9 include SerializeLabelsNicely
10 yaml_properties :uri, :cur_offset, :usual, :archived, :id, :labels
14 ## uri_or_fp is horrific. need to refactor.
15 def initialize uri_or_fp, start_offset=0, usual=true, archived=false, id=nil, labels=nil
17 @labels = Set.new((labels || []) - LabelManager::RESERVED_LABELS)
21 uri = URI(Source.expand_filesystem_uri(uri_or_fp))
22 raise ArgumentError, "not an mbox uri" unless uri.scheme == "mbox"
23 raise ArgumentError, "mbox URI ('#{uri}') cannot have a host: #{uri.host}" if uri.host
24 raise ArgumentError, "mbox URI must have a path component" unless uri.path
25 @f = File.open uri.path
29 @path = uri_or_fp.path
32 super uri_or_fp, start_offset, usual, archived, id
35 def file_path; @path end
36 def is_source_for? uri; super || (self.uri.is_a?(String) && (URI(Source.expand_filesystem_uri(uri)) == URI(Source.expand_filesystem_uri(self.uri)))) end
38 def self.suggest_labels_for path
39 ## heuristic: use the filename as a label, unless the file
40 ## has a path that probably represents an inbox.
41 if File.dirname(path) =~ /\b(var|usr|spool)\b/
44 [File.basename(path).downcase.intern]
49 if (cur_offset ||= start_offset) > end_offset
50 raise OutOfSyncSourceError, "mbox file is smaller than last recorded message offset. Messages have probably been deleted by another client."
54 def start_offset; 0; end
55 def end_offset; File.size @f; end
57 def load_header offset
62 unless MBox::is_break_line? l
63 raise OutOfSyncSourceError, "mismatch in mbox file offset #{offset.inspect}: #{l.inspect}."
65 header = parse_raw_email_header @f
70 def load_message offset
74 ## don't use RMail::Mailbox::MBoxReader because it doesn't properly ignore
75 ## "From" at the start of a message body line.
78 string << l until @f.eof? || MBox::is_break_line?(l = @f.gets)
79 RMail::Parser.read string
80 rescue RMail::Parser::Error => e
81 raise FatalSourceError, "error parsing mbox file: #{e.message}"
86 ## scan forward until we're at the valid start of a message
91 until @f.eof? || MBox::is_break_line?(l = @f.gets)
94 self.cur_offset += string.length
100 @mutex.synchronize do
102 until @f.eof? || (l = @f.gets) =~ /^\r*$/
109 def raw_message offset
111 each_raw_message_line(offset) { |l| ret << l }
115 def store_message date, from_email, &block
116 need_blank = File.exists?(@filename) && !File.zero?(@filename)
117 File.open(@filename, "a") do |f|
119 f.puts "From #{from_email} #{date.utc}"
124 ## apparently it's a million times faster to call this directly if
125 ## we're just moving messages around on disk, than reading things
126 ## into memory with raw_message.
128 ## i hoped never to have to move shit around on disk but
129 ## sup-sync-back has to do it.
130 def each_raw_message_line offset
131 @mutex.synchronize do
134 until @f.eof? || MBox::is_break_line?(l = @f.gets)
141 returned_offset = nil
142 next_offset = cur_offset
145 @mutex.synchronize do
148 ## cur_offset could be at one of two places here:
150 ## 1. before a \n and a mbox separator, if it was previously at
151 ## EOF and a new message was added; or,
152 ## 2. at the beginning of an mbox separator (in all other
155 l = @f.gets or return nil
156 if l =~ /^\s*$/ # case 1
157 returned_offset = @f.tell
158 @f.gets # now we're at a BREAK_RE, so skip past it
160 returned_offset = cur_offset
161 ## we've already skipped past the BREAK_RE, so just go
164 while(line = @f.gets)
165 break if MBox::is_break_line? line
166 next_offset = @f.tell
169 rescue SystemCallError, IOError => e
170 raise FatalSourceError, "Error reading #{@f.path}: #{e.message}"
173 self.cur_offset = next_offset
174 [returned_offset, (labels + [:unread])]