FLUSH TABLE t FOR EXPORT or ALTER TABLE t2 IMPORT TABLESPACE

 
How to repeat:

1. Create a sandbox with dbdeployer.
   $ dbdeployer deploy single mysql_8.0.17

1. Create a schema and a table.
2. I am using many CHAR fields fo generate a big dataset quickly.
   $ ./use <<< "CREATE DATABASE test_jfg"
   $ ./use test_jfg <<< "
   CREATE TABLE test_jfg.t(id INT UNSIGNED AUTO_INCREMENT PRIMARY KEY);
   ALTER TABLE test_jfg.t ADD COLUMN v VARCHAR(255);"
   $ for i in $(seq 9); do
   ./use test_jfg <<< "ALTER TABLE test_jfg.t ADD COLUMN c$i CHAR(255) DEFAULT ''"
   done

1. Generate ~2 GB of data in that table.
   $ seq -f "INSERT INTO t(v) VALUES ('%f');" 1 1000000 | sed -e '1~1000s/^/COMMIT;BEGIN;/' | ./use test_jfg

1. Contrarily to
   Bug#96475
   , I will run all next steps manually to show things are broken.

– In the mysql client, run FLUSH TABLE [...] FOR EXPORT.
mysql [localhost:8017] {msandbox} (test_jfg) > FLUSH TABLE t FOR EXPORT;
Query OK, 0 rows affected (0.00 sec)

1. In bash, copy the files.
   $ for s in cfg ibd; do cp data/test_jfg/t{,_}.$s; done

– In the . mysql client, UNLOCK TABLE and prepare import.
mysql [localhost:8017] {msandbox} (test_jfg) > UNLOCK TABLES;
Query OK, 0 rows affected (0.00 sec)

mysql [localhost:8017] {msandbox} (test_jfg) > CREATE TABLE t2 LIKE t; ALTER TABLE t2 DISCARD TABLESPACE;
Query OK, 0 rows affected (0.04 sec)

Query OK, 0 rows affected (0.01 sec)

1. In bash, copy the files.
   $ for s in cfg ibd; do cp data/test_jfg/t{_,2}.$s; done

– In the mysql client, import the table.
mysql [localhost:8017] {msandbox} (test_jfg) > ALTER TABLE t2 IMPORT TABLESPACE;
ERROR 1817 (HY000): Index corrupt: Clustered index validation failed, due to data file corruption.

– In above, we can see that things fail, but I would expect them to succeed.