PT-1776: Adding in slave-user and slave-password for dsn table based …

[mbenshoof]

…recursion

[it-percona]

All committers have signed the CLA.

[svetasmirnova]

Thank you for your contribution.

You modified the auto-generated code. It will be removed once we run the update-modules utility. Please put your changes into the MasterSlave package.

[svetasmirnova]

Another issue with this proposal is that it modifies DSN table entries without checking if they have user and password defined. This could cause issues when users have, say, 10 replicas with the same user/password pairs and two with different ones.

We also need a test case to accept this PR.

[svetasmirnova]

There is also similar PR at #381

[mbenshoof]

@svetasmirnova - Modified the logic to check for user/pass specified in DSN string. Only sets the CLI passed options if they are not set in the DSN table.

For the test case, I've been using dbdeployer for a quick primary/replica environment and setting different variations of DSN string in the table (with both user/pass, with only one, with neither) to validate it doesn't overwrite, but sets if missing

[svetasmirnova]

@mbenshoof

Thank you for the changes; I will review them in the new year.

For the test case, I've been using dbdeployer for a quick primary/replica environment and setting different variations of the DSN string in the table (with both user/pass, with only one, with neither) to validate it doesn't overwrite, but sets if missing

We need automatic test cases to run them at each release and ensure your changes are not overwritten by another contribution. We use a solution similar to MySQL Sandbox and dbdeployer for automatic test cases. Please check directory t and CONTRIBUTING.md.