From 60fd39ff24975486da7d02cdf07abae31c525529 Mon Sep 17 00:00:00 2001
From: John Denker <jsd@av8n.com>
Date: Fri, 20 Jul 2012 12:15:59 -0700
Subject: much smarter about exit status conventions

---
 tools/skrewt.c | 52 +++++++++++++++++++++++++++++++++++-----------------
 1 file changed, 35 insertions(+), 17 deletions(-)

(limited to 'tools/skrewt.c')

diff --git a/tools/skrewt.c b/tools/skrewt.c
index 2ca32a4..d0289a6 100644
--- a/tools/skrewt.c
+++ b/tools/skrewt.c
@@ -19,7 +19,7 @@ void usage(const int sts){
 "Usage: skrewt [options]\n"
 "\n"
 "  Scrutinizes email.  Reads stdin, copies it to stdout.\n"
-"  Exit result 0 means good, 1 means rejection (spam).\n"
+"  Exit result 0 means good, 21 means rejection (spam).\n"
 "  Writes reason for rejection to stderr.\n"
 "\n"
 "  Typically used as a filter in a pipeline, along with spamc -E\n"
@@ -33,10 +33,20 @@ void usage(const int sts){
   exit(sts);
 }
 
-// exit codes, compatible with spamassassin (not with qmail-queue)
-const int sa_good(0);
-const int sa_spam(1);
-const int sa_usage(64);
+//  error exit codes, mostly as stated in qmail.c
+#define bar \
+foo(good, 0) ;\
+foo(spam, 21) ;\
+foo(permerr, 31) ;\
+foo(usage, 39) ;\
+foo(greylisting, 70) ;\
+foo(syserr, 71) ;\
+foo(comerr, 74) ;
+
+#define foo(name, num) const int ex_ ## name = num
+bar
+#undef foo
+
 
 /////////////////////////////////////////////////////////
 // Case insensitive comparison of strings
@@ -106,7 +116,7 @@ int prefix(const string shorter, const string longer){
 }
 
 void exeunt(const int sts){
-  if (sts == sa_good) exit(sts);
+  if (sts == ex_good) exit(sts);
 
   const char* foo = getenv("HI_Q_GROUP");
   if (!foo) exit(sts);
@@ -144,18 +154,18 @@ int main(int _argc, const char** _argv){
     if (prefix(arg, "-maxsize")) {
       if (!argc) {
         cerr << "Option -maxsize requires an argument" << endl;
-        exit(sa_usage);
+        exit(ex_usage);
       }
       maxsize = atoi(*argv); argv++; argc--;
     }
     if (arg.substr(0,1) == "-") {
       cerr << "Unrecognized option '" << arg << "'" << endl;
       cerr << "For help, try:  " << progname << " -help" << endl;
-      exit(sa_usage);
+      exit(ex_usage);
     } else {
       cerr << "Extraneous verbiage '" << arg << "'" << endl;
       cerr << "For help, try:  " << progname << " -help" << endl;
-      exit(sa_usage);
+      exit(ex_usage);
     }
   }
 
@@ -172,6 +182,10 @@ int main(int _argc, const char** _argv){
       string header;
       if (getline(cin, header).fail()) continue;
       msgsize += header.length()+1;
+      if (msgsize > maxsize) {
+        cerr << "skrewt rejection: bigger than " << maxsize << endl;
+        exeunt(ex_spam);
+      }
       for (;;) {
         if (cin.eof()) break;
         if (cin.bad()) return 1;
@@ -184,18 +198,22 @@ int main(int _argc, const char** _argv){
         msgsize += line.length()+1;
         if (msgsize > maxsize) {
           cerr << "skrewt rejection: bigger than " << maxsize << endl;
-          exeunt(sa_spam);
+          exeunt(ex_spam);
         }
         header += "\n" + line;
       }
-      if (header.length() == 0) {
+      int len = header.length();
+      if (len && header[len-1] == '\r') len--;  // reduced length, not counting <cr>
+      if (len == 0) {
         if (!gotdate) {
           cerr << "skrewt rejection: no date" << endl;
-          exeunt(sa_spam);              // disallow mail with no date
+          exeunt(ex_spam);              // disallow mail with no date
         }
         inheads = 0;
+        //cerr << "end of headers" << endl;
       }
       else {
+// here if it's a header line
         string headword;
         string rest;
         size_t where = header.find(":");
@@ -239,18 +257,18 @@ int main(int _argc, const char** _argv){
         } else if (headword == "subject") {
           if (rest.find("-please-bounce-this-") != string::npos) {
             cerr << "skrewt rejection: by request" << endl;
-            exeunt(sa_spam);
+            exeunt(ex_spam);
           }
         }
       }
-      cout << header << endl;
+      //cout << header.length() << " ... " << header << endl;
     } else {
       string line;
       if (!getline(cin, line).fail()) {
         msgsize += line.length()+1;
         if (msgsize > maxsize) {
           cerr << "skrewt rejection: bigger than " << maxsize << endl;
-          exeunt(sa_spam);
+          exeunt(ex_spam);
         }
         if (line == "--" + boundary) {
           inheads = 1;
@@ -266,8 +284,8 @@ int main(int _argc, const char** _argv){
   if (0) cerr << "textlines: " << textlines << endl;
   if (!textlines) {
     cerr << "skrewt rejection: no text" << endl;
-    exeunt(sa_spam);
+    exeunt(ex_spam);
   }
   cerr << "skrewt normal completion" << endl;
-  exit(sa_good);
+  exit(ex_good);
 }
-- 
cgit v1.2.3


