mid debug

4 files changed, 185 insertions, 89 deletions

diff --git a/tools/libskrewt.c b/tools/libskrewt.c
index 68fe741..dc131f9 100644
--- a/tools/libskrewt.c
+++ b/tools/libskrewt.c
@@ -7,7 +7,7 @@ using namespace std;
 
 void parse_content(const string type_spec_line,
         string &maintype, string &boundary) {
-  //xxx cerr << "parser called with: " << type_spec_line << endl;
+  cerr << "parser called with: " << type_spec_line << endl;
   string get_type(type_spec_line);
 
   size_t where = get_type.find_first_of(" \t;\n");
@@ -61,65 +61,96 @@ int skrewt::krunch_rfrom(){
     return ex_syserr;
   }
   parse >> proximta_rDNS;
-  parse >> word;
-  if (word == "(HELO" /*)*/) {
-    parse >> proximta_HELO;
-    proximta_HELO = rtrim(proximta_HELO, "()");
+  for (;;) {
     parse >> word;
-  } else {
-    proximta_HELO = proximta_rDNS;
-  }
-  size_t len = word.length();
-  if (len<2 || word[0] != '(' || word[len-1] != ')') {
-    cerr << progid << " bad 'Received: from' line ;;; '"
-        << word << "'" << endl;
-    return ex_syserr;
+    size_t len = word.length();
+    if (len < 2) {
+      cerr << progid << " bad 'Received: from' line ;;; short '"
+          << word << "'" << endl;
+      return ex_syserr;
+    }
+    if (word == "by") break;
+    if (word == "(HELO" /*)*/) {
+      parse >> proximta_HELO;
+      proximta_HELO = rtrim(proximta_HELO, "()");
+      continue;
+    }
+    if (word[0] != '(' || word[len-1] != ')') {
+      cerr << progid << " bad 'Received: from' line ;;; no parens: '"
+          << word << "'" << endl;
+      return ex_syserr;
+    }
+    proximta_IP = word.substr(1, len-2);
+    size_t where = proximta_IP.find("@");
+    if (where != string::npos){
+      proximta_AuthUser = proximta_IP.substr(0, where);
+      proximta_IP = proximta_IP.substr(1+where);
+    }
   }
-  proximta_IP = word.substr(1, len-2);
-  size_t where = proximta_IP.find("@");
-  if (where != string::npos){
-    proximta_AuthUser = proximta_IP.substr(0, where);
-    proximta_IP = proximta_IP.substr(1+where);
+// provide some kind of default? maybe not.
+  if (0) if (proximta_HELO == "") {
+    proximta_HELO = proximta_rDNS + " (+-)";
   }
 
   return 0;
 }
 
-int skrewt::headers(istream& xin){
-  //xxxx cerr << progid << " begins" << endl;
-  for (;;){             // outer loop over all records in the header
-    if (xin.eof()) break;
-    if (xin.bad()) return 1;
+xstr skrewt::getLine(istream& xin){
+  string rslt;
+  if (lookahead.err == 0) {
+    rslt = lookahead.str;
+    lookahead.err = 1;
+    return {0, rslt};
+  }
+  if (getline(xin, rslt).fail()) return {1, ""};
+  else return {0, rslt};
+}
 
-    string line;
-// on fail, go back to top of outer loop and check for eof versus bad
-    if (getline(xin, line).fail()) continue;
-    msgsize += line.length()+1;
+xstr skrewt::getRecord(istream& xin){
+    xstr line = getLine(xin);
+    if (line.err) return line;
+    msgsize += line.str.length()+1;
     if (msgsize > maxsize) {
-      cerr << progid << " rejection: bigger than " << maxsize << endl;
+      cerr << progid << " header rejection: bigger than " << maxsize << endl;
       exeunt(ex_spam);
     }
-    bigbuf.push_back(line);
-    string headrec = noCR(line);       // for a folded record, this is the first line
+    bigbuf.push_back(line.str);
+// for a folded record, this is the first line:
+    string headrec = noCR(line.str);
 
