Test multiple starting rules

.rspec

@@ -1,3 +1,2 @@
---format documentation
 --color
 --require spec_helper

Gemfile.lock

@@ -2,39 +2,41 @@ GEM
   remote: https://rubygems.org/
   specs:
     base64 (0.3.0)
-    date (3.4.1)
+    date (3.5.1)
     diff-lcs (1.6.2)
     docile (1.4.1)
-    erb (5.0.2)
+    erb (6.0.1)
-    psych (5.2.6)
+    psych (5.3.1)
       date
       stringio
-    rake (13.3.0)
+    rake (13.3.1)
-    rdoc (6.14.2)
+    rdoc (7.1.0)
       erb
       psych (>= 4.0.0)
+      tsort
     redcarpet (3.6.1)
-    rspec (3.13.1)
+    rspec (3.13.2)
       rspec-core (~> 3.13.0)
       rspec-expectations (~> 3.13.0)
       rspec-mocks (~> 3.13.0)
-    rspec-core (3.13.5)
+    rspec-core (3.13.6)
       rspec-support (~> 3.13.0)
     rspec-expectations (3.13.5)
       diff-lcs (>= 1.2.0, < 2.0)
       rspec-support (~> 3.13.0)
-    rspec-mocks (3.13.5)
+    rspec-mocks (3.13.7)
       diff-lcs (>= 1.2.0, < 2.0)
       rspec-support (~> 3.13.0)
-    rspec-support (3.13.4)
+    rspec-support (3.13.6)
     simplecov (0.22.0)
       docile (~> 1.1)
       simplecov-html (~> 0.11)
       simplecov_json_formatter (~> 0.1)
     simplecov-html (0.13.2)
     simplecov_json_formatter (0.1.4)
-    stringio (3.1.7)
+    stringio (3.2.0)
     syntax (1.2.2)
+    tsort (0.2.0)
 
 PLATFORMS
   ruby

README.md

@@ -6,7 +6,7 @@ Propane is a LALR Parser Generator (LPG) which:
   * generates a built-in lexer to tokenize input
   * supports UTF-8 lexer inputs
   * generates a table-driven shift/reduce parser to parse input in linear time
-  * targets C or D language outputs
+  * targets C, C++, or D language outputs
   * optionally supports automatic full AST generation
   * is MIT-licensed
   * is distributable as a standalone Ruby script

Rakefile.rb

@@ -1,5 +1,7 @@
 require "rake/clean"
 require "rspec/core/rake_task"
+require "simplecov"
+require "stringio"
 
 CLEAN.include %w[spec/run gen .yardoc yard coverage dist]
 
@@ -12,6 +14,18 @@ RSpec::Core::RakeTask.new(:spec, :example_pattern) do |task, args|
     task.rspec_opts = %W[-e "#{args.example_pattern}" -f documentation]
   end
 end
+task :spec do |task, args|
+  unless ENV["dist_specs"]
+    original_stdout = $stdout
+    sio = StringIO.new
+    $stdout = sio
+    SimpleCov.collate Dir["coverage/.resultset.json"]
+    $stdout = original_stdout
+    sio.string.lines.each do |line|
+      $stdout.write(line) unless line =~ /Coverage report generated for/
+    end
+  end
+end
 
 # dspec task is useful to test the distributable release script, but is not
 # useful for coverage information.

assets/parser.c.erb

@@ -695,11 +695,13 @@ typedef struct
 } state_value_t;
 
 /** Common AST node structure. */
-typedef struct
+typedef struct ASTNode_s
 {
     <%= @grammar.prefix %>position_t position;
     <%= @grammar.prefix %>position_t end_position;
-    void * fields[];
+    uint16_t n_fields;
+    uint8_t is_token;
+    struct ASTNode_s * fields[];
 } ASTNode;
 
 /** Parser shift table. */
@@ -805,7 +807,7 @@ static void state_values_stack_push(state_values_stack_t * stack)
     if (current_length >= current_capacity)
     {
         size_t const new_capacity = current_capacity * 2u;
-        state_value_t * new_entries = malloc(new_capacity * sizeof(state_value_t));
+        state_value_t * new_entries = (state_value_t *)malloc(new_capacity * sizeof(state_value_t));
         memcpy(new_entries, stack->entries, current_length * sizeof(state_value_t));
         free(stack->entries);
         stack->capacity = new_capacity;
@@ -922,6 +924,8 @@ static size_t check_reduce(size_t state_id, <%= @grammar.prefix %>token_t token)
  *
  * @param context
  *   Lexer/parser context structure.
+ * @start_state_id
+ *   ID of the state in which to start.
  *
  * @retval P_SUCCESS
  *   The parser successfully matched the input text. The parse result value
@@ -934,7 +938,7 @@ static size_t check_reduce(size_t state_id, <%= @grammar.prefix %>token_t token)
  * @reval P_UNEXPECTED_INPUT
  *   Input text does not match any lexer pattern.
  */
-size_t <%= @grammar.prefix %>parse(<%= @grammar.prefix %>context_t * context)
+static size_t parse_from(<%= @grammar.prefix %>context_t * context, size_t start_state_id)
 {
     <%= @grammar.prefix %>token_info_t token_info;
     <%= @grammar.prefix %>token_t token = INVALID_TOKEN_ID;
@@ -947,6 +951,7 @@ size_t <%= @grammar.prefix %>parse(<%= @grammar.prefix %>context_t * context)
 <% end %>
     state_values_stack_init(&statevalues);
     state_values_stack_push(&statevalues);
+    state_values_stack_index(&statevalues, -1)->state_id = start_state_id;
     size_t result;
     for (;;)
     {
@@ -972,7 +977,7 @@ size_t <%= @grammar.prefix %>parse(<%= @grammar.prefix %>context_t * context)
             {
                 /* Successful parse. */
 <% if @grammar.ast %>
-                context->parse_result = (<%= @grammar.ast_prefix %><%= @grammar.start_rule %><%= @grammar.ast_suffix %> *)state_values_stack_index(&statevalues, -1)->ast_node;
+                context->parse_result = state_values_stack_index(&statevalues, -1)->ast_node;
 <% else %>
                 context->parse_result = state_values_stack_index(&statevalues, -1)->pvalue;
 <% end %>
@@ -989,9 +994,11 @@ size_t <%= @grammar.prefix %>parse(<%= @grammar.prefix %>context_t * context)
             {
                 /* We shifted a token, mark it consumed. */
 <% if @grammar.ast %>
-                <%= @grammar.ast_prefix %>Token<%= @grammar.ast_suffix %> * token_ast_node = malloc(sizeof(<%= @grammar.ast_prefix %>Token<%= @grammar.ast_suffix %>));
+                <%= @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 %>));
                 token_ast_node->position = token_info.position;
                 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;
@@ -1028,31 +1035,30 @@ size_t <%= @grammar.prefix %>parse(<%= @grammar.prefix %>context_t * context)
             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;
-                ASTNode * node = (ASTNode *)malloc(sizeof(ASTNode) + n_fields * sizeof(void *));
+                size_t bytes = sizeof(ASTNode) + n_fields * sizeof(void *);
+                ASTNode * node = (ASTNode *)malloc(bytes);
+                memset(node, 0, bytes);
                 node->position = INVALID_POSITION;
                 node->end_position = INVALID_POSITION;
-                for (size_t i = 0; i < n_fields; i++)
+                node->n_fields = n_fields;
-                {
-                    node->fields[i] = NULL;
-                }
                 if (parser_reduce_table[reduce_index].rule_set_node_field_index_map == NULL)
                 {
                     for (size_t i = 0; i < parser_reduce_table[reduce_index].n_states; i++)
                     {
-                        node->fields[i] = state_values_stack_index(&statevalues, -(int)parser_reduce_table[reduce_index].n_states + (int)i)->ast_node;
+                        node->fields[i] = (ASTNode *)state_values_stack_index(&statevalues, -(int)parser_reduce_table[reduce_index].n_states + (int)i)->ast_node;
                     }
                 }
                 else
                 {
                     for (size_t i = 0; i < parser_reduce_table[reduce_index].n_states; i++)
                     {
-                        node->fields[parser_reduce_table[reduce_index].rule_set_node_field_index_map[i]] = state_values_stack_index(&statevalues, -(int)parser_reduce_table[reduce_index].n_states + (int)i)->ast_node;
+                        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;
                     }
                 }
                 bool position_found = false;
                 for (size_t i = 0; i < n_fields; i++)
                 {
-                    ASTNode * child = (ASTNode *)node->fields[i];
+                    ASTNode * child = node->fields[i];
                     if ((child != NULL) && <%= @grammar.prefix %>position_valid(child->position))
                     {
                         if (!position_found)
@@ -1074,6 +1080,7 @@ size_t <%= @grammar.prefix %>parse(<%= @grammar.prefix %>context_t * context)
             memset(&reduced_parser_value2, 0, sizeof(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)
             {
+                state_values_stack_free(&statevalues);
                 return P_USER_TERMINATED;
             }
             reduced_parser_value = reduced_parser_value2;
@@ -1097,6 +1104,19 @@ size_t <%= @grammar.prefix %>parse(<%= @grammar.prefix %>context_t * context)
     return result;
 }
 
