nostr/damus
1
0
Fork 0
mirror of git://jb55.com/damus synced 2024-09-18 19:23:49 +00:00
Code Issues Packages Projects Releases Wiki Activity
7ecb9aad62
damus/nostrdb/flatcc/portable/pbase64.h
William Casarin 1f5f1e28a4 nostrdb: pull latest, adding flatcc and lmdb
2023-08-25 19:05:34 -07:00

449 lines
16 KiB
C
Raw Blame History

#ifndef PBASE64_H
#define PBASE64_H
#ifdef __cplusplus
extern "C" {
#endif
#include <stdlib.h>
/* Guarded to allow inclusion of pstdint.h first, if stdint.h is not supported. */
#ifndef UINT8_MAX
#include <stdint.h>
#endif
#define BASE64_EOK 0
/* 0 or mure full blocks decoded, remaining content may be parsed with fresh buffer. */
#define BASE64_EMORE 1
/* The `src_len` argument is required when encoding. */
#define BASE64_EARGS 2
/* Unsupported mode, or modifier not supported by mode when encoding. */
#define BASE64_EMODE 3
/* Decoding ends at invalid tail length - either by source length or by non-alphabet symbol. */
#define BASE64_ETAIL 4
/* Decoding ends at valid tail length but last byte has non-zero bits where it shouldn't have. */
#define BASE64_EDIRTY 5
static inline const char *base64_strerror(int err);
/* All codecs are URL safe. Only Crockford allow for non-canocical decoding. */
enum {
/* Most common base64 codec, but not url friendly. */
base64_mode_rfc4648 = 0,
/* URL safe version, '+' -> '-', '/' -> '_'. */
base64_mode_url = 1,
/*
* Skip ' ', '\r', and '\n' - we do not allow tab because common
* uses of base64 such as PEM do not allow tab.
*/
base64_dec_modifier_skipspace = 32,
/* Padding is excluded by default. Not allowed for zbase64. */
base64_enc_modifier_padding = 128,
/* For internal use or to decide codec of mode. */
base64_modifier_mask = 32 + 64 + 128,
};
/* Encoded size with or without padding. */
static inline size_t base64_encoded_size(size_t len, int mode);
/*
* Decoded size assuming no padding.
* If `len` does include padding, the actual size may be less
* when decoding, but never more.
*/
static inline size_t base64_decoded_size(size_t len);
/*
* `dst` must hold ceil(len * 4 / 3) bytes.
* `src_len` points to length of source and is updated with length of
* parse on both success and failure. If `dst_len` is not null
* it is used to store resulting output lengt withh length of decoded
* output on both success and failure.
* If `hyphen` is non-zero a hyphen is encoded every `hyphen` output bytes.
* `mode` selects encoding alphabet defaulting to Crockfords base64.
* Returns 0 on success.
*
* A terminal space can be added with `dst[dst_len++] = ' '` after the
* encode call. All non-alphabet can be used as terminators except the
* padding character '='. The following characters will work as
* terminator for all modes: { '\0', '\n', ' ', '\t' }. A terminator is
* optional when the source length is given to the decoder. Note that
* crockford also reserves a few extra characters for checksum but the
* checksum must be separate from the main buffer and is not supported
* by this library.
*/
static inline int base64_encode(uint8_t *dst, const uint8_t *src, size_t *dst_len, size_t *src_len, int mode);
/*
* Decodes according to mode while ignoring encoding modifiers.
* `src_len` and `dst_len` are optional pointers. If `src_len` is set it
* must contain the length of the input, otherwise the input must be
* terminated with a non-alphabet character or valid padding (a single
* padding character is accepted) - if the src_len output is needed but
* not the input due to guaranteed termination, then set it to
* (size_t)-1. `dst_len` must contain length of output buffer if present
* and parse will fail with BASE64_EMORE after decoding a block multiple
* if dst_len is exhausted - the parse can thus be resumed after
* draining destination. `src_len` and `dst_len` are updated with parsed
* and decoded length, when present, on both success and failure.
