prism 0.23.0 → 0.24.0

data/lib/prism/translation/parser/compiler.rb

@@ -1477,7 +1477,7 @@ module Prism
         # ^^^^^
         def visit_string_node(node)
           if node.opening&.start_with?("<<")
-            children, closing = visit_heredoc(InterpolatedStringNode.new(node.opening_loc, [node.copy(opening_loc: nil, closing_loc: nil, location: node.content_loc)], node.closing_loc, node.location))
+            children, closing = visit_heredoc(InterpolatedStringNode.new(node.send(:source), node.opening_loc, [node.copy(opening_loc: nil, closing_loc: nil, location: node.content_loc)], node.closing_loc, node.location))
             builder.string_compose(token(node.opening_loc), children, closing)
           elsif node.opening == "?"
             builder.character([node.unescaped, srange(node.location)])

data/lib/prism/translation/parser/rubocop.rb

@@ -9,18 +9,26 @@ require "prism/translation/parser"
 module Prism
   module Translation
     class Parser
-      # This is the special version number that should be used in rubocop
+      # This is the special version numbers that should be used in RuboCop
       # configuration files to trigger using prism.
+
+      # For Ruby 3.3
       VERSION_3_3 = 80_82_73_83_77.33
 
+      # For Ruby 3.4
+      VERSION_3_4 = 80_82_73_83_77.34
+
       # This module gets prepended into RuboCop::AST::ProcessedSource.
       module ProcessedSource
         # Redefine parser_class so that we can inject the prism parser into the
         # list of known parsers.
         def parser_class(ruby_version)
           if ruby_version == Prism::Translation::Parser::VERSION_3_3
-            require "prism/translation/parser"
-            Prism::Translation::Parser
+            require "prism/translation/parser33"
+            Prism::Translation::Parser33
+          elsif ruby_version == Prism::Translation::Parser::VERSION_3_4
+            require "prism/translation/parser34"
+            Prism::Translation::Parser34
           else
             super
           end

data/lib/prism/translation/parser.rb

@@ -43,7 +43,7 @@ module Prism
         source = source_buffer.source
 
         offset_cache = build_offset_cache(source)
-        result = unwrap(Prism.parse(source, filepath: source_buffer.name), offset_cache)
+        result = unwrap(Prism.parse(source, filepath: source_buffer.name, version: convert_for_prism(version)), offset_cache)
 
         build_ast(result.value, offset_cache)
       ensure
@@ -56,7 +56,7 @@ module Prism
         source = source_buffer.source
 
         offset_cache = build_offset_cache(source)
-        result = unwrap(Prism.parse(source, filepath: source_buffer.name), offset_cache)
+        result = unwrap(Prism.parse(source, filepath: source_buffer.name, version: convert_for_prism(version)), offset_cache)
 
         [
           build_ast(result.value, offset_cache),
@@ -75,7 +75,7 @@ module Prism
         offset_cache = build_offset_cache(source)
         result =
           begin
-            unwrap(Prism.parse_lex(source, filepath: source_buffer.name), offset_cache)
+            unwrap(Prism.parse_lex(source, filepath: source_buffer.name, version: convert_for_prism(version)), offset_cache)
           rescue ::Parser::SyntaxError
             raise if !recover
           end
@@ -168,6 +168,18 @@ module Prism
         )
       end
 
+      # Converts the version format handled by Parser to the format handled by Prism.
+      def convert_for_prism(version)
+        case version
+        when 33
+          "3.3.0"
+        when 34
+          "3.4.0"
+        else
+          "latest"
+        end
+      end
+
       require_relative "parser/compiler"
       require_relative "parser/lexer"
 

data/lib/prism/translation/parser33.rb

@@ -0,0 +1,12 @@
+require_relative "parser"
+
+module Prism
+  module Translation
+    # This class is the entry-point for Ruby 3.3 of `Prism::Translation::Parser`.
+    class Parser33 < Parser
+      def version # :nodoc:
+        33
+      end
+    end
+  end
+end

data/lib/prism/translation/parser34.rb

@@ -0,0 +1,12 @@
+require_relative "parser"
+
+module Prism
+  module Translation
+    # This class is the entry-point for Ruby 3.4 of `Prism::Translation::Parser`.
+    class Parser34 < Parser
+      def version # :nodoc:
+        34
+      end
+    end
+  end
+end

data/lib/prism/translation/ripper.rb

@@ -222,6 +222,44 @@ module Prism
         on_break(on_args_add_block(args_val, false))
       end
 
