lib/sup/message.rb in sup-0.3 vs lib/sup/message.rb in sup-0.4
- old
+ new
@@ -1,6 +1,5 @@
-require 'tempfile'
require 'time'
require 'iconv'
module Redwood
@@ -28,58 +27,67 @@
def reify_subj s; subj_is_reply?(s) ? s : "Re: " + s; end
end
QUOTE_PATTERN = /^\s{0,4}[>|\}]/
BLOCK_QUOTE_PATTERN = /^-----\s*Original Message\s*----+$/
- QUOTE_START_PATTERN = /(^\s*Excerpts from)|(^\s*In message )|(^\s*In article )|(^\s*Quoting )|((wrote|writes|said|says)\s*:\s*$)/
+ QUOTE_START_PATTERN = /\w.*:$/
SIG_PATTERN = /(^-- ?$)|(^\s*----------+\s*$)|(^\s*_________+\s*$)|(^\s*--~--~-)|(^\s*--\+\+\*\*==)/
MAX_SIG_DISTANCE = 15 # lines from the end
DEFAULT_SUBJECT = ""
DEFAULT_SENDER = "(missing sender)"
attr_reader :id, :date, :from, :subj, :refs, :replytos, :to, :source,
:cc, :bcc, :labels, :list_address, :recipient_email, :replyto,
- :source_info, :chunks, :list_subscribe, :list_unsubscribe
+ :source_info, :list_subscribe, :list_unsubscribe
- bool_reader :dirty, :source_marked_read
+ bool_reader :dirty, :source_marked_read, :snippet_contains_encrypted_content
## if you specify a :header, will use values from that. otherwise,
## will try and load the header from the source.
def initialize opts
@source = opts[:source] or raise ArgumentError, "source can't be nil"
@source_info = opts[:source_info] or raise ArgumentError, "source_info can't be nil"
- @snippet = opts[:snippet] || ""
- @have_snippet = !opts[:snippet].nil?
+ @snippet = opts[:snippet]
+ @snippet_contains_encrypted_content = false
+ @have_snippet = !(opts[:snippet].nil? || opts[:snippet].empty?)
@labels = [] + (opts[:labels] || [])
@dirty = false
+ @encrypted = false
@chunks = nil
+ ## we need to initialize this. see comments in parse_header as to
+ ## why.
+ @refs = []
+
parse_header(opts[:header] || @source.load_header(@source_info))
end
def parse_header header
header.each { |k, v| header[k.downcase] = v }
-
+
+ fakeid = nil
+ fakename = nil
+
@id =
if header["message-id"]
sanitize_message_id header["message-id"]
else
- returning("sup-faked-" + Digest::MD5.hexdigest(raw_header)) do |id|
- Redwood::log "faking message-id for message from #@from: #{id}"
- end
+ fakeid = "sup-faked-" + Digest::MD5.hexdigest(raw_header)
end
@from =
if header["from"]
PersonManager.person_for header["from"]
else
- name = "Sup Auto-generated Fake Sender <sup@fake.sender.example.com>"
- Redwood::log "faking from for message #@id: #{name}"
- PersonManager.person_for name
+ fakename = "Sup Auto-generated Fake Sender <sup@fake.sender.example.com>"
+ PersonManager.person_for fakename
end
+ Redwood::log "faking message-id for message from #@from: #{id}" if fakeid
+ Redwood::log "faking from for message #@id: #{fakename}" if fakename
+
date = header["date"]
@date =
case date
when Time
date
@@ -96,11 +104,17 @@
@subj = header.member?("subject") ? header["subject"].gsub(/\s+/, " ").gsub(/\s+$/, "") : DEFAULT_SUBJECT
@to = PersonManager.people_for header["to"]
@cc = PersonManager.people_for header["cc"]
@bcc = PersonManager.people_for header["bcc"]
- @refs = (header["references"] || "").scan(/<(.+?)>/).map { |x| sanitize_message_id x.first }
+
+ ## before loading our full header from the source, we can actually
+ ## have some extra refs set by the UI. (this happens when the user
+ ## joins threads manually). so we will merge the current refs values
+ ## in here.
