wip

assets/parser.c.erb

@@ -0,0 +1,938 @@
+#include "<%= File.basename(output_file).sub(%r{\.[a-z]+$}, "") %>.h"
+#include <stdbool.h>
+#include <stdlib.h>
+#include <string.h>
+
+/**************************************************************************
+ * User code blocks
+ *************************************************************************/
+
+<% @grammar.code_blocks.each do |code| %>
+<%= code %>
+<% end %>
+
+/**************************************************************************
+ * Private types
+ *************************************************************************/
+
+<% if @grammar.prefix.upcase != "P_" %>
+/* Result codes. */
+#define P_SUCCESS 0u
+#define P_DECODE_ERROR 1u
+#define P_UNEXPECTED_INPUT 2u
+#define P_UNEXPECTED_TOKEN 3u
+#define P_DROP 4u
+#define P_EOF 5u
+<% end %>
+
+/* An invalid ID value. */
+#define INVALID_ID ((size_t)-1)
+
+/**************************************************************************
+ * State initialization
+ *************************************************************************/
+
+/**
+ * Initialize lexer/parser context structure.
+ *
+ * @param[out] context
+ *   Lexer/parser context structure.
+ * @param input
+ *   Text input.
+ * @param input_length
+ *   Text input length.
+ */
+void <%= @grammar.prefix %>context_init(<%= @grammar.prefix %>context_t * context, uint8_t const * input, size_t input_length)
+{
+    /* New default-initialized context structure. */
+    <%= @grammar.prefix %>context_t newcontext = {0};
+
+    /* Lexer initialization. */
+    newcontext.input = input;
+    newcontext.input_length = input_length;
+    newcontext.mode = <%= @lexer.mode_id("default") %>;
+
+    /* Copy to the user's context structure. */
+    *context = newcontext;
+}
+
+/**************************************************************************
+ * Decoder
+ *************************************************************************/
+
+/**
+ * Decode a UTF-8 code point.
+ *
+ * @param input
+ *   Text input to decode.
+ * @param input_length
+ *   Input text length.
+ * @param[out] out_code_point
+ *   The decoded code point is stored here if the return value is P_SUCCESS.
+ * @param[out] out_code_point_length
+ *   The number of bytes the code point used is stored here if the return value
+ *   is P_SUCCESS.
+ *
+ * @retval P_SUCCESS on a successful code point decode
+ * @retval P_DECODE_ERROR when an encoding error is observed
+ * @retval P_EOF when the end of the text input is reached
+ */
+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)
+{
+    if (input_length == 0u)
+    {
+        return P_EOF;
+    }
+    char c = input[0];
+    <%= @grammar.prefix %>code_point_t code_point;
+    uint8_t code_point_length;
+    if ((c & 0x80u) == 0u)
+    {
+        code_point = c;
+        code_point_length = 1u;
+    }
+    else
+    {
+        uint8_t following_bytes;
+        if ((c & 0xE0u) == 0xC0u)
+        {
+            code_point = c & 0x1Fu;
+            following_bytes = 1u;
+        }
+        else if ((c & 0xF0u) == 0xE0u)
+        {
+            code_point = c & 0x0Fu;
+            following_bytes = 2u;
+        }
+        else if ((c & 0xF8u) == 0xF0u)
+        {
+            code_point = c & 0x07u;
+            following_bytes = 3u;
+        }
+        else if ((c & 0xFCu) == 0xF8u)
+        {
+            code_point = c & 0x03u;
+            following_bytes = 4u;
+        }
+        else if ((c & 0xFEu) == 0xFCu)
+        {
+            code_point = c & 0x01u;
+            following_bytes = 5u;
+        }
+        else
+        {
+            return P_DECODE_ERROR;
+        }
+        if (input_length <= following_bytes)
+        {
+            return P_DECODE_ERROR;
+        }
+        code_point_length = (uint8_t)(following_bytes + 1u);
+        for (size_t i = 0u; i < following_bytes; i++)
+        {
+            char b = input[i + 1u];
+            if ((b & 0xC0u) != 0x80u)
+            {
+                return P_DECODE_ERROR;
+            }
+            code_point = (code_point << 6u) | (b & 0x3Fu);
+        }
+    }
+    *out_code_point = code_point;
+    *out_code_point_length = code_point_length;
+    return P_SUCCESS;
+}
+
+/**************************************************************************
+ * Lexer
+ *************************************************************************/
+
+/** Lexer state ID type. */
+typedef <%= get_type_for(@lexer.state_table.size) %> lexer_state_id_t;
+
+/** Invalid lexer state ID. */
+#define INVALID_LEXER_STATE_ID <%= @lexer.state_table.size %>u
+
+/** Lexer user code ID type. */
+<% user_code_id_count = (@grammar.patterns.map(&:code_id).compact.max || 0) + 1 %>
+typedef <%= get_type_for(user_code_id_count) %> lexer_user_code_id_t;
+
+/** Invalid lexer user code ID. */
+#define INVALID_USER_CODE_ID <%= user_code_id_count %>u
+
+/**
+ * Lexer transition table entry.
