cannam@226: /*
cannam@226:   Copyright 2011-2016 David Robillard <http://drobilla.net>
cannam@226: 
cannam@226:   Permission to use, copy, modify, and/or distribute this software for any
cannam@226:   purpose with or without fee is hereby granted, provided that the above
cannam@226:   copyright notice and this permission notice appear in all copies.
cannam@226: 
cannam@226:   THIS SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
cannam@226:   WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
cannam@226:   MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
cannam@226:   ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
cannam@226:   WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
cannam@226:   ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
cannam@226:   OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
cannam@226: */
cannam@226: 
cannam@226: #include "serd_internal.h"
cannam@226: 
cannam@226: #include <math.h>
cannam@226: 
cannam@226: SERD_API
cannam@226: const uint8_t*
cannam@226: serd_strerror(SerdStatus st)
cannam@226: {
cannam@226: 	switch (st) {
cannam@226: 	case SERD_SUCCESS:        return (const uint8_t*)"Success";
cannam@226: 	case SERD_FAILURE:        return (const uint8_t*)"Non-fatal failure";
cannam@226: 	case SERD_ERR_UNKNOWN:    return (const uint8_t*)"Unknown error";
cannam@226: 	case SERD_ERR_BAD_SYNTAX: return (const uint8_t*)"Invalid syntax";
cannam@226: 	case SERD_ERR_BAD_ARG:    return (const uint8_t*)"Invalid argument";
cannam@226: 	case SERD_ERR_NOT_FOUND:  return (const uint8_t*)"Not found";
cannam@226: 	case SERD_ERR_ID_CLASH:   return (const uint8_t*)"Blank node ID clash";
cannam@226: 	case SERD_ERR_BAD_CURIE:  return (const uint8_t*)"Invalid CURIE";
cannam@226: 	case SERD_ERR_INTERNAL:   return (const uint8_t*)"Internal error";
cannam@226: 	}
cannam@226: 	return (const uint8_t*)"Unknown error";  // never reached
cannam@226: }
cannam@226: 
cannam@226: SERD_API
cannam@226: size_t
cannam@226: serd_strlen(const uint8_t* str, size_t* n_bytes, SerdNodeFlags* flags)
cannam@226: {
cannam@226: 	size_t        n_chars = 0;
cannam@226: 	size_t        i       = 0;
cannam@226: 	SerdNodeFlags f       = 0;
cannam@226: 	for (; str[i]; ++i) {
cannam@226: 		if ((str[i] & 0xC0) != 0x80) {
cannam@226: 			// Does not start with `10', start of a new character
cannam@226: 			++n_chars;
cannam@226: 			switch (str[i]) {
cannam@226: 			case '\r': case '\n':
cannam@226: 				f |= SERD_HAS_NEWLINE;
cannam@226: 				break;
cannam@226: 			case '"':
cannam@226: 				f |= SERD_HAS_QUOTE;
cannam@226: 			}
cannam@226: 		}
cannam@226: 	}
cannam@226: 	if (n_bytes) {
cannam@226: 		*n_bytes = i;
cannam@226: 	}
cannam@226: 	if (flags) {
cannam@226: 		*flags = f;
cannam@226: 	}
cannam@226: 	return n_chars;
cannam@226: }
cannam@226: 
cannam@226: static inline double
cannam@226: read_sign(const char** sptr)
cannam@226: {
cannam@226: 	double sign = 1.0;
cannam@226: 	switch (**sptr) {
cannam@226: 	case '-': sign = -1.0;
cannam@226: 	case '+': ++(*sptr);
cannam@226: 	default:  return sign;
cannam@226: 	}
cannam@226: }
cannam@226: 
cannam@226: SERD_API
cannam@226: double
cannam@226: serd_strtod(const char* str, char** endptr)
cannam@226: {
cannam@226: 	double result = 0.0;