-    for (;;) {        // inner loop to build a multi-line record e.g. folded record:
+// inner loop to build a multi-line record e.g. folded record:
+    for (;;) {
       if (xin.eof()) break;
-      if (xin.bad()) return 1;
-      char ch;
-      if (xin.get(ch).fail()) continue;
-      xin.putback(ch);
-      if (ch != ' ' && ch != '\t') break;
-      string line;
-// on fail, go back to top of inner loop and check for eof versus bad
-      if (getline(xin, line).fail()) continue;
-      msgsize += line.length()+1;
+      if (xin.bad()) return {1, ""};
+      xstr more = getLine(xin);
+      if (more.err) continue;
+      char ch(0);
+      if (more.str != "") ch = more.str[0];
+      if (ch != ' ' && ch != '\t') {
+        lookahead = more;
+        break;
+      }
+      msgsize += line.str.length()+1;
       if (msgsize > maxsize) {
-        cerr << progid << " rejection: bigger than " << maxsize << endl;
+        cerr << progid
+          << " folded header rejection: bigger than " << maxsize << endl;
         exeunt(ex_spam);
       }
-      bigbuf.push_back(line);
-      headrec += "\n" + noCR(line);
+      bigbuf.push_back(line.str);
+      headrec += "\n" + noCR(line.str);
     }
+    return {0, headrec};
+}
+
+int skrewt::headers(istream& xin){
+  //xxxx cerr << progid << " begins" << endl;
+  for (;;){             // outer loop over all records in the header
+    if (xin.eof()) break;
+    if (xin.bad()) return 1;
+
+    xstr rec = getRecord(xin);
+    if (rec.err) break;
+    string headrec = rec.str;
 // here with a fully assembled header record
 // headrec (unlike line) contains no DOS CR characters
     int len = headrec.length();
@@ -146,7 +177,12 @@ int skrewt::headers(istream& xin){
       return_path = rest;
     } else if (headword == "message-id") {
       message_id = rest;
-    } else if (headword == "received") {
+    }
+// Now pick up the proximal "received: from" line,
+// i.e. the latest in time,
+// i.e. the topmost in the file,
+// i.e. the one that was just now put there by qmail-smtpd:
+    else if (headword == "received") {
       if (!received_from.length() && prefix("from ", rest)){
         received_from = rest;
       }
@@ -157,7 +193,7 @@ int skrewt::headers(istream& xin){
     } else if (headword == "content-type") {
       content_type = rest;
     } else if (headword == "delivered-to") {
-      delivered_to = rest;
+      delivered_to.push_back(rest);
     }
     recno++;
     if (0) if (recno <= 6) cerr << progid << "#" << recno
@@ -176,30 +212,39 @@ int skrewt::dump_headers(std::ostream& xout){
 
 int skrewt::interstage(){
   if (saw_blank_line) {/* ignore */}
-// Note that the headers are in reverse-chronological order:
-  cerr << progid <<" Return-path: " << return_path <<endl;
+// Note that the headers are in reverse-chronological order.
+// Also note that there probably is no return-path
+// at the time skrewt normally runs, since it is upstream
+// of qmail-queue.
+  cerr << progid << " Return-path: "
+    << (return_path.length() ? return_path : "[not set yet]") <<endl;
 
   { // parse the 'Received: from' line:
-    cerr << "        Received: " << received_from <<endl;
+    cerr << progid << " Received: " << received_from <<endl;
     int rslt = krunch_rfrom();
     if (rslt) return rslt;
-    cerr << "         rDNS:     " << proximta_rDNS << endl;
-    cerr << "         HELO:     " << proximta_HELO << endl;
-    cerr << "         IP:       "   << proximta_IP << endl;
-    cerr << "         AuthUser: "   << proximta_AuthUser << endl;
-    cerr << "         Mid       '"  << message_id << "'" << endl;
+    cerr << progid << " === rDNS:     "  << proximta_rDNS << endl;
+    cerr << progid << " === HELO:     "  << proximta_HELO << endl;
+    cerr << progid << " === IP:       "  << proximta_IP << endl;
+    cerr << progid << " === Mid       '" << message_id << "'" << endl;
+    cerr << progid << " === AuthUser: "  << proximta_AuthUser << endl;
   }
 