+ *
+ * An incoming code point matching the range for a transition entry will cause
+ * the lexer to progress to the destination state.
+ */
+typedef struct
+{
+    /** First code point in the range for this transition. */
+    <%= @grammar.prefix %>code_point_t first;
+
+    /** Last code point in the range for this transition. */
+    <%= @grammar.prefix %>code_point_t last;
+
+    /** Destination lexer state ID for this transition. */
+    lexer_state_id_t destination_state;
+} lexer_transition_t;
+
+/** Lexer state table entry. */
+typedef struct
+{
+    /** Index to the transition table for this state. */
+    <%= get_type_for(@lexer.transition_table.size - 1) %> transition_table_index;
+
+    /** Number of transition table entries for this state. */
+    <%= get_type_for(@lexer.state_table.map {|ste| ste[:n_transitions]}.max) %> n_transitions;
+
+    /** Lexer token formed at this state. */
+    <%= @grammar.prefix %>token_t token;
+
+    /** Lexer user code ID to execute at this state. */
+    lexer_user_code_id_t code_id;
+
+    /** Whether this state matches a lexer pattern. */
+    bool accepts;
+} lexer_state_t;
+
+/** Lexer mode table entry. */
+typedef struct
+{
+    /** Offset in the state table to be used for this mode. */
+    uint32_t state_table_offset;
+} lexer_mode_t;
+
+/**
+ * Lexer match info structure.
+ *
+ * This structure holds output values from the lexer upon a successful pattern
+ * match.
+ */
+typedef struct
+{
+    /** Number of bytes of input text used to match. */
+    size_t length;
+
+    /** Input text position delta. */
+    <%= @grammar.prefix %>position_t delta_position;
+
+    /** Accepting lexer state from the match. */
+    lexer_state_t const * accepting_state;
+} lexer_match_info_t;
+
+/** Lexer transition table. */
+static lexer_transition_t lexer_transition_table[] = {
+<% @lexer.transition_table.each do |transition_table_entry| %>
+    {<%= transition_table_entry[:first] %>u, <%= transition_table_entry[:last] %>u, <%= transition_table_entry[:destination] %>u},
+<% end %>
+};
+
+/** Lexer state table. */
+static lexer_state_t lexer_state_table[] = {
+<% @lexer.state_table.each do |state_table_entry| %>
+    {<%= state_table_entry[:transition_table_index] %>u, <%= state_table_entry[:n_transitions] %>u, <%= state_table_entry[:token] || "INVALID_TOKEN_ID" %>, <%= state_table_entry[:code_id] || "INVALID_USER_CODE_ID" %>, <%= state_table_entry[:accepts] %>},
+<% end %>
+};
+
+/** Lexer mode table. */
+static lexer_mode_t lexer_mode_table[] = {
+<% @lexer.mode_table.each do |mode_table_entry| %>
+    {<%= mode_table_entry[:state_table_offset] %>},
+<% end %>
+};
+
+/**
+ * Execute user code associated with a lexer pattern.
+ *
+ * @param context
+ *   Lexer/parser context structure.
+ * @param code_id
+ *   The ID of the user code block to execute.
+ * @param match
+ *   Matched text for this pattern.
+ * @param match_length
+ *   Matched text length.
+ * @param out_token_info
+ *   Lexer token info in progress.
+ *
+ * @return Token to accept, or invalid token if the user code does
+ *   not explicitly return a token.
+ */
+static <%= @grammar.prefix %>token_t lexer_user_code(<%= @grammar.prefix %>context_t * context,
+    lexer_user_code_id_t code_id, uint8_t const * match,
+    size_t match_length, <%= @grammar.prefix %>token_info_t * out_token_info)
+{
+    switch (code_id)
+    {
+<% @grammar.patterns.each do |pattern| %>
+<%   if pattern.code_id %>
+    case <%= pattern.code_id %>u: {
+<%= expand_code(pattern.code, false, nil, pattern) %>
+    } break;
+<%   end %>
+<% end %>
+    default: break;
+    }
+
+    return INVALID_TOKEN_ID;
+}
+
+/**
+ * Check if there is a transition from the current lexer state to another
+ * based on the given input code point.
+ *
+ * @param current_state
+ *   Current lexer state.
+ * @param code_point
+ *   Input code point.
+ *
+ * @return Lexer state to transition to, or INVALID_LEXER_STATE_ID if none.
+ */
+static lexer_state_id_t check_lexer_transition(uint32_t current_state, uint32_t code_point)
+{
+    uint32_t transition_table_index = lexer_state_table[current_state].transition_table_index;
+    for (uint32_t i = 0u; i < lexer_state_table[current_state].n_transitions; i++)
+    {
+        if ((lexer_transition_table[transition_table_index + i].first <= code_point) &&
+            (code_point <= lexer_transition_table[transition_table_index + i].last))
+        {
+            return lexer_transition_table[transition_table_index + i].destination_state;
+        }
+    }
+    return INVALID_LEXER_STATE_ID;
+}
+
+/**
+ * Find the longest lexer pattern match at the current position.
