fix: migrator run with nil schema (#5795)

[black-06]

• Do only one thing
• Non breaking API changes
• Tested

What did this pull request do?

fix DropColumn RenameColumn HasColumn with string table param.
Close #5795

User Case Description

It should work instead of panic

migrator := DB.Migrator()
migrator.RenameColumn("column_structs", "new_name", "new_new_name")
migrator.HasColumn("column_structs", "new_new_name")
migrator.DropColumn("column_structs", "new_new_name")

[a631807682]

This is a breaking change, because we use if stmt.Schema == nil in many places to judge

[black-06]

This is a breaking change, because we use if stmt.Schema == nil in many places to judge

Okay, I'll modify it

[a631807682]

Test cases need to wait for drivers to be merged.
Can tests cover changes? For example, AlterColumn/AddColumn needs to provide datatype, so it does not support specifying by table name, and an error needs to be returned at this time.

[black-06]

Test cases need to wait for drivers to be merged. Can tests cover changes? For example, AlterColumn/AddColumn needs to provide datatype, so it does not support specifying by table name, and an error needs to be returned at this time.

modified~

[a631807682]

The driver changes have been merged, I re-triggered the ci but the test still fails, is it related to this change?

[black-06]

The driver changes have been merged, I re-triggered the ci but the test still fails, is it related to this change?

I will check it.

[black-06]

When RenameIndex(&IndexStruct{}, "idx_index_structs_name", "idx_users_name_1") in sqlite, it actually exec

CREATE INDEX `idx_users_name_1` ON `index_structs`(`name`)

This is a bug, right? Because idx_index_structs_name still exist.
We need to fix it first

[a631807682]

When RenameIndex(&IndexStruct{}, "idx_index_structs_name", "idx_users_name_1") in sqlite, it actually exec

CREATE INDEX `idx_users_name_1` ON `index_structs`(`name`)

This is a bug, right? Because idx_index_structs_name still exist. We need to fix it first

yes