request is nil in middleware on missing or incorrect content_type

We have a Logger that extends from the Grape middleman base. This logger need to have access to the request from env["api.request"] in the after method, which when a http request is sent with no content_type defined env["api.request"].request is nil which raises an error.

We are currently mounted with the use method

The following is the middle man call trace. Our middleware is API::Logger.

# Middleware call trace

## GOOD, correct content_type
### Before
API::Logger
Grape::Middleware::Versioner::Path
Grape::Middleware::Formatter

- Calls app endpoint here, logs sql whatnot

### After
Grape::Middleware::Formatter
Grape::Middleware::Versioner::Path
API::Logger

## BAD, missing content_type
### Before
API::Logger
Grape::Middleware::Versioner::Path
Grape::Middleware::Formatter

- Does NOT hit api

### After
Grape::Middleware::Versioner::Path
API::Logger
- raises error because request is missing from env["api.endpoint"]

One weird part in the bad calls is it goes up.

API::Logger
Grape::Middleware::Versioner::Path
Grape::Middleware::Formatter

but then one way down.

Grape::Middleware::Versioner::Path
API::Logger

My question is the request deleted if it does not match the content_type required or is this a bug?

No source files or tests are named. Start by reproducing a request with missing or incorrect content_type and trace Grape::Middleware::Formatter, Grape::Middleware::Versioner::Path, and API::Logger around env["api.request"]. Establish the expected middleware and request behavior, then verify it with regression coverage.