+      # Visit an AliasMethodNode.
+      def visit_alias_method_node(node)
+        # For both the old and new name, if there is a colon in the symbol
+        # name (e.g. 'alias :foo :bar') then we do *not* emit the [:symbol] wrapper around
+        # the lexer token (e.g. :@ident) inside [:symbol_literal]. But if there
+        # is no colon (e.g. 'alias foo bar') then we *do* still emit the [:symbol] wrapper.
+
+        if node.new_name.is_a?(SymbolNode) && !node.new_name.opening
+          new_name_val = visit_symbol_literal_node(node.new_name, no_symbol_wrapper: true)
+        else
+          new_name_val = visit(node.new_name)
+        end
+        if node.old_name.is_a?(SymbolNode) && !node.old_name.opening
+          old_name_val = visit_symbol_literal_node(node.old_name, no_symbol_wrapper: true)
+        else
+          old_name_val = visit(node.old_name)
+        end
+
+        on_alias(new_name_val, old_name_val)
+      end
+
+      # Visit an AliasGlobalVariableNode.
+      def visit_alias_global_variable_node(node)
+        on_var_alias(visit(node.new_name), visit(node.old_name))
+      end
+
+      # Visit a GlobalVariableReadNode.
+      def visit_global_variable_read_node(node)
+        bounds(node.location)
+        on_gvar(node.name.to_s)
+      end
+
+      # Visit a BackReferenceReadNode.
+      def visit_back_reference_read_node(node)
+        bounds(node.location)
+        on_backref(node.name.to_s)
+      end
+
       # Visit an AndNode.
       def visit_and_node(node)
         visit_binary_operator(node)
@@ -326,23 +364,7 @@ module Prism
 
       # Visit an InterpolatedStringNode node.
       def visit_interpolated_string_node(node)
-        parts = node.parts.map do |part|
-          case part
-          when StringNode
-            bounds(part.content_loc)
-            on_tstring_content(part.content)
-          when EmbeddedStatementsNode
-            on_string_embexpr(visit(part))
-          else
-            raise NotImplementedError, "Unexpected node type in InterpolatedStringNode"
-          end
-        end
-
-        string_list = parts.inject(on_string_content) do |items, item|
-          on_string_add(items, item)
-        end
-
-        on_string_literal(string_list)
+        on_string_literal(visit_enumerated_node(node))
       end
 
       # Visit an EmbeddedStatementsNode node.
@@ -352,15 +374,12 @@ module Prism
 
       # Visit a SymbolNode node.
       def visit_symbol_node(node)
-        if (opening = node.opening) && (['"', "'"].include?(opening[-1]) || opening.start_with?("%s"))
-          bounds(node.value_loc)
-          tstring_val = on_tstring_content(node.value.to_s)
-          return on_dyna_symbol(on_string_add(on_string_content, tstring_val))
-        end
+        visit_symbol_literal_node(node)
+      end
 
-        bounds(node.value_loc)
-        ident_val = on_ident(node.value.to_s)
-        on_symbol_literal(on_symbol(ident_val))
+      # Visit an InterpolatedSymbolNode node.
+      def visit_interpolated_symbol_node(node)
+        on_dyna_symbol(visit_enumerated_node(node))
       end
 
       # Visit a StatementsNode node.
@@ -459,6 +478,25 @@ module Prism
         end
       end
 
+      # Visit an InterpolatedStringNode or an InterpolatedSymbolNode node.
+      def visit_enumerated_node(node)
+        parts = node.parts.map do |part|
+          case part
+          when StringNode
+            bounds(part.content_loc)
+            on_tstring_content(part.content)
+          when EmbeddedStatementsNode
+            on_string_embexpr(visit(part))
+          else
+            raise NotImplementedError, "Unexpected node type in visit_enumerated_node"
+          end
+        end
+
+        parts.inject(on_string_content) do |items, item|
+          on_string_add(items, item)
+        end
+      end
+
       # Visit an operation-and-assign node, such as +=.
       def visit_binary_op_assign(node, operator: node.operator)
         bounds(node.name_loc)
@@ -487,6 +525,87 @@ module Prism
         on_assign(on_aref_field(visit(node.receiver), args_val), assign_val)
       end
 
