Adds Message-Ids to an external database as they flow through Postfix, keeping
the most recent time seen with each.

The goal is to maintain a database of all Message-Ids sent from my system in
the past N days. A content-filter can look for mail delivery status
notifications which do not contain a Message-Id found in this database. Those
can be safely assumed to be backscatter email and rejected.

A cron job could remove stale entries from the database.

This patch overshoots slightly and tracks _all_ Message-Ids seen by Postfix.
That's probably harmless.

diff -ru postfix-2.2.8/src/cleanup/cleanup.h postfix-2.2.8-msgid/src/cleanup/cleanup.h
--- postfix-2.2.8/src/cleanup/cleanup.h	2005-02-12 14:53:25.000000000 -0800
+++ postfix-2.2.8-msgid/src/cleanup/cleanup.h	2006-01-29 09:23:09.000000000 -0800
@@ -93,6 +93,11 @@
 extern MAPS *cleanup_send_bcc_maps;
 extern MAPS *cleanup_rcpt_bcc_maps;
 
+/*
+ * Message-ID database.
+ */
+extern DICT *cleanup_messageid_db;
+
  /*
   * Address canonicalization fine control.
   */
diff -ru postfix-2.2.8/src/cleanup/cleanup_init.c postfix-2.2.8-msgid/src/cleanup/cleanup_init.c
--- postfix-2.2.8/src/cleanup/cleanup_init.c	2004-10-25 13:59:02.000000000 -0700
+++ postfix-2.2.8-msgid/src/cleanup/cleanup_init.c	2006-01-29 10:07:57.000000000 -0800
@@ -122,6 +122,7 @@
 char   *var_send_bcc_maps;		/* sender auto-bcc maps */
 char   *var_rcpt_bcc_maps;		/* recipient auto-bcc maps */
 char   *var_remote_rwr_domain;		/* header-only surrogate */
+char   *var_messageid_db;		/* message-id database */
 
 CONFIG_INT_TABLE cleanup_int_table[] = {
     VAR_HOPCOUNT_LIMIT, DEF_HOPCOUNT_LIMIT, &var_hopcount_limit, 1, 0,
@@ -165,6 +166,7 @@
     VAR_SEND_BCC_MAPS, DEF_SEND_BCC_MAPS, &var_send_bcc_maps, 0, 0,
     VAR_RCPT_BCC_MAPS, DEF_RCPT_BCC_MAPS, &var_rcpt_bcc_maps, 0, 0,
     VAR_REM_RWR_DOMAIN, DEF_REM_RWR_DOMAIN, &var_remote_rwr_domain, 0, 0,
+    VAR_MESSAGEID_DB, DEF_MESSAGEID_DB, &var_messageid_db, 0, 0,
     0,
 };
 
@@ -188,6 +190,11 @@
 MAPS   *cleanup_send_bcc_maps;
 MAPS   *cleanup_rcpt_bcc_maps;
 
+/*
+ * Message-ID database.
+ */
+DICT   *cleanup_messageid_db;
+
  /*
   * Address extension propagation restrictions.
   */
@@ -286,6 +293,11 @@
 	cleanup_rcpt_bcc_maps =
 	    maps_create(VAR_RCPT_BCC_MAPS, var_rcpt_bcc_maps,
 			DICT_FLAG_LOCK);
+    if (*var_messageid_db)
+	cleanup_messageid_db =
+	    dict_open(var_messageid_db, O_RDWR|O_CREAT|O_APPEND,
+		      DICT_FLAG_DUP_REPLACE | DICT_FLAG_LOCK
+		      | DICT_FLAG_SYNC_UPDATE | DICT_FLAG_PARANOID);
 
     flush_init();
 }
diff -ru postfix-2.2.8/src/cleanup/cleanup_message.c postfix-2.2.8-msgid/src/cleanup/cleanup_message.c
--- postfix-2.2.8/src/cleanup/cleanup_message.c	2005-11-26 17:48:44.000000000 -0800
+++ postfix-2.2.8-msgid/src/cleanup/cleanup_message.c	2006-01-29 09:28:16.000000000 -0800
@@ -85,6 +85,20 @@
 
