base64 impl now escapes dots. does not really respect buflen

This commit is contained in:
Erik Ekman 2007-06-10 18:55:18 +00:00
parent 82ac77da7e
commit ba8f2cfac1
2 changed files with 204 additions and 4 deletions

View File

@ -19,6 +19,7 @@
#include <string.h> #include <string.h>
#include "encoding.h" #include "encoding.h"
#include "common.h"
#include "base64.h" #include "base64.h"
static const char cb64[] = static const char cb64[] =
@ -26,6 +27,11 @@ static const char cb64[] =
static unsigned char rev64[128]; static unsigned char rev64[128];
static int reverse_init = 0; static int reverse_init = 0;
#define REV64(x) rev64[(int) (x)]
#define MODE (cb64[62])
#define P62 (cb64[62])
#define P63 (cb64[63])
static struct encoder base64_encoder = static struct encoder base64_encoder =
{ {
"BASE64", "BASE64",
@ -44,7 +50,136 @@ struct encoder
int int
base64_handles_dots() base64_handles_dots()
{ {
return 1; return 0;
}
static void
findesc(int *count, unsigned char *esc, char c1, char c2, char c3, char c4)
{
int min1 = 0;
int min2 = 0;
int num1 = 0xFF;
int num2 = 0xFE;
int i;
/* check if no more escapes needed */
if (count[62] == 0 && count[63] == 0) {
esc[0] = P62;
esc[1] = P62;
return;
}
for (i = 0; i < 62; i++) {
if (i == c1 || i == c2 || i == c3 || i == c4) {
continue;
}
if (count[i] < num1) {
min2 = min1;
num2 = num1;
min1 = i;
num1 = count[i];
} else if (count[i] < num2) {
min2 = i;
num2 = count[i];
}
}
esc[0] = cb64[min1];
esc[1] = cb64[min2];
}
static void
escape_chars(char *buf, size_t buflen)
{
int counter[64];
int escapes;
int reset;
int i;
unsigned char temp[4096];
unsigned char *r;
unsigned char *w;
unsigned char *e;
unsigned char esc[2];
memset(counter, 0, sizeof(counter));
esc[0] = P62;
esc[1] = P63;
/* first, find the number of times each token is used */
r = (unsigned char *) buf;
w = temp;
while (*r) {
counter[REV64(*r)]++;
*w++ = *r++;
}
/* check if work needed */
if (counter[62] == 0 && counter[63] == 0)
return;
r = temp;
w = (unsigned char *) buf;
reset = 1;
escapes = 0;
/* check a block for esc chars */
while (*r) {
if (reset == 0 && escapes == 0 && (
r[0] == esc[0] || r[1] == esc[0] ||r[2] == esc[0] ||r[2] == esc[0] ||
r[0] == esc[1] || r[1] == esc[1] ||r[2] == esc[1] ||r[2] == esc[1])) {
/* last set of escape chars were unused.
* if we reset last escape switch then we dont have to switch now */
/* change last ecape switch to 999 (RESET) */
e[1] = MODE;
e[2] = MODE;
/* store default esc chars */
esc[0] = P62;
esc[1] = P63;
reset = 1;
}
if (r[0] == esc[0] || r[1] == esc[0] ||r[2] == esc[0] ||r[2] == esc[0] ||
r[0] == esc[1] || r[1] == esc[1] ||r[2] == esc[1] ||r[2] == esc[1]) {
/* switch escape chars */
escapes = 0;
reset = 0;
/* find 2 suitable escape chars */
findesc(counter, esc, REV64(r[0]), REV64(r[1]), REV64(r[2]), REV64(r[3]));
/* store escape switch position */
e = w;
/* write new escape chars */
*w++ = MODE;
*w++ = esc[0];
*w++ = esc[1];
}
for (i = 0; i < 4; i++) {
if (r[i])
counter[REV64(r[i])]--;
}
for (i = 0; i < 4; i++) {
if (r[i] == P62) {
r[i] = esc[0];
escapes++;
} else if (r[i] == P63) {
r[i] = esc[1];
escapes++;
}
}
*w++ = *r++;
*w++ = *r++;
*w++ = *r++;
*w++ = *r++;
}
} }
int int
@ -52,6 +187,7 @@ base64_encode(char *buf, size_t *buflen, const void *data, size_t size)
{ {
size_t newsize; size_t newsize;
size_t maxsize; size_t maxsize;
unsigned char c;
unsigned char *s; unsigned char *s;
unsigned char *p; unsigned char *p;
unsigned char *q; unsigned char *q;
@ -59,6 +195,14 @@ base64_encode(char *buf, size_t *buflen, const void *data, size_t size)
memset(buf, 0, *buflen); memset(buf, 0, *buflen);
if (!reverse_init) {
for (i = 0; i < 64; i++) {
c = cb64[i];
rev64[(int) c] = i;
}
reverse_init = 1;
}
/* how many chars can we encode within the buf */ /* how many chars can we encode within the buf */
maxsize = 3 * (*buflen / 4 - 1) - 1; maxsize = 3 * (*buflen / 4 - 1) - 1;
/* how big will the encoded data be */ /* how big will the encoded data be */
@ -82,6 +226,8 @@ base64_encode(char *buf, size_t *buflen, const void *data, size_t size)
} }
*p = 0; *p = 0;
escape_chars(buf, *buflen);
/* store number of bytes from data that was used */ /* store number of bytes from data that was used */
*buflen = size; *buflen = size;
@ -89,7 +235,6 @@ base64_encode(char *buf, size_t *buflen, const void *data, size_t size)
} }
#define DECODE_ERROR 0xffffffff #define DECODE_ERROR 0xffffffff
#define REV64(x) rev64[(int) (x)]
static int static int
decode_token(const unsigned char *t, unsigned char *data, size_t len) decode_token(const unsigned char *t, unsigned char *data, size_t len)
@ -123,6 +268,9 @@ base64_decode(void *buf, size_t *buflen, const char *str, size_t slen)
size_t maxsize; size_t maxsize;
const char *p; const char *p;
unsigned char c; unsigned char c;
unsigned char block[4];
unsigned char prot62;
unsigned char prot63;
int len; int len;
int i; int i;
@ -143,9 +291,33 @@ base64_decode(void *buf, size_t *buflen, const char *str, size_t slen)
slen = maxsize; slen = maxsize;
} }
prot62 = P62;
prot63 = P63;
q = buf; q = buf;
for (p = str; *p && strchr(cb64, *p); p += 4) { for (p = str; *p; p += 4) {
len = decode_token((unsigned char *) p, (unsigned char *) q, slen); if (*p == MODE) {
p++;
if (p[0] == MODE && p[1] == MODE) {
/* reset escape chars */
prot62 = P62;
prot63 = P63;
p += 2;
} else {
prot62 = *p++;
prot63 = *p++;
}
}
for (i = 0; i < 4; i++) {
block[i] = p[i];
if (prot62 == block[i]) {
block[i] = P62;
} else if (prot63 == block[i]) {
block[i] = P63;
}
}
len = decode_token(block, (unsigned char *) q, slen);
q += len; q += len;
slen -= 4; slen -= 4;

