Preserve files when reprocessing

When using S3 for storage, their data consistency model makes the
current behaviour of Attachment#reprocess! unreliable.

Apparently, Amazon can not guarantee that PUT's called after DELETE's
is actually executed in that order.

Hence, I propose preserving files, when doing a reprocces.

lib/paperclip/attachment.rb

@@ -314,6 +314,7 @@ def hash_key(style_name = default_style)
     # the post-process again.
     def reprocess!(*style_args)
       saved_only_process, @options[:only_process] = @options[:only_process], style_args
+      saved_preserve_files, @options[:preserve_files] = @options[:preserve_files], true
       begin
         assign(self)
         save
@@ -323,6 +324,7 @@ def reprocess!(*style_args)
         false
       ensure
         @options[:only_process] = saved_only_process
+        @options[:preserve_files] = saved_preserve_files
       end
     end