-  sepofra my_spf;
-  try {
-    my_spf.check(proximta_IP,
-        proximta_HELO,
-        return_path,
-        "junk", 0/* verbosity */);
-    cerr << "*** " << my_spf.explain() << endl;
-  } catch (bad_thing foo) {
-    cerr << "Caught bad thing: " << foo.what() << endl;
-    return ex_syserr;
+  if (proximta_AuthUser == "")
+   if (proximta_IP != "") {
+    sepofra my_spf;
+    try {
+      my_spf.check(proximta_IP,
+          proximta_HELO,
+          return_path,
+          "junk", 0/* verbosity */);
+      cerr << progid << " " << my_spf.explain() << endl;
+// keep a copy of the result:
+      spf_result = my_spf.result;
+    } catch (bad_thing foo) {
+      cerr << "Caught bad thing: " << foo.what() << endl;
+      return ex_syserr;
+    }
   }
 
 // The logic here is:  In order:
@@ -311,11 +356,12 @@ int skrewt::interstage(){
     whitelisted++;
   }
 
-  if (delivered_to.length()){
-    cerr << progid <<  " Delivered-to: <<<" << delivered_to << ">>>" << endl;
-  }
-  if (toLower(trim(delivered_to)) == "jean@av8n.com") {
-    badnews.push_back("Looping Delivered-to: " + delivered_to);
+  for (list<string>::const_iterator foo = delivered_to.begin();
+        foo != delivered_to.end(); foo++){
+    cerr << progid <<  " Delivered-to: <<<" << *foo << ">>>" << endl;
+    if (toLower(trim(*foo)) == "jean@av8n.com") {
+      badnews.push_back("Looping Delivered-to: " + *foo);
+    }
   }
 
   if (subject.find("-please-bounce-this-") != string::npos) {
@@ -333,7 +379,7 @@ int skrewt::interstage(){
   if (badnews.size() && !whitelisted){
     cerr << progid << " " << join(", ", badnews) << endl;
     if (error_exit){
-      cerr << progid << " '" << from
+      cerr << progid << " ... badnews from '" << from
              << "' to '" << to
              << "'" << endl;
       exeunt(ex_spam);
@@ -342,22 +388,35 @@ int skrewt::interstage(){
   return 0;
 }
 
+class conner{
+public:
+  string tp;
+  string bdy;
+};
+
 int skrewt::body(std::istream& xin, std::ostream& xout){
-  string main_contype;
-  if (content_type.length())
-    parse_content(content_type, main_contype, boundary);
-// some slightly-useful booleans:
-  int currently_text = main_contype == "text";
-  int main_multipart = main_contype == "multipart";
+  list<conner> sitch;
+  if (content_type.length()) {
+    string tp, bdy;
+    parse_content(content_type, tp, bdy);
+    sitch.push_front({tp, bdy});
+  } else {
+    // assume single-part, all text if not otherwise specified:
+    sitch.push_front({"text", ""});
+  }
 
 // early-stage thinking has been done.
 // Now spew the rest of the message
 // Note that multipart messages can have sub-headers.
 
-  //xxxx cerr << "body begins: " << main_contype << " " << currently_text << " " << boundary << endl;
+  cerr << "body begins: tp: " << sitch.front().tp
+       << "  bdy: " << sitch.front().bdy
+       << "  depth: " << sitch.size()
+       << endl;
   int in_subheads(0);
   int textlines(0);
 
+  string sub_conntype("");
   for (;;){             // outer loop over all lines in the body
     if (xin.eof()) break;
     if (xin.bad()) return 1;
@@ -371,14 +430,27 @@ int skrewt::body(std::istream& xin, std::ostream& xout){
     }
     bigbuf.push_back(line);
     xout << line << endl;
+    if (0) cerr << "+++" << line << endl;
     if (in_subheads){
-      if (line == "" || line == "\r") in_subheads = 0;
+      if (line == "" || line == "\r") {
+        in_subheads = 0;
+        cerr << "+++ end of subhead" << endl;
+        string tp, bdy;
+        parse_content(sub_conntype, tp, bdy);
+        cerr << tp << "   '" << bdy << "'" << endl;
+        sitch.push_front({tp, bdy});
+        cerr << "payload mode: tp: " << sitch.front().tp
+             << "  bdy: " << sitch.front().bdy
+             << "  depth: " << sitch.size()
+             << endl;
+      }
     }
-    if (in_subheads){
-        string sub_contype;
+    if (/* still */in_subheads){
         string junk;
+// FIXME
 // in principle could worry about folded headers,
 // but in this application it doesn't actually matter
+
         string headword;
         string rest;
         size_t where = line.find(":");
@@ -388,17 +460,26 @@ int skrewt::body(std::istream& xin, std::ostream& xout){
         }
         headword = toLower(headword);
         if (headword == "content-type") {
-          parse_content(rest, sub_contype, junk);
-          currently_text = sub_contype == "text";
-          //xxxx cerr << "setting contype '" << sub_contype << "' " << currently_text << " ... " << textlines << endl;
+          sub_conntype = rest;
         }
     } else {
-      if (main_multipart && line == "--" + boundary) {
-        //xxxx cerr << "found subhead boundary" << endl;
+      if (sitch.front().tp == "multipart"
+           && line == "--" + sitch.front().bdy) {
+     cerr << "boundary: begin subhead block" << endl;
+        in_subheads = 1;
+        sub_conntype = "";
+        continue;
+      }
+      if (sitch.front().tp == "multipart"
+           && line == "--" + sitch.front().bdy + "--") {
+        sitch.pop_front();
+        cerr << "found subhead termination --> "
+          << sitch.front().tp
+          << endl;
         in_subheads = 1;
         continue;
       }
-      if (currently_text) textlines++;
+      if (sitch.front().tp == "text") textlines++;
     }
   }
 