+ *
+ * @param context
+ *   Lexer/parser context structure.
+ * @param[out] out_token_info
+ *   The lexed token information is stored here if the return value is
+ *   P_SUCCESS.
+ *
+ * @reval P_SUCCESS
+ *   A token was successfully lexed.
+ * @reval P_DECODE_ERROR
+ *   The decoder encountered invalid text encoding.
+ * @reval P_UNEXPECTED_INPUT
+ *   Input text does not match any lexer pattern.
+ * @retval P_EOF
+ *   The end of the text input was reached.
+ */
+static size_t find_longest_match(<%= @grammar.prefix %>context_t * context,
+    lexer_match_info_t * out_match_info, size_t * out_unexpected_input_length)
+{
+    lexer_match_info_t longest_match = {0};
+    lexer_match_info_t attempt_match = {0};
+    *out_match_info = longest_match;
+    uint32_t current_state = lexer_mode_table[context->mode].state_table_offset;
+    for (;;)
+    {
+        size_t const input_index = context->input_index + attempt_match.length;
+        uint8_t const * input = &context->input[input_index];
+        size_t input_length = context->input_length - input_index;
+        <%= @grammar.prefix %>code_point_t code_point;
+        uint8_t code_point_length;
+        size_t result = <%= @grammar.prefix %>decode_code_point(input, input_length, &code_point, &code_point_length);
+        switch (result)
+        {
+        case P_SUCCESS:
+            lexer_state_id_t transition_state = check_lexer_transition(current_state, code_point);
+            if (transition_state != INVALID_LEXER_STATE_ID)
+            {
+                attempt_match.length += code_point_length;
+                if (code_point == '\n')
+                {
+                    attempt_match.delta_position.row++;
+                    attempt_match.delta_position.col = 0u;
+                }
+                else
+                {
+                    attempt_match.delta_position.col++;
+                }
+                current_state = transition_state;
+                if (lexer_state_table[current_state].accepts)
+                {
+                    attempt_match.accepting_state = &lexer_state_table[current_state];
+                    longest_match = attempt_match;
+                }
+            }
+            else if (longest_match.length > 0)
+            {
+                *out_match_info = longest_match;
+                return P_SUCCESS;
+            }
+            else
+            {
+                *out_unexpected_input_length = attempt_match.length + code_point_length;
+                return P_UNEXPECTED_INPUT;
+            }
+            break;
+
+        case P_EOF:
+            /* We hit EOF. */
+            if (longest_match.length > 0)
+            {
+                /* We have a match, so use it. */
+                *out_match_info = longest_match;
+                return P_SUCCESS;
+            }
+            else if (attempt_match.length != 0)
+            {
+                /* There is a partial match - error! */
+                *out_unexpected_input_length = attempt_match.length;
+                return P_UNEXPECTED_INPUT;
+            }
+            else
+            {
+                /* Valid EOF return. */
+                return P_EOF;
+            }
+            break;
+
+        case P_DECODE_ERROR:
+            /* If we see a decode error, we may be partially in the middle of
+             * matching a pattern, so return the attempted match info so that
+             * the input text position can be updated. */
+            *out_match_info = attempt_match;
+            return result;
+
+        default:
+            return result;
+        }
+    }
+}
+
+/**
+ * Attempt to lex the next token in the input stream.
+ *
+ * @param context
+ *   Lexer/parser context structure.
+ * @param[out] out_token_info
+ *   The lexed token information is stored here if the return value is
+ *   P_SUCCESS.
+ *
+ * @reval P_SUCCESS
+ *   A token was successfully lexed.
+ * @reval P_DECODE_ERROR
+ *   The decoder encountered invalid text encoding.
+ * @reval P_UNEXPECTED_INPUT
+ *   Input text does not match any lexer pattern.
+ * @retval P_DROP
+ *   A drop pattern was matched so the lexer should continue.