+ refs = (header["references"] || "").scan(/<(.+?)>/).map { |x| sanitize_message_id x.first }
+ @refs = (@refs + refs).uniq
@replytos = (header["in-reply-to"] || "").scan(/<(.+?)>/).map { |x| sanitize_message_id x.first }
@replyto = PersonManager.person_for header["reply-to"]
@list_address =
if header["list-post"]
@@ -114,11 +128,16 @@
@list_subscribe = header["list-subscribe"]
@list_unsubscribe = header["list-unsubscribe"]
end
private :parse_header
- def snippet; @snippet || chunks && @snippet; end
+ def add_ref ref
+ @refs << ref
+ @dirty = true
+ end
+
+ def snippet; @snippet || (chunks && @snippet); end
def is_list_message?; !@list_address.nil?; end
def is_draft?; @source.is_a? DraftLoader; end
def draft_filename
raise "not a draft" unless is_draft?
@source.fn_for_offset @source_info
@@ -150,15 +169,20 @@
def labels= l
@labels = l
@dirty = true
end
+ def chunks
+ load_from_source!
+ @chunks
+ end
+
## this is called when the message body needs to actually be loaded.
def load_from_source!
@chunks ||=
if @source.has_errors?
- [Chunk::Text.new(error_message(@source.error.message.split("\n")))]
+ [Chunk::Text.new(error_message(@source.error.message).split("\n"))]
else
begin
## we need to re-read the header because it contains information
## that we don't store in the index. actually i think it's just
## the mailing list address (if any), so this is kinda overkill.
@@ -173,11 +197,11 @@
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
@source.error ||= e
Redwood::report_broken_sources :force_to_top => true
- [Chunk::Text.new(error_message(e.message))]
+ [Chunk::Text.new(error_message(e.message).split("\n"))]
end
end
end
def error_message msg
@@ -273,11 +297,10 @@
## of the gruesome slaughterhouse and sausage factory that is a
## mime-encoded message, but need only see the delicious end
## product.
def multipart_signed_to_chunks m
-# Redwood::log ">> multipart SIGNED: #{m.header['Content-Type']}: #{m.body.size}"
if m.body.size != 2
Redwood::log "warning: multipart/signed with #{m.body.size} parts (expecting 2)"
return
end
@@ -285,25 +308,26 @@
if signature.multipart?
Redwood::log "warning: multipart/signed with payload multipart #{payload.multipart?} and signature multipart #{signature.multipart?}"
return
end
+ ## this probably will never happen
if payload.header.content_type == "application/pgp-signature"
Redwood::log "warning: multipart/signed with payload content type #{payload.header.content_type}"
return
end
if signature.header.content_type != "application/pgp-signature"
- Redwood::log "warning: multipart/signed with signature content type #{signature.header.content_type}"
+ ## unknown signature type; just ignore.
+ #Redwood::log "warning: multipart/signed with signature content type #{signature.header.content_type}"
return
end
[CryptoManager.verify(payload, signature), message_to_chunks(payload)].flatten.compact
end
def multipart_encrypted_to_chunks m
- Redwood::log ">> multipart ENCRYPTED: #{m.header['Content-Type']}: #{m.body.size}"
if m.body.size != 2
Redwood::log "warning: multipart/encrypted with #{m.body.size} parts (expecting 2)"
return
end
@@ -322,15 +346,15 @@
Redwood::log "warning: multipart/encrypted with control content type #{signature.header.content_type}"
return
end
decryptedm, sig, notice = CryptoManager.decrypt payload
- children = message_to_chunks(decryptedm) if decryptedm
+ children = message_to_chunks(decryptedm, true) if decryptedm
[notice, sig, children].flatten.compact
end
- def message_to_chunks m, sibling_types=[]
+ def message_to_chunks m, encrypted=false, sibling_types=[]
if m.multipart?
chunks =
case m.header.content_type
when "multipart/signed"
multipart_signed_to_chunks m
@@ -338,11 +362,11 @@
multipart_encrypted_to_chunks m
end
unless chunks
sibling_types = m.body.map { |p| p.header.content_type }
- chunks = m.body.map { |p| message_to_chunks p, sibling_types }.flatten.compact
+ chunks = m.body.map { |p| message_to_chunks p, encrypted, sibling_types }.flatten.compact
end
chunks
elsif m.header.content_type == "message/rfc822"
payload = RMail::Parser.read(m.body)
@@ -357,42 +381,51 @@
elsif m.header["Content-Type"] && m.header["Content-Type"] =~ /name="?(.*?[^\\])("|;|$)/
$1
## haven't found one, but it's a non-text message. fake
## it.
+ ##
+ ## TODO: make this less lame.