From 51f3d88572ae0f8eea40996db28ece9cdd0dae18 Mon Sep 17 00:00:00 2001
From: John Denker <jsd@av8n.com>
Date: Fri, 20 Jul 2012 13:59:53 -0700
Subject: gaaack -- fix  horrible bug

---
 tools/skrewt.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

(limited to 'tools/skrewt.c')

diff --git a/tools/skrewt.c b/tools/skrewt.c
index d0289a6..44e885b 100644
--- a/tools/skrewt.c
+++ b/tools/skrewt.c
@@ -261,7 +261,8 @@ int main(int _argc, const char** _argv){
           }
         }
       }
-      //cout << header.length() << " ... " << header << endl;
+      //xxxx  cout << header.length() << " ... ";
+      cout << header << endl;
     } else {
       string line;
       if (!getline(cin, line).fail()) {
-- 
cgit v1.2.3


From 6c56d656bd407bc5fc4dd713aacb553ebefee892 Mon Sep 17 00:00:00 2001
From: John Denker <jsd@av8n.com>
Date: Sun, 22 Jul 2012 14:44:49 -0700
Subject: regularize log/progress messages

---
 tools/skrewt.c | 33 +++++++++++++++++++++++++--------
 1 file changed, 25 insertions(+), 8 deletions(-)

(limited to 'tools/skrewt.c')

diff --git a/tools/skrewt.c b/tools/skrewt.c
index 44e885b..6de3dd9 100644
--- a/tools/skrewt.c
+++ b/tools/skrewt.c
@@ -11,6 +11,7 @@
 #include <signal.h>
 
 #include <stdio.h>              /* perror */
+#include <sstream>
 
 using namespace std;
 
@@ -134,6 +135,15 @@ void exeunt(const int sts){
   exit(sts);
 }
 
+string basename(const string path){
+  size_t where = path.rfind("/");
+  if (where != string::npos) return path.substr(1+where);
+  return path;
+}
+
+string progname, progid;
+int mypid;
+
 ////////////////////////////////////////////////////////////
 int main(int _argc, const char** _argv){
 ////  pid_t pid = getpid();
@@ -141,7 +151,14 @@ int main(int _argc, const char** _argv){
 ////  cout << getpgid(pid) << endl;
   int argc(_argc);
   const char **argv(_argv);
-  string progname(*argv); argv++; argc--;
+  {
+    progname = *argv++; argc--;
+    mypid = getpid();
+    stringstream binder;
+    binder << basename(progname) << "[" << mypid << "]";
+    progid = binder.str();
+  }
+
 
   int maxsize(1000000);
 
@@ -183,7 +200,7 @@ int main(int _argc, const char** _argv){
       if (getline(cin, header).fail()) continue;
       msgsize += header.length()+1;
       if (msgsize > maxsize) {
-        cerr << "skrewt rejection: bigger than " << maxsize << endl;
+        cerr << progid << " rejection: bigger than " << maxsize << endl;
         exeunt(ex_spam);
       }
       for (;;) {
@@ -197,7 +214,7 @@ int main(int _argc, const char** _argv){
         if (getline(cin, line).fail()) continue;
         msgsize += line.length()+1;
         if (msgsize > maxsize) {
-          cerr << "skrewt rejection: bigger than " << maxsize << endl;
+          cerr << progid << " rejection: bigger than " << maxsize << endl;
           exeunt(ex_spam);
         }
         header += "\n" + line;
@@ -206,7 +223,7 @@ int main(int _argc, const char** _argv){
       if (len && header[len-1] == '\r') len--;  // reduced length, not counting <cr>
       if (len == 0) {
         if (!gotdate) {
-          cerr << "skrewt rejection: no date" << endl;
+          cerr << progid << " rejection: no date" << endl;
           exeunt(ex_spam);              // disallow mail with no date
         }
         inheads = 0;
@@ -256,7 +273,7 @@ int main(int _argc, const char** _argv){
           gotdate++;
         } else if (headword == "subject") {
           if (rest.find("-please-bounce-this-") != string::npos) {
-            cerr << "skrewt rejection: by request" << endl;
+            cerr << progid << " rejection: by request" << endl;
             exeunt(ex_spam);
           }
         }
@@ -268,7 +285,7 @@ int main(int _argc, const char** _argv){
       if (!getline(cin, line).fail()) {
         msgsize += line.length()+1;
         if (msgsize > maxsize) {
-          cerr << "skrewt rejection: bigger than " << maxsize << endl;
+          cerr << progid << " rejection: bigger than " << maxsize << endl;
           exeunt(ex_spam);
         }
         if (line == "--" + boundary) {
@@ -284,9 +301,9 @@ int main(int _argc, const char** _argv){
   }
   if (0) cerr << "textlines: " << textlines << endl;
   if (!textlines) {
-    cerr << "skrewt rejection: no text" << endl;
+    cerr << progid << " rejection: no text" << endl;
     exeunt(ex_spam);
   }
-  cerr << "skrewt normal completion" << endl;
+  cerr << progid << " normal completion" << endl;
   exit(ex_good);
 }
-- 
cgit v1.2.3