+size_t <%= @grammar.prefix %>parse(<%= @grammar.prefix %>context_t * context)
+{
+    return parse_from(context, 0u);
+}
+
+<% @grammar.start_rules.each_with_index do |start_rule, i| %>
+
+size_t <%= @grammar.prefix %>parse_<%= start_rule %>(<%= @grammar.prefix %>context_t * context)
+{
+    return parse_from(context, <%= i %>u);
+}
+<% end %>
+
 /**
  * Get the parse result value.
  *
@@ -1106,17 +1126,28 @@ size_t <%= @grammar.prefix %>parse(<%= @grammar.prefix %>context_t * context)
  * @return Parse result value.
  */
 <% if @grammar.ast %>
-<%= @grammar.ast_prefix %><%= @grammar.start_rule %><%= @grammar.ast_suffix %> * <%= @grammar.prefix %>result(<%= @grammar.prefix %>context_t * context)
+<%= @grammar.ast_prefix %><%= @grammar.start_rules[0] %><%= @grammar.ast_suffix %> * <%= @grammar.prefix %>result(<%= @grammar.prefix %>context_t * context)
+{
+    return (<%= @grammar.ast_prefix %><%= @grammar.start_rules[0] %><%= @grammar.ast_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)
+{
+    return (<%= @grammar.ast_prefix %><%= start_rule %><%= @grammar.ast_suffix %> *) context->parse_result;
+}
+<%   end %>
 <% else %>
 <%= start_rule_type[1] %> <%= @grammar.prefix %>result(<%= @grammar.prefix %>context_t * context)
-<% end %>
 {
-<% if @grammar.ast %>
-    return context->parse_result;
-<% else %>
     return context->parse_result.v_<%= start_rule_type[0] %>;
-<% end %>
 }
+<%   @grammar.start_rules.each_with_index do |start_rule, i| %>
+<%= start_rule_type(i)[1] %> <%= @grammar.prefix %>result_<%= start_rule %>(<%= @grammar.prefix %>context_t * context)
+{
+    return context->parse_result.v_<%= start_rule_type(i)[0] %>;
+}
+<%   end %>
+<% end %>
 
 /**
  * Get the current text input position.
@@ -1153,3 +1184,45 @@ size_t <%= @grammar.prefix %>user_terminate_code(<%= @grammar.prefix %>context_t
 {
     return context->token;
 }
+<% if @grammar.ast %>
+
+static void free_ast_node(ASTNode * node)
+{
+    if (node->is_token)
+    {
+<%   if @grammar.free_token_node %>
+        <%= @grammar.free_token_node %>((<%= @grammar.ast_prefix %>Token<%= @grammar.ast_suffix %> *) node);
+<%   end %>
+        /* TODO: free value_t */
+    }
+    else if (node->n_fields > 0u)
+    {
+        for (size_t i = 0u; i < node->n_fields; i++)
+        {
+            if (node->fields[i] != NULL)
+            {
+                free_ast_node(node->fields[i]);
+            }
+        }
+    }
+    free(node);
+}
+
+/**
+ * Free all AST node memory.
+ */
+void <%= @grammar.prefix %>free_ast(<%= @grammar.ast_prefix %><%= @grammar.start_rules[0] %><%= @grammar.ast_suffix %> * ast)
+{
+    free_ast_node((ASTNode *)ast);
+}
+<%   @grammar.start_rules.each_with_index do |start_rule, i| %>
+
+/**
+ * Free all AST node memory.
+ */
+void <%= @grammar.prefix %>free_ast_<%= start_rule %>(<%= @grammar.ast_prefix %><%= start_rule %><%= @grammar.ast_suffix %> * ast)
+{
+    free_ast_node((ASTNode *)ast);
+}
+<%   end %>
+<% end %>

assets/parser.d.erb

@@ -173,7 +173,7 @@ public struct <%= @grammar.prefix %>context_t
 
     /** Parse result value. */
 <% if @grammar.ast %>
-    <%= @grammar.ast_prefix %><%= @grammar.start_rule %><%= @grammar.ast_suffix %> * parse_result;
+    void * parse_result;
 <% else %>
     <%= @grammar.prefix %>value_t parse_result;
 <% end %>
@@ -985,6 +985,8 @@ private size_t check_reduce(size_t state_id, <%= @grammar.prefix %>token_t token
  *
  * @param context
  *   Lexer/parser context structure.
+ * @start_state_id
+ *   ID of the state in which to start.
  *
  * @retval P_SUCCESS
  *   The parser successfully matched the input text. The parse result value
@@ -997,11 +999,12 @@ private size_t check_reduce(size_t state_id, <%= @grammar.prefix %>token_t token
  * @reval P_UNEXPECTED_INPUT
  *   Input text does not match any lexer pattern.
  */
-public size_t <%= @grammar.prefix %>parse(<%= @grammar.prefix %>context_t * context)
+private size_t parse_from(<%= @grammar.prefix %>context_t * context, size_t start_state_id)
 {
     <%= @grammar.prefix %>token_info_t token_info;
     <%= @grammar.prefix %>token_t token = INVALID_TOKEN_ID;
     state_value_t[] statevalues = new state_value_t[](1);
+    statevalues[0].state_id = start_state_id;
     size_t reduced_rule_set = INVALID_ID;
 <% if @grammar.ast %>
     void * reduced_parser_node;
@@ -1031,7 +1034,7 @@ public size_t <%= @grammar.prefix %>parse(<%= @grammar.prefix %>context_t * cont
             {
                 /* Successful parse. */
 <% if @grammar.ast %>
-                context.parse_result = cast(<%= @grammar.ast_prefix %><%= @grammar.start_rule %><%= @grammar.ast_suffix %> *)statevalues[$-1].ast_node;
+                context.parse_result = statevalues[$-1].ast_node;
 <% else %>
                 context.parse_result = statevalues[$-1].pvalue;
 <% end %>
@@ -1147,6 +1150,19 @@ public size_t <%= @grammar.prefix %>parse(<%= @grammar.prefix %>context_t * cont
     }
 }
 
+public size_t <%= @grammar.prefix %>parse(<%= @grammar.prefix %>context_t * context)
+{
+    return parse_from(context, 0u);
+}
+
+<% @grammar.start_rules.each_with_index do |start_rule, i| %>
+
+public size_t <%= @grammar.prefix %>parse_<%= start_rule %>(<%= @grammar.prefix %>context_t * context)
+{
+    return parse_from(context, <%= i %>u);
+}
+<% end %>
+
 /**
  * Get the parse result value.
  *
@@ -1156,17 +1172,28 @@ public size_t <%= @grammar.prefix %>parse(<%= @grammar.prefix %>context_t * cont
  * @return Parse result value.
  */
 <% if @grammar.ast %>
-public <%= @grammar.ast_prefix %><%= @grammar.start_rule %><%= @grammar.ast_suffix %> * <%= @grammar.prefix %>result(<%= @grammar.prefix %>context_t * context)
+public <%= @grammar.ast_prefix %><%= @grammar.start_rules[0] %><%= @grammar.ast_suffix %> * <%= @grammar.prefix %>result(<%= @grammar.prefix %>context_t * context)
+{
+    return cast(<%= @grammar.ast_prefix %><%= @grammar.start_rules[0] %><%= @grammar.ast_suffix %> *)context.parse_result;
+}
+<%   @grammar.start_rules.each_with_index do |start_rule, i| %>
+public <%= @grammar.ast_prefix %><%= start_rule %><%= @grammar.ast_suffix %> * <%= @grammar.prefix %>result_<%= start_rule %>(<%= @grammar.prefix %>context_t * context)
+{
+    return cast(<%= @grammar.ast_prefix %><%= start_rule %><%= @grammar.ast_suffix %> *)context.parse_result;
+}
+<%   end %>
 <% else %>
 public <%= start_rule_type[1] %> <%= @grammar.prefix %>result(<%= @grammar.prefix %>context_t * context)
-<% end %>
 {
-<% if @grammar.ast %>
-    return context.parse_result;
-<% else %>
     return context.parse_result.v_<%= start_rule_type[0] %>;
-<% end %>
 }
+<%   @grammar.start_rules.each_with_index do |start_rule, i| %>
+public <%= start_rule_type(i)[1] %> <%= @grammar.prefix %>result_<%= start_rule %>(<%= @grammar.prefix %>context_t * context)
+{
+    return context.parse_result.v_<%= start_rule_type(i)[0] %>;
+}
+<%   end %>
+<% end %>
 
