v4.7.0

Allow user code to access specific pvalue values, especially when
multiple ptypes are in use.

CHANGELOG.md

@@ -1,3 +1,97 @@
+## 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 (6.0.1)
+    erb (6.0.4)
-    psych (5.3.1)
+    psych (5.4.0)
       date
       stringio
-    rake (13.3.1)
+    rake (13.4.2)
-    rdoc (7.1.0)
+    rdoc (7.2.0)
       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 (3.13.7)
+    rspec-mocks (3.13.8)
       diff-lcs (>= 1.2.0, < 2.0)
       rspec-support (~> 3.13.0)
-    rspec-support (3.13.6)
+    rspec-support (3.13.7)
     simplecov (0.22.0)
       docile (~> 1.1)
       simplecov-html (~> 0.11)
@@ -51,4 +51,4 @@ DEPENDENCIES
   syntax
 
 BUNDLED WITH
-   2.3.7
+  4.0.14

LICENSE.txt

@@ -1,6 +1,6 @@
 The MIT License (MIT)
 
-Copyright (c) 2010-2024 Josh Holtrop
+Copyright (c) 2010-2026 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

@@ -7,7 +7,9 @@ Propane is a LALR Parser Generator (LPG) which:
   * supports UTF-8 lexer inputs
   * generates a table-driven shift/reduce parser to parse input in linear time
   * targets C, C++, or D language outputs
-  * optionally supports automatic full AST generation
+  * optionally supports automatic full parse tree 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
 

UPGRADING.md

@@ -0,0 +1,21 @@
+## 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

@@ -43,30 +43,52 @@ const char * <%= @grammar.prefix %>token_names[] = {
  *************************************************************************/
 
 /**
- * Initialize lexer/parser context structure.
+ * Allocate and 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.
  */
-void <%= @grammar.prefix %>context_init(<%= @grammar.prefix %>context_t * context, uint8_t const * input, size_t input_length)
+<%= @grammar.prefix %>context_t * <%= @grammar.prefix %>context_new(uint8_t const * input, size_t input_length)
 {
-    /* New default-initialized context structure. */
+<% if @cpp %>
-    <%= @grammar.prefix %>context_t newcontext;
+    <%= @grammar.prefix %>context_t * context = new <%= @grammar.prefix %>context_t();
-    memset(&newcontext, 0, sizeof(newcontext));
+<% else %>
+    <%= @grammar.prefix %>context_t * context = (<%= @grammar.prefix %>context_t *)calloc(1, sizeof(<%= @grammar.prefix %>context_t));
+<% end %>
 
     /* Lexer initialization. */
-    newcontext.input = input;
+    context->input = input;
-    newcontext.input_length = input_length;
+    context->input_length = input_length;
-    newcontext.text_position.row = 1u;
+    context->text_position.row = 1u;
-    newcontext.text_position.col = 1u;
+    context->text_position.col = 1u;
-    newcontext.mode = <%= @lexer.mode_id("default") %>;
+    context->mode = <%= @lexer.mode_id("default") %>;
 
-    /* Copy to the user's context structure. */
+    return context;
-    *context = newcontext;
+}
+
+/**
+ * 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 @cpp %>
+    delete context;
+<% else %>
+    free(context);
+<% end %>
 }
 
 /**************************************************************************
@@ -464,6 +486,22 @@ 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 = &context->input[context->input_index];
@@ -502,17 +540,6 @@ 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;
@@ -638,7 +665,7 @@ typedef struct
      * reduce action.
      */
     parser_state_id_t n_states;
-<% if @grammar.ast %>
+<% if @grammar.tree %>
 
     /**
      * Map of rule components to rule set child fields.
@@ -646,7 +673,7 @@ typedef struct
     uint16_t const * rule_set_node_field_index_map;
 
     /**
-     * Number of rule set AST node fields.
+     * Number of rule set tree node fields.
      */
     uint16_t rule_set_node_field_array_size;
 
@@ -685,24 +712,28 @@ typedef struct
     /** Parser state ID. */
     size_t state_id;
 
+<% if @grammar.tree %>
+    /** tree node. */
+    void * tree_node;
+<% 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. */
+<% if @grammar.tree %>
-typedef struct ASTNode_s
+/** Common tree node structure. */
+typedef struct TreeNode_s
 {
     <%= @grammar.prefix %>position_t position;
     <%= @grammar.prefix %>position_t end_position;
     uint16_t n_fields;
     uint8_t is_token;
-    struct ASTNode_s * fields[];
+    struct TreeNode_s * fields[];
-} ASTNode;
+} TreeNode;
+<% end %>
 
 /** Parser shift table. */
 static const shift_t parser_shift_table[] = {
@@ -711,7 +742,7 @@ static const shift_t parser_shift_table[] = {
 <%   end %>
 };
 
-<% if @grammar.ast %>
+<% if @grammar.tree %>
 <%   @grammar.rules.each do |rule| %>
 <%     unless rule.flat_rule_set_node_field_index_map? %>
 const uint16_t r_<%= rule.name.gsub("$", "_") %><%= rule.id %>_node_field_index_map[<%= rule.rule_set_node_field_index_map.size %>] = {<%= rule.rule_set_node_field_index_map.map {|v| v.to_s}.join(", ") %>};
@@ -726,14 +757,14 @@ static const reduce_t parser_reduce_table[] = {
         <%= reduce[:token_id] %>u, /* Token: <%= reduce[:token] ? reduce[:token].name : "(any)" %> */
         <%= reduce[:rule_id] %>u, /* Rule ID */
         <%= reduce[:rule_set_id] %>u, /* Rule set ID (<%= reduce[:rule].rule_set.name %>) */
-<%     if @grammar.ast %>
+<%     if @grammar.tree %>
         <%= reduce[:n_states] %>u, /* Number of states */
 <%       if reduce[:rule].flat_rule_set_node_field_index_map? %>
         NULL, /* No rule set node field index map (flat map) */
 <%       else %>
         &r_<%= reduce[:rule].name.gsub("$", "_") %><%= reduce[:rule].id %>_node_field_index_map[0], /* Rule set node field index map */
 <%       end %>
-        <%= reduce[:rule].rule_set.ast_fields.size %>, /* Number of AST fields */
+        <%= reduce[:rule].rule_set.tree_fields.size %>, /* Number of tree fields */
         <%= reduce[:propagate_optional_target] %>}, /* Propagate optional target? */
 <%     else %>
         <%= reduce[:n_states] %>u},
@@ -841,7 +872,60 @@ static void state_values_stack_free(state_values_stack_t * stack)
      free(stack->entries);
 }
 
-<% unless @grammar.ast %>
+<% 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.
  *
@@ -852,7 +936,7 @@ static void state_values_stack_free(state_values_stack_t * stack)
  * @retval P_USER_TERMINATED
  *   User requested to terminate parsing.
  */
-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)
+static size_t parser_user_code(<%= @grammar.tree ? "void" : "#{@grammar.prefix}value_t" %> * _pvalue, uint32_t rule, state_values_stack_t * statevalues, uint32_t n_states, <%= @grammar.prefix %>context_t * context)
 {
     switch (rule)
     {
@@ -944,9 +1028,11 @@ static size_t parse_from(<%= @grammar.prefix %>context_t * context, size_t start
     <%= @grammar.prefix %>token_t token = INVALID_TOKEN_ID;
     state_values_stack_t statevalues;
     size_t reduced_rule_set = INVALID_ID;
-<% if @grammar.ast %>
+<% if @grammar.tree %>
     void * 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);
@@ -957,7 +1043,7 @@ static size_t parse_from(<%= @grammar.prefix %>context_t * context, size_t start
     {
         if (token == INVALID_TOKEN_ID)
         {
-            size_t lexer_result = <%= @grammar.prefix %>lex(context, &token_info);
+            size_t lexer_result = <%= lex_fn %>(context, &token_info);
             if (lexer_result != P_SUCCESS)
             {
                 result = lexer_result;
@@ -976,8 +1062,8 @@ 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.ast %>
+<% if @grammar.tree %>
-                context->parse_result = state_values_stack_index(&statevalues, -1)->ast_node;
+                context->parse_result = state_values_stack_index(&statevalues, -1)->tree_node;
 <% else %>
                 context->parse_result = state_values_stack_index(&statevalues, -1)->pvalue;
 <% end %>
@@ -989,31 +1075,41 @@ static size_t parse_from(<%= @grammar.prefix %>context_t * context, size_t start
         {
             /* We have something to shift. */
             state_values_stack_push(&statevalues);
-            state_values_stack_index(&statevalues, -1)->state_id = shift_state;
+            state_value_t * new_state_info = state_values_stack_index(&statevalues, -1);
+            new_state_info->state_id = shift_state;
             if (reduced_rule_set == INVALID_ID)
             {
                 /* We shifted a token, mark it consumed. */
-<% if @grammar.ast %>
+<% if @grammar.tree %>
-                <%= @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 %>));
+<%   if @cpp %>
-                token_ast_node->position = token_info.position;
+                <%= @grammar.tree_prefix %>Token<%= @grammar.tree_suffix %> * token_tree_node = new <%= @grammar.tree_prefix %>Token<%= @grammar.tree_suffix %>();
-                token_ast_node->end_position = token_info.end_position;
-                token_ast_node->n_fields = 0u;
-                token_ast_node->is_token = 1u;
-                token_ast_node->token = token;
-                token_ast_node->pvalue = token_info.pvalue;
-                state_values_stack_index(&statevalues, -1)->ast_node = token_ast_node;
 <%   else %>
-                state_values_stack_index(&statevalues, -1)->pvalue = token_info.pvalue;
+                <%= @grammar.tree_prefix %>Token<%= @grammar.tree_suffix %> * token_tree_node = (<%= @grammar.tree_prefix %>Token<%= @grammar.tree_suffix %> *)malloc(sizeof(<%= @grammar.tree_prefix %>Token<%= @grammar.tree_suffix %>));
+<%   end %>
+                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 = 1u;
+                token_tree_node->token = token;
+                token_tree_node->pvalue = token_info.pvalue;
+<%= expand_code(@grammar.on_token_node, false, nil, nil) %>
+                new_state_info->tree_node = token_tree_node;
+<% else %>
+                new_state_info->position = token_info.position;
+                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.ast %>
+<% if @grammar.tree %>
-                state_values_stack_index(&statevalues, -1)->ast_node = reduced_parser_node;
+                new_state_info->tree_node = reduced_parser_node;
 <% else %>
-                state_values_stack_index(&statevalues, -1)->pvalue = reduced_parser_value;
+                new_state_info->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;
@@ -1027,16 +1123,16 @@ static size_t parse_from(<%= @grammar.prefix %>context_t * context, size_t start
         if (reduce_index != INVALID_ID)
         {
             /* We have something to reduce. */
-<% if @grammar.ast %>
+<% if @grammar.tree %>
             if (parser_reduce_table[reduce_index].propagate_optional_target)
             {
-                reduced_parser_node = state_values_stack_index(&statevalues, -1)->ast_node;
+                reduced_parser_node = state_values_stack_index(&statevalues, -1)->tree_node;
             }
             else if (parser_reduce_table[reduce_index].n_states > 0)
             {
                 size_t n_fields = parser_reduce_table[reduce_index].rule_set_node_field_array_size;
-                size_t bytes = sizeof(ASTNode) + n_fields * sizeof(void *);
+                size_t bytes = sizeof(TreeNode) + n_fields * sizeof(void *);
-                ASTNode * node = (ASTNode *)malloc(bytes);
+                TreeNode * node = (TreeNode *)malloc(bytes);
                 memset(node, 0, bytes);
                 node->position = INVALID_POSITION;
                 node->end_position = INVALID_POSITION;
@@ -1045,20 +1141,20 @@ static size_t parse_from(<%= @grammar.prefix %>context_t * context, size_t start
                 {
                     for (size_t i = 0; i < parser_reduce_table[reduce_index].n_states; i++)
                     {
-                        node->fields[i] = (ASTNode *)state_values_stack_index(&statevalues, -(int)parser_reduce_table[reduce_index].n_states + (int)i)->ast_node;
+                        node->fields[i] = (TreeNode *)state_values_stack_index(&statevalues, -(int)parser_reduce_table[reduce_index].n_states + (int)i)->tree_node;
                     }
                 }
                 else
                 {
                     for (size_t i = 0; i < parser_reduce_table[reduce_index].n_states; i++)
                     {
-                        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;
+                        node->fields[parser_reduce_table[reduce_index].rule_set_node_field_index_map[i]] = (TreeNode *)state_values_stack_index(&statevalues, -(int)parser_reduce_table[reduce_index].n_states + (int)i)->tree_node;
                     }
                 }
                 bool position_found = false;
                 for (size_t i = 0; i < n_fields; i++)
                 {
-                    ASTNode * child = node->fields[i];
+                    TreeNode * child = node->fields[i];
                     if ((child != NULL) && <%= @grammar.prefix %>position_valid(child->position))
                     {
                         if (!position_found)
@@ -1075,6 +1171,13 @@ static size_t parse_from(<%= @grammar.prefix %>context_t * context, size_t start
             {
                 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));
@@ -1084,6 +1187,16 @@ 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);
@@ -1125,15 +1238,15 @@ size_t <%= @grammar.prefix %>parse_<%= start_rule %>(<%= @grammar.prefix %>conte
  *
  * @return Parse result value.
  */
-<% if @grammar.ast %>
+<% if @grammar.tree %>
-<%= @grammar.ast_prefix %><%= @grammar.start_rules[0] %><%= @grammar.ast_suffix %> * <%= @grammar.prefix %>result(<%= @grammar.prefix %>context_t * context)
+<%= @grammar.tree_prefix %><%= @grammar.start_rules[0] %><%= @grammar.tree_suffix %> * <%= @grammar.prefix %>result(<%= @grammar.prefix %>context_t * context)
 {
-    return (<%= @grammar.ast_prefix %><%= @grammar.start_rules[0] %><%= @grammar.ast_suffix %> *) context->parse_result;
+    return (<%= @grammar.tree_prefix %><%= @grammar.start_rules[0] %><%= @grammar.tree_suffix %> *) context->parse_result;
 }
 <%   @grammar.start_rules.each_with_index do |start_rule, i| %>
-<%= @grammar.ast_prefix %><%= start_rule %><%= @grammar.ast_suffix %> * <%= @grammar.prefix %>result_<%= start_rule %>(<%= @grammar.prefix %>context_t * context)
+<%= @grammar.tree_prefix %><%= start_rule %><%= @grammar.tree_suffix %> * <%= @grammar.prefix %>result_<%= start_rule %>(<%= @grammar.prefix %>context_t * context)
 {
-    return (<%= @grammar.ast_prefix %><%= start_rule %><%= @grammar.ast_suffix %> *) context->parse_result;
+    return (<%= @grammar.tree_prefix %><%= start_rule %><%= @grammar.tree_suffix %> *) context->parse_result;
 }
 <%   end %>
 <% else %>
@@ -1184,16 +1297,19 @@ size_t <%= @grammar.prefix %>user_terminate_code(<%= @grammar.prefix %>context_t
 {
     return context->token;
 }
-<% if @grammar.ast %>
+<% if @grammar.tree %>
 
-static void free_ast_node(ASTNode * node)
+static void tree_delete(TreeNode * node)
 {
     if (node->is_token)
     {
-<%   if @grammar.free_token_node %>
+        <%= @grammar.tree_prefix %>Token<%= @grammar.tree_suffix %> * token_tree_node = (<%= @grammar.tree_prefix %>Token<%= @grammar.tree_suffix %> *)node;
-        <%= @grammar.free_token_node %>((<%= @grammar.ast_prefix %>Token<%= @grammar.ast_suffix %> *) node);
+<%= expand_code(@grammar.free_token_node, false, nil, nil) %>
+<%   if @cpp %>
+        delete token_tree_node;
+<%   else %>
+        free(token_tree_node);
 <%   end %>
-        /* TODO: free value_t */
     }
     else if (node->n_fields > 0u)
     {
@@ -1201,28 +1317,28 @@ static void free_ast_node(ASTNode * node)
         {
             if (node->fields[i] != NULL)
             {
-                free_ast_node(node->fields[i]);
+                tree_delete(node->fields[i]);
-            }
             }
         }
         free(node);
     }
+}
 