* Returns 0 on success. Invalid characters are not considered errors -
* they simply terminate the parse, however, if the termination is not
* at a block multiple or a valid partial block length then BASE64_ETAIL
* without output holding the last full block, if any. BASE64_ETAIL is also
* returned if the a valid length holds non-zero unused tail bits.
*/
static inline int base64_decode(uint8_t *dst, const uint8_t *src, size_t *dst_len, size_t *src_len, int mode);
static inline const char *base64_strerror(int err)
{
switch (err) {
case BASE64_EOK: return "ok";
case BASE64_EARGS: return "invalid argument";
case BASE64_EMODE: return "invalid mode";
case BASE64_EMORE: return "destination full";
case BASE64_ETAIL: return "invalid tail length";
case BASE64_EDIRTY: return "invalid tail content";
default: return "unknown error";
}
}
static inline size_t base64_encoded_size(size_t len, int mode)
{
size_t k = len % 3;
size_t n = (len * 4 / 3 + 3) & ~(size_t)3;
int pad = mode & base64_enc_modifier_padding;
if (!pad) {
switch (k) {
case 2:
n -= 1;
break;
case 1:
n -= 2;
break;
default:
break;
}
}
return n;
}
static inline size_t base64_decoded_size(size_t len)
{
size_t k = len % 4;
size_t n = len / 4 * 3;
switch (k) {
case 3:
return n + 2;
case 2:
return n + 1;
case 1: /* Not valid without padding. */
case 0:
default:
return n;
}
}
static inline int base64_encode(uint8_t *dst, const uint8_t *src, size_t *dst_len, size_t *src_len, int mode)
{
const uint8_t *rfc4648_alphabet = (const uint8_t *)
"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
const uint8_t *url_alphabet = (const uint8_t *)
"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-_";
const uint8_t *T;
uint8_t *dst_base = dst;
int pad = mode & base64_enc_modifier_padding;
size_t len = 0;
int ret = BASE64_EMODE;
if (!src_len) {
ret = BASE64_EARGS;
goto done;
}
len = *src_len;
mode = mode & ~base64_modifier_mask;
switch (mode) {
case base64_mode_rfc4648:
T = rfc4648_alphabet;
break;
case base64_mode_url:
T = url_alphabet;
break;
default:
/* Invalid mode. */
goto done;
}
ret = BASE64_EOK;
/* Encodes 4 destination bytes from 3 source bytes. */
while (len >= 3) {
dst[0] = T[((src[0] >> 2))];
dst[1] = T[((src[0] << 4) & 0x30) | (src[1] >> 4)];
dst[2] = T[((src[1] << 2) & 0x3c) | (src[2] >> 6)];
dst[3] = T[((src[2] & 0x3f))];
len -= 3;
dst += 4;
src += 3;
}
/* Encodes 8 destination bytes from 1 to 4 source bytes, if any. */
switch(len) {
case 2:
dst[0] = T[((src[0] >> 2))];
dst[1] = T[((src[0] << 4) & 0x30) | (src[1] >> 4)];
dst[2] = T[((src[1] << 2) & 0x3c)];
dst += 3;
if (pad) {
*dst++ = '=';
}
break;
case 1:
dst[0] = T[((src[0] >> 2))];
dst[1] = T[((src[0] << 4) & 0x30)];
dst += 2;
if (pad) {
*dst++ = '=';
*dst++ = '=';
}
break;
default:
pad = 0;
break;
}
len = 0;
done:
if (dst_len) {
*dst_len = (size_t)(dst - dst_base);
}
if (src_len) {
*src_len -= len;
}
return ret;
}
static inline int base64_decode(uint8_t *dst, const uint8_t *src, size_t *dst_len, size_t *src_len, int mode)
{
static const uint8_t cinvalid = 64;
static const uint8_t cignore = 65;
static const uint8_t cpadding = 66;
/*
* 0..63: 6-bit encoded value.
* 64: flags non-alphabet symbols.
* 65: codes for ignored symbols.
* 66: codes for pad symbol '='.
* All codecs consider padding an optional terminator and if present
* consumes as many pad bytes as possible up to block termination,
* but does not fail if a block is not full.