View File

@ -31,6 +31,34 @@ static struct tuple
} testpairs[] = { } testpairs[] = {
{ "iodinetestingtesting", "Aw8KAw4LDgvZDgLUz2rLC2rPBMC" }, { "iodinetestingtesting", "Aw8KAw4LDgvZDgLUz2rLC2rPBMC" },
{ "abc123", "ywjJmtiZ" }, { "abc123", "ywjJmtiZ" },
{
"\xFF\xEF\x7C\xEF\xAE\x78\xDF\x6D\x74\xCF\x2C\x70\xBE\xEB\x6C\xAE\xAA\x68"
"\x9E\x69\x64\x8E\x28\x60\x7D\xE7\x5C\x6D\xA6\x58\x5D\x65\x54\x4D\x24\x50"
"\x3C\xE3\x4C\x2C\xA2\x48\x1C\x61\x44\x0C\x20\x40\x3F\x3F\x3C\xEF\xAE\x78"
"\xDF\x6D\x74\xCF\x2C\x70\xBE\xEB\x6C\xAE\xAA\x68\x9E\x69\x64\x8E\x28\x60"
"\x7D\xE7\x5C\x6D\xA6\x58\x5D\x65\x54\x4D\x24\x50\x3C\xE3\x4C\x2C\xA2\x48"
"\x1C\x61\x44\x0C\x20\x40\xFF\xEF\x7C\xEF\xAE\x78\xDF\x6D\x74\xCF\x2C\x70"
"\xBE\xEB\x6C\xAE\xAA\x68\x9E\x69\x64\x8E\x28\x60\x7D\xE7\x5C\x6D\xA6\x58"
"\x5D\x65\x54\x4D\x24\x50\x3C\xE3\x4C\x2C\xA2\x48\x1C\x61\x44\x0C\x20\x40",
"9abba876543210-ZYXWVUTSRQPONMLKJIHGFEDCBAzyxwvutsrqponmlkjihgfe999dcbapZ"
"776543210-ZYXWVUTSRQPONMLKJIHGFEDCBAzyxwvutsrqponmlkjihgfedcba9abba87654"
"3210-ZYXWVUTSRQPONMLKJIHGFEDCBAzyxwvutsrqponmlkjihgfe999dcba"
},
{
"\xFF\xEF\x7C\xEF\xAE\x78\xDF\x6D\x74\xCF\x2C\x70\xBE\xEB\x6C\xAE\xAA\x68"
"\x9E\x69\x64\x8E\x28\x60\x7D\xE7\x5C\x6D\xA6\x58\x5D\x65\x54\x4D\x24\x50"
"\x3C\xE3\x4C\x2C\xA2\x48\x1C\x61\x44\x0C\x20\x40\x3F\x3F\x3C\xEF\xAE\x78"
"\xDF\x6D\x74\xCF\x2C\x70\xBE\xEB\x6C\xAE\xA1\x61\x91\x61\x61\x81\x28\x60"
"\x7D\xE7\x5C\x6D\xA6\x58\x5D\x65\x54\x4D\x24\x50\x3C\xE3\x4C\x2C\xA2\x48"
"\x1C\x61\x44\x0C\x20\x40\xFF\xEF\x7C\xEF\xAE\x78\xDF\x6D\x74\xCF\x2C\x70"
"\xBE\xEB\x6C\xAE\xA1\x61\x91\x61\x61\x81\x28\x60\x7D\xE7\x5C\x6D\xA6\x58"
"\x5D\x65\x54\x4D\x24\x50\x3C\xE3\x4C\x2C\xA2\x48\x1C\x61\x44\x0C\x20\x40",
"9IJJI876543210-ZYXWVUTSRQPONMLK9LMJIHGFEDCBAzyxwvutsrqponmlkjihgfedcbapZ"
"776543210-ZYXWVUTSRQfHKwfHGsHGFEDCBAzyxwvutsrqponmlkjihgfedcbaML87654321"
"0-ZYXWVUTSRQfHKwfHGsHGFEDCBAzyxwvutsrqponmlkjihgfedcba"
},
{ NULL, NULL } { NULL, NULL }
}; };