Created
February 13, 2010 10:50
-
-
Save bensummers/303378 to your computer and use it in GitHub Desktop.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
# Apple Binary Property List serializer for Ruby 1.8. | |
require 'iconv' | |
module AppleBinaryPropertyList | |
MIME_TYPE = 'application/octet-stream' # Don't know what to use, so use a very generic type for now | |
CFData = Struct.new(:data) # For marking strings as binary data which will be decoded as a CFData object | |
# Convert a Ruby data structure into an OS X binary property list file. (.plist) | |
# Works as you'd expect. Integers are limited to 4 bytes, even though the format implies longer values can be written. | |
# Strings are assumed to be in UTF-8 format. Symbols are written as strings. | |
def self.convert(data) | |
write("", data) | |
end | |
# Alternative interface which writes data to the out object using << | |
def self.write(out, data) | |
# Find out how many objects there are, so we know how big the references are | |
count = count_objects(data) | |
ref_format, ref_size = int_format_and_size(count) | |
# Now serialize all the objects | |
values = Array.new | |
append_values(data, values, ref_format) | |
# Write header, then the values, calculating offsets as they're written | |
out << 'bplist00' | |
offset = 8 | |
offsets = Array.new | |
values.each do |v| | |
offsets << offset | |
out << v | |
offset += v.length | |
end | |
# How big should the offset ints be? | |
# Decoder gets upset if the size can't fit the entire file, even if it's not strictly needed, so add the length of the last value. | |
offset_format, offset_size = int_format_and_size(offsets.last + values.last.length) | |
# Write the offsets | |
out << offsets.pack(offset_format) | |
# Write trailer | |
out << [0,0,offset_size,ref_size, 0,values.length, 0,0, 0,offset].pack("NnCCNNNNNN") | |
out | |
end | |
private | |
def self.count_objects(data) | |
case data | |
when Array | |
data.inject(1) { |sum,x| sum + count_objects(x) } | |
when Hash | |
# Note: Assumes that the keys aren't a Hash or Array | |
data.length + count_objects(data.values) | |
else | |
1 | |
end | |
end | |
def self.append_values(data, values, ref_format) | |
case data | |
# Constant values | |
when nil | |
# values << "\x00" | |
raise "Can't store a nil in a binary plist. While the format supports it, decoders don't like it." | |
when false | |
values << "\x08" | |
when true | |
values << "\x09" | |
when Integer | |
raise "Integer out of range to write in binary plist" if data < -2147483648 || data > 0x7FFFFFFF | |
values << packed_int(data) | |
when Float | |
values << "\x23#{[data].pack("d").reverse}" | |
when Symbol | |
append_values(data.to_s, values, ref_format) | |
when String | |
if data =~ /[\x80-\xff]/ | |
# Has high bits set, so is UTF-8 and must be reencoded for the plist file | |
c = Iconv.iconv(PLIST_TEXT_ENCODING, INPUT_TEXT_ENCODING, data).join | |
values << "#{objhdr_with_length(0x60, c.length / 2)}#{c}" | |
else | |
# Just ASCII | |
o = objhdr_with_length(0x50, data.length) | |
o << data | |
values << o | |
end | |
when AppleBinaryPropertyList::CFData | |
o = objhdr_with_length(0x40, data.data.length) | |
o << data.data | |
values << o | |
when Time | |
v = data.getutc.to_f - DATE_EPOCH_OFFSET_APPLE_UNIX | |
values << "\x33#{[v].pack("d").reverse}" | |
when Hash | |
o = objhdr_with_length(0xd0, data.length) | |
values << o # now, so we get the refs of other objects right | |
ks = Array.new | |
vs = Array.new | |
data.each do |k,v| | |
ks << values.length | |
append_values(k, values, ref_format) | |
vs << values.length | |
append_values(v, values, ref_format) | |
end | |
o << ks.pack(ref_format) | |
o << vs.pack(ref_format) | |
when Array | |
o = objhdr_with_length(0xa0, data.length) | |
values << o # now, so we get the refs of other objects right | |
refs = Array.new | |
data.each do |e| | |
refs << values.length # index in array of object we're about to write | |
append_values(e, values, ref_format) | |
end | |
o << refs.pack(ref_format) | |
else | |
raise "Couldn't serialize value of class #{data.class.name}" | |
end | |
end | |
def self.int_format_and_size(i) | |
if i > 0xffff | |
['N*',4] | |
elsif i > 0xff | |
['n*',2] | |
else | |
['C*',1] | |
end | |
end | |
def self.packed_int(data) | |
if data < 0 | |
# Need to use 64 bits for negative numbers. | |
[0x13,0xffffffff,data].pack("CNN") | |
elsif data > 0xffff | |
[0x12,data].pack("CN") | |
elsif data > 0xff | |
[0x11,data].pack("Cn") | |
else | |
[0x10,data].pack("CC") | |
end | |
end | |
def self.objhdr_with_length(id, length) | |
if length < 0xf | |
(id + length).chr | |
else | |
(id + 0xf).chr + packed_int(length) | |
end | |
end | |
# Difference between Apple and UNIX timestamps | |
DATE_EPOCH_OFFSET_APPLE_UNIX = 978307200 | |
# Text encoding | |
INPUT_TEXT_ENCODING = 'UTF-8' | |
PLIST_TEXT_ENCODING = 'UTF-16BE' | |
end |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment