Type every node field and mark on-error-only types explicitly

* For Loader.java, do not deserialize the AST if there are errors, so then Java nodes only have non-error types for fields.

rakelib/typecheck.rake

@@ -59,6 +59,7 @@ namespace :typecheck do
       --ignore=test/
       --ignore=rakelib/
       --ignore=Rakefile
+      --ignore=top-100-gems/
       # Treat all files as "typed: true" by default
       --typed=true
       # Use the typed-override file to revert some files to "typed: false"

rust/ruby-prism/build.rs

@@ -47,11 +47,27 @@ enum NodeFieldType {
     Double,
 }
 
+#[derive(Debug, Deserialize)]
+#[allow(dead_code)]
+struct OnErrorType {
+    #[serde(rename = "on error")]
+    kind: String,
+}
+
+#[derive(Debug, Deserialize)]
+#[serde(untagged)]
+#[allow(dead_code)]
+enum UnionKind {
+    OnSuccess(String),
+    OnError(OnErrorType),
+}
+
 #[derive(Debug, Deserialize)]
 #[serde(untagged)]
+#[allow(dead_code)]
 enum NodeFieldKind {
     Concrete(String),
-    Union(Vec<String>),
+    Union(Vec<UnionKind>),
 }
 
 #[derive(Debug, Deserialize)]
@@ -126,6 +142,13 @@ fn struct_name(name: &str) -> String {
     result
 }
 
+fn kind_to_type(kind: &String) -> String {
+    match kind.as_str() {
+        "non-void expression" | "pattern expression" | "Node" => String::new(),
+        _ => kind.to_string(),
+    }
+}
+
 /// Returns the name of the C type from the given node name.
 fn type_name(name: &str) -> String {
     let mut result = String::with_capacity(8 + name.len());
@@ -263,30 +286,34 @@ fn write_node(file: &mut File, flags: &[Flags], node: &Node) -> Result<(), Box<d
         writeln!(file, "    #[must_use]")?;
 
         match field.field_type {
-            NodeFieldType::Node => {
-                if let Some(NodeFieldKind::Concrete(kind)) = &field.kind {
+            NodeFieldType::Node => match &field.kind {
+                Some(NodeFieldKind::Concrete(raw_kind)) if !kind_to_type(raw_kind).is_empty() => {
+                    let kind = kind_to_type(raw_kind);
                     writeln!(file, "    pub fn {}(&self) -> {}<'pr> {{", field.name, kind)?;
-                    writeln!(file, "        let node: *mut pm{}_t = unsafe {{ (*self.pointer).{} }};", struct_name(kind), field.name)?;
+                    writeln!(file, "        let node: *mut pm{}_t = unsafe {{ (*self.pointer).{} }};", struct_name(&kind), field.name)?;
                     writeln!(file, "        {} {{ parser: self.parser, pointer: node, marker: PhantomData }}", kind)?;
                     writeln!(file, "    }}")?;
-                } else {
+                },
+                _ => {
                     writeln!(file, "    pub fn {}(&self) -> Node<'pr> {{", field.name)?;
                     writeln!(file, "        let node: *mut pm_node_t = unsafe {{ (*self.pointer).{} }};", field.name)?;
                     writeln!(file, "        Node::new(self.parser, node)")?;
                     writeln!(file, "    }}")?;
-                }
+                },
             },
-            NodeFieldType::OptionalNode => {
-                if let Some(NodeFieldKind::Concrete(kind)) = &field.kind {
+            NodeFieldType::OptionalNode => match &field.kind {
+                Some(NodeFieldKind::Concrete(raw_kind)) if !kind_to_type(raw_kind).is_empty() => {
+                    let kind = kind_to_type(raw_kind);
                     writeln!(file, "    pub fn {}(&self) -> Option<{}<'pr>> {{", field.name, kind)?;
-                    writeln!(file, "        let node: *mut pm{}_t = unsafe {{ (*self.pointer).{} }};", struct_name(kind), field.name)?;
+                    writeln!(file, "        let node: *mut pm{}_t = unsafe {{ (*self.pointer).{} }};", struct_name(&kind), field.name)?;
                     writeln!(file, "        if node.is_null() {{")?;
                     writeln!(file, "            None")?;
                     writeln!(file, "        }} else {{")?;
                     writeln!(file, "            Some({} {{ parser: self.parser, pointer: node, marker: PhantomData }})", kind)?;
                     writeln!(file, "        }}")?;
                     writeln!(file, "    }}")?;
-                } else {
+                },
+                _ => {
                     writeln!(file, "    pub fn {}(&self) -> Option<Node<'pr>> {{", field.name)?;
                     writeln!(file, "        let node: *mut pm_node_t = unsafe {{ (*self.pointer).{} }};", field.name)?;
                     writeln!(file, "        if node.is_null() {{")?;
@@ -295,7 +322,7 @@ fn write_node(file: &mut File, flags: &[Flags], node: &Node) -> Result<(), Box<d
                     writeln!(file, "            Some(Node::new(self.parser, node))")?;
                     writeln!(file, "        }}")?;
                     writeln!(file, "    }}")?;
-                }
+                },
             },
             NodeFieldType::NodeList => {
                 writeln!(file, "    pub fn {}(&self) -> NodeList<'pr> {{", field.name)?;
@@ -473,16 +500,18 @@ fn write_visit(file: &mut File, config: &Config) -> Result<(), Box<dyn std::erro
             for field in &node.fields {
                 match field.field_type {
                     NodeFieldType::Node => {
-                        if let Some(NodeFieldKind::Concrete(kind)) = &field.kind {
-                            writeln!(file, "    visitor.visit{}(&node.{}());", struct_name(kind), field.name)?;
+                        if let Some(NodeFieldKind::Concrete(raw_kind)) = &field.kind {
+                            let kind = kind_to_type(raw_kind);
+                            writeln!(file, "    visitor.visit{}(&node.{}());", struct_name(&kind), field.name)?;
                         } else {
                             writeln!(file, "    visitor.visit(&node.{}());", field.name)?;
                         }
                     },
                     NodeFieldType::OptionalNode => {
-                        if let Some(NodeFieldKind::Concrete(kind)) = &field.kind {
+                        if let Some(NodeFieldKind::Concrete(raw_kind)) = &field.kind {
+                            let kind = kind_to_type(raw_kind);
                             writeln!(file, "    if let Some(node) = node.{}() {{", field.name)?;
-                            writeln!(file, "        visitor.visit{}(&node);", struct_name(kind))?;
+                            writeln!(file, "        visitor.visit{}(&node);", struct_name(&kind))?;
                             writeln!(file, "    }}")?;
                         } else {
                             writeln!(file, "    if let Some(node) = node.{}() {{", field.name)?;
@@ -762,7 +791,7 @@ impl TryInto<i32> for Integer<'_> {{
         let length = unsafe {{ (*self.pointer).length }};
 