+      # In an alias statement Ripper will emit @kw instead of @ident if the object
+      # being aliased is a Ruby keyword. For instance, in the line "alias :foo :if",
+      # the :if is treated as a lexer keyword. So we need to know what symbols are
+      # also keywords.
+      RUBY_KEYWORDS = [
+        "alias",
+        "and",
+        "begin",
+        "BEGIN",
+        "break",
+        "case",
+        "class",
+        "def",
+        "defined?",
+        "do",
+        "else",
+        "elsif",
+        "end",
+        "END",
+        "ensure",
+        "false",
+        "for",
+        "if",
+        "in",
+        "module",
+        "next",
+        "nil",
+        "not",
+        "or",
+        "redo",
+        "rescue",
+        "retry",
+        "return",
+        "self",
+        "super",
+        "then",
+        "true",
+        "undef",
+        "unless",
+        "until",
+        "when",
+        "while",
+        "yield",
+        "__ENCODING__",
+        "__FILE__",
+        "__LINE__",
+      ]
+
+      # Ripper has several methods of emitting a symbol literal. Inside an alias
+      # sometimes it suppresses the [:symbol] wrapper around ident. If the symbol
+      # is also the name of a keyword (e.g. :if) it will emit a :@kw wrapper, not
+      # an :@ident wrapper, with similar treatment for constants and operators.
+      def visit_symbol_literal_node(node, no_symbol_wrapper: false)
+        if (opening = node.opening) && (['"', "'"].include?(opening[-1]) || opening.start_with?("%s"))
+          bounds(node.value_loc)
+          str_val = node.value.to_s
+          if str_val == ""
+            return on_dyna_symbol(on_string_content)
+          else
+            tstring_val = on_tstring_content(str_val)
+            return on_dyna_symbol(on_string_add(on_string_content, tstring_val))
+          end
+        end
+
+        bounds(node.value_loc)
+        node_name = node.value.to_s
+        if RUBY_KEYWORDS.include?(node_name)
+          token_val = on_kw(node_name)
+        elsif node_name.length == 0
+          raise NotImplementedError
+        elsif /[[:upper:]]/.match(node_name[0])
+          token_val = on_const(node_name)
+        elsif /[[:punct:]]/.match(node_name[0])
+          token_val = on_op(node_name)
+        else
+          token_val = on_ident(node_name)
+        end
+        sym_val = no_symbol_wrapper ? token_val : on_symbol(token_val)
+        on_symbol_literal(sym_val)
+      end
+
       # Visit a node that represents a number. We need to explicitly handle the
       # unary - operator.
       def visit_number(node)

data/prism.gemspec

@@ -2,7 +2,7 @@
 
 Gem::Specification.new do |spec|
   spec.name = "prism"
-  spec.version = "0.23.0"
+  spec.version = "0.24.0"
   spec.authors = ["Shopify"]
   spec.email = ["ruby@shopify.com"]
 
@@ -87,6 +87,8 @@ Gem::Specification.new do |spec|
     "lib/prism/serialize.rb",
     "lib/prism/translation.rb",
     "lib/prism/translation/parser.rb",
+    "lib/prism/translation/parser33.rb",
+    "lib/prism/translation/parser34.rb",
     "lib/prism/translation/parser/compiler.rb",
     "lib/prism/translation/parser/lexer.rb",
     "lib/prism/translation/parser/rubocop.rb",

data/src/prettyprint.c

@@ -44,9 +44,9 @@ prettyprint_source(pm_buffer_t *output_buffer, const uint8_t *source, size_t len
 
