Test multiple starting rules

.github/workflows/run-tests.yml

@@ -17,7 +17,7 @@ jobs:
     steps:
     - name: Install dependencies (Linux)
       if: runner.os == 'Linux'
-      run: sudo apt-get update && sudo apt-get install -y gcc gdc ldc valgrind
+      run: sudo apt-get update && sudo apt-get install -y gcc gdc ldc
 
     - name: Install dependencies (macOS)
       if: runner.os == 'macOS'
@@ -31,9 +31,6 @@ jobs:
       with:
         ruby-version: ${{ matrix.ruby-version }}
 
-    - name: Set up Rust
-      uses: dtolnay/rust-toolchain@stable
-
     - name: Install dependencies
       run: bundle install
 

CHANGELOG.md

@@ -1,169 +1,3 @@
-## v5.1.0
-
-### New Features
-
-- Add a `node_id()` accessor to the C++ and D tree node handle types, for node
-  identity comparison. This matches the existing `p_node_id()` macro (C) and
-  `node_id()` method (Rust).
-
-## v5.0.0
-
-### New Features
-
-- Add Rust target language output.
-- Add Rust language detection in propane.vim.
-
-### API Changes
-
-- The matched text argument passed to lexer user code blocks is now named
-  `match_text` instead of `match`, since `match` is a keyword in Rust. The
-  `match_length` argument (C and C++) is unchanged.
-- Tree generation mode now stores all tree nodes in a compact arena owned by
-  the parser context (a flat node array plus a shared child-link array).
-  This replaces the previous design of one heap allocation per node with
-  layout-punned typed structs.
-- Tree nodes are now referenced by lightweight handles rather than pointers.
-  `p_result()` and the field accessors now return handle values in tree
-  generation mode.
-- The whole tree is freed together with the context by `p_context_delete()`.
-  The `p_tree_delete()` / `p_tree_delete_XXX()` functions have been removed;
-  tree node handles are only valid while the context is alive.
-- Tree node field access changed per target language:
-  - C: per-field accessor functions (e.g. `p_Start_pItems(node)`) plus tree
-    walk macros (e.g. `p_tree_walk_Start(node, pItems, pItem, pToken1, token)`),
-    and generic accessors `p_node_valid()`, `p_node_position()`,
-    `p_node_end_position()`, `p_node_n_fields()`, `p_node_data()`, `p_node_id()`.
-  - C++: handle methods called with `()` (e.g. `node.pItems().pToken1().token()`),
-    plus the same C-style functions/macros for convenience.
-  - D: `@property` accessors preserving the previous field-access syntax
-    (e.g. `node.pItems.pToken1.token`); null checks use `.valid` instead of
-    `is null`.
-- Tree-mode parser rule user code: `$$` and `$1` etc. now yield node handles.
-  Reference child fields through the target-language accessors described above
-  rather than through struct pointer members.
-
-### Improvements
-
-- Improve D language detection in propane.vim
-- Speed up specs
-
-## v4.8.1
-
-### Fixes
-
-- Fix tree node struct type forward-declarations for C/C++
-
-## v4.8.0
-
-### New Features
-
-- Add `p_parse_inner_XXX()` APIs that accept a caller-provided set of follow
-  tokens. These behave the same as `p_parse_XXX()` by parsing starting at the
-  given start rule, but instead of expecting the rest of the input to match
-  the start rule they allow specifying a set of tokens that may follow the
-  start rule.
-- Add `p_set_position()` API to set the current text position stored in the
-  context. Useful for setting the initial text position to something other
-  than `(1, 1)` for a nested parse operation.
-- Add `p_input_index()` API to get the current input text byte offset.
-- Add `p_set_input_index()` API to set the current input text byte offset.
-  Useful together with `p_set_position()` to rewind the input part-way through
-  a parse in order to re-read an earlier section of the input.
-
-## v4.7.0
-
-### New Features
-
-- Support parser rule user code blocks in tree generation mode.
-
-### Fixes
-
-- propane.vim: do not highlight rule components as propane keywords
-
-## v4.6.0
-
-### New Features
-
-- Add lexer user code API to access matched input text positions
-- Track rule component text positions and add parser user code API to access
-
-### Fixes
-
-- Fixed a few user guide and source comments related to text input positions
-
-## v4.5.0
-
-### New Features
-
-- Add `noline` grammar statement to skip emitting `#line` directives
-- Attempt to autodetect target language (D/C++) in extra/vim/syntax/propane.vim
-
-### Fixes
-
-- Fix #line reset directives
-- Update keyword list in extra/vim/syntax/propane.vim
-- Fix propane.vim keyword detection
-
-## v4.4.0
-
-### New Features
-
-- Add p_value_get() / p_value_get_XXX() accessors
-
-## v4.3.0
-
-### New Features
-
-- Use #line for user code blocks to report input grammar position for errors.
-
-## v4.2.0
-
-### New Features
-
-- Add support for a custom lex function.
-
-## v4.1.0
-
-### New Features
-
-- Add `p_context_delete()` and `p_tree_delete()` for D targets.
-
-## v4.0.0
-
-### New Features
-
-- Add `context_user_fields` statement to allow custom context user fields.
-- Add `token_user_fields` statement to allow custom token user fields.
-- Add `on_token_node` statement to allow custom code when constructing token nodes.
-- Add `free_token_node` statement to allow custom code when freeing token nodes.
-- Add `p_context_delete()`.
-- Allow `drop` patterns to execute lexer user code blocks.
-
-### Breaking Changes
-
-- Replace `p_context_init()` with `p_context_new()` and `p_context_delete()`.
-- Renamed `p_free_tree()` to `p_tree_delete()`.
-- The `free_token_node` statement now takes a user code block instead of a
-  function name parameter.
-
-## v3.0.0
-
-### New Features
-
-- Add support for multiple starting rules (#38)
-- Add `p_free_tree()` functions to reclaim generated tree memory
-- Add `free_token_node` grammar statement to reclaim user-allocated memory stored in a Token tree node `pvalue` field
-- Add valgrind memory leak tests to unit tests
-- Fix build issues for C++ to officially support C++ target output
-
-### Improvements
-
-- Document `p_lex()` and `p_token_info_t` in user guide (#37)
-
-### Breaking Changes
-
-- Rename AST generation mode to tree generation mode (see [UPGRADING.md](UPGRADING.md))
-
 ## v2.3.0
 
 ### New Features

Gemfile.lock

@@ -5,12 +5,12 @@ GEM
     date (3.5.1)
     diff-lcs (1.6.2)
     docile (1.4.1)
-    erb (6.0.4)
+    erb (6.0.1)
-    psych (5.4.0)
+    psych (5.3.1)
       date
       stringio
-    rake (13.4.2)
+    rake (13.3.1)
-    rdoc (7.2.0)
+    rdoc (7.1.0)
       erb
       psych (>= 4.0.0)
       tsort
@@ -24,10 +24,10 @@ GEM
     rspec-expectations (3.13.5)
       diff-lcs (>= 1.2.0, < 2.0)
       rspec-support (~> 3.13.0)
-    rspec-mocks (3.13.8)
+    rspec-mocks (3.13.7)
       diff-lcs (>= 1.2.0, < 2.0)
       rspec-support (~> 3.13.0)
-    rspec-support (3.13.7)
+    rspec-support (3.13.6)
     simplecov (0.22.0)
       docile (~> 1.1)
       simplecov-html (~> 0.11)
@@ -51,4 +51,4 @@ DEPENDENCIES
   syntax
 
 BUNDLED WITH
-  4.0.14
+   2.3.7

LICENSE.txt

@@ -1,6 +1,6 @@
 The MIT License (MIT)
 
-Copyright (c) 2010-2026 Josh Holtrop
+Copyright (c) 2010-2024 Josh Holtrop
 
 Permission is hereby granted, free of charge, to any person obtaining a copy
 of this software and associated documentation files (the "Software"), to deal

README.md

@@ -6,10 +6,8 @@ Propane is a LALR Parser Generator (LPG) which:
   * generates a built-in lexer to tokenize input
   * supports UTF-8 lexer inputs
   * generates a table-driven shift/reduce parser to parse input in linear time
-  * targets C, C++, D, or Rust language outputs
+  * targets C, C++, or D language outputs
-  * optionally supports automatic full parse tree generation
+  * optionally supports automatic full AST generation
-  * supports starting parsing from multiple start rules
-  * tracks input text start and end positions for all matched tokens/rules
   * is MIT-licensed
   * is distributable as a standalone Ruby script
 
@@ -69,7 +67,7 @@ token times /\*/;
 token power /\*\*/;
 token integer /\d+/ <<
   ulong v;
-  foreach (c; match_text)
+  foreach (c; match)
   {
     v *= 10;
     v += (c - '0');

Rakefile.rb

@@ -1,4 +1,3 @@
-require "fileutils"
 require "rake/clean"
 require "rspec/core/rake_task"
 require "simplecov"
@@ -12,10 +11,7 @@ end
 
 RSpec::Core::RakeTask.new(:spec, :example_pattern) do |task, args|
   if args.example_pattern
-    ENV["partial_specs"] = "1"
     task.rspec_opts = %W[-e "#{args.example_pattern}" -f documentation]
-  else
-    FileUtils.rm_rf("coverage")
   end
 end
 task :spec do |task, args|
@@ -23,7 +19,7 @@ task :spec do |task, args|
     original_stdout = $stdout
     sio = StringIO.new
     $stdout = sio
-    SimpleCov.collate Dir["coverage/parts/*/.resultset.json"]
+    SimpleCov.collate Dir["coverage/.resultset.json"]
     $stdout = original_stdout
     sio.string.lines.each do |line|
       $stdout.write(line) unless line =~ /Coverage report generated for/
@@ -31,15 +27,6 @@ task :spec do |task, args|
   end
 end
 
-task :valgrind do
-  begin
-    ENV["spec-valgrind"] = "1"
-    Rake::Task[:spec].execute
-  ensure
-    ENV.delete("spec-valgrind")
-  end
-end
-
 # dspec task is useful to test the distributable release script, but is not
 # useful for coverage information.
 desc "Dist Specs"
@@ -56,4 +43,4 @@ task :user_guide do
   system("ruby", "-Ilib", "rb/gen_user_guide.rb")
 end
 
-task :all => [:valgrind, :dspec, :user_guide]
+task :all => [:spec, :dspec, :user_guide]

UPGRADING.md

@@ -1,110 +0,0 @@
-## v5.0.0
-
-The generated API for tree generation mode (`tree;`) has been changed
-significantly for this version.
-Aside from the lexer user code block matched text rename described below, the
-lexer/parser value APIs for non-tree grammars are unchanged.
-
-### Lexer user code block matched text
-
-The matched text argument passed to lexer user code blocks has been renamed
-from `match` to `match_text` for all target languages.
-
-- C, C++, and D: rename references to `match` in lexer user code blocks to
-  `match_text` (for example `$$ = match[0];` becomes `$$ = match_text[0];`).
-
-The `match_length` argument (C, C++) is unchanged.
-
-### Tree memory management
-
-- Remove all calls to `p_tree_delete()` / `p_tree_delete_XXX()`. Tree nodes now
-  live in the parser context and are freed by `p_context_delete()`.
-- Tree node handles (returned by `p_result()` and the field accessors) are only
-  valid while the context is alive. Do not use them after `p_context_delete()`.
-
-### Tree node field access
-
-Tree nodes are now referenced by handle values instead of pointers, and field
-access differs per target language:
-
-- C: replace `node->field` with the accessor function `p_TYPE_field(node)`, or
-  use the tree walk macro `p_tree_walk_TYPE(node, field1, field2, ...)`. Replace
-  `x != NULL` / `x == NULL` node checks with `p_node_valid(x)` /
-  `!p_node_valid(x)`. Read positions with `p_node_position(node)` /
-  `p_node_end_position(node)`, token payload with `p_TYPE_token(node)` /
-  `p_TYPE_pvalue(node)` or `p_node_data(node)->field`, and compare node identity
-  with `p_node_id(a) == p_node_id(b)`.
-- C++: replace `node->field` with the handle method `node.field()`. Use
-  `node.valid()`, `node.position()`, `node.token()`, `node.pvalue()`, and
-  `node.data()->field` for user token fields. (The C-style functions and macros
-  above are also available.)
-- D: replace pointer declarations (`Start * s`) with value handles (`Start s`)
-  and replace `x !is null` / `x is null` with `x.valid` / `!x.valid`. Field
-  access syntax (`node.field.field`) is otherwise unchanged.
-
-### Tree-mode parser rule user code
-
-In tree generation mode `$$` and `$1`, `$2`, ... now expand to node handles.
-Reference child fields through the target-language accessors above (for example
-`$$->pA->pToken1->pvalue` becomes `p_tree_walk_Start($$, pA, pToken1, pvalue)`
-in C, `$$.pA().pToken1().pvalue()` in C++, and `$$.pA.pToken1.pvalue` in D).
-
-### Pointers into tree node storage
-
-Tree nodes previously each had their own allocation, so a pointer to a node
-stayed valid for the life of the tree. They are now held in a single array
-which is reallocated as it grows, so a pointer or reference into that array may
-be invalidated whenever a new node is created.
-
-New nodes are created while parsing, so this matters for a pointer taken in a
-tree-mode parser rule user code block, which runs before the parse has
-finished. Keep the node handle instead, which stores a node ID rather than an
-address and stays valid, and obtain the pointer from it when it is needed.
-
-For example, replace a saved pointer:
-
-```
-context_user_fields <<
-    p_node_data_t * saved;
->>
-Items -> Items a << ${context.saved} = p_node_data($$); >>
-```
-
-with a saved handle:
-
-```
-context_user_fields <<
-    Items saved_node;
->>
-Items -> Items a << ${context.saved_node} = $$; >>
-```
-
-```
-p_node_data_t * data = p_node_data(context->saved_node);
-```
-
-Once parsing has finished, no further nodes are created, so a pointer obtained
-after `p_parse()` returns stays valid until the context is deleted, as long as
-no further parsing is performed with the same context.
-
-## v4.0.0
-
-### API Changes
-
-- Replace any calls to `p_context_init()` with `p_context_new()`.
-- Replace any references to the address of a statically allocated context
-  structure with the pointer returned from `p_context_init()` (e.g. `&context`
-  -> `context`).
-- Add a call to `p_context_delete()` (for C or C++) after lexing/parsing to
-  reclaim context memory.
-- Rename `p_free_tree()` calls to `p_tree_delete()`.
-- Change `free_token_node` statement calls from taking a function name argument
-  to taking a user code block.
-
-## v3.0.0
-
-### Grammar Changes
-
-- Rename `ast;` statement to `tree;`.
-- Rename `ast_prefix;` statement to `tree_prefix;`.
-- Rename `ast_suffix;` statement to `tree_suffix;`.

assets/parser.c.erb

@@ -43,82 +43,30 @@ const char * <%= @grammar.prefix %>token_names[] = {
  *************************************************************************/
 
 /**
- * Allocate and initialize lexer/parser context structure.
+ * Initialize lexer/parser context structure.
- *
- * Deinitialize and deallocate with <%= @grammar.prefix %>context_delete().
  *
+ * @param[out] context
+ *   Lexer/parser context structure.
  * @param input
  *   Text input.
  * @param input_length
  *   Text input length.
- *
- * @return Context structure for lexer/parser.
  */
-<%= @grammar.prefix %>context_t * <%= @grammar.prefix %>context_new(uint8_t const * input, size_t input_length)
+void <%= @grammar.prefix %>context_init(<%= @grammar.prefix %>context_t * context, uint8_t const * input, size_t input_length)
 {
-<% if @cpp %>
+    /* New default-initialized context structure. */
-    <%= @grammar.prefix %>context_t * context = new <%= @grammar.prefix %>context_t();
+    <%= @grammar.prefix %>context_t newcontext;
-<% else %>
+    memset(&newcontext, 0, sizeof(newcontext));
-    <%= @grammar.prefix %>context_t * context = (<%= @grammar.prefix %>context_t *)calloc(1, sizeof(<%= @grammar.prefix %>context_t));
-<% end %>
 
     /* Lexer initialization. */
-    context->input = input;
+    newcontext.input = input;
-    context->input_length = input_length;
+    newcontext.input_length = input_length;
-    context->text_position.row = 1u;
+    newcontext.text_position.row = 1u;
-    context->text_position.col = 1u;
+    newcontext.text_position.col = 1u;
-    context->mode = <%= @lexer.mode_id("default") %>;
+    newcontext.mode = <%= @lexer.mode_id("default") %>;
-<% if @grammar.tree %>
 
-    /* Reserve node ID 0 as the null tree node. */
+    /* Copy to the user's context structure. */
-<%   if @cpp %>
+    *context = newcontext;
-    context-><%= @grammar.prefix %>tree_nodes.resize(1);
-<%   else %>
-    context-><%= @grammar.prefix %>tree_nodes_capacity = 16u;
-    context-><%= @grammar.prefix %>tree_nodes = (<%= @grammar.prefix %>node_data_t *)malloc(16u * sizeof(<%= @grammar.prefix %>node_data_t));
-    memset(&context-><%= @grammar.prefix %>tree_nodes[0], 0, sizeof(<%= @grammar.prefix %>node_data_t));
-    context-><%= @grammar.prefix %>tree_nodes_length = 1u;
-<%   end %>
-<% end %>
-
-    return context;
-}
-
-/**
- * Deinitialize and deallocate lexer/parser context structure.
- *
- * For C++, destructors will be called for any context user fields. However, if
- * pointers are used to store allocated resources, the user should free them
- * before calling this function.
- *
- * @param context
- *   Lexer/parser context structure allocated with <%= @grammar.prefix %>context_new().
- */
-void <%= @grammar.prefix %>context_delete(<%= @grammar.prefix %>context_t * context)
-{
-<% if @grammar.tree && @grammar.free_token_node != "" %>
-<%   if @cpp %>
-    for (size_t i = 0u; i < context-><%= @grammar.prefix %>tree_nodes.size(); i++)
-<%   else %>
-    for (size_t i = 0u; i < context-><%= @grammar.prefix %>tree_nodes_length; i++)
-<%   end %>
-    {
-        if (context-><%= @grammar.prefix %>tree_nodes[i].is_token)
-        {
-            <%= @grammar.prefix %>node_data_t * token_tree_node = &context-><%= @grammar.prefix %>tree_nodes[i];
-<%= expand_code(@grammar.free_token_node, false, nil, nil) %>
-        }
-    }
-<% end %>
-<% if @cpp %>
-    delete context;
-<% else %>
-<%   if @grammar.tree %>
-    free(context-><%= @grammar.prefix %>tree_nodes);
-    free(context-><%= @grammar.prefix %>tree_children);
-<%   end %>
-    free(context);
-<% end %>
 }
 
 /**************************************************************************
@@ -319,7 +267,7 @@ static lexer_mode_t lexer_mode_table[] = {
  *   Lexer/parser context structure.
  * @param code_id
  *   The ID of the user code block to execute.
- * @param match_text
+ * @param match
  *   Matched text for this pattern.
  * @param match_length
  *   Matched text length.
@@ -330,7 +278,7 @@ static lexer_mode_t lexer_mode_table[] = {
  *   not explicitly return a token.
  */
 static <%= @grammar.prefix %>token_t lexer_user_code(<%= @grammar.prefix %>context_t * context,
-    lexer_user_code_id_t code_id, uint8_t const * match_text,
+    lexer_user_code_id_t code_id, uint8_t const * match,
     size_t match_length, <%= @grammar.prefix %>token_info_t * out_token_info)
 {
     switch (code_id)
@@ -516,27 +464,11 @@ static size_t attempt_lex_token(<%= @grammar.prefix %>context_t * context, <%= @
     case P_SUCCESS:
         {
         <%= @grammar.prefix %>token_t token_to_accept = match_info.accepting_state->token;
-        /* Calculate the token length and start/end positions before invoking
-         * the lexer user code so that the user code can access them. The
-         * context input text position tracking is not updated until after the
-         * user code has run so that it is left unchanged if the user code
-         * requests to terminate the lexer. */
-        token_info.length = match_info.length;
-        if (match_info.end_delta_position.row != 0u)
-        {
-            token_info.end_position.row = token_info.position.row + match_info.end_delta_position.row;
-            token_info.end_position.col = match_info.end_delta_position.col;
-        }
-        else
-        {
-            token_info.end_position.row = token_info.position.row;
-            token_info.end_position.col = token_info.position.col + match_info.end_delta_position.col;
-        }
         if (match_info.accepting_state->code_id != INVALID_USER_CODE_ID)
         {
-            uint8_t const * match_text = &context->input[context->input_index];
+            uint8_t const * match = &context->input[context->input_index];
             <%= @grammar.prefix %>token_t user_code_token = lexer_user_code(context,
-                match_info.accepting_state->code_id, match_text, match_info.length, &token_info);
+                match_info.accepting_state->code_id, match, match_info.length, &token_info);
             /* A TERMINATE_TOKEN_ID return code from lexer_user_code() means
              * that the user code is requesting to terminate the lexer. */
             if (user_code_token == TERMINATE_TOKEN_ID)
@@ -570,6 +502,17 @@ static size_t attempt_lex_token(<%= @grammar.prefix %>context_t * context, <%= @
             return P_DROP;
         }
         token_info.token = token_to_accept;
+        token_info.length = match_info.length;
+        if (match_info.end_delta_position.row != 0u)
+        {
+            token_info.end_position.row = token_info.position.row + match_info.end_delta_position.row;
+            token_info.end_position.col = match_info.end_delta_position.col;
+        }
+        else
+        {
+            token_info.end_position.row = token_info.position.row;
+            token_info.end_position.col = token_info.position.col + match_info.end_delta_position.col;
+        }
         *out_token_info = token_info;
         }
         return P_SUCCESS;
@@ -695,7 +638,7 @@ typedef struct
      * reduce action.
      */
     parser_state_id_t n_states;
-<% if @grammar.tree %>
+<% if @grammar.ast %>
 
     /**
      * Map of rule components to rule set child fields.
@@ -703,7 +646,7 @@ typedef struct
     uint16_t const * rule_set_node_field_index_map;
 
     /**
-     * Number of rule set tree node fields.
+     * Number of rule set AST node fields.
      */
     uint16_t rule_set_node_field_array_size;
 
@@ -742,17 +685,25 @@ typedef struct
     /** Parser state ID. */
     size_t state_id;
 
-<% if @grammar.tree %>
-    /** Tree node ID. */
-    <%= @grammar.prefix %>node_id_t node_id;
-<% else %>
-    <%= @grammar.prefix %>position_t position;
-    <%= @grammar.prefix %>position_t end_position;
     /** Parser value from this state. */
     <%= @grammar.prefix %>value_t pvalue;
+
+<% if @grammar.ast %>
+    /** AST node. */
+    void * ast_node;
 <% end %>
 } state_value_t;
 
+/** Common AST node structure. */
+typedef struct ASTNode_s
+{
+    <%= @grammar.prefix %>position_t position;
+    <%= @grammar.prefix %>position_t end_position;
+    uint16_t n_fields;
+    uint8_t is_token;
+    struct ASTNode_s * fields[];
+} ASTNode;
+
 /** Parser shift table. */
 static const shift_t parser_shift_table[] = {
 <%   @parser.shift_table.each do |shift| %>
@@ -760,7 +711,7 @@ static const shift_t parser_shift_table[] = {
 <%   end %>
 };
 
-<% if @grammar.tree %>
+<% if @grammar.ast %>
 <%   @grammar.rules.each do |rule| %>
 <%     unless rule.flat_rule_set_node_field_index_map? %>
 const uint16_t r_<%= rule.name.gsub("$", "_") %><%= rule.id %>_node_field_index_map[<%= rule.rule_set_node_field_index_map.size %>] = {<%= rule.rule_set_node_field_index_map.map {|v| v.to_s}.join(", ") %>};
@@ -775,14 +726,14 @@ static const reduce_t parser_reduce_table[] = {
         <%= reduce[:token_id] %>u, /* Token: <%= reduce[:token] ? reduce[:token].name : "(any)" %> */
         <%= reduce[:rule_id] %>u, /* Rule ID */
         <%= reduce[:rule_set_id] %>u, /* Rule set ID (<%= reduce[:rule].rule_set.name %>) */
-<%     if @grammar.tree %>
+<%     if @grammar.ast %>
         <%= reduce[:n_states] %>u, /* Number of states */
 <%       if reduce[:rule].flat_rule_set_node_field_index_map? %>
         NULL, /* No rule set node field index map (flat map) */
 <%       else %>
         &r_<%= reduce[:rule].name.gsub("$", "_") %><%= reduce[:rule].id %>_node_field_index_map[0], /* Rule set node field index map */
 <%       end %>
-        <%= reduce[:rule].rule_set.tree_fields.size %>, /* Number of tree fields */
+        <%= reduce[:rule].rule_set.ast_fields.size %>, /* Number of AST fields */
         <%= reduce[:propagate_optional_target] %>}, /* Propagate optional target? */
 <%     else %>
         <%= reduce[:n_states] %>u},
@@ -890,131 +841,7 @@ static void state_values_stack_free(state_values_stack_t * stack)
      free(stack->entries);
 }
 
