X-Git-Url: https://git.cworth.org/git?a=blobdiff_plain;f=lib%2Fsup%2Fmbox.rb;h=0d941b1d8ae341b86bcf6303e752c3c4bb1c36e5;hb=5bf815cacd0bd5b5b3b5e42a85fed9b03326269a;hp=5b346ef5c1b0725b149b106f098f07057b2324b9;hpb=974c3731a0a91fca5dfef766d519ea108f0cbb68;p=sup

diff --git a/lib/sup/mbox.rb b/lib/sup/mbox.rb
index 5b346ef..0d941b1 100644
--- a/lib/sup/mbox.rb
+++ b/lib/sup/mbox.rb
@@ -1,58 +1,25 @@
 require "sup/mbox/loader"
 require "sup/mbox/ssh-file"
 require "sup/mbox/ssh-loader"
+require "sup/rfc2047"
 
 module Redwood
 
-## some utility functions
 module MBox
-  BREAK_RE = /^From \S+/
+  BREAK_RE = /^From \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):\s+(.*)$/i,
-        /^(To):\s+(.*)$/i,
-        /^(Cc):\s+(.*)$/i,
-        /^(Bcc):\s+(.*)$/i,
-        /^(Subject):\s+(.*)$/i,
-        /^(Date):\s+(.*)$/i,
-        /^(Message-Id):\s+<(.*)>$/i,
-        /^(References):\s+(.*)$/i,
-        /^(In-Reply-To):\s+(.*)$/i,
-        /^(Reply-To):\s+(.*)$/i,
-        /^(List-Post):\s+(.*)$/i,
-        /^(Status):\s+(.*)$/i: header[last = $1] = $2
-
-      ## these next three can occur multiple times, and we want the
-      ## first one
-      when /^(Delivered-To):\s+(.*)$/i,
-        /^(X-Original-To):\s+(.*)$/i,
-        /^(Envelope-To):\s+(.*)$/i: header[last = $1.downcase] ||= $2
-
-      when /^$/: break
-      when /:/: last = nil
-      else
-        header[last] += line.chomp.gsub(/^\s+/, "") if last
-      end
+  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
+      Redwood::log "found invalid date in potential mbox split line, not splitting: #{l.inspect}"
+      false
     end
-    header
   end
-  
-  def read_body f
-    body = []
-    f.each_line do |l|
-      break if l =~ BREAK_RE
-      body << l.chomp
-    end
-    body
-  end
-
-  module_function :read_header, :read_body
+  module_function :is_break_line?
 end
 end