[mod_auth] fix base64_decode (#2484)
git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2871 152afb58-edef-0310-8abb-c4023f1b3aa9
This commit is contained in:
parent
c26b50d9ad
commit
ce4bc0c0f7
1
NEWS
1
NEWS
|
@ -12,6 +12,7 @@ NEWS
|
|||
* reject non ASCII characters in HTTP header names
|
||||
* [mod_auth] use crypt() on encrypted password instead of extracting salt first (fixes #2483)
|
||||
* [mod_auth] add htpasswd -s (SHA1) support if openssl is used (needs openssl for SHA1). This doesn't use any salt, md5 with salt is probably better.
|
||||
* [mod_auth] fix base64_decode (#2484)
|
||||
|
||||
- 1.4.32 - 2012-11-21
|
||||
* Code cleanup with clang/sparse (fixes #2437, thx kibi)
|
||||
|
|
|
@ -91,7 +91,8 @@ static const short base64_reverse_table[256] = {
|
|||
|
||||
static unsigned char * base64_decode(buffer *out, const char *in) {
|
||||
unsigned char *result;
|
||||
int ch, j = 0, k;
|
||||
unsigned int j = 0; /* current output character (position) that is decoded. can contain partial result */
|
||||
unsigned int group = 0; /* how many base64 digits in the current group were decoded already. each group has up to 4 digits */
|
||||
size_t i;
|
||||
|
||||
size_t in_len = strlen(in);
|
||||
|
@ -100,51 +101,64 @@ static unsigned char * base64_decode(buffer *out, const char *in) {
|
|||
|
||||
result = (unsigned char *)out->ptr;
|
||||
|
||||
ch = in[0];
|
||||
/* run through the whole string, converting as we go */
|
||||
for (i = 0; i < in_len; i++) {
|
||||
ch = (unsigned char) in[i];
|
||||
unsigned char c = (unsigned char) in[i];
|
||||
short ch;
|
||||
|
||||
if (ch == '\0') break;
|
||||
if (c == '\0') break;
|
||||
|
||||
if (ch == base64_pad) break;
|
||||
if (c == base64_pad) {
|
||||
/* pad character can only come after 2 base64 digits in a group */
|
||||
if (group < 2) return NULL;
|
||||
break;
|
||||
}
|
||||
|
||||
ch = base64_reverse_table[ch];
|
||||
if (ch < 0) continue;
|
||||
ch = base64_reverse_table[c];
|
||||
if (ch < 0) continue; /* skip invalid characters */
|
||||
|
||||
switch(i % 4) {
|
||||
switch(group) {
|
||||
case 0:
|
||||
result[j] = ch << 2;
|
||||
group = 1;
|
||||
break;
|
||||
case 1:
|
||||
result[j++] |= ch >> 4;
|
||||
result[j] = (ch & 0x0f) << 4;
|
||||
group = 2;
|
||||
break;
|
||||
case 2:
|
||||
result[j++] |= ch >>2;
|
||||
result[j] = (ch & 0x03) << 6;
|
||||
group = 3;
|
||||
break;
|
||||
case 3:
|
||||
result[j++] |= ch;
|
||||
group = 0;
|
||||
break;
|
||||
}
|
||||
}
|
||||
k = j;
|
||||
/* mop things up if we ended on a boundary */
|
||||
if (ch == base64_pad) {
|
||||
switch(i % 4) {
|
||||
case 0:
|
||||
case 1:
|
||||
return NULL;
|
||||
case 2:
|
||||
k++;
|
||||
case 3:
|
||||
result[k++] = 0;
|
||||
}
|
||||
}
|
||||
result[k] = '\0';
|
||||
|
||||
out->used = k;
|
||||
switch(group) {
|
||||
case 0:
|
||||
/* ended on boundary */
|
||||
break;
|
||||
case 1:
|
||||
/* need at least 2 base64 digits per group */
|
||||
return NULL;
|
||||
case 2:
|
||||
/* have 2 base64 digits in last group => one real octect, two zeroes padded */
|
||||
case 3:
|
||||
/* have 3 base64 digits in last group => two real octects, one zero padded */
|
||||
|
||||
/* for both cases the current index already is on the first zero padded octet
|
||||
* - check it really is zero (overlapping bits) */
|
||||
if (0 != result[j]) return NULL;
|
||||
break;
|
||||
}
|
||||
|
||||
result[j] = '\0';
|
||||
out->used = j;
|
||||
|
||||
return result;
|
||||
}
|
||||
|
|
Loading…
Reference in New Issue