Bug #111540 binlog compress optimize

Description:
When the compressed size of the ROW event being processed is higher than
it's uncompressed counter part, it is better to abandon this compression.
Have the following benefits:
1. Saves storage space
2. The slave node reduces decompression costs

How to repeat:
Refer to the repetition scheme in binlog_compression_size_higher.test:
# 2. Create a table with a blob column.
CREATE TABLE t1 (id bigint NOT NULL AUTO_INCREMENT PRIMARY KEY,
                 x mediumblob NOT NULL) ENGINE=InnoDB;

# 3. Prepare an insert statement and populate the blob columns with
#    uncompressable data.
PREPARE s FROM "INSERT INTO t1 (x) VALUES (?)";
SET @a = LOAD_FILE('../../std_data/binlog_compression.gz');
EXECUTE s USING @a;

Suggested fix:
just as the below patch

a simple patch to describe the optimization

(*) I confirm the code being submitted is offered under the terms of the OCA, and that I am authorized to contribute it.

Contribution: binlog_compress_optimize.patch (text/plain), 1.46 KiB.

Hello alex xing,

Thank you very much for your patch contribution, we appreciate it!

regards,
Umesh