-<% if @grammar.tree %>
+<% unless @grammar.ast %>
-/* Tree arena helpers. */
-
-/**
- * Allocate a new (zeroed) tree node in the context arena.
- *
- * @return The new node ID.
- */
-static <%= @grammar.prefix %>node_id_t tree_new_node(<%= @grammar.prefix %>context_t * context)
-{
-<% if @cpp %>
-    <%= @grammar.prefix %>node_id_t id = (<%= @grammar.prefix %>node_id_t)context-><%= @grammar.prefix %>tree_nodes.size();
-    context-><%= @grammar.prefix %>tree_nodes.emplace_back();
-    return id;
-<% else %>
-    if (context-><%= @grammar.prefix %>tree_nodes_length >= context-><%= @grammar.prefix %>tree_nodes_capacity)
-    {
-        size_t new_capacity = context-><%= @grammar.prefix %>tree_nodes_capacity * 2u;
-        <%= @grammar.prefix %>node_data_t * new_nodes = (<%= @grammar.prefix %>node_data_t *)malloc(new_capacity * sizeof(<%= @grammar.prefix %>node_data_t));
-        memcpy(new_nodes, context-><%= @grammar.prefix %>tree_nodes, context-><%= @grammar.prefix %>tree_nodes_length * sizeof(<%= @grammar.prefix %>node_data_t));
-        free(context-><%= @grammar.prefix %>tree_nodes);
-        context-><%= @grammar.prefix %>tree_nodes = new_nodes;
-        context-><%= @grammar.prefix %>tree_nodes_capacity = new_capacity;
-    }
-    <%= @grammar.prefix %>node_id_t id = (<%= @grammar.prefix %>node_id_t)context-><%= @grammar.prefix %>tree_nodes_length;
-    memset(&context-><%= @grammar.prefix %>tree_nodes[id], 0, sizeof(<%= @grammar.prefix %>node_data_t));
-    context-><%= @grammar.prefix %>tree_nodes_length += 1u;
-    return id;
-<% end %>
-}
-
-/**
- * Reserve n contiguous (zeroed) child slots in the shared children array.
- *
- * @return The offset of the first reserved slot.
- */
-static <%= @grammar.prefix %>node_id_t tree_reserve_children(<%= @grammar.prefix %>context_t * context, size_t n)
-{
-<% if @cpp %>
-    <%= @grammar.prefix %>node_id_t offset = (<%= @grammar.prefix %>node_id_t)context-><%= @grammar.prefix %>tree_children.size();
-    context-><%= @grammar.prefix %>tree_children.resize(context-><%= @grammar.prefix %>tree_children.size() + n);
-    return offset;
-<% else %>
-    size_t offset = context-><%= @grammar.prefix %>tree_children_length;
-    size_t needed = offset + n;
-    if (needed > context-><%= @grammar.prefix %>tree_children_capacity)
-    {
-        size_t new_capacity = context-><%= @grammar.prefix %>tree_children_capacity ? context-><%= @grammar.prefix %>tree_children_capacity : 1u;
-        while (new_capacity < needed)
-        {
-            new_capacity *= 2u;
-        }
-        <%= @grammar.prefix %>node_id_t * new_children = (<%= @grammar.prefix %>node_id_t *)malloc(new_capacity * sizeof(<%= @grammar.prefix %>node_id_t));
-        if (context-><%= @grammar.prefix %>tree_children != NULL)
-        {
-            memcpy(new_children, context-><%= @grammar.prefix %>tree_children, context-><%= @grammar.prefix %>tree_children_length * sizeof(<%= @grammar.prefix %>node_id_t));
-            free(context-><%= @grammar.prefix %>tree_children);
-        }
-        context-><%= @grammar.prefix %>tree_children = new_children;
-        context-><%= @grammar.prefix %>tree_children_capacity = new_capacity;
-    }
-    memset(&context-><%= @grammar.prefix %>tree_children[offset], 0, n * sizeof(<%= @grammar.prefix %>node_id_t));
-    context-><%= @grammar.prefix %>tree_children_length = needed;
-    return (<%= @grammar.prefix %>node_id_t)offset;
-<% end %>
-}
-
-/* Tree node field accessor functions. */
-<%= c_tree_accessor_defs %>
-<% end %>
-
-<% unless @grammar.tree %>
-/**
- * Get the rule position (start or end) for the currently matched rule.
- */
-static <%= @grammar.prefix %>position_t get_rule_position(state_values_stack_t * statevalues, size_t i, size_t n_states, bool get_end)
-{
-    if (n_states > 0u)
-    {
-        if (i == 0u)
-        {
-            if (get_end)
-            {
-                int stack_index = -1;
-                for (size_t j = 0u; j < n_states; j++)
-                {
-                    state_value_t * sv = state_values_stack_index(statevalues, stack_index - (int)j);
-                    if (<%= @grammar.prefix %>position_valid(sv->end_position))
-                    {
-                        return sv->end_position;
-                    }
-                }
-            }
-            else
-            {
-                int stack_index = -(int)n_states;
-                for (size_t j = 0u; j < n_states; j++)
-                {
-                    state_value_t * sv = state_values_stack_index(statevalues, stack_index + (int)j);
-                    if (<%= @grammar.prefix %>position_valid(sv->position))
-                    {
-                        return sv->position;
-                    }
-                }
-            }
-        }
-        else
-        {
-            if (get_end)
-            {
-                return state_values_stack_index(statevalues, -1 - (int)n_states + (int)i)->end_position;
-            }
-            else
-            {
-                return state_values_stack_index(statevalues, -1 - (int)n_states + (int)i)->position;
-            }
-        }
-    }
-    <%= @grammar.prefix %>position_t empty_pos;
-    memset(&empty_pos, 0, sizeof(empty_pos));
-    return empty_pos;
-}
-<% end %>
-
-<% if !@grammar.tree || @grammar.parser_user_code_used? %>
 /**
  * Execute user code associated with a parser rule.
  *
@@ -1025,7 +852,7 @@ static <%= @grammar.prefix %>position_t get_rule_position(state_values_stack_t *
  * @retval P_USER_TERMINATED
  *   User requested to terminate parsing.
  */
-static size_t parser_user_code(<%= @grammar.tree ? "#{@grammar.prefix}node_id_t _node_id" : "#{@grammar.prefix}value_t * _pvalue" %>, uint32_t rule, state_values_stack_t * statevalues, uint32_t n_states, <%= @grammar.prefix %>context_t * context)
+static size_t parser_user_code(<%= @grammar.prefix %>value_t * _pvalue, uint32_t rule, state_values_stack_t * statevalues, uint32_t n_states, <%= @grammar.prefix %>context_t * context)
 {
     switch (rule)
     {
@@ -1075,7 +902,7 @@ static size_t check_shift(size_t state_id, size_t symbol_id)
  * @param token
  *   Incoming token.
  *
- * @return Reduce table index to reduce with, or INVALID_ID if none.
+ * @return State to reduce to, or INVALID_ID if none.
  */
 static size_t check_reduce(size_t state_id, <%= @grammar.prefix %>token_t token)
 {
@@ -1097,17 +924,8 @@ static size_t check_reduce(size_t state_id, <%= @grammar.prefix %>token_t token)
  *
  * @param context
  *   Lexer/parser context structure.
- * @param start_state_id
+ * @start_state_id
  *   ID of the state in which to start.
- * @param start_rule_set_id
- *   Rule set ID for the requested start rule. Only used when
- *   @p follow_tokens is non-NULL, to gate follow-token shift success.
- * @param follow_tokens
- *   Optional array of caller-provided follow tokens (tokens expected to
- *   appear immediately after the start rule in some outer context). Used to
- *   drive the "parse inner" retry logic. May be NULL for a standard parse.
- * @param n_follow_tokens
- *   Number of entries in @p follow_tokens.
  *
  * @retval P_SUCCESS
  *   The parser successfully matched the input text. The parse result value
@@ -1120,20 +938,15 @@ static size_t check_reduce(size_t state_id, <%= @grammar.prefix %>token_t token)
  * @reval P_UNEXPECTED_INPUT
  *   Input text does not match any lexer pattern.
  */
-static size_t parse_from(<%= @grammar.prefix %>context_t * context, size_t start_state_id,
+static size_t parse_from(<%= @grammar.prefix %>context_t * context, size_t start_state_id)
-    size_t start_rule_set_id,
-    <%= @grammar.prefix %>token_t const * follow_tokens, size_t n_follow_tokens)
 {
     <%= @grammar.prefix %>token_info_t token_info;
     <%= @grammar.prefix %>token_t token = INVALID_TOKEN_ID;
     state_values_stack_t statevalues;
     size_t reduced_rule_set = INVALID_ID;
-    size_t last_shifted_rule_set_id = INVALID_ID;
+<% if @grammar.ast %>
-<% if @grammar.tree %>
+    void * reduced_parser_node;
-    <%= @grammar.prefix %>node_id_t reduced_parser_node;
 <% else %>
-    <%= @grammar.prefix %>position_t reduced_position;
-    <%= @grammar.prefix %>position_t reduced_end_position;
     <%= @grammar.prefix %>value_t reduced_parser_value;
 <% end %>
     state_values_stack_init(&statevalues);
@@ -1144,7 +957,7 @@ static size_t parse_from(<%= @grammar.prefix %>context_t * context, size_t start
     {
         if (token == INVALID_TOKEN_ID)
         {
-            size_t lexer_result = <%= lex_fn %>(context, &token_info);
+            size_t lexer_result = <%= @grammar.prefix %>lex(context, &token_info);
             if (lexer_result != P_SUCCESS)
             {
                 result = lexer_result;
@@ -1152,18 +965,6 @@ static size_t parse_from(<%= @grammar.prefix %>context_t * context, size_t start
             }
             token = token_info.token;
         }
-        /* For a "parse inner" operation, determine once per iteration whether
-         * the current token is a member of the caller-provided follow token
-         * set. Used by both the shift-side and reduce-side retries below. */
-        bool token_is_follow = false;
-        for (size_t i = 0u; i < n_follow_tokens; i++)
-        {
-            if (token == follow_tokens[i])
-            {
-                token_is_follow = true;
-                break;
-            }
-        }
         size_t shift_state = INVALID_ID;
         if (reduced_rule_set != INVALID_ID)
         {
@@ -1175,83 +976,44 @@ static size_t parse_from(<%= @grammar.prefix %>context_t * context, size_t start
             if ((shift_state != INVALID_ID) && (token == TOKEN___EOF))
             {
                 /* Successful parse. */
-<% if @grammar.tree %>
+<% if @grammar.ast %>
-                context->parse_result = state_values_stack_index(&statevalues, -1)->node_id;
+                context->parse_result = state_values_stack_index(&statevalues, -1)->ast_node;
 <% else %>
                 context->parse_result = state_values_stack_index(&statevalues, -1)->pvalue;
 <% end %>
                 result = P_SUCCESS;
                 break;
             }
-            if ((shift_state == INVALID_ID) && token_is_follow)
-            {
-                /* For a "parse inner" operation, if the incoming token is one
-                 * of the caller's follow tokens, retry the shift as
-                 * TOKEN___EOF. Only consider the parse complete if the reduced
-                 * start rule is the only thing on the parse stack (i.e. the
-                 * initial state plus a single shifted start rule set entry). */
-                size_t retry_shift_state = check_shift(state_values_stack_index(&statevalues, -1)->state_id, TOKEN___EOF);
-                if ((retry_shift_state != INVALID_ID) &&
-                    (statevalues.length == 2u) &&
-                    (last_shifted_rule_set_id == start_rule_set_id))
-                {
-                    /* Successful parse via follow token. Rewind the input
-                     * position so that the follow token is not consumed from
-                     * the input stream and remains available for a subsequent
-                     * call to <%= @grammar.prefix %>lex() or a
-                     * <%= @grammar.prefix %>parse*() function. */
-                    context->input_index -= token_info.length;
-                    context->text_position = token_info.position;
-<% if @grammar.tree %>
-                    context->parse_result = state_values_stack_index(&statevalues, -1)->node_id;
-<% else %>
-                    context->parse_result = state_values_stack_index(&statevalues, -1)->pvalue;
-<% end %>
-                    result = P_SUCCESS;
-                    break;
-                }
-            }
         }
         if (shift_state != INVALID_ID)
         {
-            /* We have something to shift. Track the last shifted rule set ID
+            /* We have something to shift. */
-             * (INVALID_ID if we just shifted a token) so the follow-token
-             * shift retry can gate success on the reduced start rule being the
-             * only thing on top of the initial state. */
-            last_shifted_rule_set_id = reduced_rule_set;
             state_values_stack_push(&statevalues);
-            state_value_t * new_state_info = state_values_stack_index(&statevalues, -1);
+            state_values_stack_index(&statevalues, -1)->state_id = shift_state;
-            new_state_info->state_id = shift_state;
             if (reduced_rule_set == INVALID_ID)
             {
                 /* We shifted a token, mark it consumed. */
-<% if @grammar.tree %>
+<% if @grammar.ast %>
-                <%= @grammar.prefix %>node_id_t token_node_id = tree_new_node(context);
+                <%= @grammar.ast_prefix %>Token<%= @grammar.ast_suffix %> * token_ast_node = (<%= @grammar.ast_prefix %>Token<%= @grammar.ast_suffix %> *)malloc(sizeof(<%= @grammar.ast_prefix %>Token<%= @grammar.ast_suffix %>));
-                <%= @grammar.prefix %>node_data_t * token_tree_node = &context-><%= @grammar.prefix %>tree_nodes[token_node_id];
+                token_ast_node->position = token_info.position;
-                token_tree_node->position = token_info.position;
+                token_ast_node->end_position = token_info.end_position;
-                token_tree_node->end_position = token_info.end_position;
+                token_ast_node->n_fields = 0u;
-                token_tree_node->n_fields = 0u;
+                token_ast_node->is_token = 1u;
-                token_tree_node->is_token = 1u;
+                token_ast_node->token = token;
-                token_tree_node->token = token;
+                token_ast_node->pvalue = token_info.pvalue;
-                token_tree_node->pvalue = token_info.pvalue;
+                state_values_stack_index(&statevalues, -1)->ast_node = token_ast_node;
-<%= expand_code(@grammar.on_token_node, false, nil, nil) %>
-                new_state_info->node_id = token_node_id;
 <% else %>
-                new_state_info->position = token_info.position;
+                state_values_stack_index(&statevalues, -1)->pvalue = token_info.pvalue;
-                new_state_info->end_position = token_info.end_position;
-                new_state_info->pvalue = token_info.pvalue;
 <% end %>
                 token = INVALID_TOKEN_ID;
             }
             else
             {
                 /* We shifted a RuleSet. */
-<% if @grammar.tree %>
+<% if @grammar.ast %>
-                new_state_info->node_id = reduced_parser_node;
+                state_values_stack_index(&statevalues, -1)->ast_node = reduced_parser_node;
 <% else %>
-                new_state_info->pvalue = reduced_parser_value;
+                state_values_stack_index(&statevalues, -1)->pvalue = reduced_parser_value;
-                new_state_info->position = reduced_position;
-                new_state_info->end_position = reduced_end_position;
                 <%= @grammar.prefix %>value_t new_parse_result;
                 memset(&new_parse_result, 0, sizeof(new_parse_result));
                 reduced_parser_value = new_parse_result;
@@ -1262,77 +1024,57 @@ static size_t parse_from(<%= @grammar.prefix %>context_t * context, size_t start
         }
 
         size_t reduce_index = check_reduce(state_values_stack_index(&statevalues, -1)->state_id, token);
-        if ((reduce_index == INVALID_ID) && token_is_follow)
-        {
-            /* For a "parse inner" operation, if the incoming token is one of
-             * the caller's follow tokens, retry the reduce lookup as
-             * TOKEN___EOF. Whatever reduce_index results (if any) is used
-             * regardless of which rule set it reduces to; this allows chains
-             * of reductions leading up to the start rule. */
-            reduce_index = check_reduce(state_values_stack_index(&statevalues, -1)->state_id, TOKEN___EOF);
-        }
         if (reduce_index != INVALID_ID)
         {
             /* We have something to reduce. */
-<% if @grammar.tree %>
+<% if @grammar.ast %>
             if (parser_reduce_table[reduce_index].propagate_optional_target)
             {
-                reduced_parser_node = state_values_stack_index(&statevalues, -1)->node_id;
+                reduced_parser_node = state_values_stack_index(&statevalues, -1)->ast_node;
             }
             else if (parser_reduce_table[reduce_index].n_states > 0)
             {
-                uint16_t n_fields = parser_reduce_table[reduce_index].rule_set_node_field_array_size;
+                size_t n_fields = parser_reduce_table[reduce_index].rule_set_node_field_array_size;
-                /* Reserve child slots. New slots are zero-initialized
+                size_t bytes = sizeof(ASTNode) + n_fields * sizeof(void *);
-                 * (null node ID) so absent optional children remain null. */
+                ASTNode * node = (ASTNode *)malloc(bytes);
-                <%= @grammar.prefix %>node_id_t child_offset = tree_reserve_children(context, n_fields);
+                memset(node, 0, bytes);
+                node->position = INVALID_POSITION;
+                node->end_position = INVALID_POSITION;
+                node->n_fields = n_fields;
                 if (parser_reduce_table[reduce_index].rule_set_node_field_index_map == NULL)
                 {
                     for (size_t i = 0; i < parser_reduce_table[reduce_index].n_states; i++)
                     {
-                        context-><%= @grammar.prefix %>tree_children[child_offset + i] = state_values_stack_index(&statevalues, -(int)parser_reduce_table[reduce_index].n_states + (int)i)->node_id;
+                        node->fields[i] = (ASTNode *)state_values_stack_index(&statevalues, -(int)parser_reduce_table[reduce_index].n_states + (int)i)->ast_node;
                     }
                 }
                 else
                 {
                     for (size_t i = 0; i < parser_reduce_table[reduce_index].n_states; i++)
                     {
-                        context-><%= @grammar.prefix %>tree_children[child_offset + parser_reduce_table[reduce_index].rule_set_node_field_index_map[i]] = state_values_stack_index(&statevalues, -(int)parser_reduce_table[reduce_index].n_states + (int)i)->node_id;
+                        node->fields[parser_reduce_table[reduce_index].rule_set_node_field_index_map[i]] = (ASTNode *)state_values_stack_index(&statevalues, -(int)parser_reduce_table[reduce_index].n_states + (int)i)->ast_node;
                     }
                 }
-                <%= @grammar.prefix %>node_id_t node_id = tree_new_node(context);
-                <%= @grammar.prefix %>node_data_t * node = &context-><%= @grammar.prefix %>tree_nodes[node_id];
-                node->position = INVALID_POSITION;
-                node->end_position = INVALID_POSITION;
-                node->child_offset = child_offset;
-                node->n_fields = n_fields;
-                node->is_token = 0u;
                 bool position_found = false;
-                for (uint16_t i = 0; i < n_fields; i++)
+                for (size_t i = 0; i < n_fields; i++)
                 {
-                    <%= @grammar.prefix %>node_id_t child_id = context-><%= @grammar.prefix %>tree_children[child_offset + i];
+                    ASTNode * child = node->fields[i];
-                    if ((child_id != 0u) && <%= @grammar.prefix %>position_valid(context-><%= @grammar.prefix %>tree_nodes[child_id].position))
+                    if ((child != NULL) && <%= @grammar.prefix %>position_valid(child->position))
                     {
                         if (!position_found)
                         {
-                            node->position = context-><%= @grammar.prefix %>tree_nodes[child_id].position;
+                            node->position = child->position;
                             position_found = true;
                         }
-                        node->end_position = context-><%= @grammar.prefix %>tree_nodes[child_id].end_position;
+                        node->end_position = child->end_position;
                     }
                 }
-                reduced_parser_node = node_id;
+                reduced_parser_node = node;
             }
             else
             {
-                reduced_parser_node = 0u;
+                reduced_parser_node = NULL;
             }
-<%   if @grammar.parser_user_code_used? %>
-            if (parser_user_code(reduced_parser_node, parser_reduce_table[reduce_index].rule, &statevalues, parser_reduce_table[reduce_index].n_states, context) == P_USER_TERMINATED)
-            {
-                state_values_stack_free(&statevalues);
-                return P_USER_TERMINATED;
-            }
-<%   end %>
 <% else %>
             <%= @grammar.prefix %>value_t reduced_parser_value2;
             memset(&reduced_parser_value2, 0, sizeof(reduced_parser_value2));
@@ -1342,16 +1084,6 @@ static size_t parse_from(<%= @grammar.prefix %>context_t * context, size_t start
                 return P_USER_TERMINATED;
             }
             reduced_parser_value = reduced_parser_value2;
-            if (parser_reduce_table[reduce_index].n_states > 0u)
-            {
-                reduced_position = get_rule_position(&statevalues, 0u, parser_reduce_table[reduce_index].n_states, false);
-                reduced_end_position = get_rule_position(&statevalues, 0u, parser_reduce_table[reduce_index].n_states, true);
-            }
-            else
-            {
-                memset(&reduced_position, 0, sizeof(reduced_position));
-                memset(&reduced_end_position, 0, sizeof(reduced_end_position));
-            }
 <% end %>
             reduced_rule_set = parser_reduce_table[reduce_index].rule_set;
             state_values_stack_pop(&statevalues, parser_reduce_table[reduce_index].n_states);
@@ -1374,20 +1106,14 @@ static size_t parse_from(<%= @grammar.prefix %>context_t * context, size_t start
 
 size_t <%= @grammar.prefix %>parse(<%= @grammar.prefix %>context_t * context)
 {
-    return parse_from(context, 0u, <%= @parser.rule_sets[@grammar.start_rules[0]].id %>u, NULL, 0u);
+    return parse_from(context, 0u);
 }
 
 <% @grammar.start_rules.each_with_index do |start_rule, i| %>
 
 size_t <%= @grammar.prefix %>parse_<%= start_rule %>(<%= @grammar.prefix %>context_t * context)
 {
-    return parse_from(context, <%= i %>u, <%= @parser.rule_sets[start_rule].id %>u, NULL, 0u);
+    return parse_from(context, <%= i %>u);
-}
-
-size_t <%= @grammar.prefix %>parse_inner_<%= start_rule %>(<%= @grammar.prefix %>context_t * context,
-    <%= @grammar.prefix %>token_t const * follow_tokens, size_t n_follow_tokens)
-{
-    return parse_from(context, <%= i %>u, <%= @parser.rule_sets[start_rule].id %>u, follow_tokens, n_follow_tokens);
 }
 <% end %>
 
@@ -1399,15 +1125,15 @@ size_t <%= @grammar.prefix %>parse_inner_<%= start_rule %>(<%= @grammar.prefix %
  *
  * @return Parse result value.
  */
-<% if @grammar.tree %>
+<% if @grammar.ast %>
-<%= h_type(@grammar.start_rules[0]) %> <%= @grammar.prefix %>result(<%= @grammar.prefix %>context_t * context)
+<%= @grammar.ast_prefix %><%= @grammar.start_rules[0] %><%= @grammar.ast_suffix %> * <%= @grammar.prefix %>result(<%= @grammar.prefix %>context_t * context)
 {
-    return <%= tree_handle(h_type(@grammar.start_rules[0]), "context->parse_result") %>;
+    return (<%= @grammar.ast_prefix %><%= @grammar.start_rules[0] %><%= @grammar.ast_suffix %> *) context->parse_result;
 }
 <%   @grammar.start_rules.each_with_index do |start_rule, i| %>
-<%= h_type(start_rule) %> <%= @grammar.prefix %>result_<%= start_rule %>(<%= @grammar.prefix %>context_t * context)
+<%= @grammar.ast_prefix %><%= start_rule %><%= @grammar.ast_suffix %> * <%= @grammar.prefix %>result_<%= start_rule %>(<%= @grammar.prefix %>context_t * context)
 {
-    return <%= tree_handle(h_type(start_rule), "context->parse_result") %>;
+    return (<%= @grammar.ast_prefix %><%= start_rule %><%= @grammar.ast_suffix %> *) context->parse_result;
 }
 <%   end %>
 <% else %>
@@ -1436,58 +1162,6 @@ size_t <%= @grammar.prefix %>parse_inner_<%= start_rule %>(<%= @grammar.prefix %
     return context->text_position;
 }
 
-/**
- * Set the current text input position.
- *
- * This can be used to set the initial text position to something other than
- * (1, 1) for a nested parse operation so that error positions reported by
- * subsequent lexer/parser calls are relative to a larger enclosing document.
- *
- * @param context
- *   Lexer/parser context structure.
- * @param position
- *   Text position to set.
- */
-void <%= @grammar.prefix %>set_position(<%= @grammar.prefix %>context_t * context, <%= @grammar.prefix %>position_t position)
-{
-    context->text_position = position;
-}
-
-/**
- * Get the current input text byte offset.
- *
- * @param context
- *   Lexer/parser context structure.
- *
- * @return Current input text byte offset (measured from the start of the
- *   input text passed to <%= @grammar.prefix %>context_new()).
- */
-size_t <%= @grammar.prefix %>input_index(<%= @grammar.prefix %>context_t * context)
-{
-    return context->input_index;
-}
-
-/**
- * Set the current input text byte offset.
- *
- * This moves the lexer's read cursor to the given byte offset (measured from
- * the start of the input text passed to <%= @grammar.prefix %>context_new()).
- * It can be used together with <%= @grammar.prefix %>set_position() to rewind
- * the input part-way through a parse in order to re-read an earlier section of
- * the input. The byte offset is not validated; the caller is responsible for
- * providing an offset within the bounds of the input text. A value previously
- * returned by <%= @grammar.prefix %>input_index() is a suitable argument.
- *
- * @param context
- *   Lexer/parser context structure.
- * @param input_index
- *   Input text byte offset to set.
- */
-void <%= @grammar.prefix %>set_input_index(<%= @grammar.prefix %>context_t * context, size_t input_index)
-{
-    context->input_index = input_index;
-}
-
 /**
  * Get the user terminate code.
  *
@@ -1510,3 +1184,45 @@ size_t <%= @grammar.prefix %>user_terminate_code(<%= @grammar.prefix %>context_t
 {
     return context->token;
 }
+<% if @grammar.ast %>
+
+static void free_ast_node(ASTNode * node)
+{
+    if (node->is_token)
+    {
+<%   if @grammar.free_token_node %>
+        <%= @grammar.free_token_node %>((<%= @grammar.ast_prefix %>Token<%= @grammar.ast_suffix %> *) node);
+<%   end %>
+        /* TODO: free value_t */
+    }
+    else if (node->n_fields > 0u)
+    {
+        for (size_t i = 0u; i < node->n_fields; i++)
+        {
+            if (node->fields[i] != NULL)
+            {
+                free_ast_node(node->fields[i]);
+            }
+        }
+    }
+    free(node);
+}
+
+/**
+ * Free all AST node memory.
+ */
+void <%= @grammar.prefix %>free_ast(<%= @grammar.ast_prefix %><%= @grammar.start_rules[0] %><%= @grammar.ast_suffix %> * ast)
+{
+    free_ast_node((ASTNode *)ast);
+}
+<%   @grammar.start_rules.each_with_index do |start_rule, i| %>
+
+/**
+ * Free all AST node memory.
+ */
+void <%= @grammar.prefix %>free_ast_<%= start_rule %>(<%= @grammar.ast_prefix %><%= start_rule %><%= @grammar.ast_suffix %> * ast)
+{
+    free_ast_node((ASTNode *)ast);
+}
+<%   end %>
+<% end %>

assets/parser.d.erb

@@ -9,7 +9,7 @@ module <%= @grammar.modulename %>;
 <% end %>
 
 import core.memory;
-import core.stdc.stdlib : malloc, free;
+import core.stdc.stdlib : malloc;
 
 /**************************************************************************
  * User code blocks
@@ -59,10 +59,10 @@ public alias <%= @grammar.prefix %>code_point_t = uint;
  */
 public struct <%= @grammar.prefix %>position_t
 {
-    /** Input text row (1-based). */
+    /** Input text row (0-based). */
     uint row;
 
-    /** Input text column (1-based). */
+    /** Input text column (0-based). */
     uint col;
 
     /** Invalid position value. */
@@ -75,7 +75,7 @@ public struct <%= @grammar.prefix %>position_t
     }
 }
 
-<% if @grammar.tree %>
+<% if @grammar.ast %>
 /** Parser values type. */
 public alias <%= @grammar.prefix %>value_t = <%= @grammar.ptype %>;
 <% else %>
@@ -86,139 +86,41 @@ public union <%= @grammar.prefix %>value_t
     <%= typestring %> v_<%= name %>;
 <%   end %>
 }
-
-/** Parser value constructor(s). */
-<%   @grammar.ptypes.each do |name, typestring| %>
-public <%= @grammar.prefix %>value_t <%= @grammar.prefix %>value<%= name == "default" ? "" : "_#{name}" %>(T)(T v)
-{
-    return <%= @grammar.prefix %>value_t(v_<%= name %>: v);
-}
 <% end %>
 
-/** Parser value accessor(s). */
+<% if @grammar.ast %>
-<%   @grammar.ptypes.each do |name, typestring| %>
+/** Common AST node structure. */
-public <%= typestring %> <%= @grammar.prefix %>value_get<%= name == "default" ? "" : "_#{name}" %>(<%= @grammar.prefix %>value_t * pvalue)
+private struct ASTNode
-{
-    return pvalue.v_<%= name %>;
-}
-<%   end %>
-<% end %>
-
-<% if @grammar.tree %>
-/** Tree node ID type (index into the context node arena). ID 0 is null. */
-public alias <%= @grammar.prefix %>node_id_t = uint;
-
-/**
- * Tree node record.
- *
- * All tree nodes are stored contiguously in the context node arena. Child
- * links are stored in a shared children array: a node's children
- * occupy children[child_offset .. child_offset + n_fields]. Token payload
- * fields (token, pvalue, and any user fields) are only meaningful when
- * is_token is true.
- */
-private struct <%= @grammar.prefix %>node_data_t
 {
     <%= @grammar.prefix %>position_t position;
     <%= @grammar.prefix %>position_t end_position;
-    <%= @grammar.prefix %>node_id_t child_offset;
+    void *[0] fields;
-    ushort n_fields;
+}
-    bool is_token;
+
+/** AST node types. @{ */
+public struct <%= @grammar.ast_prefix %>Token<%= @grammar.ast_suffix %>
+{
+    /* ASTNode fields must be present in the same order here. */
+    <%= @grammar.prefix %>position_t position;
+    <%= @grammar.prefix %>position_t end_position;
     <%= @grammar.prefix %>token_t token;
     <%= @grammar.prefix %>value_t pvalue;
-<%= @grammar.token_user_fields %>
 }
 
-/** Tree node handle types. @{ */
+<%   @parser.rule_sets.each do |name, rule_set| %>
-
+<%     next if name.start_with?("$") %>
-/** Token tree node handle. */
+<%     next if rule_set.optional? %>
-public struct <%= @grammar.tree_prefix %>Token<%= @grammar.tree_suffix %>
+public struct <%= @grammar.ast_prefix %><%= name %><%= @grammar.ast_suffix %>
 {
-    private <%= @grammar.prefix %>context_t * __context;
+    <%= @grammar.prefix %>position_t position;
-    private <%= @grammar.prefix %>node_id_t __id;
+    <%= @grammar.prefix %>position_t end_position;
-
+<%     rule_set.ast_fields.each do |fields| %>
-    this(<%= @grammar.prefix %>context_t * context, <%= @grammar.prefix %>node_id_t id)
+    union
     {
-        this.__context = context;
-        this.__id = id;
-    }
-
-    /** Return whether this handle refers to a valid (non-null) node. */
-    @property bool valid()
-    {
-        return __id != 0u;
-    }
-
-    /** Return the node ID (for identity comparison). */
-    @property <%= @grammar.prefix %>node_id_t node_id()
-    {
-        return __id;
-    }
-
-    /** Access the underlying node record (token, pvalue, and user fields). */
-    @property ref <%= @grammar.prefix %>node_data_t __node()
-    {
-        return __context.<%= @grammar.prefix %>tree_nodes[__id];
-    }
-
-    alias __node this;
-}
-
-<%   tree_node_rule_sets.each do |rule_set| %>
-/** <%= rule_set.name %> tree node handle. */
-public struct <%= @grammar.tree_prefix %><%= rule_set.name %><%= @grammar.tree_suffix %>
-{
-    private <%= @grammar.prefix %>context_t * __context;
-    private <%= @grammar.prefix %>node_id_t __id;
-
-    this(<%= @grammar.prefix %>context_t * context, <%= @grammar.prefix %>node_id_t id)
-    {
-        this.__context = context;
-        this.__id = id;
-    }
-
-    /** Return whether this handle refers to a valid (non-null) node. */
-    @property bool valid()
-    {
-        return __id != 0u;
-    }
-
-    /** Return the node ID (for identity comparison). */
-    @property <%= @grammar.prefix %>node_id_t node_id()
-    {
-        return __id;
-    }
-
-    /** Text position of the first code point spanned by this node. */
-    @property <%= @grammar.prefix %>position_t position()
-    {
-        return __context.<%= @grammar.prefix %>tree_nodes[__id].position;
-    }
-
-    /** Text position of the last code point spanned by this node. */
-    @property <%= @grammar.prefix %>position_t end_position()
-    {
-        return __context.<%= @grammar.prefix %>tree_nodes[__id].end_position;
-    }
-
-    /** Number of child fields in this node. */
-    @property ushort n_fields()
-    {
-        return __id ? __context.<%= @grammar.prefix %>tree_nodes[__id].n_fields : cast(ushort)0u;
-    }
-<%     rule_set.tree_fields.each_with_index do |fields, i| %>
 <%       fields.each do |field_name, type| %>
-
+        <%= type %> * <%= field_name %>;
-    /** Access the <%= field_name %> child node. */
-    @property <%= type %> <%= field_name %>()
-    {
-        if (__id == 0u)
-        {
-            return <%= type %>(__context, 0u);
-        }
-        return <%= type %>(__context, __context.<%= @grammar.prefix %>tree_children[__context.<%= @grammar.prefix %>tree_nodes[__id].child_offset + <%= i %>u]);
-    }
 <%       end %>
+    }
 <%     end %>
 }
 
@@ -270,14 +172,8 @@ public struct <%= @grammar.prefix %>context_t
     /* Parser context data. */
 
     /** Parse result value. */
-<% if @grammar.tree %>
+<% if @grammar.ast %>
-    <%= @grammar.prefix %>node_id_t parse_result;
+    void * parse_result;
-
-    /** Tree node arena. Node ID 0 is reserved as the null node. */
-    <%= @grammar.prefix %>node_data_t[] <%= @grammar.prefix %>tree_nodes;
-
-    /** Shared tree child links. */
-    <%= @grammar.prefix %>node_id_t[] <%= @grammar.prefix %>tree_children;
 <% else %>
     <%= @grammar.prefix %>value_t parse_result;
 <% end %>
@@ -287,8 +183,6 @@ public struct <%= @grammar.prefix %>context_t
 
     /** User terminate code. */
     size_t user_terminate_code;
-
-<%= @grammar.context_user_fields %>
 }
 
 /**************************************************************************
@@ -328,52 +222,26 @@ private enum size_t INVALID_ID = cast(size_t)-1;
  *************************************************************************/
 
 /**
- * Allocate and initialize lexer/parser context structure.
+ * Initialize lexer/parser context structure.
- *
- * Deinitialize and deallocate with <%= @grammar.prefix %>context_delete().
  *
+ * @param[out] context
+ *   Lexer/parser context structure.
  * @param input
  *   Text input.
- *
- * @return Context structure for lexer/parser.
  */
-<%= @grammar.prefix %>context_t * <%= @grammar.prefix %>context_new(string input)
+public void <%= @grammar.prefix %>context_init(<%= @grammar.prefix %>context_t * context, string input)
 {
     /* New default-initialized context structure. */
-    <%= @grammar.prefix %>context_t * context = new <%= @grammar.prefix %>context_t;
+    <%= @grammar.prefix %>context_t newcontext;
 
     /* Lexer initialization. */
-    context.input = input;
+    newcontext.input = input;
-    context.text_position.row = 1u;
+    newcontext.text_position.row = 1u;
-    context.text_position.col = 1u;
+    newcontext.text_position.col = 1u;
-    context.mode = <%= @lexer.mode_id("default") %>;
+    newcontext.mode = <%= @lexer.mode_id("default") %>;
-<% if @grammar.tree %>
 
-    /* Reserve node ID 0 as the null tree node. */
+    /* Copy to the user's context structure. */
-    context.<%= @grammar.prefix %>tree_nodes = new <%= @grammar.prefix %>node_data_t[](1);
+    *context = newcontext;
-<% end %>
-
-    return context;
-}
-
-/**
- * Deinitialize and deallocate lexer/parser context structure.
- *
- * @param context
- *   Lexer/parser context structure allocated with <%= @grammar.prefix %>context_new().
- */
-void <%= @grammar.prefix %>context_delete(<%= @grammar.prefix %>context_t * context)
-{
-<% if @grammar.tree && @grammar.free_token_node != "" %>
-    foreach (ref node; context.<%= @grammar.prefix %>tree_nodes)
-    {
-        if (node.is_token)
-        {
-            <%= @grammar.prefix %>node_data_t * token_tree_node = &node;
-<%= expand_code(@grammar.free_token_node, false, nil, nil) %>
-        }
-    }
-<% end %>
 }
 
 /**************************************************************************
@@ -572,7 +440,7 @@ private immutable lexer_mode_t[] lexer_mode_table = [
  *   Lexer/parser context structure.
  * @param code_id
  *   The ID of the user code block to execute.
- * @param match_text
+ * @param match
  *   Matched text for this pattern.
  * @param out_token_info
  *   Lexer token info in progress.
@@ -581,7 +449,7 @@ private immutable lexer_mode_t[] lexer_mode_table = [
  *   not explicitly return a token.
  */
 private <%= @grammar.prefix %>token_t lexer_user_code(<%= @grammar.prefix %>context_t * context,
-    lexer_user_code_id_t code_id, string match_text,
+    lexer_user_code_id_t code_id, string match,
     <%= @grammar.prefix %>token_info_t * out_token_info)
 {
     switch (code_id)
@@ -759,27 +627,11 @@ private size_t attempt_lex_token(<%= @grammar.prefix %>context_t * context, <%=
     {
     case P_SUCCESS:
         <%= @grammar.prefix %>token_t token_to_accept = match_info.accepting_state.token;
-        /* Calculate the token length and start/end positions before invoking
-         * the lexer user code so that the user code can access them. The
-         * context input text position tracking is not updated until after the
-         * user code has run so that it is left unchanged if the user code
-         * requests to terminate the lexer. */
-        token_info.length = match_info.length;
-        if (match_info.end_delta_position.row != 0u)
-        {
-            token_info.end_position.row = token_info.position.row + match_info.end_delta_position.row;
-            token_info.end_position.col = match_info.end_delta_position.col;
-        }
-        else
-        {
-            token_info.end_position.row = token_info.position.row;
-            token_info.end_position.col = token_info.position.col + match_info.end_delta_position.col;
-        }
         if (match_info.accepting_state.code_id != INVALID_USER_CODE_ID)
         {
-            string match_text = context.input[context.input_index..(context.input_index + match_info.length)];
+            string match = context.input[context.input_index..(context.input_index + match_info.length)];
             <%= @grammar.prefix %>token_t user_code_token = lexer_user_code(context,
-                match_info.accepting_state.code_id, match_text, &token_info);
+                match_info.accepting_state.code_id, match, &token_info);
             /* A TERMINATE_TOKEN_ID return code from lexer_user_code() means
              * that the user code is requesting to terminate the lexer. */
             if (user_code_token == TERMINATE_TOKEN_ID)
@@ -813,6 +665,17 @@ private size_t attempt_lex_token(<%= @grammar.prefix %>context_t * context, <%=
             return P_DROP;
         }
         token_info.token = token_to_accept;
+        token_info.length = match_info.length;
+        if (match_info.end_delta_position.row != 0u)
+        {
+            token_info.end_position.row = token_info.position.row + match_info.end_delta_position.row;
+            token_info.end_position.col = match_info.end_delta_position.col;
+        }
+        else
+        {
+            token_info.end_position.row = token_info.position.row;
+            token_info.end_position.col = token_info.position.col + match_info.end_delta_position.col;
+        }
         *out_token_info = token_info;
         return P_SUCCESS;
 