 /**
- * Free all AST node memory.
+ * Free all tree node memory.
  */
-void <%= @grammar.prefix %>free_ast(<%= @grammar.ast_prefix %><%= @grammar.start_rules[0] %><%= @grammar.ast_suffix %> * ast)
+void <%= @grammar.prefix %>tree_delete(<%= @grammar.tree_prefix %><%= @grammar.start_rules[0] %><%= @grammar.tree_suffix %> * tree)
 {
-    free_ast_node((ASTNode *)ast);
+    tree_delete((TreeNode *)tree);
 }
 <%   @grammar.start_rules.each_with_index do |start_rule, i| %>
 
 /**
- * Free all AST node memory.
+ * Free all tree node memory.
  */
-void <%= @grammar.prefix %>free_ast_<%= start_rule %>(<%= @grammar.ast_prefix %><%= start_rule %><%= @grammar.ast_suffix %> * ast)
+void <%= @grammar.prefix %>tree_delete_<%= start_rule %>(<%= @grammar.tree_prefix %><%= start_rule %><%= @grammar.tree_suffix %> * tree)
 {
-    free_ast_node((ASTNode *)ast);
+    tree_delete((TreeNode *)tree);
 }
 <%   end %>
 <% end %>

assets/parser.d.erb

@@ -9,7 +9,7 @@ module <%= @grammar.modulename %>;
 <% end %>
 
 import core.memory;
-import core.stdc.stdlib : malloc;
+import core.stdc.stdlib : malloc, free;
 
 /**************************************************************************
  * User code blocks
@@ -59,10 +59,10 @@ public alias <%= @grammar.prefix %>code_point_t = uint;
  */
 public struct <%= @grammar.prefix %>position_t
 {
-    /** Input text row (0-based). */
+    /** Input text row (1-based). */
     uint row;
 
-    /** Input text column (0-based). */
+    /** Input text column (1-based). */
     uint col;
 
     /** Invalid position value. */
@@ -75,7 +75,7 @@ public struct <%= @grammar.prefix %>position_t
     }
 }
 
-<% if @grammar.ast %>
+<% if @grammar.tree %>
 /** Parser values type. */
 public alias <%= @grammar.prefix %>value_t = <%= @grammar.ptype %>;
 <% else %>
@@ -86,35 +86,58 @@ 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 %>
 
-<% if @grammar.ast %>
+/** Parser value accessor(s). */
-/** Common AST node structure. */
+<%   @grammar.ptypes.each do |name, typestring| %>
-private struct ASTNode
+public <%= typestring %> <%= @grammar.prefix %>value_get<%= name == "default" ? "" : "_#{name}" %>(<%= @grammar.prefix %>value_t * pvalue)
+{
+    return pvalue.v_<%= name %>;
+}
+<%   end %>
+<% end %>
+
+<% if @grammar.tree %>
+/** Common tree node structure. */
+private struct TreeNode
 {
     <%= @grammar.prefix %>position_t position;
     <%= @grammar.prefix %>position_t end_position;
+    ushort n_fields;
+    bool is_token;
     void *[0] fields;
 }
 
-/** AST node types. @{ */
+/** Tree node types. @{ */
-public struct <%= @grammar.ast_prefix %>Token<%= @grammar.ast_suffix %>
+public struct <%= @grammar.tree_prefix %>Token<%= @grammar.tree_suffix %>
 {
-    /* ASTNode fields must be present in the same order here. */
+    /* TreeNode fields must be present in the same order here. */
     <%= @grammar.prefix %>position_t position;
     <%= @grammar.prefix %>position_t end_position;
+    ushort n_fields;
+    bool is_token;
     <%= @grammar.prefix %>token_t token;
     <%= @grammar.prefix %>value_t pvalue;
+<%= @grammar.token_user_fields %>
 }
 
 <%   @parser.rule_sets.each do |name, rule_set| %>
 <%     next if name.start_with?("$") %>
 <%     next if rule_set.optional? %>
-public struct <%= @grammar.ast_prefix %><%= name %><%= @grammar.ast_suffix %>
+public struct <%= @grammar.tree_prefix %><%= name %><%= @grammar.tree_suffix %>
 {
     <%= @grammar.prefix %>position_t position;
     <%= @grammar.prefix %>position_t end_position;
-<%     rule_set.ast_fields.each do |fields| %>
+    ushort n_fields;
+    bool is_token;
+<%     rule_set.tree_fields.each do |fields| %>
     union
     {
 <%       fields.each do |field_name, type| %>
@@ -172,7 +195,7 @@ public struct <%= @grammar.prefix %>context_t
     /* Parser context data. */
 
     /** Parse result value. */
-<% if @grammar.ast %>
+<% if @grammar.tree %>
     void * parse_result;
 <% else %>
     <%= @grammar.prefix %>value_t parse_result;
@@ -183,6 +206,8 @@ public struct <%= @grammar.prefix %>context_t
 
     /** User terminate code. */
     size_t user_terminate_code;
+
+<%= @grammar.context_user_fields %>
 }
 
 /**************************************************************************
@@ -222,26 +247,39 @@ private enum size_t INVALID_ID = cast(size_t)-1;
  *************************************************************************/
 
 /**
- * Initialize lexer/parser context structure.
+ * Allocate and 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.
  */
-public void <%= @grammar.prefix %>context_init(<%= @grammar.prefix %>context_t * context, string input)
+<%= @grammar.prefix %>context_t * <%= @grammar.prefix %>context_new(string input)
 {
     /* New default-initialized context structure. */
-    <%= @grammar.prefix %>context_t newcontext;
+    <%= @grammar.prefix %>context_t * context = new <%= @grammar.prefix %>context_t;
 
     /* Lexer initialization. */
-    newcontext.input = input;
+    context.input = input;
-    newcontext.text_position.row = 1u;
+    context.text_position.row = 1u;
-    newcontext.text_position.col = 1u;
+    context.text_position.col = 1u;
-    newcontext.mode = <%= @lexer.mode_id("default") %>;
+    context.mode = <%= @lexer.mode_id("default") %>;
 
-    /* Copy to the user's context structure. */
+    return context;
-    *context = newcontext;
+}
+
+/**
+ * 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)
+{
 }
 
 /**************************************************************************
@@ -627,6 +665,22 @@ 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 = context.input[context.input_index..(context.input_index + match_info.length)];
@@ -665,17 +719,6 @@ 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;
 
@@ -797,7 +840,7 @@ private struct reduce_t
      * reduce action.
      */
     parser_state_id_t n_states;
-<% if @grammar.ast %>
+<% if @grammar.tree %>
 
     /**
      * Map of rule components to rule set child fields.
@@ -805,7 +848,7 @@ private struct reduce_t
     immutable(ushort) * rule_set_node_field_index_map;
 
     /**
-     * Number of rule set AST node fields.
+     * Number of rule set tree node fields.
      */
     ushort rule_set_node_field_array_size;
 
@@ -844,12 +887,14 @@ private struct state_value_t
     /** Parser state ID. */
     size_t state_id;
 
+<% if @grammar.tree %>
+    /** Tree node. */
+    void * tree_node;
+<% 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)
@@ -865,7 +910,7 @@ private immutable shift_t[] parser_shift_table = [
 <%   end %>
 ];
 
-<% if @grammar.ast %>
+<% if @grammar.tree %>
 <%   @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("$", "_") %><%= rule.id %>_node_field_index_map = [<%= rule.rule_set_node_field_index_map.map {|v| v.to_s}.join(", ") %>];
@@ -880,14 +925,14 @@ private immutable reduce_t[] parser_reduce_table = [
         <%= reduce[:token_id] %>u, /* Token: <%= reduce[:token] ? reduce[:token].name : "(any)" %> */
         <%= reduce[:rule_id] %>u, /* Rule ID */
         <%= reduce[:rule_set_id] %>u, /* Rule set ID (<%= reduce[:rule].rule_set.name %>) */
-<%     if @grammar.ast %>
+<%     if @grammar.tree %>
         <%= reduce[:n_states] %>u, /* Number of states */
 <%       if reduce[:rule].flat_rule_set_node_field_index_map? %>
         null, /* No rule set node field index map (flat map) */
 <%       else %>
         &r_<%= reduce[:rule].name.gsub("$", "_") %><%= reduce[:rule].id %>_node_field_index_map[0], /* Rule set node field index map */
 <%       end %>
-        <%= reduce[:rule].rule_set.ast_fields.size %>, /* Number of AST fields */
+        <%= reduce[:rule].rule_set.tree_fields.size %>, /* Number of tree fields */
         <%= reduce[:propagate_optional_target] %>), /* Propagate optional target? */
 <%     else %>
         <%= reduce[:n_states] %>u), /* Number of states */
@@ -902,7 +947,56 @@ private immutable parser_state_t[] parser_state_table = [
 <%   end %>
 ];
 
-<% unless @grammar.ast %>
+<% unless @grammar.tree %>
+/**
+ * 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.
  *
@@ -913,7 +1007,7 @@ private immutable parser_state_t[] parser_state_table = [
  * @retval P_USER_TERMINATED
  *   User requested to terminate parsing.
  */
-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)
+private size_t parser_user_code(<%= @grammar.tree ? "void" : "#{@grammar.prefix}value_t" %> * _pvalue, uint rule, state_value_t[] statevalues, uint n_states, <%= @grammar.prefix %>context_t * context)
 {
     switch (rule)
     {
@@ -1006,16 +1100,18 @@ private size_t parse_from(<%= @grammar.prefix %>context_t * context, size_t star
     state_value_t[] statevalues = new state_value_t[](1);
     statevalues[0].state_id = start_state_id;
     size_t reduced_rule_set = INVALID_ID;
-<% if @grammar.ast %>
+<% if @grammar.tree %>
     void * 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 = <%= @grammar.prefix %>lex(context, &token_info);
+            size_t lexer_result = <%= lex_fn %>(context, &token_info);
             if (lexer_result != P_SUCCESS)
             {
                 return lexer_result;
@@ -1033,8 +1129,8 @@ 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.ast %>
+<% if @grammar.tree %>
-                context.parse_result = statevalues[$-1].ast_node;
+                context.parse_result = statevalues[$-1].tree_node;
 <% else %>
                 context.parse_result = statevalues[$-1].pvalue;
 <% end %>
@@ -1048,10 +1144,13 @@ private size_t parse_from(<%= @grammar.prefix %>context_t * context, size_t star
             if (reduced_rule_set == INVALID_ID)
             {
                 /* We shifted a token, mark it consumed. */
-<% if @grammar.ast %>
+<% if @grammar.tree %>
-                <%= @grammar.ast_prefix %>Token<%= @grammar.ast_suffix %> * token_ast_node = new <%= @grammar.ast_prefix %>Token<%= @grammar.ast_suffix %>(token_info.position, token_info.end_position, token, token_info.pvalue);
+                <%= @grammar.tree_prefix %>Token<%= @grammar.tree_suffix %> * token_tree_node = new <%= @grammar.tree_prefix %>Token<%= @grammar.tree_suffix %>(token_info.position, token_info.end_position, 0u, true, token, token_info.pvalue);
-                statevalues[$-1].ast_node = token_ast_node;
+<%= expand_code(@grammar.on_token_node, false, nil, nil) %>
+                statevalues[$-1].tree_node = token_tree_node;
 <% else %>
+                statevalues[$-1].position = token_info.position;
+                statevalues[$-1].end_position = token_info.end_position;
                 statevalues[$-1].pvalue = token_info.pvalue;
 <% end %>
                 token = INVALID_TOKEN_ID;
@@ -1059,10 +1158,12 @@ private size_t parse_from(<%= @grammar.prefix %>context_t * context, size_t star
             else
             {
                 /* We shifted a RuleSet. */
-<% if @grammar.ast %>
+<% if @grammar.tree %>
-                statevalues[$-1].ast_node = reduced_parser_node;
+                statevalues[$-1].tree_node = reduced_parser_node;
 <% else %>
                 statevalues[$-1].pvalue = reduced_parser_value;
+                statevalues[$-1].position = reduced_position;
+                statevalues[$-1].end_position = reduced_end_position;
                 <%= @grammar.prefix %>value_t new_parse_result;
                 reduced_parser_value = new_parse_result;
 <% end %>
@@ -1075,19 +1176,21 @@ private size_t parse_from(<%= @grammar.prefix %>context_t * context, size_t star
         if (reduce_index != INVALID_ID)
         {
             /* We have something to reduce. */
-<% if @grammar.ast %>
+<% if @grammar.tree %>
             if (parser_reduce_table[reduce_index].propagate_optional_target)
             {
-                reduced_parser_node = statevalues[$ - 1].ast_node;
+                reduced_parser_node = statevalues[$ - 1].tree_node;
             }
             else if (parser_reduce_table[reduce_index].n_states > 0)
             {
                 size_t n_fields = parser_reduce_table[reduce_index].rule_set_node_field_array_size;
-                size_t node_size = ASTNode.sizeof + n_fields * (void *).sizeof;
+                size_t node_size = TreeNode.sizeof + n_fields * (void *).sizeof;
-                ASTNode * node = cast(ASTNode *)malloc(node_size);
+                TreeNode * node = cast(TreeNode *)malloc(node_size);
                 GC.addRange(node, node_size);
                 node.position = <%= @grammar.prefix %>position_t.INVALID;
                 node.end_position = <%= @grammar.prefix %>position_t.INVALID;
+                node.n_fields = cast(ushort)n_fields;
+                node.is_token = false;
                 foreach (i; 0..n_fields)
                 {
                     node.fields[i] = null;
@@ -1096,20 +1199,20 @@ private size_t parse_from(<%= @grammar.prefix %>context_t * context, size_t star
                 {
                     foreach (i; 0..parser_reduce_table[reduce_index].n_states)
                     {
-                        node.fields[i] = statevalues[$ - parser_reduce_table[reduce_index].n_states + i].ast_node;
+                        node.fields[i] = statevalues[$ - parser_reduce_table[reduce_index].n_states + i].tree_node;
                     }
                 }
                 else
                 {
                     foreach (i; 0..parser_reduce_table[reduce_index].n_states)
                     {
-                        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;
+                        node.fields[parser_reduce_table[reduce_index].rule_set_node_field_index_map[i]] = statevalues[$ - parser_reduce_table[reduce_index].n_states + i].tree_node;
                     }
                 }
                 bool position_found = false;
                 foreach (i; 0..n_fields)
                 {
-                    ASTNode * child = cast(ASTNode *)node.fields[i];
+                    TreeNode * child = cast(TreeNode *)node.fields[i];
                     if (child && child.position.valid)
                     {
                         if (!position_found)
@@ -1126,6 +1229,12 @@ private size_t parse_from(<%= @grammar.prefix %>context_t * context, size_t star
             {
                 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)
@@ -1133,6 +1242,16 @@ 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;
@@ -1171,15 +1290,15 @@ public size_t <%= @grammar.prefix %>parse_<%= start_rule %>(<%= @grammar.prefix
  *
  * @return Parse result value.
  */
-<% if @grammar.ast %>
+<% if @grammar.tree %>
-public <%= @grammar.ast_prefix %><%= @grammar.start_rules[0] %><%= @grammar.ast_suffix %> * <%= @grammar.prefix %>result(<%= @grammar.prefix %>context_t * context)
+public <%= @grammar.tree_prefix %><%= @grammar.start_rules[0] %><%= @grammar.tree_suffix %> * <%= @grammar.prefix %>result(<%= @grammar.prefix %>context_t * context)
 {
-    return cast(<%= @grammar.ast_prefix %><%= @grammar.start_rules[0] %><%= @grammar.ast_suffix %> *)context.parse_result;
+    return cast(<%= @grammar.tree_prefix %><%= @grammar.start_rules[0] %><%= @grammar.tree_suffix %> *)context.parse_result;
 }
 <%   @grammar.start_rules.each_with_index do |start_rule, i| %>
-public <%= @grammar.ast_prefix %><%= start_rule %><%= @grammar.ast_suffix %> * <%= @grammar.prefix %>result_<%= start_rule %>(<%= @grammar.prefix %>context_t * context)
+public <%= @grammar.tree_prefix %><%= start_rule %><%= @grammar.tree_suffix %> * <%= @grammar.prefix %>result_<%= start_rule %>(<%= @grammar.prefix %>context_t * context)
 {
-    return cast(<%= @grammar.ast_prefix %><%= start_rule %><%= @grammar.ast_suffix %> *)context.parse_result;
+    return cast(<%= @grammar.tree_prefix %><%= start_rule %><%= @grammar.tree_suffix %> *)context.parse_result;
 }
 <%   end %>
 <% else %>
@@ -1195,6 +1314,35 @@ public <%= start_rule_type(i)[1] %> <%= @grammar.prefix %>result_<%= start_rule
 <%   end %>
 <% end %>
 
