Gianni Gambetti

Unexpected `after_commit` and `after_transaction` behaviour with ActiveRecord Persistence

1

**Describe the bug** The `after_commit` and `after_transaction` event callbacks don't behave the same way around nested transactions with ActiveRecord persistence. `after_commit` runs whenever the top level transaction commits, while `after_transaction`...

Shallow cloning by timestamp?

3

Currently `go-git` exposes shallow clones via `CloneOptions.Depth` to specify the number of commits to retrieve. It appears that the underlying plumbing required to support timestamp ranges is implemented (eg: [DeepenSince](https://github.com/go-git/go-git/blob/5dabd83e3712e2554745c736b55df405a0ba4f33/plumbing/protocol/packp/capability/capability.go#L148)),...

Better `plumbing.transport.Transport` isolation?

1

Currently `plumbing.transport.Transport` items are retrieved for remotes via [`plumbing.transport.client.NewClient`](https://github.com/go-git/go-git/blob/v5.4.2/plumbing/transport/client/client.go#L48). This uses a global [`client.Protocols` map](https://github.com/go-git/go-git/blob/v5.4.2/plumbing/transport/client/client.go#L19-L25), which creates complications for managing protocols across application packages and goroutines. I'm wondering if there's...

Running Rack Attack in production systems sometimes requires debugging misbehaving rules. One of the common ways to find misbehaving application components in Ruby is by monkey-patching in tracing information (eg:...