@@ -934,7 +797,7 @@ private struct reduce_t
      * reduce action.
      */
     parser_state_id_t n_states;
-<% if @grammar.tree %>
+<% if @grammar.ast %>
 
     /**
      * Map of rule components to rule set child fields.
@@ -942,7 +805,7 @@ private struct reduce_t
     immutable(ushort) * rule_set_node_field_index_map;
 
     /**
-     * Number of rule set tree node fields.
+     * Number of rule set AST node fields.
      */
     ushort rule_set_node_field_array_size;
 
@@ -981,14 +844,12 @@ private struct state_value_t
     /** Parser state ID. */
     size_t state_id;
 
-<% if @grammar.tree %>
-    /** Tree node ID. */
-    <%= @grammar.prefix %>node_id_t node_id;
-<% else %>
-    <%= @grammar.prefix %>position_t position;
-    <%= @grammar.prefix %>position_t end_position;
     /** Parser value from this state. */
     <%= @grammar.prefix %>value_t pvalue;
+
+<% if @grammar.ast %>
+    /** AST node. */
+    void * ast_node;
 <% end %>
 
     this(size_t state_id)
@@ -1004,7 +865,7 @@ private immutable shift_t[] parser_shift_table = [
 <%   end %>
 ];
 
-<% if @grammar.tree %>
+<% if @grammar.ast %>
 <%   @grammar.rules.each do |rule| %>
 <%     unless rule.flat_rule_set_node_field_index_map? %>
 immutable ushort[<%= rule.rule_set_node_field_index_map.size %>] r_<%= rule.name.gsub("$", "_") %><%= rule.id %>_node_field_index_map = [<%= rule.rule_set_node_field_index_map.map {|v| v.to_s}.join(", ") %>];
@@ -1019,14 +880,14 @@ private immutable reduce_t[] parser_reduce_table = [
         <%= reduce[:token_id] %>u, /* Token: <%= reduce[:token] ? reduce[:token].name : "(any)" %> */
         <%= reduce[:rule_id] %>u, /* Rule ID */
         <%= reduce[:rule_set_id] %>u, /* Rule set ID (<%= reduce[:rule].rule_set.name %>) */
-<%     if @grammar.tree %>
+<%     if @grammar.ast %>
         <%= reduce[:n_states] %>u, /* Number of states */
 <%       if reduce[:rule].flat_rule_set_node_field_index_map? %>
         null, /* No rule set node field index map (flat map) */
 <%       else %>
         &r_<%= reduce[:rule].name.gsub("$", "_") %><%= reduce[:rule].id %>_node_field_index_map[0], /* Rule set node field index map */
 <%       end %>
-        <%= reduce[:rule].rule_set.tree_fields.size %>, /* Number of tree fields */
+        <%= reduce[:rule].rule_set.ast_fields.size %>, /* Number of AST fields */
         <%= reduce[:propagate_optional_target] %>), /* Propagate optional target? */
 <%     else %>
         <%= reduce[:n_states] %>u), /* Number of states */
@@ -1041,56 +902,7 @@ private immutable parser_state_t[] parser_state_table = [
 <%   end %>
 ];
 
-<% unless @grammar.tree %>
+<% unless @grammar.ast %>
-/**
- * Get the rule position (start or end) for the currently matched rule.
- */
-private <%= @grammar.prefix %>position_t get_rule_position(state_value_t[] statevalues, size_t i, size_t n_states, bool get_end)
-{
-    if (n_states > 0u)
-    {
-        if (i == 0u)
-        {
-            if (get_end)
-            {
-                for (size_t j = 0u; j < n_states; j++)
-                {
-                    state_value_t * sv = &statevalues[$-1-j];
-                    if (sv.end_position.valid)
-                    {
-                        return sv.end_position;
-                    }
-                }
-            }
-            else
-            {
-                for (size_t j = 0u; j < n_states; j++)
-                {
-                    state_value_t * sv = &statevalues[$-n_states+j];
-                    if (sv.position.valid)
-                    {
-                        return sv.position;
-                    }
-                }
-            }
-        }
-        else
-        {
-            if (get_end)
-            {
-                return statevalues[$-1-n_states+i].end_position;
-            }
-            else
-            {
-                return statevalues[$-1-n_states+i].position;
-            }
-        }
-    }
-    return <%= @grammar.prefix %>position_t.INVALID;
-}
-<% end %>
-
-<% if !@grammar.tree || @grammar.parser_user_code_used? %>
 /**
  * Execute user code associated with a parser rule.
  *
@@ -1101,7 +913,7 @@ private <%= @grammar.prefix %>position_t get_rule_position(state_value_t[] state
  * @retval P_USER_TERMINATED
  *   User requested to terminate parsing.
  */
-private size_t parser_user_code(<%= @grammar.tree ? "#{@grammar.prefix}node_id_t _node_id" : "#{@grammar.prefix}value_t * _pvalue" %>, uint rule, state_value_t[] statevalues, uint n_states, <%= @grammar.prefix %>context_t * context)
+private size_t parser_user_code(<%= @grammar.prefix %>value_t * _pvalue, uint rule, state_value_t[] statevalues, uint n_states, <%= @grammar.prefix %>context_t * context)
 {
     switch (rule)
     {
@@ -1151,7 +963,7 @@ private size_t check_shift(size_t state_id, size_t symbol_id)
  * @param token
  *   Incoming token.
  *
- * @return Reduce table index to reduce with, or INVALID_ID if none.
+ * @return State to reduce to, or INVALID_ID if none.
  */
 private size_t check_reduce(size_t state_id, <%= @grammar.prefix %>token_t token)
 {
@@ -1173,16 +985,8 @@ private size_t check_reduce(size_t state_id, <%= @grammar.prefix %>token_t token
  *
  * @param context
  *   Lexer/parser context structure.
- * @param start_state_id
+ * @start_state_id
  *   ID of the state in which to start.
- * @param start_rule_set_id
- *   Rule set ID for the requested start rule. Only used when
- *   @p follow_tokens is non-empty, to gate follow-token shift success.
- * @param follow_tokens
- *   Optional slice of caller-provided follow tokens (tokens expected to
- *   appear immediately after the start rule in some outer context). Used to
- *   drive the "parse inner" retry logic. May be null/empty for a standard
- *   parse.
  *
  * @retval P_SUCCESS
  *   The parser successfully matched the input text. The parse result value
@@ -1195,46 +999,29 @@ private size_t check_reduce(size_t state_id, <%= @grammar.prefix %>token_t token
  * @reval P_UNEXPECTED_INPUT
  *   Input text does not match any lexer pattern.
  */
-private size_t parse_from(<%= @grammar.prefix %>context_t * context, size_t start_state_id,
+private size_t parse_from(<%= @grammar.prefix %>context_t * context, size_t start_state_id)
-    size_t start_rule_set_id,
-    const(<%= @grammar.prefix %>token_t)[] follow_tokens)
 {
     <%= @grammar.prefix %>token_info_t token_info;
     <%= @grammar.prefix %>token_t token = INVALID_TOKEN_ID;
     state_value_t[] statevalues = new state_value_t[](1);
     statevalues[0].state_id = start_state_id;
     size_t reduced_rule_set = INVALID_ID;
-    size_t last_shifted_rule_set_id = INVALID_ID;
+<% if @grammar.ast %>
-<% if @grammar.tree %>
+    void * reduced_parser_node;
-    <%= @grammar.prefix %>node_id_t reduced_parser_node;
 <% else %>
-    <%= @grammar.prefix %>position_t reduced_position;
-    <%= @grammar.prefix %>position_t reduced_end_position;
     <%= @grammar.prefix %>value_t reduced_parser_value;
 <% end %>
     for (;;)
     {
         if (token == INVALID_TOKEN_ID)
         {
-            size_t lexer_result = <%= lex_fn %>(context, &token_info);
+            size_t lexer_result = <%= @grammar.prefix %>lex(context, &token_info);
             if (lexer_result != P_SUCCESS)
             {
                 return lexer_result;
             }
             token = token_info.token;
         }
-        /* For a "parse inner" operation, determine once per iteration whether
-         * the current token is a member of the caller-provided follow token
-         * set. Used by both the shift-side and reduce-side retries below. */
-        bool token_is_follow = false;
-        foreach (eof_token; follow_tokens)
-        {
-            if (token == eof_token)
-            {
-                token_is_follow = true;
-                break;
-            }
-        }
         size_t shift_state = INVALID_ID;
         if (reduced_rule_set != INVALID_ID)
         {
@@ -1246,67 +1033,25 @@ private size_t parse_from(<%= @grammar.prefix %>context_t * context, size_t star
             if ((shift_state != INVALID_ID) && (token == TOKEN___EOF))
             {
                 /* Successful parse. */
-<% if @grammar.tree %>
+<% if @grammar.ast %>
-                context.parse_result = statevalues[$-1].node_id;
+                context.parse_result = statevalues[$-1].ast_node;
 <% else %>
                 context.parse_result = statevalues[$-1].pvalue;
 <% end %>
                 return P_SUCCESS;
             }
-            if ((shift_state == INVALID_ID) && token_is_follow)
-            {
-                /* For a "parse inner" operation, if the incoming token is one
-                 * of the caller's follow tokens, retry the shift as
-                 * TOKEN___EOF. Only consider the parse complete if the reduced
-                 * start rule is the only thing on the parse stack (i.e. the
-                 * initial state plus a single shifted start rule set entry). */
-                size_t retry_shift_state = check_shift(statevalues[$-1].state_id, TOKEN___EOF);
-                if ((retry_shift_state != INVALID_ID) &&
-                    (statevalues.length == 2u) &&
-                    (last_shifted_rule_set_id == start_rule_set_id))
-                {
-                    /* Successful parse via follow token. Rewind the input
-                     * position so that the follow token is not consumed from
-                     * the input stream and remains available for a subsequent
-                     * call to <%= @grammar.prefix %>lex() or a
-                     * <%= @grammar.prefix %>parse*() function. */
-                    context.input_index -= token_info.length;
-                    context.text_position = token_info.position;
-<% if @grammar.tree %>
-                    context.parse_result = statevalues[$-1].node_id;
-<% else %>
-                    context.parse_result = statevalues[$-1].pvalue;
-<% end %>
-                    return P_SUCCESS;
-                }
-            }
         }
         if (shift_state != INVALID_ID)
         {
-            /* We have something to shift. Track the last shifted rule set ID
+            /* We have something to shift. */
-             * (INVALID_ID if we just shifted a token) so the follow-token
-             * shift retry can gate success on the reduced start rule being the
-             * only thing on top of the initial state. */
-            last_shifted_rule_set_id = reduced_rule_set;
             statevalues ~= state_value_t(shift_state);
             if (reduced_rule_set == INVALID_ID)
             {
                 /* We shifted a token, mark it consumed. */
-<% if @grammar.tree %>
+<% if @grammar.ast %>
-                <%= @grammar.prefix %>node_id_t token_node_id = cast(<%= @grammar.prefix %>node_id_t)context.<%= @grammar.prefix %>tree_nodes.length;
+                <%= @grammar.ast_prefix %>Token<%= @grammar.ast_suffix %> * token_ast_node = new <%= @grammar.ast_prefix %>Token<%= @grammar.ast_suffix %>(token_info.position, token_info.end_position, token, token_info.pvalue);
-                context.<%= @grammar.prefix %>tree_nodes ~= <%= @grammar.prefix %>node_data_t.init;
+                statevalues[$-1].ast_node = token_ast_node;
-                <%= @grammar.prefix %>node_data_t * token_tree_node = &context.<%= @grammar.prefix %>tree_nodes[token_node_id];
-                token_tree_node.position = token_info.position;
-                token_tree_node.end_position = token_info.end_position;
-                token_tree_node.n_fields = 0u;
-                token_tree_node.is_token = true;
-                token_tree_node.token = token;
-                token_tree_node.pvalue = token_info.pvalue;
-<%= expand_code(@grammar.on_token_node, false, nil, nil) %>
-                statevalues[$-1].node_id = token_node_id;
 <% else %>
-                statevalues[$-1].position = token_info.position;
-                statevalues[$-1].end_position = token_info.end_position;
                 statevalues[$-1].pvalue = token_info.pvalue;
 <% end %>
                 token = INVALID_TOKEN_ID;
@@ -1314,12 +1059,10 @@ private size_t parse_from(<%= @grammar.prefix %>context_t * context, size_t star
             else
             {
                 /* We shifted a RuleSet. */
-<% if @grammar.tree %>
+<% if @grammar.ast %>
-                statevalues[$-1].node_id = reduced_parser_node;
+                statevalues[$-1].ast_node = reduced_parser_node;
 <% else %>
                 statevalues[$-1].pvalue = reduced_parser_value;
-                statevalues[$-1].position = reduced_position;
-                statevalues[$-1].end_position = reduced_end_position;
                 <%= @grammar.prefix %>value_t new_parse_result;
                 reduced_parser_value = new_parse_result;
 <% end %>
@@ -1329,78 +1072,60 @@ private size_t parse_from(<%= @grammar.prefix %>context_t * context, size_t star
         }
 
         size_t reduce_index = check_reduce(statevalues[$-1].state_id, token);
-        if ((reduce_index == INVALID_ID) && token_is_follow)
-        {
-            /* For a "parse inner" operation, if the incoming token is one of
-             * the caller's follow tokens, retry the reduce lookup as
-             * TOKEN___EOF. Whatever reduce_index results (if any) is used
-             * regardless of which rule set it reduces to; this allows chains
-             * of reductions leading up to the start rule. */
-            reduce_index = check_reduce(statevalues[$-1].state_id, TOKEN___EOF);
-        }
         if (reduce_index != INVALID_ID)
         {
             /* We have something to reduce. */
-<% if @grammar.tree %>
+<% if @grammar.ast %>
             if (parser_reduce_table[reduce_index].propagate_optional_target)
             {
-                reduced_parser_node = statevalues[$ - 1].node_id;
+                reduced_parser_node = statevalues[$ - 1].ast_node;
             }
             else if (parser_reduce_table[reduce_index].n_states > 0)
             {
-                ushort n_fields = parser_reduce_table[reduce_index].rule_set_node_field_array_size;
+                size_t n_fields = parser_reduce_table[reduce_index].rule_set_node_field_array_size;
-                /* Reserve child slots. New slots are zero-initialized
+                size_t node_size = ASTNode.sizeof + n_fields * (void *).sizeof;
-                 * (null node ID) so absent optional children remain null. */
+                ASTNode * node = cast(ASTNode *)malloc(node_size);
-                <%= @grammar.prefix %>node_id_t child_offset = cast(<%= @grammar.prefix %>node_id_t)context.<%= @grammar.prefix %>tree_children.length;
+                GC.addRange(node, node_size);
-                context.<%= @grammar.prefix %>tree_children.length += n_fields;
+                node.position = <%= @grammar.prefix %>position_t.INVALID;
+                node.end_position = <%= @grammar.prefix %>position_t.INVALID;
+                foreach (i; 0..n_fields)
+                {
+                    node.fields[i] = null;
+                }
                 if (parser_reduce_table[reduce_index].rule_set_node_field_index_map is null)
                 {
                     foreach (i; 0..parser_reduce_table[reduce_index].n_states)
                     {
-                        context.<%= @grammar.prefix %>tree_children[child_offset + i] = statevalues[$ - parser_reduce_table[reduce_index].n_states + i].node_id;
+                        node.fields[i] = statevalues[$ - parser_reduce_table[reduce_index].n_states + i].ast_node;
                     }
                 }
                 else
                 {
                     foreach (i; 0..parser_reduce_table[reduce_index].n_states)
                     {
-                        context.<%= @grammar.prefix %>tree_children[child_offset + parser_reduce_table[reduce_index].rule_set_node_field_index_map[i]] = statevalues[$ - parser_reduce_table[reduce_index].n_states + i].node_id;
+                        node.fields[parser_reduce_table[reduce_index].rule_set_node_field_index_map[i]] = statevalues[$ - parser_reduce_table[reduce_index].n_states + i].ast_node;
                     }
                 }
-                <%= @grammar.prefix %>node_id_t node_id = cast(<%= @grammar.prefix %>node_id_t)context.<%= @grammar.prefix %>tree_nodes.length;
-                context.<%= @grammar.prefix %>tree_nodes ~= <%= @grammar.prefix %>node_data_t.init;
-                <%= @grammar.prefix %>node_data_t * node = &context.<%= @grammar.prefix %>tree_nodes[node_id];
-                node.position = <%= @grammar.prefix %>position_t.INVALID;
-                node.end_position = <%= @grammar.prefix %>position_t.INVALID;
-                node.child_offset = child_offset;
-                node.n_fields = n_fields;
-                node.is_token = false;
                 bool position_found = false;
                 foreach (i; 0..n_fields)
                 {
-                    <%= @grammar.prefix %>node_id_t child_id = context.<%= @grammar.prefix %>tree_children[child_offset + i];
+                    ASTNode * child = cast(ASTNode *)node.fields[i];
-                    if (child_id != 0u && context.<%= @grammar.prefix %>tree_nodes[child_id].position.valid)
+                    if (child && child.position.valid)
                     {
                         if (!position_found)
                         {
-                            node.position = context.<%= @grammar.prefix %>tree_nodes[child_id].position;
+                            node.position = child.position;
                             position_found = true;
                         }
-                        node.end_position = context.<%= @grammar.prefix %>tree_nodes[child_id].end_position;
+                        node.end_position = child.end_position;
                     }
                 }
-                reduced_parser_node = node_id;
+                reduced_parser_node = node;
             }
             else
             {
-                reduced_parser_node = 0u;
+                reduced_parser_node = null;
             }
-<%   if @grammar.parser_user_code_used? %>
-            if (parser_user_code(reduced_parser_node, parser_reduce_table[reduce_index].rule, statevalues, parser_reduce_table[reduce_index].n_states, context) == P_USER_TERMINATED)
-            {
-                return P_USER_TERMINATED;
-            }
-<%   end %>
 <% else %>
             <%= @grammar.prefix %>value_t reduced_parser_value2;
             if (parser_user_code(&reduced_parser_value2, parser_reduce_table[reduce_index].rule, statevalues, parser_reduce_table[reduce_index].n_states, context) == P_USER_TERMINATED)
@@ -1408,16 +1133,6 @@ private size_t parse_from(<%= @grammar.prefix %>context_t * context, size_t star
                 return P_USER_TERMINATED;
             }
             reduced_parser_value = reduced_parser_value2;
-            if (parser_reduce_table[reduce_index].n_states > 0u)
-            {
-                reduced_position = get_rule_position(statevalues, 0u, parser_reduce_table[reduce_index].n_states, false);
-                reduced_end_position = get_rule_position(statevalues, 0u, parser_reduce_table[reduce_index].n_states, true);
-            }
-            else
-            {
-                reduced_position = <%= @grammar.prefix %>position_t.INVALID;
-                reduced_end_position = <%= @grammar.prefix %>position_t.INVALID;
-            }
 <% end %>
             reduced_rule_set = parser_reduce_table[reduce_index].rule_set;
             statevalues.length -= parser_reduce_table[reduce_index].n_states;
@@ -1437,20 +1152,14 @@ private size_t parse_from(<%= @grammar.prefix %>context_t * context, size_t star
 
 public size_t <%= @grammar.prefix %>parse(<%= @grammar.prefix %>context_t * context)
 {
-    return parse_from(context, 0u, <%= @parser.rule_sets[@grammar.start_rules[0]].id %>u, null);
+    return parse_from(context, 0u);
 }
 
 <% @grammar.start_rules.each_with_index do |start_rule, i| %>
 
 public size_t <%= @grammar.prefix %>parse_<%= start_rule %>(<%= @grammar.prefix %>context_t * context)
 {
-    return parse_from(context, <%= i %>u, <%= @parser.rule_sets[start_rule].id %>u, null);
+    return parse_from(context, <%= i %>u);
-}
-
-public size_t <%= @grammar.prefix %>parse_inner_<%= start_rule %>(<%= @grammar.prefix %>context_t * context,
-    const(<%= @grammar.prefix %>token_t)[] follow_tokens)
-{
-    return parse_from(context, <%= i %>u, <%= @parser.rule_sets[start_rule].id %>u, follow_tokens);
 }
 <% end %>
 
@@ -1462,15 +1171,15 @@ public size_t <%= @grammar.prefix %>parse_inner_<%= start_rule %>(<%= @grammar.p
  *
  * @return Parse result value.
  */
-<% if @grammar.tree %>
+<% if @grammar.ast %>
-public <%= @grammar.tree_prefix %><%= @grammar.start_rules[0] %><%= @grammar.tree_suffix %> <%= @grammar.prefix %>result(<%= @grammar.prefix %>context_t * context)
+public <%= @grammar.ast_prefix %><%= @grammar.start_rules[0] %><%= @grammar.ast_suffix %> * <%= @grammar.prefix %>result(<%= @grammar.prefix %>context_t * context)
 {
-    return <%= @grammar.tree_prefix %><%= @grammar.start_rules[0] %><%= @grammar.tree_suffix %>(context, context.parse_result);
+    return cast(<%= @grammar.ast_prefix %><%= @grammar.start_rules[0] %><%= @grammar.ast_suffix %> *)context.parse_result;
 }
 <%   @grammar.start_rules.each_with_index do |start_rule, i| %>
-public <%= @grammar.tree_prefix %><%= start_rule %><%= @grammar.tree_suffix %> <%= @grammar.prefix %>result_<%= start_rule %>(<%= @grammar.prefix %>context_t * context)
+public <%= @grammar.ast_prefix %><%= start_rule %><%= @grammar.ast_suffix %> * <%= @grammar.prefix %>result_<%= start_rule %>(<%= @grammar.prefix %>context_t * context)
 {
-    return <%= @grammar.tree_prefix %><%= start_rule %><%= @grammar.tree_suffix %>(context, context.parse_result);
+    return cast(<%= @grammar.ast_prefix %><%= start_rule %><%= @grammar.ast_suffix %> *)context.parse_result;
 }
 <%   end %>
 <% else %>
@@ -1499,58 +1208,6 @@ public <%= @grammar.prefix %>position_t <%= @grammar.prefix %>position(<%= @gram
     return context.text_position;
 }
 
-/**
- * Set the current text input position.
- *
- * This can be used to set the initial text position to something other than
- * (1, 1) for a nested parse operation so that error positions reported by
- * subsequent lexer/parser calls are relative to a larger enclosing document.
- *
- * @param context
- *   Lexer/parser context structure.
- * @param position
- *   Text position to set.
- */
-public void <%= @grammar.prefix %>set_position(<%= @grammar.prefix %>context_t * context, <%= @grammar.prefix %>position_t position)
-{
-    context.text_position = position;
-}
-
-/**
- * Get the current input text byte offset.
- *
- * @param context
- *   Lexer/parser context structure.
- *
- * @return Current input text byte offset (measured from the start of the
- *   input text passed to <%= @grammar.prefix %>context_new()).
- */
-public size_t <%= @grammar.prefix %>input_index(<%= @grammar.prefix %>context_t * context)
-{
-    return context.input_index;
-}
-
-/**
- * Set the current input text byte offset.
- *
- * This moves the lexer's read cursor to the given byte offset (measured from
- * the start of the input text passed to <%= @grammar.prefix %>context_new()).
- * It can be used together with <%= @grammar.prefix %>set_position() to rewind
- * the input part-way through a parse in order to re-read an earlier section of
- * the input. The byte offset is not validated; the caller is responsible for
- * providing an offset within the bounds of the input text. A value previously
- * returned by <%= @grammar.prefix %>input_index() is a suitable argument.
- *
- * @param context
- *   Lexer/parser context structure.
- * @param input_index
- *   Input text byte offset to set.
- */
-public void <%= @grammar.prefix %>set_input_index(<%= @grammar.prefix %>context_t * context, size_t input_index)
-{
-    context.input_index = input_index;
-}
-
 /**
  * Get the user terminate code.
  *

assets/parser.h.erb

@@ -8,9 +8,6 @@
 
 #include <stdint.h>
 #include <stddef.h>
-<% if @cpp %>
-#include <vector>
-<% end %>
 
 /**************************************************************************
  * Public types
@@ -48,10 +45,10 @@ typedef uint32_t <%= @grammar.prefix %>code_point_t;
  */
 typedef struct
 {
-    /** Input text row (1-based). */
+    /** Input text row (0-based). */
     uint32_t row;
 
-    /** Input text column (1-based). */
+    /** Input text column (0-based). */
     uint32_t col;
 } <%= @grammar.prefix %>position_t;
 
@@ -61,7 +58,7 @@ typedef struct
 /** User header code blocks. */
 <%= @grammar.code_blocks.fetch("header", "") %>
 
-<% if @grammar.tree %>
+<% if @grammar.ast %>
 /** Parser values type. */
 typedef <%= @grammar.ptype %> <%= @grammar.prefix %>value_t;
 <% else %>
@@ -72,48 +69,49 @@ typedef union
     <%= typestring %> v_<%= name %>;
 <%   end %>
 } <%= @grammar.prefix %>value_t;
-
-/** Parser value constructor(s). */
-<%   @grammar.ptypes.each do |name, typestring| %>
-static inline <%= @grammar.prefix %>value_t <%= @grammar.prefix %>value<%= name == "default" ? "" : "_#{name}" %>(<%= typestring %> v)
-{
-    return (<%= @grammar.prefix %>value_t){.v_<%= name %> = v};
-}
 <% end %>
 
-/** Parser value accessor(s). */
+<% if @grammar.ast %>
-<%   @grammar.ptypes.each do |name, typestring| %>
+/** AST node types. @{ */
-static inline <%= typestring %> <%= @grammar.prefix %>value_get<%= name == "default" ? "" : "_#{name}" %>(<%= @grammar.prefix %>value_t const * pvalue)
+typedef struct <%= @grammar.ast_prefix %>Token<%= @grammar.ast_suffix %>
-{
-    return pvalue->v_<%= name %>;
-}
-<%   end %>
-<% end %>
-
-<% if @grammar.tree %>
-/** Tree node ID type (index into the context node arena). ID 0 is null. */
-typedef uint32_t <%= @grammar.prefix %>node_id_t;
-
-/**
- * Tree node record.
- *
- * All tree nodes are stored contiguously in the context node arena. Child
- * links are stored in a shared children array: a node's children
- * occupy children[child_offset .. child_offset + n_fields]. Token payload
- * fields (token, pvalue, and any user fields) are only meaningful when
- * is_token is nonzero.
- */
-typedef struct
 {
+    <% # ASTNode fields must be present in the same order here. # %>
     <%= @grammar.prefix %>position_t position;
     <%= @grammar.prefix %>position_t end_position;
-    <%= @grammar.prefix %>node_id_t child_offset;
     uint16_t n_fields;
     uint8_t is_token;
     <%= @grammar.prefix %>token_t token;
     <%= @grammar.prefix %>value_t pvalue;
-<%= @grammar.token_user_fields %>
+} <%= @grammar.ast_prefix %>Token<%= @grammar.ast_suffix %>;
-} <%= @grammar.prefix %>node_data_t;
+
+<%   @parser.rule_sets.each do |name, rule_set| %>
+<%     next if name.start_with?("$") %>
+<%     next if rule_set.optional? %>
+struct <%= name %>;
+<%   end %>
+
+<%   @parser.rule_sets.each do |name, rule_set| %>
+<%     next if name.start_with?("$") %>
+<%     next if rule_set.optional? %>
+typedef struct <%= @grammar.ast_prefix %><%= name %><%= @grammar.ast_suffix %>
+{
+    <% # ASTNode fields must be present in the same order here. # %>
+    <%= @grammar.prefix %>position_t position;
+    <%= @grammar.prefix %>position_t end_position;
+    uint16_t n_fields;
+    uint8_t is_token;
+<%     rule_set.ast_fields.each do |fields| %>
+    union
+    {
+<%       fields.each do |field_name, type| %>
+        struct <%= type %> * <%= field_name %>;
+<%       end %>
+    };
+<%     end %>
+} <%= @grammar.ast_prefix %><%= name %><%= @grammar.ast_suffix %>;
+
+<%   end %>
+/** @} */
 <% end %>
 
 /** Lexed token information. */
@@ -135,19 +133,13 @@ typedef struct
     <%= @grammar.prefix %>value_t pvalue;
 } <%= @grammar.prefix %>token_info_t;
 
-typedef struct <%= @grammar.prefix %>context_s <%= @grammar.prefix %>context_t;
-
-<% if @grammar.tree %>
-<%= c_tree_handle_types_header %>
-
-<% end %>
 /**
  * Lexer and parser context.
  *
  * The user must allocate an instance of this structure and pass it to any
  * public API function.
  */
-struct <%= @grammar.prefix %>context_s
+typedef struct
 {
     /* Lexer context data. */
 
@@ -169,26 +161,8 @@ struct <%= @grammar.prefix %>context_s
     /* Parser context data. */
 
     /** Parse result value. */
-<% if @grammar.tree %>
+<% if @grammar.ast %>
-    <%= @grammar.prefix %>node_id_t parse_result;
+    void * parse_result;
-
-<%   if @cpp %>
-    /** Tree node arena. Node ID 0 is reserved as the null node. */
-    std::vector<<%= @grammar.prefix %>node_data_t> <%= @grammar.prefix %>tree_nodes;
-
-    /** Shared tree child links. */
-    std::vector<<%= @grammar.prefix %>node_id_t> <%= @grammar.prefix %>tree_children;
-<%   else %>
-    /** Tree node arena. Node ID 0 is reserved as the null node. */
-    <%= @grammar.prefix %>node_data_t * <%= @grammar.prefix %>tree_nodes;
-    size_t <%= @grammar.prefix %>tree_nodes_length;
-    size_t <%= @grammar.prefix %>tree_nodes_capacity;
-
-    /** Shared tree child links. */
-    <%= @grammar.prefix %>node_id_t * <%= @grammar.prefix %>tree_children;
-    size_t <%= @grammar.prefix %>tree_children_length;
-    size_t <%= @grammar.prefix %>tree_children_capacity;
-<%   end %>
 <% else %>
     <%= @grammar.prefix %>value_t parse_result;
 <% end %>
@@ -198,13 +172,7 @@ struct <%= @grammar.prefix %>context_s
 
     /** User terminate code. */
     size_t user_terminate_code;
-
+} <%= @grammar.prefix %>context_t;
-<%= @grammar.context_user_fields %>
-};
-
-<% if @grammar.tree %>
-<%= c_tree_types_header %>
-<% end %>
 
 /**************************************************************************
  * Public data
@@ -213,9 +181,7 @@ struct <%= @grammar.prefix %>context_s
 /** Token names. */
 extern const char * <%= @grammar.prefix %>token_names[];
 
-<%= @grammar.prefix %>context_t * <%= @grammar.prefix %>context_new(uint8_t const * input, size_t input_length);
+void <%= @grammar.prefix %>context_init(<%= @grammar.prefix %>context_t * context, uint8_t const * input, size_t input_length);
-
-void <%= @grammar.prefix %>context_delete(<%= @grammar.prefix %>context_t * context);
 
 size_t <%= @grammar.prefix %>decode_code_point(uint8_t const * input, size_t input_length,
         <%= @grammar.prefix %>code_point_t * out_code_point, uint8_t * out_code_point_length);
@@ -225,14 +191,12 @@ size_t <%= @grammar.prefix %>lex(<%= @grammar.prefix %>context_t * context, <%=
 size_t <%= @grammar.prefix %>parse(<%= @grammar.prefix %>context_t * context);
 <%   @grammar.start_rules.each_with_index do |start_rule, i| %>
 size_t <%= @grammar.prefix %>parse_<%= start_rule %>(<%= @grammar.prefix %>context_t * context);
-size_t <%= @grammar.prefix %>parse_inner_<%= start_rule %>(<%= @grammar.prefix %>context_t * context,
-        <%= @grammar.prefix %>token_t const * follow_tokens, size_t n_follow_tokens);
 <%   end %>
 