+<% if @grammar.tree %>
+private void tree_delete(TreeNode * node)
+{
+    if (!node.is_token)
+    {
+        for (size_t i = 0u; i < node.n_fields; i++)
+        {
+            if (node.fields[i])
+            {
+                tree_delete(cast(TreeNode *)node.fields[i]);
+            }
+        }
+        GC.removeRange(node);
+        free(node);
+    }
+}
+
+void <%= @grammar.prefix %>tree_delete(<%= @grammar.tree_prefix %><%= @grammar.start_rules[0] %><%= @grammar.tree_suffix %> * tree)
+{
+    tree_delete(cast(TreeNode *)tree);
+}
+<%   @grammar.start_rules.each_with_index do |start_rule, i| %>
+void <%= @grammar.prefix %>tree_delete_<%= start_rule %>(<%= @grammar.tree_prefix %><%= start_rule %><%= @grammar.tree_suffix %> * tree)
+{
+    tree_delete(cast(TreeNode *)tree);
+}
+<%   end %>
+<% end %>
+
 /**
  * Get the current text input position.
  *

assets/parser.h.erb

@@ -45,10 +45,10 @@ typedef uint32_t <%= @grammar.prefix %>code_point_t;
  */
 typedef struct
 {
-    /** Input text row (0-based). */
+    /** Input text row (1-based). */
     uint32_t row;
 
-    /** Input text column (0-based). */
+    /** Input text column (1-based). */
     uint32_t col;
 } <%= @grammar.prefix %>position_t;
 
@@ -58,7 +58,7 @@ typedef struct
 /** User header code blocks. */
 <%= @grammar.code_blocks.fetch("header", "") %>
 
-<% if @grammar.ast %>
+<% if @grammar.tree %>
 /** Parser values type. */
 typedef <%= @grammar.ptype %> <%= @grammar.prefix %>value_t;
 <% else %>
@@ -69,20 +69,37 @@ 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 %>
 
-<% if @grammar.ast %>
+/** Parser value accessor(s). */
-/** AST node types. @{ */
+<%   @grammar.ptypes.each do |name, typestring| %>
-typedef struct <%= @grammar.ast_prefix %>Token<%= @grammar.ast_suffix %>
+static inline <%= typestring %> <%= @grammar.prefix %>value_get<%= name == "default" ? "" : "_#{name}" %>(<%= @grammar.prefix %>value_t const * pvalue)
 {
-    <% # ASTNode fields must be present in the same order here. # %>
+    return pvalue->v_<%= name %>;
+}
+<%   end %>
+<% end %>
+
+<% if @grammar.tree %>
+/** Tree node types. @{ */
+typedef struct <%= @grammar.tree_prefix %>Token<%= @grammar.tree_suffix %>
+{
+    <% # TreeNode 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;
+<%= @grammar.token_user_fields %>
     <%= @grammar.prefix %>token_t token;
     <%= @grammar.prefix %>value_t pvalue;
-} <%= @grammar.ast_prefix %>Token<%= @grammar.ast_suffix %>;
+} <%= @grammar.tree_prefix %>Token<%= @grammar.tree_suffix %>;
 
 <%   @parser.rule_sets.each do |name, rule_set| %>
 <%     next if name.start_with?("$") %>
@@ -93,14 +110,14 @@ struct <%= name %>;
 <%   @parser.rule_sets.each do |name, rule_set| %>
 <%     next if name.start_with?("$") %>
 <%     next if rule_set.optional? %>
-typedef struct <%= @grammar.ast_prefix %><%= name %><%= @grammar.ast_suffix %>
+typedef struct <%= @grammar.tree_prefix %><%= name %><%= @grammar.tree_suffix %>
 {
-    <% # ASTNode fields must be present in the same order here. # %>
+    <% # TreeNode 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| %>
+<%     rule_set.tree_fields.each do |fields| %>
     union
     {
 <%       fields.each do |field_name, type| %>
@@ -108,7 +125,7 @@ typedef struct <%= @grammar.ast_prefix %><%= name %><%= @grammar.ast_suffix %>
 <%       end %>
     };
 <%     end %>
-} <%= @grammar.ast_prefix %><%= name %><%= @grammar.ast_suffix %>;
+} <%= @grammar.tree_prefix %><%= name %><%= @grammar.tree_suffix %>;
 
 <%   end %>
 /** @} */
@@ -161,7 +178,7 @@ typedef struct
     /* Parser context data. */
 
     /** Parse result value. */
-<% if @grammar.ast %>
+<% if @grammar.tree %>
     void * parse_result;
 <% else %>
     <%= @grammar.prefix %>value_t parse_result;
@@ -172,6 +189,8 @@ typedef struct
 
     /** User terminate code. */
     size_t user_terminate_code;
+
+<%= @grammar.context_user_fields %>
 } <%= @grammar.prefix %>context_t;
 
 /**************************************************************************
@@ -181,7 +200,9 @@ typedef struct
 /** Token names. */
 extern const char * <%= @grammar.prefix %>token_names[];
 
-void <%= @grammar.prefix %>context_init(<%= @grammar.prefix %>context_t * context, uint8_t const * input, size_t input_length);
+<%= @grammar.prefix %>context_t * <%= @grammar.prefix %>context_new(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);
@@ -193,10 +214,10 @@ size_t <%= @grammar.prefix %>parse(<%= @grammar.prefix %>context_t * context);
 size_t <%= @grammar.prefix %>parse_<%= start_rule %>(<%= @grammar.prefix %>context_t * context);
 <%   end %>
 
-<% if @grammar.ast %>
+<% if @grammar.tree %>
-<%= @grammar.ast_prefix %><%= @grammar.start_rules[0] %><%= @grammar.ast_suffix %> * <%= @grammar.prefix %>result(<%= @grammar.prefix %>context_t * context);
+<%= @grammar.tree_prefix %><%= @grammar.start_rules[0] %><%= @grammar.tree_suffix %> * <%= @grammar.prefix %>result(<%= @grammar.prefix %>context_t * context);
 <%   @grammar.start_rules.each_with_index do |start_rule, i| %>
-<%= @grammar.ast_prefix %><%= start_rule %><%= @grammar.ast_suffix %> * <%= @grammar.prefix %>result_<%= start_rule %>(<%= @grammar.prefix %>context_t * context);
+<%= @grammar.tree_prefix %><%= start_rule %><%= @grammar.tree_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);
@@ -205,10 +226,10 @@ size_t <%= @grammar.prefix %>parse_<%= start_rule %>(<%= @grammar.prefix %>conte
 <%   end %>
 <% end %>
 
-<% if @grammar.ast %>
+<% if @grammar.tree %>
-void <%= @grammar.prefix %>free_ast(<%= @grammar.ast_prefix %><%= @grammar.start_rules[0] %><%= @grammar.ast_suffix %> * ast);
+void <%= @grammar.prefix %>tree_delete(<%= @grammar.tree_prefix %><%= @grammar.start_rules[0] %><%= @grammar.tree_suffix %> * tree);
 <%   @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);
+void <%= @grammar.prefix %>tree_delete_<%= start_rule %>(<%= @grammar.tree_prefix %><%= start_rule %><%= @grammar.tree_suffix %> * tree);
 <%   end %>
 <% end %>
 

extra/vim/syntax/propane.vim

@@ -8,7 +8,11 @@ if exists("b:current_syntax")
 endif
 
 if !exists("b:propane_subtype")
+  if search('\<import\s\+\%(std\|core\)\.', 'nw') > 0
     let b:propane_subtype = "d"
+  else
+    let b:propane_subtype = "cpp"
+  endif
 endif
 
 exe "syn include @propaneTarget syntax/".b:propane_subtype.".vim"
@@ -16,18 +20,32 @@ 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 "?"
-syn keyword propaneKeyword ast ast_prefix ast_suffix drop module prefix ptype start token tokenid
+" The right-hand side of a rule (after '->' up to '<<' or ';') lists symbol
+" 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))
+        grammar = Grammar.new(File.read(input_file), input_file)
         generator = Generator.new(grammar, output_file, log_file, options)
         generator.generate
       rescue Error => e

lib/propane/generator.rb

@@ -13,8 +13,13 @@ class Propane
       @language =
         if output_file.end_with?(".d")
           "d"
-        else
+        elsif output_file.end_with?(".c")
           "c"
+        elsif output_file =~ %r{\.(cc|cpp|cxx)$}
+          @cpp = true
+          "c"
+        else
+          raise Error.new("Could not determine target language from output file name (#{output_file})")
         end
       @options = options
       process_grammar!
@@ -33,7 +38,13 @@ class Propane
           output_file = @output_file
         end
         erb = ERB.new(template, trim_mode: "<>")
-        result = erb.result(binding.clone)
+        result = erb.result(binding.clone).lines.each_with_index.map do |line, i|
+          if 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
@@ -174,7 +185,8 @@ class Propane
                 end
               end
               @grammar.rules << Rule.new(component, [], nil, ptypename, rule.line_number)
-              @grammar.rules << Rule.new(component, [c], "$$ = $1;\n", ptypename, rule.line_number)
+              optcode = @grammar.tree ? nil : "$$ = $1;\n"
+              @grammar.rules << Rule.new(component, [c], optcode, ptypename, rule.line_number)
               optional_rules_added << component
             end
           end
@@ -265,8 +277,34 @@ class Propane
           "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}"
+        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}"
+        end
+      end
       if parser
         code = code.gsub(/\$\$/) do |match|
+          if @grammar.tree
+            case @language
+            when "c"
+              "((#{@grammar.tree_prefix}#{rule.name}#{@grammar.tree_suffix} *)_pvalue)"
+            when "d"
+              "(cast(#{@grammar.tree_prefix}#{rule.name}#{@grammar.tree_suffix} *)_pvalue)"
+            end
+          else
             case @language
             when "c"
               "_pvalue->v_#{rule.ptypename}"
@@ -274,31 +312,33 @@ class Propane
               "_pvalue.v_#{rule.ptypename}"
             end
           end
-        code = code.gsub(/\$(\d+)/) do |match|
-          index = $1.to_i
-          case @language
-          when "c"
-            "state_values_stack_index(statevalues, -1 - (int)n_states + #{index})->pvalue.v_#{rule.components[index - 1].ptypename}"
-          when "d"
-            "statevalues[$-1-n_states+#{index}].pvalue.v_#{rule.components[index - 1].ptypename}"
         end
+        code = code.gsub(/\$(\d+)/) 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)"
+        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]
-            case @language
+            # Field aliases are just a named reference to a positional rule
-            when "c"
+            # component, so reuse the same expansion as `$1', `$2', etc. Note
-              "state_values_stack_index(statevalues, -(int)n_states + #{index})->pvalue.v_#{rule.components[index].ptypename}"
+            # that rule.aliases stores a 0-based component index, so add 1 to
-            when "d"
+            # convert it to the 1-based index used for positional references.
-              "statevalues[$-n_states+#{index}].pvalue.v_#{rule.components[index].ptypename}"
+            parser_component_reference(rule, index + 1)
-            end
           else
             raise Error.new("Field alias '#{aliasname}' not found")
           end
         end
       else
         code = code.gsub(/\$\$/) do |match|
-          if @grammar.ast
+          if @grammar.tree
             case @language
             when "c"
               "out_token_info->pvalue"
@@ -314,6 +354,22 @@ class Propane
             end
           end
         end
+        code = code.gsub(/\$\{position\}/) do |match|
+          case @language
+          when "c"
+            "out_token_info->position"
+          when "d"
+            "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"
+          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)
@@ -331,6 +387,53 @@ class Propane
       code
     end
 
+    # Expand a positional reference to a parser rule component.
+    #
+    # This is used to expand `$1', `$2', etc. as well as field aliases (which
+    # are just named references to a positional rule component).
+    #
+    # @param rule [Rule]
+    #   The Rule containing the user code.
+    # @param index [Integer]
+    #   1-based index of the rule component to reference.
+    #
+    # @return [String]
+    #   Expanded rule component reference.
+    def parser_component_reference(rule, index)
+      component = rule.components[index - 1]
+      if @grammar.tree
+        # In tree mode a component reference yields a pointer to that
+        # component's tree node. An optional component propagates its target
+        # node (or null), so use the optional target's node type.
+        if component.is_a?(RuleSet) && component.optional?
+          component = component.option_target
+        end
+        node_name = component.is_a?(Token) ? "Token" : component.name
+        typename = "#{@grammar.tree_prefix}#{node_name}#{@grammar.tree_suffix}"
+        case @language
+        when "c"
+          "((#{typename} *)state_values_stack_index(statevalues, -1 - (int)n_states + #{index})->tree_node)"
+        when "d"
+          "(cast(#{typename} *)statevalues[$-1-n_states+#{index}].tree_node)"
+        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}"
+        end
+      end
+    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>]

lib/propane/grammar.rb

@@ -5,9 +5,11 @@ 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 :ast
+    attr_reader :context_user_fields
-    attr_reader :ast_prefix
+    attr_reader :lex_fn
-    attr_reader :ast_suffix
+    attr_reader :tree
+    attr_reader :tree_prefix
+    attr_reader :tree_suffix
     attr_reader :free_token_node
     attr_reader :modulename
     attr_reader :patterns
@@ -17,8 +19,11 @@ 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)
+    def initialize(input, filename)
+      @filename = filename
       @patterns = []
       @start_rules = []
       @tokens = []
@@ -30,10 +35,13 @@ class Propane
       @input = input.gsub("\r\n", "\n")
       @ptypes = {"default" => "void *"}
       @prefix = "p_"
-      @ast = false
+      @tree = false
-      @ast_prefix = ""
+      @tree_prefix = ""
-      @ast_suffix = ""
+      @tree_suffix = ""
-      @free_token_node = nil
+      @free_token_node = ""
+      @context_user_fields = nil
+      @on_token_node = ""
+      @token_user_fields = nil
       parse_grammar!
       @start_rules << "Start" if @start_rules.empty?
     end
@@ -50,6 +58,10 @@ class Propane
       @tokens.size + 1
     end
 
+    def parser_user_code_used?
+      @rules.any? {|r| r.code}
+    end
+
     private
 
     def parse_grammar!
@@ -62,11 +74,15 @@ class Propane
       if parse_white_space!
       elsif parse_comment_line!
       elsif @modeline.nil? && parse_mode_label!
-      elsif parse_ast_statement!
+      elsif parse_context_user_fields_statement!
-      elsif parse_ast_prefix_statement!
+      elsif parse_lex_fn!
-      elsif parse_ast_suffix_statement!
+      elsif parse_tree_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!
@@ -76,6 +92,7 @@ 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) + "..."
@@ -98,27 +115,37 @@ class Propane
       consume!(/#.*\n/)
     end
 
-    def parse_ast_statement!
+    def parse_context_user_fields_statement!
-      if consume!(/ast\s*;/)
+      if md = consume!(/context_user_fields\b\s*/)
-        @ast = true
+        unless code = parse_code_block!
+          raise Error.new("Line #{@line_number}: expected code block")
+        end
+        @context_user_fields ||= ""
+        @context_user_fields += code
       end
     end
 
-    def parse_ast_prefix_statement!
+    def parse_lex_fn!
-      if md = consume!(/ast_prefix\s+(\w+)\s*;/)
+      if md = consume!(/lex_fn\b\s*(\w+)\s*;/)
-        @ast_prefix = md[1]
+        @lex_fn = md[1]
       end
     end
 
-    def parse_ast_suffix_statement!
+    def parse_tree_statement!
-      if md = consume!(/ast_suffix\s+(\w+)\s*;/)
+      if consume!(/tree\s*;/)
-        @ast_suffix = md[1]
+        @tree = true
       end
     end
 
-    def parse_free_token_node_statement!
+    def parse_tree_prefix_statement!
-      if md = consume!(/free_token_node\s+(\w+)\s*;/)
+      if md = consume!(/tree_prefix\s+(\w+)\s*;/)
-        @free_token_node = md[1]
+        @tree_prefix = md[1]
+      end
+    end
+
+    def parse_tree_suffix_statement!
+      if md = consume!(/tree_suffix\s+(\w+)\s*;/)
+        @tree_suffix = md[1]
       end
     end
 