From 9b5fbc11bb1d96dd598ebfad5539660f75571835 Mon Sep 17 00:00:00 2001
From: John Denker <jsd@av8n.com>
Date: Mon, 23 Jul 2012 18:01:34 -0700
Subject: progress toward cleaning up skrewt

---
 tools/skrewt.c | 265 +++++++++++++++++++++++++++++++++++----------------------
 1 file changed, 163 insertions(+), 102 deletions(-)

(limited to 'tools/skrewt.c')

diff --git a/tools/skrewt.c b/tools/skrewt.c
index 6de3dd9..ed0e627 100644
--- a/tools/skrewt.c
+++ b/tools/skrewt.c
@@ -12,6 +12,7 @@
 
 #include <stdio.h>              /* perror */
 #include <sstream>
+#include <vector>
 
 using namespace std;
 
@@ -144,6 +145,58 @@ string basename(const string path){
 string progname, progid;
 int mypid;
 
+
+/* Content-Type: text/plain; charset="us-ascii"                                         */
+/* Content-Type: multipart/mixed; boundary="1170861315-1262462055-1341954763=:92165"    */
+void parse_content(const string type_spec_line, string &maintype, string &boundary) {
+      string mainline(type_spec_line);
+
+      string get_type(toLower(mainline));
+      size_t where = get_type.find_first_of(" \t;\n");
+      if (where == string::npos) {
+        // keep whole string
+      }
+      else {
+        get_type = get_type.substr(0,where);
+      }
+      where = get_type.find("/");
+      if (where == string::npos){
+        maintype = "";
+      } else {
+        maintype = get_type.substr(0, where);
+      }
+}
+
+#ifdef xxxxxxxxxxxxxxxxxxxx
+      if
+      if (0) cerr << "type (" << get_type << ")  "
+            << (text_type?"text":"nope") << endl;
+
+
+      string srch = "boundary=";
+      where = headrec.find(srch);
+      if (where != string::npos) {
+        where += srch.length();
+        boundary = headrec.substr(where);
+        if (boundary[0] == '"') {
+          boundary = boundary.substr(1);
+          where = boundary.find_first_of("\"");
+        } else {
+          where = boundary.find_first_of(" \t;\n");
+        }
+        if (where == string::npos) {
+          /* do nothing, boundary=boundary as a whole */
+        } else {
+          boundary = boundary.substr(0, where);
+        }
+      }
+    }
+#endif
+
+
+
+
+
 ////////////////////////////////////////////////////////////
 int main(int _argc, const char** _argv){
 ////  pid_t pid = getpid();
@@ -159,8 +212,7 @@ int main(int _argc, const char** _argv){
     progid = binder.str();
   }
 
-
-  int maxsize(1000000);
+  int maxsize(1000*1000);
 
   while (argc) {
     string arg(*argv); argv++; argc--;
@@ -186,121 +238,130 @@ int main(int _argc, const char** _argv){
     }
   }
 