+ */
+static size_t attempt_lex_token(<%= @grammar.prefix %>context_t * context, <%= @grammar.prefix %>token_info_t * out_token_info)
+{
+    <%= @grammar.prefix %>token_info_t token_info = {0};
+    token_info.position = context->text_position;
+    token_info.token = INVALID_TOKEN_ID;
+    *out_token_info = token_info; // TODO: remove
+    lexer_match_info_t match_info;
+    size_t unexpected_input_length;
+    size_t result = find_longest_match(context, &match_info, &unexpected_input_length);
+    switch (result)
+    {
+    case P_SUCCESS:
+        <%= @grammar.prefix %>token_t token_to_accept = match_info.accepting_state->token;
+        if (match_info.accepting_state->code_id != INVALID_USER_CODE_ID)
+        {
+            uint8_t const * match = &context->input[context->input_index];
+            <%= @grammar.prefix %>token_t user_code_token = lexer_user_code(context,
+                match_info.accepting_state->code_id, match, match_info.length, &token_info);
+            /* An invalid token returned from lexer_user_code() means that the
+             * user code did not explicitly return a token. So only override
+             * the token to return if the user code does explicitly return a
+             * token. */
+            if (user_code_token != INVALID_TOKEN_ID)
+            {
+                token_to_accept = user_code_token;
+            }
+        }
+
+        /* Update the input position tracking. */
+        context->input_index += match_info.length;
+        context->text_position.row += match_info.delta_position.row;
+        if (match_info.delta_position.row != 0u)
+        {
+            context->text_position.col = match_info.delta_position.col;
+        }
+        else
+        {
+            context->text_position.col += match_info.delta_position.col;
+        }
+
+        if (token_to_accept == INVALID_TOKEN_ID)
+        {
+            return P_DROP;
+        }
+        token_info.token = token_to_accept;
+        token_info.length = match_info.length;
+        *out_token_info = token_info;
+        return P_SUCCESS;
+
+    case P_EOF:
+        token_info.token = TOKEN___EOF;
+        *out_token_info = token_info;
+        return P_SUCCESS;
+
+    case P_DECODE_ERROR:
+        /* Update the input position tracking. */
+        context->input_index += match_info.length;
+        context->text_position.row += match_info.delta_position.row;
+        if (match_info.delta_position.row != 0u)
+        {
+            context->text_position.col = match_info.delta_position.col;
+        }
+        else
+        {
+            context->text_position.col += match_info.delta_position.col;
+        }
+        return result;
+
+    default:
+        return result;
+    }
+}
+
+/**
+ * Lex the next token in the input stream.
+ *
+ * @param context
+ *   Lexer/parser context structure.
+ * @param[out] out_token_info
+ *   The lexed token information is stored here if the return value is
+ *   P_SUCCESS.
+ *
+ * @reval P_SUCCESS
+ *   A token was successfully lexed.
+ * @reval P_DECODE_ERROR
+ *   The decoder encountered invalid text encoding.
+ * @reval P_UNEXPECTED_INPUT
+ *   Input text does not match any lexer pattern.
+ */
+size_t <%= @grammar.prefix %>lex(<%= @grammar.prefix %>context_t * context, <%= @grammar.prefix %>token_info_t * out_token_info)
+{
+    for (;;)
+    {
+        size_t result = attempt_lex_token(context, out_token_info);
+        if (result != P_DROP)
+        {
+            return result;
+        }
+    }
+}
+
+/**************************************************************************
+ * Parser
+ *************************************************************************/
+
+/** Reduce ID type. */
+typedef <%= get_type_for(@parser.reduce_table.size) %> reduce_id_t;
+
+/**
+ * A symbol ID can hold either a token ID or a rule set ID.
+ *
+ * Token IDs and rule set IDs share the same namespace, with rule set IDs
+ * beginning after token IDs end.
+ */
+typedef <%= get_type_for(@parser.rule_sets.map(&:last).map(&:id).max) %> symbol_id_t;
+
+/** Parser state ID type. */
+typedef <%= get_type_for(@parser.state_table.size) %> parser_state_id_t;
+
+/** Parser rule ID type. */
+typedef <%= get_type_for(@grammar.rules.size) %> rule_id_t;
+
+/** Parser shift ID type. */
+typedef <%= get_type_for(@parser.shift_table.size) %> shift_id_t;
+
+/** Shift table entry. */
+typedef struct
+{
+    /** Token or rule set ID. */
+    symbol_id_t symbol_id;
+
+    /** Parser state to shift to. */
+    parser_state_id_t state_id;
+} shift_t;
+
+/** Reduce table entry. */
+typedef struct
+{
+    /** Lookahead token. */
+    <%= @grammar.prefix %>token_t token;
+
+    /**
+     * Rule ID.
+     *
+     * This is used to execute the parser user code block associated with a
+     * grammar rule.
+     */
+    rule_id_t rule;
+
+    /**
+     * Rule set ID.
+     *
+     * This is used as the new top symbol ID of the parse stack after this
+     * reduce action.
+     */
+    symbol_id_t rule_set;
+
+    /**
+     * Number of states leading to this reduce action.
+     *
+     * This is the number of entries popped from the parse stack after this
+     * reduce action.
+     */
+    parser_state_id_t n_states;
+} reduce_t;
+
+/** Parser state entry. */
+typedef struct
+{
+    /** First shift table entry for this parser state. */
+    shift_id_t shift_table_index;
+
+    /** Number of shift table entries for this parser state. */
+    shift_id_t n_shift_entries;
+
+    /** First reduce table entry for this parser state. */
+    reduce_id_t reduce_table_index;
+
+    /** Number of reduce table entries for this parser state. */
+    reduce_id_t n_reduce_entries;
+} parser_state_t;
+
+/**
+ * Structure to hold a state ID and value pair.
+ *
+ * A stack of these structures makes up the parse stack.