@@ -132,12 +159,40 @@ 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 @ast
+          if @tree
-            raise Error.new("Multiple ptypes are unsupported in AST mode")
+            raise Error.new("Multiple ptypes are unsupported in tree mode")
           end
           name = md[1]
         end
@@ -151,8 +206,8 @@ class Propane
         md = consume!(/(#{IDENTIFIER_REGEX})\s*/, "expected token name")
         name = md[1]
         if md = consume!(/\((#{IDENTIFIER_REGEX})\)\s*/)
-          if @ast
+          if @tree
-            raise Error.new("Multiple ptypes are unsupported in AST mode")
+            raise Error.new("Multiple ptypes are unsupported in tree mode")
           end
           ptypename = md[1]
         end
@@ -175,8 +230,8 @@ class Propane
         md = consume!(/(#{IDENTIFIER_REGEX})\s*/, "expected token name")
         name = md[1]
         if md = consume!(/\((#{IDENTIFIER_REGEX})\)\s*/)
-          if @ast
+          if @tree
-            raise Error.new("Multiple ptypes are unsupported in AST mode")
+            raise Error.new("Multiple ptypes are unsupported in tree mode")
           end
           ptypename = md[1]
         end
@@ -195,8 +250,10 @@ class Propane
           raise Error.new("Line #{@line_number}: expected pattern to follow `drop'")
         end
         consume!(/\s+/)
-        consume!(/;/, "expected `;'")
+        unless code = parse_code_block!
-        @patterns << Pattern.new(pattern: pattern, line_number: @line_number, modes: get_modes_from_modeline)
+          consume!(/;/, "expected `;' or code block")
+        end
+        @patterns << Pattern.new(pattern: pattern, line_number: @line_number, code: code, modes: get_modes_from_modeline)
         @modeline = nil
         true
       end
@@ -205,18 +262,14 @@ class Propane
     def parse_rule_statement!
       if md = consume!(/(#{IDENTIFIER_REGEX})\s*(?:\((#{IDENTIFIER_REGEX})\))?\s*->\s*/)
         rule_name, ptypename = *md[1, 2]
-        if @ast && ptypename
+        if @tree && ptypename
-          raise Error.new("Multiple ptypes are unsupported in AST mode")
+          raise Error.new("Multiple ptypes are unsupported in tree 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
@@ -227,8 +280,8 @@ class Propane
       if pattern = parse_pattern!
         consume!(/\s+/)
         if md = consume!(/\((#{IDENTIFIER_REGEX})\)\s*/)
-          if @ast
+          if @tree
-            raise Error.new("Multiple ptypes are unsupported in AST mode")
+            raise Error.new("Multiple ptypes are unsupported in tree mode")
           end
           ptypename = md[1]
         end
@@ -253,8 +306,14 @@ class Propane
     def parse_code_block_statement!
       if md = consume!(/<<([a-z]*)(.*?)>>\n/m)
         name, code = md[1..2]
-        code.sub!(/\A\n/, "")
+        code = code.chomp
-        code += "\n" unless code.end_with?("\n")
+        unless @noline
+          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
@@ -272,6 +331,13 @@ 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 = ""
@@ -285,6 +351,8 @@ 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
@@ -293,9 +361,14 @@ class Propane
 
     def parse_code_block!
       if md = consume!(/<<(.*?)>>\n/m)
-        code = md[1]
+        code = md[1].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
         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 AST construction.
+    #   node field pointer array. This is used for tree 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>]
-    #   AST fields.
+    #   tree fields.
-    attr_reader :ast_fields
+    attr_reader :tree_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.ast
+      if grammar.tree
-        build_ast_fields(grammar)
+        build_tree_fields(grammar)
       end
     end
 
     private
 
-    # Build the set of AST fields for this RuleSet.
+    # Build the set of tree fields for this RuleSet.
     #
     # This is an Array of Hashes. Each entry in the Array corresponds to a
-    # field location in the AST node. The entry is a Hash. It could have one or
+    # field location in the tree 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_ast_fields(grammar)
+    def build_tree_fields(grammar)
-      field_ast_node_indexes = {}
+      field_tree_node_indexes = {}
       field_indexes_across_all_rules = {}
-      # Stores the index into @ast_fields by field alias name.
+      # Stores the index into @tree_fields by field alias name.
       field_aliases = {}
-      @ast_fields = []
+      @tree_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.ast_prefix}#{node_name}#{grammar.ast_suffix}"
+          struct_name = "#{grammar.tree_prefix}#{node_name}#{grammar.tree_suffix}"
           field_name = "p#{node_name}#{i + 1}"
-          unless field_ast_node_indexes[field_name]
+          unless field_tree_node_indexes[field_name]
-            field_ast_node_indexes[field_name] = @ast_fields.size
+            field_tree_node_indexes[field_name] = @tree_fields.size
-            @ast_fields << {field_name => struct_name}
+            @tree_fields << {field_name => struct_name}
           end
           rule.aliases.each do |alias_name, index|
             if index == i
-              alias_ast_fields_index = field_ast_node_indexes[field_name]
+              alias_tree_fields_index = field_tree_node_indexes[field_name]
-              if field_aliases[alias_name] && field_aliases[alias_name] != alias_ast_fields_index
+              if field_aliases[alias_name] && field_aliases[alias_name] != alias_tree_fields_index
-                raise Error.new("Error: conflicting AST node field positions for alias `#{alias_name}` in rule #{rule.name} defined on line #{rule.line_number}")
+                raise Error.new("Error: conflicting tree node field positions for alias `#{alias_name}` in rule #{rule.name} defined on line #{rule.line_number}")
               end
-              field_aliases[alias_name] = alias_ast_fields_index
+              field_aliases[alias_name] = alias_tree_fields_index
-              @ast_fields[alias_ast_fields_index][alias_name] = @ast_fields[alias_ast_fields_index].first[1]
+              @tree_fields[alias_tree_fields_index][alias_name] = @tree_fields[alias_tree_fields_index].first[1]
             end
           end
           field_indexes_across_all_rules[node_name] ||= Set.new
-          field_indexes_across_all_rules[node_name] << field_ast_node_indexes[field_name]
+          field_indexes_across_all_rules[node_name] << field_tree_node_indexes[field_name]
-          rule.rule_set_node_field_index_map[i] = field_ast_node_indexes[field_name]
+          rule.rule_set_node_field_index_map[i] = field_tree_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.
-          @ast_fields[indexes_across_all_rules.first]["p#{node_name}"] =
+          @tree_fields[indexes_across_all_rules.first]["p#{node_name}"] =
-            "#{grammar.ast_prefix}#{node_name}#{grammar.ast_suffix}"
+            "#{grammar.tree_prefix}#{node_name}#{grammar.tree_suffix}"
         end
       end
     end

lib/propane/version.rb

@@ -1,3 +1,3 @@
 class Propane
-  VERSION = "2.3.0"
+  VERSION = "4.7.0"
 end

spec/propane/grammar_spec.rb

@@ -28,7 +28,7 @@ B -> <<
   b = 0;
 >>
 EOF
-      grammar = Grammar.new(input)
+      grammar = Grammar.new(input, "test.propane")
       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 "Code for the token\n"
+      expect(o.code).to eq %[#line 12 "test.propane"\nCode for the token\n#linereset\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 "  a = 42;\n"
+      expect(o.code).to eq %[#line 20 "test.propane"\n  a = 42;\n#linereset\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 "  b = 0;\n"
+      expect(o.code).to eq %[#line 24 "test.propane"\n  b = 0;\n#linereset\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)
+      grammar = Grammar.new(input, "test.propane")
       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 "  a = b;\n  return c;\n"
+      expect(o.code).to eq %[#line 2 "test.propane"\n  a = b;\n  return c;\n#linereset\n]
 
       o = grammar.tokens.find {|token| token.name == "code2"}
       expect(o).to_not be_nil
@@ -130,7 +130,42 @@ EOF
 
       o = grammar.patterns.find {|pattern| pattern.token == o}
       expect(o).to_not be_nil
-      expect(o.code).to eq %[  writeln("Hello there");\n]
+      expect(o.code).to eq %[#line 7 "test.propane"\n  writeln("Hello there");\n#linereset\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
@@ -144,7 +179,7 @@ m2: /bar/ <<
 drop /q/;
 m3: drop /r/;
 EOF
-      grammar = Grammar.new(input)
+      grammar = Grammar.new(input, "test.propane")
 
       o = grammar.tokens.find {|token| token.name == "a"}
       expect(o).to_not be_nil
@@ -197,7 +232,7 @@ tokenid int(integer);
 Start (node) -> R;
 R -> abc int;
 EOF
-      grammar = Grammar.new(input)
+      grammar = Grammar.new(input, "test.propane")
 
       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)
+  grammar = Propane::Grammar.new(grammar, "test.propane")
   token_dfa = Propane::Lexer::DFA.new(grammar.patterns)
   test_lexer = TestLexer.new(token_dfa)
   test_lexer.lex(input)

spec/propane_spec.rb

@@ -14,6 +14,7 @@ describe Propane do
   end
 
   def run_propane(options = {})
+    options[:language] ||= "d"
     @statics[:build_test_id] ||= 0
     @statics[:build_test_id] += 1
     if ENV["dist_specs"]
@@ -94,17 +95,15 @@ EOF
     File.binwrite("spec/run/.stdout", stdout)
     stderr.sub!(/^.*modules passed unittests\n/, "")
     results = Results.new(stdout, stderr, status)
-    if %w[c cpp].include?(options[:language])
     stdout, stderr, status = Open3.capture3("valgrind --leak-check=full --show-leak-kinds=all --track-origins=yes --verbose spec/run/testparser")
     vgout = stdout + stderr
     File.binwrite("spec/run/.vgout", vgout)
-      vgout.scan(/lost: (\d+) bytes/) do |match|
+    vgout.scan(/(?:definitely|indirectly) lost: (\d+) bytes/) do |match|
       bytes = $1.to_i
       if bytes > 0
         raise "Valgrind detected memory leak"
       end
     end
-    end
     results
   end
 
@@ -245,20 +244,20 @@ EOF
     expect(results.status).to_not eq 0
   end
 
-  it "errors when an alias is in different positions for different rules in a rule set when AST mode is enabled" do
+  it "errors when an alias is in different positions for different rules in a rule set when tree mode is enabled" do
     write_grammar <<EOF
-ast;
+tree;
 token a;
 token b;
 Start -> a:foo b;
 Start -> b b:foo;
 EOF
     results = run_propane(extra_args: %w[-w], capture: true)
-    expect(results.stderr).to match %r{Error: conflicting AST node field positions for alias `foo`}
+    expect(results.stderr).to match %r{Error: conflicting tree node field positions for alias `foo`}
     expect(results.status).to_not eq 0
   end
 
-  it "does not error when an alias is in different positions for different rules in a rule set when AST mode is not enabled" do
+  it "does not error when an alias is in different positions for different rules in a rule set when tree mode is not enabled" do
     write_grammar <<EOF
 token a;
 token b;
@@ -746,6 +745,78 @@ EOF
         ])
       end
 
+      it "executes user code associated with a parser rule in tree mode" do
+        case language
+        when "c", "cpp"
+          write_grammar <<EOF
+tree;
+context_user_fields <<
+    int start_n_fields;
+    int start_a_value;
+    int a_value;
+    int b_value;
+    p_token_t b_token;
+    int c_is_null;
+    int c_field_is_null;
+    int alias_a_value;
+    int alias_b_value;
+>>
+ptype int;
+token a << $$ = 11; >>
+token b << $$ = 22; >>
+Start -> A:ay B:bee C <<
+  ${context.start_n_fields} = $$->n_fields;
+  ${context.start_a_value} = $$->pA->pToken1->pvalue;
+  ${context.a_value} = $1->pToken1->pvalue;
+  ${context.b_value} = $2->pToken1->pvalue;
+  ${context.b_token} = $2->pToken1->token;
+  ${context.c_field_is_null} = ($$->pC == NULL) ? 1 : 0;
+  ${context.alias_a_value} = ${ay}->pToken1->pvalue;
+  ${context.alias_b_value} = ${bee}->pToken1->pvalue;
+>>
+A -> a;
+B -> b;
+C -> << ${context.c_is_null} = ($$ == NULL) ? 1 : 0; >>
+EOF
+        when "d"
+          write_grammar <<EOF
+tree;
+context_user_fields <<
+    int start_n_fields;
+    int start_a_value;
+    int a_value;
+    int b_value;
+    p_token_t b_token;
+    int c_is_null;
+    int c_field_is_null;
+    int alias_a_value;
+    int alias_b_value;
+>>
+ptype int;
+token a << $$ = 11; >>
+token b << $$ = 22; >>
+Start -> A:ay B:bee C <<
+  ${context.start_n_fields} = $$.n_fields;
+  ${context.start_a_value} = $$.pA.pToken1.pvalue;
+  ${context.a_value} = $1.pToken1.pvalue;
+  ${context.b_value} = $2.pToken1.pvalue;
+  ${context.b_token} = $2.pToken1.token;
+  ${context.c_field_is_null} = ($$.pC is null) ? 1 : 0;
+  ${context.alias_a_value} = ${ay}.pToken1.pvalue;
+  ${context.alias_b_value} = ${bee}.pToken1.pvalue;
+>>
+A -> a;
+B -> b;
+C -> << ${context.c_is_null} = ($$ is null) ? 1 : 0; >>
+EOF
+        end
+        run_propane(language: language)
+        compile("spec/test_parser_user_code_tree.#{language}", language: language)
+        results = run_test(language: language)
+        expect(results.stderr).to eq ""
+        expect(results.status).to eq 0
+      end
+
       it "parses lists" do
         write_grammar <<EOF
 ptype #{language == "d" ? "uint" : "uint32_t"};
@@ -911,6 +982,57 @@ EOF
         expect(results.status).to eq 0
       end
 
+      it "allows lexer code blocks to access the matched token positions" do
+        case language
+        when "c", "cpp"
+          write_grammar <<EOF
+<<
+#include <stdio.h>
+>>
+context_user_fields <<
+    p_position_t last_start;
+    p_position_t last_end;
+>>
+drop /\\s+/;
+token word /[a-z]+/ <<
+    ${context.last_start} = ${position};
+    ${context.last_end} = ${end_position};
+>>
+token stop /!/ <<
+    $terminate(42);
+>>
+Start -> Words;
+Words -> ;
+Words -> word Words;
+Words -> stop Words;
+EOF
+        when "d"
+          write_grammar <<EOF
+context_user_fields <<
+    p_position_t last_start;
+    p_position_t last_end;
+>>
+drop /\\s+/;
+token word /[a-z]+/ <<
+    ${context.last_start} = ${position};
+    ${context.last_end} = ${end_position};
+>>
+token stop /!/ <<
+    $terminate(42);
+>>
+Start -> Words;
+Words -> ;
+Words -> word Words;
+Words -> stop Words;
+EOF
+        end
+        run_propane(language: language)
+        compile("spec/test_lexer_positions.#{language}", language: language)
+        results = run_test(language: language)
+        expect(results.stderr).to eq ""
+        expect(results.status).to eq 0
+      end
+
       it "allows the user to terminate the parser" do
         write_grammar <<EOF
 token a;
@@ -996,9 +1118,9 @@ EOF
         run_propane(language: language)
       end
 
-      it "generates an AST" do
+      it "generates a tree" do
         write_grammar <<EOF
-ast;
+tree;
 
 ptype int;
 
@@ -1032,17 +1154,17 @@ One -> one;
 Two -> two;
 EOF
         run_propane(language: language)
-        compile("spec/test_ast.#{language}", language: language)
+        compile("spec/test_tree.#{language}", language: language)
         results = run_test(language: language)
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
       end
 
-      it "supports AST node prefix and suffix" do
+      it "supports tree node prefix and suffix" do
         write_grammar <<EOF
-ast;
+tree;
-ast_prefix P ;
+tree_prefix P ;
-ast_suffix  S;
+tree_suffix  S;
 
 ptype int;
 
@@ -1076,7 +1198,7 @@ One -> one;
 Two -> two;
 EOF
         run_propane(language: language)
-        compile("spec/test_ast_ps.#{language}", language: language)
+        compile("spec/test_tree_ps.#{language}", language: language)
         results = run_test(language: language)
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
@@ -1092,15 +1214,15 @@ EOF
         compile("spec/test_start_rule.#{language}", language: language)
       end
 