-  int inheads(1);
+  int saw_blank_line(0);
   string boundary("x-xx-x");
-  int text_type(1);
-  int textlines(0);
-  int gotdate(0);
+  string date;
+  string subject;
+  string content_type;
+  string message_id;
   int msgsize(0);
-  for (;;){
+  vector<string> bigbuf;
+  cerr <<  "hi there" << endl;
+
+  for (;;){             // outer loop over all records in the header
     if (cin.eof()) break;
     if (cin.bad()) return 1;
-    if (inheads) {
-      string header;
-      if (getline(cin, header).fail()) continue;
-      msgsize += header.length()+1;
+
+    string headrec;
+// on fail, go back to top of outer loop and check for eof versus bad
+    if (getline(cin, headrec).fail()) continue;
+    msgsize += headrec.length()+1;
+    if (msgsize > maxsize) {
+      cerr << progid << " rejection: bigger than " << maxsize << endl;
+      exeunt(ex_spam);
+    }
+    cout << headrec << endl;
+    bigbuf.push_back(headrec);  // for a folded record, this is the first line
+
+    for (;;) {        // inner loop to build a multi-line record e.g. folded record:
+      if (cin.eof()) break;
+      if (cin.bad()) return 1;
+      char ch;
+      if (cin.get(ch).fail()) continue;
+      cin.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(cin, line).fail()) continue;
+      msgsize += line.length()+1;
       if (msgsize > maxsize) {
         cerr << progid << " rejection: bigger than " << maxsize << endl;
         exeunt(ex_spam);
       }
-      for (;;) {
-        if (cin.eof()) break;
-        if (cin.bad()) return 1;
-        char ch;
-        if (cin.get(ch).fail()) continue;
-        cin.putback(ch);
-        if (ch != ' ' && ch != '\t') break;
-        string line;
-        if (getline(cin, line).fail()) continue;
-        msgsize += line.length()+1;
-        if (msgsize > maxsize) {
-          cerr << progid << " rejection: bigger than " << maxsize << endl;
-          exeunt(ex_spam);
-        }
-        header += "\n" + line;
-      }
-      int len = header.length();
-      if (len && header[len-1] == '\r') len--;  // reduced length, not counting <cr>
-      if (len == 0) {
-        if (!gotdate) {
-          cerr << progid << " rejection: no date" << endl;
-          exeunt(ex_spam);              // disallow mail with no date
-        }
-        inheads = 0;
-        //cerr << "end of headers" << endl;
+      cout << line << endl;
+      bigbuf.push_back(line);
+      string cooked(line);
+      if (cooked.length()){
+        string::iterator ptr = cooked.end()-1;
+        if (*ptr == '\r') cooked.erase(ptr);
       }
-      else {
+      headrec += "\n" + cooked;
+    }
+// here with a fully assembled header record
+    int len = headrec.length();
+    if (len && headrec[len-1] == '\r') len--;  // reduced length, not counting <cr>
+    if (len == 0) {
+      saw_blank_line = 1;
+      break;            // no more headers in this message
+    }
+
 // here if it's a header line
-        string headword;
-        string rest;
-        size_t where = header.find(":");
-        if (where != string::npos) {
-          headword = header.substr(0, where);
-          rest = ltrim(header.substr(1+where));
-        }
-        headword = toLower(headword);
-        if (headword == "content-type") {
-          string the_type = rest;
-          size_t where = the_type.find_first_of(" \t;\n");
-          if (where == string::npos) {
-            /* do nothing */
-          }
-          else {
-            the_type = the_type.substr(0,where);
-          }
-          the_type = toLower(the_type);
-          text_type = (the_type.find("text/") == 0);
-          if (0) cerr << "type (" << the_type << ")  "
-                << (text_type?"text":"nope") << endl;
-          string srch = "boundary=";
-          where = header.find(srch);
-          if (where != string::npos) {
-            where += srch.length();
-            boundary = header.substr(where);
-            if (boundary[0] == '"') {
-              boundary = boundary.substr(1);
-              where = boundary.find_first_of("\"");
-            } else {
-              where = boundary.find_first_of(" \t;\n");
-            }
-            if (where == string::npos) {
-              /* do nothing, boundary=boundary as a whole */
-            } else {
-              boundary = boundary.substr(0, where);
-            }
-          }
-        } else if (headword == "date") {
-          gotdate++;
-        } else if (headword == "subject") {
-          if (rest.find("-please-bounce-this-") != string::npos) {
-            cerr << progid << " rejection: by request" << endl;
-            exeunt(ex_spam);
-          }
-        }
-      }
-      //xxxx  cout << header.length() << " ... ";
-      cout << header << endl;
+    string headword;
+    string rest;
+    size_t where = headrec.find(":");
+    if (where != string::npos) {
+      headword = headrec.substr(0, where);
+      rest = ltrim(headrec.substr(1+where));
+    }
+    headword = toLower(headword);
+    if (0){
+    } else if (headword == "date") {
+      date = rest;
+    } else if (headword == "subject") {
+      subject = rest;
+    } else if (headword == "content-type") {
+      content_type = rest;
+    }
+    //xxxx  cout << headrec.length() << " ... ";
+  }
+  cerr << "headers are done. Delimited: " << saw_blank_line << endl;
+
+// Headers are done.
+// Do some early-stage thinking.
+
+  if (subject.find("-please-bounce-this-") != string::npos) {
+    cerr << progid << " rejection: by request" << endl;
+    exeunt(ex_spam);
+  }
+
+  if (!date.length()) {
+    cerr << progid << " rejection: no date" << endl;
+    exeunt(ex_spam);              // disallow mail with no date
+  }
+
+  string contype;
+  int textlines(0);
+  parse_content(content_type, contype, boundary);
+  int is_text = contype == "text";
+
+// early-stage thinking has been done.
+// Now spew the rest of the message
+  cerr << "body begins: " << contype << " " << is_text << endl;
+  int inheads(0);
+  for (;;){             // outer loop over all lines in the body
+    if (cin.eof()) break;
+    if (cin.bad()) return 1;
+    string line;
+// on fail, go back to top of outer loop and check for eof versus bad
+    if (getline(cin, line).fail()) continue;
+    msgsize += line.length()+1;
+    if (msgsize > maxsize) {
+      cerr << progid << " rejection: bigger than " << maxsize << endl;
+      exeunt(ex_spam);
+    }
+    bigbuf.push_back(line);
+    cout << line << endl;
+    if (line == "--" + boundary) {
+      inheads = 1;
     } else {
-      string line;
-      if (!getline(cin, line).fail()) {
-        msgsize += line.length()+1;
-        if (msgsize > maxsize) {
-          cerr << progid << " rejection: bigger than " << maxsize << endl;
-          exeunt(ex_spam);
-        }
-        if (line == "--" + boundary) {
-          inheads = 1;
-        } else {
-          if (text_type) {
-            if (ltrim(line).length()) textlines++;
-          }
-        }
-        cout << line << endl;
+      if (is_text) {
+        if (ltrim(line).length()) textlines++;
       }
     }
   }
+
   if (0) cerr << "textlines: " << textlines << endl;
-  if (!textlines) {
+  if (000 && !textlines) {
     cerr << progid << " rejection: no text" << endl;
     exeunt(ex_spam);
   }
-- 
cgit v1.2.3


From 0f0fc6dbb7e7354ac662951f124b0a64c54b741b Mon Sep 17 00:00:00 2001
From: John Denker <jsd@av8n.com>
Date: Mon, 23 Jul 2012 21:37:35 -0700
Subject: get line-counter to work

---
 tools/skrewt.c | 134 +++++++++++++++++++++++++++++++++------------------------
 1 file changed, 77 insertions(+), 57 deletions(-)