 /**
  * Get the current text input position.

assets/parser.h.erb

@@ -75,9 +75,11 @@ typedef union
 /** AST node types. @{ */
 typedef struct <%= @grammar.ast_prefix %>Token<%= @grammar.ast_suffix %>
 {
-    /* ASTNode fields must be present in the same order here. */
+    <% # ASTNode fields must be present in the same order here. # %>
     <%= @grammar.prefix %>position_t position;
     <%= @grammar.prefix %>position_t end_position;
+    uint16_t n_fields;
+    uint8_t is_token;
     <%= @grammar.prefix %>token_t token;
     <%= @grammar.prefix %>value_t pvalue;
 } <%= @grammar.ast_prefix %>Token<%= @grammar.ast_suffix %>;
@@ -93,8 +95,11 @@ struct <%= name %>;
 <%     next if rule_set.optional? %>
 typedef struct <%= @grammar.ast_prefix %><%= name %><%= @grammar.ast_suffix %>
 {
+    <% # ASTNode fields must be present in the same order here. # %>
     <%= @grammar.prefix %>position_t position;
     <%= @grammar.prefix %>position_t end_position;
+    uint16_t n_fields;
+    uint8_t is_token;
 <%     rule_set.ast_fields.each do |fields| %>
     union
     {
@@ -157,7 +162,7 @@ typedef struct
 
     /** Parse result value. */
 <% if @grammar.ast %>
-    <%= @grammar.ast_prefix %><%= @grammar.start_rule %><%= @grammar.ast_suffix %> * parse_result;
+    void * parse_result;
 <% else %>
     <%= @grammar.prefix %>value_t parse_result;
 <% end %>
@@ -184,11 +189,27 @@ size_t <%= @grammar.prefix %>decode_code_point(uint8_t const * input, size_t inp
 size_t <%= @grammar.prefix %>lex(<%= @grammar.prefix %>context_t * context, <%= @grammar.prefix %>token_info_t * out_token_info);
 
 size_t <%= @grammar.prefix %>parse(<%= @grammar.prefix %>context_t * context);
+<%   @grammar.start_rules.each_with_index do |start_rule, i| %>
+size_t <%= @grammar.prefix %>parse_<%= start_rule %>(<%= @grammar.prefix %>context_t * context);
+<%   end %>
 
 <% if @grammar.ast %>
-<%= @grammar.ast_prefix %><%= @grammar.start_rule %><%= @grammar.ast_suffix %> * <%= @grammar.prefix %>result(<%= @grammar.prefix %>context_t * context);
+<%= @grammar.ast_prefix %><%= @grammar.start_rules[0] %><%= @grammar.ast_suffix %> * <%= @grammar.prefix %>result(<%= @grammar.prefix %>context_t * context);
+<%   @grammar.start_rules.each_with_index do |start_rule, i| %>
+<%= @grammar.ast_prefix %><%= start_rule %><%= @grammar.ast_suffix %> * <%= @grammar.prefix %>result_<%= start_rule %>(<%= @grammar.prefix %>context_t * context);
+<%   end %>
 <% else %>
 <%= start_rule_type[1] %> <%= @grammar.prefix %>result(<%= @grammar.prefix %>context_t * context);
+<%   @grammar.start_rules.each_with_index do |start_rule, i| %>
+<%= start_rule_type(i)[1] %> <%= @grammar.prefix %>result_<%= start_rule %>(<%= @grammar.prefix %>context_t * context);
+<%   end %>
+<% end %>
+
+<% if @grammar.ast %>
+void <%= @grammar.prefix %>free_ast(<%= @grammar.ast_prefix %><%= @grammar.start_rules[0] %><%= @grammar.ast_suffix %> * ast);
+<%   @grammar.start_rules.each_with_index do |start_rule, i| %>
+void <%= @grammar.prefix %>free_ast_<%= start_rule %>(<%= @grammar.ast_prefix %><%= start_rule %><%= @grammar.ast_suffix %> * ast);
+<%   end %>
 <% end %>
 
 <%= @grammar.prefix %>position_t <%= @grammar.prefix %>position(<%= @grammar.prefix %>context_t * context);

doc/user_guide.md

@@ -13,7 +13,7 @@ Propane is a LALR Parser Generator (LPG) which:
   * generates a built-in lexer to tokenize input
   * supports UTF-8 lexer inputs
   * generates a table-driven shift/reduce parser to parse input in linear time
-  * targets C or D language outputs
+  * targets C, C++, or D language outputs
   * optionally supports automatic full AST generation
   * tracks input text start and end positions for all matched tokens/rules
   * is MIT-licensed
@@ -326,6 +326,36 @@ assert(itemsmore.pItem.pItem.pItem !is null);
 assert(itemsmore.pItem.pItem.pItem.pToken1 !is null);
 ```
 
+## Freeing user-allocated memory in token node `pvalue`: the `free_token_node` statement
+
+If user lexer code block allocates memory to store in a token node's `pvalue`,
+the `free_token_node` grammar statement can be used to specify the name of a
+function which will be called during the `p_free_ast()` call to free the memory
+associated with a token node.
+
+Example:
+
+```
+<<
+static void free_token(Token * token)
+{
+    free(token->pvalue);
+}
+>>
+ast;
+free_token_node free_token;
+ptype int *;
+token a <<
+  $$ = (int *)malloc(sizeof(int));
+  *$$ = 1;
+>>
+token b <<
+  $$ = (int *)malloc(sizeof(int));
+  *$$ = 2;
+>>
+Start -> a:a b:b;
+```
+
 ##> Specifying tokens - the `token` statement
 
 The `token` statement allows defining a lexer token and a pattern to match that
@@ -708,6 +738,22 @@ Example:
 start MyStartRule;
 ```
 
+Multiple start rules can be specified, either with multiple `start` statements
+or one `start` statement listing multiple start rules.
+Example:
+
+```
+start Module ModuleItem Statement Expression;
+```
+
+When multiple start rules are specified, multiple `p_parse_*()` functions,
+`p_result_*()`, and `p_free_ast_*()` functions (in AST mode) are generated.
+A default `p_parse()`, `p_result()`, `p_free_ast()` are generated corresponding
+to the first start rule.
+Additionally, each start rule causes the generation of another version of each
+of these functions, for example `p_parse_Statement()`, `p_result_Statement()`,
+and `p_free_ast_Statement()`.
+
 ##> Specifying the parser module name - the `module` statement
 
 The `module` statement can be used to specify the module name for a generated
@@ -751,7 +797,7 @@ Some example uses of this functionality could be to:
 
   * Detect integer overflow when lexing an integer literal constant.
   * Detect and report an error as soon as possible during parsing before continuing to parse any more of the input.
-  * Determine whether parsing should stop and instead be performed using a different parser version.
+  * Determine whether parsing should stop and instead be retried using a different parser version.
 
 To terminate parsing from a lexer or parser user code block, use the
 `$terminate(code)` function, passing an integer expression argument.
@@ -787,7 +833,7 @@ Propane generates the following result code constants:
 * `P_EOF`: The lexer reached the end of the input string.
 * `P_USER_TERMINATED`: A parser user code block has requested to terminate the parser.
 