         if length == 0 {{
-            i32::try_from(unsafe {{ (*self.pointer).value }}).map_or(Err(()), |value| 
+            i32::try_from(unsafe {{ (*self.pointer).value }}).map_or(Err(()), |value|
                 if negative {{
                     Ok(-value)
                 }} else {{

src/prism.c

@@ -2985,6 +2985,7 @@ pm_index_and_write_node_create(pm_parser_t *parser, pm_call_node_t *target, cons
 
     pm_index_arguments_check(parser, target->arguments, target->block);
 
+    assert(!target->block || PM_NODE_TYPE_P(target->block, PM_BLOCK_ARGUMENT_NODE));
     *node = (pm_index_and_write_node_t) {
         {
             .type = PM_INDEX_AND_WRITE_NODE,
@@ -3000,7 +3001,7 @@ pm_index_and_write_node_create(pm_parser_t *parser, pm_call_node_t *target, cons
         .opening_loc = target->opening_loc,
         .arguments = target->arguments,
         .closing_loc = target->closing_loc,
-        .block = target->block,
+        .block = (pm_block_argument_node_t *) target->block,
         .operator_loc = PM_LOCATION_TOKEN_VALUE(operator),
         .value = value
     };
@@ -3060,6 +3061,7 @@ pm_index_operator_write_node_create(pm_parser_t *parser, pm_call_node_t *target,
 
     pm_index_arguments_check(parser, target->arguments, target->block);
 
+    assert(!target->block || PM_NODE_TYPE_P(target->block, PM_BLOCK_ARGUMENT_NODE));
     *node = (pm_index_operator_write_node_t) {
         {
             .type = PM_INDEX_OPERATOR_WRITE_NODE,
@@ -3075,7 +3077,7 @@ pm_index_operator_write_node_create(pm_parser_t *parser, pm_call_node_t *target,
         .opening_loc = target->opening_loc,
         .arguments = target->arguments,
         .closing_loc = target->closing_loc,
-        .block = target->block,
+        .block = (pm_block_argument_node_t *) target->block,
         .binary_operator = pm_parser_constant_id_location(parser, operator->start, operator->end - 1),
         .binary_operator_loc = PM_LOCATION_TOKEN_VALUE(operator),
         .value = value
@@ -3137,6 +3139,7 @@ pm_index_or_write_node_create(pm_parser_t *parser, pm_call_node_t *target, const
 
     pm_index_arguments_check(parser, target->arguments, target->block);
 