(limited to 'tools/skrewt.c')

diff --git a/tools/skrewt.c b/tools/skrewt.c
index ed0e627..d2e1bbc 100644
--- a/tools/skrewt.c
+++ b/tools/skrewt.c
@@ -145,56 +145,52 @@ string basename(const string path){
 string progname, progid;
 int mypid;
 
-
 /* Content-Type: text/plain; charset="us-ascii"                                         */
 /* Content-Type: multipart/mixed; boundary="1170861315-1262462055-1341954763=:92165"    */
+//
 void parse_content(const string type_spec_line, string &maintype, string &boundary) {
-      string mainline(type_spec_line);
-
-      string get_type(toLower(mainline));
-      size_t where = get_type.find_first_of(" \t;\n");
-      if (where == string::npos) {
-        // keep whole string
-      }
-      else {
-        get_type = get_type.substr(0,where);
-      }
-      where = get_type.find("/");
-      if (where == string::npos){
-        maintype = "";
-      } else {
-        maintype = get_type.substr(0, where);
-      }
-}
-
-#ifdef xxxxxxxxxxxxxxxxxxxx
-      if
-      if (0) cerr << "type (" << get_type << ")  "
-            << (text_type?"text":"nope") << endl;
-
-
-      string srch = "boundary=";
-      where = headrec.find(srch);
-      if (where != string::npos) {
-        where += srch.length();
-        boundary = headrec.substr(where);
-        if (boundary[0] == '"') {
-          boundary = boundary.substr(1);
-          where = boundary.find_first_of("\"");
-        } else {
-          where = boundary.find_first_of(" \t;\n");
-        }
-        if (where == string::npos) {
-          /* do nothing, boundary=boundary as a whole */
-        } else {
-          boundary = boundary.substr(0, where);
-        }
-      }
-    }
-#endif
+  cerr << "parser called with: " << type_spec_line << endl;
+  string get_type(type_spec_line);
 
+  size_t where = get_type.find_first_of(" \t;\n");
+  string rest;
+  if (where == string::npos) {
+    // keep whole string
+  }
+  else {
+    rest = get_type.substr(where+1);
+    get_type = get_type.substr(0,where);
+  }
+  where = get_type.find("/");
+  if (where == string::npos){
+    maintype = "";
+    cerr << "could not find / in " << get_type << endl;
+  } else {
+    maintype = get_type.substr(0, where);
+  }
 
+// now need to find boundary
 
+  string srch = "boundary=";
+  where = rest.find(srch);
+  if (where != string::npos) {
+    where += srch.length();
+    boundary = rest.substr(where);
+    if (boundary[0] == '"') {
+      boundary = boundary.substr(1);
+      where = boundary.find_first_of("\"");
+    } else {
+      where = boundary.find_first_of(" \t;\n");
+    }
+    if (where == string::npos) {
+      /* do nothing, boundary=boundary as a whole */
+    } else {
+      boundary = boundary.substr(0, where);
+    }
+  } else {
+    //xxxxxxx cerr << "boundary= not found in " << type_spec_line << endl;
+  }
+}
 
 
 ////////////////////////////////////////////////////////////
@@ -329,15 +325,17 @@ int main(int _argc, const char** _argv){
     exeunt(ex_spam);              // disallow mail with no date
   }
 
-  string contype;
-  int textlines(0);
-  parse_content(content_type, contype, boundary);
-  int is_text = contype == "text";
+  string main_contype;
+  parse_content(content_type, main_contype, boundary);
+  int currently_text = main_contype == "text";
+  int main_multipart = main_contype == "multipart";
 
 // early-stage thinking has been done.
 // Now spew the rest of the message
-  cerr << "body begins: " << contype << " " << is_text << endl;
-  int inheads(0);
+  cerr << "body begins: " << main_contype << " " << currently_text << " " << boundary << endl;
+  int in_subheads(0);
+  int textlines(0);
+
   for (;;){             // outer loop over all lines in the body
     if (cin.eof()) break;
     if (cin.bad()) return 1;
@@ -351,19 +349,41 @@ int main(int _argc, const char** _argv){
     }
     bigbuf.push_back(line);
     cout << line << endl;
-    if (line == "--" + boundary) {
-      inheads = 1;
+    if (in_subheads){
+      if (line == "" || line == "\r") in_subheads = 0;
+    }
+    if (in_subheads){
+        string sub_contype;
+        string junk;
+// 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(":");
+        if (where != string::npos) {
+          headword = line.substr(0, where);
+          rest = ltrim(line.substr(1+where));
+        }
+        headword = toLower(headword);
+        if (headword == "content-type") {
+          parse_content(rest, sub_contype, junk);
+          currently_text = sub_contype == "text";
+          cerr << "setting contype '" << sub_contype << "' " << currently_text << " ... " << textlines << endl;
+        }
     } else {
-      if (is_text) {
-        if (ltrim(line).length()) textlines++;
+      if (main_multipart && line == "--" + boundary) {
+        //xxxx cerr << "found subhead boundary" << endl;
+        in_subheads = 1;
+        continue;
       }
+      if (currently_text) textlines++;
     }
   }
 
-  if (0) cerr << "textlines: " << textlines << endl;
-  if (000 && !textlines) {
+  if (1) cerr << "textlines: " << textlines << endl;
+  if (1 && !textlines) {
     cerr << progid << " rejection: no text" << endl;
-    exeunt(ex_spam);
+//    exeunt(ex_spam);
   }
   cerr << progid << " normal completion" << endl;
   exit(ex_good);
-- 
cgit v1.2.3


