avro

AVRO-818. C: Fix data file corruption bug

This patch fixes the C library's data file writer to produce correct
files when there are enough records to produce multiple blocks in the
file.  The logic in datafile.c is that we have a fixed-buffer in-memory
avro_writer_t instance.  When you append records to the data file, they
go into this memory buffer.  If we get an error serializing into the
memory buffer, it's presumably because we've filled it, so we write out
the memory buffer's contents as a new block in the file, clear the
buffer, and try again.

The problem is that the failed serialization into the memory buffer
isn't atomic; some of the serialization will have made it into the
buffer before we discover that there's not enough room.  Before, this
incomplete record would then make it into the file.  To fix this, we
keep track of the size of the in-memory buffer after the most recent
successfully serialized record, and use this as the block size when we
write a block to disk.  This ensures that even if there's any incomplete
records at the end of the memory buffer, we don't include them in the
block.


git-svn-id: https://svn.apache.org/repos/asf/avro/trunk@1102332 13f79535-47bb-0310-9956-ffa450edef68