-      it "allows specifying a different start rule with AST generation" do
+      it "allows specifying a different start rule with tree generation" do
         write_grammar <<EOF
-ast;
+tree;
 token hi;
 start Top;
 Top -> hi;
 EOF
         run_propane(language: language)
-        compile("spec/test_start_rule_ast.#{language}", language: language)
+        compile("spec/test_start_rule_tree.#{language}", language: language)
       end
 
       it "allows marking a rule component as optional" do
@@ -1167,10 +1289,10 @@ EOF
         ])
       end
 
-      it "allows marking a rule component as optional in AST generation mode" do
+      it "allows marking a rule component as optional in tree generation mode" do
         if language == "d"
           write_grammar <<EOF
-ast;
+tree;
 
 <<
 import std.stdio;
@@ -1186,7 +1308,7 @@ R -> d c;
 EOF
         else
           write_grammar <<EOF
-ast;
+tree;
 
 <<
 #include <stdio.h>
@@ -1202,16 +1324,16 @@ R -> d c;
 EOF
         end
         run_propane(language: language)
-        compile("spec/test_optional_rule_component_ast.#{language}", language: language)
+        compile("spec/test_optional_rule_component_tree.#{language}", language: language)
         results = run_test(language: language)
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
       end
 
-      it "allows naming an optional rule component in AST generation mode" do
+      it "allows naming an optional rule component in tree generation mode" do
         if language == "d"
           write_grammar <<EOF
-ast;
+tree;
 
 <<
 import std.stdio;
@@ -1227,7 +1349,7 @@ R -> d c;
 EOF
         else
           write_grammar <<EOF
-ast;
+tree;
 
 <<
 #include <stdio.h>
@@ -1243,15 +1365,15 @@ R -> d c;
 EOF
         end
         run_propane(language: language)
-        compile("spec/test_named_optional_rule_component_ast.#{language}", language: language)
+        compile("spec/test_named_optional_rule_component_tree.#{language}", language: language)
         results = run_test(language: language)
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
       end
 
-      it "stores token and rule positions in AST nodes" do
+      it "stores token and rule positions in tree nodes" do
         write_grammar <<EOF
-ast;
+tree;
 
 token a;
 token bb;
@@ -1263,7 +1385,7 @@ T -> bb;
 T -> c;
 EOF
         run_propane(language: language)
-        compile("spec/test_ast_token_positions.#{language}", language: language)
+        compile("spec/test_tree_token_positions.#{language}", language: language)
         results = run_test(language: language)
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
@@ -1271,7 +1393,7 @@ EOF
 
       it "stores invalid positions for empty rule matches" do
         write_grammar <<EOF
-ast;
+tree;
 
 token a;
 token bb;
@@ -1283,15 +1405,15 @@ T -> a A;
 A -> bb? c?;
 EOF
         run_propane(language: language)
-        compile("spec/test_ast_invalid_positions.#{language}", language: language)
+        compile("spec/test_tree_invalid_positions.#{language}", language: language)
         results = run_test(language: language)
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
       end
 
-      it "allows specifying field aliases in AST mode" do
+      it "allows specifying field aliases in tree mode" do
         write_grammar <<EOF
-ast;
+tree;
 
 token a;
 token b;
@@ -1303,15 +1425,15 @@ T -> b;
 T -> c;
 EOF
         run_propane(language: language)
-        compile("spec/test_ast_field_aliases.#{language}", language: language)
+        compile("spec/test_tree_field_aliases.#{language}", language: language)
         results = run_test(language: language)
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
       end
 
-      it "aliases the correct field when multiple rules are in a rule set in AST mode" do
+      it "aliases the correct field when multiple rules are in a rule set in tree mode" do
         write_grammar <<EOF
-ast;
+tree;
 
 token a;
 token b;
@@ -1326,13 +1448,13 @@ T -> b;
 T -> c;
 EOF
         run_propane(language: language)
-        compile("spec/test_ast_field_aliases.#{language}", language: language)
+        compile("spec/test_tree_field_aliases.#{language}", language: language)
         results = run_test(language: language)
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
       end
 
-      it "allows specifying field aliases when AST mode is not enabled" do
+      it "allows specifying field aliases when tree mode is not enabled" do
         if language == "d"
           write_grammar <<EOF
 <<
@@ -1378,7 +1500,7 @@ EOF
         expect(results.stdout).to match /first is foo1.*second is bar2/m
       end
 
-      it "aliases the correct field when multiple rules are in a rule set when AST mode is not enabled" do
+      it "aliases the correct field when multiple rules are in a rule set when tree mode is not enabled" do
         if language == "d"
           write_grammar <<EOF
 <<
@@ -1430,11 +1552,14 @@ EOF
         expect(results.stdout).to match /first is foo1.*second is bar2/m
       end
 
-      it "does not free memory allocated for AST nodes" do
+      # D garbage collector was freeing memory for nodes when it should not
+      # have been. This was due to pointers being in the TreeNode that was
+      # allocated by C malloc() function instead of allocated using D.
+      it "does not free memory allocated for tree nodes" do
         ext = language == "cpp" ? "c" : language
-        write_grammar(File.read("spec/ast_node_memory_remains.#{ext}.propane"))
+        write_grammar(File.read("spec/tree_node_memory_remains.#{ext}.propane"))
         run_propane(language: language)
-        compile("spec/test_ast_node_memory_remains.#{language}", language: language)
+        compile("spec/test_tree_node_memory_remains.#{language}", language: language)
         results = run_test(language: language)
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
@@ -1460,9 +1585,9 @@ EOF
         expect(results.status).to eq 0
       end
 
-      it "allows multiple starting rules in AST mode" do
+      it "allows multiple starting rules in tree mode" do
     write_grammar <<EOF
-ast;
+tree;
 ptype int;
 token a << $$ = 1; >>
 token b << $$ = 2; >>
@@ -1475,23 +1600,19 @@ Bs -> b:b Bs:bs;
 start Start R Bs;
 EOF
         run_propane(language: language)
-        compile("spec/test_starting_rules_ast.#{language}", language: language)
+        compile("spec/test_starting_rules_tree.#{language}", language: language)
         results = run_test(language: language)
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
       end
 
       if %w[c cpp].include?(language)
-        it "allows a user function to free token node memory in AST mode" do
+        it "allows a user function to free token node memory in tree mode" do
           write_grammar <<EOF
-<<
+tree;
-static void free_token(Token * token)
+free_token_node <<
-{
+    free(${token.pvalue});
-    free(token->pvalue);
-}
 >>
-ast;
-free_token_node free_token;
 ptype int *;
 token a <<
   $$ = (int *)malloc(sizeof(int));
@@ -1504,12 +1625,444 @@ token b <<
 Start -> a:a b:b;
 EOF
           run_propane(language: language)
-          compile("spec/test_free_ast_token_node_memory.#{language}", language: language)
+          compile("spec/test_tree_delete_token_node_memory.#{language}", language: language)
           results = run_test(language: language)
           expect(results.stderr).to eq ""
           expect(results.status).to eq 0
         end
       end
+
+      it "executes code blocks associated with drop statements" do
+        if language == "d"
+          write_grammar <<EOF
+<<
+import std.stdio;
+>>
+drop /\\s+/;
+drop /#(.*)\\n/ <<
+  stderr.write("comment: ", match);
+>>
+token a;
+Start -> a;
+EOF
+        else
+          write_grammar <<EOF
+<<
+#include <stdio.h>
+#include <string.h>
+>>
+drop /\\s+/;
+drop /#(.*)\\n/ <<
+  fprintf(stderr, "comment: %.*s", (int)match_length, match);
+>>
+token a;
+Start -> a;
+EOF
+        end
+        run_propane(language: language)
+        compile("spec/test_drop_code_block.#{language}", language: language)
+        results = run_test(language: language)
+        expect(results.stderr).to match %r{comment: # comment 1\n.*comment: #    comment 2}m
+        expect(results.status).to eq 0
+      end
+
+      it "allows user-defined context fields" do
+        if language == "d"
+          write_grammar <<EOF
+context_user_fields <<
+    string comments;
+    uint acount;
+>>
+drop /\\s+/;
+drop /#(.*)\\n/ <<
+    ${context.comments} ~= match;
+>>
+token a <<
+    ${context.acount}++;
+>>
+Start -> As;
+As -> ;
+As -> a As;
+EOF
+        else
+          write_grammar <<EOF
+<<
+#include <string.h>
+#include <stdlib.h>
+>>
+context_user_fields <<
+    char * comments;
+    unsigned int acount;
+>>
+drop /\\s+/;
+drop /#(.*)\\n/ <<
+    size_t cur_len = 0u;
+    if (${context.comments} != NULL)
+        cur_len = strlen(${context.comments});
+    char * commentsnew = (char *)malloc(cur_len + match_length + 1);
+    if (${context.comments} != NULL)
+        memcpy(commentsnew, ${context.comments}, cur_len);
+    memcpy(&commentsnew[cur_len], match, match_length);
+    commentsnew[cur_len + match_length] = '\\0';
+    if (${context.comments} != NULL)
+    {
+        free(${context.comments});
+    }
+    ${context.comments} = commentsnew;
+>>
+token a <<
+    ${context.acount}++;
+>>
+Start -> As;
+As -> ;
+As -> a As;
+EOF
+        end
+        run_propane(language: language)
+        compile("spec/test_user_context_fields.#{language}", language: language)
+        results = run_test(language: language)
+        expect(results.stderr).to include %r{comments: # comment 1\n#    comment 2}
+        expect(results.stderr).to include %r{acount: 11\n}
+        expect(results.status).to eq 0
+      end
+
+      it "allows custom token user fields" do
+        if language == "d"
+          write_grammar <<EOF
+context_user_fields <<
+    string comments;
+>>
+token_user_fields <<
+    string comments;
+>>
+on_token_node <<
+    ${token.comments} = ${context.comments};
+    ${context.comments} = "";
+>>
+tree;
+drop /\\s+/;
+drop /#(.*)\\n/ <<
+    ${context.comments} ~= match;
+>>
+token id /\\w+/;
+Start -> IDs;
+IDs -> ;
+IDs -> id:id IDs;
+EOF
+        elsif language == "c"
+          write_grammar <<EOF
+<<
+#include <string.h>
+#include <stdlib.h>
+>>
+context_user_fields <<
+    char * comments;
+>>
+token_user_fields <<
+    char * comments;
+>>
+free_token_node <<
+    free(${token.comments});
+>>
+on_token_node <<
+    ${token.comments} = ${context.comments};
+    ${context.comments} = (char *)malloc(1);
+    ${context.comments}[0] = '\\0';
+>>
+tree;
+drop /\\s+/;
+drop /#(.*)\\n/ <<
+    size_t cur_len = 0u;
+    if (${context.comments} != NULL)
+        cur_len = strlen(${context.comments});
+    char * commentsnew = (char *)malloc(cur_len + match_length + 1);
+    if (${context.comments} != NULL)
+        memcpy(commentsnew, ${context.comments}, cur_len);
+    memcpy(&commentsnew[cur_len], match, match_length);
+    commentsnew[cur_len + match_length] = '\\0';
+    if (${context.comments} != NULL)
+    {
+        free(${context.comments});
+    }
+    ${context.comments} = commentsnew;
+>>
+token id /\\w+/;
+Start -> IDs;
+IDs -> ;
+IDs -> id:id IDs;
+EOF
+        else # C++
+          write_grammar <<EOF
+<<header
+#include <string>
+>>
+context_user_fields <<
+    std::string comments;
+>>
+token_user_fields <<
+    std::string comments;
+>>
+on_token_node <<
+    ${token.comments} = ${context.comments};
+    ${context.comments} = "";
+>>
+tree;
+drop /\\s+/;
+drop /#(.*)\\n/ <<
+    ${context.comments} += std::string((const char *)match, match_length);
+>>
+token id /\\w+/;
+Start -> IDs;
+IDs -> ;
+IDs -> id:id IDs;
+EOF
+        end
+        run_propane(language: language)
+        compile("spec/test_token_user_fields.#{language}", language: language)
+        results = run_test(language: language)
+        expect(results.status).to eq 0
+      end
+
+      it "allows a custom lex function" do
+        if language == "d"
+          write_grammar <<EOF
+<<
+private size_t mylexfn(p_context_t * context, p_token_info_t * out_token_info)
+{
+    static size_t count;
+    size_t result = P_SUCCESS;
+    if (count > 0)
+    {
+        out_token_info.token = TOKEN_a;
+        out_token_info.pvalue = p_value(count);
+        count--;
+    }
+    else
+    {
+        result = p_lex(context, out_token_info);
+        if (out_token_info.token == TOKEN_c)
+        {
+            count = 3;
+        }
+    }
+    return result;
+}
+>>
+
+ptype size_t;
+lex_fn mylexfn;
+
+token a << $$ = 7; >>
+token b << $$ = 8; >>
+token c << $$ = 9; >>
+Start -> << $$ = 0; >>
+Start -> Start ID << $$ = ($1 << 4) | $2; >>
+ID -> a << $$ = $1; >>
+ID -> b << $$ = $1; >>
+ID -> c << $$ = $1; >>
+EOF
+        else
+          write_grammar <<EOF
+<<
+static size_t mylexfn(p_context_t * context, p_token_info_t * out_token_info)
+{
+    static size_t count;
+    size_t result = P_SUCCESS;
+    if (count > 0)
+    {
+        out_token_info->token = TOKEN_a;
+        out_token_info->pvalue = p_value(count);
+        count--;
+    }
+    else
+    {
+        result = p_lex(context, out_token_info);
+        if (out_token_info->token == TOKEN_c)
+        {
+            count = 3;
+        }
+    }
+    return result;
+}
+>>
+
+ptype size_t;
+lex_fn mylexfn;
+
+token a << $$ = 7; >>
+token b << $$ = 8; >>
+token c << $$ = 9; >>
+Start -> << $$ = 0; >>
+Start -> Start ID << $$ = ($1 << 4) | $2; >>
+ID -> a << $$ = $1; >>
+ID -> b << $$ = $1; >>
+ID -> c << $$ = $1; >>
+EOF
+        end
+        run_propane(language: language)
+        compile("spec/test_custom_lex_fn.#{language}", language: language)
+        results = run_test(language: language)
+        expect(results.status).to eq 0
+      end
+
+      it "provides accessor functions to extract user values from a parser value" do
+        if language == "d"
+          write_grammar <<EOF
+ptype int;
+ptype float = float;
+ptype string = string;
+
+drop /\\s+/;
+token num /\\d+/ << $$ = 42; >>
+token flt (float) /f/ << $$ = 1.5; >>
+token str (string) /s/ << $$ = "hello"; >>
+Start -> num flt str;
+EOF
+        else
+          write_grammar <<EOF
+ptype int;
+ptype float = float;
+ptype string = char *;
+
+drop /\\s+/;
+token num /\\d+/ << $$ = 42; >>
+token flt (float) /f/ << $$ = 1.5; >>
+token str (string) /s/ << $$ = (char *)"hello"; >>
+Start -> num flt str;
+EOF
+        end
+        run_propane(language: language)
+        compile("spec/test_value_accessors.#{language}", language: language)
+        results = run_test(language: language)
+        expect(results.stderr).to eq ""
+        expect(results.status).to eq 0
+      end
+
+      it "allows accessing rule and component text positions" do
+        if language == "d"
+          write_grammar <<EOF
+<<
+import std.stdio;
+>>
+drop /\\s+/;
+token tok1;
+token tok2;
+token ident /[a-zA-Z_]\\w*/;
+token num /\\d+/;
+Num -> num;
+Start -> ident Num <<
+    writeln("ident start: ", ${1.position}.row, ", ", ${1.position}.col);
+    writeln("ident end: ", ${1.end_position}.row, ", ", ${1.end_position}.col);
+    writeln("Num start: ", ${2.position}.row, ", ", ${2.position}.col);
+    writeln("Num end: ", ${2.end_position}.row, ", ", ${2.end_position}.col);
+    writeln("Start start: ", ${$.position}.row, ", ", ${$.position}.col);
+    writeln("Start end: ", ${$.end_position}.row, ", ", ${$.end_position}.col);
+>>
+R -> Empty tok2 <<
+    writeln("Empty start: ", ${1.position}.row, ", ", ${1.position}.col);
+    writeln("Empty end: ", ${1.end_position}.row, ", ", ${1.end_position}.col);
+    writeln("tok2 start: ", ${2.position}.row, ", ", ${2.position}.col);
+    writeln("tok2 end: ", ${2.end_position}.row, ", ", ${2.end_position}.col);
+    writeln("R start: ", ${$.position}.row, ", ", ${$.position}.col);
+    writeln("R end: ", ${$.end_position}.row, ", ", ${$.end_position}.col);
+>>
+R -> tok1 Empty <<
+    writeln("tok1 start: ", ${1.position}.row, ", ", ${1.position}.col);
+    writeln("tok1 end: ", ${1.end_position}.row, ", ", ${1.end_position}.col);
+    writeln("Empty start: ", ${2.position}.row, ", ", ${2.position}.col);
+    writeln("Empty end: ", ${2.end_position}.row, ", ", ${2.end_position}.col);
+    writeln("R2 start: ", ${$.position}.row, ", ", ${$.position}.col);
+    writeln("R2 end: ", ${$.end_position}.row, ", ", ${$.end_position}.col);
+>>
+Empty -> ;
+Start -> R <<
+    writeln("StartR start: ", ${$.position}.row, ", ", ${$.position}.col);
+    writeln("StartR end: ", ${$.end_position}.row, ", ", ${$.end_position}.col);
+>>
+Start -> Empty <<
+    writeln("StartEmpty start: ", ${$.position}.row, ", ", ${$.position}.col);
+    writeln("StartEmpty end: ", ${$.end_position}.row, ", ", ${$.end_position}.col);
+>>
+EOF
+        else
+          write_grammar <<EOF
+<<
+#include <stdio.h>
+>>
+drop /\\s+/;
+token tok1;
+token tok2;
+token ident /[a-zA-Z_]\\w*/;
+token num /\\d+/;
+Num -> num;
+token pct /%/;
+Start -> ident Num <<
+    printf("ident start: %d, %d\\n", ${1.position}.row, ${1.position}.col);
+    printf("ident end: %d, %d\\n", ${1.end_position}.row, ${1.end_position}.col);
+    printf("Num start: %d, %d\\n", ${2.position}.row, ${2.position}.col);
+    printf("Num end: %d, %d\\n", ${2.end_position}.row, ${2.end_position}.col);
+    printf("Start start: %d, %d\\n", ${$.position}.row, ${$.position}.col);
+    printf("Start end: %d, %d\\n", ${$.end_position}.row, ${$.end_position}.col);
+>>
+R -> Empty tok2 <<
+    printf("Empty start: %d, %d\\n", ${1.position}.row, ${1.position}.col);
+    printf("Empty end: %d, %d\\n", ${1.end_position}.row, ${1.end_position}.col);
+    printf("tok2 start: %d, %d\\n", ${2.position}.row, ${2.position}.col);
+    printf("tok2 end: %d, %d\\n", ${2.end_position}.row, ${2.end_position}.col);
+    printf("R start: %d, %d\\n", ${$.position}.row, ${$.position}.col);
+    printf("R end: %d, %d\\n", ${$.end_position}.row, ${$.end_position}.col);
+>>
+R -> tok1 Empty <<
+    printf("tok1 start: %d, %d\\n", ${1.position}.row, ${1.position}.col);
+    printf("tok1 end: %d, %d\\n", ${1.end_position}.row, ${1.end_position}.col);
+    printf("Empty start: %d, %d\\n", ${2.position}.row, ${2.position}.col);
+    printf("Empty end: %d, %d\\n", ${2.end_position}.row, ${2.end_position}.col);
+    printf("R2 start: %d, %d\\n", ${$.position}.row, ${$.position}.col);
+    printf("R2 end: %d, %d\\n", ${$.end_position}.row, ${$.end_position}.col);
+>>
+Empty -> ;
+Start -> R <<
+    printf("StartR start: %d, %d\\n", ${$.position}.row, ${$.position}.col);
+    printf("StartR end: %d, %d\\n", ${$.end_position}.row, ${$.end_position}.col);
+>>
+Start -> Empty <<
+    printf("StartEmpty start: %d, %d\\n", ${$.position}.row, ${$.position}.col);
+    printf("StartEmpty end: %d, %d\\n", ${$.end_position}.row, ${$.end_position}.col);
+>>
+EOF
+        end
+        run_propane(language: language)
+        compile("spec/test_positions.#{language}", language: language)
+        results = run_test(language: language)
+        expect(results.stderr).to eq ""
+        expect(results.status).to eq 0
+        expect(results.stdout).to eq <<EOF
+ident start: 1, 5
+ident end: 1, 9
+Num start: 3, 9
+Num end: 3, 12
+Start start: 1, 5
+Start end: 3, 12
+
+Empty start: 0, 0
+Empty end: 0, 0
+tok2 start: 2, 2
+tok2 end: 2, 5
+R start: 2, 2
+R end: 2, 5
+StartR start: 2, 2
+StartR end: 2, 5
+
+tok1 start: 1, 3
+tok1 end: 1, 6
+Empty start: 0, 0
+Empty end: 0, 0
+R2 start: 1, 3
+R2 end: 1, 6
+StartR start: 1, 3
+StartR end: 1, 6
+EOF
+      end
     end
   end
 end