From e2390efc92d66a50c6ce13d027aac8c6ecc02e89 Mon Sep 17 00:00:00 2001
From: John Denker <jsd@av8n.com>
Date: Tue, 24 Jul 2012 11:28:34 -0700
Subject: remove print statements

---
 tools/skrewt.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

(limited to 'tools/skrewt.c')

diff --git a/tools/skrewt.c b/tools/skrewt.c
index d2e1bbc..ad99e67 100644
--- a/tools/skrewt.c
+++ b/tools/skrewt.c
@@ -149,7 +149,7 @@ int mypid;
 /* Content-Type: multipart/mixed; boundary="1170861315-1262462055-1341954763=:92165"    */
 //
 void parse_content(const string type_spec_line, string &maintype, string &boundary) {
-  cerr << "parser called with: " << type_spec_line << endl;
+  //xxx cerr << "parser called with: " << type_spec_line << endl;
   string get_type(type_spec_line);
 
   size_t where = get_type.find_first_of(" \t;\n");
@@ -310,7 +310,7 @@ int main(int _argc, const char** _argv){
     }
     //xxxx  cout << headrec.length() << " ... ";
   }
-  cerr << "headers are done. Delimited: " << saw_blank_line << endl;
+  //xxx cerr << "headers are done. Delimited: " << saw_blank_line << endl;
 
 // Headers are done.
 // Do some early-stage thinking.
@@ -368,7 +368,7 @@ int main(int _argc, const char** _argv){
         if (headword == "content-type") {
           parse_content(rest, sub_contype, junk);
           currently_text = sub_contype == "text";
-          cerr << "setting contype '" << sub_contype << "' " << currently_text << " ... " << textlines << endl;
+          //xxxx cerr << "setting contype '" << sub_contype << "' " << currently_text << " ... " << textlines << endl;
         }
     } else {
       if (main_multipart && line == "--" + boundary) {
-- 
cgit v1.2.3


From 6e8083ff4ffe3fd2b6d337386637a2b5c1378cf7 Mon Sep 17 00:00:00 2001
From: John Denker <jsd@av8n.com>
Date: Wed, 25 Jul 2012 13:05:03 -0700
Subject: fix a bunch of DOS-CR bugs

---
 tools/skrewt.c | 121 ++++++++++++++++++++++++++++++++++++++++++---------------
 1 file changed, 89 insertions(+), 32 deletions(-)

(limited to 'tools/skrewt.c')

diff --git a/tools/skrewt.c b/tools/skrewt.c
index ad99e67..a43fd13 100644
--- a/tools/skrewt.c
+++ b/tools/skrewt.c
@@ -1,4 +1,4 @@
-///////////////////
+//////////////////
 // skrewt.c
 //
 // scrutinize email
@@ -13,6 +13,7 @@
 #include <stdio.h>              /* perror */
 #include <sstream>
 #include <vector>
+#include <list>
 
 using namespace std;
 
@@ -28,6 +29,7 @@ void usage(const int sts){
 "  Options\n"
 "    -help              print this msg (and exit immediately).\n"
 "    -maxsize ii        msg size in bytes; anything bigger will be rejected.\n"
+"    -error-exit        exit early if errors have been detected.\n"
 "\n"
 "  Messages containing the string '-please-bounce-this-' will be rejected.\n"
 "  Messages with no date will be rejected.\n"
@@ -36,7 +38,7 @@ void usage(const int sts){
 }
 
 //  error exit codes, mostly as stated in qmail.c
-#define bar \
+#define ErrorCodes \
 foo(good, 0) ;\
 foo(spam, 21) ;\
 foo(permerr, 31) ;\
@@ -46,7 +48,7 @@ foo(syserr, 71) ;\
 foo(comerr, 74) ;
 
 #define foo(name, num) const int ex_ ## name = num
-bar
+ErrorCodes
 #undef foo
 
 
@@ -104,12 +106,23 @@ string toLower(const std::string& a){
 }
 
 ////////////////
-string ltrim(string foo){
+string ltrim(const string foo){
   size_t where = foo.find_first_not_of(" \t\r\n");
   if (where == foo.npos) return foo;
   return foo.substr(where);
 }
 
+string noCR(const string bar){
+  string foo(bar);
+  int len = foo.length();
+  if (len){
+    if (foo[len-1] == '\r') {
+      foo.erase(len-1);
+    }
+  }
+  return foo;
+}
+
 ////////////////
 // little utility to help with argument parsing:
 //
@@ -117,7 +130,8 @@ int prefix(const string shorter, const string longer){
   return shorter == longer.substr(0, shorter.length());
 }
 
-void exeunt(const int sts){
+void maybe_exeunt(const int sts, const int really){
+  if (!really) return;
   if (sts == ex_good) exit(sts);
 
   const char* foo = getenv("HI_Q_GROUP");
@@ -136,6 +150,10 @@ void exeunt(const int sts){
   exit(sts);
 }
 
+void exeunt(const int sts){
+  maybe_exeunt(sts, 1);
+}
+
 string basename(const string path){
   size_t where = path.rfind("/");
   if (where != string::npos) return path.substr(1+where);
@@ -192,6 +210,15 @@ void parse_content(const string type_spec_line, string &maintype, string &bounda
   }
 }
 
