diff options
Diffstat (limited to 'app/models/incoming_message.rb')
| -rw-r--r-- | app/models/incoming_message.rb | 35 |
1 files changed, 20 insertions, 15 deletions
diff --git a/app/models/incoming_message.rb b/app/models/incoming_message.rb index bcf0b6ec9..6db145348 100644 --- a/app/models/incoming_message.rb +++ b/app/models/incoming_message.rb @@ -35,7 +35,7 @@ require 'htmlentities' require 'rexml/document' require 'zip/zip' -require 'iconv' unless RUBY_VERSION >= '1.9' +require 'iconv' unless String.method_defined?(:encode) class IncomingMessage < ActiveRecord::Base extend MessageProminence @@ -294,7 +294,7 @@ class IncomingMessage < ActiveRecord::Base emails = ascii_chars.scan(MySociety::Validate.email_find_regexp) # Convert back to UCS-2, making a mask at the same time - if RUBY_VERSION >= '1.9' + if String.method_defined?(:encode) emails.map! do |email| # We want the ASCII representation of UCS-2 [email[0].encode('UTF-16LE').force_encoding('US-ASCII'), @@ -385,6 +385,10 @@ class IncomingMessage < ActiveRecord::Base multiline_original_message = '(' + '''>>>.* \d\d/\d\d/\d\d\d\d\s+\d\d:\d\d(?::\d\d)?\s*>>>''' + ')' text.gsub!(/^(#{multiline_original_message}\n.*)$/m, replacement) + # On Thu, Nov 28, 2013 at 9:08 AM, A User + # <[1]request-7-skm40s2ls@xxx.xxxx> wrote: + text.gsub!(/^( On [^\n]+\n\s*\<[^>\n]+\> (wrote|said):\s*\n.*)$/m, replacement) + # Single line sections text.gsub!(/^(>.*\n)/, replacement) text.gsub!(/^(On .+ (wrote|said):\n)/, replacement) @@ -516,7 +520,7 @@ class IncomingMessage < ActiveRecord::Base # should instead tell elinks to respect the source # charset use_charset = "utf-8" - if RUBY_VERSION.to_f >= 1.9 + if String.method_defined?(:encode) begin text.encode('utf-8') rescue Encoding::UndefinedConversionError, Encoding::InvalidByteSequenceError @@ -547,7 +551,7 @@ class IncomingMessage < ActiveRecord::Base end def _sanitize_text(text) - if RUBY_VERSION.to_f >= 1.9 + if String.method_defined?(:encode) begin # Test if it's good UTF-8 text.encode('utf-8') @@ -792,27 +796,28 @@ class IncomingMessage < ActiveRecord::Base return self.cached_attachment_text_clipped end - def _get_attachment_text_internal + def _extract_text # Extract text from each attachment - text = '' - attachments = self.get_attachments_for_display - for attachment in attachments - text += MailHandler.get_attachment_text_one_file(attachment.content_type, + self.get_attachments_for_display.reduce(''){ |memo, attachment| + memo += MailHandler.get_attachment_text_one_file(attachment.content_type, attachment.body, attachment.charset) - end + } + end + + def _get_attachment_text_internal + text = self._extract_text # Remove any bad characters - if RUBY_VERSION >= '1.9' - text.encode("utf-8", :invalid => :replace, - :undef => :replace, - :replace => "") + if String.method_defined?(:encode) + # handle "problematic" encoding + text.encode!('UTF-16', 'UTF-8', :invalid => :replace, :undef => :replace, :replace => '') + text.encode('UTF-8', 'UTF-16') else Iconv.conv('utf-8//IGNORE', 'utf-8', text) end end - # Returns text for indexing def get_text_for_indexing_full return get_body_for_quoting + "\n\n" + get_attachment_text_full |