 static inline void
 prettyprint_location(pm_buffer_t *output_buffer, const pm_parser_t *parser, const pm_location_t *location) {
-    pm_line_column_t start = pm_newline_list_line_column(&parser->newline_list, location->start);
-    pm_line_column_t end = pm_newline_list_line_column(&parser->newline_list, location->end);
-    pm_buffer_append_format(output_buffer, "(%lu,%lu)-(%lu,%lu)", (unsigned long) start.line, (unsigned long) start.column, (unsigned long) end.line, (unsigned long) end.column);
+    pm_line_column_t start = pm_newline_list_line_column(&parser->newline_list, location->start, parser->start_line);
+    pm_line_column_t end = pm_newline_list_line_column(&parser->newline_list, location->end, parser->start_line);
+    pm_buffer_append_format(output_buffer, "(%" PRIi32 ",%" PRIu32 ")-(%" PRIi32 ",%" PRIu32 ")", start.line, start.column, end.line, end.column);
 }
 
 static inline void

data/src/prism.c

@@ -51,6 +51,7 @@ debug_context(pm_context_t context) {
         case PM_CONTEXT_IF: return "IF";
         case PM_CONTEXT_MAIN: return "MAIN";
         case PM_CONTEXT_MODULE: return "MODULE";
+        case PM_CONTEXT_NONE: return "NONE";
         case PM_CONTEXT_PARENS: return "PARENS";
         case PM_CONTEXT_POSTEXE: return "POSTEXE";
         case PM_CONTEXT_PREDICATE: return "PREDICATE";
@@ -14205,7 +14206,6 @@ parse_strings(pm_parser_t *parser, pm_node_t *current) {
             // If we get here, then we have an end of a label immediately
             // after a start. In that case we'll create an empty symbol
             // node.
-            pm_token_t opening = not_provided(parser);
             pm_token_t content = parse_strings_empty_content(parser->previous.start);
             pm_symbol_node_t *symbol = pm_symbol_node_create(parser, &opening, &content, &parser->previous);
 
@@ -16769,7 +16769,18 @@ parse_assignment_values(pm_parser_t *parser, pm_binding_power_t previous_binding
     if (is_single_value && match1(parser, PM_TOKEN_KEYWORD_RESCUE_MODIFIER)) {
         pm_token_t rescue = parser->current;
         parser_lex(parser);
-        pm_node_t *right = parse_expression(parser, binding_power, false, PM_ERR_RESCUE_MODIFIER_VALUE);
+
+        bool accepts_command_call_inner = false;
+
+        // RHS can accept command call iff the value is a call with arguments but without paranthesis.
+        if (PM_NODE_TYPE_P(value, PM_CALL_NODE)) {
+            pm_call_node_t *call_node = (pm_call_node_t *)value;
+            if ((call_node->arguments != NULL) && (call_node->opening_loc.start == NULL)) {
+                accepts_command_call_inner = true;
+            }
+        }
+
+        pm_node_t *right = parse_expression(parser, binding_power, accepts_command_call_inner, PM_ERR_RESCUE_MODIFIER_VALUE);
 
         return (pm_node_t *) pm_rescue_modifier_node_create(parser, value, &rescue, right);
     }
@@ -18115,7 +18126,7 @@ typedef struct {
     pm_diagnostic_t *error;
 
     /** The start line of the diagnostic message. */
-    uint32_t line;
+    int32_t line;
 
     /** The column start of the diagnostic message. */
     uint32_t column_start;
@@ -18147,12 +18158,13 @@ typedef struct {
 #define PM_COLOR_RESET "\033[0m"
 
 static inline pm_error_t *
-pm_parser_errors_format_sort(const pm_list_t *error_list, const pm_newline_list_t *newline_list) {
+pm_parser_errors_format_sort(const pm_parser_t *parser, const pm_list_t *error_list, const pm_newline_list_t *newline_list) {
     pm_error_t *errors = calloc(error_list->size, sizeof(pm_error_t));
+    int32_t start_line = parser->start_line;
 
     for (pm_diagnostic_t *error = (pm_diagnostic_t *) error_list->head; error != NULL; error = (pm_diagnostic_t *) error->node.next) {
-        pm_line_column_t start = pm_newline_list_line_column(newline_list, error->location.start);
-        pm_line_column_t end = pm_newline_list_line_column(newline_list, error->location.end);
+        pm_line_column_t start = pm_newline_list_line_column(newline_list, error->location.start, start_line);
+        pm_line_column_t end = pm_newline_list_line_column(newline_list, error->location.end, start_line);
 
