What

Adds a new cache backend that supports s3 for cache storage using the minio client.

Why

s3 will work as a durable storage layer capability of storing larger amounts of data than the currently implemented cache backends.

We're using the Minio s3 client due to it being lighter than the AWS SDK.

Tests

We've added integration tests via a minio test container.

Performed the following steps for manual validation against an AWS s3 bucket:

1. Create a test s3 bucket

aws s3 mb --region us-west-2 s3://sftpc-test-bucket 

1. Created a sftpc.hcl config

host "https://w3.org" {}

s3 {
  bucket = "sftpc-test-bucket"
  region = "us-west-2"
}

3. Start the proxy

go build ./cmd/sfptcd && ./sfptcd --config sfptc.hcl

4. Validate the behaviour

curl http://127.0.0.1:8080/w3.org/TR/html5/
# ...snip...
aws s3 ls s3://sftpc-test-bucket/ --recursive
2026-01-14 09:39:31     152586 2e/2ecdce061e0b582a42737e8eba56a67c6bdcb157c440c9e500a870e7c427a16d