spec/test_basic_math_grammar.c

@@ -5,25 +5,29 @@
 int main()
 {
     char const * input = "1 + 2 * 3 + 4";
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((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";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((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";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((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";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((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;
-    p_context_init(&context, input);
+    context = p_context_new(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";
-    p_context_init(&context, input);
+    context = p_context_new(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";
-    p_context_init(&context, input);
+    context = p_context_new(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";
-    p_context_init(&context, input);
+    context = p_context_new(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_custom_lex_fn.c

@@ -0,0 +1,15 @@
+#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

@@ -0,0 +1,18 @@
+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_drop_code_block.c

@@ -0,0 +1,15 @@
+#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

@@ -0,0 +1,16 @@
+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_error_positions.c

@@ -5,38 +5,43 @@
 int main()
 {
     char const * input = "a 42";
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((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";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((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";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((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";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((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";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((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;
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
 
     input = "a\n123\na  a";
-    p_context_init(&context, input);
+    context = p_context_new(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";
-    p_context_init(&context, input);
+    context = p_context_new(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";
-    p_context_init(&context, input);
+    context = p_context_new(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";
-    p_context_init(&context, input);
+    context = p_context_new(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_field_aliases.c

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

spec/test_lexer.c

@@ -38,73 +38,75 @@ int main()
 
     p_token_info_t token_info;
     char const * input = "5 + 4 * \n677 + 567";
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((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);
 
-    p_context_init(&context, (uint8_t const *)"", 0u);
+    context = p_context_new((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;
-    p_context_init(&context, input);
+    context = p_context_new(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));
 
-    p_context_init(&context, "");
+    context = p_context_new("");
-    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_match_text.c

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

spec/test_lexer_modes.c

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

spec/test_lexer_multiple_modes.c

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

spec/test_lexer_positions.c

@@ -0,0 +1,50 @@
+#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

@@ -0,0 +1,42 @@
+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_result_value.c

@@ -5,15 +5,17 @@
 int main()
 {
     char const * input = "x";
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((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";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((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;
-    p_context_init(&context, input);
+    context = p_context_new(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`;
-    p_context_init(&context, input);
+    context = p_context_new(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_unknown_character.c

@@ -5,14 +5,16 @@
 int main()
 {
     char const * input = "x";
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((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";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((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;
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert(p_parse(&context) == P_UNEXPECTED_INPUT);
+    assert(p_parse(context) == P_UNEXPECTED_INPUT);
 
     input = `123`;
-    p_context_init(&context, input);
+    context = p_context_new(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_match_backslashes.c

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

spec/test_multiple_parsers.c

@@ -6,14 +6,16 @@
 int main()
 {
     char const * input1 = "a\n1";
-    myp1_context_t context1;
+    myp1_context_t * context1;
-    myp1_context_init(&context1, (uint8_t const *)input1, strlen(input1));
+    context1 = myp1_context_new((uint8_t const *)input1, strlen(input1));
-    assert(myp1_parse(&context1) == MYP1_SUCCESS);
+    assert(myp1_parse(context1) == MYP1_SUCCESS);
+    myp1_context_delete(context1);
 
     char const * input2 = "bcb";
-    myp2_context_t context2;
+    myp2_context_t * context2;
-    myp2_context_init(&context2, (uint8_t const *)input2, strlen(input2));
+    context2 = myp2_context_new((uint8_t const *)input2, strlen(input2));
-    assert(myp2_parse(&context2) == MYP2_SUCCESS);
+    assert(myp2_parse(context2) == MYP2_SUCCESS);
+    myp2_context_delete(context2);
 
     return 0;
 }

spec/test_multiple_parsers.d

@@ -10,12 +10,12 @@ int main()
 unittest
 {
     string input1 = "a\n1";
-    myp1_context_t context1;
+    myp1_context_t * context1;
-    myp1_context_init(&context1, input1);
+    context1 = myp1_context_new(input1);
-    assert(myp1_parse(&context1) == MYP1_SUCCESS);
+    assert(myp1_parse(context1) == MYP1_SUCCESS);
 
     string input2 = "bcb";
-    myp2_context_t context2;
+    myp2_context_t * context2;
-    myp2_context_init(&context2, input2);
+    context2 = myp2_context_new(input2);
-    assert(myp2_parse(&context2) == MYP2_SUCCESS);
+    assert(myp2_parse(context2) == MYP2_SUCCESS);
 }

spec/test_named_optional_rule_component_tree.c

@@ -6,10 +6,10 @@
 int main()
 {
     char const * input = "b";
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    Start * start = p_result(&context);
+    Start * start = p_result(context);
     assert(start->a == NULL);
     assert(start->pToken2 != NULL);
     assert_eq(TOKEN_b, start->pToken2->token);
@@ -17,12 +17,13 @@ int main()
     assert(start->pR == NULL);
     assert(start->r == NULL);
 
-    p_free_ast(start);
+    p_tree_delete(start);
+    p_context_delete(context);
 
     input = "abcd";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    start = p_result(&context);
+    start = p_result(context);
     assert(start->a != NULL);
     assert_eq(TOKEN_a, start->pToken1->token);
     assert(start->pToken2 != NULL);
@@ -33,18 +34,20 @@ int main()
     assert(start->pR == start->r);
     assert_eq(TOKEN_c, start->pR->pToken1->token);
 
-    p_free_ast(start);
+    p_tree_delete(start);
+    p_context_delete(context);
 
     input = "bdc";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    start = p_result(&context);
+    start = p_result(context);
     assert(start->a == NULL);
     assert(start->pToken2 != NULL);
     assert(start->r != NULL);
     assert_eq(TOKEN_d, start->pR->pToken1->token);
 
-    p_free_ast(start);
+    p_tree_delete(start);
+    p_context_delete(context);
 
     return 0;
 }

spec/test_named_optional_rule_component_tree.d

@@ -10,10 +10,9 @@ int main()
 unittest
 {
     string input = "b";
-    p_context_t context;
+    p_context_t * context = p_context_new(input);
-    p_context_init(&context, input);
+    assert(p_parse(context) == P_SUCCESS);
-    assert(p_parse(&context) == P_SUCCESS);
+    Start * start = p_result(context);
-    Start * start = p_result(&context);
     assert(start.pToken1 is null);
     assert(start.pToken2 !is null);
     assert_eq(TOKEN_b, start.pToken2.token);
@@ -21,10 +20,12 @@ unittest
     assert(start.pR is null);
     assert(start.r is null);
 
+    p_tree_delete(start);
+
     input = "abcd";
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    start = p_result(&context);
+    start = p_result(context);
     assert(start.pToken1 != null);
     assert_eq(TOKEN_a, start.pToken1.token);
     assert(start.pToken2 != null);
@@ -35,12 +36,16 @@ unittest
     assert(start.pR == start.r);
     assert_eq(TOKEN_c, start.pR.pToken1.token);
 
+    p_tree_delete(start);
+
     input = "bdc";
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    start = p_result(&context);
+    start = p_result(context);
     assert(start.pToken1 is null);
     assert(start.pToken2 !is null);
     assert(start.pR !is null);
     assert_eq(TOKEN_d, start.pR.pToken1.token);
+
+    p_tree_delete(start);
 }

spec/test_optional_rule_component.c

@@ -5,17 +5,20 @@
 int main()
 {
     char const * input = "b";
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
+    p_context_delete(context);
 
     input = "abcd";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
+    p_context_delete(context);
 
     input = "abdc";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((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_optional_rule_component.d

@@ -9,15 +9,15 @@ int main()
 unittest
 {
     string input = "b";
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
 
     input = "abcd";
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
 
     input = "abdc";
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
 }

spec/test_optional_rule_component_tree.c

@@ -6,22 +6,23 @@
 int main()
 {
     char const * input = "b";
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    Start * start = p_result(&context);
+    Start * start = p_result(context);
     assert(start->pToken1 == NULL);
     assert(start->pToken2 != NULL);
     assert_eq(TOKEN_b, start->pToken2->token);
     assert(start->pR3 == NULL);
     assert(start->pR == NULL);
 
-    p_free_ast(start);
+    p_tree_delete(start);
+    p_context_delete(context);
 
     input = "abcd";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    start = p_result(&context);
+    start = p_result(context);
     assert(start->pToken1 != NULL);
     assert_eq(TOKEN_a, start->pToken1->token);
     assert(start->pToken2 != NULL);
@@ -30,18 +31,20 @@ int main()
     assert(start->pR == start->pR3);
     assert_eq(TOKEN_c, start->pR->pToken1->token);
 
-    p_free_ast(start);
+    p_tree_delete(start);
+    p_context_delete(context);
 
     input = "bdc";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    start = p_result(&context);
+    start = p_result(context);
     assert(start->pToken1 == NULL);
     assert(start->pToken2 != NULL);
     assert(start->pR != NULL);
     assert_eq(TOKEN_d, start->pR->pToken1->token);
 
-    p_free_ast(start);
+    p_tree_delete(start);
+    p_context_delete(context);
 
     return 0;
 }

spec/test_optional_rule_component_tree.d

@@ -10,20 +10,21 @@ int main()
 unittest
 {
     string input = "b";
-    p_context_t context;
+    p_context_t * context = p_context_new(input);
-    p_context_init(&context, input);
+    assert(p_parse(context) == P_SUCCESS);
-    assert(p_parse(&context) == P_SUCCESS);
+    Start * start = p_result(context);
-    Start * start = p_result(&context);
     assert(start.pToken1 is null);
     assert(start.pToken2 !is null);
     assert_eq(TOKEN_b, start.pToken2.token);
     assert(start.pR3 is null);
     assert(start.pR is null);
 
+    p_tree_delete(start);
+
     input = "abcd";
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    start = p_result(&context);
+    start = p_result(context);
     assert(start.pToken1 != null);
     assert_eq(TOKEN_a, start.pToken1.token);
     assert(start.pToken2 != null);
@@ -32,12 +33,16 @@ unittest
     assert(start.pR == start.pR3);
     assert_eq(TOKEN_c, start.pR.pToken1.token);
 
+    p_tree_delete(start);
+
     input = "bdc";
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    start = p_result(&context);
+    start = p_result(context);
     assert(start.pToken1 is null);
     assert(start.pToken2 !is null);
     assert(start.pR !is null);
     assert_eq(TOKEN_d, start.pR.pToken1.token);
+
+    p_tree_delete(start);
 }

spec/test_parser_identical_rules_lookahead.c

@@ -5,13 +5,15 @@
 int main()
 {
     char const * input = "aba";
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
+    p_context_delete(context);
 
     input = "abb";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((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_parser_identical_rules_lookahead.d

@@ -9,11 +9,11 @@ int main()
 unittest
 {
     string input = "aba";
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
 
     input = "abb";
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
 }

spec/test_parser_rule_from_multiple_states.c

@@ -5,20 +5,23 @@
 int main()
 {
     char const * input = "a";
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((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 == 2);
+    assert(p_position(context).col == 2);
-    assert(context.token == TOKEN___EOF);
+    assert(context->token == TOKEN___EOF);
+    p_context_delete(context);
 
     input = "a b";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
+    p_context_delete(context);
 
     input = "bb";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((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_parser_rule_from_multiple_states.d

@@ -9,17 +9,17 @@ int main()
 unittest
 {
     string input = "a";
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert(p_parse(&context) == P_UNEXPECTED_TOKEN);
+    assert(p_parse(context) == P_UNEXPECTED_TOKEN);
-    assert(p_position(&context) == p_position_t(1, 2));
+    assert(p_position(context) == p_position_t(1, 2));
     assert(context.token == TOKEN___EOF);
 
     input = "a b";
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
 
     input = "bb";
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
 }

spec/test_parser_rule_user_code.c

@@ -5,9 +5,10 @@
 int main()
 {
     char const * input = "ab";
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((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_parser_rule_user_code.d