-<% if @grammar.tree %>
+<% if @grammar.ast %>
-<%= h_type(@grammar.start_rules[0]) %> <%= @grammar.prefix %>result(<%= @grammar.prefix %>context_t * context);
+<%= @grammar.ast_prefix %><%= @grammar.start_rules[0] %><%= @grammar.ast_suffix %> * <%= @grammar.prefix %>result(<%= @grammar.prefix %>context_t * context);
 <%   @grammar.start_rules.each_with_index do |start_rule, i| %>
-<%= h_type(start_rule) %> <%= @grammar.prefix %>result_<%= start_rule %>(<%= @grammar.prefix %>context_t * context);
+<%= @grammar.ast_prefix %><%= start_rule %><%= @grammar.ast_suffix %> * <%= @grammar.prefix %>result_<%= start_rule %>(<%= @grammar.prefix %>context_t * context);
 <%   end %>
 <% else %>
 <%= start_rule_type[1] %> <%= @grammar.prefix %>result(<%= @grammar.prefix %>context_t * context);
@@ -241,14 +205,15 @@ size_t <%= @grammar.prefix %>parse_inner_<%= start_rule %>(<%= @grammar.prefix %
 <%   end %>
 <% end %>
 
+<% if @grammar.ast %>
+void <%= @grammar.prefix %>free_ast(<%= @grammar.ast_prefix %><%= @grammar.start_rules[0] %><%= @grammar.ast_suffix %> * ast);
+<%   @grammar.start_rules.each_with_index do |start_rule, i| %>
+void <%= @grammar.prefix %>free_ast_<%= start_rule %>(<%= @grammar.ast_prefix %><%= start_rule %><%= @grammar.ast_suffix %> * ast);
+<%   end %>
+<% end %>
+
 <%= @grammar.prefix %>position_t <%= @grammar.prefix %>position(<%= @grammar.prefix %>context_t * context);
 
-void <%= @grammar.prefix %>set_position(<%= @grammar.prefix %>context_t * context, <%= @grammar.prefix %>position_t position);
-
-size_t <%= @grammar.prefix %>input_index(<%= @grammar.prefix %>context_t * context);
-
-void <%= @grammar.prefix %>set_input_index(<%= @grammar.prefix %>context_t * context, size_t input_index);
-
 size_t <%= @grammar.prefix %>user_terminate_code(<%= @grammar.prefix %>context_t * context);
 
 <%= @grammar.prefix %>token_t <%= @grammar.prefix %>token(<%= @grammar.prefix %>context_t * context);

extra/vim/syntax/propane.vim

@@ -7,46 +7,8 @@ if exists("b:current_syntax")
   finish
 endif
 
-" Guess the language of the user code blocks from their contents so that the
-" matching syntax file can be included below. b:propane_subtype may also be set
-" before this file is sourced to select the language explicitly.
 if !exists("b:propane_subtype")
-  " Rust markers. Each keyword requires the syntax that follows it in Rust so
-  " that a plain identifier of the same name in another language does not match
-  " (`int fn = 3;' in C, for example). Type names are only accepted within a
-  " `ptype' statement for the same reason.
-  let s:rust = '\<let\s\+\%(mut\s\+\)\?\w'
-  let s:rust .= '\|\<fn\s\+\w\+\s*('
-  let s:rust .= '\|&mut\>\|\<pub\s\+\w\|\<impl\s\+\w'
-  let s:rust .= '\|#\[\|\<use\s\+\%(std\|core\)::'
-  let s:rust .= '\|\<ptype\>[^;]*\<\%(isize\|usize\|i8\|i16\|i32\|i64\|i128'
-  let s:rust .= '\|u8\|u16\|u32\|u64\|u128\|f32\|f64\|String\)\>'
-  " D markers. These are spellings that have no valid C, C++, or Rust
-  " equivalent, so `import' is deliberately not among them: it is a D keyword
-  " but is also a C++20 module declaration.
-  let s:d = '\<foreach\%(_reverse\)\?\s*([^)]*;'
-  let s:d .= '\|\~=\|\<static\s\+if\s*(\|\<version\s*(\s*\w\+\s*)'
-  let s:d .= '\|\<scope\s*(\s*\%(exit\|failure\|success\)\s*)'
-  let s:d .= '\|\<\%(unittest\|mixin\|immutable\|__gshared\|invariant\)\>'
-  let s:d .= '\|\<alias\s\+\w\+\s*=\|\<enum\s\+\w\+\s*='
-  let s:d .= '\|@\%(property\|safe\|trusted\|system\|nogc\|disable\)\>'
-  let s:d .= '\|\<is\s\+null\>\|\<cast\s*(\s*\w\+\s*)'
-  let s:d .= '\|\<write\%(ln\|fln\|f\)\s*('
-  let s:d .= '\|\<\%(dchar\|dstring\|wstring\|cent\|ucent\)\>'
-  " A module import on its own is ambiguous between D and C++20, so only take
-  " it as D when nothing else in the file looks like C++.
-  let s:import = '\<import\s\+[A-Za-z_][A-Za-z0-9_.]*\s*;'
-  let s:cpp = '::\|\<template\s*<\|\<namespace\>\|\<nullptr\>\|#include\s*[<"]'
-  if search(s:rust, 'nw') > 0
-    let b:propane_subtype = "rust"
-  elseif search(s:d, 'nw') > 0
   let b:propane_subtype = "d"
-  elseif search(s:import, 'nw') > 0 && search(s:cpp, 'nw') == 0
-    let b:propane_subtype = "d"
-  else
-    let b:propane_subtype = "cpp"
-  endif
-  unlet s:rust s:d s:import s:cpp
 endif
 
 exe "syn include @propaneTarget syntax/".b:propane_subtype.".vim"
@@ -54,32 +16,18 @@ exe "syn include @propaneTarget syntax/".b:propane_subtype.".vim"
 syn region propaneTarget matchgroup=propaneDelimiter start="<<" end=">>$" contains=@propaneTarget keepend
 
 syn match propaneComment "#.*"
+syn match propaneOperator "->"
 syn match propaneFieldAlias ":[a-zA-Z0-9_]\+" contains=propaneFieldOperator
 syn match propaneFieldOperator ":" contained
 syn match propaneOperator "?"
-" The right-hand side of a rule (after '->' up to '<<' or ';') lists symbol
+syn keyword propaneKeyword ast ast_prefix ast_suffix drop module prefix ptype start token tokenid
-" names that may coincide with propane keywords (e.g. 'start', 'token',
-" 'tree'). Wrap it in a region that excludes keyword matches so those names
-" are not highlighted as keywords. The '<<' is left unconsumed so the
-" propaneTarget region can still match it.
-syn region propaneRuleRhs matchgroup=propaneOperator start="->" end="\ze<<" end=";" contains=propaneFieldAlias,propaneRuleOperator,propaneComment keepend
-syn match propaneRuleOperator "?" contained
-" Keywords that introduce a user-defined name. The name is consumed by
-" propaneName via nextgroup so a name matching a keyword (e.g. 'token start')
-" is not highlighted as a keyword. These must be a match (not syn keyword)
-" because a syn keyword always wins over a contained nextgroup match.
-syn match propaneNameDecl "\<\%(tokenid\|token\|lex_fn\|module\|start\|tree_prefix\|tree_suffix\)\>" nextgroup=propaneName skipwhite
-syn match propaneName "\<\h\w*\>" contained
-syn match propaneKeyword "\<\%(context_user_fields\|drop\|free_token_node\|noline\|on_token_node\|prefix\|ptype\|token_user_fields\|tree\)\>"
 
 syn region propaneRegex start="/" end="/" skip="\v\\\\|\\/"
 
 hi def link propaneComment Comment
 hi def link propaneKeyword Keyword
-hi def link propaneNameDecl Keyword
 hi def link propaneRegex String
 hi def link propaneOperator Operator
-hi def link propaneRuleOperator Operator
 hi def link propaneFieldOperator Operator
 hi def link propaneDelimiter Delimiter
 hi def link propaneFieldAlias Identifier

lib/propane.rb

@@ -33,7 +33,7 @@ class Propane
 
     def run(input_file, output_file, log_file, options)
       begin
-        grammar = Grammar.new(File.read(input_file), input_file)
+        grammar = Grammar.new(File.read(input_file))
         generator = Generator.new(grammar, output_file, log_file, options)
         generator.generate
       rescue Error => e

lib/propane/generator.rb

@@ -13,15 +13,8 @@ class Propane
       @language =
         if output_file.end_with?(".d")
           "d"
-        elsif output_file.end_with?(".c")
-          "c"
-        elsif output_file =~ %r{\.(cc|cpp|cxx)$}
-          @cpp = true
-          "c"
-        elsif output_file.end_with?(".rs")
-          "rust"
         else
-          raise Error.new("Could not determine target language from output file name (#{output_file})")
+          "c"
         end
       @options = options
       process_grammar!
@@ -33,37 +26,14 @@ class Propane
         extensions += %w[h]
       end
       extensions.each do |extension|
-        template_language = @language == "rust" ? "rs" : @language
+        template = Assets.get("parser.#{extension || @language}.erb")
-        template = Assets.get("parser.#{extension || template_language}.erb")
         if extension
           output_file = @output_file.sub(%r{\.[a-z]+$}, ".#{extension}")
         else
           output_file = @output_file
         end
         erb = ERB.new(template, trim_mode: "<>")
-        # Rust has no #line directive support. For a Rust target the directives
+        result = erb.result(binding.clone)
-        # that the grammar embeds around user code blocks are replaced with
-        # comments naming the grammar file and line number the code came from,
-        # so that the origin of a section of user code can still be found by
-        # reading up from a compiler diagnostic pointing into the generated
-        # module.
-        user_code_origin = nil
-        result = erb.result(binding.clone).lines.each_with_index.map do |line, i|
-          if @language == "rust"
-            if md = line.match(/^#line (\d+) "([^"]*)"/)
-              user_code_origin = "#{md[2]} line #{md[1]}"
-              line.sub(/^#line \d+ "[^"]*"/, %[/* Begin user code from #{user_code_origin}. */])
-            elsif line == "#linereset\n"
-              %[/* End user code from #{user_code_origin}. */\n]
-            else
-              line
-            end
-          elsif line == "#linereset\n"
-            %[#line #{i + 2} "#{output_file}"\n]
-          else
-            line
-          end
-        end.join
         File.open(output_file, "wb") do |fh|
           fh.write(result)
         end
@@ -204,8 +174,7 @@ class Propane
                 end
               end
               @grammar.rules << Rule.new(component, [], nil, ptypename, rule.line_number)
-              optcode = @grammar.tree ? nil : "$$ = $1;\n"
+              @grammar.rules << Rule.new(component, [c], "$$ = $1;\n", ptypename, rule.line_number)
-              @grammar.rules << Rule.new(component, [c], optcode, ptypename, rule.line_number)
               optional_rules_added << component
             end
           end
@@ -294,88 +263,47 @@ class Propane
           "context->user_terminate_code = (#{user_terminate_code}); return #{retval};"
         when "d"
           "context.user_terminate_code = (#{user_terminate_code}); return #{retval};"
-        when "rust"
-          "context.user_terminate_code = (#{user_terminate_code}); return #{retval};"
-        end
-      end
-      code = code.gsub(/\$\{context\.(\w+)\}/) do |match|
-        fieldname = $1
-        case @language
-        when "c"
-          "context->#{fieldname}"
-        when "d"
-          "context.#{fieldname}"
-        when "rust"
-          "context.#{fieldname}"
-        end
-      end
-      code = code.gsub(/\$\{token\.(\w+)\}/) do |match|
-        fieldname = $1
-        case @language
-        when "c"
-          "token_tree_node->#{fieldname}"
-        when "d"
-          "token_tree_node.#{fieldname}"
-        when "rust"
-          "token_tree_node.#{fieldname}"
         end
       end
       if parser
         code = code.gsub(/\$\$/) do |match|
-          if @grammar.tree
-            typename = "#{@grammar.tree_prefix}#{rule.name}#{@grammar.tree_suffix}"
-            case @language
-            when "c"
-              tree_handle(typename, "_node_id")
-            when "d"
-              tree_handle(typename, "_node_id")
-            when "rust"
-              tree_handle(typename, "_node_id")
-            end
-          else
           case @language
           when "c"
             "_pvalue->v_#{rule.ptypename}"
           when "d"
             "_pvalue.v_#{rule.ptypename}"
-            when "rust"
-              "(*_pvalue.v_#{rule.ptypename}_mut())"
-            end
           end
         end
         code = code.gsub(/\$(\d+)/) do |match|
-          parser_component_reference(rule, $1.to_i)
-        end
-        code = code.gsub(/\$\{(\$|\d+)\.position\}/) do |match|
           index = $1.to_i
-          "get_rule_position(statevalues, #{index}, n_states, false)"
+          case @language
+          when "c"
+            "state_values_stack_index(statevalues, -1 - (int)n_states + #{index})->pvalue.v_#{rule.components[index - 1].ptypename}"
+          when "d"
+            "statevalues[$-1-n_states+#{index}].pvalue.v_#{rule.components[index - 1].ptypename}"
           end
-        code = code.gsub(/\$\{(\$|\d+)\.end_position\}/) do |match|
-          index = $1.to_i
-          "get_rule_position(statevalues, #{index}, n_states, true)"
         end
         code = code.gsub(/\$\{(\w+)\}/) do |match|
           aliasname = $1
           if index = rule.aliases[aliasname]
-            # Field aliases are just a named reference to a positional rule
+            case @language
-            # component, so reuse the same expansion as `$1', `$2', etc. Note
+            when "c"
-            # that rule.aliases stores a 0-based component index, so add 1 to
+              "state_values_stack_index(statevalues, -(int)n_states + #{index})->pvalue.v_#{rule.components[index].ptypename}"
-            # convert it to the 1-based index used for positional references.
+            when "d"
-            parser_component_reference(rule, index + 1)
+              "statevalues[$-n_states+#{index}].pvalue.v_#{rule.components[index].ptypename}"
+            end
           else
             raise Error.new("Field alias '#{aliasname}' not found")
           end
         end
       else
         code = code.gsub(/\$\$/) do |match|
-          if @grammar.tree
+          if @grammar.ast
             case @language
             when "c"
               "out_token_info->pvalue"
             when "d"
               "out_token_info.pvalue"
-            when "rust"
-              "out_token_info.pvalue"
             end
           else
             case @language
@@ -383,31 +311,9 @@ class Propane
               "out_token_info->pvalue.v_#{pattern.ptypename}"
             when "d"
               "out_token_info.pvalue.v_#{pattern.ptypename}"
-            when "rust"
-              "(*out_token_info.pvalue.v_#{pattern.ptypename}_mut())"
             end
           end
         end
-        code = code.gsub(/\$\{position\}/) do |match|
-          case @language
-          when "c"
-            "out_token_info->position"
-          when "d"
-            "out_token_info.position"
-          when "rust"
-            "out_token_info.position"
-          end
-        end
-        code = code.gsub(/\$\{end_position\}/) do |match|
-          case @language
-          when "c"
-            "out_token_info->end_position"
-          when "d"
-            "out_token_info.end_position"
-          when "rust"
-            "out_token_info.end_position"
-          end
-        end
         code = code.gsub(/\$mode\(([a-zA-Z_][a-zA-Z_0-9]*)\)/) do |match|
           mode_name = $1
           mode_id = @lexer.mode_id(mode_name)
@@ -419,411 +325,12 @@ class Propane
             "context->mode = #{mode_id}u"
           when "d"
             "context.mode = #{mode_id}u"
-          when "rust"
-            "context.mode = #{mode_id}"
           end
         end
       end
       code
     end
 
-    # Expand a positional reference to a parser rule component.
-    #
-    # This is used to expand `$1', `$2', etc. as well as field aliases (which
-    # are just named references to a positional rule component).
-    #
-    # @param rule [Rule]
-    #   The Rule containing the user code.
-    # @param index [Integer]
-    #   1-based index of the rule component to reference.
-    #
-    # @return [String]
-    #   Expanded rule component reference.
-    def parser_component_reference(rule, index)
-      component = rule.components[index - 1]
-      if @grammar.tree
-        # In tree mode a component reference yields a handle to that
-        # component's tree node. An optional component propagates its target
-        # node (or null), so use the optional target's node type.
-        if component.is_a?(RuleSet) && component.optional?
-          component = component.option_target
-        end
-        node_name = component.is_a?(Token) ? "Token" : component.name
-        typename = "#{@grammar.tree_prefix}#{node_name}#{@grammar.tree_suffix}"
-        case @language
-        when "c"
-          tree_handle(typename, "state_values_stack_index(statevalues, -1 - (int)n_states + #{index})->node_id")
-        when "d"
-          tree_handle(typename, "statevalues[$-1-n_states+#{index}].node_id")
-        when "rust"
-          tree_handle(typename, "statevalues[statevalues.len() - 1 - n_states + #{index}].node_id")
-        end
-      else
-        case @language
-        when "c"
-          "state_values_stack_index(statevalues, -1 - (int)n_states + #{index})->pvalue.v_#{component.ptypename}"
-        when "d"
-          "statevalues[$-1-n_states+#{index}].pvalue.v_#{component.ptypename}"
-        when "rust"
-          "statevalues[statevalues.len() - 1 - n_states + #{index}].pvalue.get_v_#{component.ptypename}()"
-        end
-      end
-    end
-
-    # Construct a tree node handle expression for the target language.
-    #
-    # A handle is a small value pairing the parser context with a node ID
-    # (an index into the context's node arena). All handle types share this
-    # layout; the distinct types exist for documentation and, in C, to drive
-    # the tree walk macro's type threading.
-    #
-    # @param typename [String]
-    #   Handle type name.
-    # @param id_expr [String]
-    #   Expression yielding the node ID.
-    # @param parenthesize [Boolean]
-    #   Whether to parenthesize the expression. Parentheses are required where
-    #   the expression is substituted into a user code block, since the
-    #   expression could be followed there by a field access or appear in a
-    #   position where a bare Rust struct literal is not accepted. They are
-    #   unnecessary where the expression stands alone, and Rust warns about
-    #   them there, so this can be disabled for those uses.
-    #
-    # @return [String]
-    #   Handle constructor expression.
-    def tree_handle(typename, id_expr, parenthesize = true)
-      if @cpp
-        "(#{typename}{context, #{id_expr}})"
-      elsif @language == "c"
-        "((#{typename}){context, #{id_expr}})"
-      elsif @language == "rust"
-        expr = "#{typename} { context, id: #{id_expr} }"
-        parenthesize ? "(#{expr})" : expr
-      else
-        "#{typename}(context, #{id_expr})"
-      end
-    end
-
-    # Get the list of non-optional, non-internal rule sets that get a tree node
-    # handle type generated for them.
-    #
-    # @return [Array<Propane::RuleSet>]
-    #   Rule sets with generated tree node handle types.
-    def tree_node_rule_sets
-      @parser.rule_sets.reject do |name, rule_set|
-        name.start_with?("$") || rule_set.optional?
-      end.map {|name, rule_set| rule_set}
-    end
-
-    # Maximum number of chained fields supported by a single C tree walk macro
-    # invocation. Deeper navigation can be expressed by nesting walk calls.
-    C_TREE_WALK_MAX = 16
-
-    # Get the tree node handle type name for a node name.
-    #
-    # @param name [String]
-    #   Rule set name, or "Token".
-    #
-    # @return [String]
-    #   Handle type name.
-    def h_type(name)
-      "#{@grammar.tree_prefix}#{name}#{@grammar.tree_suffix}"
-    end
-
-    # Get the list of all tree node handle type names (Token plus rule sets).
-    #
-    # @return [Array<String>]
-    #   Handle type names.
-    def tree_handle_types
-      [h_type("Token")] + tree_node_rule_sets.map {|rs| h_type(rs.name)}
-    end
-
-    # Enumerate the navigation fields of a rule set's tree node.
-    #
-    # @yield [rtype, field_name, child_type, slot]
-    #   Handle type name, field accessor name, child handle type, and child
-    #   slot index.
-    def each_tree_field(rule_set)
-      rtype = h_type(rule_set.name)
-      rule_set.tree_fields.each_with_index do |fields, slot|
-        fields.each do |field_name, child_type|
-          yield rtype, field_name, child_type, slot
-        end
-      end
-    end
-
-    # Generate the tree node handle type declarations for the header.
-    #
-    # These are emitted before the context structure definition so that a
-    # context_user_fields block can declare a field of a handle type.
-    #
-    # @return [String]
-    #   Handle type declarations.
-    def c_tree_handle_types_header
-      @cpp ? cpp_tree_handle_types_header : c_only_tree_handle_types_header
-    end
-
-    # Generate the remainder of the tree node section for the header.
-    #
-    # This is emitted after the context structure definition since it
-    # dereferences the context and so requires the complete type.
-    #
-    # @return [String]
-    #   Accessors, macros, and out-of-line handle method definitions.
-    def c_tree_types_header
-      @cpp ? cpp_tree_types_header : c_only_tree_types_header
-    end
-
-    # Generate the C (non-C++) tree node handle type section for the header.
-    def c_only_tree_handle_types_header
-      p = @grammar.prefix
-      out = []
-      out << "/** Tree node handle types. @{ */"
-      tree_handle_types.each do |t|
-        out << "typedef struct { #{p}context_t * __context; #{p}node_id_t __id; } #{t};"
-      end
-      out << ""
-      out.join("\n")
-    end
-
-    def c_only_tree_types_header
-      out = []
-      out << c_common_accessors_header
-      out << "/** @} */"
-      out.join("\n")
-    end
-
-    # Generate the C-style (function + macro) tree node accessors shared by the
-    # C and C++ headers. In C++ these are provided in addition to the handle
-    # methods so that C-style code (and the tree walk macros) also works.
-    def c_common_accessors_header
-      p = @grammar.prefix
-      out = []
-      out << "/** Generic tree node accessors (usable on any handle type). */"
-      out << "#define #{p}node_valid(h) ((h).__id != 0u)"
-      out << "#define #{p}node_id(h) ((h).__id)"
-      out << "#define #{p}node_data(h) (&(h).__context->#{p}tree_nodes[(h).__id])"
-      out << "#define #{p}node_position(h) ((h).__context->#{p}tree_nodes[(h).__id].position)"
-      out << "#define #{p}node_end_position(h) ((h).__context->#{p}tree_nodes[(h).__id].end_position)"
-      out << "#define #{p}node_n_fields(h) ((h).__id ? (h).__context->#{p}tree_nodes[(h).__id].n_fields : (uint16_t)0u)"
-      out << ""
-      out << "/** Tree node field accessor functions. */"
-      out << "#{p}token_t #{p}#{h_type("Token")}_token(#{h_type("Token")} node);"
-      out << "#{p}value_t #{p}#{h_type("Token")}_pvalue(#{h_type("Token")} node);"
-      tree_node_rule_sets.each do |rule_set|
-        each_tree_field(rule_set) do |rtype, field_name, child_type, slot|
-          out << "#{child_type} #{p}#{rtype}_#{field_name}(#{rtype} node);"
-        end
-      end
-      out << ""
-      out << c_tree_walk_macros
-      out.join("\n")
-    end
-
-    # Generate the C tree walk macro machinery.
-    def c_tree_walk_macros
-      p = @grammar.prefix
-      max = C_TREE_WALK_MAX
-      out = []
-      out << "/* Tree walk macros: p_tree_walk_<Type>(handle, field, ...). */"
-      out << "#define #{p}CAT_(a, b) a##b"
-      out << "#define #{p}CAT(a, b) #{p}CAT_(a, b)"
-      out << "#define #{p}TA(t, f) #{p}CAT(#{p}CAT(#{p}CAT(#{p}TYPEAFTER_, t), _), f)"
-      out << "#define #{p}ACC(t, f) #{p}CAT(#{p}CAT(#{p}CAT(#{p}, t), _), f)"
-      argn = (1..max).map {|i| "_#{i}"}.join(", ")
-      rseq = (0..max).to_a.reverse.join(", ")
-      out << "#define #{p}ARG_N(#{argn}, N, ...) N"
-      out << "#define #{p}NARG(...) #{p}ARG_N(__VA_ARGS__, #{rseq})"
-      (1..max).each do |n|
-        fparams = (1..n).map {|k| "f#{k}"}.join(", ")
-        call = "h"
-        (1..n).each do |k|
-          texpr = "R"
-          (1...k).each {|j| texpr = "#{p}TA(#{texpr}, f#{j})"}
-          call = "#{p}ACC(#{texpr}, f#{k})(#{call})"
-        end
-        out << "#define #{p}tree_walk_#{n}(R, h, #{fparams}) #{call}"
-      end
-      out << "#define #{p}tree_walk_dispatch(R, h, ...) #{p}CAT(#{p}tree_walk_, #{p}NARG(__VA_ARGS__))(R, h, __VA_ARGS__)"
-      # Type transition map (navigation fields only).
-      tree_node_rule_sets.each do |rule_set|
-        each_tree_field(rule_set) do |rtype, field_name, child_type, slot|
-          out << "#define #{p}TYPEAFTER_#{rtype}_#{field_name} #{child_type}"
-        end
-      end
-      # Per-handle-type walk entry points.
-      tree_handle_types.each do |t|
-        out << "#define #{p}tree_walk_#{t}(...) #{p}tree_walk_dispatch(#{t}, __VA_ARGS__)"
-      end
-      out.join("\n")
-    end
-
-    # Generate the C tree node accessor function definitions for the source.
-    #
-    # @return [String]
-    #   Accessor function definitions.
-    def c_tree_accessor_defs
-      p = @grammar.prefix
-      tt = h_type("Token")
-      out = []
-      out << "#{p}token_t #{p}#{tt}_token(#{tt} node)"
-      out << "{"
-      out << "    return node.__context->#{p}tree_nodes[node.__id].token;"
-      out << "}"
-      out << ""
-      out << "#{p}value_t #{p}#{tt}_pvalue(#{tt} node)"
-      out << "{"
-      out << "    return node.__context->#{p}tree_nodes[node.__id].pvalue;"
-      out << "}"
-      tree_node_rule_sets.each do |rule_set|
-        each_tree_field(rule_set) do |rtype, field_name, child_type, slot|
-          out << ""
-          out << "#{child_type} #{p}#{rtype}_#{field_name}(#{rtype} node)"
-          out << "{"
-          out << "    #{child_type} result;"
-          out << "    result.__context = node.__context;"
-          out << "    if (node.__id == 0u)"
-          out << "    {"
-          out << "        result.__id = 0u;"
-          out << "        return result;"
-          out << "    }"
-          out << "    result.__id = node.__context->#{p}tree_children[node.__context->#{p}tree_nodes[node.__id].child_offset + #{slot}u];"
-          out << "    return result;"
-          out << "}"
-        end
-      end
-      out.join("\n")
-    end
-
-    # Generate the C++ tree node handle class declarations for the header.
-    # Only valid() and node_id() are defined inline; every other method
-    # dereferences the context, which is still an incomplete type here, so
-    # those are declared and defined out of line once the context is
-    # complete.
-    def cpp_tree_handle_types_header
-      p = @grammar.prefix
-      out = []
-      out << "/** Tree node handle types. @{ */"
-      tree_handle_types.each {|t| out << "struct #{t};"}
-      out << ""
-      tt = h_type("Token")
-      out << "struct #{tt}"
-      out << "{"
-      out << "    #{p}context_t * __context;"
-      out << "    #{p}node_id_t __id;"
-      out << "    bool valid() const { return __id != 0u; }"
-      out << "    #{p}node_id_t node_id() const { return __id; }"
-      out << "    #{p}node_data_t * data() const;"
-      out << "    #{p}position_t position() const;"
-      out << "    #{p}position_t end_position() const;"
-      out << "    uint16_t n_fields() const;"
-      out << "    #{p}token_t token() const;"
-      out << "    #{p}value_t pvalue() const;"
-      out << "};"
-      out << ""
-      tree_node_rule_sets.each do |rule_set|
-        rtype = h_type(rule_set.name)
-        out << "struct #{rtype}"
-        out << "{"
-        out << "    #{p}context_t * __context;"
-        out << "    #{p}node_id_t __id;"
-        out << "    bool valid() const { return __id != 0u; }"
-        out << "    #{p}node_id_t node_id() const { return __id; }"
-        out << "    #{p}node_data_t * data() const;"
-        out << "    #{p}position_t position() const;"
-        out << "    #{p}position_t end_position() const;"
-        out << "    uint16_t n_fields() const;"
-        each_tree_field(rule_set) do |rt, field_name, child_type, slot|
-          out << "    #{child_type} #{field_name}() const;"
-        end
-        out << "};"
-        out << ""
-      end
-      out.join("\n")
-    end
-
-    # Generate the out-of-line C++ handle method definitions plus the C-style
-    # accessors. Emitted after the context structure definition.
-    def cpp_tree_types_header
-      p = @grammar.prefix
-      out = []
-      # Common node methods, now that the context type is complete.
-      ([h_type("Token")] + tree_node_rule_sets.map {|rs| h_type(rs.name)}).each do |ht|
-        out << "inline #{p}node_data_t * #{ht}::data() const { return &__context->#{p}tree_nodes[__id]; }"
-        out << "inline #{p}position_t #{ht}::position() const { return __context->#{p}tree_nodes[__id].position; }"
-        out << "inline #{p}position_t #{ht}::end_position() const { return __context->#{p}tree_nodes[__id].end_position; }"
-        out << "inline uint16_t #{ht}::n_fields() const { return __id ? __context->#{p}tree_nodes[__id].n_fields : (uint16_t)0u; }"
-      end
-      tt = h_type("Token")
-      out << "inline #{p}token_t #{tt}::token() const { return __context->#{p}tree_nodes[__id].token; }"
-      out << "inline #{p}value_t #{tt}::pvalue() const { return __context->#{p}tree_nodes[__id].pvalue; }"
-      out << ""
-      # Out-of-line navigation method bodies (all handle types now complete).
-      tree_node_rule_sets.each do |rule_set|
-        rtype = h_type(rule_set.name)
-        each_tree_field(rule_set) do |rt, field_name, child_type, slot|
-          out << "inline #{child_type} #{rtype}::#{field_name}() const"
-          out << "{"
-          out << "    if (__id == 0u)"
-          out << "    {"
-          out << "        return #{child_type}{__context, 0u};"
-          out << "    }"
-          out << "    return #{child_type}{__context, __context->#{p}tree_children[__context->#{p}tree_nodes[__id].child_offset + #{slot}u]};"
-          out << "}"
-        end
-      end
-      out << ""
-      out << "/*"
-      out << " * C-style function and macro accessors, provided in addition to the handle"
-      out << " * methods above so that C-style code and the tree walk macros also work."
-      out << " */"
-      out << c_common_accessors_header
-      out << "/** @} */"
-      out.join("\n")
-    end
-
-    # Rust keywords that must be escaped as raw identifiers when used as a
-    # generated identifier (e.g. a field alias named `type`).
-    RUST_KEYWORDS = %w[
-      as break const continue dyn else enum extern false fn for if impl in let
-      loop match mod move mut pub ref return static struct trait true type
-      unsafe use where while async await abstract become box do final macro
-      override priv typeof unsized virtual yield try gen
-    ]
-
-    # Escape a name as a Rust raw identifier if it is a reserved keyword.
-    #
-    # @param name [String]
-    #   Identifier name.
-    #
-    # @return [String]
-    #   Name, escaped as a raw identifier if necessary.
-    def rust_ident(name)
-      RUST_KEYWORDS.include?(name) ? "r##{name}" : name
-    end
-
-    # Map a ptype type string to a valid Rust type.
-    #
-    # The default ptype is a C "void *"; for Rust with no declared ptype we use
-    # the unit type instead.
-    #
-    # @param typestring [String]
-    #   ptype type string.
-    #
-    # @return [String]
-    #   Rust type string.
-    def rust_ptype(typestring)
-      typestring == "void *" ? "()" : typestring
-    end
-
-    # Get the lex function to use.
-    #
-    # @return [String]
-    #   Lex function to use.
-    def lex_fn
-      @grammar.lex_fn || "#{@grammar.prefix}lex"
-    end
-
     # Get the parser value type for the start rule.
     #
     # @return [Array<String>]
@@ -849,8 +356,6 @@ class Propane
           "uint8_t"
         when "d"
           "ubyte"
-        when "rust"
-          "u8"
         end
       elsif max <= 0xFFFF
         case @language
@@ -858,15 +363,11 @@ class Propane
           "uint16_t"
         when "d"
           "ushort"
-        when "rust"
-          "u16"
         end
       else
         case @language
         when "c"
           "uint32_t"
-        when "rust"
-          "u32"
         else
           "uint"
         end

lib/propane/grammar.rb

@@ -5,11 +5,9 @@ class Propane
     # Reserve identifiers beginning with a double-underscore for internal use.
     IDENTIFIER_REGEX = /(?:[a-zA-Z]|_[a-zA-Z0-9])[a-zA-Z_0-9]*/
 
