[#6824] Handle context in Scan() without breaking external library compatibility

[zeze1004]

issue: #6824

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

What did this pull request do?

I initially thought that adding context as an argument to the Scan() method would solve the issue, but it causes compile errors because external libraries do not expect Scan to accept context.

I’m wondering if it’s the right approach to modify Scan to directly accept context, or if there’s a better way to handle this issue without changing the method signature.

package sqlite
...
type Migrator struct {
	migrator.Migrator
}

func (m *Migrator) RunWithoutForeignKey(fc func() error) error {
	var enabled int
	m.DB.Raw("PRAGMA foreign_keys").Scan(&enabled)   // Scan() error ⭐ ⭐ ⭐  
	if enabled == 1 {
		m.DB.Exec("PRAGMA foreign_keys = OFF")
		defer m.DB.Exec("PRAGMA foreign_keys = ON")
	}

	return fc()
}

User Case Description

I’m a bit concerned about how I’m approaching this. Specifically, I’m wondering if adding context.Context internally in the Scan method is the right move. Since external libraries like sqlite might be using this method, I’m not sure if this could cause issues.

Questions:

Is it okay to add context.Context handling like this, or is there a better way I should consider?
Could this change cause compatibility issues with libraries that don’t expect the context.Context?
Should I be thinking about this problem differently, especially regarding method signatures?

I’d really appreciate any guidance or feedback on this. Thanks!

[jinzhu]

Just use it like this? https://gorm.io/docs/context.html#Single-Session-Mode