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

spec/test_parser_user_code_tree.c

@@ -0,0 +1,39 @@
+#include "testparser.h"
+#include <assert.h>
+#include <string.h>
+#include "testutils.h"
+
+int main()
+{
+    char const * input = "ab";
+    p_context_t * context;
+    context = p_context_new((uint8_t const *)input, strlen(input));
+    assert_eq(P_SUCCESS, p_parse(context));
+
+    /* The parser user code recorded values accessed via $$, $1, and $2 while
+     * the tree node for the Start rule was being formed. */
+    assert_eq(3, context->start_n_fields);
+    assert_eq(11, context->start_a_value);
+    assert_eq(11, context->a_value);
+    assert_eq(22, context->b_value);
+    assert_eq(TOKEN_b, context->b_token);
+
+    /* The empty-matched rule C has a null $$ tree node, and its field in the
+     * Start node is null as well. */
+    assert_eq(1, context->c_is_null);
+    assert_eq(1, context->c_field_is_null);
+
+    /* Field aliases reference the same component tree nodes as the positional
+     * references. */
+    assert_eq(11, context->alias_a_value);
+    assert_eq(22, context->alias_b_value);
+
+    Start * start = p_result(context);
+    assert(start->pA != NULL);
+    assert(start->pB != NULL);
+    assert(start->pC == NULL);
+    p_tree_delete(start);
+    p_context_delete(context);
+
+    return 0;
+}

spec/test_parser_user_code_tree.d

@@ -0,0 +1,39 @@
+import testparser;
+import std.stdio;
+import testutils;
+
+int main()
+{
+    return 0;
+}
+
+unittest
+{
+    string input = "ab";
+    p_context_t * context = p_context_new(input);
+    assert_eq(P_SUCCESS, p_parse(context));
+
+    /* The parser user code recorded values accessed via $$, $1, and $2 while
+     * the tree node for the Start rule was being formed. */
+    assert_eq(3, context.start_n_fields);
+    assert_eq(11, context.start_a_value);
+    assert_eq(11, context.a_value);
+    assert_eq(22, context.b_value);
+    assert_eq(TOKEN_b, context.b_token);
+
+    /* The empty-matched rule C has a null $$ tree node, and its field in the
+     * Start node is null as well. */
+    assert_eq(1, context.c_is_null);
+    assert_eq(1, context.c_field_is_null);
+
+    /* Field aliases reference the same component tree nodes as the positional
+     * references. */
+    assert_eq(11, context.alias_a_value);
+    assert_eq(22, context.alias_b_value);
+
+    Start * start = p_result(context);
+    assert(start.pA !is null);
+    assert(start.pB !is null);
+    assert(start.pC is null);
+    p_tree_delete(start);
+}

spec/test_parsing_json.c

@@ -6,51 +6,58 @@
 int main()
 {
     char const * input = "";
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
+    p_context_delete(context);
 
     input = "{}";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    assert(p_result(&context)->id == JSON_OBJECT);
+    assert(p_result(context)->id == JSON_OBJECT);
+    p_context_delete(context);
 
     input = "[]";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    assert(p_result(&context)->id == JSON_ARRAY);
+    assert(p_result(context)->id == JSON_ARRAY);
+    p_context_delete(context);
 
     input = "-45.6";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    assert(p_result(&context)->id == JSON_NUMBER);
+    assert(p_result(context)->id == JSON_NUMBER);
-    assert(p_result(&context)->number == -45.6);
+    assert(p_result(context)->number == -45.6);
+    p_context_delete(context);
 
     input = "2E-2";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    assert(p_result(&context)->id == JSON_NUMBER);
+    assert(p_result(context)->id == JSON_NUMBER);
-    assert(p_result(&context)->number == 0.02);
+    assert(p_result(context)->number == 0.02);
+    p_context_delete(context);
 
     input = "{\"hi\":true}";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    JSONValue * o = p_result(&context);
+    JSONValue * o = p_result(context);
     assert(o->id == JSON_OBJECT);
     assert_eq(1, o->object.size);
     assert(strcmp(o->object.entries[0].name, "hi") == 0);
     assert(o->object.entries[0].value->id == JSON_TRUE);
+    p_context_delete(context);
 
     input = "{\"ff\": false, \"nn\": null}";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    o = p_result(&context);
+    o = p_result(context);
     assert(o->id == JSON_OBJECT);
     assert_eq(2, o->object.size);
     assert(strcmp(o->object.entries[0].name, "ff") == 0);
     assert(o->object.entries[0].value->id == JSON_FALSE);
     assert(strcmp(o->object.entries[1].name, "nn") == 0);
     assert(o->object.entries[1].value->id == JSON_NULL);
+    p_context_delete(context);
 
     return 0;
 }

spec/test_parsing_json.d

@@ -10,45 +10,45 @@ int main()
 unittest
 {
     string input = ``;
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
 
     input = `{}`;
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    assert(cast(JSONObject)p_result(&context));
+    assert(cast(JSONObject)p_result(context));
 
     input = `[]`;
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    assert(cast(JSONArray)p_result(&context));
+    assert(cast(JSONArray)p_result(context));
 
     input = `-45.6`;
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    assert(cast(JSONNumber)p_result(&context));
+    assert(cast(JSONNumber)p_result(context));
-    assert((cast(JSONNumber)p_result(&context)).value == -45.6);
+    assert((cast(JSONNumber)p_result(context)).value == -45.6);
 
     input = `2E-2`;
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    assert(cast(JSONNumber)p_result(&context));
+    assert(cast(JSONNumber)p_result(context));
-    assert((cast(JSONNumber)p_result(&context)).value == 0.02);
+    assert((cast(JSONNumber)p_result(context)).value == 0.02);
 
     input = `{"hi":true}`;
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    assert(cast(JSONObject)p_result(&context));
+    assert(cast(JSONObject)p_result(context));
-    JSONObject o = cast(JSONObject)p_result(&context);
+    JSONObject o = cast(JSONObject)p_result(context);
     assert(o.value["hi"]);
     assert(cast(JSONTrue)o.value["hi"]);
 
     input = `{"ff": false, "nn": null}`;
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    assert(cast(JSONObject)p_result(&context));
+    assert(cast(JSONObject)p_result(context));
-    o = cast(JSONObject)p_result(&context);
+    o = cast(JSONObject)p_result(context);
     assert(o.value["ff"]);
     assert(cast(JSONFalse)o.value["ff"]);
     assert(o.value["nn"]);

spec/test_parsing_lists.c

@@ -5,20 +5,23 @@
 int main()
 {
     char const * input = "a";
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((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 = "";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    assert(p_result(&context) == 0u);
+    assert(p_result(context) == 0u);
+    p_context_delete(context);
 
     input = "aaaaaaaaaaaaaaaa";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    assert(p_result(&context) == 16u);
+    assert(p_result(context) == 16u);
+    p_context_delete(context);
 
     return 0;
 }

spec/test_parsing_lists.d

@@ -9,18 +9,18 @@ int main()
 unittest
 {
     string input = "a";
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    assert(p_result(&context) == 1u);
+    assert(p_result(context) == 1u);
 
     input = "";
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    assert(p_result(&context) == 0u);
+    assert(p_result(context) == 0u);
 
     input = "aaaaaaaaaaaaaaaa";
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    assert(p_result(&context) == 16u);
+    assert(p_result(context) == 16u);
 }

spec/test_pattern.c

@@ -6,15 +6,17 @@
 int main()
 {
     char const * input = "abcdef";
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((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 = "defabcdef";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((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_pattern.d

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

spec/test_positions.c

@@ -0,0 +1,29 @@
+#include "testparser.h"
+#include <assert.h>
+#include <string.h>
+#include <stdio.h>
+
+int main()
+{
+    char const * input = "    Hello\n\n        4200\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);
+
+    printf("\n");
+
+    input = "\n tok2";
+    context = p_context_new((uint8_t const *)input, strlen(input));
+    assert(p_parse(context) == P_SUCCESS);
+    p_context_delete(context);
+
+    printf("\n");
+
+    input = "  tok1";
+    context = p_context_new((uint8_t const *)input, strlen(input));
+    assert(p_parse(context) == P_SUCCESS);
+    p_context_delete(context);
+
+    return 0;
+}

spec/test_positions.d

@@ -0,0 +1,27 @@
+import testparser;
+import std.stdio;
+
+int main()
+{
+    return 0;
+}
+
+unittest
+{
+    string input = "    Hello\n\n        4200\n";
+    p_context_t * context;
+    context = p_context_new(input);
+    assert(p_parse(context) == P_SUCCESS);
+
+    writeln();
+
+    input = "\n tok2";
+    context = p_context_new(input);
+    assert(p_parse(context) == P_SUCCESS);
+
+    writeln();
+
+    input = "  tok1";
+    context = p_context_new(input);
+    assert(p_parse(context) == P_SUCCESS);
+}

spec/test_return_token_from_pattern.c

@@ -5,9 +5,10 @@
 int main()
 {
     char const * input = "defghidef";
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((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_return_token_from_pattern.d

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

spec/test_start_rule_ast.c

@@ -1,19 +0,0 @@
-#include "testparser.h"
-#include <assert.h>
-#include <string.h>
-#include "testutils.h"
-
-int main()
-{
-    char const * input = "hi";
-    p_context_t context;
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
-    assert_eq(P_SUCCESS, p_parse(&context));
-    Top * top = p_result(&context);
-    assert(top->pToken != NULL);
-    assert_eq(TOKEN_hi, top->pToken->token);
-
-    p_free_ast(top);
-
-    return 0;
-}

spec/test_start_rule_tree.c

@@ -0,0 +1,20 @@
+#include "testparser.h"
+#include <assert.h>
+#include <string.h>
+#include "testutils.h"
+
+int main()
+{
+    char const * input = "hi";
+    p_context_t * context;
+    context = p_context_new((uint8_t const *)input, strlen(input));
+    assert_eq(P_SUCCESS, p_parse(context));
+    Top * top = p_result(context);
+    assert(top->pToken != NULL);
+    assert_eq(TOKEN_hi, top->pToken->token);
+
+    p_tree_delete(top);
+    p_context_delete(context);
+
+    return 0;
+}

spec/test_start_rule_tree.d

@@ -10,10 +10,10 @@ int main()
 unittest
 {
     string input = "hi";
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert_eq(P_SUCCESS, p_parse(&context));
+    assert_eq(P_SUCCESS, p_parse(context));
-    Top * top = p_result(&context);
+    Top * top = p_result(context);
     assert(top.pToken !is null);
     assert_eq(TOKEN_hi, top.pToken.token);
 }

spec/test_starting_rules.c

@@ -6,22 +6,25 @@
 int main()
 {
     char const * input = "bbbb";
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    int result = p_result(&context);
+    int result = p_result(context);
     assert_eq(8, result);
+    p_context_delete(context);
 
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert(p_parse_Bs(&context) == P_SUCCESS);
+    assert(p_parse_Bs(context) == P_SUCCESS);
-    result = p_result_Bs(&context);
+    result = p_result_Bs(context);
     assert_eq(8, result);
+    p_context_delete(context);
 
     input = "c";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert(p_parse_R(&context) == P_SUCCESS);
+    assert(p_parse_R(context) == P_SUCCESS);
-    result = p_result_R(&context);
+    result = p_result_R(context);
     assert_eq(3, result);
+    p_context_delete(context);
 
     return 0;
 }

spec/test_starting_rules.d

@@ -10,20 +10,20 @@ int main()
 unittest
 {
     string input = "bbbb";
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    int result = p_result(&context);
+    int result = p_result(context);
     assert(result == 8);
 
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert(p_parse_Bs(&context) == P_SUCCESS);
+    assert(p_parse_Bs(context) == P_SUCCESS);
-    result = p_result_Bs(&context);
+    result = p_result_Bs(context);
     assert(result == 8);
 
     input = "c";
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert(p_parse_R(&context) == P_SUCCESS);
+    assert(p_parse_R(context) == P_SUCCESS);
-    result = p_result_R(&context);
+    result = p_result_R(context);
     assert(result == 3);
 }

spec/test_starting_rules_ast.c

@@ -1,37 +0,0 @@
-#include "testparser.h"
-#include <assert.h>
-#include <string.h>
-#include "testutils.h"
-
-int main()
-{
-    char const * input = "bbbb";
-    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_not_null(start->bs);
-    assert_not_null(start->bs->b);
-    assert_not_null(start->bs->bs->b);
-    assert_not_null(start->bs->bs->bs->b);
-    assert_not_null(start->bs->bs->bs->bs->b);
-    p_free_ast(start);
-
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
-    assert(p_parse_Bs(&context) == P_SUCCESS);
-    Bs * bs = p_result_Bs(&context);
-    assert_not_null(bs->b);
-    assert_not_null(bs->bs->b);
-    assert_not_null(bs->bs->bs->b);
-    assert_not_null(bs->bs->bs->bs->b);
-    p_free_ast_Bs(bs);
-
-    input = "c";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
-    assert(p_parse_R(&context) == P_SUCCESS);
-    R * r = p_result_R(&context);
-    assert_not_null(r->c);
-    p_free_ast_R(r);
-
-    return 0;
-}

spec/test_starting_rules_ast.d

@@ -1,36 +0,0 @@
-import testparser;
-import std.stdio;
-import testutils;
-
-int main()
-{
-    return 0;
-}
-
-unittest
-{
-    string input = "bbbb";
-    p_context_t context;
-    p_context_init(&context, input);
-    assert(p_parse(&context) == P_SUCCESS);
-    Start * start = p_result(&context);
-    assert(start.bs);
-    assert(start.bs.b);
-    assert(start.bs.bs.b);
-    assert(start.bs.bs.bs.b);
-    assert(start.bs.bs.bs.bs.b);
-
-    p_context_init(&context, input);
-    assert(p_parse_Bs(&context) == P_SUCCESS);
-    Bs * bs = p_result_Bs(&context);
-    assert(bs.b);
-    assert(bs.bs.b);
-    assert(bs.bs.bs.b);
-    assert(bs.bs.bs.bs.b);
-
-    input = "c";
-    p_context_init(&context, input);
-    assert(p_parse_R(&context) == P_SUCCESS);
-    R * r = p_result_R(&context);
-    assert(r.c);
-}

spec/test_starting_rules_tree.c

@@ -0,0 +1,40 @@
+#include "testparser.h"
+#include <assert.h>
+#include <string.h>
+#include "testutils.h"
+
+int main()
+{
+    char const * input = "bbbb";
+    p_context_t * context;
+    context = p_context_new((uint8_t const *)input, strlen(input));
+    assert(p_parse(context) == P_SUCCESS);
+    Start * start = p_result(context);
+    assert_not_null(start->bs);
+    assert_not_null(start->bs->b);
+    assert_not_null(start->bs->bs->b);
+    assert_not_null(start->bs->bs->bs->b);
+    assert_not_null(start->bs->bs->bs->bs->b);
+    p_tree_delete(start);
+    p_context_delete(context);
+
+    context = p_context_new((uint8_t const *)input, strlen(input));
+    assert(p_parse_Bs(context) == P_SUCCESS);
+    Bs * bs = p_result_Bs(context);
+    assert_not_null(bs->b);
+    assert_not_null(bs->bs->b);
+    assert_not_null(bs->bs->bs->b);
+    assert_not_null(bs->bs->bs->bs->b);
+    p_tree_delete_Bs(bs);
+    p_context_delete(context);
+
+    input = "c";
+    context = p_context_new((uint8_t const *)input, strlen(input));
+    assert(p_parse_R(context) == P_SUCCESS);
+    R * r = p_result_R(context);
+    assert_not_null(r->c);
+    p_tree_delete_R(r);
+    p_context_delete(context);
+
+    return 0;
+}