+    assert(!target->block || PM_NODE_TYPE_P(target->block, PM_BLOCK_ARGUMENT_NODE));
     *node = (pm_index_or_write_node_t) {
         {
             .type = PM_INDEX_OR_WRITE_NODE,
@@ -3152,7 +3155,7 @@ pm_index_or_write_node_create(pm_parser_t *parser, pm_call_node_t *target, const
         .opening_loc = target->opening_loc,
         .arguments = target->arguments,
         .closing_loc = target->closing_loc,
-        .block = target->block,
+        .block = (pm_block_argument_node_t *) target->block,
         .operator_loc = PM_LOCATION_TOKEN_VALUE(operator),
         .value = value
     };
@@ -3205,6 +3208,7 @@ pm_index_target_node_create(pm_parser_t *parser, pm_call_node_t *target) {
 
     pm_index_arguments_check(parser, target->arguments, target->block);
 
+    assert(!target->block || PM_NODE_TYPE_P(target->block, PM_BLOCK_ARGUMENT_NODE));
     *node = (pm_index_target_node_t) {
         {
             .type = PM_INDEX_TARGET_NODE,
@@ -3216,7 +3220,7 @@ pm_index_target_node_create(pm_parser_t *parser, pm_call_node_t *target) {
         .opening_loc = target->opening_loc,
         .arguments = target->arguments,
         .closing_loc = target->closing_loc,
-        .block = target->block
+        .block = (pm_block_argument_node_t *) target->block,
     };
 
     // Here we're going to free the target, since it is no longer necessary.
@@ -3231,7 +3235,7 @@ pm_index_target_node_create(pm_parser_t *parser, pm_call_node_t *target) {
  * Allocate and initialize a new CapturePatternNode node.
  */
 static pm_capture_pattern_node_t *
-pm_capture_pattern_node_create(pm_parser_t *parser, pm_node_t *value, pm_node_t *target, const pm_token_t *operator) {
+pm_capture_pattern_node_create(pm_parser_t *parser, pm_node_t *value, pm_local_variable_target_node_t *target, const pm_token_t *operator) {
     pm_capture_pattern_node_t *node = PM_NODE_ALLOC(parser, pm_capture_pattern_node_t);
 
     *node = (pm_capture_pattern_node_t) {
@@ -3240,7 +3244,7 @@ pm_capture_pattern_node_create(pm_parser_t *parser, pm_node_t *value, pm_node_t
             .node_id = PM_NODE_IDENTIFY(parser),
             .location = {
                 .start = value->location.start,
-                .end = target->location.end
+                .end = target->base.location.end
             },
         },
         .value = value,
@@ -4032,14 +4036,25 @@ pm_find_pattern_node_create(pm_parser_t *parser, pm_node_list_t *nodes) {
     pm_find_pattern_node_t *node = PM_NODE_ALLOC(parser, pm_find_pattern_node_t);
 
     pm_node_t *left = nodes->nodes[0];
+    assert(PM_NODE_TYPE_P(left, PM_SPLAT_NODE));
+    pm_splat_node_t *left_splat_node = (pm_splat_node_t *) left;
+
     pm_node_t *right;
 
     if (nodes->size == 1) {
         right = (pm_node_t *) pm_missing_node_create(parser, left->location.end, left->location.end);
     } else {
         right = nodes->nodes[nodes->size - 1];
+        assert(PM_NODE_TYPE_P(right, PM_SPLAT_NODE));
     }
 
+#if PRISM_SERIALIZE_ONLY_SEMANTICS_FIELDS
+    // FindPatternNode#right is typed as SplatNode in this case, so replace the potential MissingNode with a SplatNode.
+    // The resulting AST will anyway be ignored, but this file still needs to compile.
+    pm_splat_node_t *right_splat_node = PM_NODE_TYPE_P(right, PM_SPLAT_NODE) ? (pm_splat_node_t *) right : left_splat_node;
+#else
+    pm_node_t *right_splat_node = right;
+#endif
     *node = (pm_find_pattern_node_t) {
         {
             .type = PM_FIND_PATTERN_NODE,
@@ -4050,8 +4065,8 @@ pm_find_pattern_node_create(pm_parser_t *parser, pm_node_list_t *nodes) {
             },
         },
         .constant = NULL,
-        .left = left,
-        .right = right,
+        .left = left_splat_node,
+        .right = right_splat_node,
         .requireds = { 0 },
         .opening_loc = PM_OPTIONAL_LOCATION_NOT_PROVIDED_VALUE,
         .closing_loc = PM_OPTIONAL_LOCATION_NOT_PROVIDED_VALUE
@@ -17401,7 +17416,7 @@ parse_pattern_primitives(pm_parser_t *parser, pm_constant_id_list_t *captures, p
         }
 
