From e1f31376b7a43fe777104f8e57296a6a8605dfb1 Mon Sep 17 00:00:00 2001
From: Julian Berman <Julian@GrayVines.com>
Date: Fri, 31 Jan 2025 11:01:29 -0500
Subject: [PATCH] Cope with the suite permuting of remotes.

This new-but-similarly-messy logic comes modified from Bowtie
(which itself may need to be modified there, we'll see).

`jsonschema_suite remotes` is somewhat semantically broken (as it
doesn't really help with not emitting schemas which aren't valid under
the version under test), so it's time to give up and use the other
filesystem mechanism of walking the remotes directory.

Refs: json-schema-org/JSON-Schema-Test-Suite#753
---
 jsonschema/tests/_suite.py | 75 +++++++++++++++++++++-----------------
 1 file changed, 42 insertions(+), 33 deletions(-)

diff --git a/jsonschema/tests/_suite.py b/jsonschema/tests/_suite.py
index 0da6503c..d61d3827 100644
--- a/jsonschema/tests/_suite.py
+++ b/jsonschema/tests/_suite.py
@@ -10,7 +10,6 @@
 import json
 import os
 import re
-import subprocess
 import sys
 import unittest
 
@@ -21,11 +20,14 @@
 if TYPE_CHECKING:
     from collections.abc import Iterable, Mapping, Sequence
 
+    from referencing.jsonschema import Schema
     import pyperf
 
 from jsonschema.validators import _VALIDATORS
 import jsonschema
 
+MAGIC_REMOTE_URL = "http://localhost:1234"
+
 _DELIMITERS = re.compile(r"[\W\- ]+")
 
 
@@ -51,38 +53,7 @@ def _find_suite():
 class Suite:
 
     _root: Path = field(factory=_find_suite)
-    _remotes: referencing.jsonschema.SchemaRegistry = field(init=False)
-
-    def __attrs_post_init__(self):
-        jsonschema_suite = self._root.joinpath("bin", "jsonschema_suite")
-        argv = [sys.executable, str(jsonschema_suite), "remotes"]
-        remotes = subprocess.check_output(argv).decode("utf-8")
-
-        resources = json.loads(remotes)
 
-        li = "http://localhost:1234/locationIndependentIdentifierPre2019.json"
-        li4 = "http://localhost:1234/locationIndependentIdentifierDraft4.json"
-
-        registry = Registry().with_resources(
-            [
-                (
-                    li,
-                    referencing.jsonschema.DRAFT7.create_resource(
-                        contents=resources.pop(li),
-                    ),
-                ),
-                (
-                    li4,
-                    referencing.jsonschema.DRAFT4.create_resource(
-                        contents=resources.pop(li4),
-                    ),
-                ),
-            ],
-        ).with_contents(
-            resources.items(),
-            default_specification=referencing.jsonschema.DRAFT202012,
-        )
-        object.__setattr__(self, "_remotes", registry)
 
     def benchmark(self, runner: pyperf.Runner):  # pragma: no cover
         for name, Validator in _VALIDATORS.items():
@@ -92,10 +63,18 @@ def benchmark(self, runner: pyperf.Runner):  # pragma: no cover
             )
 
     def version(self, name) -> Version:
+        Validator = _VALIDATORS[name]
+        uri: str = Validator.ID_OF(Validator.META_SCHEMA)  # type: ignore[assignment]
+        specification = referencing.jsonschema.specification_with(uri)
+
+        registry = Registry().with_contents(
+            remotes_in(root=self._root / "remotes", name=name, uri=uri),
+            default_specification=specification,
+        )
         return Version(
             name=name,
             path=self._root / "tests" / name,
-            remotes=self._remotes,
+            remotes=registry,
         )
 
 
@@ -187,6 +166,36 @@ def benchmark(self, runner: pyperf.Runner, **kwargs):  # pragma: no cover
             )
 
 
+def remotes_in(
+    root: Path,
+    name: str,
+    uri: str,
+) -> Iterable[tuple[str, Schema]]:
+    # This messy logic is because the test suite is terrible at indicating
+    # what remotes are needed for what drafts, and mixes in schemas which
+    # have no $schema and which are invalid under earlier versions, in with
+    # other schemas which are needed for tests.
+
+    for each in root.rglob("*.json"):
+        schema = json.loads(each.read_text())
+
+        relative = str(each.relative_to(root)).replace("\\", "/")
+
+        if (
+            ( # invalid boolean schema
+                name in {"draft3", "draft4"}
+                and each.stem == "tree"
+            ) or
+            (  # draft<NotThisDialect>/*.json
+                "$schema" not in schema
+                and relative.startswith("draft")
+                and not relative.startswith(name)
+            )
+        ):
+            continue
+        yield f"{MAGIC_REMOTE_URL}/{relative}", schema
+
+
 @frozen(repr=False)
 class _Test: