Merge pull request 'schemagen: fix nullable root types (#13)' (#29) from weaselbot/weaseljson:weaselbot/issue-13 into main

Reviewed-on: weaselab/weaseljson#29

contrib/schemagen/CMakeLists.txt

@@ -46,6 +46,49 @@ target_link_libraries(schemagen_big PRIVATE ${PROJECT_NAME})
 target_compile_options(schemagen_big PRIVATE -Wno-switch-enum)
 add_dependencies(schemagen_big schemagen_big_h)
 
+set(NULLABLE_OBJECT_SCHEMA
+    ${CMAKE_CURRENT_SOURCE_DIR}/nullable_object.schema.json)
+set(NULLABLE_STRING_SCHEMA
+    ${CMAKE_CURRENT_SOURCE_DIR}/nullable_string.schema.json)
+set(NULLABLE_ARRAY_SCHEMA
+    ${CMAKE_CURRENT_SOURCE_DIR}/nullable_array.schema.json)
+set(NULLABLE_OBJECT_H ${CMAKE_CURRENT_BINARY_DIR}/nullable_object.h)
+set(NULLABLE_STRING_H ${CMAKE_CURRENT_BINARY_DIR}/nullable_string.h)
+set(NULLABLE_ARRAY_H ${CMAKE_CURRENT_BINARY_DIR}/nullable_array.h)
+
+add_custom_command(
+  OUTPUT ${NULLABLE_OBJECT_H}
+  COMMAND ${Python3_EXECUTABLE} ${SCHEMAGEN_SCRIPT} ${NULLABLE_OBJECT_SCHEMA} -o
+          ${NULLABLE_OBJECT_H} --namespace nullable_object
+  DEPENDS ${SCHEMAGEN_SCRIPT} ${NULLABLE_OBJECT_SCHEMA}
+  COMMENT "Generating nullable_object.h")
+
+add_custom_command(
+  OUTPUT ${NULLABLE_STRING_H}
+  COMMAND ${Python3_EXECUTABLE} ${SCHEMAGEN_SCRIPT} ${NULLABLE_STRING_SCHEMA} -o
+          ${NULLABLE_STRING_H} --namespace nullable_string
+  DEPENDS ${SCHEMAGEN_SCRIPT} ${NULLABLE_STRING_SCHEMA}
+  COMMENT "Generating nullable_string.h")
+
+add_custom_command(
+  OUTPUT ${NULLABLE_ARRAY_H}
+  COMMAND ${Python3_EXECUTABLE} ${SCHEMAGEN_SCRIPT} ${NULLABLE_ARRAY_SCHEMA} -o
+          ${NULLABLE_ARRAY_H} --namespace nullable_array
+  DEPENDS ${SCHEMAGEN_SCRIPT} ${NULLABLE_ARRAY_SCHEMA}
+  COMMENT "Generating nullable_array.h")
+
+add_custom_target(
+  schemagen_nullable_h DEPENDS ${NULLABLE_OBJECT_H} ${NULLABLE_STRING_H}
+                               ${NULLABLE_ARRAY_H})
+
+add_executable(schemagen_nullable_root
+               ${CMAKE_CURRENT_SOURCE_DIR}/test_nullable_root.cpp)
+target_include_directories(schemagen_nullable_root
+                           PRIVATE include ${CMAKE_CURRENT_BINARY_DIR})
+target_link_libraries(schemagen_nullable_root PRIVATE ${PROJECT_NAME})
+target_compile_options(schemagen_nullable_root PRIVATE -Wno-switch-enum)
+add_dependencies(schemagen_nullable_root schemagen_nullable_h)
+
 add_test(
   NAME schemagen_example
   COMMAND schemagen_example
@@ -55,3 +98,8 @@ add_test(
   NAME schemagen_big
   COMMAND schemagen_big
   WORKING_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR})
+
+add_test(
+  NAME schemagen_nullable_root
+  COMMAND schemagen_nullable_root
+  WORKING_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR})

contrib/schemagen/nullable_array.schema.json

@@ -0,0 +1,9 @@
+{
+  "type": [
+    "array",
+    "null"
+  ],
+  "items": {
+    "type": "integer"
+  }
+}

contrib/schemagen/nullable_object.schema.json

@@ -0,0 +1,15 @@
+{
+  "type": [
+    "object",
+    "null"
+  ],
+  "additionalProperties": false,
+  "required": [
+    "x"
+  ],
+  "properties": {
+    "x": {
+      "type": "string"
+    }
+  }
+}