+string join(const string sep, const list<string> stuff){
+  string rslt;
+  for (list<string>::const_iterator ptr = stuff.begin();
+        ptr != stuff.end(); ptr++){
+    if (rslt.length()) rslt += sep;
+    rslt += *ptr;
+  }
+  return rslt;
+}
 
 ////////////////////////////////////////////////////////////
 int main(int _argc, const char** _argv){
@@ -209,6 +236,8 @@ int main(int _argc, const char** _argv){
   }
 
   int maxsize(1000*1000);
+  int error_exit(0);
+  int mid_required(0);
 
   while (argc) {
     string arg(*argv); argv++; argc--;
@@ -216,14 +245,18 @@ int main(int _argc, const char** _argv){
     if (prefix(arg, "-help")) {
       usage(0);
     }
-    if (prefix(arg, "-maxsize")) {
+    if (0) {
+    } else if (prefix(arg, "-mid-required")) {
+      mid_required++;
+    } else if (prefix(arg, "-error-exit")) {
+      error_exit++;
+    } else if (prefix(arg, "-maxsize")) {
       if (!argc) {
         cerr << "Option -maxsize requires an argument" << endl;
         exit(ex_usage);
       }
       maxsize = atoi(*argv); argv++; argc--;
-    }
-    if (arg.substr(0,1) == "-") {
+    } else if (arg.substr(0,1) == "-") {
       cerr << "Unrecognized option '" << arg << "'" << endl;
       cerr << "For help, try:  " << progname << " -help" << endl;
       exit(ex_usage);
@@ -236,28 +269,32 @@ int main(int _argc, const char** _argv){
 
   int saw_blank_line(0);
   string boundary("x-xx-x");
-  string date;
+  string to;
+  string from;
   string subject;
-  string content_type;
+  string date;
   string message_id;
+  string content_type;
   int msgsize(0);
   vector<string> bigbuf;
-  cerr <<  "hi there" << endl;
+  int recno(0);
 
+  //xxxx cerr << progid << " begins" << endl;
   for (;;){             // outer loop over all records in the header
     if (cin.eof()) break;
     if (cin.bad()) return 1;
 
-    string headrec;
+    string line;
 // on fail, go back to top of outer loop and check for eof versus bad
-    if (getline(cin, headrec).fail()) continue;
-    msgsize += headrec.length()+1;
+    if (getline(cin, line).fail()) continue;
+    msgsize += line.length()+1;
     if (msgsize > maxsize) {
       cerr << progid << " rejection: bigger than " << maxsize << endl;
       exeunt(ex_spam);
     }
-    cout << headrec << endl;
-    bigbuf.push_back(headrec);  // for a folded record, this is the first line
+    cout << line << endl;
+    bigbuf.push_back(line);
+    string headrec = noCR(line);       // for a folded record, this is the first line
 
     for (;;) {        // inner loop to build a multi-line record e.g. folded record:
       if (cin.eof()) break;
@@ -276,12 +313,7 @@ int main(int _argc, const char** _argv){
       }
       cout << line << endl;
       bigbuf.push_back(line);
-      string cooked(line);
-      if (cooked.length()){
-        string::iterator ptr = cooked.end()-1;
-        if (*ptr == '\r') cooked.erase(ptr);
-      }
-      headrec += "\n" + cooked;
+      headrec += "\n" + noCR(line);
     }
 // here with a fully assembled header record
     int len = headrec.length();
@@ -301,6 +333,12 @@ int main(int _argc, const char** _argv){
     }
     headword = toLower(headword);
     if (0){
+    } else if (headword == "from") {
+      from = rest;
+    } else if (headword == "to") {
+      to = rest;
+    } else if (headword == "message-id") {
+      message_id = rest;
     } else if (headword == "date") {
       date = rest;
     } else if (headword == "subject") {
@@ -309,30 +347,49 @@ int main(int _argc, const char** _argv){
       content_type = rest;
     }
     //xxxx  cout << headrec.length() << " ... ";
+    recno++;
+    if (0) if (recno <= 6) cerr << progid << "#" << recno
+        << " " << headrec << endl;
   }
-  //xxx cerr << "headers are done. Delimited: " << saw_blank_line << endl;
+  cerr << progid <<" Mid '" << message_id << "'" << endl;
 
 // Headers are done.
 // Do some early-stage thinking.
 
+  list<string> badnews;
+
   if (subject.find("-please-bounce-this-") != string::npos) {
-    cerr << progid << " rejection: by request" << endl;
-    exeunt(ex_spam);
+    badnews.push_back("by request");
   }
 
   if (!date.length()) {
-    cerr << progid << " rejection: no date" << endl;
-    exeunt(ex_spam);              // disallow mail with no date
+    badnews.push_back("no date");
+  }
+
+  if (mid_required && !message_id.length()) {
+    badnews.push_back("no message-id");
+  }
+
+  if (badnews.size()){
+    cerr << progid << " " << join(", ", badnews) << endl;
+    if (error_exit){
+      cerr << progid << " '" << from
+             << "' to '" << to
+             << "'" << endl;
+      exeunt(ex_spam);
+    }
   }
 
   string main_contype;
-  parse_content(content_type, main_contype, boundary);
+  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";
 
 // early-stage thinking has been done.
 // Now spew the rest of the message
-  cerr << "body begins: " << main_contype << " " << currently_text << " " << boundary << endl;
+  //xxxx cerr << "body begins: " << main_contype << " " << currently_text << " " << boundary << endl;
   int in_subheads(0);
   int textlines(0);
 