-    attr_reader :context_user_fields
+    attr_reader :ast
-    attr_reader :lex_fn
+    attr_reader :ast_prefix
-    attr_reader :tree
+    attr_reader :ast_suffix
-    attr_reader :tree_prefix
-    attr_reader :tree_suffix
     attr_reader :free_token_node
     attr_reader :modulename
     attr_reader :patterns
@@ -19,11 +17,8 @@ class Propane
     attr_reader :code_blocks
     attr_reader :ptypes
     attr_reader :prefix
-    attr_reader :on_token_node
-    attr_reader :token_user_fields
 
-    def initialize(input, filename)
+    def initialize(input)
-      @filename = filename
       @patterns = []
       @start_rules = []
       @tokens = []
@@ -35,13 +30,10 @@ class Propane
       @input = input.gsub("\r\n", "\n")
       @ptypes = {"default" => "void *"}
       @prefix = "p_"
-      @tree = false
+      @ast = false
-      @tree_prefix = ""
+      @ast_prefix = ""
-      @tree_suffix = ""
+      @ast_suffix = ""
-      @free_token_node = ""
+      @free_token_node = nil
-      @context_user_fields = nil
-      @on_token_node = ""
-      @token_user_fields = nil
       parse_grammar!
       @start_rules << "Start" if @start_rules.empty?
     end
@@ -58,10 +50,6 @@ class Propane
       @tokens.size + 1
     end
 
-    def parser_user_code_used?
-      @rules.any? {|r| r.code}
-    end
-
     private
 
     def parse_grammar!
@@ -74,15 +62,11 @@ class Propane
       if parse_white_space!
       elsif parse_comment_line!
       elsif @modeline.nil? && parse_mode_label!
-      elsif parse_context_user_fields_statement!
+      elsif parse_ast_statement!
-      elsif parse_lex_fn!
+      elsif parse_ast_prefix_statement!
-      elsif parse_tree_statement!
+      elsif parse_ast_suffix_statement!
-      elsif parse_tree_prefix_statement!
-      elsif parse_tree_suffix_statement!
       elsif parse_free_token_node_statement!
       elsif parse_module_statement!
-      elsif parse_on_token_node_statement!
-      elsif parse_token_user_fields_statement!
       elsif parse_ptype_statement!
       elsif parse_pattern_statement!
       elsif parse_start_statement!
@@ -92,7 +76,6 @@ class Propane
       elsif parse_rule_statement!
       elsif parse_code_block_statement!
       elsif parse_prefix_statement!
-      elsif parse_noline_statement!
       else
         if @input.size > 25
           @input = @input.slice(0..20) + "..."
@@ -115,37 +98,27 @@ class Propane
       consume!(/#.*\n/)
     end
 
-    def parse_context_user_fields_statement!
+    def parse_ast_statement!
-      if md = consume!(/context_user_fields\b\s*/)
+      if consume!(/ast\s*;/)
-        unless code = parse_code_block!
+        @ast = true
-          raise Error.new("Line #{@line_number}: expected code block")
-        end
-        @context_user_fields ||= ""
-        @context_user_fields += code
       end
     end
 
-    def parse_lex_fn!
+    def parse_ast_prefix_statement!
-      if md = consume!(/lex_fn\b\s*(\w+)\s*;/)
+      if md = consume!(/ast_prefix\s+(\w+)\s*;/)
-        @lex_fn = md[1]
+        @ast_prefix = md[1]
       end
     end
 
-    def parse_tree_statement!
+    def parse_ast_suffix_statement!
-      if consume!(/tree\s*;/)
+      if md = consume!(/ast_suffix\s+(\w+)\s*;/)
-        @tree = true
+        @ast_suffix = md[1]
       end
     end
 
-    def parse_tree_prefix_statement!
+    def parse_free_token_node_statement!
-      if md = consume!(/tree_prefix\s+(\w+)\s*;/)
+      if md = consume!(/free_token_node\s+(\w+)\s*;/)
-        @tree_prefix = md[1]
+        @free_token_node = md[1]
-      end
-    end
-
-    def parse_tree_suffix_statement!
-      if md = consume!(/tree_suffix\s+(\w+)\s*;/)
-        @tree_suffix = md[1]
       end
     end
 
@@ -159,40 +132,12 @@ class Propane
       end
     end
 
-    def parse_on_token_node_statement!
-      if md = consume!(/on_token_node\b\s*/)
-        unless code = parse_code_block!
-          raise Error.new("Line #{@line_number}: expected code block")
-        end
-        @on_token_node += code
-      end
-    end
-
-    def parse_token_user_fields_statement!
-      if md = consume!(/token_user_fields\b\s*/)
-        unless code = parse_code_block!
-          raise Error.new("Line #{@line_number}: expected code block")
-        end
-        @token_user_fields ||= ""
-        @token_user_fields += code
-      end
-    end
-
-    def parse_free_token_node_statement!
-      if md = consume!(/free_token_node\b\s*/)
-        unless code = parse_code_block!
-          raise Error.new("Line #{@line_number}: expected code block")
-        end
-        @free_token_node += code
-      end
-    end
-
     def parse_ptype_statement!
       if consume!(/ptype\s+/)
         name = "default"
         if md = consume!(/(#{IDENTIFIER_REGEX})\s*=\s*/)
-          if @tree
+          if @ast
-            raise Error.new("Multiple ptypes are unsupported in tree mode")
+            raise Error.new("Multiple ptypes are unsupported in AST mode")
           end
           name = md[1]
         end
@@ -206,8 +151,8 @@ class Propane
         md = consume!(/(#{IDENTIFIER_REGEX})\s*/, "expected token name")
         name = md[1]
         if md = consume!(/\((#{IDENTIFIER_REGEX})\)\s*/)
-          if @tree
+          if @ast
-            raise Error.new("Multiple ptypes are unsupported in tree mode")
+            raise Error.new("Multiple ptypes are unsupported in AST mode")
           end
           ptypename = md[1]
         end
@@ -230,8 +175,8 @@ class Propane
         md = consume!(/(#{IDENTIFIER_REGEX})\s*/, "expected token name")
         name = md[1]
         if md = consume!(/\((#{IDENTIFIER_REGEX})\)\s*/)
-          if @tree
+          if @ast
-            raise Error.new("Multiple ptypes are unsupported in tree mode")
+            raise Error.new("Multiple ptypes are unsupported in AST mode")
           end
           ptypename = md[1]
         end
@@ -250,10 +195,8 @@ class Propane
           raise Error.new("Line #{@line_number}: expected pattern to follow `drop'")
         end
         consume!(/\s+/)
-        unless code = parse_code_block!
+        consume!(/;/, "expected `;'")
-          consume!(/;/, "expected `;' or code block")
+        @patterns << Pattern.new(pattern: pattern, line_number: @line_number, modes: get_modes_from_modeline)
-        end
-        @patterns << Pattern.new(pattern: pattern, line_number: @line_number, code: code, modes: get_modes_from_modeline)
         @modeline = nil
         true
       end
@@ -262,14 +205,18 @@ class Propane
     def parse_rule_statement!
       if md = consume!(/(#{IDENTIFIER_REGEX})\s*(?:\((#{IDENTIFIER_REGEX})\))?\s*->\s*/)
         rule_name, ptypename = *md[1, 2]
-        if @tree && ptypename
+        if @ast && ptypename
-          raise Error.new("Multiple ptypes are unsupported in tree mode")
+          raise Error.new("Multiple ptypes are unsupported in AST mode")
         end
         md = consume!(/((?:#{IDENTIFIER_REGEX}\??(?::#{IDENTIFIER_REGEX})?\s*)*)\s*/, "expected rule component list")
         components = md[1].strip.split(/\s+/)
+        if @ast
+          consume!(/;/, "expected `;'")
+        else
           unless code = parse_code_block!
             consume!(/;/, "expected `;' or code block")
           end
+        end
         @rules << Rule.new(rule_name, components, code, ptypename, @line_number)
         @modeline = nil
         true
@@ -280,8 +227,8 @@ class Propane
       if pattern = parse_pattern!
         consume!(/\s+/)
         if md = consume!(/\((#{IDENTIFIER_REGEX})\)\s*/)
-          if @tree
+          if @ast
-            raise Error.new("Multiple ptypes are unsupported in tree mode")
+            raise Error.new("Multiple ptypes are unsupported in AST mode")
           end
           ptypename = md[1]
         end
@@ -306,14 +253,8 @@ class Propane
     def parse_code_block_statement!
       if md = consume!(/<<([a-z]*)(.*?)>>\n/m)
         name, code = md[1..2]
-        code = code.chomp
+        code.sub!(/\A\n/, "")
-        unless @noline
+        code += "\n" unless code.end_with?("\n")
-          if code.start_with?("\n")
-            code = %[#line #{@line_number + 1} "#{@filename}"#{code}\n#linereset\n]
-          else
-            code = %[#line #{@line_number} "#{@filename}"\n#{code}\n#linereset\n]
-          end
-        end
         if @code_blocks[name]
           @code_blocks[name] += code
         else
@@ -331,13 +272,6 @@ class Propane
       end
     end
 
-    def parse_noline_statement!
-      if md = consume!(/noline\s*;/)
-        @noline = true
-        true
-      end
-    end
-
     def parse_pattern!
       if md = consume!(%r{/})
         pattern = ""
@@ -351,8 +285,6 @@ class Propane
             end
           elsif md = consume!(%r{(.)})
             pattern += md[1]
-          elsif @input == "" || @input.start_with?("\n")
-            raise Error.new("Line #{@line_number}: Unterminated pattern; expected `/`")
           end
         end
         pattern
@@ -361,14 +293,9 @@ class Propane
 
     def parse_code_block!
       if md = consume!(/<<(.*?)>>\n/m)
-        code = md[1].chomp
+        code = md[1]
-        unless @noline
+        code.sub!(/\A\n/, "")
-          if code.start_with?("\n")
+        code += "\n" unless code.end_with?("\n")
-            code = %[#line #{@line_number + 1} "#{@filename}"#{code}\n#linereset\n]
-          else
-            code = %[#line #{@line_number} "#{@filename}"\n#{code}\n#linereset\n]
-          end
-        end
         code
       end
     end

lib/propane/rule.rb

@@ -36,7 +36,7 @@ class Propane
 
     # @return [Array<Integer>]
     #   Map this rule's components to their positions in the parent RuleSet's
-    #   node field pointer array. This is used for tree construction.
+    #   node field pointer array. This is used for AST construction.
     attr_accessor :rule_set_node_field_index_map
 
     # Construct a Rule.

lib/propane/rule_set.rb

@@ -4,8 +4,8 @@ class Propane
   class RuleSet
 
     # @return [Array<Hash>]
-    #   tree fields.
+    #   AST fields.
-    attr_reader :tree_fields
+    attr_reader :ast_fields
 
     # @return [Integer]
     #   ID of the RuleSet.
@@ -100,28 +100,28 @@ class Propane
 
     # Finalize a RuleSet after adding all Rules to it.
     def finalize(grammar)
-      if grammar.tree
+      if grammar.ast
-        build_tree_fields(grammar)
+        build_ast_fields(grammar)
       end
     end
 
     private
 
-    # Build the set of tree fields for this RuleSet.
+    # Build the set of AST fields for this RuleSet.
     #
     # This is an Array of Hashes. Each entry in the Array corresponds to a
-    # field location in the tree node. The entry is a Hash. It could have one or
+    # field location in the AST node. The entry is a Hash. It could have one or
     # two keys. It will always have the field name with a positional suffix as
     # a key. It may also have the field name without the positional suffix if
     # that field only exists in one position across all Rules in the RuleSet.
     #
     # @return [void]
-    def build_tree_fields(grammar)
+    def build_ast_fields(grammar)
-      field_tree_node_indexes = {}
+      field_ast_node_indexes = {}
       field_indexes_across_all_rules = {}
-      # Stores the index into @tree_fields by field alias name.
+      # Stores the index into @ast_fields by field alias name.
       field_aliases = {}
-      @tree_fields = []
+      @ast_fields = []
       @rules.each do |rule|
         rule.components.each_with_index do |component, i|
           if component.is_a?(RuleSet) && component.optional?
@@ -132,25 +132,25 @@ class Propane
           else
             node_name = component.name
           end
-          struct_name = "#{grammar.tree_prefix}#{node_name}#{grammar.tree_suffix}"
+          struct_name = "#{grammar.ast_prefix}#{node_name}#{grammar.ast_suffix}"
           field_name = "p#{node_name}#{i + 1}"
-          unless field_tree_node_indexes[field_name]
+          unless field_ast_node_indexes[field_name]
-            field_tree_node_indexes[field_name] = @tree_fields.size
+            field_ast_node_indexes[field_name] = @ast_fields.size
-            @tree_fields << {field_name => struct_name}
+            @ast_fields << {field_name => struct_name}
           end
           rule.aliases.each do |alias_name, index|
             if index == i
-              alias_tree_fields_index = field_tree_node_indexes[field_name]
+              alias_ast_fields_index = field_ast_node_indexes[field_name]
-              if field_aliases[alias_name] && field_aliases[alias_name] != alias_tree_fields_index
+              if field_aliases[alias_name] && field_aliases[alias_name] != alias_ast_fields_index
-                raise Error.new("Error: conflicting tree node field positions for alias `#{alias_name}` in rule #{rule.name} defined on line #{rule.line_number}")
+                raise Error.new("Error: conflicting AST node field positions for alias `#{alias_name}` in rule #{rule.name} defined on line #{rule.line_number}")
               end
-              field_aliases[alias_name] = alias_tree_fields_index
+              field_aliases[alias_name] = alias_ast_fields_index
-              @tree_fields[alias_tree_fields_index][alias_name] = @tree_fields[alias_tree_fields_index].first[1]
+              @ast_fields[alias_ast_fields_index][alias_name] = @ast_fields[alias_ast_fields_index].first[1]
             end
           end
           field_indexes_across_all_rules[node_name] ||= Set.new
-          field_indexes_across_all_rules[node_name] << field_tree_node_indexes[field_name]
+          field_indexes_across_all_rules[node_name] << field_ast_node_indexes[field_name]
-          rule.rule_set_node_field_index_map[i] = field_tree_node_indexes[field_name]
+          rule.rule_set_node_field_index_map[i] = field_ast_node_indexes[field_name]
         end
       end
       field_indexes_across_all_rules.each do |node_name, indexes_across_all_rules|
@@ -158,8 +158,8 @@ class Propane
           # If this field was only seen in one position across all rules,
           # then add an alias to the positional field name that does not
           # include the position.
-          @tree_fields[indexes_across_all_rules.first]["p#{node_name}"] =
+          @ast_fields[indexes_across_all_rules.first]["p#{node_name}"] =
-            "#{grammar.tree_prefix}#{node_name}#{grammar.tree_suffix}"
+            "#{grammar.ast_prefix}#{node_name}#{grammar.ast_suffix}"
         end
       end
     end

lib/propane/version.rb

@@ -1,3 +1,3 @@
 class Propane
-  VERSION = "5.1.0"
+  VERSION = "2.3.0"
 end

spec/ast_node_memory_remains.c.propane

@@ -1,5 +1,5 @@
-tree;
+ast;
-tree_prefix P;
+ast_prefix P;
 
 <<header
 #include <stdio.h>
@@ -46,7 +46,7 @@ token hex_int_l /0[xX][0-9a-fA-F][0-9a-fA-F_]*/ <<
 
 # Identifier.
 token ident /\$?[a-zA-Z_][a-zA-Z_0-9]*\??/ <<
-  $$.s = match_text;
+  $$.s = match;
   $mode(default);
   return $token(ident);
 >>

spec/ast_node_memory_remains.d.propane

@@ -1,5 +1,5 @@
-tree;
+ast;
-tree_prefix P;
+ast_prefix P;
 
 <<
 import std.bigint;
@@ -42,8 +42,8 @@ token semicolon /;/;
 
 # Integer literals.
 token hex_int_l /0[xX][0-9a-fA-F][0-9a-fA-F_]*/ <<
-  $$.bi = BigInt(match_text[0..3]);
+  $$.bi = BigInt(match[0..3]);
-  foreach (c; match_text[3..$])
+  foreach (c; match[3..$])
   {
     if (('0' <= c) && (c <= '9'))
     {
@@ -65,13 +65,13 @@ token hex_int_l /0[xX][0-9a-fA-F][0-9a-fA-F_]*/ <<
 
 # Identifier.
 token ident /\$?[a-zA-Z_][a-zA-Z_0-9]*\??/ <<
-  if (match_text[0] == '$')
+  if (match[0] == '$')
   {
-    $$.s = match_text[1..$];
+    $$.s = match[1..$];
   }
   else
   {
-    $$.s = match_text;
+    $$.s = match;
   }
   $mode(default);
   return $token(ident);

spec/json_parser.c.propane

@@ -21,46 +21,46 @@ token number /-?(0|[1-9][0-9]*)(\.[0-9]+)?([eE][-+]?[0-9]+)?/ <<
     double n = 0.0;
     bool negative = false;
     size_t i = 0u;
-    if (match_text[i] == '-')
+    if (match[i] == '-')
     {
         negative = true;
         i++;
     }
-    while ('0' <= match_text[i] && match_text[i] <= '9')
+    while ('0' <= match[i] && match[i] <= '9')
     {
         n *= 10.0;
-        n += (match_text[i] - '0');
+        n += (match[i] - '0');
         i++;
     }
-    if (match_text[i] == '.')
+    if (match[i] == '.')
     {
         i++;
         double mult = 0.1;
-        while ('0' <= match_text[i] && match_text[i] <= '9')
+        while ('0' <= match[i] && match[i] <= '9')
         {
-            n += mult * (match_text[i] - '0');
+            n += mult * (match[i] - '0');
             mult /= 10.0;
             i++;
         }
     }
-    if (match_text[i] == 'e' || match_text[i] == 'E')
+    if (match[i] == 'e' || match[i] == 'E')
     {
         bool exp_negative = false;
         i++;
-        if (match_text[i] == '-')
+        if (match[i] == '-')
         {
             exp_negative = true;
             i++;
         }
-        else if (match_text[i] == '+')
+        else if (match[i] == '+')
         {
             i++;
         }
         long exp = 0.0;
-        while ('0' <= match_text[i] && match_text[i] <= '9')
+        while ('0' <= match[i] && match[i] <= '9')
         {
             exp *= 10;
-            exp += (match_text[i] - '0');
+            exp += (match[i] - '0');
             i++;
         }
         if (exp_negative)
@@ -120,11 +120,11 @@ string: /\\t/ <<
 >>
 string: /\\u[0-9a-fA-F]{4}/ <<
   /* Not actually going to encode the code point for this example... */
-  char s[] = {'{', (char)match_text[2], (char)match_text[3], (char)match_text[4], (char)match_text[5], '}', 0};
+  char s[] = {'{', (char)match[2], (char)match[3], (char)match[4], (char)match[5], '}', 0};
   str_append(&string_value, s);
 >>
 string: /[^\\]/ <<
-  char s[] = {(char)match_text[0], 0};
+  char s[] = {(char)match[0], 0};
   str_append(&string_value, s);
 >>
 Start -> Value <<

spec/json_parser.d.propane

@@ -20,46 +20,46 @@ token number /-?(0|[1-9][0-9]*)(\.[0-9]+)?([eE][-+]?[0-9]+)?/ <<
     double n;
     bool negative;
     size_t i = 0u;
-    if (match_text[i] == '-')
+    if (match[i] == '-')
     {
         negative = true;
         i++;
     }
-    while ('0' <= match_text[i] && match_text[i] <= '9')
+    while ('0' <= match[i] && match[i] <= '9')
     {
         n *= 10.0;
-        n += (match_text[i] - '0');
+        n += (match[i] - '0');
         i++;
     }
-    if (match_text[i] == '.')
+    if (match[i] == '.')
     {
         i++;
         double mult = 0.1;
-        while ('0' <= match_text[i] && match_text[i] <= '9')
+        while ('0' <= match[i] && match[i] <= '9')
         {
-            n += mult * (match_text[i] - '0');
+            n += mult * (match[i] - '0');
             mult /= 10.0;
             i++;
         }
     }
-    if (match_text[i] == 'e' || match_text[i] == 'E')
+    if (match[i] == 'e' || match[i] == 'E')
     {
         bool exp_negative;
         i++;
-        if (match_text[i] == '-')
+        if (match[i] == '-')
         {
             exp_negative = true;
             i++;
         }
-        else if (match_text[i] == '+')
+        else if (match[i] == '+')
         {
             i++;
         }
         long exp;
-        while ('0' <= match_text[i] && match_text[i] <= '9')
+        while ('0' <= match[i] && match[i] <= '9')
         {
             exp *= 10;
-            exp += (match_text[i] - '0');
+            exp += (match[i] - '0');
             i++;
         }
         if (exp_negative)
@@ -117,10 +117,10 @@ string: /\\t/ <<
 >>
 string: /\\u[0-9a-fA-F]{4}/ <<
   /* Not actually going to encode the code point for this example... */
-  string_value ~= "{" ~ match_text[2..6] ~ "}";
+  string_value ~= "{" ~ match[2..6] ~ "}";
 >>
 string: /[^\\]/ <<
-  string_value ~= match_text;
+  string_value ~= match;
 >>
 Start -> Value <<
   $$ = $1;

spec/json_parser.rust.propane

@@ -1,176 +0,0 @@
-<<
-pub const JSON_OBJECT: usize = 0;
-pub const JSON_ARRAY: usize = 1;
-pub const JSON_NUMBER: usize = 2;
-pub const JSON_STRING: usize = 3;
-pub const JSON_TRUE: usize = 4;
-pub const JSON_FALSE: usize = 5;
-pub const JSON_NULL: usize = 6;
-
-#[derive(Clone, Default)]
-pub enum JSONValue {
-    #[default]
-    Null,
-    Object(Vec<(String, JSONValue)>),
-    Array(Vec<JSONValue>),
-    Number(f64),
-    StringVal(String),
-    True,
-    False,
-}
-
-impl JSONValue {
-    pub fn id(&self) -> usize {
-        match self {
-            JSONValue::Object(_) => JSON_OBJECT,
-            JSONValue::Array(_) => JSON_ARRAY,
-            JSONValue::Number(_) => JSON_NUMBER,
-            JSONValue::StringVal(_) => JSON_STRING,
-            JSONValue::True => JSON_TRUE,
-            JSONValue::False => JSON_FALSE,
-            JSONValue::Null => JSON_NULL,
-        }
-    }
-    pub fn number(&self) -> f64 {
-        if let JSONValue::Number(n) = self { *n } else { 0.0 }
-    }
-    pub fn string(&self) -> &str {
-        if let JSONValue::StringVal(s) = self { s.as_str() } else { "" }
-    }
-    pub fn object_len(&self) -> usize {
-        if let JSONValue::Object(e) = self { e.len() } else { 0 }
-    }
-    pub fn array_len(&self) -> usize {
-        if let JSONValue::Array(e) = self { e.len() } else { 0 }
-    }
-}
->>
-
-context_user_fields <<
-    pub string_value: String,
->>
-
-ptype JSONValue;
-
-drop /\s+/;
-token lbrace /\{/;
-token rbrace /\}/;
-token lbracket /\[/;
-token rbracket /\]/;
-token comma /,/;
-token colon /:/;
-token number /-?(0|[1-9][0-9]*)(\.[0-9]+)?([eE][-+]?[0-9]+)?/ <<
-    let n: f64 = std::str::from_utf8(match_text).unwrap().parse().unwrap();
-    $$ = JSONValue::Number(n);
->>
-token true <<
-  $$ = JSONValue::True;
->>
-token false <<
-  $$ = JSONValue::False;
->>
-token null <<
-  $$ = JSONValue::Null;
->>
-/"/ <<
-  $mode(string);
-  ${context.string_value} = String::new();
->>
-string: token string /"/ <<
-  $$ = JSONValue::StringVal(std::mem::take(&mut ${context.string_value}));
-  $mode(default);
->>
-string: /\\"/ <<
-  ${context.string_value}.push('"');
->>
-string: /\\\\/ <<
-  ${context.string_value}.push('\\');
->>
-string: /\\\// <<
-  ${context.string_value}.push('/');
->>
-string: /\\b/ <<
-  ${context.string_value}.push('\u{0008}');
->>
-string: /\\f/ <<
-  ${context.string_value}.push('\u{000C}');
->>
-string: /\\n/ <<
-  ${context.string_value}.push('\n');
->>
-string: /\\r/ <<
-  ${context.string_value}.push('\r');
->>
-string: /\\t/ <<
-  ${context.string_value}.push('\t');
->>
-string: /\\u[0-9a-fA-F]{4}/ <<
-  /* Not actually going to encode the code point for this example... */
-  let s: String = ['{', match_text[2] as char, match_text[3] as char, match_text[4] as char, match_text[5] as char, '}'].iter().collect();
-  ${context.string_value}.push_str(&s);
->>
-string: /[^\\]/ <<
-  ${context.string_value}.push(match_text[0] as char);
->>
-Start -> Value <<
-  $$ = $1;
->>
-Value -> string <<
-  $$ = $1;
->>
-Value -> number <<
-  $$ = $1;
->>
-Value -> Object <<
-  $$ = $1;
->>
-Value -> Array <<
-  $$ = $1;
->>
-Value -> true <<
-  $$ = $1;
->>
-Value -> false <<
-  $$ = $1;
->>
-Value -> null <<
-  $$ = $1;
->>
-Object -> lbrace rbrace <<
-  $$ = JSONValue::Object(Vec::new());
->>
-Object -> lbrace KeyValues rbrace <<
-  $$ = $2;
->>
-KeyValues -> KeyValue <<
-  $$ = $1;
->>
-KeyValues -> KeyValues comma KeyValue <<
-  let mut obj = $1;
-  if let JSONValue::Object(kve) = $3 {
-      if let JSONValue::Object(entries) = &mut obj {
-          entries.extend(kve);
-      }
-  }
-  $$ = obj;
->>
-KeyValue -> string colon Value <<
-  let name = if let JSONValue::StringVal(s) = $1 { s } else { String::new() };
-  $$ = JSONValue::Object(vec![(name, $3)]);
->>
-Array -> lbracket rbracket <<
-  $$ = JSONValue::Array(Vec::new());
->>
-Array -> lbracket Values rbracket <<
-  $$ = $2;
->>
-Values -> Value <<
-  $$ = $1;
->>
-Values -> Values comma Value <<
-  let mut arr = $1;
-  if let JSONValue::Array(elems) = &mut arr {
-      elems.push($3);
-  }
-  $$ = arr;
->>

spec/macros.c.propane

@@ -1,25 +0,0 @@
-<<
-#include <stdlib.h>
-size_t mylexfn(p_context_t * context, p_token_info_t * out_token_info);
-void record(int v);
->>
-
-ptype int;
-
-lex_fn mylexfn;
-
-drop /\s+/;
-token lbrace /\{/;
-token rbrace /\}/;
-token plus /\+/;
-token macro;
-token macroname /@[a-zA-Z_]\w*/;
-token num /\d+/ << char b[100]; memcpy(b, match_text, match_length); b[match_length] = '\0'; $$ = atoi(b); >>
-
-Start -> Statements;
-Statements -> ;
-Statements -> Statement Statements;
-Statement -> Add;
-Statement -> MacroStart;
-Add -> num plus num << $$ = $1 + $3; record($$); >>
-MacroStart -> macro macroname lbrace;

spec/macros.d.propane

@@ -1,31 +0,0 @@
-<<
-import test_macros;
->>
-
-ptype int;
-
-lex_fn mylexfn;
-
-drop /\s+/;
-token lbrace /\{/;
-token rbrace /\}/;
-token plus /\+/;
-token macro;
-token macroname /@[a-zA-Z_]\w*/;
-token num /\d+/ <<
-    int n = 0;
-    foreach (c; match_text)
-    {
-        n *= 10;
-        n += (c - '0');
-    }
-    $$ = n;
->>
-
-Start -> Statements;
-Statements -> ;
-Statements -> Statement Statements;
-Statement -> Add;
-Statement -> MacroStart;
-Add -> num plus num << $$ = $1 + $3; record($$); >>
-MacroStart -> macro macroname lbrace;

spec/macros.rust.propane

@@ -1,80 +0,0 @@
-<<
-fn mylexfn(context: &mut p_context_t, out_token_info: &mut p_token_info_t) -> usize {
-    loop {
-        if context.expanding {
-            let ei = context.expand_i;
-            context.expand_i += 1;
-            if context.expand_i >= context.token_infos.len() {
-                context.expanding = false;
-            }
-            *out_token_info = context.token_infos[ei].clone();
-            return P_SUCCESS;
-        }
-
-        let lex_result = p_lex(context, out_token_info);
-        if lex_result != P_SUCCESS {
-            return lex_result;
-        }
-
-        if out_token_info.token == TOKEN_macro {
-            context.defining = true;
-        } else if out_token_info.token == TOKEN_macroname {
-            if !context.defining {
-                context.expanding = true;
-                context.expand_i = 0;
-                continue;
-            }
-        } else if out_token_info.token == TOKEN_lbrace {
-            if context.defining {
-                /* Capture the macro body tokens (up to the closing '}'). */
-                let mut infos: Vec<p_token_info_t> = Vec::new();
-                loop {
-                    let mut ti = p_token_info_t::default();
-                    assert_eq!(P_SUCCESS, p_lex(context, &mut ti));
-                    if ti.token == TOKEN_rbrace {
-                        break;
-                    }
-                    infos.push(ti);
-                }
-                context.token_infos = infos;
-                context.defining = false;
-            }
-        } else {
-            context.defining = false;
-        }
-        return lex_result;
-    }
-}
->>
-
-context_user_fields <<
-    pub defining: bool,
-    pub expanding: bool,
-    pub expand_i: usize,
-    pub token_infos: Vec<p_token_info_t>,
-    pub nums: Vec<i64>,
->>
-
-ptype i64;
-
-lex_fn mylexfn;
-
-drop /\s+/;
-token lbrace /\{/;
-token rbrace /\}/;
-token plus /\+/;
-token macro;
-token macroname /@[a-zA-Z_]\w*/;
-token num /\d+/ <<
-  let mut v: i64 = 0;
-  for c in match_text { v = v * 10 + (*c - b'0') as i64; }
-  $$ = v;
->>
-
-Start -> Statements;
-Statements -> ;
-Statements -> Statement Statements;
-Statement -> Add;
-Statement -> MacroStart;
-Add -> num plus num << $$ = $1 + $3; ${context.nums}.push($$); >>
-MacroStart -> macro macroname lbrace;

spec/parse_inner_nested.c.propane

@@ -1,19 +0,0 @@
-<<
-#include <stdlib.h>
-#include <string.h>
-size_t mylexfn(p_context_t * context, p_token_info_t * out_token_info);
->>
-
-ptype int;
-
-lex_fn mylexfn;
-
-drop /\s+/;
-token lparen /\(/;
-token rparen /\)/;
-token plus /\+/;
-token num /\d+/ << char b[32]; memcpy(b, match_text, match_length); b[match_length] = '\0'; $$ = atoi(b); >>
-
-Start -> Expr << $$ = $1; >>
-Expr -> num << $$ = $1; >>
-Expr -> Expr plus num << $$ = $1 + $3; >>