         parse_pattern_capture(parser, captures, constant_id, &PM_LOCATION_TOKEN_VALUE(&parser->previous));
-        pm_node_t *target = (pm_node_t *) pm_local_variable_target_node_create(
+        pm_local_variable_target_node_t *target = pm_local_variable_target_node_create(
             parser,
             &PM_LOCATION_TOKEN_VALUE(&parser->previous),
             constant_id,

templates/include/prism/ast.h.erb

@@ -218,6 +218,6 @@ typedef enum pm_<%= flag.human %> {
  * to specify that through the environment. It will never be true except for in
  * those build systems.
  */
-#define PRISM_SERIALIZE_ONLY_SEMANTICS_FIELDS <%= Prism::Template::SERIALIZE_ONLY_SEMANTICS_FIELDS %>
+#define PRISM_SERIALIZE_ONLY_SEMANTICS_FIELDS <%= Prism::Template::SERIALIZE_ONLY_SEMANTICS_FIELDS ? 1 : 0 %>
 
 #endif

templates/java/org/prism/Loader.java.erb

@@ -127,16 +127,21 @@ public class Loader {
         int constantPoolLength = loadVarUInt();
         this.constantPool = new ConstantPool(this, source.bytes, constantPoolBufferOffset, constantPoolLength);
 
-        Nodes.Node node = loadNode();
+        Nodes.Node node;
+        if (errors.length == 0) {
+            node = loadNode();
 
-        int left = constantPoolBufferOffset - buffer.position();
-        if (left != 0) {
-            throw new Error("Expected to consume all bytes while deserializing but there were " + left + " bytes left");
-        }
+            int left = constantPoolBufferOffset - buffer.position();
+            if (left != 0) {
+                throw new Error("Expected to consume all bytes while deserializing but there were " + left + " bytes left");
+            }
 
-        boolean[] newlineMarked = new boolean[1 + source.getLineCount()];
-        MarkNewlinesVisitor visitor = new MarkNewlinesVisitor(source, newlineMarked);
-        node.accept(visitor);
+            boolean[] newlineMarked = new boolean[1 + source.getLineCount()];
+            MarkNewlinesVisitor visitor = new MarkNewlinesVisitor(source, newlineMarked);
+            node.accept(visitor);
+        } else {
+            node = null;
+        }
 
         return new ParseResult(node, magicComments, dataLocation, errors, warnings, source);
     }

templates/lib/prism/dsl.rb.erb

@@ -70,10 +70,10 @@ module Prism
     def <%= node.human %>(<%= ["source: default_source", "node_id: 0", "location: default_location", "flags: 0", *node.fields.map { |field|
       case field
       when Prism::Template::NodeField
-        if !field.kind?
+        kind = field.specific_kind || field.union_kind&.first
+        if kind.nil?
           "#{field.name}: default_node(source, location)"
         else
-          kind = field.specific_kind || field.union_kind.first
           "#{field.name}: #{kind.gsub(/(?<=.)[A-Z]/, "_\\0").downcase}(source: source)"
         end
       when Prism::Template::ConstantField

templates/lib/prism/node.rb.erb

@@ -225,7 +225,7 @@ module Prism
       @flags = flags
       <%- node.fields.each do |field| -%>
       <%- if Prism::Template::CHECK_FIELD_KIND && field.respond_to?(:check_field_kind) -%>
-      raise <%= field.name %>.inspect unless <%= field.check_field_kind %>
+      raise "<%= node.name %>#<%= field.name %> was of unexpected type:\n#{<%= field.name %>.inspect}" unless <%= field.check_field_kind %>
       <%- end -%>
       @<%= field.name %> = <%= field.name %>
       <%- end -%>

templates/rbi/prism/dsl.rbi.erb

@@ -14,10 +14,10 @@ module Prism::DSL
   ].concat(node.fields.map { |field|
     case field
     when Prism::Template::NodeField
-      if !field.kind?
+      kind = field.specific_kind || field.union_kind&.first
+      if kind.nil?
         [field.name, "default_node(source, location)", field.rbi_class]
       else
-        kind = field.specific_kind || field.union_kind.first
         [field.name, %Q{#{kind.gsub(/(?<=.)[A-Z]/, "_\\0").downcase}(source: source)}, field.rbi_class]
       end
     when Prism::Template::OptionalNodeField