nikody

nikody

I noticed that the follow_redirect: true option used in HTTPoison (and therefore Hackney) when combined with an Authorization header (in my case basic auth) makes subsequent requests with the same Authorization headers. Curl, for example, does not seem to do that, at least not by default for the -L option, it looks like the --location-trusted option is to be used precisely when the user wants to reuse basic auth headers on redirects.

I want to use the Authorization header only on the first request and follow redirects without it. I stumbled upon an issue with Twilio media URLs with this where they require authentication for a media URL that redirects twice - to their CDN and then S3 - but the last URL, the S3 one, returns an error (unsupported authorization type) if a request with an Authorization header is made (or at least with the one required for the first URL), it works without any authorization though.

My question is - is it possible to specify to only use the headers passed to HTTPoison.get!/3 for the first request and follow redirects without them. I guess the question is if it’s possible in Hackney in general, and if not, is there a not too complex way to override this?

Showing Posts 1 to 2

fireproofsocks

fireproofsocks

You can manually do the redirects, adapting the headers/options on each request, but it’s kinda busywork that follow_redirect: true hopes to avoid unless you happen upon a case like this.

Here’s some code I came up with once upon a time to deal with a similar scenario where specific insight was needed to know info about the hops involved in the redirect.

A couple things to point out:

  1. You need to build in a safety valve, i.e. a maximum number of redirects, otherwise you could end up in an endless redirect.
  2. The redirect targets need to be qualified, e.g. by prepending the full hostname, so this recursive process keeps a little bit if state.

Here’s my code – put together into a single module. Hopefully this is useful to you: you can see where you could manipulate the options between successive requests.

defmodule RecursiveClient do
  @moduledoc """
  This module augments the functionality available in another HTTP client (e.g.
  `HTTPoison`) so we can know more about requests that _redirect_. Specifically,
  we want to know the final URL where a series of redirects landed.

  Whereas `HTTPoison` only discloses the originally requested URL (when `:follow_redirect`
  is true), this module keeps track of *both* the originally requested URL and the
  the ultimate URL that finally yielded a successful response.

  Note: only `:get` requests are currently supported.
  """

  use HTTPoison.Base

  import HTTPClients

  alias HTTPoison.Response

  @default_client HTTPoison
  @hard_limit_max_redirects 10
  @custom_response_header "X-Original-Request-URL"

  @doc """
  See `RecursiveClient.get/3`
  """
  @impl true
  def get(url), do: get(url, [], [])

  @doc """
  See `RecursiveClient.get/3`
  """
  @impl true
  def get(url, headers), do: get(url, headers, [])

  @doc """
  Recursively makes a GET request to the given `url` following any redirects up to
  `max_redirect` times while keeping track of the URL that initiated the sequence
  of requests as well as the URL that eventually yielded content.

  ## Options

    - `:client` an HTTP client module implementing the `HTTPoison.Base`
      callbacks. Default: `#{@default_client}`
    - `:max_redirect` non-negative integer providing a hard-stop on the number
      of redirects this will follow. Default `#{@hard_limit_max_redirects}`

  ## Examples

      iex> RecursiveClient.get("https://blog.ahrefs.com", [], client: HTTPoison, max_redirect: 3)
      {:ok, %HTTPoison.Response{
        body: _,
        headers: [{"#{@custom_response_header}", "https://blog.ahrefs.com"} | _],
        request: _,
        request_url: "https://ahrefs.com/blog/",
        status_code: 200
      }}

      iex> {:ok, response} = RecursiveClient.get("https://blog.ahrefs.com")
      iex> RecursiveClient.requested_resolved(response)
      {:ok, "https://blog.ahrefs.com", "https://ahrefs.com/blog/"}
  """
  @impl true
  def get(url, headers, options) do
    {http_client, options} = Keyword.pop(options, :client, @default_client)
    {max_redirect, options} = Keyword.pop(options, :max_redirect, @hard_limit_max_redirects)
    options = Keyword.put(options, :follow_redirect, false)

    case make_recursive_request({http_client, :get, [url, headers, options]}, max_redirect, 1) do
      {:ok, response} -> {:ok, append_custom_response_header(url, response)}
      {:error, _error} -> {:error, %HTTPoison.Error{id: url, reason: :ehostunreach}}
    end
  end

  defp make_recursive_request(_, max_redirect, n_redirect)
       when n_redirect > max_redirect do
    {:error, "Too many redirects"}
  end

  # The first of the args is assumed to be the URL to request.
  # We shift it off because we may need to replace it with a new URL
  defp make_recursive_request(
         {client, function, [url | other_args] = args},
         max_redirect,
         n_redirect
       ) do
    case is_valid_url?(url) do
      true ->
        client
        |> make_request(function, args)
        |> handle_response(fn new_request_url ->
          make_recursive_request(
            {client, function, [new_request_url | other_args]},
            max_redirect,
            n_redirect + 1
          )
        end)

      false ->
        {:error, "Invalid url: #{inspect(url)}"}
    end
  end

  defp handle_response(
         {:error, %Response{status_code: status_code} = response},
         redirect_callback
       )
       when status_code >= 300 and status_code <= 399 do
    case target_url(response) do
      {:ok, new_request_url} ->
        redirect_callback.(new_request_url)