spec/parse_inner_nested.d.propane

@@ -1,17 +0,0 @@
-<<
-import test_parse_inner_nested;
->>
-
-ptype int;
-
-lex_fn mylexfn;
-
-drop /\s+/;
-token lparen /\(/;
-token rparen /\)/;
-token plus /\+/;
-token num /\d+/ << int n = 0; foreach (ch; match_text) { n *= 10; n += (ch - '0'); } $$ = n; >>
-
-Start -> Expr << $$ = $1; >>
-Expr -> num << $$ = $1; >>
-Expr -> Expr plus num << $$ = $1 + $3; >>

spec/parse_inner_nested.rust.propane

@@ -1,41 +0,0 @@
-<<
-fn mylexfn(context: &mut p_context_t, out_token_info: &mut p_token_info_t) -> usize {
-    let result = p_lex(context, out_token_info);
-    if result != P_SUCCESS {
-        return result;
-    }
-    if out_token_info.token == TOKEN_lparen {
-        /* Reentrant nested parse of the parenthesized sub-expression. */
-        let inner_result = p_parse_inner_Start(context, &[TOKEN_rparen]);
-        if inner_result != P_SUCCESS {
-            return inner_result;
-        }
-        let value = p_result_Start(context);
-        /* p_parse_inner rewound the input so ')' was not consumed; consume it. */
-        let mut rparen_info = p_token_info_t::default();
-        assert_eq!(P_SUCCESS, p_lex(context, &mut rparen_info));
-        assert_eq!(TOKEN_rparen, rparen_info.token);
-        out_token_info.token = TOKEN_num;
-        out_token_info.pvalue = p_value(value);
-    }
-    P_SUCCESS
-}
->>
-
-ptype i64;
-
-lex_fn mylexfn;
-
-drop /\s+/;
-token lparen /\(/;
-token rparen /\)/;
-token plus /\+/;
-token num /\d+/ <<
-  let mut v: i64 = 0;
-  for c in match_text { v = v * 10 + (*c - b'0') as i64; }
-  $$ = v;
->>
-
-Start -> Expr << $$ = $1; >>
-Expr -> num << $$ = $1; >>
-Expr -> Expr plus num << $$ = $1 + $3; >>

spec/parse_inner_nested_tree.c.propane

@@ -1,17 +0,0 @@
-<<
-size_t mylexfn(p_context_t * context, p_token_info_t * out_token_info);
->>
-
-tree;
-
-lex_fn mylexfn;
-
-drop /\s+/;
-token lparen /\(/;
-token rparen /\)/;
-token plus /\+/;
-token num /\d+/;
-
-Start -> Expr;
-Expr -> num;
-Expr -> Expr plus num;

spec/parse_inner_nested_tree.d.propane

@@ -1,17 +0,0 @@
-<<
-import test_parse_inner_nested_tree;
->>
-
-tree;
-
-lex_fn mylexfn;
-
-drop /\s+/;
-token lparen /\(/;
-token rparen /\)/;
-token plus /\+/;
-token num /\d+/;
-
-Start -> Expr;
-Expr -> num;
-Expr -> Expr plus num;

spec/parse_inner_nested_tree.rust.propane

@@ -1,44 +0,0 @@
-<<
-fn mylexfn(context: &mut p_context_t, out_token_info: &mut p_token_info_t) -> usize {
-    let result = p_lex(context, out_token_info);
-    if result != P_SUCCESS {
-        return result;
-    }
-    if out_token_info.token == TOKEN_lparen {
-        let start_position = out_token_info.position;
-        let inner_result = p_parse_inner_Start(context, &[TOKEN_rparen]);
-        if inner_result != P_SUCCESS {
-            return inner_result;
-        }
-        /* Read the inner subtree's span before re-borrowing context to lex. */
-        let inner = p_result_Start(context);
-        assert!(inner.valid());
-        let inner_start_col = inner.position().col;
-        let inner_end_col = inner.end_position().col;
-        let mut rparen_info = p_token_info_t::default();
-        assert_eq!(P_SUCCESS, p_lex(context, &mut rparen_info));
-        assert_eq!(TOKEN_rparen, rparen_info.token);
-        assert_eq!(start_position.col + 1, inner_start_col);
-        assert_eq!(rparen_info.position.col - 1, inner_end_col);
-        /* Synthesize a num token spanning the whole "( ... )" group. */
-        out_token_info.token = TOKEN_num;
-        out_token_info.position = start_position;
-        out_token_info.end_position = rparen_info.end_position;
-    }
-    P_SUCCESS
-}
->>
-
-tree;
-
-lex_fn mylexfn;
-
-drop /\s+/;
-token lparen /\(/;
-token rparen /\)/;
-token plus /\+/;
-token num /\d+/;
-
-Start -> Expr;
-Expr -> num;
-Expr -> Expr plus num;

spec/propane/grammar_spec.rb

@@ -28,7 +28,7 @@ B -> <<
   b = 0;
 >>
 EOF
-      grammar = Grammar.new(input, "test.propane")
+      grammar = Grammar.new(input)
       expect(grammar.modulename).to eq "a.b"
       expect(grammar.ptype).to eq "XYZ  *"
       expect(grammar.ptypes).to eq("default" => "XYZ  *")
@@ -62,7 +62,7 @@ EOF
       expect(o).to_not be_nil
       expect(o.pattern).to eq "token_with_code"
       expect(o.line_number).to eq 11
-      expect(o.code).to eq %[#line 12 "test.propane"\nCode for the token\n#linereset\n]
+      expect(o.code).to eq "Code for the token\n"
 
       o = grammar.tokens.find {|token| token.name == "token_with_no_pattern"}
       expect(o).to_not be_nil
@@ -83,7 +83,7 @@ EOF
       expect(o.name).to eq "A"
       expect(o.components).to eq %w[B]
       expect(o.line_number).to eq 19
-      expect(o.code).to eq %[#line 20 "test.propane"\n  a = 42;\n#linereset\n]
+      expect(o.code).to eq "  a = 42;\n"
 
       o = grammar.rules[1]
       expect(o.name).to eq "B"
@@ -95,7 +95,7 @@ EOF
       expect(o.name).to eq "B"
       expect(o.components).to eq []
       expect(o.line_number).to eq 23
-      expect(o.code).to eq %[#line 24 "test.propane"\n  b = 0;\n#linereset\n]
+      expect(o.code).to eq "  b = 0;\n"
     end
 
     it "parses code segments with semicolons" do
@@ -113,7 +113,7 @@ tokenid token_with_no_pattern;
 
 prefix myparser_;
 EOF
-      grammar = Grammar.new(input, "test.propane")
+      grammar = Grammar.new(input)
       expect(grammar.prefix).to eq "myparser_"
 
       o = grammar.tokens.find {|token| token.name == "code1"}
@@ -122,7 +122,7 @@ EOF
 
       o = grammar.patterns.find {|pattern| pattern.token == o}
       expect(o).to_not be_nil
-      expect(o.code).to eq %[#line 2 "test.propane"\n  a = b;\n  return c;\n#linereset\n]
+      expect(o.code).to eq "  a = b;\n  return c;\n"
 
       o = grammar.tokens.find {|token| token.name == "code2"}
       expect(o).to_not be_nil
@@ -130,42 +130,7 @@ EOF
 
       o = grammar.patterns.find {|pattern| pattern.token == o}
       expect(o).to_not be_nil
-      expect(o.code).to eq %[#line 7 "test.propane"\n  writeln("Hello there");\n#linereset\n]
+      expect(o.code).to eq %[  writeln("Hello there");\n]
-    end
-
-    it "does not emit #line directives with noline statement" do
-      input = <<EOF
-noline;
-
-token code1 <<
-  a = b;
-  return c;
->>
-
-token code2 <<
-  writeln("Hello there");
->>
-
-tokenid token_with_no_pattern;
-
-prefix myparser_;
-EOF
-      grammar = Grammar.new(input, "test.propane")
-      expect(grammar.prefix).to eq "myparser_"
-
-      o = grammar.tokens.find {|token| token.name == "code1"}
-      expect(o).to_not be_nil
-
-      o = grammar.patterns.find {|pattern| pattern.token == o}
-      expect(o).to_not be_nil
-      expect(o.code).to eq %[\n  a = b;\n  return c;]
-
-      o = grammar.tokens.find {|token| token.name == "code2"}
-      expect(o).to_not be_nil
-
-      o = grammar.patterns.find {|pattern| pattern.token == o}
-      expect(o).to_not be_nil
-      expect(o.code).to eq %[\n  writeln("Hello there");]
     end
 
     it "supports mode labels" do
@@ -179,7 +144,7 @@ m2: /bar/ <<
 drop /q/;
 m3: drop /r/;
 EOF
-      grammar = Grammar.new(input, "test.propane")
+      grammar = Grammar.new(input)
 
       o = grammar.tokens.find {|token| token.name == "a"}
       expect(o).to_not be_nil
@@ -232,7 +197,7 @@ tokenid int(integer);
 Start (node) -> R;
 R -> abc int;
 EOF
-      grammar = Grammar.new(input, "test.propane")
+      grammar = Grammar.new(input)
 
       o = grammar.tokens.find {|token| token.name == "abc"}
       expect(o).to_not be_nil

spec/propane/lexer/dfa_spec.rb

@@ -51,7 +51,7 @@ class TestLexer
 end
 
 def run(grammar, input)
-  grammar = Propane::Grammar.new(grammar, "test.propane")
+  grammar = Propane::Grammar.new(grammar)
   token_dfa = Propane::Lexer::DFA.new(grammar.patterns)
   test_lexer = TestLexer.new(token_dfa)
   test_lexer.lex(input)

spec/rewind.c.propane

@@ -1,23 +0,0 @@
-<<
-#include <stdlib.h>
-#include <string.h>
-size_t mylexfn(p_context_t * context, p_token_info_t * out_token_info);
-void record(int value);
->>
-
-ptype int;
-
-lex_fn mylexfn;
-
-drop /\s+/;
-token repeat /repeat/;
-token lbrace /\{/;
-token rbrace /\}/;
-token plus /\+/;
-token num /\d+/ << char b[32]; memcpy(b, match_text, match_length); b[match_length] = '\0'; $$ = atoi(b); >>
-
-Start -> Statements;
-Statements -> ;
-Statements -> Statement Statements;
-Statement -> Add;
-Add -> num plus num << record($1 + $3); >>

spec/rewind.d.propane

@@ -1,20 +0,0 @@
-<<
-import test_rewind;
->>
-
-ptype int;
-
-lex_fn mylexfn;
-
-drop /\s+/;
-token repeat /repeat/;
-token lbrace /\{/;
-token rbrace /\}/;
-token plus /\+/;
-token num /\d+/ << int n = 0; foreach (ch; match_text) { n *= 10; n += (ch - '0'); } $$ = n; >>
-
-Start -> Statements;
-Statements -> ;
-Statements -> Statement Statements;
-Statement -> Add;
-Add -> num plus num << record($1 + $3); >>

spec/rewind.rust.propane

@@ -1,67 +0,0 @@
-<<
-fn mylexfn(context: &mut p_context_t, out_token_info: &mut p_token_info_t) -> usize {
-    loop {
-        let result = p_lex(context, out_token_info);
-        if result != P_SUCCESS {
-            return result;
-        }
-        if out_token_info.token == TOKEN_repeat {
-            let mut count_info = p_token_info_t::default();
-            assert_eq!(P_SUCCESS, p_lex(context, &mut count_info));
-            assert_eq!(TOKEN_num, count_info.token);
-            let mut brace_info = p_token_info_t::default();
-            assert_eq!(P_SUCCESS, p_lex(context, &mut brace_info));
-            assert_eq!(TOKEN_lbrace, brace_info.token);
-            context.remaining = p_value_get(&count_info.pvalue);
-            context.body_index = p_input_index(context);
-            context.body_position = p_position(context);
-            continue;
-        }
-        if out_token_info.token == TOKEN_rbrace {
-            if context.remaining > 1 {
-                context.remaining -= 1;
-                let bi = context.body_index;
-                let bp = context.body_position;
-                p_set_input_index(context, bi);
-                p_set_position(context, bp);
-                continue;
-            }
-            context.remaining = 0;
-            continue;
-        }
-        if out_token_info.token == TOKEN_num {
-            context.num_cols.push(out_token_info.position.col);
-        }
-        return result;
-    }
-}
->>
-
-context_user_fields <<
-    pub nums: Vec<i64>,
-    pub num_cols: Vec<u32>,
-    pub remaining: i64,
-    pub body_index: usize,
-    pub body_position: p_position_t,
->>
-
-ptype i64;
-
-lex_fn mylexfn;
-
-drop /\s+/;
-token repeat /repeat/;
-token lbrace /\{/;
-token rbrace /\}/;
-token plus /\+/;
-token num /\d+/ <<
-  let mut v: i64 = 0;
-  for c in match_text { v = v * 10 + (*c - b'0') as i64; }
-  $$ = v;
->>
-
-Start -> Statements;
-Statements -> ;
-Statements -> Statement Statements;
-Statement -> Add;
-Add -> num plus num << ${context.nums}.push($1 + $3); >>

spec/spec_helper.rb

@@ -15,10 +15,6 @@ unless ENV["dist_specs"]
       command_name "RSpec"
     end
     project_name "Propane"
-    # Keep this process's results separate from the propane subprocess results
-    # so that nothing has to merge on the fly; the spec Rake task collates all
-    # of the parts once the suite is done.
-    coverage_dir "coverage/parts/rspec"
     merge_timeout 3600
     formatter(MyFormatter)
   end

spec/test_ast.c

@@ -0,0 +1,61 @@
+#include "testparser.h"
+#include <assert.h>
+#include <string.h>
+#include "testutils.h"
+
+int main()
+{
+    char const * input = "a, ((b)), b";
+    p_context_t context;
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert_eq(P_SUCCESS, p_parse(&context));
+    Start * start = p_result(&context);
+    assert(start->pItems1 != NULL);
+    assert(start->pItems != NULL);
+    Items * items = start->pItems;
+    assert(items->pItem != NULL);
+    assert(items->pItem->pToken1 != NULL);
+    assert_eq(TOKEN_a, items->pItem->pToken1->token);
+    assert_eq(11, items->pItem->pToken1->pvalue);
+    assert(items->pItemsMore != NULL);
+    ItemsMore * itemsmore = items->pItemsMore;
+    assert(itemsmore->pItem != NULL);
+    assert(itemsmore->pItem->pItem != NULL);
+    assert(itemsmore->pItem->pItem->pItem != NULL);
+    assert(itemsmore->pItem->pItem->pItem->pToken1 != NULL);
+    assert_eq(TOKEN_b, itemsmore->pItem->pItem->pItem->pToken1->token);
+    assert_eq(22, itemsmore->pItem->pItem->pItem->pToken1->pvalue);
+    assert(itemsmore->pItemsMore != NULL);
+    itemsmore = itemsmore->pItemsMore;
+    assert(itemsmore->pItem != NULL);
+    assert(itemsmore->pItem->pToken1 != NULL);
+    assert_eq(TOKEN_b, itemsmore->pItem->pToken1->token);
+    assert_eq(22, itemsmore->pItem->pToken1->pvalue);
+    assert(itemsmore->pItemsMore == NULL);
+
+    p_free_ast(start);
+
+    input = "";
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert_eq(P_SUCCESS, p_parse(&context));
+    start = p_result(&context);
+    assert(start->pItems == NULL);
+
+    p_free_ast(start);
+
+    input = "2 1";
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert_eq(P_SUCCESS, p_parse(&context));
+    start = p_result(&context);
+    assert(start->pItems != NULL);
+    assert(start->pItems->pItem != NULL);
+    assert(start->pItems->pItem->pDual != NULL);
+    assert(start->pItems->pItem->pDual->pTwo1 != NULL);
+    assert(start->pItems->pItem->pDual->pOne2 != NULL);
+    assert(start->pItems->pItem->pDual->pTwo2 == NULL);
+    assert(start->pItems->pItem->pDual->pOne1 == NULL);
+
+    p_free_ast(start);
+
+    return 0;
+}

spec/test_ast.d

@@ -0,0 +1,57 @@
+import testparser;
+import std.stdio;
+import testutils;
+
+int main()
+{
+    return 0;
+}
+
+unittest
+{
+    string input = "a, ((b)), b";
+    p_context_t context;
+    p_context_init(&context, input);
+    assert_eq(P_SUCCESS, p_parse(&context));
+    Start * start = p_result(&context);
+    assert(start.pItems1 !is null);
+    assert(start.pItems !is null);
+    Items * items = start.pItems;
+    assert(items.pItem !is null);
+    assert(items.pItem.pToken1 !is null);
+    assert_eq(TOKEN_a, items.pItem.pToken1.token);
+    assert_eq(11, items.pItem.pToken1.pvalue);
+    assert(items.pItemsMore !is null);
+    ItemsMore * itemsmore = items.pItemsMore;
+    assert(itemsmore.pItem !is null);
+    assert(itemsmore.pItem.pItem !is null);
+    assert(itemsmore.pItem.pItem.pItem !is null);
+    assert(itemsmore.pItem.pItem.pItem.pToken1 !is null);
+    assert_eq(TOKEN_b, itemsmore.pItem.pItem.pItem.pToken1.token);
+    assert_eq(22, itemsmore.pItem.pItem.pItem.pToken1.pvalue);
+    assert(itemsmore.pItemsMore !is null);
+    itemsmore = itemsmore.pItemsMore;
+    assert(itemsmore.pItem !is null);
+    assert(itemsmore.pItem.pToken1 !is null);
+    assert_eq(TOKEN_b, itemsmore.pItem.pToken1.token);
+    assert_eq(22, itemsmore.pItem.pToken1.pvalue);
+    assert(itemsmore.pItemsMore is null);
+
+    input = "";
+    p_context_init(&context, input);
+    assert_eq(P_SUCCESS, p_parse(&context));
+    start = p_result(&context);
+    assert(start.pItems is null);
+
+    input = "2 1";
+    p_context_init(&context, input);
+    assert_eq(P_SUCCESS, p_parse(&context));
+    start = p_result(&context);
+    assert(start.pItems !is null);
+    assert(start.pItems.pItem !is null);
+    assert(start.pItems.pItem.pDual !is null);
+    assert(start.pItems.pItem.pDual.pTwo1 !is null);
+    assert(start.pItems.pItem.pDual.pOne2 !is null);
+    assert(start.pItems.pItem.pDual.pTwo2 is null);
+    assert(start.pItems.pItem.pDual.pOne1 is null);
+}

spec/test_ast_field_aliases.c

@@ -0,0 +1,21 @@
+#include "testparser.h"
+#include <assert.h>
+#include <string.h>
+#include "testutils.h"
+
+int main()
+{
+    char const * input = "\na\nb\nc";
+    p_context_t context;
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_SUCCESS);
+    Start * start = p_result(&context);
+
+    assert_eq(TOKEN_a, start->first->pToken->token);
+    assert_eq(TOKEN_b, start->second->pToken->token);
+    assert_eq(TOKEN_c, start->third->pToken->token);
+
+    p_free_ast(start);
+
+    return 0;
+}

spec/test_ast_field_aliases.d

@@ -10,13 +10,12 @@ int main()
 unittest
 {
     string input = "\na\nb\nc";
-    p_context_t * context = p_context_new(input);
+    p_context_t context;
-    assert(p_parse(context) == P_SUCCESS);
+    p_context_init(&context, input);
-    Start start = p_result(context);
+    assert(p_parse(&context) == P_SUCCESS);
+    Start * start = p_result(&context);
 
     assert_eq(TOKEN_a, start.first.pToken.token);
     assert_eq(TOKEN_b, start.second.pToken.token);
     assert_eq(TOKEN_c, start.third.pToken.token);
-
-    p_context_delete(context);
 }

spec/test_ast_invalid_positions.c

@@ -0,0 +1,110 @@
+#include "testparser.h"
+#include <assert.h>
+#include <string.h>
+#include "testutils.h"
+
+int main()
+{
+    char const * input = "\na\n  bb ccc";
+    p_context_t context;
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_SUCCESS);
+    Start * start = p_result(&context);
+
+    assert_eq(2, start->pT1->pToken->position.row);
+    assert_eq(1, start->pT1->pToken->position.col);
+    assert_eq(2, start->pT1->pToken->end_position.row);
+    assert_eq(1, start->pT1->pToken->end_position.col);
+    assert(p_position_valid(start->pT1->pA->position));
+    assert_eq(3, start->pT1->pA->position.row);
+    assert_eq(3, start->pT1->pA->position.col);
+    assert_eq(3, start->pT1->pA->end_position.row);
+    assert_eq(8, start->pT1->pA->end_position.col);
+    assert_eq(2, start->pT1->position.row);
+    assert_eq(1, start->pT1->position.col);
+    assert_eq(3, start->pT1->end_position.row);
+    assert_eq(8, start->pT1->end_position.col);
+
+    assert_eq(2, start->position.row);
+    assert_eq(1, start->position.col);
+    assert_eq(3, start->end_position.row);
+    assert_eq(8, start->end_position.col);
+
+    p_free_ast(start);
+
+    input = "a\nbb";
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_SUCCESS);
+    start = p_result(&context);
+
+    assert_eq(1, start->pT1->pToken->position.row);
+    assert_eq(1, start->pT1->pToken->position.col);
+    assert_eq(1, start->pT1->pToken->end_position.row);
+    assert_eq(1, start->pT1->pToken->end_position.col);
+    assert(p_position_valid(start->pT1->pA->position));
+    assert_eq(2, start->pT1->pA->position.row);
+    assert_eq(1, start->pT1->pA->position.col);
+    assert_eq(2, start->pT1->pA->end_position.row);
+    assert_eq(2, start->pT1->pA->end_position.col);
+    assert_eq(1, start->pT1->position.row);
+    assert_eq(1, start->pT1->position.col);
+    assert_eq(2, start->pT1->end_position.row);
+    assert_eq(2, start->pT1->end_position.col);
+
+    assert_eq(1, start->position.row);
+    assert_eq(1, start->position.col);
+    assert_eq(2, start->end_position.row);
+    assert_eq(2, start->end_position.col);
+
+    p_free_ast(start);
+
+    input = "a\nc\nc";
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_SUCCESS);
+    start = p_result(&context);
+
+    assert_eq(1, start->pT1->pToken->position.row);
+    assert_eq(1, start->pT1->pToken->position.col);
+    assert_eq(1, start->pT1->pToken->end_position.row);
+    assert_eq(1, start->pT1->pToken->end_position.col);
+    assert(p_position_valid(start->pT1->pA->position));
+    assert_eq(2, start->pT1->pA->position.row);
+    assert_eq(1, start->pT1->pA->position.col);
+    assert_eq(3, start->pT1->pA->end_position.row);
+    assert_eq(1, start->pT1->pA->end_position.col);
+    assert_eq(1, start->pT1->position.row);
+    assert_eq(1, start->pT1->position.col);
+    assert_eq(3, start->pT1->end_position.row);
+    assert_eq(1, start->pT1->end_position.col);
+
+    assert_eq(1, start->position.row);
+    assert_eq(1, start->position.col);
+    assert_eq(3, start->end_position.row);
+    assert_eq(1, start->end_position.col);
+
+    p_free_ast(start);
+
+    input = "a";
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_SUCCESS);
+    start = p_result(&context);
+
+    assert_eq(1, start->pT1->pToken->position.row);
+    assert_eq(1, start->pT1->pToken->position.col);
+    assert_eq(1, start->pT1->pToken->end_position.row);
+    assert_eq(1, start->pT1->pToken->end_position.col);
+    assert(!p_position_valid(start->pT1->pA->position));
+    assert_eq(1, start->pT1->position.row);
+    assert_eq(1, start->pT1->position.col);
+    assert_eq(1, start->pT1->end_position.row);
+    assert_eq(1, start->pT1->end_position.col);
+
+    assert_eq(1, start->position.row);
+    assert_eq(1, start->position.col);
+    assert_eq(1, start->end_position.row);
+    assert_eq(1, start->end_position.col);
+
+    p_free_ast(start);
+
+    return 0;
+}

spec/test_ast_invalid_positions.d

@@ -10,9 +10,10 @@ int main()
 unittest
 {
     string input = "\na\n  bb ccc";
-    p_context_t * context = p_context_new(input);
+    p_context_t context;
-    assert(p_parse(context) == P_SUCCESS);
+    p_context_init(&context, input);
-    Start start = p_result(context);
+    assert(p_parse(&context) == P_SUCCESS);
+    Start * start = p_result(&context);
 
     assert_eq(2, start.pT1.pToken.position.row);
     assert_eq(1, start.pT1.pToken.position.col);
@@ -33,12 +34,10 @@ unittest
     assert_eq(3, start.end_position.row);
     assert_eq(8, start.end_position.col);
 
-    p_context_delete(context);
-
     input = "a\nbb";
-    context = p_context_new(input);
+    p_context_init(&context, input);
-    assert(p_parse(context) == P_SUCCESS);
+    assert(p_parse(&context) == P_SUCCESS);
-    start = p_result(context);
+    start = p_result(&context);
 
     assert_eq(1, start.pT1.pToken.position.row);
     assert_eq(1, start.pT1.pToken.position.col);
@@ -59,12 +58,10 @@ unittest
     assert_eq(2, start.end_position.row);
     assert_eq(2, start.end_position.col);
 
-    p_context_delete(context);
-
     input = "a\nc\nc";
-    context = p_context_new(input);
+    p_context_init(&context, input);
-    assert(p_parse(context) == P_SUCCESS);
+    assert(p_parse(&context) == P_SUCCESS);
-    start = p_result(context);
+    start = p_result(&context);
 
     assert_eq(1, start.pT1.pToken.position.row);
     assert_eq(1, start.pT1.pToken.position.col);
@@ -85,12 +82,10 @@ unittest
     assert_eq(3, start.end_position.row);
     assert_eq(1, start.end_position.col);
 
-    p_context_delete(context);
-
     input = "a";
-    context = p_context_new(input);
+    p_context_init(&context, input);
-    assert(p_parse(context) == P_SUCCESS);
+    assert(p_parse(&context) == P_SUCCESS);
-    start = p_result(context);
+    start = p_result(&context);
 
     assert_eq(1, start.pT1.pToken.position.row);
     assert_eq(1, start.pT1.pToken.position.col);
@@ -106,6 +101,4 @@ unittest
     assert_eq(1, start.position.col);
     assert_eq(1, start.end_position.row);
     assert_eq(1, start.end_position.col);
-
-    p_context_delete(context);
 }

spec/test_ast_node_memory_remains.c

@@ -369,50 +369,50 @@ int main(int argc, char * argv[])
         {"size_t_to_ulong", TOKEN_ulong},
         {"main", TOKEN_int},
     };
-    p_context_t * context;
+    p_context_t context;
-    context = p_context_new((const uint8_t *)input, strlen(input));
+    p_context_init(&context, (const uint8_t *)input, strlen(input));
-    size_t result = p_parse(context);
+    size_t result = p_parse(&context);
     assert_eq(P_SUCCESS, result);
-    PModule pmod = p_result(context);
+    PModule * pmod = p_result(&context);
-    PModuleItems pmis = p_PModule_pModuleItems(pmod);
+    PModuleItems * pmis = pmod->pModuleItems;
-    PFunctionDefinition * pfds;
+    PFunctionDefinition ** pfds;
     size_t n_pfds = 0u;
-    while (p_node_valid(pmis))
+    while (pmis != NULL)
     {
-        PModuleItem pmi = p_PModuleItems_pModuleItem(pmis);
+        PModuleItem * pmi = pmis->pModuleItem;
-        if (p_node_valid(p_PModuleItem_pFunctionDefinition(pmi)))
+        if (pmi->pFunctionDefinition != NULL)
         {
             n_pfds++;
         }
-        pmis = p_PModuleItems_pModuleItems(pmis);
+        pmis = pmis->pModuleItems;
     }
-    pfds = (PFunctionDefinition *)malloc(n_pfds * sizeof(PFunctionDefinition));
+    pfds = (PFunctionDefinition **)malloc(n_pfds * sizeof(PModuleItems *));
-    pmis = p_PModule_pModuleItems(pmod);
+    pmis = pmod->pModuleItems;
     size_t pfd_i = n_pfds;
-    while (p_node_valid(pmis))
+    while (pmis != NULL)
     {
-        PModuleItem pmi = p_PModuleItems_pModuleItem(pmis);
+        PModuleItem * pmi = pmis->pModuleItem;
-        PFunctionDefinition pfd = p_PModuleItem_pFunctionDefinition(pmi);
+        PFunctionDefinition * pfd = pmi->pFunctionDefinition;
-        if (p_node_valid(pfd))
+        if (pfd != NULL)
         {
             pfd_i--;
             assert(pfd_i < n_pfds);
             pfds[pfd_i] = pfd;
         }
-        pmis = p_PModuleItems_pModuleItems(pmis);
+        pmis = pmis->pModuleItems;
     }
     assert_eq(51, n_pfds);
     for (size_t i = 0; i < n_pfds; i++)
     {
-        if (strncmp(expected[i].name, (const char *)p_node_data(p_PFunctionDefinition_name(pfds[i]))->pvalue.s, strlen(expected[i].name)) != 0 ||
+        if (strncmp(expected[i].name, (const char *)pfds[i]->name->pvalue.s, strlen(expected[i].name)) != 0 ||
-            (expected[i].token != p_tree_walk_PFunctionDefinition(pfds[i], returntype, pType, pTypeBase, pToken1, token)))
+            (expected[i].token != pfds[i]->returntype->pType->pTypeBase->pToken1->token))
         {
-            fprintf(stderr, "Index %lu: expected %s/%u, got %u\n", i, expected[i].name, expected[i].token, p_tree_walk_PFunctionDefinition(pfds[i], returntype, pType, pTypeBase, pToken1, token));
+            fprintf(stderr, "Index %lu: expected %s/%u, got %u\n", i, expected[i].name, expected[i].token, pfds[i]->returntype->pType->pTypeBase->pToken1->token);
         }
     }
 
     free(pfds);
-    p_context_delete(context);
+    p_free_ast(pmod);
 
     return 0;
 }

spec/test_ast_node_memory_remains.d

@@ -374,23 +374,23 @@ def main() -> int
         Expected("size_t_to_ulong", TOKEN_ulong),
         Expected("main", TOKEN_int),
     ];
-    p_context_t * context;
+    p_context_t context;
-    context = p_context_new(input);
+    p_context_init(&context, input);
-    size_t result = p_parse(context);
+    size_t result = p_parse(&context);
     assert_eq(P_SUCCESS, result);
-    PModule pmod = p_result(context);
+    PModule * pmod = p_result(&context);
-    PModuleItems pmis = pmod.pModuleItems;
+    PModuleItems * pmis = pmod.pModuleItems;
-    PFunctionDefinition[] pfds;
+    PFunctionDefinition *[] pfds;
-    while (pmis.valid)
+    while (pmis !is null)
     {
-        PModuleItem pmi = pmis.pModuleItem;
+        PModuleItem * pmi = pmis.pModuleItem;
-        if (!pmi.valid)
+        if (pmi is null)
         {
             stderr.writeln("pmi is null!!!?");
             assert(0);
         }
-        PFunctionDefinition pfd = pmi.pFunctionDefinition;
+        PFunctionDefinition * pfd = pmi.pFunctionDefinition;
-        if (pfd.valid)
+        if (pfd !is null)
         {
             pfds = [pfd] ~ pfds;
         }
@@ -405,5 +405,4 @@ def main() -> int
             stderr.writeln("Index ", i, ": expected ", expected[i].name, "/", expected[i].token, ", got ", pfds[i].name.pvalue.s, "/", pfds[i].returntype.pType.pTypeBase.pToken1.token);
         }
     }
-    p_context_delete(context);
 }