         // We're going to insert this error into the array in sorted order. We
         // do this by finding the first error that has a line number greater
@@ -18163,8 +18175,8 @@ pm_parser_errors_format_sort(const pm_list_t *error_list, const pm_newline_list_
             (index < error_list->size) &&
             (errors[index].error != NULL) &&
             (
-                (errors[index].line < ((uint32_t) start.line)) ||
-                (errors[index].line == ((uint32_t) start.line) && errors[index].column_start < ((uint32_t) start.column))
+                (errors[index].line < start.line) ||
+                ((errors[index].line == start.line) && (errors[index].column_start < start.column))
             )
         ) index++;
 
@@ -18177,18 +18189,18 @@ pm_parser_errors_format_sort(const pm_list_t *error_list, const pm_newline_list_
         // Finally, we'll insert the error into the array.
         uint32_t column_end;
         if (start.line == end.line) {
-            column_end = (uint32_t) end.column;
+            column_end = end.column;
         } else {
-            column_end = (uint32_t) (newline_list->offsets[start.line] - newline_list->offsets[start.line - 1] - 1);
+            column_end = (uint32_t) (newline_list->offsets[start.line - start_line + 1] - newline_list->offsets[start.line - start_line] - 1);
         }
 
         // Ensure we have at least one column of error.
-        if (((uint32_t) start.column) == column_end) column_end++;
+        if (start.column == column_end) column_end++;
 
         errors[index] = (pm_error_t) {
             .error = error,
-            .line = (uint32_t) start.line,
-            .column_start = (uint32_t) start.column,
+            .line = start.line,
+            .column_start = start.column,
             .column_end = column_end
         };
     }
@@ -18197,17 +18209,19 @@ pm_parser_errors_format_sort(const pm_list_t *error_list, const pm_newline_list_
 }
 
 static inline void
-pm_parser_errors_format_line(const pm_parser_t *parser, const pm_newline_list_t *newline_list, const char *number_prefix, size_t line, pm_buffer_t *buffer) {
-    const uint8_t *start = &parser->start[newline_list->offsets[line - 1]];
+pm_parser_errors_format_line(const pm_parser_t *parser, const pm_newline_list_t *newline_list, const char *number_prefix, int32_t line, pm_buffer_t *buffer) {
+    size_t index = (size_t) (line - parser->start_line);
+
+    const uint8_t *start = &parser->start[newline_list->offsets[index]];
     const uint8_t *end;
 
-    if (line >= newline_list->size) {
+    if (index >= newline_list->size - 1) {
         end = parser->end;
     } else {
-        end = &parser->start[newline_list->offsets[line]];
+        end = &parser->start[newline_list->offsets[index + 1]];
     }
 
-    pm_buffer_append_format(buffer, number_prefix, (uint32_t) line);
+    pm_buffer_append_format(buffer, number_prefix, line);
     pm_buffer_append_string(buffer, (const char *) start, (size_t) (end - start));
 
     if (end == parser->end && end[-1] != '\n') {
@@ -18225,25 +18239,26 @@ pm_parser_errors_format(const pm_parser_t *parser, pm_buffer_t *buffer, bool col
 
     // First, we're going to sort all of the errors by line number using an
     // insertion sort into a newly allocated array.
+    const int32_t start_line = parser->start_line;
     const pm_newline_list_t *newline_list = &parser->newline_list;
-    pm_error_t *errors = pm_parser_errors_format_sort(error_list, newline_list);
+    pm_error_t *errors = pm_parser_errors_format_sort(parser, error_list, newline_list);
 