-Result codes are returned by the functions `p_decode_code_point()`, `p_lex()`, and `p_parse()`.
+Result codes are returned by the API functions `p_decode_code_point()`, `p_lex()`, and `p_parse()`.
 
 ##> Types
 
@@ -807,7 +853,7 @@ A pointer to this instance is passed to the generated functions.
 
 ### `p_position_t`
 
-The `p_position_t` structure contains two fields `row` and `col`.
+The `p_position_t` structure contains two fields: `row` and `col`.
 These fields contain the 1-based row and column describing a parser position.
 
 For D targets, the `p_position_t` structure can be checked for validity by
@@ -817,6 +863,16 @@ For C targets, the `p_position_t` structure can be checked for validity by
 calling `p_position_valid(pos)` where `pos` is a `p_position_t` structure
 instance.
 
+### `p_token_info_t`
+
+The `p_token_info_t` structure contains the following fields:
+
+* `position` (`p_position_t`) holds the text position of the first code point in the token.
+* `end_position` (`p_position_t`) holds the text position of the last code point in the token.
+* `length` (`size_t`) holds the number of input bytes used by the token.
+* `token` (`p_token_t`) holds the token ID of the lexed token
+* `pvalue` (`p_value_t`) holds the parser value associated with the token.
+
 ### AST Node Types
 
 If AST generation mode is enabled, a structure type for each rule will be
@@ -927,6 +983,44 @@ p_context_t context;
 p_context_init(&context, input);
 ```
 
+### `p_lex`
+
+The `p_lex()` function is the main entry point to the lexer.
+It is normally called automatically by the generated parser to retrieve the
+next input token for the parser and does not need to be called by the user.
+However, the user may initialize a context and call `p_lex()` to use the
+generated lexer in a standalone mode.
+
+Example:
+
+```
+p_context_t context;
+p_context_init(&context, input, input_length);
+p_token_info_t token_info;
+size_t result = p_lex(&context, &token_info);
+switch (result)
+{
+case P_DECODE_ERROR:
+    /* UTF-8 decode error */
+    break;
+case P_UNEXPECTED_INPUT:
+    /* Input text does not match any lexer pattern. */
+    break;
+case P_USER_TERMINATED:
+    /* Lexer user code block requested to terminate the lexer. */
+    break;
+case P_SUCCESS:
+    /*
+     * token_info.position holds the text position of the first code point in the token.
+     * token_info.end_position holds the text position of the last code point in the token.
+     * token_info.length holds the number of input bytes used by the token.
+     * token_info.token holds the token ID of the lexed token
+     * token_info.pvalue holds the parser value associated with the token.
+     */
+    break;
+}
+```
+
 ### `p_parse`
 
 The `p_parse()` function is the main entry point to the parser.
@@ -940,6 +1034,16 @@ p_context_init(&context, input, input_length);
 size_t result = p_parse(&context);
 ```
 
+When multiple start rules are specified, a separate parse function is generated
+for each which starts parsing at the given rule.
+For example, if `Statement` is specified as a start rule:
+
+```
+size_t result = p_parse_Statement(&context);
+```
+
+In this case, the parser will start parsing with the `Statement` rule.
+
 ### `p_position_valid`
 
 The `p_position_valid()` function is only generated for C targets.
@@ -978,6 +1082,23 @@ if (p_parse(&context) == P_SUCCESS)
 If AST generation mode is active, then the `p_result()` function returns a
 `Start *` pointing to the `Start` AST structure.
 
+When multiple start rules are specified, a separate result function is generated
+for each which returns the parse result for the corresponding rule.
+For example, if `Statement` is specified as a start rule:
+
+```
+p_context_t context;
+p_context_init(&context, input, input_length);
+size_t result = p_parse(&context);
+if (p_parse_Statement(&context) == P_SUCCESS)
+{
+    result = p_result_Statement(&context);
+}
+```
+
+In this case, the parser will start parsing with the `Statement` rule and the
+parse result from the `Statement` rule will be returned.
+
 ### `p_position`
 
