Author: martinkl
Date: Thu Jul 3 15:10:26 2014
New Revision: 1607667
URL: http://svn.apache.org/r1607667
Log:
AVRO-1499. Ruby: Fix corruption of data files under Ruby 2.0+. Contributed by
Willem van Bergen.
Modified:
avro/trunk/CHANGES.txt
avro/trunk/lang/ruby/lib/avro/data_file.rb
Modified: avro/trunk/CHANGES.txt
URL:
http://svn.apache.org/viewvc/avro/trunk/CHANGES.txt?rev=1607667&r1=1607666&r2=1607667&view=diff
==============================================================================
--- avro/trunk/CHANGES.txt (original)
+++ avro/trunk/CHANGES.txt Thu Jul 3 15:10:26 2014
@@ -107,6 +107,9 @@ Trunk (not yet released)
AVRO-1525. Java: ReflectData cannot resolve union with fixed. (tomwhite)
+ AVRO-1499. Ruby: Fix encoding issue that caused corrupted data files
+ to be written under Ruby 2.0+. (Willem van Bergen and martinkl)
+
Avro 1.7.6 (15 January 2014)
NEW FEATURES
Modified: avro/trunk/lang/ruby/lib/avro/data_file.rb
URL:
http://svn.apache.org/viewvc/avro/trunk/lang/ruby/lib/avro/data_file.rb?rev=1607667&r1=1607666&r2=1607667&view=diff
==============================================================================
--- avro/trunk/lang/ruby/lib/avro/data_file.rb (original)
+++ avro/trunk/lang/ruby/lib/avro/data_file.rb Thu Jul 3 15:10:26 2014
@@ -20,7 +20,8 @@ module Avro
module DataFile
VERSION = 1
MAGIC = "Obj" + [VERSION].pack('c')
- MAGIC_SIZE = MAGIC.size
+ MAGIC.force_encoding('BINARY') if MAGIC.respond_to?(:force_encoding)
+ MAGIC_SIZE = MAGIC.respond_to?(:bytesize) ? MAGIC.bytesize : MAGIC.size
SYNC_SIZE = 16
SYNC_INTERVAL = 4000 * SYNC_SIZE
META_SCHEMA = Schema.parse('{"type": "map", "values": "bytes"}')
@@ -98,6 +99,7 @@ module Avro
@encoder = IO::BinaryEncoder.new(@writer)
@datum_writer = datum_writer
@buffer_writer = StringIO.new('', 'w')
+ @buffer_writer.set_encoding('BINARY') if
@buffer_writer.respond_to?(:set_encoding)
@buffer_encoder = IO::BinaryEncoder.new(@buffer_writer)
@block_count = 0
@@ -181,7 +183,7 @@ module Avro
# write number of items in block and block size in bytes
encoder.write_long(block_count)
to_write = codec.compress(buffer_writer.string)
- encoder.write_long(to_write.size)
+ encoder.write_long(to_write.respond_to?(:bytesize) ?
to_write.bytesize : to_write.size)
# write block contents
writer.write(to_write)
