bsfilter

http://nabeken.tdiary.net/20030518.html#p03
という歴史があります。

以下が，出来損ないのReceived:を全て学習対象にするpatchです．

余計なお世話ですが，序でにEnvelope SenderとEnvelope Recipientとの正規表現を修正しました．（特定char setを含むことのみの判定なので，アドレスの構造を正規表現で書いているわけではありません）
また，MTAのqueue IDの正規表現も改めました．

Received:については，MTAのqueue IDとタイムスタンプとをカットして行を継続しているだけです．元々の"envelope-from"と記されるケースがよくわかっていないので，中途半端な状態です．

Rubyを全く理解（しようと）していない輩が体感的に書いたAdHocなpatchなので，非常に恥かしい限りなのですが，情報共有のため&feed backのために晒します．;-)

*** bsfilter.1.55~ Thu Mar 25 00:42:51 2004
--- bsfilter.1.55 Thu Apr 1 00:48:34 2004
***************
*** 819,829 ****
str = str.chomp
if (str =~ /\A(\S+?):\s*(.*)/)
current = $1.downcase
! if (current == "received")
! headers[current] = $2.sub(/[\r\n]*\z/, '')
! else
! headers[current] = (headers[current] || "") + " " + $2.sub(/[\r\n]*\z/, '')
! end
elsif (str =~ /\Afrom\s+(\S+)/i)
headers["ufrom"] = $1
elsif (str =~ /\A\r*\z/)
--- 819,825 ----
str = str.chomp
if (str =~ /\A(\S+?):\s*(.*)/)
current = $1.downcase
! headers[current] = (headers[current] || "") + " " + $2.sub(/[\r\n]*\z/, '')
elsif (str =~ /\Afrom\s+(\S+)/i)
headers["ufrom"] = $1
elsif (str =~ /\A\r*\z/)
***************
*** 831,837 ****
elsif (! current)
break
else
! headers[current] += str.sub(/[\r\n]*\z/, '').sub(/\A\s*/, '')
end
end
if ((headers["content-type"] =~ /\bboundary=\s*"(.*?)"/i) ||
--- 827,837 ----
elsif (! current)
break
else
! if (current == "received")
! headers[current] += " " + str.sub(/;\s.*[\r\n]*\z/, '').sub(/\A\s*/, '')
! else
! headers[current] += " " + str.sub(/[\r\n]*\z/, '').sub(/\A\s*/, '')
! end
end
end
if ((headers["content-type"] =~ /\bboundary=\s*"(.*?)"/i) ||
***************
*** 979,1000 ****
head_db = TokenDB::new(lang)
reg_token = Regexp::compile("\\b\\d[\\d\\.]+\\d\\b|[\\w#{$mark_in_token}]+")

- if (headers["received"])
- str = headers["received"]
- str =~ /envelope\-from\s+([\w@\.\-]+)/
- efrom = $1
- str =~ /for\s+<([\w@\.\-]+)>/
- foraddress = $1
- str.sub!(/(\bid|;).*/im, '')
- str.sub!(/\(qmail[^\)]*\)/, '')
- str += " " + efrom if efrom
- str += " " + foraddress if foraddress
- headers["received"] = str
- end
headers.each do |header, content|
case header
when "ufrom", "from", "to", "cc", "subject", "reply-to", "return-path",
"content-transfer-encoding", "content-type", "content-disposition", "charset", "received"
if (lang == "ja")
content.gsub!(/=\?utf\-8\?([bq])\?(\S*)\?=/i) do |s|
b_or_q = $1
--- 979,998 ----
head_db = TokenDB::new(lang)
reg_token = Regexp::compile("\\b\\d[\\d\\.]+\\d\\b|[\\w#{$mark_in_token}]+")

headers.each do |header, content|
case header
when "ufrom", "from", "to", "cc", "subject", "reply-to", "return-path",
"content-transfer-encoding", "content-type", "content-disposition", "charset", "received"
+ if (header == "received")
+ content =~ /envelope-from\s+([\w@!#$\%&'*+\-\/=?^_`{|}~\.]+)/
+ efrom = $1
+ content =~ /for\s+<([\w@!#$\%&'*+\-\/=?^_`{|}~\.]+)>/
+ foraddress = $1
+ content.gsub!(/(\bid\s+\w+)/im, '')
+ content.gsub!(/\(qmail[^\)]*\)/, '')
+ content += " " + efrom if efrom
+ content += " " + foraddress if foraddress
+ end
if (lang == "ja")
content.gsub!(/=\?utf\-8\?([bq])\?(\S*)\?=/i) do |s|
b_or_q = $1