Expose redirect history on responses

[ZilvinasKucinskas]

Priority: P1

Describe the issue

When redirects are enabled, Wreq::Response#url exposes only the final URL. Callers cannot inspect which statuses and locations were followed, how many hops occurred, or where a cross-origin transition happened.

The underlying Rust response already stores redirect history entries containing the status, previous URI, resolved destination URI, and redirect-response headers. The Ruby binding should expose that existing data as structured Ruby values without inventing fields that the native history does not retain.

Proposed Ruby API

response = client.get("https://example.com/old")

response.url
# => "https://example.com/new"

response.history.each do |hop|
  puts [hop.status, hop.previous_url, hop.url].inspect
end

response.history.map(&:to_h)
# => [
#   {
#     status: 301,
#     previous_url: "https://example.com/old",
#     url: "https://example.com/new",
#     headers: #<Wreq::Headers ...>
#   }
# ]

history should always return an array, empty when no redirects were followed. Each entry can be an immutable Wreq::RedirectHistoryEntry value object with readers and to_h.

Recommended initial fields:

• status: integer status of the redirect response;
• previous_url: URI requested before the redirect;
• url: resolved destination URI retained by Rust wreq; and
• headers: duplicate-preserving headers of the redirect response.

Avoid storing response bodies. The first API should not promise the raw Location field separately, request-method rewriting, per-hop timings, or other information Rust wreq does not retain in HistoryEntry.

Acceptance criteria

• Response#history returns an ordered, immutable view of followed redirects.
• No-redirect responses return an empty array.
• Relative redirects expose the resolved destination retained upstream.
• Cross-origin redirects and the final URL are represented correctly through the retained URI fields.
• Intermediate headers preserve repeated values.
• History does not retain intermediate body buffers or prevent connection cleanup.
• Sensitive URL components are redacted from inspect and default serialization.
• Tests cover no redirect, one hop, multiple hops, relative locations, cross-origin redirects, and redirects disabled.

Relevant public references

• wreq-ruby already advertises redirect policy support and exposes the final URL.
• HTTPX documents follow-redirect behavior and request rewriting.
• The underlying wreq redirect implementation already places structured History entries in response extensions.
• wreq-python exposes these exact fields as status, url, previous, and headers.

Relevant wreq-ruby source

• src/client/resp.rs
• lib/wreq_ruby/response.rb