require 'time'
-require 'iconv'
module Redwood
-class MessageFormatError < StandardError; end
-
## a Message is what's threaded.
##
## it is also where the parsing for quotes and signatures is done, but
## specific module that would detect and link to /ruby-talk:\d+/
## sequences in the text of an email. (how sweet would that be?)
##
-## this class cathces all source exceptions. if the underlying source throws
-## an error, it is caught and handled.
+## this class catches all source exceptions. if the underlying source
+## throws an error, it is caught and handled.
class Message
SNIPPET_LEN = 80
QUOTE_PATTERN = /^\s{0,4}[>|\}]/
BLOCK_QUOTE_PATTERN = /^-----\s*Original Message\s*----+$/
- QUOTE_START_PATTERN = /\w.*:$/
SIG_PATTERN = /(^-- ?$)|(^\s*----------+\s*$)|(^\s*_________+\s*$)|(^\s*--~--~-)|(^\s*--\+\+\*\*==)/
MAX_SIG_DISTANCE = 15 # lines from the end
@snippet = opts[:snippet]
@snippet_contains_encrypted_content = false
@have_snippet = !(opts[:snippet].nil? || opts[:snippet].empty?)
- @labels = [] + (opts[:labels] || [])
+ @labels = (opts[:labels] || []).to_set_of_symbols
@dirty = false
@encrypted = false
@chunks = nil
## why.
@refs = []
- parse_header(opts[:header] || @source.load_header(@source_info))
+ #parse_header(opts[:header] || @source.load_header(@source_info))
end
def parse_header header
+ ## forcibly decode these headers from and to the current encoding,
+ ## which serves to strip out characters that aren't displayable
+ ## (and which would otherwise be screwing up the display)
+ %w(from to subject cc bcc).each do |f|
+ header[f] = Iconv.easy_decode($encoding, $encoding, header[f]) if header[f]
+ end
+
@id = if header["message-id"]
mid = header["message-id"] =~ /<(.+?)>/ ? $1 : header["message-id"]
sanitize_message_id mid
#Redwood::log "faking non-existent message-id for message from #{from}: #{id}"
id
end
-
+
@from = Person.from_address(if header["from"]
header["from"]
else
name = "Sup Auto-generated Fake Sender <sup@fake.sender.example.com>"
- Redwood::log "faking non-existent sender for message #@id: #{name}"
+ #Redwood::log "faking non-existent sender for message #@id: #{name}"
name
end)
begin
Time.parse date
rescue ArgumentError => e
- Redwood::log "faking mangled date header for #{@id} (orig #{header['date'].inspect} gave error: #{e.message})"
+ #Redwood::log "faking mangled date header for #{@id} (orig #{header['date'].inspect} gave error: #{e.message})"
Time.now
end
else
- Redwood::log "faking non-existent date header for #{@id}"
+ #Redwood::log "faking non-existent date header for #{@id}"
Time.now
end
@list_subscribe = header["list-subscribe"]
@list_unsubscribe = header["list-unsubscribe"]
end
- private :parse_header
+
+ ## Expected index entry format:
+ ## :message_id, :subject => String
+ ## :date => Time
+ ## :refs, :replytos => Array of String
+ ## :from => Person
+ ## :to, :cc, :bcc => Array of Person
+ def load_from_index! entry
+ @id = entry[:message_id]
+ @from = entry[:from]
+ @date = entry[:date]
+ @subj = entry[:subject]
+ @to = entry[:to]
+ @cc = entry[:cc]
+ @bcc = entry[:bcc]
+ @refs = (@refs + entry[:refs]).uniq
+ @replytos = entry[:replytos]
+
+ @replyto = nil
+ @list_address = nil
+ @recipient_email = nil
+ @source_marked_read = false
+ @list_subscribe = nil
+ @list_unsubscribe = nil
+ end
def add_ref ref
@refs << ref
def has_label? t; @labels.member? t; end
def add_label t
return if @labels.member? t
- @labels.push t
+ @labels = (@labels + [t]).to_set_of_symbols
@dirty = true
end
def remove_label t
end
def labels= l
- @labels = l
+ @labels = l.to_set_of_symbols
@dirty = true
end
## this is called when the message body needs to actually be loaded.
def load_from_source!
@chunks ||=
- if @source.has_errors?
+ if @source.respond_to?(:has_errors?) && @source.has_errors?
