bashclub/piler
1
0
Fork 0
mirror of https://bitbucket.org/jsuto/piler.git synced 2025-06-13 13:17:03 +02:00
Code Issues Packages Projects Releases Wiki Activity

src: decoder and parser fix

Browse Source 
Signed-off-by: Janos SUTO <sj@acts.hu>
This commit is contained in:
Janos SUTO
2017-11-08 11:50:28 +01:00
parent 25aeff66ba
commit 1192fc3218
4 changed files with 134 additions and 95 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
src/decoder.c
View File

@ -78,6 +78,8 @@ inline void utf8_encode_char(unsigned char c, unsigned char *buf, int buflen, in
* U+0000..U+007F 00..7F
* U+0080..U+07FF C2..DF 80..BF
* U+0800..U+0FFF E0 A0..BF 80..BF
*
* FIXME: See http://www.unicode.org/versions/Unicode5.2.0/ch03.pdf#G7404 for valid sequences
*/
if(c <= 0x7F){

136
src/parser_utils.c
View File

@ -196,6 +196,7 @@ time_t parse_date_header(char *datestr){
else if(strncasecmp(s, "Sat", 3) == 0) tm.tm_wday = 6;
else if(strncasecmp(s, "Sun", 3) == 0) tm.tm_wday = 0;
if(len <= 2 && tm.tm_mday == 0){ tm.tm_mday = atoi(s); continue; }
if(len <= 2 && tm.tm_mon == -1){ tm.tm_mon = atoi(s) - 1; continue; }
@ -313,17 +314,19 @@ int extract_boundary(char *p, struct parser_state *state){
void fixupEncodedHeaderLine(char *buf, int buflen){
char *sb, *sq, *p, *q, *r, *s, *e, *start, *end;
char *p, *q, *r, *s, *e, *end;
/*
* I thought SMALLBUFSIZE would be enough for v, encoding and tmpbuf(2*),
* but then I saw a 6-7000 byte long subject line, so I've switched to MAXBUFSIZE
*/
char v[MAXBUFSIZE], puf[MAXBUFSIZE], encoding[MAXBUFSIZE], tmpbuf[2*MAXBUFSIZE];
char v[MAXBUFSIZE], u[MAXBUFSIZE], puf[MAXBUFSIZE], encoding[MAXBUFSIZE], tmpbuf[2*MAXBUFSIZE];
int need_encoding, ret;
if(buflen < 5) return;
memset(puf, 0, sizeof(puf));
memset(encoding, 0, sizeof(encoding));
q = buf;
@ -332,69 +335,89 @@ void fixupEncodedHeaderLine(char *buf, int buflen){
p = v;
memset(encoding, 0, sizeof(encoding));
do {
start = strstr(p, "=?");
if(start){
*start = '\0';
if(strlen(p) > 0){
strncat(puf, p, sizeof(puf)-strlen(puf)-1);
memset(u, 0, sizeof(u));
/*
* We can't use split_str(p, "=?", ...) it will fail with the following pattern
* =?UTF-8?B?SG9neWFuIMOtcmp1bmsgcGFuYXN6bGV2ZWxldD8=?=
*
* Also the below patter requires special care:
* =?gb2312?B?<something>?==?gb2312?Q?<something else>?=
*/
r = strstr(p, "=?");
if(r){
p = r + 2;
end = strstr(p, "?=");
if(end){
*end = '\0';
}
start++;
snprintf(u, sizeof(u)-1, "%s", p);
e = strchr(start+2, '?');
if(e){
*e = '\0';
snprintf(encoding, sizeof(encoding)-1, "%s", start+1);
*e = '?';
}
s = NULL;
sb = strcasestr(start, "?B?"); if(sb) s = sb;
sq = strcasestr(start, "?Q?"); if(sq) s = sq;
if(s){
end = strstr(s+3, "?=");
if(end){
*end = '\0';
if(sb){ decodeBase64(s+3); }
if(sq){ decodeQP(s+3); r = s + 3; for(; *r; r++){ if(*r == '_') *r = ' '; } }
/* encode everything if it's not utf-8 encoded */
need_encoding = 0;
ret = ERR;
if(strlen(encoding) > 2 && strcasecmp(encoding, "utf-8")){
need_encoding = 1;
ret = utf8_encode(s+3, strlen(s+3), &tmpbuf[0], sizeof(tmpbuf), encoding);
}
if(need_encoding == 1 && ret == OK)
strncat(puf, tmpbuf, sizeof(puf)-strlen(puf)-1);
else
strncat(puf, s+3, sizeof(puf)-strlen(puf)-1);
p = end + 2;
}
}
else {
strncat(puf, start, sizeof(puf)-strlen(puf)-1);
break;
if(end) {
p = end + 2;
}
}
else {
strncat(puf, p, sizeof(puf)-strlen(puf)-1);
break;
snprintf(u, sizeof(u)-1, "%s", p);
p = NULL;
}
if(u[0] == 0) continue;
memset(encoding, 0, sizeof(encoding));
// Check if it's either ?B? or ?Q? encoding ...
s = strcasestr(u, "?B?");
if(s){
decodeBase64(s+3);
}
else {
s = strcasestr(u, "?Q?");
if(s){
decodeQP(s+3);
r = s + 3;
for(; *r; r++){
if(*r == '_') *r = ' ';
}
}
}
// ... if it is, then get the encoding
if(s){
e = strchr(u, '?');
if(e){
*e = '\0';
snprintf(encoding, sizeof(encoding)-1, "%s", u);
*e = '?';
need_encoding = 0;
ret = ERR;
if(encoding[0] && strcasecmp(encoding, "utf-8")){
need_encoding = 1;
ret = utf8_encode(s+3, strlen(s+3), &tmpbuf[0], sizeof(tmpbuf), encoding);
}
if(need_encoding == 1 && ret == OK)
strncat(puf, tmpbuf, sizeof(puf)-strlen(puf)-1);
else
strncat(puf, s+3, sizeof(puf)-strlen(puf)-1);
}
else {
memset(encoding, 0, sizeof(encoding));
strncat(puf, u, sizeof(puf)-strlen(puf)-1);
}
}
else {
strncat(puf, u, sizeof(puf)-strlen(puf)-1);
}
} while(p);
if(q) strncat(puf, " ", sizeof(puf)-strlen(puf)-1);
if(q && encoding[0] == 0) strncat(puf, " ", sizeof(puf)-strlen(puf)-1);
} while(q);
@ -599,6 +622,7 @@ void translateLine(unsigned char *p, struct parser_state *state){
prev = *p;
}
if(state->message_state == MSG_SUBJECT && (*p == '%' || *p == '_' || *p == '&') ){ continue; }
if(state->message_state == MSG_CONTENT_TYPE && *p == '_' ){ continue; }
@ -658,8 +682,7 @@ int does_it_seem_like_an_email_address(char *email){
*/
void reassembleToken(char *p){
unsigned int i;
int k=0;
unsigned int i, k=0;
for(i=0; i<strlen(p); i++){
@ -959,3 +982,4 @@ void fix_plus_sign_in_email_address(char *puf, char **at_sign, unsigned int *len
*at_sign = r;
}
}