 The `p_position()` function can be used to retrieve the parser position where
@@ -1050,6 +1171,31 @@ assert(code_point == 0x1F9E1u);
 assert(code_point_length == 4u);
 ```
 
+### `p_free_ast`
+
+The `p_free_ast()` function can be used to free the memory used by the AST.
+It should be passed the same value that is returned by `p_result()`.
+
+The `p_free_ast()` function is only available for C/C++ output targets.
+
+Note that if any lexer user code block allocates memory to store in a token's
+`pvalue`, in order to properly free this memory a `free_token_node` function
+should be specified in the grammar file.
+If specified, the `free_token_node` function will be called during the
+`p_free_ast()` process to allow user code to free any memory associated with
+a token node's `pvalue`.
+
+When multiple start rules are specified, a separate `p_free_ast` function is
+generated for each which frees the AST resulting from parsing the given rule.
+For example, if `Statement` is specified as a start rule:
+
+```
+p_free_ast_Statement(statement_ast);
+```
+
+In this case, Propane will free a `Statement` AST structure returned by the
+`p_parse_Statement(&context)` function.
+
 ##> Data
 
 ### `p_token_names`

lib/propane/generator.rb

@@ -11,24 +11,28 @@ class Propane
         @log = StringIO.new
       end
       @language =
-        if output_file =~ /\.([a-z]+)$/
+        if output_file.end_with?(".d")
-          $1
-        else
           "d"
+        else
+          "c"
         end
       @options = options
       process_grammar!
     end
 
     def generate
-      extensions = [@language]
+      extensions = [nil]
       if @language == "c"
         extensions += %w[h]
       end
       extensions.each do |extension|
-        template = Assets.get("parser.#{extension}.erb")
+        template = Assets.get("parser.#{extension || @language}.erb")
-        erb = ERB.new(template, trim_mode: "<>")
+        if extension
           output_file = @output_file.sub(%r{\.[a-z]+$}, ".#{extension}")
+        else
+          output_file = @output_file
+        end
+        erb = ERB.new(template, trim_mode: "<>")
         result = erb.result(binding.clone)
         File.open(output_file, "wb") do |fh|
           fh.write(result)
@@ -67,12 +71,15 @@ class Propane
         end
         tokens_by_name[token.name] = token
       end
-      # Check for user start rule.
+      # Create real start rule(s).
-      unless @grammar.rules.find {|rule| rule.name == @grammar.start_rule}
+      real_start_rules = @grammar.start_rules.map do |start_rule|
-        raise Error.new("Start rule `#{@grammar.start_rule}` not found")
+        unless @grammar.rules.find {|rule| rule.name == start_rule}
+          raise Error.new("Start rule `#{start_rule}` not found")
         end
-      # Add "real" start rule.
+        Rule.new("$#{start_rule}", [start_rule, "$EOF"], nil, nil, nil)
-      @grammar.rules.unshift(Rule.new("$Start", [@grammar.start_rule, "$EOF"], nil, nil, nil))
+      end
+      # Add real start rules before user-given rules.
+      @grammar.rules = real_start_rules + @grammar.rules
       # Generate and add rules for optional components.
       generate_optional_component_rules!(tokens_by_name)
       # Build rule sets.
@@ -328,9 +335,9 @@ class Propane
     #
     # @return [Array<String>]
     #   Start rule parser value type name and type string.
-    def start_rule_type
+    def start_rule_type(start_rule_index = 0)
       start_rule = @grammar.rules.find do |rule|
-        rule.name == @grammar.start_rule
+        rule.name == @grammar.start_rules[start_rule_index]
       end
       [start_rule.ptypename, @grammar.ptypes[start_rule.ptypename]]
     end

lib/propane/grammar.rb

@@ -8,10 +8,11 @@ class Propane
     attr_reader :ast
     attr_reader :ast_prefix
     attr_reader :ast_suffix
+    attr_reader :free_token_node
     attr_reader :modulename
     attr_reader :patterns
-    attr_reader :rules
+    attr_accessor :rules
-    attr_reader :start_rule
+    attr_reader :start_rules
     attr_reader :tokens
     attr_reader :code_blocks
     attr_reader :ptypes
@@ -19,7 +20,7 @@ class Propane
 
     def initialize(input)
       @patterns = []
-      @start_rule = "Start"
+      @start_rules = []
       @tokens = []
       @rules = []
       @code_blocks = {}
@@ -32,7 +33,9 @@ class Propane
       @ast = false
       @ast_prefix = ""
       @ast_suffix = ""
+      @free_token_node = nil
       parse_grammar!
+      @start_rules << "Start" if @start_rules.empty?
     end
 
     def ptype
@@ -62,6 +65,7 @@ class Propane
       elsif parse_ast_statement!
       elsif parse_ast_prefix_statement!
       elsif parse_ast_suffix_statement!
+      elsif parse_free_token_node_statement!
       elsif parse_module_statement!
       elsif parse_ptype_statement!
       elsif parse_pattern_statement!
@@ -112,6 +116,12 @@ class Propane
       end
     end
 
+    def parse_free_token_node_statement!
+      if md = consume!(/free_token_node\s+(\w+)\s*;/)
+        @free_token_node = md[1]
+      end
+    end
+
     def parse_module_statement!
       if consume!(/module\s+/)
         md = consume!(/([\w.]+)\s*/, "expected module name")
@@ -232,8 +242,11 @@ class Propane
     end
 
     def parse_start_statement!
-      if md = consume!(/start\s+(\w+)\s*;/)
+      if md = consume!(/start\s+([\w\s]*);/)
-        @start_rule = md[1]
+        start_rules = md[1].split(/\s+/).map(&:strip)
+        start_rules.each do |start_rule|
+          @start_rules << start_rule unless @start_rules.include?(start_rule)
+        end
       end
     end
 

lib/propane/parser.rb

@@ -16,11 +16,20 @@ class Propane
       @warnings = Set.new
       @errors = Set.new
       @options = options
-      start_item = Item.new(grammar.rules.first, 0)
+      start_items = grammar.rules[0...grammar.start_rules.length].map do |start_rule|
-      eval_item_sets = Set[ItemSet.new([start_item])]
+        Item.new(start_rule, 0)
+      end
+      start_item_sets = start_items.map {|item| ItemSet.new([item])}
+      eval_item_sets = Set[*start_item_sets]
 
       while eval_item_sets.size > 0
-        item_set = eval_item_sets.first
+        item_set =
+          if start_item_sets.size > 0
+            # Ensure we evaluate start_item_sets first in order
+            start_item_sets.slice!(0)
+          else
+            eval_item_sets.first
+          end
         eval_item_sets.delete(item_set)
         unless @item_sets_set.include?(item_set)
           item_set.id = @item_sets.size

spec/json_parser.c.propane

@@ -120,11 +120,11 @@ string: /\\t/ <<
 >>
 string: /\\u[0-9a-fA-F]{4}/ <<
   /* Not actually going to encode the code point for this example... */
-  char s[] = {'{', match[2], match[3], match[4], match[5], '}', 0};
+  char s[] = {'{', (char)match[2], (char)match[3], (char)match[4], (char)match[5], '}', 0};
   str_append(&string_value, s);
 >>
 string: /[^\\]/ <<
-  char s[] = {match[0], 0};
+  char s[] = {(char)match[0], 0};
   str_append(&string_value, s);
 >>
 Start -> Value <<

spec/json_types.c

@@ -5,7 +5,7 @@
 
 JSONValue * JSONValue_new(size_t id)
 {
-    JSONValue * jv = calloc(1, sizeof(JSONValue));
+    JSONValue * jv = (JSONValue *)calloc(1, sizeof(JSONValue));
     jv->id = id;
     return jv;
 }
@@ -29,7 +29,7 @@ void JSONObject_append(JSONValue * object, char const * name, JSONValue * value)
         }
     }
     size_t const new_size = size + 1;
-    void * new_entries = malloc(sizeof(object->object.entries[0]) * new_size);
+    JSONObjectEntry * new_entries = (JSONObjectEntry *)malloc(sizeof(object->object.entries[0]) * new_size);
     if (size > 0)
     {
         memcpy(new_entries, object->object.entries, size * sizeof(object->object.entries[0]));
@@ -52,7 +52,7 @@ void JSONArray_append(JSONValue * array, JSONValue * value)
 {
     size_t const size = array->array.size;
     size_t const new_size = size + 1;
-    JSONValue ** new_entries = malloc(sizeof(JSONValue *) * new_size);
+    JSONValue ** new_entries = (JSONValue **)malloc(sizeof(JSONValue *) * new_size);
     if (array->array.size > 0)
     {
         memcpy(new_entries, array->array.entries, sizeof(JSONValue *) * size);

spec/json_types.h

@@ -11,6 +11,12 @@
 #define JSON_FALSE 5u
 #define JSON_NULL 6u
 
+typedef struct JSONObjectEntry_s
+{
+    char const * name;
+    struct JSONValue_s * value;
+} JSONObjectEntry;
+
 typedef struct JSONValue_s
 {
     size_t id;
@@ -19,11 +25,7 @@ typedef struct JSONValue_s
         struct
         {
             size_t size;
-            struct
+            JSONObjectEntry * entries;
-            {
-                char const * name;
-                struct JSONValue_s * value;
-            } * entries;
         } object;
         struct
         {

spec/propane_spec.rb

@@ -65,26 +65,47 @@ EOF
   end
 
   def compile(test_files, options = {})
-    test_files = Array(test_files)