[Chunk::Text.new(error_message(@source.error.message).split("\n"))]
else
begin
## so i will keep this.
parse_header @source.load_header(@source_info)
message_to_chunks @source.load_message(@source_info)
- rescue SourceError, SocketError, MessageFormatError => e
+ rescue SourceError, SocketError => e
Redwood::log "problem getting messages from #{@source}: #{e.message}"
## we need force_to_top here otherwise this window will cover
## up the error message one
to.map { |p| p.indexable_content },
cc.map { |p| p.indexable_content },
bcc.map { |p| p.indexable_content },
- chunks.select { |c| c.is_a? Chunk::Text }.map { |c| c.lines },
- Message.normalize_subj(subj),
+ indexable_chunks.map { |c| c.lines },
+ indexable_subject,
].flatten.compact.join " "
end
+ def indexable_body
+ indexable_chunks.map { |c| c.lines }.flatten.compact.join " "
+ end
+
+ def indexable_chunks
+ chunks.select { |c| c.is_a? Chunk::Text }
+ end
+
+ def indexable_subject
+ Message.normalize_subj(subj)
+ end
+
def quotable_body_lines
chunks.find_all { |c| c.quotable? }.map { |c| c.lines }.flatten
end
"Subject: #{@subj}"]
end
+ def self.build_from_source source, source_info
+ m = Message.new :source => source, :source_info => source_info
+ m.load_from_source!
+ m
+ end
+
private
## here's where we handle decoding mime attachments. unfortunately
[notice, sig, children].flatten.compact
end
+ ## takes a RMail::Message, breaks it into Chunk:: classes.
def message_to_chunks m, encrypted=false, sibling_types=[]
if m.multipart?
chunks =
elsif m.header["Content-Type"] && m.header["Content-Type"] !~ /^text\/plain/
extension =
case m.header["Content-Type"]
- when /text\/html/: "html"
- when /image\/(.*)/: $1
+ when /text\/html/ then "html"
+ when /image\/(.*)/ then $1
end
["sup-attachment-#{Time.now.to_i}-#{rand 10000}", extension].join(".")
# Lowercase the filename because searches are easier that way
@attachments.push filename.downcase unless filename =~ /^sup-attachment-/
add_label :attachment unless filename =~ /^sup-attachment-/
- [Chunk::Attachment.new(m.header.content_type, filename, m, sibling_types)]
+ content_type = m.header.content_type || "application/unknown" # sometimes RubyMail gives us nil
+ [Chunk::Attachment.new(content_type, filename, m, sibling_types)]
## otherwise, it's body text
else
- body = Message.convert_from m.decode, m.charset if m.body
+ ## if there's no charset, use the current encoding as the charset.
+ ## this ensures that the body is normalized to avoid non-displayable
+ ## characters
+ body = Iconv.easy_decode($encoding, m.charset || $encoding, m.decode) if m.body
text_to_chunks((body || "").normalize_whitespace.split("\n"), encrypted)
end
end
end
- def self.convert_from body, charset
- begin
- raise MessageFormatError, "RubyMail decode returned a null body" unless body
- return body unless charset
- Iconv.easy_decode($encoding, charset, body)
- rescue Errno::EINVAL, Iconv::InvalidEncoding, Iconv::IllegalSequence, MessageFormatError => e
- Redwood::log "warning: error (#{e.class.name}) decoding message body from #{charset}: #{e.message}"
- File.open(File.join(BASE_DIR,"unable-to-decode.txt"), "w") { |f| f.write body }
- body
- end
- end
-
## parse the lines of text into chunk objects. the heuristics here
## need tweaking in some nice manner. TODO: move these heuristics
## into the classes themselves.
when :text
newstate = nil
- if line =~ QUOTE_PATTERN || (line =~ QUOTE_START_PATTERN && nextline =~ QUOTE_PATTERN)
+ ## the following /:$/ followed by /\w/ is an attempt to detect the
+ ## start of a quote. this is split into two regexen because the
+ ## original regex /\w.*:$/ had very poor behavior on long lines
+ ## like ":a:a:a:a:a" that occurred in certain emails.
+ if line =~ QUOTE_PATTERN || (line =~ /:$/ && line =~ /\w/ && nextline =~ QUOTE_PATTERN)
newstate = :quote
elsif line =~ SIG_PATTERN && (lines.length - i) < MAX_SIG_DISTANCE
newstate = :sig
projects / sup / blobdiff