+ */
+typedef struct
+{
+    /** Parser state ID. */
+    size_t state_id;
+
+    /** Parser value from this state. */
+    <%= @grammar.prefix %>value_t pvalue;
+} state_value_t;
+
+/** Parser shift table. */
+static const shift_t parser_shift_table[] = {
+<%   @parser.shift_table.each do |shift| %>
+    {<%= shift[:symbol_id] %>u, <%= shift[:state_id] %>u},
+<%   end %>
+};
+
+/** Parser reduce table. */
+static const reduce_t parser_reduce_table[] = {
+<%   @parser.reduce_table.each do |reduce| %>
+    {<%= reduce[:token_id] %>u, <%= reduce[:rule_id] %>u, <%= reduce[:rule_set_id] %>u, <%= reduce[:n_states] %>u},
+<%   end %>
+};
+
+/** Parser state table. */
+static const parser_state_t parser_state_table[] = {
+<%   @parser.state_table.each do |state| %>
+    {<%= state[:shift_index] %>u, <%= state[:n_shifts] %>u, <%= state[:reduce_index] %>u, <%= state[:n_reduces] %>u},
+<%   end %>
+};
+
+/* state_values stack functionality */
+
+/** state_values stack type. */
+typedef struct
+{
+    size_t length;
+    size_t capacity;
+    state_value_t * entries;
+} state_values_stack_t;
+
+/**
+ * Initialize state_values stack structure.
+ *
+ * @param stack
+ *   state_values stack structure.
+ */
+static void state_values_stack_init(state_values_stack_t * stack)
+{
+    const size_t initial_capacity = 10u;
+    stack->length = 0u;
+    stack->capacity = initial_capacity;
+    stack->entries = (state_value_t *)malloc(initial_capacity * sizeof(state_value_t));
+}
+
+/**
+ * Index a state_values stack.
+ *
+ * @param stack
+ *   state_values stack structure.
+ * @param index
+ *   Index to the stack.
+ *
+ * @return Pointer to the state value structure at the given index.
+ */
+static state_value_t * state_values_stack_index(state_values_stack_t * stack, int index)
+{
+    if (index >= 0)
+    {
+        return &stack->entries[index];
+    }
+    else
+    {
+        return &stack->entries[stack->length - (size_t)(unsigned int)(-index)];
+    }
+}
+
+/**
+ * Push a new state_value to the state_values stack.
+ *
+ * @param stack
+ *   state_values stack structure.
+ */
+static void state_values_stack_push(state_values_stack_t * stack)
+{
+    size_t const current_capacity = stack->capacity;
+    size_t const current_length = stack->length;
+    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));
+        memcpy(new_entries, stack->entries, current_length * sizeof(state_value_t));
+        free(stack->entries);
+        stack->capacity = new_capacity;
+        stack->entries = new_entries;
+    }
+    memset(&stack->entries[current_length], 0, sizeof(state_value_t));
+    stack->length = current_length + 1u;
+}
+
+/**
+ * Pop entries from a state_values stack.
+ *
+ * @param stack
+ *   state_values stack structure.
+ * @param n
+ *   Number of states to pop.
+ */
+static void state_values_stack_pop(state_values_stack_t * stack, size_t n)
+{
+    stack->length -= n;
+}
+
+/**
+ * Free memory for a state_values stack structure.
+ *
+ * @param stack
+ *   state_values stack structure.
+ */
+static void state_values_stack_free(state_values_stack_t * stack)
+{
+     free(stack->entries);
+}
+
+/**
+ * Execute user code associated with a parser rule.
+ *
+ * @param rule The ID of the rule.
+ *
+ * @return Parse value.
+ */
+static <%= @grammar.prefix %>value_t parser_user_code(uint32_t rule, state_values_stack_t * statevalues, uint32_t n_states)
+{
+    <%= @grammar.prefix %>value_t _pvalue = {0};
+
+    switch (rule)
+    {
+<% @grammar.rules.each do |rule| %>
+<%   if rule.code %>
+    case <%= rule.id %>u: {
+<%= expand_code(rule.code, true, rule, nil) %>
+    } break;
+<%   end %>
+<% end %>
+    default: break;
+    }
+
+    return _pvalue;
+}
+
+/**
+ * Check if the parser should shift to a new state.
+ *
+ * @param state_id
+ *   Parser state ID.
+ * @param symbol_id
+ *   Incoming token/rule set ID.
+ *
+ * @return State to shift to, or INVALID_ID if none.
+ */
+static size_t check_shift(size_t state_id, size_t symbol_id)
+{
+    uint32_t start = parser_state_table[state_id].shift_table_index;
+    uint32_t end = start + parser_state_table[state_id].n_shift_entries;
+    for (uint32_t i = start; i < end; i++)
+    {
+        if (parser_shift_table[i].symbol_id == symbol_id)
+        {
+            return parser_shift_table[i].state_id;
+        }
+    }
+    return INVALID_ID;
+}
+
+/**
+ * Check if the parser should reduce to a new state.
+ *
+ * @param state_id
+ *   Parser state ID.
+ * @param token
+ *   Incoming token.
