From 7a2ac3ef7ca35a6f9d26fa2b1890019d63a079f1 Mon Sep 17 00:00:00 2001
From: Dmitry Dygalo <dmitry@dygalo.dev>
Date: Wed, 22 Jan 2025 11:02:13 +0100
Subject: [PATCH] fix: Resolving external references nested inside local
 references

Signed-off-by: Dmitry Dygalo <dmitry@dygalo.dev>
---
 CHANGELOG.md                                  |  4 ++
 crates/jsonschema-py/CHANGELOG.md             |  4 ++
 crates/jsonschema-referencing/src/registry.rs | 20 +++++++--
 crates/jsonschema/src/keywords/ref_.rs        | 44 +++++++++++++++++++
 4 files changed, 68 insertions(+), 4 deletions(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index c804cce2..028fa637 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -2,6 +2,10 @@
 
 ## [Unreleased]
 
+### Fixed
+
+- Resolving external references that nested inside local references. [#671](https://github.com/Stranger6667/jsonschema/issues/671)
+
 ## [0.28.1] - 2024-12-31
 
 ### Fixed
diff --git a/crates/jsonschema-py/CHANGELOG.md b/crates/jsonschema-py/CHANGELOG.md
index 0dd56b45..cfb1a442 100644
--- a/crates/jsonschema-py/CHANGELOG.md
+++ b/crates/jsonschema-py/CHANGELOG.md
@@ -2,6 +2,10 @@
 
 ## [Unreleased]
 
+### Fixed
+
+- Resolving external references that nested inside local references. [#671](https://github.com/Stranger6667/jsonschema/issues/671)
+
 ## [0.28.1] - 2024-12-31
 
 ### Fixed
diff --git a/crates/jsonschema-referencing/src/registry.rs b/crates/jsonschema-referencing/src/registry.rs
index 8f0a6158..b073d615 100644
--- a/crates/jsonschema-referencing/src/registry.rs
+++ b/crates/jsonschema-referencing/src/registry.rs
@@ -444,21 +444,32 @@ fn collect_external_resources(
     collected: &mut AHashSet<Uri<String>>,
     seen: &mut AHashSet<u64>,
 ) -> Result<(), Error> {
+    // URN schemes are not supported for external resolution
     if base.scheme().as_str() == "urn" {
         return Ok(());
     }
     for key in ["$ref", "$schema"] {
         if let Some(reference) = contents.get(key).and_then(Value::as_str) {
-            if reference.starts_with('#')
-                || reference.starts_with("https://json-schema.org/draft/2020-12/")
+            // Skip well-known schema references
+            if reference.starts_with("https://json-schema.org/draft/2020-12/")
                 || reference.starts_with("https://json-schema.org/draft/2019-09/")
                 || reference.starts_with("http://json-schema.org/draft-07/")
                 || reference.starts_with("http://json-schema.org/draft-06/")
                 || reference.starts_with("http://json-schema.org/draft-04/")
             {
-                // Not an external resource
-                return Ok(());
+                continue;
+            }
+            // Handle local references separately as they may have nested references to external resources
+            if reference.starts_with('#') {
+                if reference == "#" {
+                    continue;
+                }
+                if let Some(referenced) = contents.pointer(reference.trim_start_matches('#')) {
+                    collect_external_resources(base, referenced, collected, seen)?;
+                }
+                continue;
             }
+
             let mut hasher = AHasher::default();
             (base.as_str(), reference).hash(&mut hasher);
             let hash = hasher.finish();
@@ -466,6 +477,7 @@ fn collect_external_resources(
                 // Reference has already been seen
                 return Ok(());
             }
+
             let resolved = if reference.contains('#') && base.has_fragment() {
                 uri::resolve_against(&uri::DEFAULT_ROOT_URI.borrow(), reference)?
             } else {
diff --git a/crates/jsonschema/src/keywords/ref_.rs b/crates/jsonschema/src/keywords/ref_.rs
index 6d186e88..88fcf426 100644
--- a/crates/jsonschema/src/keywords/ref_.rs
+++ b/crates/jsonschema/src/keywords/ref_.rs
@@ -607,4 +607,48 @@ mod tests {
         let validator = crate::validator_for(&json!({"$ref": "#"})).expect("Invalid schema");
         assert!(validator.is_valid(&json!(42)));
     }
+
+    #[test]
+    fn test_nested_external_reference() {
+        let schema = json!({
+            "$id": "foo://schema_1.json",
+            "$ref": "#/$defs/a/b",
+            "$defs": {
+                "a": {
+                    "b": {
+                        "description": "nested schema with external ref",
+                        "$ref": "foo://schema_2.json"
+                    }
+                }
+            }
+        });
+
+        struct NestedRetrieve;
+
+        impl Retrieve for NestedRetrieve {
+            fn retrieve(
+                &self,
+                uri: &Uri<&str>,
+            ) -> Result<Value, Box<dyn std::error::Error + Send + Sync>> {
+                match uri.as_str() {
+                    "foo://schema_2.json" => Ok(json!({
+                        "$id": "foo://schema_2.json",
+                        "type": "string"
+                    })),
+                    _ => panic!("Unexpected URI: {}", uri.path()),
+                }
+            }
+        }
+
+        let validator = match crate::options()
+            .with_retriever(NestedRetrieve)
+            .build(&schema)
+        {
+            Ok(validator) => validator,
+            Err(error) => panic!("Failed to build validator: {}", error),
+        };
+
+        assert!(validator.is_valid(&json!("test")));
+        assert!(!validator.is_valid(&json!(42)));
+    }
 }