From 572d258c2aafbcf2dfaae5490316bb1cfaa80627 Mon Sep 17 00:00:00 2001
From: SJ <sj@acts.hu>
Date: Wed, 22 Aug 2012 08:15:53 +0200
Subject: [PATCH] added min_word_len option to discard/skip too short words in
 email body

---
 etc/example.conf | 3 ++-
 src/cfg.c        | 1 +
 src/cfg.h        | 2 ++
 src/parser.c     | 2 +-
 4 files changed, 6 insertions(+), 2 deletions(-)

diff --git a/etc/example.conf b/etc/example.conf
index 6af079a1..0ed00ceb 100644
--- a/etc/example.conf
+++ b/etc/example.conf
@@ -58,7 +58,8 @@ piler_header_field=X-piler: piler already archived this email
 ; to messages without message-id.
 archive_emails_not_having_message_id=0
 
-
+; minimum word length in mail body to index
+min_word_len=1
 
 ; comma separated list of your domains. piler uses this information to determine
 ; the direction of the given email
diff --git a/src/cfg.c b/src/cfg.c
index 4502b554..53b95844 100644
--- a/src/cfg.c
+++ b/src/cfg.c
@@ -75,6 +75,7 @@ struct _parse_rule config_parse_rules[] =
    { "memcached_servers", "string", (void*) string_parser, offsetof(struct __config, memcached_servers), "127.0.0.1", MAXVAL-1},
    { "memcached_to_db_interval", "integer", (void*) int_parser, offsetof(struct __config, memcached_to_db_interval), "900", sizeof(int)},
    { "memcached_ttl", "integer", (void*) int_parser, offsetof(struct __config, memcached_ttl), "86400", sizeof(int)},
+   { "min_word_len", "integer", (void*) int_parser, offsetof(struct __config, min_word_len), "1", sizeof(int)},
    { "mydomains", "string", (void*) string_parser, offsetof(struct __config, mydomains), "", MAXVAL-1},
    { "mysqlhost", "string", (void*) string_parser, offsetof(struct __config, mysqlhost), "", MAXVAL-1},
    { "mysqlport", "integer", (void*) int_parser, offsetof(struct __config, mysqlport), "", sizeof(int)},
diff --git a/src/cfg.h b/src/cfg.h
index 24940e65..52e4f7c2 100644
--- a/src/cfg.h
+++ b/src/cfg.h
@@ -64,6 +64,8 @@ struct __config {
 
    int archive_emails_not_having_message_id;
 
+   int min_word_len;
+
    int debug;
 };
 
diff --git a/src/parser.c b/src/parser.c
index 0d208b63..eb3c8d68 100644
--- a/src/parser.c
+++ b/src/parser.c
@@ -534,7 +534,7 @@ int parse_line(char *buf, struct _state *state, struct session_data *sdata, int
             }
          }
       }
-      else if(state->message_state == MSG_BODY && state->bodylen < BIGBUFSIZE-len-1){
+      else if(state->message_state == MSG_BODY && len >= cfg->min_word_len && state->bodylen < BIGBUFSIZE-len-1){
          memcpy(&(state->b_body[state->bodylen]), puf, len);
          state->bodylen += len;
       }