+ *
+ * @return State to reduce to, or INVALID_ID if none.
+ */
+static size_t check_reduce(size_t state_id, <%= @grammar.prefix %>token_t token)
+{
+    size_t start = parser_state_table[state_id].reduce_table_index;
+    size_t end = start + parser_state_table[state_id].n_reduce_entries;
+    for (size_t i = start; i < end; i++)
+    {
+        if ((parser_reduce_table[i].token == token) ||
+            (parser_reduce_table[i].token == INVALID_TOKEN_ID))
+        {
+            return i;
+        }
+    }
+    return INVALID_ID;
+}
+
+/**
+ * Run the parser.
+ *
+ * @param context
+ *   Lexer/parser context structure.
+ *
+ * @retval P_SUCCESS
+ *   The parser successfully matched the input text. The parse result value
+ *   can be accessed with <%= @grammar.prefix %>result().
+ * @retval P_UNEXPECTED_TOKEN
+ *   An unexpected token was encountered that does not match any grammar rule.
+ *   The value context->token holds the unexpected token.
+ * @reval P_DECODE_ERROR
+ *   The decoder encountered invalid text encoding.
+ * @reval P_UNEXPECTED_INPUT
+ *   Input text does not match any lexer pattern.
+ */
+size_t <%= @grammar.prefix %>parse(<%= @grammar.prefix %>context_t * context)
+{
+    <%= @grammar.prefix %>token_info_t token_info;
+    <%= @grammar.prefix %>token_t token = INVALID_TOKEN_ID;
+    state_values_stack_t statevalues;
+    size_t reduced_rule_set = INVALID_ID;
+    <%= @grammar.prefix %>value_t reduced_parser_value;
+    state_values_stack_init(&statevalues);
+    state_values_stack_push(&statevalues);
+    size_t result;
+    for (;;)
+    {
+        if (token == INVALID_TOKEN_ID)
+        {
+            size_t lexer_result = <%= @grammar.prefix %>lex(context, &token_info);
+            if (lexer_result != P_SUCCESS)
+            {
+                result = lexer_result;
+                break;
+            }
+            token = token_info.token;
+        }
+        size_t shift_state = INVALID_ID;
+        if (reduced_rule_set != INVALID_ID)
+        {
+            shift_state = check_shift(state_values_stack_index(&statevalues, -1)->state_id, reduced_rule_set);
+        }
+        if (shift_state == INVALID_ID)
+        {
+            shift_state = check_shift(state_values_stack_index(&statevalues, -1)->state_id, token);
+            if ((shift_state != INVALID_ID) && (token == TOKEN___EOF))
+            {
+                /* Successful parse. */
+                context->parse_result = state_values_stack_index(&statevalues, -1)->pvalue;
+                result = P_SUCCESS;
+                break;
+            }
+        }
+        if (shift_state != INVALID_ID)
+        {
+            /* We have something to shift. */
+            state_values_stack_push(&statevalues);
+            state_values_stack_index(&statevalues, -1)->state_id = shift_state;
+            if (reduced_rule_set == INVALID_ID)
+            {
+                /* We shifted a token, mark it consumed. */
+                token = INVALID_TOKEN_ID;
+                state_values_stack_index(&statevalues, -1)->pvalue = token_info.pvalue;
+            }
+            else
+            {
+                /* We shifted a RuleSet. */
+                state_values_stack_index(&statevalues, -1)->pvalue = reduced_parser_value;
+                <%= @grammar.prefix %>value_t new_parse_result = {0};
+                reduced_parser_value = new_parse_result;
+                reduced_rule_set = INVALID_ID;
+            }
+            continue;
+        }
+
+        size_t reduce_index = check_reduce(state_values_stack_index(&statevalues, -1)->state_id, token);
+        if (reduce_index != INVALID_ID)
+        {
+            /* We have something to reduce. */
+            reduced_parser_value = parser_user_code(parser_reduce_table[reduce_index].rule, &statevalues, parser_reduce_table[reduce_index].n_states);
+            reduced_rule_set = parser_reduce_table[reduce_index].rule_set;
+            state_values_stack_pop(&statevalues, parser_reduce_table[reduce_index].n_states);
+            continue;
+        }
+
+        /* A token was successfully lexed, so the input text position was
+         * advanced. However, this is an unexpected token, so we want to reset
+         * the context text position to point to the token rather than the text
+         * after it, so that if the caller wants to report the error position,
+         * it will point to the correct position of the unexpected token. */
+        context->text_position = token_info.position;
+        context->token = token;
+        result = P_UNEXPECTED_TOKEN;
+        break;
+    }
+    state_values_stack_free(&statevalues);
+    return result;
+}
+
+/**
+ * Get the parse result value.
+ *
+ * @param context
+ *   Lexer/parser context structure.
+ *
+ * @return Parse result value.
+ */
+<%= start_rule_type[1] %> <%= @grammar.prefix %>result(<%= @grammar.prefix %>context_t * context)
+{
+    return context->parse_result.v_<%= start_rule_type[0] %>;
+}
+
+/**
+ * Get the current text input position.