+    test_files = Array(test_files).map do |test_file|
+      if !File.exist?(test_file) && test_file.end_with?(".cpp")
+        test_file.sub(%r{\.cpp$}, ".c")
+      else
+        test_file
+      end
+    end
     options[:parsers] ||= [""]
     parsers = options[:parsers].map do |name|
       "spec/run/testparser#{name}.#{options[:language]}"
     end
     case options[:language]
     when "c"
-      result = system(*%w[gcc -Wall -o spec/run/testparser -Ispec -Ispec/run], *parsers, *test_files, "spec/testutils.c", "-lm")
+      command = [*%w[gcc -g -Wall -o spec/run/testparser -Ispec -Ispec/run], *parsers, *test_files, "spec/testutils.c", "-lm"]
+    when "cpp"
+      command = [*%w[g++ -g -x c++ -Wall -o spec/run/testparser -Ispec -Ispec/run], *parsers, *test_files, "spec/testutils.c", "-lm"]
     when "d"
-      result = system(*%w[ldc2 -g --unittest -of spec/run/testparser -Ispec], *parsers, *test_files, "spec/testutils.d")
+      command = [*%w[ldc2 -g --unittest -of spec/run/testparser -Ispec], *parsers, *test_files, "spec/testutils.d"]
     end
+    result = system(*command)
     expect(result).to be_truthy
   end
 
-  def run_test
+  def run_test(options)
     stdout, stderr, status = Open3.capture3("spec/run/testparser")
     File.binwrite("spec/run/.stderr", stderr)
     File.binwrite("spec/run/.stdout", stdout)
     stderr.sub!(/^.*modules passed unittests\n/, "")
-    Results.new(stdout, stderr, status)
+    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|
+        bytes = $1.to_i
+        if bytes > 0
+          raise "Valgrind detected memory leak"
+        end
+      end
+    end
+    results
   end
 
   def lines(str)
@@ -261,7 +282,7 @@ EOF
     expect(results.status).to_not eq 0
   end
 
-  %w[d c].each do |language|
+  %w[d c cpp].each do |language|
 
     context "#{language.upcase} language" do
 
@@ -277,14 +298,14 @@ Foo -> plus <<>>
 EOF
         run_propane(language: language)
         compile("spec/test_lexer.#{language}", language: language)
-        results = run_test
+        results = run_test(language: language)
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
       end
 
       it "detects a lexer error when an unknown character is seen" do
         case language
-        when "c"
+        when "c", "cpp"
           write_grammar <<EOF
 ptype int;
 token int /\\d+/ <<
@@ -315,7 +336,7 @@ EOF
         end
         run_propane(language: language)
         compile("spec/test_lexer_unknown_character.#{language}", language: language)
-        results = run_test
+        results = run_test(language: language)
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
       end
@@ -338,7 +359,7 @@ EOF
 
       it "generates a parser that does basic math - user guide example" do
         case language
-        when "c"
+        when "c", "cpp"
           write_grammar <<EOF
 <<
 #include <math.h>
@@ -409,7 +430,7 @@ EOF
         end
         run_propane(language: language)
         compile("spec/test_basic_math_grammar.#{language}", language: language)
-        results = run_test
+        results = run_test(language: language)
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
       end
@@ -435,7 +456,7 @@ R2 -> a b;
 EOF
         run_propane(language: language)
         compile("spec/test_parser_identical_rules_lookahead.#{language}", language: language)
-        results = run_test
+        results = run_test(language: language)
         expect(results.status).to eq 0
       end
 
@@ -450,13 +471,13 @@ R1 -> b;
 EOF
         run_propane(language: language)
         compile("spec/test_parser_rule_from_multiple_states.#{language}", language: language)
-        results = run_test
+        results = run_test(language: language)
         expect(results.status).to eq 0
       end
 
       it "executes user code when matching lexer token" do
         case language
-        when "c"
+        when "c", "cpp"
           write_grammar <<EOF
 <<
 #include <stdio.h>
@@ -485,7 +506,7 @@ EOF
         end
         run_propane(language: language)
         compile("spec/test_user_code.#{language}", language: language)
-        results = run_test
+        results = run_test(language: language)
         expect(results.status).to eq 0
         verify_lines(results.stdout, [
           "abc!",
@@ -498,7 +519,7 @@ EOF
 
       it "supports a pattern statement" do
         case language
-        when "c"
+        when "c", "cpp"
           write_grammar <<EOF
 <<
 #include <stdio.h>
@@ -521,7 +542,7 @@ EOF
         end
         run_propane(language: language)
         compile("spec/test_pattern.#{language}", language: language)
-        results = run_test
+        results = run_test(language: language)
         expect(results.status).to eq 0
         verify_lines(results.stdout, [
           "def!",
@@ -534,7 +555,7 @@ EOF
 
       it "supports returning tokens from pattern code blocks" do
         case language
-        when "c"
+        when "c", "cpp"
           write_grammar <<EOF
 <<
 #include <stdio.h>
@@ -563,7 +584,7 @@ EOF
         end
         run_propane(language: language)
         compile("spec/test_return_token_from_pattern.#{language}", language: language)
-        results = run_test
+        results = run_test(language: language)
         expect(results.status).to eq 0
         verify_lines(results.stdout, [
           "def!",
@@ -574,7 +595,7 @@ EOF
 
       it "supports lexer modes" do
         case language
-        when "c"
+        when "c", "cpp"
           write_grammar <<EOF
 <<
 #include <stdio.h>
@@ -621,7 +642,7 @@ EOF
         end
         run_propane(language: language)
         compile("spec/test_lexer_modes.#{language}", language: language)
-        results = run_test
+        results = run_test(language: language)
         expect(results.status).to eq 0
         verify_lines(results.stdout, [
           "begin string mode",
@@ -635,7 +656,7 @@ EOF
 
       it "multiple lexer modes may apply to a pattern" do
         case language
-        when "c"
+        when "c", "cpp"
           write_grammar <<EOF
 <<
 #include <stdio.h>
@@ -679,7 +700,7 @@ EOF
         end
         run_propane(language: language)
         compile("spec/test_lexer_multiple_modes.#{language}", language: language)
-        results = run_test
+        results = run_test(language: language)
         expect(results.status).to eq 0
         verify_lines(results.stdout, [
           "ident: d",
@@ -691,7 +712,7 @@ EOF
 
       it "executes user code associated with a parser rule" do
         case language
-        when "c"
+        when "c", "cpp"
           write_grammar <<EOF
 <<
 #include <stdio.h>
@@ -716,7 +737,7 @@ EOF
         end
         run_propane(language: language)
         compile("spec/test_parser_rule_user_code.#{language}", language: language)
-        results = run_test
+        results = run_test(language: language)
         expect(results.status).to eq 0
         verify_lines(results.stdout, [
           "A!",
@@ -727,7 +748,7 @@ EOF
 
       it "parses lists" do
         write_grammar <<EOF
-ptype #{language == "c" ? "uint32_t" : "uint"};
+ptype #{language == "d" ? "uint" : "uint32_t"};
 token a;
 Start -> As << $$ = $1; >>
 As -> << $$ = 0u; >>
@@ -735,7 +756,7 @@ As -> As a << $$ = $1 + 1u; >>
 EOF
         run_propane(language: language)
         compile("spec/test_parsing_lists.#{language}", language: language)
-        results = run_test
+        results = run_test(language: language)
         expect(results.status).to eq 0
         expect(results.stderr).to eq ""
       end
@@ -762,14 +783,14 @@ EOF
 
       it "provides matched text to user code blocks" do
         case language
-        when "c"
+        when "c", "cpp"
           write_grammar <<EOF
 <<
 #include <stdio.h>
 #include <stdlib.h>
 >>
 token id /[a-zA-Z_][a-zA-Z0-9_]*/ <<
-  char * t = malloc(match_length + 1);
+  char * t = (char *)malloc(match_length + 1);
   strncpy(t, (char *)match, match_length);
   printf("Matched token is %s\\n", t);
   free(t);
@@ -789,7 +810,7 @@ EOF
         end
         run_propane(language: language)
         compile("spec/test_lexer_match_text.#{language}", language: language)
-        results = run_test
+        results = run_test(language: language)
         expect(results.status).to eq 0
         verify_lines(results.stdout, [
           "Matched token is identifier_123",
@@ -799,7 +820,7 @@ EOF
 
       it "allows storing a result value for the lexer" do
         case language
-        when "c"
+        when "c", "cpp"
           write_grammar <<EOF
 ptype uint64_t;
 token word /[a-z]+/ <<
@@ -822,7 +843,7 @@ EOF
         end
         run_propane(language: language)
         compile("spec/test_lexer_result_value.#{language}", language: language)
-        results = run_test
+        results = run_test(language: language)
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
       end
