[🐛 Bug]: HttpClient breaks URIs due to let's guess what part of the URI needs encoding

[joerg1985]

Description

#17968 did add a URI encoding which is functional not correct (see example below) and i my mind not useful at all, either a URI has been correctly encoded by the caller or the caller did use the HttpClient incorrect. We should not try to hide this by a let's guess what part of the URI needs encoding approach.

The original issue has been fixed with f3ccc60 so there is no need for this at all.

Will propably raise an PR for this, this is just for documentation or people to stop me (@diemol)

Reproducible Code

// Before PR 17968 the seResponse is also HTTP 200

var seResponse = new JdkHttpClient.Factory().createClient(URI.create("https://www.example.com").toURL())
   .execute(new HttpRequest(HttpMethod.GET, "/#test"));

var jdkResponse = HttpClient.newHttpClient().send(java.net.http.HttpRequest.newBuilder(
   URI.create("https://www.example.com/#test")).build(),
   HttpResponse.BodyHandlers.discarding());

   if (jdkResponse.statusCode() != seResponse.getStatus()) {
      throw new IllegalStateException(jdkResponse.statusCode() + " vs. " + seResponse.getStatus());
}

---

ℹ️ Last known working version: 4.48.0

[jasminritik26]

Encoding # into %23 in HttpRequest breaks fragment routing, so reverting #1