contrib/schemagen/nullable_string.schema.json

@@ -0,0 +1,6 @@
+{
+  "type": [
+    "string",
+    "null"
+  ]
+}

contrib/schemagen/test_nullable_root.cpp

@@ -0,0 +1,158 @@
+// Regression test for issue #13: nullable root types.
+#include <cassert>
+#include <cstdio>
+#include <string>
+
+#include "nullable_array.h"
+#include "nullable_object.h"
+#include "nullable_string.h"
+
+static int failures = 0;
+#define CHECK(cond)                                                            \
+  do {                                                                         \
+    if (!(cond)) {                                                             \
+      printf("FAIL %s:%d: %s\n", __FILE__, __LINE__, #cond);                   \
+      ++failures;                                                              \
+    }                                                                          \
+  } while (0)
+
+static WeaselJsonStatus parseStrided(nullable_object::RootBuilder &b,
+                                     std::string in) {
+  for (size_t i = 0; i < in.size(); ++i) {
+    char c = in[i];
+    WeaselJsonStatus s = b.feed(&c, 1);
+    if (s != WeaselJson_AGAIN)
+      return s;
+  }
+  return b.finish();
+}
+
+static WeaselJsonStatus parseStrided(nullable_string::RootBuilder &b,
+                                     std::string in) {
+  for (size_t i = 0; i < in.size(); ++i) {
+    char c = in[i];
+    WeaselJsonStatus s = b.feed(&c, 1);
+    if (s != WeaselJson_AGAIN)
+      return s;
+  }
+  return b.finish();
+}
+
+static WeaselJsonStatus parseStrided(nullable_array::RootBuilder &b,
+                                     std::string in) {
+  for (size_t i = 0; i < in.size(); ++i) {
+    char c = in[i];
+    WeaselJsonStatus s = b.feed(&c, 1);
+    if (s != WeaselJson_AGAIN)
+      return s;
+  }
+  return b.finish();
+}
+
+static void expectReject(nullable_object::RootBuilder &b, std::string in,
+                         const char *what) {
+  WeaselJsonStatus s = parseStrided(b, in);
+  if (s == WeaselJson_REJECT) {
+    printf("ok   reject: %s\n", what);
+  } else {
+    printf("FAIL expected reject (%s) got status %d for: %s\n", what, s,
+           in.c_str());
+    ++failures;
+  }
+}
+
+int main() {
+  // ---- nullable root object: valid document ----
+  {
+    nullable_object::RootBuilder b;
+    WeaselJsonStatus s = parseStrided(b, R"({"x":"hello"})");
+    CHECK(s == WeaselJson_OK);
+    if (s == WeaselJson_OK) {
+      nullable_object::Root r = b.take();
+      CHECK(r.has_value());
+      CHECK(r->x == "hello");
+      printf("ok   nullable root object accepts object\n");
+    }
+  }
+
+  // ---- nullable root object: null document ----
+  {
+    nullable_object::RootBuilder b;
+    WeaselJsonStatus s = parseStrided(b, "null");
+    CHECK(s == WeaselJson_OK);
+    if (s == WeaselJson_OK) {
+      nullable_object::Root r = b.take();
+      CHECK(!r.has_value());
+      printf("ok   nullable root object accepts null\n");
+    }
+  }
+
+  // ---- nullable root object: schema checks still run ----
+  {
+    nullable_object::RootBuilder b;
+    expectReject(b, R"({"x":"hello","extra":1})",
+                 "unknown key in strict nullable root object");
+  }
+  {
+    nullable_object::RootBuilder b;
+    expectReject(b, R"({})", "missing required field in nullable root object");
+  }
+
+  // ---- nullable root string: valid value ----
+  {
+    nullable_string::RootBuilder b;
+    WeaselJsonStatus s = parseStrided(b, R"("hello")");
+    CHECK(s == WeaselJson_OK);
+    if (s == WeaselJson_OK) {
+      nullable_string::Root r = b.take();
+      CHECK(r.has_value());
+      CHECK(*r == "hello");
+      printf("ok   nullable root string accepts string\n");
+    }
+  }
+
+  // ---- nullable root string: null value ----
+  {
+    nullable_string::RootBuilder b;
+    WeaselJsonStatus s = parseStrided(b, "null");
+    CHECK(s == WeaselJson_OK);
+    if (s == WeaselJson_OK) {
+      nullable_string::Root r = b.take();
+      CHECK(!r.has_value());
+      printf("ok   nullable root string accepts null\n");
+    }
+  }
+
+  // ---- nullable root array: valid value ----
+  {
+    nullable_array::RootBuilder b;
+    WeaselJsonStatus s = parseStrided(b, "[1,2,3]");
+    CHECK(s == WeaselJson_OK);
+    if (s == WeaselJson_OK) {
+      nullable_array::Root r = b.take();
+      CHECK(r.has_value());
+      CHECK(r->size() == 3);
+      CHECK((*r)[0] == 1 && (*r)[1] == 2 && (*r)[2] == 3);
+      printf("ok   nullable root array accepts array\n");
+    }
+  }
+
+  // ---- nullable root array: null value ----
+  {
+    nullable_array::RootBuilder b;
+    WeaselJsonStatus s = parseStrided(b, "null");
+    CHECK(s == WeaselJson_OK);
+    if (s == WeaselJson_OK) {
+      nullable_array::Root r = b.take();
+      CHECK(!r.has_value());
+      printf("ok   nullable root array accepts null\n");
+    }
+  }
+
+  if (failures == 0) {
+    printf("\nALL TESTS PASSED\n");
+    return 0;
+  }
+  printf("\n%d FAILURE(S)\n", failures);
+  return 1;
+}