@@ -345,7 +402,7 @@ int main(int _argc, const char** _argv){
     msgsize += line.length()+1;
     if (msgsize > maxsize) {
       cerr << progid << " rejection: bigger than " << maxsize << endl;
-      exeunt(ex_spam);
+      maybe_exeunt(ex_spam, error_exit);
     }
     bigbuf.push_back(line);
     cout << line << endl;
@@ -380,10 +437,10 @@ int main(int _argc, const char** _argv){
     }
   }
 
-  if (1) cerr << "textlines: " << textlines << endl;
+  if (0) cerr << "textlines: " << textlines << endl;
   if (1 && !textlines) {
     cerr << progid << " rejection: no text" << endl;
-//    exeunt(ex_spam);
+//    maybe_exeunt(ex_spam, error_exit);
   }
   cerr << progid << " normal completion" << endl;
   exit(ex_good);
-- 
cgit v1.2.3


From d2564d25e802d1ee3230cf045c4940e836b5c6a2 Mon Sep 17 00:00:00 2001
From: John Denker <jsd@av8n.com>
Date: Sun, 29 Jul 2012 16:50:11 -0700
Subject: split ltgrey (and libltgrey) off from greylist; put some utility
 functions into their own file.

---
 tools/skrewt.c | 17 +++--------------
 1 file changed, 3 insertions(+), 14 deletions(-)

(limited to 'tools/skrewt.c')

diff --git a/tools/skrewt.c b/tools/skrewt.c
index a43fd13..3fee644 100644
--- a/tools/skrewt.c
+++ b/tools/skrewt.c
@@ -37,19 +37,7 @@ void usage(const int sts){
   exit(sts);
 }
 
-//  error exit codes, mostly as stated in qmail.c
-#define ErrorCodes \
-foo(good, 0) ;\
-foo(spam, 21) ;\
-foo(permerr, 31) ;\
-foo(usage, 39) ;\
-foo(greylisting, 70) ;\
-foo(syserr, 71) ;\
-foo(comerr, 74) ;
-
-#define foo(name, num) const int ex_ ## name = num
-ErrorCodes
-#undef foo
+#include "qq_exit_codes.h"
 
 
 /////////////////////////////////////////////////////////
@@ -316,8 +304,8 @@ int main(int _argc, const char** _argv){
       headrec += "\n" + noCR(line);
     }
 // here with a fully assembled header record
+// headrec (unlike line) contains no DOS CR characters
     int len = headrec.length();
-    if (len && headrec[len-1] == '\r') len--;  // reduced length, not counting <cr>
     if (len == 0) {
       saw_blank_line = 1;
       break;            // no more headers in this message
@@ -351,6 +339,7 @@ int main(int _argc, const char** _argv){
     if (0) if (recno <= 6) cerr << progid << "#" << recno
         << " " << headrec << endl;
   }
+  if (saw_blank_line) {/* ignore */}
   cerr << progid <<" Mid '" << message_id << "'" << endl;
 
 // Headers are done.
-- 
cgit v1.2.3


From 63be414b62f3234ad80607b95e8e71e33bfd8025 Mon Sep 17 00:00:00 2001
From: John Denker <jsd@av8n.com>
Date: Sun, 29 Jul 2012 17:00:15 -0700
Subject: move more stuff to utils.c ... I hate duplication of code

---
 tools/skrewt.c | 32 +-------------------------------
 1 file changed, 1 insertion(+), 31 deletions(-)

(limited to 'tools/skrewt.c')

diff --git a/tools/skrewt.c b/tools/skrewt.c
index 3fee644..6749a01 100644
--- a/tools/skrewt.c
+++ b/tools/skrewt.c
@@ -38,7 +38,7 @@ void usage(const int sts){
 }
 
 #include "qq_exit_codes.h"
-
+#include "utils.h"
 
 /////////////////////////////////////////////////////////
 // Case insensitive comparison of strings
@@ -83,23 +83,6 @@ int cmp_casefold(const std::string& a, const std::string& b) {
   return 0;
 }
 
-
-string toLower(const std::string& a){
-  string rslt = a;
-  string::iterator rr;
-  for (rr = rslt.begin(); rr != rslt.end(); rr++){
-    *rr = tolower(*rr);
-  }
-  return rslt;
-}
-
-////////////////
-string ltrim(const string foo){
-  size_t where = foo.find_first_not_of(" \t\r\n");
-  if (where == foo.npos) return foo;
-  return foo.substr(where);
-}
-
 string noCR(const string bar){
   string foo(bar);
   int len = foo.length();
@@ -111,13 +94,6 @@ string noCR(const string bar){
   return foo;
 }
 
-////////////////
-// little utility to help with argument parsing:
-//
-int prefix(const string shorter, const string longer){
-  return shorter == longer.substr(0, shorter.length());
-}
-
 void maybe_exeunt(const int sts, const int really){
   if (!really) return;
   if (sts == ex_good) exit(sts);
@@ -142,12 +118,6 @@ void exeunt(const int sts){
   maybe_exeunt(sts, 1);
 }
 
-string basename(const string path){
-  size_t where = path.rfind("/");
-  if (where != string::npos) return path.substr(1+where);
-  return path;
-}
-
 string progname, progid;
 int mypid;
 
-- 
cgit v1.2.3