
						double v;

						memcpy(&v, &input[inputpos], sizeof(double));
						add_assoc_double(return_value, n, v);
						break;
					}

					case 'x':
						/* Do nothing with input, just skip it */
						break;

					case 'X':
						if (inputpos < size) {
							inputpos = -size;
							i = arg - 1;		/* Break out of for loop */

							if (arg >= 0) {
								php_error_docref(NULL TSRMLS_CC, E_WARNING, "Type %c: outside of string", type);
							}
						}
						break;

					case '@':
						if (arg <= inputlen) {
							inputpos = arg;
						} else {
							php_error_docref(NULL TSRMLS_CC, E_WARNING, "Type %c: outside of string", type);
						}

						i = arg - 1;	/* Done, break out of for loop */
						break;
				}

				inputpos += size;
				if (inputpos < 0) {
					if (size != -1) { /* only print warning if not working with * */
						php_error_docref(NULL TSRMLS_CC, E_WARNING, "Type %c: outside of string", type);
					}
					inputpos = 0;
				}
			} else if (arg < 0) {
				/* Reached end of input for '*' repeater */
				break;
			} else {
				php_error_docref(NULL TSRMLS_CC, E_WARNING, "Type %c: not enough input, need %d, have %d", type, size, inputlen - inputpos);
				zval_dtor(return_value);
				RETURN_FALSE;
			}
		}

		formatlen--;	/* Skip '/' separator, does no harm if inputlen == 0 */
		format++;
	}
}
/* }}} */

/* {{{ PHP_MINIT_FUNCTION
 */
PHP_MINIT_FUNCTION(pack)
{
	int machine_endian_check = 1;
	int i;

	machine_little_endian = ((char *)&machine_endian_check)[0];

	if (machine_little_endian) {
		/* Where to get lo to hi bytes from */
		byte_map[0] = 0;

		for (i = 0; i < sizeof(int); i++) {
			int_map[i] = i;
		}

		machine_endian_short_map[0] = 0;
		machine_endian_short_map[1] = 1;
		big_endian_short_map[0] = 1;
		big_endian_short_map[1] = 0;
		little_endian_short_map[0] = 0;
		little_endian_short_map[1] = 1;

		machine_endian_long_map[0] = 0;
		machine_endian_long_map[1] = 1;
		machine_endian_long_map[2] = 2;
		machine_endian_long_map[3] = 3;
		big_endian_long_map[0] = 3;
		big_endian_long_map[1] = 2;
		big_endian_long_map[2] = 1;
		big_endian_long_map[3] = 0;
		little_endian_long_map[0] = 0;
		little_endian_long_map[1] = 1;
		little_endian_long_map[2] = 2;
		little_endian_long_map[3] = 3;
	}
	else {
		zval val;
		int size = sizeof(Z_LVAL(val));
		Z_LVAL(val)=0; /*silence a warning*/

		/* Where to get hi to lo bytes from */
		byte_map[0] = size - 1;

		for (i = 0; i < sizeof(int); i++) {
			int_map[i] = size - (sizeof(int) - i);
		}

		machine_endian_short_map[0] = size - 2;
		machine_endian_short_map[1] = size - 1;
		big_endian_short_map[0] = size - 2;
		big_endian_short_map[1] = size - 1;
		little_endian_short_map[0] = size - 1;
		little_endian_short_map[1] = size - 2;

		machine_endian_long_map[0] = size - 4;
		machine_endian_long_map[1] = size - 3;
		machine_endian_long_map[2] = size - 2;
		machine_endian_long_map[3] = size - 1;
		big_endian_long_map[0] = size - 4;
		big_endian_long_map[1] = size - 3;
		big_endian_long_map[2] = size - 2;
		big_endian_long_map[3] = size - 1;
		little_endian_long_map[0] = size - 1;
		little_endian_long_map[1] = size - 2;
		little_endian_long_map[2] = size - 3;
		little_endian_long_map[3] = size - 4;
	}

	return SUCCESS;
}
/* }}} */

/*
 * Local variables:
 * tab-width: 4
 * c-basic-offset: 4
 * End:
 * vim600: noet sw=4 ts=4 fdm=marker
 * vim<600: noet sw=4 ts=4
 */