@@ -837,13 +858,14 @@ Start -> a num;
 EOF
         run_propane(language: language)
         compile("spec/test_error_positions.#{language}", language: language)
-        results = run_test
+        results = run_test(language: language)
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
       end
 
       it "allows creating a JSON parser" do
-        write_grammar(File.read("spec/json_parser.#{language}.propane"))
+        ext = language == "cpp" ? "c" : language
+        write_grammar(File.read("spec/json_parser.#{ext}.propane"))
         run_propane(language: language)
         compile(["spec/test_parsing_json.#{language}", "spec/json_types.#{language}"], language: language)
       end
@@ -865,7 +887,7 @@ Start -> b c b;
 EOF
         run_propane(name: "myp2", language: language)
         compile("spec/test_multiple_parsers.#{language}", parsers: %w[myp1 myp2], language: language)
-        results = run_test
+        results = run_test(language: language)
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
       end
@@ -884,7 +906,7 @@ Any -> c;
 EOF
         run_propane(language: language)
         compile("spec/test_user_terminate_lexer.#{language}", language: language)
-        results = run_test
+        results = run_test(language: language)
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
       end
@@ -902,14 +924,14 @@ Any -> ;
 EOF
         run_propane(language: language)
         compile("spec/test_user_terminate.#{language}", language: language)
-        results = run_test
+        results = run_test(language: language)
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
       end
 
       it "matches backslash escape sequences" do
         case language
-        when "c"
+        when "c", "cpp"
           write_grammar <<EOF
 <<
   #include <stdio.h>
@@ -946,7 +968,7 @@ EOF
         end
         run_propane(language: language)
         compile("spec/test_match_backslashes.#{language}", language: language)
-        results = run_test
+        results = run_test(language: language)
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
         verify_lines(results.stdout, [
@@ -1011,7 +1033,7 @@ Two -> two;
 EOF
         run_propane(language: language)
         compile("spec/test_ast.#{language}", language: language)
-        results = run_test
+        results = run_test(language: language)
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
       end
@@ -1055,7 +1077,7 @@ Two -> two;
 EOF
         run_propane(language: language)
         compile("spec/test_ast_ps.#{language}", language: language)
-        results = run_test
+        results = run_test(language: language)
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
       end
@@ -1123,13 +1145,13 @@ Start -> a? b R? <<
   printf("b: %d\\n", $2);
   printf("R: %s\\n", $3 == NULL ? "" : $3);
 >>
-R -> c d << $$ = "cd"; >>
+R -> c d << $$ = (char *)"cd"; >>
-R (string) -> d c << $$ = "dc"; >>
+R (string) -> d c << $$ = (char *)"dc"; >>
 EOF
         end
         run_propane(language: language)
         compile("spec/test_optional_rule_component.#{language}", language: language)
-        results = run_test
+        results = run_test(language: language)
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
         verify_lines(results.stdout, [
@@ -1181,7 +1203,7 @@ EOF
         end
         run_propane(language: language)
         compile("spec/test_optional_rule_component_ast.#{language}", language: language)
-        results = run_test
+        results = run_test(language: language)
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
       end
@@ -1222,7 +1244,7 @@ EOF
         end
         run_propane(language: language)
         compile("spec/test_named_optional_rule_component_ast.#{language}", language: language)
-        results = run_test
+        results = run_test(language: language)
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
       end
@@ -1242,7 +1264,7 @@ T -> c;
 EOF
         run_propane(language: language)
         compile("spec/test_ast_token_positions.#{language}", language: language)
-        results = run_test
+        results = run_test(language: language)
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
       end
@@ -1262,7 +1284,7 @@ A -> bb? c?;
 EOF
         run_propane(language: language)
         compile("spec/test_ast_invalid_positions.#{language}", language: language)
-        results = run_test
+        results = run_test(language: language)
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
       end
@@ -1282,7 +1304,7 @@ T -> c;
 EOF
         run_propane(language: language)
         compile("spec/test_ast_field_aliases.#{language}", language: language)
-        results = run_test
+        results = run_test(language: language)
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
       end
@@ -1305,7 +1327,7 @@ T -> c;
 EOF
         run_propane(language: language)
         compile("spec/test_ast_field_aliases.#{language}", language: language)
-        results = run_test
+        results = run_test(language: language)
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
       end
@@ -1332,9 +1354,9 @@ EOF
 #include <stdio.h>
 #include <string.h>
 >>
-ptype char const *;
+ptype char *;
 token id /[a-zA-Z_][a-zA-Z0-9_]*/ <<
-  char * s = malloc(match_length + 1);
+  char * s = (char *)malloc(match_length + 1);
   strncpy(s, (char const *)match, match_length);
   s[match_length] = 0;
   $$ = s;
@@ -1343,12 +1365,14 @@ drop /\\s+/;
 Start -> id:first id:second <<
   printf("first is %s\\n", ${first});
   printf("second is %s\\n", ${second});
+  free(${first});
+  free(${second});
 >>
 EOF
         end
         run_propane(language: language)
         compile("spec/test_field_aliases.#{language}", language: language)
-        results = run_test
+        results = run_test(language: language)
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
         expect(results.stdout).to match /first is foo1.*second is bar2/m
@@ -1379,9 +1403,9 @@ EOF
 #include <stdio.h>
 #include <string.h>
 >>
-ptype char const *;
+ptype char *;
 token id /[a-zA-Z_][a-zA-Z0-9_]*/ <<
-  char * s = malloc(match_length + 1);
+  char * s = (char *)malloc(match_length + 1);
   strncpy(s, (char const *)match, match_length);
   s[match_length] = 0;
   $$ = s;
@@ -1392,26 +1416,100 @@ Start -> Foo;
 Start -> id:first id:second <<
   printf("first is %s\\n", ${first});
   printf("second is %s\\n", ${second});
+  free(${first});
+  free(${second});
 >>
 Foo -> ;
 EOF
         end
         run_propane(language: language)
         compile("spec/test_field_aliases.#{language}", language: language)
-        results = run_test
+        results = run_test(language: language)
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
         expect(results.stdout).to match /first is foo1.*second is bar2/m
       end
 
       it "does not free memory allocated for AST nodes" do
-        write_grammar(File.read("spec/ast_node_memory_remains.#{language}.propane"))
+        ext = language == "cpp" ? "c" : language
+        write_grammar(File.read("spec/ast_node_memory_remains.#{ext}.propane"))
         run_propane(language: language)
         compile("spec/test_ast_node_memory_remains.#{language}", language: language)
-        results = run_test
+        results = run_test(language: language)
+        expect(results.stderr).to eq ""
+        expect(results.status).to eq 0
+      end
+
+      it "allows multiple starting rules" do
+    write_grammar <<EOF
+ptype int;
+token a << $$ = 1; >>
+token b << $$ = 2; >>
+token c << $$ = 3; >>
+Start -> a b R;
+Start -> Bs:bs << $$ = $1; >>
+R -> c:c << $$ = $1; >>
+Bs -> << $$ = 0; >>
+Bs -> b:b Bs:bs << $$ = $1 + $2; >>
+start Start R Bs;
+EOF
+        run_propane(language: language)
+        compile("spec/test_starting_rules.#{language}", language: language)
+        results = run_test(language: language)
+        expect(results.stderr).to eq ""
+        expect(results.status).to eq 0
+      end
+
+      it "allows multiple starting rules in AST mode" do
+    write_grammar <<EOF
+ast;
+ptype int;
+token a << $$ = 1; >>
+token b << $$ = 2; >>
+token c << $$ = 3; >>
+Start -> a b R;
+Start -> Bs:bs;
+R -> c:c;
+Bs -> ;
+Bs -> b:b Bs:bs;
+start Start R Bs;
+EOF
+        run_propane(language: language)
+        compile("spec/test_starting_rules_ast.#{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
+    write_grammar <<EOF
+<<
+static void free_token(Token * token)
+{
+    free(token->pvalue);
+}
+>>
+ast;
+free_token_node free_token;
+ptype int *;
+token a <<
+  $$ = (int *)malloc(sizeof(int));
+  *$$ = 1;
+>>
+token b <<
+  $$ = (int *)malloc(sizeof(int));
+  *$$ = 2;
+>>
+Start -> a:a b:b;
+EOF
+          run_propane(language: language)
+          compile("spec/test_free_ast_token_node_memory.#{language}", language: language)
+          results = run_test(language: language)
           expect(results.stderr).to eq ""
