qmail-remote: reimplement SMTPUTF8 support

make sure to deliver the mail to non-SMTPUTF8 capable remotes unless
there are upper byte values found in headers/envelopes.
this is more-or-less the same as postfix does.

1 files changed, 91 insertions, 44 deletions

diff --git a/qmail-remote.c b/qmail-remote.c
index 7eb2714..d7c8746 100644
--- a/qmail-remote.c
+++ b/qmail-remote.c
@@ -47,7 +47,6 @@ stralloc helohost = {0};
 stralloc routes = {0};
 struct constmap maproutes;
 stralloc host = {0};
-stralloc idnhost = {0};
 stralloc sender = {0};
 stralloc auth_smtp_user = {0};
 stralloc auth_smtp_pass = {0};
@@ -74,7 +73,9 @@ char **myargv;
 
 #ifdef SMTPUTF8
 # include <idn2.h>
-int flagutf8 = 0;
+int withutf8 = 0; /* received header has WITH SMTPUTF8 */
+int utf8message = 0; /* headers contain utf8 */
+stralloc idnhost = {0};
 stralloc header = {0};
 #endif
 
@@ -244,6 +245,15 @@ unsigned long smtpcode()
 saa ehlokw = {0}; /* list of EHLO keywords and parameters */
 int maxehlokwlen = 0;
 
+/* look for <capa> among EHLO keywords */
+int get_capa(const char *capa)
+{
+  stralloc *sa = ehlokw.sa;
+  unsigned int len = ehlokw.len;
+  for (; len && case_diffs(sa->s, capa); ++sa, --len);
+  return len > 0;
+}
+
 unsigned long ehlo()
 {
   stralloc *sa;
@@ -473,12 +483,13 @@ static int tls_init(unsigned long code, struct ip_mx *current_mx)
       tls_required = 1;
   }
 
-  if (!smtps) {
     stralloc *sa = ehlokw.sa;
     unsigned int len = ehlokw.len;
-    /* look for STARTTLS among EHLO keywords */
     for ( ; len && case_diffs(sa->s, "STARTTLS"); ++sa, --len) ;
-    if (!len) {
+    /* look for STARTTLS among EHLO keywords */
+
+  if (!smtps) {
+    if (!get_capa("STARTTLS")) {
       if (!tls_required) return 0;
       out(failure_class);
       out("TLS is required, but was not offered by host");
@@ -682,9 +693,6 @@ static int tls_init(unsigned long code, struct ip_mx *current_mx)
   }
 
   if (servercert) {
-    X509 *peercert;
-    STACK_OF(GENERAL_NAME) *gens;
-
     int r = SSL_get_verify_result(myssl);
     if (r != X509_V_OK) {
       out("ZTLS unable to verify server with ");
@@ -724,41 +732,74 @@ int utf8string(unsigned char *ch, int len)
   return 0;
 }
 
-int utf8received()
+void checkutf8message()
 {
-  int r, i, received = 0;
-  char ch;
-  stralloc receivedline = {0};
+  if (utf8string((unsigned char *)sender.s, sender.len)) {
+    utf8message = 1;
+    return;
+  }
+  for (unsigned int i = 0; i < reciplist.len; ++i) {
+    if (utf8string((unsigned char *)reciplist.sa[i].s, reciplist.sa[i].len)) {
+      utf8message = 1;
+      return;
+    }
+  }
 
-  for (;;) {                                    /* we consider only our own last written header */
-    r = substdio_get(&ssin,&ch,1);
+  int state = 0;
+  int pos = 0;
+  for (;;) {
+    char ch;
+    int r = substdio_get(&ssin, &ch, 1);
     if (r == 0) break;
     if (r == -1) temp_read();
 
