parser fix to support some emojis in subject line

Signed-off-by: Janos SUTO <sj@acts.hu>

src/decoder.c

@@ -206,9 +206,13 @@ void decodeQP(char *p){
 
          i += 2;
       }
+      else if(p[i] == '_'){
+         c = ' ';
+      }
 
       p[k] = c;
       k++;
+
    }
 
    p[k] = '\0';

src/parser_utils.c

@@ -320,13 +320,12 @@ void fixupEncodedHeaderLine(char *buf, int buflen){
     * but then I saw a 6-7000 byte long subject line, so I've switched to MAXBUFSIZE
     */
    char v[MAXBUFSIZE], u[MAXBUFSIZE], puf[MAXBUFSIZE], encoding[MAXBUFSIZE], tmpbuf[2*MAXBUFSIZE];
-   int need_encoding, ret;
+   int need_encoding, ret, prev_encoded=0, n_tokens=0;
+   int b64=0, qp=0;
 
    if(buflen < 5) return;
 
    memset(puf, 0, sizeof(puf));
-   memset(encoding, 0, sizeof(encoding));
-
 
    q = buf;
 
@@ -342,18 +341,46 @@ void fixupEncodedHeaderLine(char *buf, int buflen){
           * 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:
+          * Also the below pattern requires special care:
           *    =?gb2312?B?<something>?==?gb2312?Q?<something else>?=
+          *
+          * And we have to check the following cases as well:
+          *    Happy New Year! =?utf-8?q?=F0=9F=8E=86?=
           */
 
+         b64 = qp = 0;
+         memset(encoding, 0, sizeof(encoding));
+
          r = strstr(p, "=?");
          if(r){
             p = r + 2;
+
+            e = strchr(p, '?');
+            if(e){
+               *e = '\0';
+               snprintf(encoding, sizeof(encoding)-1, "%s", p);
+               *e = '?';
+
+               s = strcasestr(e, "?B?");
+               if(s){
+                  b64 = 1;
+                  p = s + 3;
+               }
+               else {
+                  s = strcasestr(e, "?Q?");
+                  if(s){
+                     qp = 1;
+                     p = s + 3;
+                  }
+               }
+            }
+
             end = strstr(p, "?=");
             if(end){
                *end = '\0';
             }
 
+
             snprintf(u, sizeof(u)-1, "%s", p);
 
             if(end) {
@@ -367,47 +394,38 @@ void fixupEncodedHeaderLine(char *buf, int buflen){
 
          if(u[0] == 0) continue;
 
-         memset(encoding, 0, sizeof(encoding));
+         n_tokens++;
 
-         // Check if it's either ?B? or ?Q? encoding ...
+         if(b64 == 1) decodeBase64(u);
-         s = strcasestr(u, "?B?");
+         else if(qp == 1) decodeQP(u);
-         if(s){
+
-            decodeBase64(s+3);
+
-         }
+         /*
-         else {
+          * https://www.ietf.org/rfc/rfc2047.txt says that
-            s = strcasestr(u, "?Q?");
+          *
-            if(s){
+          * "When displaying a particular header field that contains multiple
-               decodeQP(s+3);
+          *  'encoded-word's, any 'linear-white-space' that separates a pair of
-               r = s + 3;
+          *  adjacent 'encoded-word's is ignored." (6.2)
-               for(; *r; r++){
+          */
-                  if(*r == '_') *r = ' ';
+         if(prev_encoded == 1 && (b64 == 1 || qp == 1)) {}
-               }
+         else if(n_tokens > 1){
-            }
+            strncat(puf, " ", sizeof(puf)-strlen(puf)-1);
          }
 
-         // ... if it is, then get the encoding
+         if(b64 == 1 || qp == 1){
-         if(s){
+            prev_encoded = 1;
-            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);
+               ret = utf8_encode(u, strlen(u), &tmpbuf[0], sizeof(tmpbuf), encoding);
             }
 
-               if(need_encoding == 1 && ret == OK)
+            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);
             }
          }
@@ -417,8 +435,6 @@ void fixupEncodedHeaderLine(char *buf, int buflen){
 
       } while(p);
 
-      if(q && encoding[0] == 0) strncat(puf, " ", sizeof(puf)-strlen(puf)-1);
-
    } while(q);
 
    snprintf(buf, buflen-1, "%s", puf);

unit_tests/check_parser_utils.c

@@ -154,6 +154,8 @@ static void test_fixupEncodedHeaderLine(){
       {"Subject: =?windows-1251?B?ze7i7uPu5O3o5SDv7uTg8OroIOTr/yDC4Pjo?=", "Subject: Новогодние подарки для Ваши"},
       {"Subject: =?utf-8?Q?Divatos,_=C3=BCde_sz=C3=ADneinek_k=C3=B6sz=C3=B6nhet=C5=91en_el?=", "Subject: Divatos, üde színeinek köszönhetően el"},
       {"=?gb2312?B?yc/Gz76pIC0gw7/fTMir0bKy6YjzuOYgKDIwMTcxMDMwLTMxKSBHQlcgUG9k?==?gb2312?Q?ium_&_Basement.docx?=", "上葡京 - 每週全巡查報告 (20171030-31) GBW Podium & Basement.docx"},
+      {"Subject: =?UTF-8?Q?=E2=98=85_JubiDu!Versandkost?= =?UTF-8?Q?enfrei-Verl=C3=A4ngerung!=E2=98=85?=", "Subject: ★ JubiDu!Versandkostenfrei-Verlängerung!★"},
+      {"Happy New Year! =?utf-8?q?=F0=9F=8E=86?=", "Happy New Year! 🎆"},
    };
 
    TEST_HEADER();

unit_tests/run.sh

@@ -5,8 +5,10 @@ set -o pipefail
 set -o nounset
 set -x
 
-LD_LIBRARY_PATH=../src ./check_parser_utils
+export LD_LIBRARY_PATH=../src
-LD_LIBRARY_PATH=../src ./check_parser
+
-LD_LIBRARY_PATH=../src ./check_rules
+./check_parser_utils
-LD_LIBRARY_PATH=../src ./check_digest
+./check_parser
-LD_LIBRARY_PATH=../src ./check_mydomains
+./check_rules
+./check_digest
+./check_mydomains