           expect(results.status).to eq 0
         end
       end
     end
   end
+end

spec/spec_helper.rb

@@ -2,6 +2,10 @@ unless ENV["dist_specs"]
   require "bundler/setup"
   require "simplecov"
 
+  class MyFormatter
+    def format(*args)
+    end
+  end
   SimpleCov.start do
     add_filter "/spec/"
     add_filter "/.bundle/"
@@ -12,6 +16,7 @@ unless ENV["dist_specs"]
     end
     project_name "Propane"
     merge_timeout 3600
+    formatter(MyFormatter)
   end
 
   RSpec.configure do |config|

spec/test_ast.c

@@ -33,12 +33,16 @@ int main()
     assert_eq(22, itemsmore->pItem->pToken1->pvalue);
     assert(itemsmore->pItemsMore == NULL);
 
+    p_free_ast(start);
+
     input = "";
     p_context_init(&context, (uint8_t const *)input, strlen(input));
     assert_eq(P_SUCCESS, p_parse(&context));
     start = p_result(&context);
     assert(start->pItems == NULL);
 
+    p_free_ast(start);
+
     input = "2 1";
     p_context_init(&context, (uint8_t const *)input, strlen(input));
     assert_eq(P_SUCCESS, p_parse(&context));
@@ -51,5 +55,7 @@ int main()
     assert(start->pItems->pItem->pDual->pTwo2 == NULL);
     assert(start->pItems->pItem->pDual->pOne1 == NULL);
 
+    p_free_ast(start);
+
     return 0;
 }

spec/test_ast_field_aliases.c

@@ -15,5 +15,7 @@ int main()
     assert_eq(TOKEN_b, start->second->pToken->token);
     assert_eq(TOKEN_c, start->third->pToken->token);
 
+    p_free_ast(start);
+
     return 0;
 }

spec/test_ast_invalid_positions.c

@@ -30,6 +30,8 @@ int main()
     assert_eq(3, start->end_position.row);
     assert_eq(8, start->end_position.col);
 
+    p_free_ast(start);
+
     input = "a\nbb";
     p_context_init(&context, (uint8_t const *)input, strlen(input));
     assert(p_parse(&context) == P_SUCCESS);
@@ -54,6 +56,8 @@ int main()
     assert_eq(2, start->end_position.row);
     assert_eq(2, start->end_position.col);
 
+    p_free_ast(start);
+
     input = "a\nc\nc";
     p_context_init(&context, (uint8_t const *)input, strlen(input));
     assert(p_parse(&context) == P_SUCCESS);
@@ -78,6 +82,8 @@ int main()
     assert_eq(3, start->end_position.row);
     assert_eq(1, start->end_position.col);
 
+    p_free_ast(start);
+
     input = "a";
     p_context_init(&context, (uint8_t const *)input, strlen(input));
     assert(p_parse(&context) == P_SUCCESS);
@@ -98,5 +104,7 @@ int main()
     assert_eq(1, start->end_position.row);
     assert_eq(1, start->end_position.col);
 
+    p_free_ast(start);
+
     return 0;
 }

spec/test_ast_node_memory_remains.c

@@ -386,7 +386,7 @@ int main(int argc, char * argv[])
         }
         pmis = pmis->pModuleItems;
     }
-    pfds = malloc(n_pfds * sizeof(PModuleItems *));
+    pfds = (PFunctionDefinition **)malloc(n_pfds * sizeof(PModuleItems *));
     pmis = pmod->pModuleItems;
     size_t pfd_i = n_pfds;
     while (pmis != NULL)
@@ -411,5 +411,8 @@ int main(int argc, char * argv[])
         }
     }
 
+    free(pfds);
+    p_free_ast(pmod);
+
     return 0;
 }

spec/test_ast_ps.c

@@ -33,12 +33,16 @@ int main()
     assert_eq(22, itemsmore->pItem->pToken1->pvalue);
     assert(itemsmore->pItemsMore == NULL);
 
+    p_free_ast(start);
+
     input = "";
     p_context_init(&context, (uint8_t const *)input, strlen(input));
     assert_eq(P_SUCCESS, p_parse(&context));
     start = p_result(&context);
     assert(start->pItems == NULL);
 
+    p_free_ast(start);
+
     input = "2 1";
     p_context_init(&context, (uint8_t const *)input, strlen(input));
     assert_eq(P_SUCCESS, p_parse(&context));
@@ -51,5 +55,7 @@ int main()
     assert(start->pItems->pItem->pDual->pTwo2 == NULL);
     assert(start->pItems->pItem->pDual->pOne1 == NULL);
 
+    p_free_ast(start);
+
     return 0;
 }

spec/test_ast_token_positions.c

@@ -43,6 +43,8 @@ int main()
     assert_eq(1, start->end_position.row);
     assert_eq(6, start->end_position.col);
 
+    p_free_ast(start);
+
     input = "\n\n  bb\nc\ncc\n\n     a";
     p_context_init(&context, (uint8_t const *)input, strlen(input));
     assert(p_parse(&context) == P_SUCCESS);
@@ -80,5 +82,7 @@ int main()
     assert_eq(7, start->end_position.row);
     assert_eq(6, start->end_position.col);
 
+    p_free_ast(start);
+
     return 0;
 }

spec/test_free_ast_token_node_memory.c

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

spec/test_named_optional_rule_component_ast.c

@@ -17,6 +17,8 @@ int main()
     assert(start->pR == NULL);
     assert(start->r == NULL);
 
+    p_free_ast(start);
+
     input = "abcd";
     p_context_init(&context, (uint8_t const *)input, strlen(input));
     assert(p_parse(&context) == P_SUCCESS);
@@ -31,6 +33,8 @@ int main()
     assert(start->pR == start->r);
     assert_eq(TOKEN_c, start->pR->pToken1->token);
 
+    p_free_ast(start);
+
     input = "bdc";
     p_context_init(&context, (uint8_t const *)input, strlen(input));
     assert(p_parse(&context) == P_SUCCESS);
@@ -40,6 +44,8 @@ int main()
     assert(start->r != NULL);
     assert_eq(TOKEN_d, start->pR->pToken1->token);
 
+    p_free_ast(start);
+
     return 0;
 }
 

spec/test_optional_rule_component_ast.c

@@ -16,6 +16,8 @@ int main()
     assert(start->pR3 == NULL);
     assert(start->pR == NULL);
 
+    p_free_ast(start);
+
     input = "abcd";
     p_context_init(&context, (uint8_t const *)input, strlen(input));
     assert(p_parse(&context) == P_SUCCESS);
@@ -28,6 +30,8 @@ int main()
     assert(start->pR == start->pR3);
     assert_eq(TOKEN_c, start->pR->pToken1->token);
 
+    p_free_ast(start);
+
     input = "bdc";
     p_context_init(&context, (uint8_t const *)input, strlen(input));
     assert(p_parse(&context) == P_SUCCESS);
@@ -37,6 +41,8 @@ int main()
     assert(start->pR != NULL);
     assert_eq(TOKEN_d, start->pR->pToken1->token);
 
+    p_free_ast(start);
+
     return 0;
 }
 

spec/test_start_rule_ast.c

@@ -13,5 +13,7 @@ int main()
     assert(top->pToken != NULL);
     assert_eq(TOKEN_hi, top->pToken->token);
 
+    p_free_ast(top);
+
     return 0;
 }

spec/test_starting_rules.c

@@ -0,0 +1,27 @@
+#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);
+    int result = p_result(&context);
+    assert_eq(8, result);
+
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse_Bs(&context) == P_SUCCESS);
+    result = p_result_Bs(&context);
+    assert_eq(8, result);
+
+    input = "c";
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse_R(&context) == P_SUCCESS);
+    result = p_result_R(&context);
+    assert_eq(3, result);
+
+    return 0;
+}

spec/test_starting_rules.d

@@ -0,0 +1,29 @@
+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);
+    int result = p_result(&context);
+    assert(result == 8);
+
+    p_context_init(&context, input);
+    assert(p_parse_Bs(&context) == P_SUCCESS);
+    result = p_result_Bs(&context);
+    assert(result == 8);
+
+    input = "c";
+    p_context_init(&context, input);
+    assert(p_parse_R(&context) == P_SUCCESS);
+    result = p_result_R(&context);
+    assert(result == 3);
+}

spec/test_starting_rules_ast.c

@@ -0,0 +1,37 @@
+#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

@@ -0,0 +1,36 @@
+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/testutils.c

@@ -14,10 +14,28 @@ void assert_eq_size_t_i(size_t expected, size_t actual, char const * file, size_
     }
 }
 
+void assert_ne_size_t_i(size_t expected, size_t actual, char const * file, size_t line)
+{
+    if (expected == actual)
+    {
+        fprintf(stderr, "%s:%lu: expected not %lu, got %lu\n", file, line, expected, actual);
+        assert(false);
+    }
+}
+
+void assert_not_null_i(void * ptr, char const * file, size_t line)
+{
+    if (ptr == NULL)
+    {
+        fprintf(stderr, "%s:%lu: expected not NULL\n", file, line);
+        assert(false);
+    }
+}
+
 void str_init(str_t * str, char const * cs)
 {
     size_t length = strlen(cs);
-    str->cs = malloc(length + 1u);
+    str->cs = (char *)malloc(length + 1u);
     strcpy(str->cs, cs);
 }
 
@@ -25,7 +43,7 @@ void str_append(str_t * str, char const * cs)
 {
     size_t length = strlen(str->cs);
     size_t length2 = strlen(cs);
-    char * new_cs = malloc(length + length2 + 1u);
+    char * new_cs = (char *)malloc(length + length2 + 1u);
     memcpy(new_cs, str->cs, length);
     strcpy(&new_cs[length], cs);
     free(str->cs);

spec/testutils.h

@@ -5,6 +5,16 @@ void assert_eq_size_t_i(size_t expected, size_t actual, char const * file, size_
 #define assert_eq(expected, actual) \
     assert_eq_size_t_i(expected, actual, __FILE__, __LINE__)
 
+void assert_ne_size_t_i(size_t expected, size_t actual, char const * file, size_t line);
+
+#define assert_ne(expected, actual) \
+    assert_ne_size_t_i(expected, actual, __FILE__, __LINE__)
+
+void assert_not_null_i(void * ptr, char const * file, size_t line);
+
+#define assert_not_null(ptr) \
+    assert_not_null_i(ptr, __FILE__, __LINE__)
+
 typedef struct
 {
     char * cs;