curl to Ruby

What this tool is used for

• Converting an API's curl example into Net::HTTP for a Ruby service.
• Reproducing a captured request inside an existing Ruby client.
• Getting the SSL and header setup right without reading the whole Net::HTTP API.
• Producing a script to run a one-off API call from the command line.
• Checking how a POST body is expressed before wiring it into a class.

Frequently Asked Questions

Why do I need use_ssl for https URLs?

Net::HTTP does not infer TLS from the scheme. Net::HTTP.new(host, 443) alone speaks plaintext to a TLS port and fails confusingly — set use_ssl = true, or use Net::HTTP.start(uri.host, uri.port, use_ssl: uri.scheme == 'https').

Should I use Net::HTTP, Faraday or HTTParty?

Net::HTTP is in the standard library and needs no gem. Faraday adds a middleware stack and swappable adapters, which suits libraries. HTTParty is the terser choice for scripts. All three end up on the same sockets.

How do I send JSON?

request.body = payload.to_json and request['Content-Type'] = 'application/json'. Net::HTTP will not serialise or set the header for you. For form-encoded bodies matching curl's -d, use request.set_form_data(hash) instead.

Does Net::HTTP follow redirects?

No — it returns the 3xx response and you must follow it yourself, typically by looping on Net::HTTPRedirection and reading response['location']. Faraday can do it via the FollowRedirects middleware.

How do I set timeouts?

open_timeout covers establishing the connection and read_timeout covers waiting for data, both on the Net::HTTP object. Defaults are 60 seconds each — finite, but far longer than most callers want.

Why prefer Net::HTTP.start over the one-line helpers?

Because the block form opens one connection and reuses it for every request inside, while `Net::HTTP.get` opens and closes a connection each time. For more than one call to the same host that difference is most of the latency.

Common errors and gotchas

• Forgetting `use_ssl = true`, which silently attempts plain HTTP against an https URL.
• Building the request without a path from the URI, which Net::HTTP needs separately from the host.
• Setting a JSON body without the content-type header, which many APIs then reject.
• Assuming redirects are followed, when Net::HTTP does not follow them by default.
• Expecting a non-2xx to raise, when Net::HTTP returns a response object you must check.