spec/test_starting_rules_tree.d

@@ -0,0 +1,41 @@
+import testparser;
+import std.stdio;
+import testutils;
+
+int main()
+{
+    return 0;
+}
+
+unittest
+{
+    string input = "bbbb";
+    p_context_t * context = p_context_new(input);
+    assert(p_parse(context) == P_SUCCESS);
+    Start * start = p_result(context);
+    assert(start.bs);
+    assert(start.bs.b);
+    assert(start.bs.bs.b);
+    assert(start.bs.bs.bs.b);
+    assert(start.bs.bs.bs.bs.b);
+
+    p_tree_delete(start);
+
+    context = p_context_new(input);
+    assert(p_parse_Bs(context) == P_SUCCESS);
+    Bs * bs = p_result_Bs(context);
+    assert(bs.b);
+    assert(bs.bs.b);
+    assert(bs.bs.bs.b);
+    assert(bs.bs.bs.bs.b);
+
+    p_tree_delete_Bs(bs);
+
+    input = "c";
+    context = p_context_new(input);
+    assert(p_parse_R(context) == P_SUCCESS);
+    R * r = p_result_R(context);
+    assert(r.c);
+
+    p_tree_delete_R(r);
+}

spec/test_token_user_fields.c

@@ -0,0 +1,46 @@
+#include "testparser.h"
+#include <assert.h>
+#include <stdio.h>
+#include <string.h>
+#include <stdlib.h>
+
+int main()
+{
+    char const * input =
+        "# c1\n"
+        "#  c2\n"
+        "\n"
+        "first\n"
+        "\n   \n  \n"
+        "  # s1\n"
+        "   #   s2\n"
+        "second\n";
+    p_context_t * context;
+    context = p_context_new((uint8_t const *)input, strlen(input));
+    assert(p_parse(context) == P_SUCCESS);
+    Start * start = p_result(context);
+    assert(start->pIDs);
+    assert(start->pIDs->id);
+#ifdef __cplusplus
+    assert(start->pIDs->id->comments == "# c1\n#  c2\n");
+#else
+    assert(start->pIDs->id->comments);
+    assert(strcmp(start->pIDs->id->comments, "# c1\n#  c2\n") == 0);
+#endif
+    assert(start->pIDs->pIDs);
+    assert(start->pIDs->pIDs->id);
+#ifdef __cplusplus
+    assert(start->pIDs->pIDs->id->comments == "# s1\n#   s2\n");
+#else
+    assert(start->pIDs->pIDs->id->comments);
+    assert(strcmp(start->pIDs->pIDs->id->comments, "# s1\n#   s2\n") == 0);
+#endif
+
+#ifndef __cplusplus
+    free(context->comments);
+#endif
+    p_context_delete(context);
+    p_tree_delete(start);
+
+    return 0;
+}

spec/test_token_user_fields.d

@@ -0,0 +1,31 @@
+import testparser;
+import std.stdio;
+
+int main()
+{
+    return 0;
+}
+
+unittest
+{
+    string input =
+        "# c1\n" ~
+        "#  c2\n" ~
+        "\n" ~
+        "first\n" ~
+        "\n   \n  \n" ~
+        "  # s1\n" ~
+        "   #   s2\n" ~
+        "second\n";
+    p_context_t * context = p_context_new(input);
+    assert(p_parse(context) == P_SUCCESS);
+    Start * start = p_result(context);
+    assert(start.pIDs);
+    assert(start.pIDs.id);
+    assert(start.pIDs.id.comments == "# c1\n#  c2\n");
+    assert(start.pIDs.pIDs);
+    assert(start.pIDs.pIDs.id);
+    assert(start.pIDs.pIDs.id.comments == "# s1\n#   s2\n");
+
+    p_tree_delete(start);
+}

spec/test_tree.c

@@ -6,10 +6,10 @@
 int main()
 {
     char const * input = "a, ((b)), b";
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert_eq(P_SUCCESS, p_parse(&context));
+    assert_eq(P_SUCCESS, p_parse(context));
-    Start * start = p_result(&context);
+    Start * start = p_result(context);
     assert(start->pItems1 != NULL);
     assert(start->pItems != NULL);
     Items * items = start->pItems;
@@ -33,20 +33,22 @@ int main()
     assert_eq(22, itemsmore->pItem->pToken1->pvalue);
     assert(itemsmore->pItemsMore == NULL);
 
-    p_free_ast(start);
+    p_tree_delete(start);
+    p_context_delete(context);
 
     input = "";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert_eq(P_SUCCESS, p_parse(&context));
+    assert_eq(P_SUCCESS, p_parse(context));
-    start = p_result(&context);
+    start = p_result(context);
     assert(start->pItems == NULL);
 
-    p_free_ast(start);
+    p_tree_delete(start);
+    p_context_delete(context);
 
     input = "2 1";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert_eq(P_SUCCESS, p_parse(&context));
+    assert_eq(P_SUCCESS, p_parse(context));
-    start = p_result(&context);
+    start = p_result(context);
     assert(start->pItems != NULL);
     assert(start->pItems->pItem != NULL);
     assert(start->pItems->pItem->pDual != NULL);
@@ -55,7 +57,8 @@ int main()
     assert(start->pItems->pItem->pDual->pTwo2 == NULL);
     assert(start->pItems->pItem->pDual->pOne1 == NULL);
 
-    p_free_ast(start);
+    p_tree_delete(start);
+    p_context_delete(context);
 
     return 0;
 }

spec/test_tree.d

@@ -10,10 +10,9 @@ int main()
 unittest
 {
     string input = "a, ((b)), b";
-    p_context_t context;
+    p_context_t * context = p_context_new(input);
-    p_context_init(&context, input);
+    assert_eq(P_SUCCESS, p_parse(context));
-    assert_eq(P_SUCCESS, p_parse(&context));
+    Start * start = p_result(context);
-    Start * start = p_result(&context);
     assert(start.pItems1 !is null);
     assert(start.pItems !is null);
     Items * items = start.pItems;
@@ -37,16 +36,20 @@ unittest
     assert_eq(22, itemsmore.pItem.pToken1.pvalue);
     assert(itemsmore.pItemsMore is null);
 
+    p_tree_delete(start);
+
     input = "";
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert_eq(P_SUCCESS, p_parse(&context));
+    assert_eq(P_SUCCESS, p_parse(context));
-    start = p_result(&context);
+    start = p_result(context);
     assert(start.pItems is null);
 
+    p_tree_delete(start);
+
     input = "2 1";
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert_eq(P_SUCCESS, p_parse(&context));
+    assert_eq(P_SUCCESS, p_parse(context));
-    start = p_result(&context);
+    start = p_result(context);
     assert(start.pItems !is null);
     assert(start.pItems.pItem !is null);
     assert(start.pItems.pItem.pDual !is null);
@@ -54,4 +57,6 @@ unittest
     assert(start.pItems.pItem.pDual.pOne2 !is null);
     assert(start.pItems.pItem.pDual.pTwo2 is null);
     assert(start.pItems.pItem.pDual.pOne1 is null);
+
+    p_tree_delete(start);
 }

spec/test_tree_delete_token_node_memory.c

@@ -6,14 +6,15 @@
 int main()
 {
     char const * input = "ab";
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert_eq(P_SUCCESS, p_parse(&context));
+    assert_eq(P_SUCCESS, p_parse(context));
-    Start * start = p_result(&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);
+    p_tree_delete(start);
+    p_context_delete(context);
 }

spec/test_tree_field_aliases.c

@@ -6,16 +6,17 @@
 int main()
 {
     char const * input = "\na\nb\nc";
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    Start * start = p_result(&context);
+    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);
+    p_tree_delete(start);
+    p_context_delete(context);
 
     return 0;
 }

spec/test_tree_field_aliases.d

@@ -10,12 +10,13 @@ int main()
 unittest
 {
     string input = "\na\nb\nc";
-    p_context_t context;
+    p_context_t * context = p_context_new(input);
-    p_context_init(&context, input);
+    assert(p_parse(context) == P_SUCCESS);
-    assert(p_parse(&context) == P_SUCCESS);
+    Start * start = p_result(context);
-    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_tree_delete(start);
 }

spec/test_tree_invalid_positions.c

@@ -6,10 +6,10 @@
 int main()
 {
     char const * input = "\na\n  bb ccc";
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    Start * start = p_result(&context);
+    Start * start = p_result(context);
 
     assert_eq(2, start->pT1->pToken->position.row);
     assert_eq(1, start->pT1->pToken->position.col);
@@ -30,12 +30,13 @@ int main()
     assert_eq(3, start->end_position.row);
     assert_eq(8, start->end_position.col);
 
-    p_free_ast(start);
+    p_tree_delete(start);
+    p_context_delete(context);
 
     input = "a\nbb";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(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);
@@ -56,12 +57,13 @@ int main()
     assert_eq(2, start->end_position.row);
     assert_eq(2, start->end_position.col);
 
-    p_free_ast(start);
+    p_tree_delete(start);
+    p_context_delete(context);
 
     input = "a\nc\nc";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(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);
@@ -82,12 +84,13 @@ int main()
     assert_eq(3, start->end_position.row);
     assert_eq(1, start->end_position.col);
 
-    p_free_ast(start);
+    p_tree_delete(start);
+    p_context_delete(context);
 
     input = "a";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(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);
@@ -104,7 +107,8 @@ int main()
     assert_eq(1, start->end_position.row);
     assert_eq(1, start->end_position.col);
 
-    p_free_ast(start);
+    p_tree_delete(start);
+    p_context_delete(context);
 
     return 0;
 }

spec/test_tree_invalid_positions.d

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

spec/test_tree_node_memory_remains.c

@@ -369,11 +369,11 @@ int main(int argc, char * argv[])
         {"size_t_to_ulong", TOKEN_ulong},
         {"main", TOKEN_int},
     };
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, (const uint8_t *)input, strlen(input));
+    context = p_context_new((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 = pmod->pModuleItems;
     PFunctionDefinition ** pfds;
     size_t n_pfds = 0u;
@@ -412,7 +412,8 @@ int main(int argc, char * argv[])
     }
 
     free(pfds);
-    p_free_ast(pmod);
+    p_tree_delete(pmod);
+    p_context_delete(context);
 
     return 0;
 }

spec/test_tree_node_memory_remains.d

@@ -374,11 +374,11 @@ def main() -> int
         Expected("size_t_to_ulong", TOKEN_ulong),
         Expected("main", TOKEN_int),
     ];
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, input);
+    context = p_context_new(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;
     PFunctionDefinition *[] pfds;
     while (pmis !is null)
@@ -405,4 +405,5 @@ 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_tree_delete(pmod);
 }

spec/test_tree_ps.c

@@ -6,10 +6,10 @@
 int main()
 {
     char const * input = "a, ((b)), b";
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert_eq(P_SUCCESS, p_parse(&context));
+    assert_eq(P_SUCCESS, p_parse(context));
-    PStartS * start = p_result(&context);
+    PStartS * start = p_result(context);
     assert(start->pItems1 != NULL);
     assert(start->pItems != NULL);
     PItemsS * items = start->pItems;
@@ -33,20 +33,22 @@ int main()
     assert_eq(22, itemsmore->pItem->pToken1->pvalue);
     assert(itemsmore->pItemsMore == NULL);
 
-    p_free_ast(start);
+    p_tree_delete(start);
+    p_context_delete(context);
 
     input = "";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert_eq(P_SUCCESS, p_parse(&context));
+    assert_eq(P_SUCCESS, p_parse(context));
-    start = p_result(&context);
+    start = p_result(context);
     assert(start->pItems == NULL);
 
-    p_free_ast(start);
+    p_tree_delete(start);
+    p_context_delete(context);
 
     input = "2 1";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert_eq(P_SUCCESS, p_parse(&context));
+    assert_eq(P_SUCCESS, p_parse(context));
-    start = p_result(&context);
+    start = p_result(context);
     assert(start->pItems != NULL);
     assert(start->pItems->pItem != NULL);
     assert(start->pItems->pItem->pDual != NULL);
@@ -55,7 +57,8 @@ int main()
     assert(start->pItems->pItem->pDual->pTwo2 == NULL);
     assert(start->pItems->pItem->pDual->pOne1 == NULL);
 
-    p_free_ast(start);
+    p_tree_delete(start);
+    p_context_delete(context);
 
     return 0;
 }

spec/test_tree_ps.d

@@ -10,10 +10,9 @@ int main()
 unittest
 {
     string input = "a, ((b)), b";
-    p_context_t context;
+    p_context_t * context = p_context_new(input);
-    p_context_init(&context, input);
+    assert_eq(P_SUCCESS, p_parse(context));
-    assert_eq(P_SUCCESS, p_parse(&context));
+    PStartS * start = p_result(context);
-    PStartS * start = p_result(&context);
     assert(start.pItems1 !is null);
     assert(start.pItems !is null);
     PItemsS * items = start.pItems;
@@ -37,16 +36,20 @@ unittest
     assert_eq(22, itemsmore.pItem.pToken1.pvalue);
     assert(itemsmore.pItemsMore is null);
 
+    p_tree_delete(start);
+
     input = "";
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert_eq(P_SUCCESS, p_parse(&context));
+    assert_eq(P_SUCCESS, p_parse(context));
-    start = p_result(&context);
+    start = p_result(context);
     assert(start.pItems is null);
 
+    p_tree_delete(start);
+
     input = "2 1";
-    p_context_init(&context, input);
+    context = p_context_new(input);
-    assert_eq(P_SUCCESS, p_parse(&context));
+    assert_eq(P_SUCCESS, p_parse(context));
-    start = p_result(&context);
+    start = p_result(context);
     assert(start.pItems !is null);
     assert(start.pItems.pItem !is null);
     assert(start.pItems.pItem.pDual !is null);
@@ -54,4 +57,6 @@ unittest
     assert(start.pItems.pItem.pDual.pOne2 !is null);
     assert(start.pItems.pItem.pDual.pTwo2 is null);
     assert(start.pItems.pItem.pDual.pOne1 is null);
+
+    p_tree_delete(start);
 }

spec/test_tree_token_positions.c

@@ -6,10 +6,10 @@
 int main()
 {
     char const * input = "abbccc";
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
-    Start * start = p_result(&context);
+    Start * start = p_result(context);
 
     assert_eq(1, start->pT1->pToken->position.row);
     assert_eq(1, start->pT1->pToken->position.col);
@@ -43,12 +43,13 @@ int main()
     assert_eq(1, start->end_position.row);
     assert_eq(6, start->end_position.col);
 
-    p_free_ast(start);
+    p_tree_delete(start);
+    p_context_delete(context);
 
     input = "\n\n  bb\nc\ncc\n\n     a";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(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);
@@ -82,7 +83,8 @@ int main()
     assert_eq(7, start->end_position.row);
     assert_eq(6, start->end_position.col);
 
-    p_free_ast(start);
+    p_tree_delete(start);
+    p_context_delete(context);
 
     return 0;
 }

spec/test_tree_token_positions.d

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

spec/test_user_code.c

@@ -6,15 +6,17 @@
 int main()
 {
     char const * input = "abcdef";
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((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 = "abcabcdef";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((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_user_code.d

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

spec/test_user_context_fields.c

@@ -0,0 +1,20 @@
+#include "testparser.h"
+#include <assert.h>
+#include <stdio.h>
+#include <string.h>
+#include <stdlib.h>
+
+int main()
+{
+    char const * input = "aaa\n\n\na\n    # comment 1\na  a    aa\n\naa\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);
+
+    fprintf(stderr, "comments: %s", context->comments);
+    fprintf(stderr, "acount: %u\n", context->acount);
+    free(context->comments);
+    p_context_delete(context);
+
+    return 0;
+}

spec/test_user_context_fields.d

@@ -0,0 +1,19 @@
+import testparser;
+import std.stdio;
+import testutils;
+
+int main()
+{
+    return 0;
+}
+
+unittest
+{
+    string input = "aaa\n\n\na\n    # comment 1\na  a    aa\n\naa\n#    comment 2\na\n";
+    p_context_t * context;
+    context = p_context_new(input);
+    assert(p_parse(context) == P_SUCCESS);
+
+    stderr.writeln("comments: ", context.comments);
+    stderr.writeln("acount: ", context.acount);
+}

spec/test_user_terminate.c

@@ -6,14 +6,16 @@
 int main()
 {
     char const * input = "aacc";
-    p_context_t context;
+    p_context_t * context;
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_parse(context) == P_SUCCESS);
+    p_context_delete(context);
 
     input = "abc";
-    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    context = p_context_new((uint8_t const *)input, strlen(input));
-    assert(p_parse(&context) == P_USER_TERMINATED);
+    assert(p_parse(context) == P_USER_TERMINATED);
-    assert(p_user_terminate_code(&context) == 4200);
+    assert(p_user_terminate_code(context) == 4200);
+    p_context_delete(context);
 
     return 0;
 }