Details
-
Bug
-
Status: Open (View Workflow)
-
Minor
-
Resolution: Unresolved
-
10.4(EOL), 10.5, 10.6, 10.10(EOL), 10.11, 11.0(EOL), 11.1(EOL), 11.2(EOL)
-
None
Description
Some replication tests use MTR variables (i.e. set with `let`) for saving and restoring global variables, but should use user variables instead, as if the value is different at the start of the test, the test will fail. For example, debug_dbug when running mtr with --debug-common.
I think the point is that the test will fail, because the .result file will contain the old value when the value is restored, like
eval SET GLOBAL debug_dbug='$old_dbug';
right?
I agree that it's good to solve the problem by using user variables instead. Just want to point out that it's also fine to solve differently, eg. --disable-query-log for example, if that is more convenient in some specific cases.