- BREAK_RE = /^From \S+/
- HEADER_RE = /\s*(.*?\S)\s*/
-
- def read_header f
- header = {}
- last = nil
-
- ## i do it in this weird way because i am trying to speed things up
- ## when scanning over large mbox files.
- while(line = f.gets)
- case line
- when /^(From):#{HEADER_RE}$/i,
- /^(To):#{HEADER_RE}$/i,
- /^(Cc):#{HEADER_RE}$/i,
- /^(Bcc):#{HEADER_RE}$/i,
- /^(Subject):#{HEADER_RE}$/i,
- /^(Date):#{HEADER_RE}$/i,
- /^(References):#{HEADER_RE}$/i,
- /^(In-Reply-To):#{HEADER_RE}$/i,
- /^(Reply-To):#{HEADER_RE}$/i,
- /^(List-Post):#{HEADER_RE}$/i,
- /^(List-Subscribe):#{HEADER_RE}$/i,
- /^(List-Unsubscribe):#{HEADER_RE}$/i,
- /^(Status):#{HEADER_RE}$/i: header[last = $1] = $2
- when /^(Message-Id):#{HEADER_RE}$/i: header[mid_field = last = $1] = $2
-
- ## these next three can occur multiple times, and we want the
- ## first one
- when /^(Delivered-To):#{HEADER_RE}$/i,
- /^(X-Original-To):#{HEADER_RE}$/i,
- /^(Envelope-To):#{HEADER_RE}$/i: header[last = $1] ||= $2
-
- when /^\r*$/: break
- when /^\S+:/: last = nil # some other header we don't care about
- else
- header[last] += " " + line.chomp.gsub(/^\s+/, "") if last
- end
- end
-
- if mid_field && header[mid_field] && header[mid_field] =~ /<(.*?)>/
- header[mid_field] = $1
+ BREAK_RE = /^From \S+ (.+)$/
+
+ def is_break_line? l
+ l =~ BREAK_RE or return false
+ time = $1
+ begin
+ ## hack -- make Time.parse fail when trying to substitute values from Time.now
+ Time.parse time, 0
+ true
+ rescue NoMethodError
+ warn "found invalid date in potential mbox split line, not splitting: #{l.inspect}"
+ false