+ *
+ * @param context
+ *   Lexer/parser context structure.
+ *
+ * @return Current text position.
+ */
+<%= @grammar.prefix %>position_t <%= @grammar.prefix %>position(<%= @grammar.prefix %>context_t * context)
+{
+    return context->text_position;
+}

assets/parser.h.erb

@@ -0,0 +1,122 @@
+/**
+ * @file
+ *
+ * This file is generated by Propane.
+ */
+
+#pragma once
+
+#include <stdint.h>
+#include <stddef.h>
+
+/**************************************************************************
+ * Public types
+ *************************************************************************/
+
+/* Result codes. */
+#define <%= @grammar.prefix.upcase %>SUCCESS 0u
+#define <%= @grammar.prefix.upcase %>DECODE_ERROR 1u
+#define <%= @grammar.prefix.upcase %>UNEXPECTED_INPUT 2u
+#define <%= @grammar.prefix.upcase %>UNEXPECTED_TOKEN 3u
+#define <%= @grammar.prefix.upcase %>DROP 4u
+#define <%= @grammar.prefix.upcase %>EOF 5u
+
+/** Token type. */
+typedef <%= get_type_for(@grammar.invalid_token_id) %> <%= @grammar.prefix %>token_t;
+
+/** Token IDs. */
+<% @grammar.tokens.each_with_index do |token, index| %>
+#define TOKEN_<%= token.code_name %> <%= index %>u
+<%   unless token.id == index %>
+<%     raise "Token ID (#{token.id}) does not match index (#{index}) for token #{token.name}!" %>
+<%   end %>
+<% end %>
+#define INVALID_TOKEN_ID <%= @grammar.invalid_token_id %>u
+
+/** Code point type. */
+typedef uint32_t <%= @grammar.prefix %>code_point_t;
+
+/** Parser values type(s). */
+typedef union
+{
+<% @grammar.ptypes.each do |name, typestring| %>
+    <%= typestring %> v_<%= name %>;
+<% end %>
+} <%= @grammar.prefix %>value_t;
+
+/**
+ * A structure to keep track of parser position.
+ *
+ * This is useful for reporting errors, etc...
+ */
+typedef struct
+{
+    /** Input text row (0-based). */
+    uint32_t row;
+
+    /** Input text column (0-based). */
+    uint32_t col;
+} <%= @grammar.prefix %>position_t;
+
+/** Lexed token information. */
+typedef struct
+{
+    /** Text position where the token was found. */
+    <%= @grammar.prefix %>position_t position;
+
+    /** Number of input bytes used by the token. */
+    size_t length;
+
+    /** Token that was lexed. */
+    <%= @grammar.prefix %>token_t token;
+
+    /** Parser value associated with the token. */
+    <%= @grammar.prefix %>value_t pvalue;
+} <%= @grammar.prefix %>token_info_t;
+
+/**
+ * Lexer and parser context.
+ *
+ * The user must allocate an instance of this structure and pass it to any
+ * public API function.
+ */
+typedef struct
+{
+    /* Lexer context data. */
+
+    /** Input text. */
+    uint8_t const * input;
+
+    /** Input text length. */
+    size_t input_length;
+
+    /** Input text index (byte offset). */
+    size_t input_index;
+
+    /** Input text position (row/column). */
+    <%= @grammar.prefix %>position_t text_position;
+
+    /** Current lexer mode. */
+    size_t mode;
+
+    /* Parser context data. */
+
+    /** Parse result value. */
+    <%= @grammar.prefix %>value_t parse_result;
+
+    /** Unexpected token received. */
+    <%= @grammar.prefix %>token_t token;
+} <%= @grammar.prefix %>context_t;
+
+void <%= @grammar.prefix %>context_init(<%= @grammar.prefix %>context_t * context, uint8_t const * input, size_t input_length);
+
+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);
+
+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);
+
+<%= start_rule_type[1] %> <%= @grammar.prefix %>result(<%= @grammar.prefix %>context_t * context);
+
+<%= @grammar.prefix %>position_t <%= @grammar.prefix %>position(<%= @grammar.prefix %>context_t * context);

lib/propane/generator.rb

@@ -11,14 +11,27 @@ class Propane
         @log = StringIO.new
       end
       @classname = @grammar.classname || File.basename(output_file).sub(%r{[^a-zA-Z0-9].*}, "").capitalize
+      @language =
+        if output_file =~ /\.([a-z]+)$/
+          $1
+        else
+          "d"
+        end
       process_grammar!