diff --git a/tools/libskrewt.h b/tools/libskrewt.h
index cfbbb79..5ae6ece 100644
--- a/tools/libskrewt.h
+++ b/tools/libskrewt.h
@@ -5,6 +5,14 @@
 #include "qq_exit_codes.h"              // a bit of a kludge
 extern std::string progid;
 
+class xstr {
+public:
+  int err;
+  std::string str;
+  xstr() : err(1), str("") {}
+  xstr(const int foo, const std::string bar) : err(foo), str(bar) {}
+};
+
 class skrewt{
 public:
   std::string received_from;         // envelope HELO among other things
@@ -12,6 +20,7 @@ public:
     std::string proximta_rDNS;
     std::string proximta_IP;
     std::string proximta_AuthUser;
+    SPF_result_t spf_result;
   std::string return_path;           // envelope MAIL FROM
   std::string boundary;
   std::string to;
@@ -20,7 +29,7 @@ public:
   std::string date;
   std::string message_id;
   std::string content_type;
-  std::string delivered_to;
+  std::list<std::string> delivered_to;
   int msgsize;
   std::vector<std::string> bigbuf;
   int saw_blank_line;
@@ -32,15 +41,20 @@ public:
 
   // constructor
   skrewt()
-  : boundary("x-xx-x"), msgsize(0), saw_blank_line(0), recno(0),
-    maxsize(1000*1000), error_exit(0), mid_required(0)
+  : spf_result(SPF_RESULT_INVALID),
+    boundary("x-xx-x"), msgsize(0), saw_blank_line(0), recno(0),
+    maxsize(1000*1000), error_exit(0), mid_required(0),
+    lookahead(1, "")
   {}
 
+  xstr getRecord(std::istream& xin);
+  xstr getLine(std::istream& xin);
   int headers(std::istream& xin);
   int dump_headers(std::ostream& xout);
   int interstage();
   int body(std::istream& xin, std::ostream& xout);
   int krunch_rfrom();
+  xstr lookahead;
 };
 
 void parse_content(const std::string type_spec_line,
diff --git a/tools/makefile b/tools/makefile
index 1cb1425..97a39b4 100644
--- a/tools/makefile
+++ b/tools/makefile
@@ -15,7 +15,7 @@ qmain = pido.c hi-q.c skrewt.c hi-test.c mail-scan.c greylist.c wripper.c
 qprogs = $(qmain:%.c=%)
 
 # sources for other main programs:
-moremain = wripper.c bash-c.c ltgrey.c fixown.c
+moremain = wripper.c bash-c.c ltgrey.c fixown.c ward.c
 moreprogs = $(moremain:%.c=%)
 
 nonmain = libltgrey.c
diff --git a/tools/ward.c b/tools/ward.c
index b6f5a15..b8bdee4 100644
--- a/tools/ward.c
+++ b/tools/ward.c
@@ -85,6 +85,7 @@ int main(int _argc, const char** _argv){
   }
 
   skrewt mysk;
+//  cerr << "maxsize: " << mysk.maxsize << endl;
 
   argParser ARGS(argc, argv);
   try {while (ARGS.size()) {