 #include "cleanup.h"
 
+/* cleanup_store_messageid - store Message-IDs for backscatter analysis */
+
+static void cleanup_store_messageid(char *msgid, time_t time)
+{
+    if (cleanup_messageid_db) {
+	char *timestr = NULL;
+	asprintf(&timestr, "%d", time);
+	if (timestr == NULL)
+	    msg_fatal("cleanup_store_messageid: insufficient memory");
+	cleanup_messageid_db->update(cleanup_messageid_db, msgid, timestr);
+	free(timestr);
+    }
+}
+
 /* cleanup_out_header - output one header as a bunch of records */
 
 static void cleanup_out_header(CLEANUP_STATE *state, VSTRING *header_buf)
@@ -533,10 +547,14 @@
      */
     else {
 	state->headers_seen |= (1 << hdr_opts->type);
-	if (hdr_opts->type == HDR_MESSAGE_ID)
+	if (hdr_opts->type == HDR_MESSAGE_ID) {
 	    msg_info("%s: message-id=%s", state->queue_id, hdrval);
-	if (hdr_opts->type == HDR_RESENT_MESSAGE_ID)
+	    cleanup_store_messageid(hdrval, state->time);
+	}
+	if (hdr_opts->type == HDR_RESENT_MESSAGE_ID) {
 	    msg_info("%s: resent-message-id=%s", state->queue_id, hdrval);
+	    cleanup_store_messageid(hdrval, state->time);
+	}
 	if (hdr_opts->type == HDR_RECEIVED)
 	    if (++state->hop_count >= var_hopcount_limit)
 		state->errs |= CLEANUP_STAT_HOPS;
@@ -583,14 +601,21 @@
      */
     if ((state->headers_seen & (1 << (state->resent[0] ?
 			   HDR_RESENT_MESSAGE_ID : HDR_MESSAGE_ID))) == 0) {
+	char *messageid = NULL;
 	tv = state->handle->ctime.tv_sec;
 	tp = gmtime(&tv);
 	strftime(time_stamp, sizeof(time_stamp), "%Y%m%d%H%M%S", tp);
-	cleanup_out_format(state, REC_TYPE_NORM, "%sMessage-Id: <%s.%s@%s>",
-		state->resent, time_stamp, state->queue_id, var_myhostname);
-	msg_info("%s: %smessage-id=<%s.%s@%s>",
-		 state->queue_id, *state->resent ? "resent-" : "",
+	asprintf(&messageid, "%s.%s@%s",
 		 time_stamp, state->queue_id, var_myhostname);
+	if (messageid == NULL)
+	    msg_fatal("cleanup_header_done_callback: insufficient memory");
+	cleanup_out_format(state, REC_TYPE_NORM, "%sMessage-Id: <%s>",
+		state->resent, messageid);
+	msg_info("%s: %smessage-id=<%s>",
+		 state->queue_id, *state->resent ? "resent-" : "",
+		 messageid);
+	cleanup_store_messageid(messageid, state->time);
+	free(messageid);
     }
 
     /*
diff -ru postfix-2.2.8/src/global/mail_params.h postfix-2.2.8-msgid/src/global/mail_params.h
--- postfix-2.2.8/src/global/mail_params.h	2005-02-27 07:06:07.000000000 -0800
+++ postfix-2.2.8-msgid/src/global/mail_params.h	2006-01-29 09:22:00.000000000 -0800
@@ -2327,6 +2327,13 @@
 #define DEF_LOC_RWR_CLIENTS		PERMIT_INET_INTERFACES
 extern char *var_local_rwr_clients;
 
+/*
+ * Where to store outgoing Message-IDs for backscatter analysis.
+ */
+#define VAR_MESSAGEID_DB		"messageid_db"
+#define DEF_MESSAGEID_DB		""
+extern char *var_messageid_db;
+
  /*
   * EHLO keyword filter.
   */