     // Now we're going to determine how we're going to format line numbers and
     // blank lines based on the maximum number of digits in the line numbers
     // that are going to be displayed.
     pm_error_format_t error_format;
-    size_t max_line_number = errors[error_list->size - 1].line;
+    int32_t max_line_number = errors[error_list->size - 1].line - start_line;
 
     if (max_line_number < 10) {
         if (colorize) {
             error_format = (pm_error_format_t) {
-                .number_prefix = PM_COLOR_GRAY "%1" PRIu32 " | " PM_COLOR_RESET,
+                .number_prefix = PM_COLOR_GRAY "%1" PRIi32 " | " PM_COLOR_RESET,
                 .blank_prefix = PM_COLOR_GRAY "  | " PM_COLOR_RESET,
                 .divider = PM_COLOR_GRAY "  ~~~~~" PM_COLOR_RESET "\n"
             };
         } else {
             error_format = (pm_error_format_t) {
-                .number_prefix = "%1" PRIu32 " | ",
+                .number_prefix = "%1" PRIi32 " | ",
                 .blank_prefix = "  | ",
                 .divider = "  ~~~~~\n"
             };
@@ -18251,13 +18266,13 @@ pm_parser_errors_format(const pm_parser_t *parser, pm_buffer_t *buffer, bool col
     } else if (max_line_number < 100) {
         if (colorize) {
             error_format = (pm_error_format_t) {
-                .number_prefix = PM_COLOR_GRAY "%2" PRIu32 " | " PM_COLOR_RESET,
+                .number_prefix = PM_COLOR_GRAY "%2" PRIi32 " | " PM_COLOR_RESET,
                 .blank_prefix = PM_COLOR_GRAY "   | " PM_COLOR_RESET,
                 .divider = PM_COLOR_GRAY "  ~~~~~~" PM_COLOR_RESET "\n"
             };
         } else {
             error_format = (pm_error_format_t) {
-                .number_prefix = "%2" PRIu32 " | ",
+                .number_prefix = "%2" PRIi32 " | ",
                 .blank_prefix = "   | ",
                 .divider = "  ~~~~~~\n"
             };
@@ -18265,13 +18280,13 @@ pm_parser_errors_format(const pm_parser_t *parser, pm_buffer_t *buffer, bool col
     } else if (max_line_number < 1000) {
         if (colorize) {
             error_format = (pm_error_format_t) {
-                .number_prefix = PM_COLOR_GRAY "%3" PRIu32 " | " PM_COLOR_RESET,
+                .number_prefix = PM_COLOR_GRAY "%3" PRIi32 " | " PM_COLOR_RESET,
                 .blank_prefix = PM_COLOR_GRAY "    | " PM_COLOR_RESET,
                 .divider = PM_COLOR_GRAY "  ~~~~~~~" PM_COLOR_RESET "\n"
             };
         } else {
             error_format = (pm_error_format_t) {
-                .number_prefix = "%3" PRIu32 " | ",
+                .number_prefix = "%3" PRIi32 " | ",
                 .blank_prefix = "    | ",
                 .divider = "  ~~~~~~~\n"
             };
@@ -18279,13 +18294,13 @@ pm_parser_errors_format(const pm_parser_t *parser, pm_buffer_t *buffer, bool col
     } else if (max_line_number < 10000) {
         if (colorize) {
             error_format = (pm_error_format_t) {
-                .number_prefix = PM_COLOR_GRAY "%4" PRIu32 " | " PM_COLOR_RESET,
+                .number_prefix = PM_COLOR_GRAY "%4" PRIi32 " | " PM_COLOR_RESET,
                 .blank_prefix = PM_COLOR_GRAY "     | " PM_COLOR_RESET,
                 .divider = PM_COLOR_GRAY "  ~~~~~~~~" PM_COLOR_RESET "\n"
             };
         } else {
             error_format = (pm_error_format_t) {
-                .number_prefix = "%4" PRIu32 " | ",
+                .number_prefix = "%4" PRIi32 " | ",
                 .blank_prefix = "     | ",
                 .divider = "  ~~~~~~~~\n"
             };
@@ -18293,13 +18308,13 @@ pm_parser_errors_format(const pm_parser_t *parser, pm_buffer_t *buffer, bool col
     } else {
         if (colorize) {
             error_format = (pm_error_format_t) {
-                .number_prefix = PM_COLOR_GRAY "%5" PRIu32 " | " PM_COLOR_RESET,
+                .number_prefix = PM_COLOR_GRAY "%5" PRIi32 " | " PM_COLOR_RESET,
                 .blank_prefix = PM_COLOR_GRAY "      | " PM_COLOR_RESET,
                 .divider = PM_COLOR_GRAY "  ~~~~~~~~" PM_COLOR_RESET "\n"
             };
         } else {
             error_format = (pm_error_format_t) {
-                .number_prefix = "%5" PRIu32 " | ",
+                .number_prefix = "%5" PRIi32 " | ",
                 .blank_prefix = "      | ",
                 .divider = "  ~~~~~~~~\n"
             };
@@ -18314,7 +18329,7 @@ pm_parser_errors_format(const pm_parser_t *parser, pm_buffer_t *buffer, bool col
     // the source before the error to give some context. We'll be careful not to
     // display the same line twice in case the errors are close enough in the
     // source.
-    uint32_t last_line = 0;
+    int32_t last_line = 0;
     const pm_encoding_t *encoding = parser->encoding;
 