spec/test_ast_ps.c

@@ -0,0 +1,61 @@
+#include "testparser.h"
+#include <assert.h>
+#include <string.h>
+#include "testutils.h"
+
+int main()
+{
+    char const * input = "a, ((b)), b";
+    p_context_t context;
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert_eq(P_SUCCESS, p_parse(&context));
+    PStartS * start = p_result(&context);
+    assert(start->pItems1 != NULL);
+    assert(start->pItems != NULL);
+    PItemsS * items = start->pItems;
+    assert(items->pItem != NULL);
+    assert(items->pItem->pToken1 != NULL);
+    assert_eq(TOKEN_a, items->pItem->pToken1->token);
+    assert_eq(11, items->pItem->pToken1->pvalue);
+    assert(items->pItemsMore != NULL);
+    PItemsMoreS * itemsmore = items->pItemsMore;
+    assert(itemsmore->pItem != NULL);
+    assert(itemsmore->pItem->pItem != NULL);
+    assert(itemsmore->pItem->pItem->pItem != NULL);
+    assert(itemsmore->pItem->pItem->pItem->pToken1 != NULL);
+    assert_eq(TOKEN_b, itemsmore->pItem->pItem->pItem->pToken1->token);
+    assert_eq(22, itemsmore->pItem->pItem->pItem->pToken1->pvalue);
+    assert(itemsmore->pItemsMore != NULL);
+    itemsmore = itemsmore->pItemsMore;
+    assert(itemsmore->pItem != NULL);
+    assert(itemsmore->pItem->pToken1 != NULL);
+    assert_eq(TOKEN_b, itemsmore->pItem->pToken1->token);
+    assert_eq(22, itemsmore->pItem->pToken1->pvalue);
+    assert(itemsmore->pItemsMore == NULL);
+
+    p_free_ast(start);
+
+    input = "";
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert_eq(P_SUCCESS, p_parse(&context));
+    start = p_result(&context);
+    assert(start->pItems == NULL);
+
+    p_free_ast(start);
+
+    input = "2 1";
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert_eq(P_SUCCESS, p_parse(&context));
+    start = p_result(&context);
+    assert(start->pItems != NULL);
+    assert(start->pItems->pItem != NULL);
+    assert(start->pItems->pItem->pDual != NULL);
+    assert(start->pItems->pItem->pDual->pTwo1 != NULL);
+    assert(start->pItems->pItem->pDual->pOne2 != NULL);
+    assert(start->pItems->pItem->pDual->pTwo2 == NULL);
+    assert(start->pItems->pItem->pDual->pOne1 == NULL);
+
+    p_free_ast(start);
+
+    return 0;
+}

spec/test_ast_ps.d

@@ -0,0 +1,57 @@
+import testparser;
+import std.stdio;
+import testutils;
+
+int main()
+{
+    return 0;
+}
+
+unittest
+{
+    string input = "a, ((b)), b";
+    p_context_t context;
+    p_context_init(&context, input);
+    assert_eq(P_SUCCESS, p_parse(&context));
+    PStartS * start = p_result(&context);
+    assert(start.pItems1 !is null);
+    assert(start.pItems !is null);
+    PItemsS * items = start.pItems;
+    assert(items.pItem !is null);
+    assert(items.pItem.pToken1 !is null);
+    assert_eq(TOKEN_a, items.pItem.pToken1.token);
+    assert_eq(11, items.pItem.pToken1.pvalue);
+    assert(items.pItemsMore !is null);
+    PItemsMoreS * itemsmore = items.pItemsMore;
+    assert(itemsmore.pItem !is null);
+    assert(itemsmore.pItem.pItem !is null);
+    assert(itemsmore.pItem.pItem.pItem !is null);
+    assert(itemsmore.pItem.pItem.pItem.pToken1 !is null);
+    assert_eq(TOKEN_b, itemsmore.pItem.pItem.pItem.pToken1.token);
+    assert_eq(22, itemsmore.pItem.pItem.pItem.pToken1.pvalue);
+    assert(itemsmore.pItemsMore !is null);
+    itemsmore = itemsmore.pItemsMore;
+    assert(itemsmore.pItem !is null);
+    assert(itemsmore.pItem.pToken1 !is null);
+    assert_eq(TOKEN_b, itemsmore.pItem.pToken1.token);
+    assert_eq(22, itemsmore.pItem.pToken1.pvalue);
+    assert(itemsmore.pItemsMore is null);
+
+    input = "";
+    p_context_init(&context, input);
+    assert_eq(P_SUCCESS, p_parse(&context));
+    start = p_result(&context);
+    assert(start.pItems is null);
+
+    input = "2 1";
+    p_context_init(&context, input);
+    assert_eq(P_SUCCESS, p_parse(&context));
+    start = p_result(&context);
+    assert(start.pItems !is null);
+    assert(start.pItems.pItem !is null);
+    assert(start.pItems.pItem.pDual !is null);
+    assert(start.pItems.pItem.pDual.pTwo1 !is null);
+    assert(start.pItems.pItem.pDual.pOne2 !is null);
+    assert(start.pItems.pItem.pDual.pTwo2 is null);
+    assert(start.pItems.pItem.pDual.pOne1 is null);
+}

spec/test_ast_token_positions.c

@@ -0,0 +1,88 @@
+#include "testparser.h"
+#include <assert.h>
+#include <string.h>
+#include "testutils.h"
+
+int main()
+{
+    char const * input = "abbccc";
+    p_context_t context;
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_SUCCESS);
+    Start * start = p_result(&context);
+
+    assert_eq(1, start->pT1->pToken->position.row);
+    assert_eq(1, start->pT1->pToken->position.col);
+    assert_eq(1, start->pT1->pToken->end_position.row);
+    assert_eq(1, start->pT1->pToken->end_position.col);
+    assert_eq(1, start->pT1->position.row);
+    assert_eq(1, start->pT1->position.col);
+    assert_eq(1, start->pT1->end_position.row);
+    assert_eq(1, start->pT1->end_position.col);
+
+    assert_eq(1, start->pT2->pToken->position.row);
+    assert_eq(2, start->pT2->pToken->position.col);
+    assert_eq(1, start->pT2->pToken->end_position.row);
+    assert_eq(3, start->pT2->pToken->end_position.col);
+    assert_eq(1, start->pT2->position.row);
+    assert_eq(2, start->pT2->position.col);
+    assert_eq(1, start->pT2->end_position.row);
+    assert_eq(3, start->pT2->end_position.col);
+
+    assert_eq(1, start->pT3->pToken->position.row);
+    assert_eq(4, start->pT3->pToken->position.col);
+    assert_eq(1, start->pT3->pToken->end_position.row);
+    assert_eq(6, start->pT3->pToken->end_position.col);
+    assert_eq(1, start->pT3->position.row);
+    assert_eq(4, start->pT3->position.col);
+    assert_eq(1, start->pT3->end_position.row);
+    assert_eq(6, start->pT3->end_position.col);
+
+    assert_eq(1, start->position.row);
+    assert_eq(1, start->position.col);
+    assert_eq(1, start->end_position.row);
+    assert_eq(6, start->end_position.col);
+
+    p_free_ast(start);
+
+    input = "\n\n  bb\nc\ncc\n\n     a";
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_SUCCESS);
+    start = p_result(&context);
+
+    assert_eq(3, start->pT1->pToken->position.row);
+    assert_eq(3, start->pT1->pToken->position.col);
+    assert_eq(3, start->pT1->pToken->end_position.row);
+    assert_eq(4, start->pT1->pToken->end_position.col);
+    assert_eq(3, start->pT1->position.row);
+    assert_eq(3, start->pT1->position.col);
+    assert_eq(3, start->pT1->end_position.row);
+    assert_eq(4, start->pT1->end_position.col);
+
+    assert_eq(4, start->pT2->pToken->position.row);
+    assert_eq(1, start->pT2->pToken->position.col);
+    assert_eq(5, start->pT2->pToken->end_position.row);
+    assert_eq(2, start->pT2->pToken->end_position.col);
+    assert_eq(4, start->pT2->position.row);
+    assert_eq(1, start->pT2->position.col);
+    assert_eq(5, start->pT2->end_position.row);
+    assert_eq(2, start->pT2->end_position.col);
+
+    assert_eq(7, start->pT3->pToken->position.row);
+    assert_eq(6, start->pT3->pToken->position.col);
+    assert_eq(7, start->pT3->pToken->end_position.row);
+    assert_eq(6, start->pT3->pToken->end_position.col);
+    assert_eq(7, start->pT3->position.row);
+    assert_eq(6, start->pT3->position.col);
+    assert_eq(7, start->pT3->end_position.row);
+    assert_eq(6, start->pT3->end_position.col);
+
+    assert_eq(3, start->position.row);
+    assert_eq(3, start->position.col);
+    assert_eq(7, start->end_position.row);
+    assert_eq(6, start->end_position.col);
+
+    p_free_ast(start);
+
+    return 0;
+}

spec/test_ast_token_positions.d

@@ -10,9 +10,10 @@ int main()
 unittest
 {
     string input = "abbccc";
-    p_context_t * context = p_context_new(input);
+    p_context_t context;
-    assert(p_parse(context) == P_SUCCESS);
+    p_context_init(&context, input);
-    Start start = p_result(context);
+    assert(p_parse(&context) == P_SUCCESS);
+    Start * start = p_result(&context);
 
     assert_eq(1, start.pT1.pToken.position.row);
     assert_eq(1, start.pT1.pToken.position.col);
@@ -46,12 +47,10 @@ unittest
     assert_eq(1, start.end_position.row);
     assert_eq(6, start.end_position.col);
 
-    p_context_delete(context);
-
     input = "\n\n  bb\nc\ncc\n\n     a";
-    context = p_context_new(input);
+    p_context_init(&context, input);
-    assert(p_parse(context) == P_SUCCESS);
+    assert(p_parse(&context) == P_SUCCESS);
-    start = p_result(context);
+    start = p_result(&context);
 
     assert_eq(3, start.pT1.pToken.position.row);
     assert_eq(3, start.pT1.pToken.position.col);
@@ -84,6 +83,4 @@ unittest
     assert_eq(3, start.position.col);
     assert_eq(7, start.end_position.row);
     assert_eq(6, start.end_position.col);
-
-    p_context_delete(context);
 }

spec/test_basic_math_grammar.c

@@ -5,29 +5,25 @@
 int main()
 {
     char const * input = "1 + 2 * 3 + 4";
-    p_context_t * context;
+    p_context_t context;
-    context = p_context_new((uint8_t const *)input, strlen(input));
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
-    assert_eq(P_SUCCESS, p_parse(context));
+    assert_eq(P_SUCCESS, p_parse(&context));
-    assert_eq(11, p_result(context));
+    assert_eq(11, p_result(&context));
-    p_context_delete(context);
 
     input = "1 * 2 ** 4 * 3";
-    context = p_context_new((uint8_t const *)input, strlen(input));
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
-    assert_eq(P_SUCCESS, p_parse(context));
+    assert_eq(P_SUCCESS, p_parse(&context));
-    assert_eq(48, p_result(context));
+    assert_eq(48, p_result(&context));
-    p_context_delete(context);
 
     input = "(1 + 2) * 3 + 4";
-    context = p_context_new((uint8_t const *)input, strlen(input));
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
-    assert_eq(P_SUCCESS, p_parse(context));
+    assert_eq(P_SUCCESS, p_parse(&context));
-    assert_eq(13, p_result(context));
+    assert_eq(13, p_result(&context));
-    p_context_delete(context);
 
     input = "(2 * 2) ** 3 + 4 + 5";
-    context = p_context_new((uint8_t const *)input, strlen(input));
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
-    assert_eq(P_SUCCESS, p_parse(context));
+    assert_eq(P_SUCCESS, p_parse(&context));
-    assert_eq(73, p_result(context));
+    assert_eq(73, p_result(&context));
-    p_context_delete(context);
 
     return 0;
 }

spec/test_basic_math_grammar.d

@@ -10,23 +10,23 @@ int main()
 unittest
 {
     string input = "1 + 2 * 3 + 4";
-    p_context_t * context;
+    p_context_t context;
-    context = p_context_new(input);
+    p_context_init(&context, input);
-    assert_eq(P_SUCCESS, p_parse(context));
+    assert_eq(P_SUCCESS, p_parse(&context));
-    assert_eq(11, p_result(context));
+    assert_eq(11, p_result(&context));
 
     input = "1 * 2 ** 4 * 3";
-    context = p_context_new(input);
+    p_context_init(&context, input);
-    assert_eq(P_SUCCESS, p_parse(context));
+    assert_eq(P_SUCCESS, p_parse(&context));
-    assert_eq(48, p_result(context));
+    assert_eq(48, p_result(&context));
 
     input = "(1 + 2) * 3 + 4";
-    context = p_context_new(input);
+    p_context_init(&context, input);
-    assert_eq(P_SUCCESS, p_parse(context));
+    assert_eq(P_SUCCESS, p_parse(&context));
-    assert_eq(13, p_result(context));
+    assert_eq(13, p_result(&context));
 
     input = "(2 * 2) ** 3 + 4 + 5";
-    context = p_context_new(input);
+    p_context_init(&context, input);
-    assert_eq(P_SUCCESS, p_parse(context));
+    assert_eq(P_SUCCESS, p_parse(&context));
-    assert_eq(73, p_result(context));
+    assert_eq(73, p_result(&context));
 }

spec/test_basic_math_grammar.rs

@@ -1,16 +0,0 @@
-use testparser::*;
-
-fn main() {
-    let cases: [(&[u8], u64); 4] = [
-        (b"1 + 2 * 3 + 4", 11),
-        (b"1 * 2 ** 4 * 3", 48),
-        (b"(1 + 2) * 3 + 4", 13),
-        (b"(2 * 2) ** 3 + 4 + 5", 73),
-    ];
-    for (input, expected) in cases {
-        let mut context = p_context_new(input);
-        assert_eq!(P_SUCCESS, p_parse(&mut context));
-        assert_eq!(expected, p_result(&context));
-        p_context_delete(context);
-    }
-}

spec/test_context_field_handle.c

@@ -1,39 +0,0 @@
-#include "testparser.h"
-#include <assert.h>
-#include <string.h>
-#include "testutils.h"
-
-int main()
-{
-    char input[128];
-    size_t i;
-    p_context_t * context;
-    Token token;
-
-    /* Enough tokens that the tree node arena is reallocated during the parse. */
-    memset(input, 0, sizeof(input));
-    for (i = 0u; i < 40u; i++)
-    {
-        input[i] = 'a';
-    }
-
-    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert_eq(P_SUCCESS, p_parse(context));
-
-    /* The handle was stored in a context user field during the parse, before
-     * the remaining nodes were created. It still refers to the same node. */
-    assert_eq(1u, context->have_first);
-    assert(p_node_valid(context->first_item));
-    token = p_Item_pToken1(context->first_item);
-    assert(p_node_valid(token));
-    assert_eq(TOKEN_a, p_Token_token(token));
-    assert_eq(7u, p_Token_pvalue(token));
-
-    /* The stored handle refers to the first Item, which starts at column 1. */
-    assert_eq(1u, p_node_position(context->first_item).row);
-    assert_eq(1u, p_node_position(context->first_item).col);
-
-    p_context_delete(context);
-
-    return 0;
-}

spec/test_context_field_handle.cpp

@@ -1,36 +0,0 @@
-#include "testparser.h"
-#include <cassert>
-#include <cstring>
-#include "testutils.h"
-
-int main()
-{
-    char input[128];
-
-    /* Enough tokens that the tree node arena is reallocated during the parse. */
-    memset(input, 0, sizeof(input));
-    for (size_t i = 0u; i < 40u; i++)
-    {
-        input[i] = 'a';
-    }
-
-    p_context_t * context = p_context_new((uint8_t const *)input, strlen(input));
-    assert_eq(P_SUCCESS, p_parse(context));
-
-    /* The handle was stored in a context user field during the parse, before
-     * the remaining nodes were created. It still refers to the same node. */
-    assert_eq(1u, context->have_first);
-    assert(context->first_item.valid());
-    Token token = context->first_item.pToken1();
-    assert(token.valid());
-    assert_eq(TOKEN_a, token.token());
-    assert_eq(7u, token.pvalue());
-
-    /* The stored handle refers to the first Item, which starts at column 1. */
-    assert_eq(1u, context->first_item.position().row);
-    assert_eq(1u, context->first_item.position().col);
-
-    p_context_delete(context);
-
-    return 0;
-}

spec/test_context_field_handle.d

@@ -1,35 +0,0 @@
-import testparser;
-import testutils;
-
-int main()
-{
-    return 0;
-}
-
-unittest
-{
-    /* Enough tokens that the tree node array is reallocated during the parse. */
-    string input;
-    foreach (i; 0 .. 40)
-    {
-        input ~= "a";
-    }
-
-    p_context_t * context = p_context_new(input);
-    assert_eq(P_SUCCESS, p_parse(context));
-
-    /* The handle was stored in a context user field during the parse, before
-     * the remaining nodes were created. It still refers to the same node. */
-    assert_eq(1, context.have_first);
-    assert(context.first_item.valid);
-    Token token = context.first_item.pToken1;
-    assert(token.valid);
-    assert_eq(TOKEN_a, token.token);
-    assert_eq(7, token.pvalue);
-
-    /* The stored handle refers to the first Item, which starts at column 1. */
-    assert_eq(1u, context.first_item.position.row);
-    assert_eq(1u, context.first_item.position.col);
-
-    p_context_delete(context);
-}

spec/test_custom_lex_fn.c

@@ -1,15 +0,0 @@
-#include "testparser.h"
-#include "testutils.h"
-#include <string.h>
-
-int main()
-{
-    char const * input = "cbacba";
-    p_context_t * context = p_context_new((uint8_t const *)input, strlen(input));
-    assert_eq(P_SUCCESS, p_parse(context));
-    size_t result = p_result(context);
-    assert_eq(0x932187932187, result);
-    p_context_delete(context);
-
-    return 0;
-}

spec/test_custom_lex_fn.d

@@ -1,18 +0,0 @@
-import testparser;
-import std.stdio;
-import testutils;
-
-int main()
-{
-    return 0;
-}
-
-unittest
-{
-    string input = "cbacba";
-    p_context_t * context = p_context_new(input);
-    assert_eq(P_SUCCESS, p_parse(context));
-    size_t result = p_result(context);
-    assert_eq(0x932187932187, result);
-    p_context_delete(context);
-}

spec/test_custom_lex_fn.rs

@@ -1,8 +0,0 @@
-use testparser::*;
-
-fn main() {
-    let mut c = p_context_new(b"cbacba");
-    assert_eq!(P_SUCCESS, p_parse(&mut c));
-    assert_eq!(0x932187932187, p_result(&c));
-    p_context_delete(c);
-}

spec/test_drop_code_block.c

@@ -1,15 +0,0 @@
-#include "testparser.h"
-#include <assert.h>
-#include <stdio.h>
-#include <string.h>
-
-int main()
-{
-    char const * input = "    # comment 1\n#    comment 2\na\n";
-    p_context_t * context;
-    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert(p_parse(context) == P_SUCCESS);
-    p_context_delete(context);
-
-    return 0;
-}

spec/test_drop_code_block.d

@@ -1,16 +0,0 @@
-import testparser;
-import std.stdio;
-import testutils;
-
-int main()
-{
-    return 0;
-}
-
-unittest
-{
-    string input = "    # comment 1\n#    comment 2\na\n";
-    p_context_t * context;
-    context = p_context_new(input);
-    assert(p_parse(context) == P_SUCCESS);
-}

spec/test_drop_code_block.rs

@@ -1,7 +0,0 @@
-use testparser::*;
-
-fn main() {
-    let mut c = p_context_new(b"    # comment 1\n#    comment 2\na\n");
-    assert_eq!(P_SUCCESS, p_parse(&mut c));
-    p_context_delete(c);
-}

spec/test_error_positions.c

@@ -5,43 +5,38 @@
 int main()
 {
     char const * input = "a 42";
-    p_context_t * context;
+    p_context_t context;
-    context = p_context_new((uint8_t const *)input, strlen(input));
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
-    assert(p_parse(context) == P_SUCCESS);
+    assert(p_parse(&context) == P_SUCCESS);
-    p_context_delete(context);
 
     input = "a\n123\na  a";
-    context = p_context_new((uint8_t const *)input, strlen(input));
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
-    assert(p_parse(context) == P_UNEXPECTED_TOKEN);
+    assert(p_parse(&context) == P_UNEXPECTED_TOKEN);
-    assert(p_position(context).row == 3);
+    assert(p_position(&context).row == 3);
-    assert(p_position(context).col == 4);
+    assert(p_position(&context).col == 4);
-    assert(p_token(context) == TOKEN_a);
+    assert(p_token(&context) == TOKEN_a);
-    p_context_delete(context);
 
     input = "12";
-    context = p_context_new((uint8_t const *)input, strlen(input));
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
-    assert(p_parse(context) == P_UNEXPECTED_TOKEN);
+    assert(p_parse(&context) == P_UNEXPECTED_TOKEN);
-    assert(p_position(context).row == 1);
+    assert(p_position(&context).row == 1);
-    assert(p_position(context).col == 1);
+    assert(p_position(&context).col == 1);
-    assert(p_token(context) == TOKEN_num);
+    assert(p_token(&context) == TOKEN_num);
-    p_context_delete(context);
 
     input = "a 12\n\nab";
-    context = p_context_new((uint8_t const *)input, strlen(input));
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
-    assert(p_parse(context) == P_UNEXPECTED_INPUT);
+    assert(p_parse(&context) == P_UNEXPECTED_INPUT);
-    assert(p_position(context).row == 3);
+    assert(p_position(&context).row == 3);
-    assert(p_position(context).col == 2);
+    assert(p_position(&context).col == 2);
-    p_context_delete(context);
 
     input = "a 12\n\na\n\n77\na   \xAA";
-    context = p_context_new((uint8_t const *)input, strlen(input));
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
-    assert(p_parse(context) == P_DECODE_ERROR);
+    assert(p_parse(&context) == P_DECODE_ERROR);
-    assert(p_position(context).row == 6);
+    assert(p_position(&context).row == 6);
-    assert(p_position(context).col == 5);
+    assert(p_position(&context).col == 5);
 
     assert(strcmp(p_token_names[TOKEN_a], "a") == 0);
     assert(strcmp(p_token_names[TOKEN_num], "num") == 0);
-    p_context_delete(context);
 
     return 0;
 }

spec/test_error_positions.d

@@ -9,31 +9,31 @@ int main()
 unittest
 {
     string input = "a 42";
-    p_context_t * context;
+    p_context_t context;
-    context = p_context_new(input);
+    p_context_init(&context, input);
-    assert(p_parse(context) == P_SUCCESS);
+    assert(p_parse(&context) == P_SUCCESS);
 
     input = "a\n123\na  a";
-    context = p_context_new(input);
+    p_context_init(&context, input);
-    assert(p_parse(context) == P_UNEXPECTED_TOKEN);
+    assert(p_parse(&context) == P_UNEXPECTED_TOKEN);
-    assert(p_position(context) == p_position_t(3, 4));
+    assert(p_position(&context) == p_position_t(3, 4));
-    assert(p_token(context) == TOKEN_a);
+    assert(p_token(&context) == TOKEN_a);
 
     input = "12";
-    context = p_context_new(input);
+    p_context_init(&context, input);
-    assert(p_parse(context) == P_UNEXPECTED_TOKEN);
+    assert(p_parse(&context) == P_UNEXPECTED_TOKEN);
-    assert(p_position(context) == p_position_t(1, 1));
+    assert(p_position(&context) == p_position_t(1, 1));
-    assert(p_token(context) == TOKEN_num);
+    assert(p_token(&context) == TOKEN_num);
 
     input = "a 12\n\nab";
-    context = p_context_new(input);
+    p_context_init(&context, input);
-    assert(p_parse(context) == P_UNEXPECTED_INPUT);
+    assert(p_parse(&context) == P_UNEXPECTED_INPUT);
-    assert(p_position(context) == p_position_t(3, 2));
+    assert(p_position(&context) == p_position_t(3, 2));
 
     input = "a 12\n\na\n\n77\na   \xAA";
-    context = p_context_new(input);
+    p_context_init(&context, input);
-    assert(p_parse(context) == P_DECODE_ERROR);
+    assert(p_parse(&context) == P_DECODE_ERROR);
-    assert(p_position(context) == p_position_t(6, 5));
+    assert(p_position(&context) == p_position_t(6, 5));
 
     assert(p_token_names[TOKEN_a] == "a");
     assert(p_token_names[TOKEN_num] == "num");

spec/test_error_positions.rs

@@ -1,35 +0,0 @@
-use testparser::*;
-
-fn main() {
-    let mut c = p_context_new(b"a 42");
-    assert_eq!(P_SUCCESS, p_parse(&mut c));
-    p_context_delete(c);
-
-    let mut c = p_context_new(b"a\n123\na  a");
-    assert_eq!(P_UNEXPECTED_TOKEN, p_parse(&mut c));
-    assert_eq!(3, p_position(&c).row);
-    assert_eq!(4, p_position(&c).col);
-    assert_eq!(TOKEN_a, p_token(&c));
-    p_context_delete(c);
-
-    let mut c = p_context_new(b"12");
-    assert_eq!(P_UNEXPECTED_TOKEN, p_parse(&mut c));
-    assert_eq!(1, p_position(&c).row);
-    assert_eq!(1, p_position(&c).col);
-    assert_eq!(TOKEN_num, p_token(&c));
-    p_context_delete(c);
-
-    let mut c = p_context_new(b"a 12\n\nab");
-    assert_eq!(P_UNEXPECTED_INPUT, p_parse(&mut c));
-    assert_eq!(3, p_position(&c).row);
-    assert_eq!(2, p_position(&c).col);
-    p_context_delete(c);
-
-    let mut c = p_context_new(b"a 12\n\na\n\n77\na   \xAA");
-    assert_eq!(P_DECODE_ERROR, p_parse(&mut c));
-    assert_eq!(6, p_position(&c).row);
-    assert_eq!(5, p_position(&c).col);
-    assert_eq!("a", p_token_names[TOKEN_a as usize]);
-    assert_eq!("num", p_token_names[TOKEN_num as usize]);
-    p_context_delete(c);
-}

spec/test_field_aliases.c

@@ -6,9 +6,8 @@
 int main()
 {
     char const * input = "foo1\nbar2";
-    p_context_t * context;
+    p_context_t context;
-    context = p_context_new((uint8_t const *)input, strlen(input));
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
-    assert(p_parse(context) == P_SUCCESS);
+    assert(p_parse(&context) == P_SUCCESS);
-    p_context_delete(context);
     return 0;
 }

spec/test_field_aliases.d

@@ -9,7 +9,7 @@ int main()
 unittest
 {
     string input = "foo1\nbar2";
-    p_context_t * context;
+    p_context_t context;
-    context = p_context_new(input);
+    p_context_init(&context, input);
-    assert(p_parse(context) == P_SUCCESS);
+    assert(p_parse(&context) == P_SUCCESS);
 }

spec/test_field_aliases.rs

@@ -1,7 +0,0 @@
-use testparser::*;
-
-fn main() {
-    let mut c = p_context_new(b"foo1\nbar2");
-    assert_eq!(P_SUCCESS, p_parse(&mut c));
-    p_context_delete(c);
-}

spec/test_free_ast_token_node_memory.c

@@ -0,0 +1,19 @@
+#include "testparser.h"
+#include <assert.h>
+#include <string.h>
+#include "testutils.h"
+
+int main()
+{
+    char const * input = "ab";
+    p_context_t context;
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert_eq(P_SUCCESS, p_parse(&context));
+    Start * start = p_result(&context);
+    assert(start->a != NULL);
+    assert(*start->a->pvalue == 1);
+    assert(start->b != NULL);
+    assert(*start->b->pvalue == 2);
+
+    p_free_ast(start);
+}

spec/test_input_index.c

@@ -1,60 +0,0 @@
-#include "testparser.h"
-#include <assert.h>
-#include <string.h>
-#include "testutils.h"
-
-int main()
-{
-    /* Grammar (simple):
-     *   drop /\\s+/;
-     *   token a; token b;
-     *   Start -> a b;
-     *
-     * Verifies that p_input_index() reports the parser/lexer's current byte
-     * offset into the input text. */
-
-    /* Fresh context: input_index starts at 0. */
-    {
-        char const * input = "ab";
-        p_context_t * context = p_context_new((uint8_t const *)input, strlen(input));
-        assert_eq(0u, p_input_index(context));
-        p_context_delete(context);
-    }
-
-    /* After each successful lex the byte offset advances past the token. */
-    {
-        char const * input = "a b";
-        p_context_t * context = p_context_new((uint8_t const *)input, strlen(input));
-        p_token_info_t token_info;
-        assert(p_lex(context, &token_info) == P_SUCCESS);
-        assert_eq((size_t)TOKEN_a, (size_t)token_info.token);
-        assert_eq(1u, p_input_index(context));
-        assert(p_lex(context, &token_info) == P_SUCCESS);
-        assert_eq((size_t)TOKEN_b, (size_t)token_info.token);
-        /* The dropped space between `a` and `b` advances input_index too. */
-        assert_eq(3u, p_input_index(context));
-        p_context_delete(context);
-    }
-
-    /* After a full successful parse, input_index has reached the end. */
-    {
-        char const * input = "ab";
-        p_context_t * context = p_context_new((uint8_t const *)input, strlen(input));
-        assert(p_parse_Start(context) == P_SUCCESS);
-        assert_eq(2u, p_input_index(context));
-        p_context_delete(context);
-    }
-
-    /* When parse_inner completes via a follow token, the follow token is not
-     * consumed, so input_index points at the start of the follow token. */
-    {
-        char const * input = "abb";
-        p_context_t * context = p_context_new((uint8_t const *)input, strlen(input));
-        p_token_t follow_tokens[] = { TOKEN_b };
-        assert(p_parse_inner_Start(context, follow_tokens, 1u) == P_SUCCESS);
-        assert_eq(2u, p_input_index(context));
-        p_context_delete(context);
-    }
-
-    return 0;
-}

spec/test_input_index.d

@@ -1,51 +0,0 @@
-import testparser;
-import std.stdio;
-import testutils;
-
-int main()
-{
-    return 0;
-}
-
-unittest
-{
-    /* See test_input_index.c for details on the grammar and cases. */
-
-    /* Fresh context: input_index starts at 0. */
-    {
-        string input = "ab";
-        p_context_t * context = p_context_new(input);
-        assert(p_input_index(context) == 0);
-    }
-
-    /* After each successful lex the byte offset advances past the token. */
-    {
-        string input = "a b";
-        p_context_t * context = p_context_new(input);
-        p_token_info_t token_info;
-        assert(p_lex(context, &token_info) == P_SUCCESS);
-        assert(token_info.token == TOKEN_a);
-        assert(p_input_index(context) == 1);
-        assert(p_lex(context, &token_info) == P_SUCCESS);
-        assert(token_info.token == TOKEN_b);
-        assert(p_input_index(context) == 3);
-    }
-
-    /* After a full successful parse, input_index has reached the end. */
-    {
-        string input = "ab";
-        p_context_t * context = p_context_new(input);
-        assert(p_parse_Start(context) == P_SUCCESS);
-        assert(p_input_index(context) == 2);
-    }
-
-    /* When parse_inner completes via a follow token, the follow token is not
-     * consumed, so input_index points at the start of the follow token. */
-    {
-        string input = "abb";
-        p_context_t * context = p_context_new(input);
-        p_token_t[] follow_tokens = [TOKEN_b];
-        assert(p_parse_inner_Start(context, follow_tokens) == P_SUCCESS);
-        assert(p_input_index(context) == 2);
-    }
-}

spec/test_input_index.rs

@@ -1,28 +0,0 @@
-use testparser::*;
-
-fn main() {
-    let c = p_context_new(b"ab");
-    assert_eq!(0, p_input_index(&c));
-    p_context_delete(c);
-
-    let mut c = p_context_new(b"a b");
-    let mut ti = p_token_info_t::default();
-    assert_eq!(P_SUCCESS, p_lex(&mut c, &mut ti));
-    assert_eq!(TOKEN_a, ti.token);
-    assert_eq!(1, p_input_index(&c));
-    assert_eq!(P_SUCCESS, p_lex(&mut c, &mut ti));
-    assert_eq!(TOKEN_b, ti.token);
-    assert_eq!(3, p_input_index(&c));
-    p_context_delete(c);
-
-    let mut c = p_context_new(b"ab");
-    assert_eq!(P_SUCCESS, p_parse_Start(&mut c));
-    assert_eq!(2, p_input_index(&c));
-    p_context_delete(c);
-
-    let mut c = p_context_new(b"abb");
-    let follow = [TOKEN_b];
-    assert_eq!(P_SUCCESS, p_parse_inner_Start(&mut c, &follow));
-    assert_eq!(2, p_input_index(&c));
-    p_context_delete(c);
-}

spec/test_lexer.c

@@ -38,75 +38,73 @@ int main()
 
     p_token_info_t token_info;
     char const * input = "5 + 4 * \n677 + 567";
-    p_context_t * context;
+    p_context_t context;
-    context = p_context_new((uint8_t const *)input, strlen(input));
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
-    assert(p_lex(context, &token_info) == P_SUCCESS);
+    assert(p_lex(&context, &token_info) == P_SUCCESS);
     assert(token_info.position.row == 1u);
     assert(token_info.position.col == 1u);
     assert(token_info.end_position.row == 1u);
     assert(token_info.end_position.col == 1u);
     assert(token_info.length == 1u);
     assert(token_info.token == TOKEN_int);
-    assert(p_lex(context, &token_info) == P_SUCCESS);
+    assert(p_lex(&context, &token_info) == P_SUCCESS);
     assert(token_info.position.row == 1u);
     assert(token_info.position.col == 3u);
     assert(token_info.end_position.row == 1u);
     assert(token_info.end_position.col == 3u);
     assert(token_info.length == 1u);
     assert(token_info.token == TOKEN_plus);
-    assert(p_lex(context, &token_info) == P_SUCCESS);
+    assert(p_lex(&context, &token_info) == P_SUCCESS);
     assert(token_info.position.row == 1u);
     assert(token_info.position.col == 5u);
     assert(token_info.end_position.row == 1u);
     assert(token_info.end_position.col == 5u);
     assert(token_info.length == 1u);
     assert(token_info.token == TOKEN_int);
-    assert(p_lex(context, &token_info) == P_SUCCESS);
+    assert(p_lex(&context, &token_info) == P_SUCCESS);
     assert(token_info.position.row == 1u);
     assert(token_info.position.col == 7u);
     assert(token_info.end_position.row == 1u);
     assert(token_info.end_position.col == 7u);
     assert(token_info.length == 1u);
     assert(token_info.token == TOKEN_times);
-    assert(p_lex(context, &token_info) == P_SUCCESS);
+    assert(p_lex(&context, &token_info) == P_SUCCESS);
     assert(token_info.position.row == 2u);
     assert(token_info.position.col == 1u);
     assert(token_info.end_position.row == 2u);
     assert(token_info.end_position.col == 3u);
     assert(token_info.length == 3u);
     assert(token_info.token == TOKEN_int);
-    assert(p_lex(context, &token_info) == P_SUCCESS);
+    assert(p_lex(&context, &token_info) == P_SUCCESS);
     assert(token_info.position.row == 2u);
     assert(token_info.position.col == 5u);
     assert(token_info.end_position.row == 2u);
     assert(token_info.end_position.col == 5u);
     assert(token_info.length == 1u);
     assert(token_info.token == TOKEN_plus);
-    assert(p_lex(context, &token_info) == P_SUCCESS);
+    assert(p_lex(&context, &token_info) == P_SUCCESS);
     assert(token_info.position.row == 2u);
     assert(token_info.position.col == 7u);
     assert(token_info.end_position.row == 2u);
     assert(token_info.end_position.col == 9u);
     assert(token_info.length == 3u);
     assert(token_info.token == TOKEN_int);
-    assert(p_lex(context, &token_info) == P_SUCCESS);
+    assert(p_lex(&context, &token_info) == P_SUCCESS);
     assert(token_info.position.row == 2u);
     assert(token_info.position.col == 10u);
     assert(token_info.end_position.row == 2u);
     assert(token_info.end_position.col == 10u);
     assert(token_info.length == 0u);
     assert(token_info.token == TOKEN___EOF);
-    p_context_delete(context);
 