elsif m.header["Content-Type"] && m.header["Content-Type"] !~ /^text\/plain/
- "sup-attachment-#{Time.now.to_i}-#{rand 10000}"
+ extension =
+ case m.header["Content-Type"]
+ when /text\/html/: "html"
+ when /image\/(.*)/: $1
+ end
+
+ ["sup-attachment-#{Time.now.to_i}-#{rand 10000}", extension].join(".")
end
## if there's a filename, we'll treat it as an attachment.
if filename
[Chunk::Attachment.new(m.header.content_type, filename, m, sibling_types)]
## otherwise, it's body text
else
- body = Message.convert_from m.decode, m.charset
- text_to_chunks body.normalize_whitespace.split("\n")
+ body = Message.convert_from m.decode, m.charset if m.body
+ text_to_chunks((body || "").normalize_whitespace.split("\n"), encrypted)
end
end
end
def self.convert_from body, charset
+ charset = "utf-8" if charset =~ /UTF_?8/i
begin
raise MessageFormatError, "RubyMail decode returned a null body" unless body
return body unless charset
- Iconv.iconv($encoding, charset, body).join
+ Iconv.iconv($encoding + "//IGNORE", charset, body + " ").join[0 .. -2]
rescue Errno::EINVAL, Iconv::InvalidEncoding, Iconv::IllegalSequence, MessageFormatError => e
Redwood::log "warning: error (#{e.class.name}) decoding message body from #{charset}: #{e.message}"
File.open("sup-unable-to-decode.txt", "w") { |f| f.write body }
body
end
end
## parse the lines of text into chunk objects. the heuristics here
## need tweaking in some nice manner. TODO: move these heuristics
## into the classes themselves.
- def text_to_chunks lines
+ def text_to_chunks lines, encrypted
state = :text # one of :text, :quote, or :sig
chunks = []
chunk_lines = []
lines.each_with_index do |line, i|
@@ -400,11 +433,11 @@
case state
when :text
newstate = nil
- if line =~ QUOTE_PATTERN || (line =~ QUOTE_START_PATTERN && (nextline =~ QUOTE_PATTERN || nextline =~ QUOTE_START_PATTERN))
+ if line =~ QUOTE_PATTERN || (line =~ QUOTE_START_PATTERN && nextline =~ QUOTE_PATTERN)
newstate = :quote
elsif line =~ SIG_PATTERN && (lines.length - i) < MAX_SIG_DISTANCE
newstate = :sig
elsif line =~ BLOCK_QUOTE_PATTERN
newstate = :block_quote
@@ -419,11 +452,11 @@
end
when :quote
newstate = nil
- if line =~ QUOTE_PATTERN || line =~ QUOTE_START_PATTERN #|| line =~ /^\s*$/
+ if line =~ QUOTE_PATTERN || (line =~ /^\s*$/ && nextline =~ QUOTE_PATTERN)
chunk_lines << line
elsif line =~ SIG_PATTERN && (lines.length - i) < MAX_SIG_DISTANCE
newstate = :sig
else
newstate = :text
@@ -440,14 +473,17 @@
end
when :block_quote, :sig
chunk_lines << line
end
-
+
if !@have_snippet && state == :text && (@snippet.nil? || @snippet.length < SNIPPET_LEN) && line !~ /[=\*#_-]{3,}/ && line !~ /^\s*$/
+ @snippet ||= ""
@snippet += " " unless @snippet.empty?
@snippet += line.gsub(/^\s+/, "").gsub(/[\r\n]/, "").gsub(/\s+/, " ")
@snippet = @snippet[0 ... SNIPPET_LEN].chomp
+ @dirty = true unless encrypted && $config[:discard_snippets_from_encrypted_messages]
+ @snippet_contains_encrypted_content = true if encrypted
end
end
## final object
case state