-    if (ch == '\n' && receivedline.len) {
-      if (!stralloc_append(&header,"\r")) temp_nomem();                 /* received.c does not add '\r' */
-      if (!stralloc_append(&header,"\n")) temp_nomem();
-      if (case_startb(receivedline.s,5,"Date:")) return 0;              /* header to quit asap */
-      if (case_startb(receivedline.s,14,"Received: from")) received++;  /* found Received header */
-      if (received) {
-        if (case_startb(receivedline.s,5,"  by ")) {
-          for (i = 6; i < receivedline.len-6; ++i)
-            if (*(receivedline.s+i) == ' ')
-              if (case_startb(receivedline.s+i+1,9,"with UTF8")) return 1;
-          return 0;
-        }
-      }
-      if (!stralloc_copys(&receivedline,"")) temp_nomem();
-    } else if (ch == '\n' && !receivedline.len) {    /* we got an empty newline. probably body start */
-      if (!stralloc_append(&header,"\r")) temp_nomem();
-      if (!stralloc_append(&header,"\n")) temp_nomem();
-      return 0;
-    } else {
-      if (!stralloc_append(&header,&ch)) temp_nomem();
-      if (!stralloc_catb(&receivedline,&ch,1)) temp_nomem();
+    if (ch == '\n' && !stralloc_cats(&header, "\r")) temp_nomem(); /* received.c does not add '\r' */
+    if (!stralloc_append(&header, &ch)) temp_nomem();
+    if (ch == '\r')
+      continue;
+    if ((unsigned char)ch > 127)
+      utf8message = 1;
+    if (utf8message) {
+      /* no need to look any further */
+      if (ch == '\n') return; /* end of header */
+      continue;
+    }
+    if (ch == '\t')
+      ch = ' ';
+
+    switch (state) {
+    case 6: /* in Received, at LF but before WITH clause */
+      if (ch == ' ') { state = 3; pos = 1; continue; }
+      state = 0;
+      /* FALL THROUGH */
+    case 0: /* start of header field */
+      if (ch == '\n') return; /* end of headers */
+      state = 1;
+      pos = 0;
+      /* FALL THROUGH */
+    case 1: /* partway through "Received:" */
+      if (ch != "RECEIVED:"[pos] && ch != "received:"[pos]) { state = 2; continue; }
+      if (++pos == 9) { state = 3; pos = 0; }
+      continue;
+    case 2: /* other header field */
+      if (ch == '\n') state = 0;
+      continue;
+    case 3: /* in Received, before WITH clause or partway though " with " */
+      if (ch == '\n') { state = 6; continue; }
+      if (ch != " WITH "[pos] && ch != " with "[pos]) { pos = 0; continue; }
+      if (++pos == 6) { state = 4; pos = 0; }
+      continue;
+    case 4: /* in Received, having seen with, before the argument */
+      if (pos == 0 && (ch == ' ' || ch == '\t')) continue;
+      if (ch != "UTF8"[pos] && ch != "utf8"[pos]) { state = 5; continue; }
+      if (++pos == 4) { withutf8 = 1; state = 5; continue; }
+      continue;
+    case 5: /* after the RECEIVED WITH argument */
+      if (ch == '\n') state = 0;
+      /* blast() assumes that it copies whole lines */
+      continue;
     }
   }
-  return 0;
 }
 #endif
 
@@ -768,7 +809,7 @@ void mail_without_auth()
   substdio_put(&smtpto,sender.s,sender.len);
   substdio_put(&smtpto,">",1);
 #ifdef SMTPUTF8
-  if (flagutf8 || utf8received())
+  if (utf8message || withutf8)
     substdio_puts(&smtpto, " SMTPUTF8");
 #endif
   substdio_puts(&smtpto,"\r\n");
@@ -826,6 +867,17 @@ static void smtp(struct ip_mx *current_mx)
     if (code != 250) quit("ZConnected to "," but my name was rejected");
   }
 
+#ifdef SMTPUTF8
+  checkutf8message();
+  int can_utf8 = get_capa("SMTPUTF8");
+  if (utf8message && !can_utf8) {
+    smtptext.len = 0;
+    quit("DConnected to "," but the remote host does not support SMTPUTF8 and the message requires it");
+  }
+  if (!utf8message && withutf8 && !can_utf8)
+    withutf8 = 0; // downgrade
+#endif
+
   int i = 0;
   if (auth_smtp_user.len && auth_smtp_pass.len)  {
     while((i += str_chr(smtptext.s+i,'\n') + 1) &&
@@ -857,7 +909,7 @@ static void smtp(struct ip_mx *current_mx)
       substdio_put(&smtpto,sender.s,sender.len);
       substdio_put(&smtpto,">",1);
 #ifdef SMTPUTF8
-      if (flagutf8 || utf8received())
+      if (utf8message || withutf8)
         substdio_puts(&smtpto, " SMTPUTF8");
 #endif
       substdio_puts(&smtpto,"\r\n");
@@ -1032,11 +1084,6 @@ char *s;
 {
   int j;
 
-#ifdef SMTPUTF8
-  if (!flagutf8)
-    flagutf8 = utf8string(s, str_len(s));
-#endif
-
   j = str_rchr(s,'@');
   if (!s[j]) {
     if (!stralloc_copys(saout,s)) temp_nomem();
@@ -1124,7 +1171,7 @@ char **argv;
     char *asciihost = 0;
     if (!stralloc_0(&host)) temp_nomem();
     --host.len;
-    switch (idn2_lookup_u8(host.s,(uint8_t**)&asciihost,IDN2_NFC_INPUT)) {
+    switch (idn2_lookup_u8((uint8_t*)host.s,(uint8_t**)&asciihost,IDN2_NFC_INPUT)) {
       case IDN2_OK: break;
       case IDN2_MALLOC: temp_nomem();
       default: perm_dns();