-    context = p_context_new((uint8_t const *)"", 0u);
+    p_context_init(&context, (uint8_t const *)"", 0u);
-    assert(p_lex(context, &token_info) == P_SUCCESS);
+    assert(p_lex(&context, &token_info) == P_SUCCESS);
     assert(token_info.position.row == 1u);
     assert(token_info.position.col == 1u);
     assert(token_info.end_position.row == 1u);
     assert(token_info.end_position.col == 1u);
     assert(token_info.length == 0u);
     assert(token_info.token == TOKEN___EOF);
-    p_context_delete(context);
 
     return 0;
 }

spec/test_lexer.d

@@ -44,26 +44,26 @@ unittest
 {
     p_token_info_t token_info;
     string input = "5 + 4 * \n677 + 567";
-    p_context_t * context;
+    p_context_t context;
-    context = p_context_new(input);
+    p_context_init(&context, input);
-    assert(p_lex(context, &token_info) == P_SUCCESS);
+    assert(p_lex(&context, &token_info) == P_SUCCESS);
     assert(token_info == p_token_info_t(p_position_t(1, 1), p_position_t(1, 1), 1, TOKEN_int));
-    assert(p_lex(context, &token_info) == P_SUCCESS);
+    assert(p_lex(&context, &token_info) == P_SUCCESS);
     assert(token_info == p_token_info_t(p_position_t(1, 3), p_position_t(1, 3), 1, TOKEN_plus));
-    assert(p_lex(context, &token_info) == P_SUCCESS);
+    assert(p_lex(&context, &token_info) == P_SUCCESS);
     assert(token_info == p_token_info_t(p_position_t(1, 5), p_position_t(1, 5), 1, TOKEN_int));
-    assert(p_lex(context, &token_info) == P_SUCCESS);
+    assert(p_lex(&context, &token_info) == P_SUCCESS);
     assert(token_info == p_token_info_t(p_position_t(1, 7), p_position_t(1, 7), 1, TOKEN_times));
-    assert(p_lex(context, &token_info) == P_SUCCESS);
+    assert(p_lex(&context, &token_info) == P_SUCCESS);
     assert(token_info == p_token_info_t(p_position_t(2, 1), p_position_t(2, 3), 3, TOKEN_int));
-    assert(p_lex(context, &token_info) == P_SUCCESS);
+    assert(p_lex(&context, &token_info) == P_SUCCESS);
     assert(token_info == p_token_info_t(p_position_t(2, 5), p_position_t(2, 5), 1, TOKEN_plus));
-    assert(p_lex(context, &token_info) == P_SUCCESS);
+    assert(p_lex(&context, &token_info) == P_SUCCESS);
     assert(token_info == p_token_info_t(p_position_t(2, 7), p_position_t(2, 9), 3, TOKEN_int));
-    assert(p_lex(context, &token_info) == P_SUCCESS);
+    assert(p_lex(&context, &token_info) == P_SUCCESS);
     assert(token_info == p_token_info_t(p_position_t(2, 10), p_position_t(2, 10), 0, TOKEN___EOF));
 
-    context = p_context_new("");
+    p_context_init(&context, "");
-    assert(p_lex(context, &token_info) == P_SUCCESS);
+    assert(p_lex(&context, &token_info) == P_SUCCESS);
     assert(token_info == p_token_info_t(p_position_t(1, 1), p_position_t(1, 1), 0, TOKEN___EOF));
 }

spec/test_lexer.rs

@@ -1,49 +0,0 @@
-use testparser::*;
-
-fn chk(ti: &p_token_info_t, row: u32, col: u32, erow: u32, ecol: u32, len: usize, token: p_token_t) {
-    assert_eq!(row, ti.position.row);
-    assert_eq!(col, ti.position.col);
-    assert_eq!(erow, ti.end_position.row);
-    assert_eq!(ecol, ti.end_position.col);
-    assert_eq!(len, ti.length);
-    assert_eq!(token, ti.token);
-}
-
-fn main() {
-    let mut cp: p_code_point_t = 0;
-    let mut cpl: u8 = 0;
-
-    assert_eq!(P_SUCCESS, p_decode_code_point(b"5", &mut cp, &mut cpl));
-    assert_eq!('5' as u32, cp);
-    assert_eq!(1, cpl);
-
-    assert_eq!(P_EOF, p_decode_code_point(b"", &mut cp, &mut cpl));
-
-    assert_eq!(P_SUCCESS, p_decode_code_point(b"\xC2\xA9", &mut cp, &mut cpl));
-    assert_eq!(0xA9, cp);
-    assert_eq!(2, cpl);
-
-    assert_eq!(P_SUCCESS, p_decode_code_point(b"\xf0\x9f\xa7\xa1", &mut cp, &mut cpl));
-    assert_eq!(0x1F9E1, cp);
-    assert_eq!(4, cpl);
-
-    assert_eq!(P_DECODE_ERROR, p_decode_code_point(b"\xf0\x9f\x27", &mut cp, &mut cpl));
-    assert_eq!(P_DECODE_ERROR, p_decode_code_point(b"\xf0\x9f\xa7\xFF", &mut cp, &mut cpl));
-    assert_eq!(P_DECODE_ERROR, p_decode_code_point(b"\xfe", &mut cp, &mut cpl));
-
-    let mut context = p_context_new(b"5 + 4 * \n677 + 567");
-    let mut ti = p_token_info_t::default();
-    assert_eq!(P_SUCCESS, p_lex(&mut context, &mut ti)); chk(&ti, 1, 1, 1, 1, 1, TOKEN_int);
-    assert_eq!(P_SUCCESS, p_lex(&mut context, &mut ti)); chk(&ti, 1, 3, 1, 3, 1, TOKEN_plus);
-    assert_eq!(P_SUCCESS, p_lex(&mut context, &mut ti)); chk(&ti, 1, 5, 1, 5, 1, TOKEN_int);
-    assert_eq!(P_SUCCESS, p_lex(&mut context, &mut ti)); chk(&ti, 1, 7, 1, 7, 1, TOKEN_times);
-    assert_eq!(P_SUCCESS, p_lex(&mut context, &mut ti)); chk(&ti, 2, 1, 2, 3, 3, TOKEN_int);
-    assert_eq!(P_SUCCESS, p_lex(&mut context, &mut ti)); chk(&ti, 2, 5, 2, 5, 1, TOKEN_plus);
-    assert_eq!(P_SUCCESS, p_lex(&mut context, &mut ti)); chk(&ti, 2, 7, 2, 9, 3, TOKEN_int);
-    assert_eq!(P_SUCCESS, p_lex(&mut context, &mut ti)); chk(&ti, 2, 10, 2, 10, 0, TOKEN___EOF);
-    p_context_delete(context);
-
-    let mut context = p_context_new(b"");
-    assert_eq!(P_SUCCESS, p_lex(&mut context, &mut ti)); chk(&ti, 1, 1, 1, 1, 0, TOKEN___EOF);
-    p_context_delete(context);
-}

spec/test_lexer_match_text.c

@@ -6,11 +6,10 @@
 int main()
 {
     char const * input = "identifier_123";
-    p_context_t * context;
+    p_context_t context;
-    context = p_context_new((uint8_t const *)input, strlen(input));
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
-    assert(p_parse(context) == P_SUCCESS);
+    assert(p_parse(&context) == P_SUCCESS);
     printf("pass1\n");
-    p_context_delete(context);
 
     return 0;
 }

spec/test_lexer_match_text.d

@@ -9,8 +9,8 @@ int main()
 unittest
 {
     string input = `identifier_123`;
-    p_context_t * context;
+    p_context_t context;
-    context = p_context_new(input);
+    p_context_init(&context, input);
-    assert(p_parse(context) == P_SUCCESS);
+    assert(p_parse(&context) == P_SUCCESS);
     writeln("pass1");
 }

spec/test_lexer_match_text.rs

@@ -1,8 +0,0 @@
-use testparser::*;
-
-fn main() {
-    let mut context = p_context_new(b"identifier_123");
-    assert_eq!(P_SUCCESS, p_parse(&mut context));
-    println!("pass1");
-    p_context_delete(context);
-}

spec/test_lexer_modes.c

@@ -6,17 +6,15 @@
 int main()
 {
     char const * input = "abc \"a string\" def";
-    p_context_t * context;
+    p_context_t context;
-    context = p_context_new((uint8_t const *)input, strlen(input));
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
-    assert(p_parse(context) == P_SUCCESS);
+    assert(p_parse(&context) == P_SUCCESS);
     printf("pass1\n");
-    p_context_delete(context);
 
     input = "abc \"abc def\" def";
-    context = p_context_new((uint8_t const *)input, strlen(input));
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
-    assert(p_parse(context) == P_SUCCESS);
+    assert(p_parse(&context) == P_SUCCESS);
     printf("pass2\n");
-    p_context_delete(context);
 
     return 0;
 }

spec/test_lexer_modes.d

@@ -9,13 +9,13 @@ int main()
 unittest
 {
     string input = `abc "a string" def`;
-    p_context_t * context;
+    p_context_t context;
-    context = p_context_new(input);
+    p_context_init(&context, input);
-    assert(p_parse(context) == P_SUCCESS);
+    assert(p_parse(&context) == P_SUCCESS);
     writeln("pass1");
 
     input = `abc "abc def" def`;
-    context = p_context_new(input);
+    p_context_init(&context, input);
-    assert(p_parse(context) == P_SUCCESS);
+    assert(p_parse(&context) == P_SUCCESS);
     writeln("pass2");
 }

spec/test_lexer_modes.rs

@@ -1,13 +0,0 @@
-use testparser::*;
-
-fn main() {
-    let mut context = p_context_new(b"abc \"a string\" def");
-    assert_eq!(P_SUCCESS, p_parse(&mut context));
-    println!("pass1");
-    p_context_delete(context);
-
-    let mut context = p_context_new(b"abc \"abc def\" def");
-    assert_eq!(P_SUCCESS, p_parse(&mut context));
-    println!("pass2");
-    p_context_delete(context);
-}

spec/test_lexer_multiple_modes.c

@@ -6,17 +6,15 @@
 int main()
 {
     char const * input = "abc.def";
-    p_context_t * context;
+    p_context_t context;
-    context = p_context_new((uint8_t const *)input, strlen(input));
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
-    assert(p_parse(context) == P_SUCCESS);
+    assert(p_parse(&context) == P_SUCCESS);
     printf("pass1\n");
-    p_context_delete(context);
 
     input = "abc .  abc";
-    context = p_context_new((uint8_t const *)input, strlen(input));
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
-    assert(p_parse(context) == P_SUCCESS);
+    assert(p_parse(&context) == P_SUCCESS);
     printf("pass2\n");
-    p_context_delete(context);
 
     return 0;
 }

spec/test_lexer_multiple_modes.d

@@ -9,13 +9,13 @@ int main()
 unittest
 {
     string input = `abc.def`;
-    p_context_t * context;
+    p_context_t context;
-    context = p_context_new(input);
+    p_context_init(&context, input);
-    assert(p_parse(context) == P_SUCCESS);
+    assert(p_parse(&context) == P_SUCCESS);
     writeln("pass1");
 
     input = `abc .  abc`;
-    context = p_context_new(input);
+    p_context_init(&context, input);
-    assert(p_parse(context) == P_SUCCESS);
+    assert(p_parse(&context) == P_SUCCESS);
     writeln("pass2");
 }

spec/test_lexer_multiple_modes.rs

@@ -1,13 +0,0 @@
-use testparser::*;
-
-fn main() {
-    let mut context = p_context_new(b"abc.def");
-    assert_eq!(P_SUCCESS, p_parse(&mut context));
-    println!("pass1");
-    p_context_delete(context);
-
-    let mut context = p_context_new(b"abc .  abc");
-    assert_eq!(P_SUCCESS, p_parse(&mut context));
-    println!("pass2");
-    p_context_delete(context);
-}

spec/test_lexer_positions.c

@@ -1,50 +0,0 @@
-#include "testparser.h"
-#include <assert.h>
-#include <string.h>
-
-int main()
-{
-    char const * input = "abc\n  defg hi\n!";
-    p_context_t * context = p_context_new((uint8_t const *)input, strlen(input));
-    p_token_info_t token_info;
-
-    /* First token "abc" on row 1, cols 1-3. */
-    assert(p_lex(context, &token_info) == P_SUCCESS);
-    assert(token_info.token == TOKEN_word);
-    assert(context->last_start.row == 1u);
-    assert(context->last_start.col == 1u);
-    assert(context->last_end.row == 1u);
-    assert(context->last_end.col == 3u);
-    /* The lexer code block observed the same positions reported to the caller. */
-    assert(context->last_start.row == token_info.position.row);
-    assert(context->last_start.col == token_info.position.col);
-    assert(context->last_end.row == token_info.end_position.row);
-    assert(context->last_end.col == token_info.end_position.col);
-
-    /* Second token "defg" on row 2, cols 3-6. */
-    assert(p_lex(context, &token_info) == P_SUCCESS);
-    assert(token_info.token == TOKEN_word);
-    assert(context->last_start.row == 2u);
-    assert(context->last_start.col == 3u);
-    assert(context->last_end.row == 2u);
-    assert(context->last_end.col == 6u);
-
-    /* Third token "hi" on row 2, cols 8-9. */
-    assert(p_lex(context, &token_info) == P_SUCCESS);
-    assert(token_info.token == TOKEN_word);
-    assert(context->last_start.row == 2u);
-    assert(context->last_start.col == 8u);
-    assert(context->last_end.row == 2u);
-    assert(context->last_end.col == 9u);
-
-    /* The "!" stop token terminates the lexer. The context input text position
-     * must not be updated when the lexer user code requests termination, so it
-     * still points at the "!" token on row 3, col 1. */
-    assert(p_lex(context, &token_info) == P_USER_TERMINATED);
-    assert(p_user_terminate_code(context) == 42u);
-    assert(context->text_position.row == 3u);
-    assert(context->text_position.col == 1u);
-
-    p_context_delete(context);
-    return 0;
-}

spec/test_lexer_positions.d

@@ -1,42 +0,0 @@
-import testparser;
-import std.stdio;
-
-int main()
-{
-    return 0;
-}
-
-unittest
-{
-    string input = "abc\n  defg hi\n!";
-    p_context_t * context = p_context_new(input);
-    p_token_info_t token_info;
-
-    /* First token "abc" on row 1, cols 1-3. */
-    assert(p_lex(context, &token_info) == P_SUCCESS);
-    assert(token_info.token == TOKEN_word);
-    assert(context.last_start == p_position_t(1, 1));
-    assert(context.last_end == p_position_t(1, 3));
-    /* The lexer code block observed the same positions reported to the caller. */
-    assert(context.last_start == token_info.position);
-    assert(context.last_end == token_info.end_position);
-
-    /* Second token "defg" on row 2, cols 3-6. */
-    assert(p_lex(context, &token_info) == P_SUCCESS);
-    assert(token_info.token == TOKEN_word);
-    assert(context.last_start == p_position_t(2, 3));
-    assert(context.last_end == p_position_t(2, 6));
-
-    /* Third token "hi" on row 2, cols 8-9. */
-    assert(p_lex(context, &token_info) == P_SUCCESS);
-    assert(token_info.token == TOKEN_word);
-    assert(context.last_start == p_position_t(2, 8));
-    assert(context.last_end == p_position_t(2, 9));
-
-    /* The "!" stop token terminates the lexer. The context input text position
-     * must not be updated when the lexer user code requests termination, so it
-     * still points at the "!" token on row 3, col 1. */
-    assert(p_lex(context, &token_info) == P_USER_TERMINATED);
-    assert(p_user_terminate_code(context) == 42u);
-    assert(context.text_position == p_position_t(3, 1));
-}

spec/test_lexer_positions.rs

@@ -1,38 +0,0 @@
-use testparser::*;
-
-fn main() {
-    let mut c = p_context_new(b"abc\n  defg hi\n!");
-    let mut ti = p_token_info_t::default();
-
-    assert_eq!(P_SUCCESS, p_lex(&mut c, &mut ti));
-    assert_eq!(TOKEN_word, ti.token);
-    assert_eq!(1, c.last_start.row);
-    assert_eq!(1, c.last_start.col);
-    assert_eq!(1, c.last_end.row);
-    assert_eq!(3, c.last_end.col);
-    assert_eq!(c.last_start.row, ti.position.row);
-    assert_eq!(c.last_start.col, ti.position.col);
-    assert_eq!(c.last_end.row, ti.end_position.row);
-    assert_eq!(c.last_end.col, ti.end_position.col);
-
-    assert_eq!(P_SUCCESS, p_lex(&mut c, &mut ti));
-    assert_eq!(TOKEN_word, ti.token);
-    assert_eq!(2, c.last_start.row);
-    assert_eq!(3, c.last_start.col);
-    assert_eq!(2, c.last_end.row);
-    assert_eq!(6, c.last_end.col);
-
-    assert_eq!(P_SUCCESS, p_lex(&mut c, &mut ti));
-    assert_eq!(TOKEN_word, ti.token);
-    assert_eq!(2, c.last_start.row);
-    assert_eq!(8, c.last_start.col);
-    assert_eq!(2, c.last_end.row);
-    assert_eq!(9, c.last_end.col);
-
-    assert_eq!(P_USER_TERMINATED, p_lex(&mut c, &mut ti));
-    assert_eq!(42, p_user_terminate_code(&c));
-    assert_eq!(3, p_position(&c).row);
-    assert_eq!(1, p_position(&c).col);
-
-    p_context_delete(c);
-}

spec/test_lexer_result_value.c

@@ -5,17 +5,15 @@
 int main()
 {
     char const * input = "x";
-    p_context_t * context;
+    p_context_t context;
-    context = p_context_new((uint8_t const *)input, strlen(input));
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
-    assert(p_parse(context) == P_SUCCESS);
+    assert(p_parse(&context) == P_SUCCESS);
-    assert(p_result(context) == 1u);
+    assert(p_result(&context) == 1u);
-    p_context_delete(context);
 
     input = "fabulous";
-    context = p_context_new((uint8_t const *)input, strlen(input));
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
-    assert(p_parse(context) == P_SUCCESS);
+    assert(p_parse(&context) == P_SUCCESS);
-    assert(p_result(context) == 8u);
+    assert(p_result(&context) == 8u);
-    p_context_delete(context);
 
     return 0;
 }

spec/test_lexer_result_value.d

@@ -9,13 +9,13 @@ int main()
 unittest
 {
     string input = `x`;
-    p_context_t * context;
+    p_context_t context;
-    context = p_context_new(input);
+    p_context_init(&context, input);
-    assert(p_parse(context) == P_SUCCESS);
+    assert(p_parse(&context) == P_SUCCESS);
-    assert(p_result(context) == 1u);
+    assert(p_result(&context) == 1u);
 
     input = `fabulous`;
-    context = p_context_new(input);
+    p_context_init(&context, input);
-    assert(p_parse(context) == P_SUCCESS);
+    assert(p_parse(&context) == P_SUCCESS);
-    assert(p_result(context) == 8u);
+    assert(p_result(&context) == 8u);
 }

spec/test_lexer_result_value.rs

@@ -1,13 +0,0 @@
-use testparser::*;
-
-fn main() {
-    let mut context = p_context_new(b"x");
-    assert_eq!(P_SUCCESS, p_parse(&mut context));
-    assert_eq!(1, p_result(&context));
-    p_context_delete(context);
-
-    let mut context = p_context_new(b"fabulous");
-    assert_eq!(P_SUCCESS, p_parse(&mut context));
-    assert_eq!(8, p_result(&context));
-    p_context_delete(context);
-}

spec/test_lexer_unknown_character.c

@@ -5,16 +5,14 @@
 int main()
 {
     char const * input = "x";
-    p_context_t * context;
+    p_context_t context;
-    context = p_context_new((uint8_t const *)input, strlen(input));
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
-    assert(p_parse(context) == P_UNEXPECTED_INPUT);
+    assert(p_parse(&context) == P_UNEXPECTED_INPUT);
-    p_context_delete(context);
 
     input = "123";
-    context = p_context_new((uint8_t const *)input, strlen(input));
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
-    assert(p_parse(context) == P_SUCCESS);
+    assert(p_parse(&context) == P_SUCCESS);
-    assert(p_result(context) == 123u);
+    assert(p_result(&context) == 123u);
-    p_context_delete(context);
 
     return 0;
 }

spec/test_lexer_unknown_character.d

@@ -9,12 +9,12 @@ int main()
 unittest
 {
     string input = `x`;
-    p_context_t * context;
+    p_context_t context;
-    context = p_context_new(input);
+    p_context_init(&context, input);
-    assert(p_parse(context) == P_UNEXPECTED_INPUT);
+    assert(p_parse(&context) == P_UNEXPECTED_INPUT);
 
     input = `123`;
-    context = p_context_new(input);
+    p_context_init(&context, input);
-    assert(p_parse(context) == P_SUCCESS);
+    assert(p_parse(&context) == P_SUCCESS);
-    assert(p_result(context) == 123u);
+    assert(p_result(&context) == 123u);
 }

spec/test_lexer_unknown_character.rs

@@ -1,12 +0,0 @@
-use testparser::*;
-
-fn main() {
-    let mut context = p_context_new(b"x");
-    assert_eq!(P_UNEXPECTED_INPUT, p_parse(&mut context));
-    p_context_delete(context);
-
-    let mut context = p_context_new(b"123");
-    assert_eq!(P_SUCCESS, p_parse(&mut context));
-    assert_eq!(123, p_result(&context));
-    p_context_delete(context);
-}

spec/test_macros.c

@@ -1,118 +0,0 @@
-#include "testparser.h"
-#include "testutils.h"
-#include <string.h>
-#include <assert.h>
-#include <stddef.h>
-#include <stdbool.h>
-
-static p_context_t * context;
-size_t n_tokens;
-p_token_info_t token_infos[10];
-
-/* Capture the macro body tokens (everything up to the closing '}') into
- * token_infos[]. Called from mylexfn() right after the definition's '{' has
- * been lexed, so the input cursor is positioned at the first body token. */
-static void capture_macro_body(void)
-{
-    n_tokens = 0u;
-    for (;;)
-    {
-        size_t result = p_lex(context, &token_infos[n_tokens]);
-        assert_eq(result, P_SUCCESS);
-        if (token_infos[n_tokens].token == TOKEN_rbrace)
-        {
-            break;
-        }
-        n_tokens++;
-        assert(n_tokens < sizeof(token_infos) / sizeof(token_infos[0]));
-    }
-}
-
-size_t mylexfn(p_context_t * context, p_token_info_t * out_token_info)
-{
-    static bool defining;
-    static bool expanding;
-    static size_t expand_i;
-
-    for (;;)
-    {
-        if (expanding)
-        {
-            size_t ei = expand_i++;
-            if (expand_i >= n_tokens)
-            {
-                expanding = false;
-            }
-            *out_token_info = token_infos[ei];
-            return P_SUCCESS;
-        }
-
-        size_t lex_result = p_lex(context, out_token_info);
-        if (lex_result != P_SUCCESS)
-        {
-            return lex_result;
-        }
-
-        switch (out_token_info->token)
-        {
-        case TOKEN_macro:
-            /* Start of a macro definition: "macro macroname { ... }". */
-            defining = true;
-            break;
-        case TOKEN_macroname:
-            if (!defining)
-            {
-                /* Use of a macro: replay its captured body tokens instead of
-                 * returning the macroname to the parser. */
-                expanding = true;
-                expand_i = 0u;
-                continue;
-            }
-            /* Definition name: pass through and keep waiting for '{'. */
-            break;
-        case TOKEN_lbrace:
-            if (defining)
-            {
-                /* Consume and store the macro body now, before the parser gets
-                 * a chance to read its lookahead token (which would otherwise
-                 * swallow the first body token). */
-                capture_macro_body();
-                defining = false;
-            }
-            break;
-        default:
-            defining = false;
-            break;
-        }
-        return lex_result;
-    }
-}
-
-size_t n_nums;
-int nums[10];
-
-void record(int v)
-{
-    nums[n_nums++] = v;
-}
-
-int main()
-{
-    char const * input =
-        "macro @m { 23 + 200 }\n"
-        "66 + 100\n"
-        "@m\n"
-        "33 + 55\n"
-        "@m\n";
-    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert(p_parse(context) == P_SUCCESS);
-    p_context_delete(context);
-
-    assert_eq(n_nums, 4);
-    assert_eq(nums[0], 166);
-    assert_eq(nums[1], 223);
-    assert_eq(nums[2], 88);
-    assert_eq(nums[3], 223);
-
-    return 0;
-}

spec/test_macros.d

@@ -1,116 +0,0 @@
-import testparser;
-import testutils;
-
-size_t n_tokens;
-p_token_info_t[10] token_infos;
-
-// Capture the macro body tokens (everything up to the closing '}') into
-// token_infos[]. Called from mylexfn() right after the definition's '{' has
-// been lexed, so the input cursor is positioned at the first body token.
-void capture_macro_body(p_context_t * context)
-{
-    n_tokens = 0u;
-    for (;;)
-    {
-        size_t result = p_lex(context, &token_infos[n_tokens]);
-        assert(result == P_SUCCESS);
-        if (token_infos[n_tokens].token == TOKEN_rbrace)
-        {
-            break;
-        }
-        n_tokens++;
-        assert(n_tokens < token_infos.length);
-    }
-}
-
-size_t mylexfn(p_context_t * context, p_token_info_t * out_token_info)
-{
-    static bool defining;
-    static bool expanding;
-    static size_t expand_i;
-
-    for (;;)
-    {
-        if (expanding)
-        {
-            size_t ei = expand_i++;
-            if (expand_i >= n_tokens)
-            {
-                expanding = false;
-            }
-            *out_token_info = token_infos[ei];
-            return P_SUCCESS;
-        }
-
-        size_t lex_result = p_lex(context, out_token_info);
-        if (lex_result != P_SUCCESS)
-        {
-            return lex_result;
-        }
-
-        switch (out_token_info.token)
-        {
-        case TOKEN_macro:
-            // Start of a macro definition: "macro macroname { ... }".
-            defining = true;
-            break;
-        case TOKEN_macroname:
-            if (!defining)
-            {
-                // Use of a macro: replay its captured body tokens instead of
-                // returning the macroname to the parser.
-                expanding = true;
-                expand_i = 0u;
-                continue;
-            }
-            // Definition name: pass through and keep waiting for '{'.
-            break;
-        case TOKEN_lbrace:
-            if (defining)
-            {
-                // Consume and store the macro body now, before the parser gets
-                // a chance to read its lookahead token (which would otherwise
-                // swallow the first body token).
-                capture_macro_body(context);
-                defining = false;
-            }
-            break;
-        default:
-            defining = false;
-            break;
-        }
-        return lex_result;
-    }
-}
-
-size_t n_nums;
-int[10] nums;
-
-void record(int v)
-{
-    nums[n_nums++] = v;
-}
-
-int main()
-{
-    return 0;
-}
-
-unittest
-{
-    string input =
-        "macro @m { 23 + 200 }\n" ~
-        "66 + 100\n" ~
-        "@m\n" ~
-        "33 + 55\n" ~
-        "@m\n";
-    p_context_t * context = p_context_new(input);
-    assert(p_parse(context) == P_SUCCESS);
-    p_context_delete(context);
-
-    assert(n_nums == 4);
-    assert(nums[0] == 166);
-    assert(nums[1] == 223);
-    assert(nums[2] == 88);
-    assert(nums[3] == 223);
-}

spec/test_macros.rs

@@ -1,9 +0,0 @@
-use testparser::*;
-
-fn main() {
-    let input = b"macro @m { 23 + 200 }\n66 + 100\n@m\n33 + 55\n@m\n";
-    let mut c = p_context_new(input);
-    assert_eq!(P_SUCCESS, p_parse(&mut c));
-    assert_eq!(vec![166, 223, 88, 223], c.nums);
-    p_context_delete(c);
-}

spec/test_match_backslashes.c

@@ -5,10 +5,9 @@
 int main()
 {
     char const * input = "\a\b\t\n\v\f\rt";
-    p_context_t * context;
+    p_context_t context;
-    context = p_context_new((uint8_t const *)input, strlen(input));
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
-    assert(p_parse(context) == P_SUCCESS);
+    assert(p_parse(&context) == P_SUCCESS);
-    p_context_delete(context);
 
     return 0;
 }

spec/test_match_backslashes.d

@@ -9,7 +9,7 @@ int main()
 unittest
 {
     string input = "\a\b\t\n\v\f\rt";
-    p_context_t * context;
+    p_context_t context;
-    context = p_context_new(input);
+    p_context_init(&context, input);
-    assert(p_parse(context) == P_SUCCESS);
+    assert(p_parse(&context) == P_SUCCESS);
 }