diff options
Diffstat (limited to 'app/models/incoming_message.rb')
| -rw-r--r-- | app/models/incoming_message.rb | 50 |
1 files changed, 33 insertions, 17 deletions
diff --git a/app/models/incoming_message.rb b/app/models/incoming_message.rb index 5c845ead3..c914edb7e 100644 --- a/app/models/incoming_message.rb +++ b/app/models/incoming_message.rb @@ -25,7 +25,7 @@ # response from the public body. # # Copyright (c) 2007 UK Citizens Online Democracy. All rights reserved. -# Email: francis@mysociety.org; WWW: http://www.mysociety.org/ +# Email: hello@mysociety.org; WWW: http://www.mysociety.org/ # TODO # Move some of the (e.g. quoting) functions here into rblib, as they feel @@ -37,7 +37,7 @@ require 'rexml/document' require 'zip/zip' require 'mapi/msg' require 'mapi/convert' - +require 'iconv' unless RUBY_VERSION >= '1.9' class IncomingMessage < ActiveRecord::Base belongs_to :info_request @@ -193,7 +193,7 @@ class IncomingMessage < ActiveRecord::Base text.gsub!(self.info_request.public_body.request_email, _("[{{public_body}} request email]", :public_body => self.info_request.public_body.short_or_long_name)) end text.gsub!(self.info_request.incoming_email, _('[FOI #{{request}} email]', :request => self.info_request.id.to_s) ) - text.gsub!(Configuration::contact_email, _("[{{site_name}} contact email]", :site_name => Configuration::site_name) ) + text.gsub!(AlaveteliConfiguration::contact_email, _("[{{site_name}} contact email]", :site_name => AlaveteliConfiguration::site_name) ) end # Replaces all email addresses in (possibly binary data) with equal length alternative ones. @@ -219,7 +219,7 @@ class IncomingMessage < ActiveRecord::Base if censored_uncompressed_text != uncompressed_text # then use the altered file (recompressed) recompressed_text = nil - if Configuration::use_ghostscript_compression == true + if AlaveteliConfiguration::use_ghostscript_compression == true command = ["gs", "-sDEVICE=pdfwrite", "-dCompatibilityLevel=1.4", "-dPDFSETTINGS=/screen", "-dNOPAUSE", "-dQUIET", "-dBATCH", "-sOutputFile=-", "-"] else command = ["pdftk", "-", "output", "-", "compress"] @@ -258,11 +258,21 @@ class IncomingMessage < ActiveRecord::Base # equivalents to the UCS-2 ascii_chars = text.gsub(/\0/, "") emails = ascii_chars.scan(MySociety::Validate.email_find_regexp) + # Convert back to UCS-2, making a mask at the same time - emails.map! {|email| [ - Iconv.conv('ucs-2le', 'ascii', email[0]), - Iconv.conv('ucs-2le', 'ascii', email[0].gsub(/[^@.]/, 'x')) - ] } + if RUBY_VERSION >= '1.9' + emails.map! do |email| + # We want the ASCII representation of UCS-2 + [email[0].encode('UTF-16LE').force_encoding('US-ASCII'), + email[0].gsub(/[^@.]/, 'x').encode('UTF-16LE').force_encoding('US-ASCII')] + end + else + emails.map! {|email| [ + Iconv.conv('ucs-2le', 'ascii', email[0]), + Iconv.conv('ucs-2le', 'ascii', email[0].gsub(/[^@.]/, 'x')) + ] } + end + # Now search and replace the UCS-2 email with the UCS-2 mask for email, mask in emails text.gsub!(email, mask) @@ -316,7 +326,7 @@ class IncomingMessage < ActiveRecord::Base text.gsub!(/(Mobile|Mob)([\s\/]*(Fax|Tel))*\s*:?[\s\d]*\d/, "[mobile number]") # Remove WhatDoTheyKnow signup links - text.gsub!(/http:\/\/#{Configuration::domain}\/c\/[^\s]+/, "[WDTK login link]") + text.gsub!(/http:\/\/#{AlaveteliConfiguration::domain}\/c\/[^\s]+/, "[WDTK login link]") # Remove things from censor rules self.info_request.apply_censor_rules_to_text!(text) @@ -534,7 +544,7 @@ class IncomingMessage < ActiveRecord::Base source_charset = 'utf-8' if source_charset.nil? text = Iconv.conv('utf-8//IGNORE', source_charset, text) + _("\n\n[ {{site_name}} note: The above text was badly encoded, and has had strange characters removed. ]", - :site_name => Configuration::site_name) + :site_name => AlaveteliConfiguration::site_name) rescue Iconv::InvalidEncoding, Iconv::IllegalSequence, Iconv::InvalidCharacter if source_charset != "utf-8" source_charset = "utf-8" @@ -605,7 +615,7 @@ class IncomingMessage < ActiveRecord::Base content_type = 'application/octet-stream' end hexdigest = Digest::MD5.hexdigest(content) - attachment = self.foi_attachments.find_or_create_by_hexdigest(:hexdigest => hexdigest) + attachment = self.foi_attachments.find_or_create_by_hexdigest(hexdigest) attachment.update_attributes(:filename => filename, :content_type => content_type, :body => content, @@ -632,14 +642,14 @@ class IncomingMessage < ActiveRecord::Base attachment_attributes = MailHandler.get_attachment_attributes(self.mail(force)) attachments = [] attachment_attributes.each do |attrs| - attachment = self.foi_attachments.find_or_create_by_hexdigest(:hexdigest => attrs[:hexdigest]) - body = attrs.delete(:body) + attachment = self.foi_attachments.find_or_create_by_hexdigest(attrs[:hexdigest]) attachment.update_attributes(attrs) - # Set the body separately as its handling can depend on the value of charset - attachment.body = body attachment.save! attachments << attachment.id end + # Reload to refresh newly created foi_attachments + self.reload + main_part = get_main_body_text_part # we don't use get_main_body_text_internal, as we want to avoid charset # conversions, since /usr/bin/uudecode needs to deal with those. @@ -748,9 +758,15 @@ class IncomingMessage < ActiveRecord::Base attachment.body, attachment.charset) end + # Remove any bad characters - text = Iconv.conv('utf-8//IGNORE', 'utf-8', text) - return text + if RUBY_VERSION >= '1.9' + text.encode("utf-8", :invalid => :replace, + :undef => :replace, + :replace => "") + else + Iconv.conv('utf-8//IGNORE', 'utf-8', text) + end end |