From 366065c0f635fc14b6b086783ffb844485f73d52 Mon Sep 17 00:00:00 2001
From: Oneric <oneric@oneric.stub>
Date: Sat, 11 Jan 2025 21:25:00 +0100
Subject: [PATCH] fetcher: split out core object fetch validation

To allow reuse for adapted key validation logic
---
 lib/pleroma/object/fetcher.ex | 41 +++++++++++++++++++++++++----------
 1 file changed, 29 insertions(+), 12 deletions(-)

diff --git a/lib/pleroma/object/fetcher.ex b/lib/pleroma/object/fetcher.ex
index 11ed57ed5..0f129f318 100644
--- a/lib/pleroma/object/fetcher.ex
+++ b/lib/pleroma/object/fetcher.ex
@@ -265,6 +265,28 @@ def fetch_and_contain_remote_object_from_id(%{"id" => id}, is_ap_id),
   def fetch_and_contain_remote_object_from_id(id, is_ap_id) when is_binary(id) do
     Logger.debug("Fetching object #{id} via AP [ap_id=#{is_ap_id}]")
 
+    fetch_and_contain_remote_ap_doc(
+      id,
+      is_ap_id,
+      fn final_uri, data -> {Containment.contain_id_to_fetch(final_uri, data), data["id"]} end
+    )
+  end
+
+  def fetch_and_contain_remote_object_from_id(_id, _is_ap_id),
+    do: {:error, :invalid_id}
+
+  # Fetches an AP document and performing variable security checks on it.
+  #
+  # Note that the received documents "id" matching the final host domain
+  # is always enforced before the custom ID check runs.
+  @spec fetch_and_contain_remote_ap_doc(
+          String.t(),
+          boolean(),
+          (String.t(), Map.t() -> {:ok | :error, String.t() | term()})
+        ) :: {:ok, Map.t()} | {:reject, term()} | {:error, term()}
+  defp fetch_and_contain_remote_ap_doc(id, is_ap_id, verify_id) do
+    Logger.debug("Dereferencing AP doc #{}")
+
     with {:valid_uri_scheme, true} <- {:valid_uri_scheme, String.starts_with?(id, "http")},
          %URI{} = uri <- URI.parse(id),
          {:mrf_reject_check, {:ok, nil}} <-
@@ -277,7 +299,7 @@ def fetch_and_contain_remote_object_from_id(id, is_ap_id) when is_binary(id) do
          true <- !is_ap_id || final_id == id,
          {:ok, data} <- safe_json_decode(body),
          {_, :ok} <- {:containment, Containment.contain_origin(final_id, data)},
-         {_, _, :ok} <- {:strict_id, data["id"], Containment.contain_id_to_fetch(final_id, data)} do
+         {_, {:ok, _}} <- {:strict_id, verify_id.(final_id, data)} do
       unless Instances.reachable?(final_id) do
         Instances.set_reachable(final_id)
       end
@@ -289,14 +311,12 @@ def fetch_and_contain_remote_object_from_id(id, is_ap_id) when is_binary(id) do
       # Similarly keys, either use a fragment ID and are a subobjects or a distinct ID
       # but for compatibility are still a subobject presenting their owning actors ID at the toplevel.
       # Refetching _once_ from the listed id, should yield a strict match afterwards.
-      {:strict_id, ap_id, _} = e ->
-        case is_ap_id do
-          false ->
-            fetch_and_contain_remote_object_from_id(ap_id, true)
-
-          true ->
-            log_fetch_error(id, e)
-            {:error, :id_mismatch}
+      {:strict_id, {_error, ap_id}} = e ->
+        if !is_ap_id and is_binary(ap_id) do
+          fetch_and_contain_remote_ap_doc(ap_id, true, verify_id)
+        else
+          log_fetch_error(id, e)
+          {:error, :id_mismatch}
         end
 
       {:mrf_reject_check, _} = e ->
@@ -327,9 +347,6 @@ def fetch_and_contain_remote_object_from_id(id, is_ap_id) when is_binary(id) do
     end
   end
 
-  def fetch_and_contain_remote_object_from_id(_id, _is_ap_id),
-    do: {:error, :invalid_id}
-
   # HOPEFULLY TEMPORARY
   # Basically none of our Tesla mocks in tests set the (supposed to
   # exist for Tesla proper) url parameter for their responses