*
* We do not currently have any ignored characters but we might
* add spaces as per MIME spec, but assuming spaces only happen
* at block boundaries this is probalby better handled by repeated
* parsing.
*/
static const uint8_t base64rfc4648_decode[256] = {
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 62, 64, 64, 64, 63,
52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 64, 64, 64, 66, 64, 64,
64, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14,
15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 64, 64, 64, 64, 64,
64, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40,
41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64
};
static const uint8_t base64url_decode[256] = {
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 62, 64, 64,
52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 64, 64, 64, 66, 64, 64,
64, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14,
15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 64, 64, 64, 64, 63,
64, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40,
41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64
};
static const uint8_t base64rfc4648_decode_skipspace[256] = {
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 65, 64, 64, 65, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
65, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 62, 64, 64, 64, 63,
52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 64, 64, 64, 66, 64, 64,
64, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14,
15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 64, 64, 64, 64, 64,
64, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40,
41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64
};
static const uint8_t base64url_decode_skipspace[256] = {
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 65, 64, 64, 65, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
65, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 62, 64, 64,
52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 64, 64, 64, 66, 64, 64,
64, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14,
15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 64, 64, 64, 64, 63,
64, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40,
41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64
};
int ret = BASE64_EOK;
size_t i, k;
uint8_t hold[4];
uint8_t *dst_base = dst;
size_t limit = (size_t)-1;
size_t len = (size_t)-1, mark;
const uint8_t *T = base64rfc4648_decode;
int skipspace = mode & base64_dec_modifier_skipspace;
if (src_len) {
len = *src_len;
}
mark = len;
mode = mode & ~base64_modifier_mask;
switch (mode) {
case base64_mode_rfc4648:
T = skipspace ? base64rfc4648_decode_skipspace : base64rfc4648_decode;
break;
case base64_mode_url:
T = skipspace ? base64url_decode_skipspace : base64url_decode;
break;
default:
ret = BASE64_EMODE;
goto done;
}
if (dst_len && *dst_len > 0) {
limit = *dst_len;
}
while(limit > 0) {
for (i = 0; i < 4; ++i) {
if (len == i) {
k = i;
len -= i;
goto tail;
}
if ((hold[i] = T[src[i]]) >= cinvalid) {
if (hold[i] == cignore) {
++src;
--len;
--i;
continue;
}
k = i;
/* Strip padding and ignore hyphen in padding, if present. */
if (hold[i] == cpadding) {
++i;
while (i < len && i < 8) {
if (T[src[i]] != cpadding && T[src[i]] != cignore) {
break;
}
++i;
}
}
len -= i;
goto tail;
}
}
if (limit < 3) {
goto more;
}
dst[0] = (uint8_t)((hold[0] << 2) | (hold[1] >> 4));
dst[1] = (uint8_t)((hold[1] << 4) | (hold[2] >> 2));
dst[2] = (uint8_t)((hold[2] << 6) | (hold[3]));
dst += 3;
src += 4;
limit -= 3;
len -= 4;
mark = len;
}
done:
if (dst_len) {
*dst_len = (size_t)(dst - dst_base);
}
if (src_len) {
*src_len -= mark;
}
return ret;
tail:
switch (k) {
case 0:
break;
case 2:
if ((hold[1] << 4) & 0xff) {
goto dirty;
}
if (limit < 1) {
goto more;
}
dst[0] = (uint8_t)((hold[0] << 2) | (hold[1] >> 4));
dst += 1;
break;
case 3:
if ((hold[2] << 6) & 0xff) {
goto dirty;
}
if (limit < 2) {
goto more;
}
dst[0] = (uint8_t)((hold[0] << 2) | (hold[1] >> 4));
dst[1] = (uint8_t)((hold[1] << 4) | (hold[2] >> 2));
dst += 2;
break;
default:
ret = BASE64_ETAIL;
goto done;
}
mark = len;
goto done;
dirty:
ret = BASE64_EDIRTY;
goto done;
more:
ret = BASE64_EMORE;
goto done;
}
#ifdef __cplusplus
}
#endif
#endif /* PBASE64_H */
Reference in New Issue View Git Blame Copy Permalink