contrib/schemagen/weaseljson_schemagen.py

@@ -493,6 +493,18 @@ class Emitter:
         self.root_ty, self.root_nullable = Builder._unpack(
             self.b.build_type(self.b.root_schema, "Root")
         )
+
+        # A nullable root object would otherwise produce
+        #   using Root = std::optional<Root>;
+        # which conflicts with the struct named Root.  Rename the inner struct.
+        if isinstance(self.root_ty, TObj) and self.root_nullable:
+            old_name = self.root_ty.name
+            new_name = self.b.unique_name("RootInner")
+            obj = self.b.objects.pop(old_name)
+            obj.name = new_name
+            self.b.objects[new_name] = obj
+            self.root_ty = TObj(new_name)
+
         break_cycles(self.b.objects)
 
         # register all array kinds (walk every field + root)
@@ -582,9 +594,7 @@ namespace {ns} {{"""
     def _kind_enum(self):
         kinds = list(self.b.objects.keys())
         kinds += [n for n, _ in self.arr_types]
-        root_is_container = (
+        root_is_container = isinstance(self.root_ty, (TObj, TArr))
-            isinstance(self.root_ty, (TObj, TArr)) and not self.root_nullable
-        )
         if not root_is_container:
             kinds.append("RootScalar")
         kinds.append("Skip")
@@ -593,9 +603,9 @@ namespace {ns} {{"""
 
     def _root_info(self):
         """Return (root_cat, root_container_kind_or_None)."""
-        if isinstance(self.root_ty, TObj) and not self.root_nullable:
+        if isinstance(self.root_ty, TObj):
             return ("Obj", f"Kind::{self.root_ty.name}")
-        if isinstance(self.root_ty, TArr) and not self.root_nullable:
+        if isinstance(self.root_ty, TArr):
             return ("Arr", f"Kind::{self.arr_kind(self.root_ty)}")
         return (self.cat(self.root_ty), None)
 
@@ -665,6 +675,11 @@ namespace {ns} {{"""
             lines.append("    }")
         root_cat, root_kind = self._root_info()
         if root_kind is None:
+            if self.root_nullable:
+                lines.append(
+                    "    case Kind::RootScalar: { if (!result_) result_.emplace(); return &*result_; }"
+                )
+            else:
                 lines.append("    case Kind::RootScalar: return &result_;")
         lines.append("    default: return nullptr;")
         lines.append("    }")
@@ -807,7 +822,12 @@ namespace {ns} {{"""
         lines = ["    if (error_) return;"]
         if root_kind is not None and root_cat == event_cat:
             lines.append("    if (stack_.empty()) {")
-            lines.append(f"      stack_.push_back(Frame{{{root_kind}, &result_}});")
+            if self.root_nullable:
+                lines.append("      result_.emplace();")
+                dest = "&*result_"
+            else:
+                dest = "&result_"
+            lines.append(f"      stack_.push_back(Frame{{{root_kind}, {dest}}});")
             if event_cat == "Obj":
                 lines.append("      {")
                 lines.append(f"        int n = fieldCount({root_kind});")
@@ -839,6 +859,9 @@ namespace {ns} {{"""
         root_cat, root_kind = self._root_info()
         begin_obj = self._begin_container("Obj", root_kind)
         begin_arr = self._begin_container("Arr", root_kind)
+        null_at_root = (
+            "done_ = true; return;" if self.root_nullable else "reject(); return;"
+        )
 
         return f"""
 class RootBuilder {{
@@ -1065,6 +1088,9 @@ private:
 
   void cbNull() {{
     if (error_) return;
+    if (stack_.empty()) {{
+      {null_at_root}
+    }}
     Frame &f = stack_.back();
     if (f.kind == Kind::Skip) return;
     SlotInfo si = slotInfoG(f);