     for (size_t index = 0; index < error_list->size; index++) {
@@ -18360,7 +18375,7 @@ pm_parser_errors_format(const pm_parser_t *parser, pm_buffer_t *buffer, bool col
         pm_buffer_append_string(buffer, error_format.blank_prefix, error_format.blank_prefix_length);
 
         size_t column = 0;
-        const uint8_t *start = &parser->start[newline_list->offsets[error->line - 1]];
+        const uint8_t *start = &parser->start[newline_list->offsets[error->line - start_line]];
 
         while (column < error->column_end) {
             if (column < error->column_start) {
@@ -18384,7 +18399,7 @@ pm_parser_errors_format(const pm_parser_t *parser, pm_buffer_t *buffer, bool col
         // Here we determine how many lines of padding to display after the
         // error, depending on where the next error is in source.
         last_line = error->line;
-        size_t next_line = (index == error_list->size - 1) ? newline_list->size : errors[index + 1].line;
+        int32_t next_line = (index == error_list->size - 1) ? ((int32_t) newline_list->size) : errors[index + 1].line;
 
         if (next_line - last_line > 1) {
             pm_buffer_append_string(buffer, "  ", 2);

data/src/util/pm_newline_list.c

@@ -51,7 +51,7 @@ pm_newline_list_append(pm_newline_list_t *list, const uint8_t *cursor) {
  * are returned.
  */
 pm_line_column_t
-pm_newline_list_line_column(const pm_newline_list_t *list, const uint8_t *cursor) {
+pm_newline_list_line_column(const pm_newline_list_t *list, const uint8_t *cursor, int32_t start_line) {
     assert(cursor >= list->start);
     size_t offset = (size_t) (cursor - list->start);
 
@@ -62,7 +62,7 @@ pm_newline_list_line_column(const pm_newline_list_t *list, const uint8_t *cursor
         size_t mid = left + (right - left) / 2;
 
         if (list->offsets[mid] == offset) {
-            return ((pm_line_column_t) { mid + 1, 0 });
+            return ((pm_line_column_t) { ((int32_t) mid) + start_line, 0 });
         }
 
         if (list->offsets[mid] < offset) {
@@ -72,7 +72,10 @@ pm_newline_list_line_column(const pm_newline_list_t *list, const uint8_t *cursor
         }
     }
 
-    return ((pm_line_column_t) { left, offset - list->offsets[left - 1] });
+    return ((pm_line_column_t) {
+        .line = ((int32_t) left) + start_line - 1,
+        .column = (uint32_t) (offset - list->offsets[left - 1])
+    });
 }
 
 /**

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: prism
 version: !ruby/object:Gem::Version
-  version: 0.23.0
+  version: 0.24.0
 platform: ruby
 authors:
 - Shopify
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2024-02-14 00:00:00.000000000 Z
+date: 2024-02-15 00:00:00.000000000 Z
 dependencies: []
 description:
 email:
@@ -95,6 +95,8 @@ files:
 - lib/prism/translation/parser/compiler.rb
 - lib/prism/translation/parser/lexer.rb
 - lib/prism/translation/parser/rubocop.rb
+- lib/prism/translation/parser33.rb
+- lib/prism/translation/parser34.rb
 - lib/prism/translation/ripper.rb
 - lib/prism/translation/ruby_parser.rb
 - lib/prism/visitor.rb