      # Couldn't determine the new request URL
      {:error, error} ->
        {:error, error}
    end
  end

  defp handle_response(response, _), do: response

  # Adds a custom response header to the Response: we stash the info
  # there so it is in a place that is compatible with the defined behaviour.
  defp append_custom_response_header(original_url, %Response{headers: headers} = response) do
    Map.put(response, :headers, [{@custom_response_header, original_url}] ++ headers)
  end

  @doc """
  Takes in a URL and determines if URL is valid or invalid.
  See https://github.com/johno/is_url/blob/master/lib/is_url.ex

  ## Examples
      iex> RecursiveClient.is_valid_url?("https://blog.ahrefs.com")
      true

      iex> RecursiveClient.is_valid_url?("invalid_url")
      false
  """
  @spec is_valid_url?(binary) :: boolean()
  def is_valid_url?(url) when is_binary(url) do
    case URI.parse(url) do
      %URI{scheme: scheme} when scheme not in ["https", "http"] -> false
      %URI{host: host} when host in ["", nil] -> false
      _ -> true
    end
  end

  @doc """
  Given a redirect response, this returns a fully qualified url to which it redirects.

  ## Examples

      iex> RecursiveClient.target_url(%HTTPoison.Response{
        body:
          ""
        headers: [
          {"Location", "https://ahrefs.com/blog/"}
        ],
        request: %HTTPoison.Request{
          body: "",
          headers: [],
          method: :get,
          options: [follow_redirect: false],
          params: %{},
          url: "https://blog.ahrefs.com"
        },
        request_url: "https://blog.ahrefs.com",
        status_code: 301
      })

      {:ok, "https://ahrefs.com/blog/"}
  """
  @spec target_url(response_struct :: HTTPoison.Response.t()) ::
          {:ok, String.t()} | {:error, any()}
  def target_url(%Response{request_url: request_url, headers: headers, status_code: status_code})
      when status_code >= 300 and status_code <= 399 do
    with {:ok, location} <- get_location(headers) do
      fully_qualify_location(request_url, location)
    end
  end

  defp fully_qualify_location(request_url, location) do
    uri1 = URI.parse(request_url)
    uri2 = URI.parse(location)

    {:ok,
     URI
     |> struct(
       Map.merge(
         uri1 |> Map.from_struct(),
         uri2
         |> Map.from_struct()
         |> Enum.reject(fn {_, v} -> is_nil(v) end)
         |> Enum.into(%{})
       )
     )
     |> URI.to_string()}
  rescue
    e in ArgumentError ->
      {:error,
       "Invalid redirect location #{inspect(location)} returned from #{request_url}; #{inspect(e)}"}
  end

  defp get_location(headers) do
    headers
    |> Enum.find_value(:error, fn
      {k, v} ->
        case String.downcase(k) do
          "location" ->
            v

          _ ->
            false
        end
    end)
    |> case do
      :error -> {:error, "Location not found"}
      location -> {:ok, location}
    end
  end
end

It’s possible some other HTTP client has more nuanced rules around its redirects that would avoid a custom solution like this for your use-case :thinking:

nikody

nikody OP

Thanks a lot! Your idea might very much help, I thought a little about it and it might be the route I go if nothing better pops up.

— All posts loaded —

Where Next? Top

Trending in Questions Top

water
I’m aware of 2 such projects, are there others? git: GitHub - team-alembic/realworld: A fullstack Phoenix LiveView application with back...
New
katta
I having some trouble figuring out if I have set myself too strict of standards for my production server. Currently I can handle 75% of r...
New
peck
Hi everyone. My team and I have been working on a fairly modest app based around video streaming and chat, but we’ve landed a customer t...
New
bradley
I really like the adapter patterns that ecto, nebulex, waffle, etc. use and would love find something similar for a key management servic...
New
unaware8150
Hello folks! So at work, we are seeing some situations where we have to define some “fixed” strings that are used across the codebase in...
New
Alvinkariuki
How Can I Optimise Compile Time Dependencies I have been building an elixir application for about 2 years now. Many modules and files ha...
New
leoschweizer
Hi, we run different LLM models across a cluster, with each node loading a configurable subset. Model-specific queues let nodes consume ...
New

Other Trending Topics Top

KristerV
Hey. Is there anyone here who creates agents in their apps? Not talking about using agents, but creating them. I’m finding it pretty diff...
New
mudasobwa
I fully migrated to my own harness from Anthropic/Gemini and I think it’s time to share it. Welcome DSH, the DeepSeek Harness, fully writ...
New
jimsynz
Beam Bots (or just BB for short) is a framework for building fault-tolerant robotics applications in Elixir using familiar OTP patterns. ...
New
juhalehtonen
There has been a thread to discuss the Stack Overflow Developer Survey on this forum every year since 2018, so here’s yet another one for...
New
budgie
I love Elixir. It’s one of 2 programming languages I’ve ever fallen in love with. But I don’t use it anymore. Serverless was the promis...
New
ausimian
Emily is an Elixir library that runs Nx computations on Apple’s MLX. Install it as the default Nx backend and Nx, defn, Axon, Nx.Serving,...
New

We're in Beta

About us Mission Statement

Options

Thread Display Mode




Thread Preview

Skip Thread Previews