auto-approve for terragrunt destroy not working

[gercograndia]

Describe the bug

After upgrading to the latest version of terragrunt, when running terragrunt destroy -auto-approve it still keeps asking for approval.

Steps To Reproduce

❯ terragrunt --version
terragrunt version v0.58.2

❯ terragrunt destroy -auto-approve --terragrunt-no-auto-approve=false
Detected dependent modules:
xxx
WARNING: Are you sure you want to continue? (y/n)

Expected behavior

If you pass the -auto-approve flag, it should destroy without asking for approval.

Using plain terraform is does without complaining:

❯ terraform --version
Terraform v1.5.7
on darwin_arm64
+ provider registry.terraform.io/hashicorp/azuread v2.48.0
+ provider registry.terraform.io/hashicorp/azurerm v3.101.0
+ provider registry.terraform.io/hashicorp/random v3.6.1
+ provider terraformprovider.z6.web.core.windows.net/eon/eoncms v5.0.1

Your version of Terraform is out of date! The latest version
is 1.8.2. You can update by downloading from https://www.terraform.io/downloads.html

❯ terraform destroy -auto-approve -var-file terragrunt-debug.tfvars.json
...
Destroy complete! Resources: 15 destroyed.

Versions

• Terragrunt version: see before
• OpenTofu/Terraform version: see before
• Environment details (Ubuntu 20.04, Windows 10, etc.): macos Sonoma

[CristianoOliveira1]

This works for me (still asks for confirmation if there are dependent modules)
terragrunt destroy -auto-approve

and for all..
terragrunt run-all --terragrunt-non-interactive destroy

[gercograndia]

So you're saying the auto-approve works as intended but if there are dependent modules there will always be a request for confirmation?

So in my view the auto-approve is doing half of its job then.

[CristianoOliveira1]

yes, was also surprised with that behaviour. but when using destroy for the whole set of modules, ( "terragrunt run-all --terragrunt-non-interactive destroy" ) it worked.

[gercograndia]

Ok, I do believe this is somewhat a workaround and the auto-approve should just auto-approve.

But glad there is a workaround at least, so we can move on.

Thanks for your input!