cannam@226: 
cannam@226: 	// Point s at the first non-whitespace character
cannam@226: 	const char* s = str;
cannam@226: 	while (is_space(*s)) { ++s; }
cannam@226: 
cannam@226: 	// Read leading sign if necessary
cannam@226: 	const double sign = read_sign(&s);
cannam@226: 
cannam@226: 	// Parse integer part
cannam@226: 	for (; is_digit(*s); ++s) {
cannam@226: 		result = (result * 10.0) + (*s - '0');
cannam@226: 	}
cannam@226: 
cannam@226: 	// Parse fractional part
cannam@226: 	if (*s == '.') {
cannam@226: 		double denom = 10.0;
cannam@226: 		for (++s; is_digit(*s); ++s) {
cannam@226: 			result += (*s - '0') / denom;
cannam@226: 			denom *= 10.0;
cannam@226: 		}
cannam@226: 	}
cannam@226: 
cannam@226: 	// Parse exponent
cannam@226: 	if (*s == 'e' || *s == 'E') {
cannam@226: 		++s;
cannam@226: 		double expt      = 0.0;
cannam@226: 		double expt_sign = read_sign(&s);
cannam@226: 		for (; is_digit(*s); ++s) {
cannam@226: 			expt = (expt * 10.0) + (*s - '0');
cannam@226: 		}
cannam@226: 		result *= pow(10, expt * expt_sign);
cannam@226: 	}
cannam@226: 
cannam@226: 	if (endptr) {
cannam@226: 		*endptr = (char*)s;
cannam@226: 	}
cannam@226: 
cannam@226: 	return result * sign;
cannam@226: }
cannam@226: 
cannam@226: /**
cannam@226:    Base64 decoding table.
cannam@226:    This is indexed by encoded characters and returns the numeric value used
cannam@226:    for decoding, shifted up by 47 to be in the range of printable ASCII.
cannam@226:    A '$' is a placeholder for characters not in the base64 alphabet.
cannam@226: */
cannam@226: static const char b64_unmap[] =
cannam@226: 	"$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$m$$$ncdefghijkl$$$$$$"
cannam@226: 	"$/0123456789:;<=>?@ABCDEFGH$$$$$$IJKLMNOPQRSTUVWXYZ[\\]^_`ab$$$$"
cannam@226: 	"$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$"
cannam@226: 	"$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$";
cannam@226: 
cannam@226: static inline uint8_t unmap(const uint8_t in) { return b64_unmap[in] - 47; }
cannam@226: 
cannam@226: /**
cannam@226:    Decode 4 base64 characters to 3 raw bytes.
cannam@226: */
cannam@226: static inline size_t
cannam@226: decode_chunk(const uint8_t in[4], uint8_t out[3])
cannam@226: {
cannam@226: 	out[0] = (uint8_t)(((unmap(in[0]) << 2))        | unmap(in[1]) >> 4);
cannam@226: 	out[1] = (uint8_t)(((unmap(in[1]) << 4) & 0xF0) | unmap(in[2]) >> 2);
cannam@226: 	out[2] = (uint8_t)(((unmap(in[2]) << 6) & 0xC0) | unmap(in[3]));
cannam@226: 	return 1 + (in[2] != '=') + ((in[2] != '=') && (in[3] != '='));
cannam@226: }
cannam@226: 
cannam@226: SERD_API
cannam@226: void*
cannam@226: serd_base64_decode(const uint8_t* str, size_t len, size_t* size)
cannam@226: {
cannam@226: 	void* buf = malloc((len * 3) / 4 + 2);
cannam@226: 	*size = 0;
cannam@226: 	for (size_t i = 0, j = 0; i < len; j += 3) {
cannam@226: 		uint8_t in[] = "====";
cannam@226: 		size_t  n_in = 0;
cannam@226: 		for (; i < len && n_in < 4; ++n_in) {
cannam@226: 			for (; i < len && !is_base64(str[i]); ++i) {}  // Skip junk
cannam@226: 			in[n_in] = str[i++];
cannam@226: 		}
cannam@226: 		if (n_in > 1) {
cannam@226: 			*size += decode_chunk(in, (uint8_t*)buf + j);
cannam@226: 		}
cannam@226: 	}
cannam@226: 	return buf;
cannam@226: }