     end
 
     def generate
-      erb = ERB.new(File.read(File.join(File.dirname(File.expand_path(__FILE__)), "../../assets/parser.d.erb")), trim_mode: "<>")
-      result = erb.result(binding.clone)
-      File.open(@output_file, "wb") do |fh|
-        fh.write(result)
+      extensions = [@language]
+      if @language == "c"
+        extensions += %w[h]
+      end
+      extensions.each do |extension|
+        erb = ERB.new(File.read(File.join(File.dirname(File.expand_path(__FILE__)), "../../assets/parser.#{extension}.erb")), trim_mode: "<>")
+        output_file = @output_file.sub(%r{\.[a-z]+$}, ".#{extension}")
+        result = erb.result(binding.clone)
+        File.open(output_file, "wb") do |fh|
+          fh.write(result)
+        end
       end
       @log.close
     end
@@ -191,11 +204,21 @@ class Propane
         end
         code = code.gsub(/\$(\d+)/) do |match|
           index = $1.to_i
-          "statevalues[$-1-n_states+#{index}].pvalue.v_#{rule.components[index - 1].ptypename}"
+          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
         end
       else
         code = code.gsub(/\$\$/) do |match|
-          "out_token_info.pvalue.v_#{pattern.ptypename}"
+          case @language
+          when "c"
+            "out_token_info->pvalue.v_#{pattern.ptypename}"
+          when "d"
+            "out_token_info.pvalue.v_#{pattern.ptypename}"
+          end
         end
         code = code.gsub(/\$mode\(([a-zA-Z_][a-zA-Z_0-9]*)\)/) do |match|
           mode_name = $1
@@ -203,7 +226,12 @@ class Propane
           unless mode_id
             raise Error.new("Lexer mode '#{mode_name}' not found")
           end
-          "context.mode = #{mode_id}u"
+          case @language
+          when "c"
+            "context->mode = #{mode_id}u"
+          when "d"
+            "context.mode = #{mode_id}u"
+          end
         end
       end
       code
@@ -229,11 +257,26 @@ class Propane
     #   Type.
     def get_type_for(max)
       if max <= 0xFF
-        "ubyte"
+        case @language
+        when "c"
+          "uint8_t"
+        when "d"
+          "ubyte"
+        end
       elsif max <= 0xFFFF
-        "ushort"
+        case @language
+        when "c"
+          "uint16_t"
+        when "d"
+          "ushort"
+        end
       else
-        "uint"
+        case @language
+        when "c"
+          "uint32_t"
+        else
+          "uint"
+        end
       end
     end
 

spec/propane_spec.rb

@@ -11,7 +11,7 @@ describe Propane do
 
   def build_parser(options = {})
     options[:name] ||= ""
-    command = %W[./propane.sh spec/run/testparser#{options[:name]}.propane spec/run/testparser#{options[:name]}.d --log spec/run/testparser#{options[:name]}.log]
+    command = %W[./propane.sh spec/run/testparser#{options[:name]}.propane spec/run/testparser#{options[:name]}.#{options[:language]} --log spec/run/testparser#{options[:name]}.log]
     if (options[:capture])
       stdout, stderr, status = Open3.capture3(*command)
       Results.new(stdout, stderr, status)
@@ -25,9 +25,14 @@ describe Propane do
     test_files = Array(test_files)
     options[:parsers] ||= [""]
     parsers = options[:parsers].map do |name|
-      "spec/run/testparser#{name}.d"
+      "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")
+    when "d"
+      result = system(*%w[ldc2 --unittest -of spec/run/testparser -Ispec], *parsers, *test_files, "spec/testutils.d")
     end
-    result = system(*%w[ldc2 --unittest -of spec/run/testparser -Ispec], *parsers, *test_files, "spec/testutils.d")
     expect(result).to be_truthy
   end
 
@@ -69,8 +74,12 @@ describe Propane do
     FileUtils.mkdir_p("spec/run")
   end
 
-  it "generates a lexer" do
-    write_grammar <<EOF
+  %w[d c].each do |language|
+
+    context "#{language.upcase} language" do
+
+      it "generates a lexer" do
+        write_grammar <<EOF
 token int /\\d+/;
 token plus /\\+/;
 token times /\\*/;
@@ -81,15 +90,33 @@ Foo -> int <<
 Foo -> plus <<
 >>
 EOF
-    build_parser
-    compile("spec/test_lexer.d")
-    results = run
-    expect(results.stderr).to eq ""
-    expect(results.status).to eq 0
-  end
+        build_parser(language: language)
+        compile("spec/test_lexer.#{language}", language: language)
+        results = run
+        expect(results.stderr).to eq ""
+        expect(results.status).to eq 0
+      end
 
-  it "detects a lexer error when an unknown character is seen" do
-    write_grammar <<EOF
+      it "detects a lexer error when an unknown character is seen" do
+        case language
+        when "c"
+          write_grammar <<EOF
+ptype int;
+token int /\\d+/ <<
+  int v = 0;
+  for (size_t i = 0u; i < match_length; i++)
+  {
+    v *= 10;
+    v += (match[i] - '0');
+  }
+  $$ = v;
+>>
+Start -> int <<
+  $$ = $1;
+>>
+EOF
+        when "d"
+          write_grammar <<EOF
 ptype int;
 token int /\\d+/ <<
   int v;
@@ -104,15 +131,16 @@ Start -> int <<
   $$ = $1;
 >>
 EOF
-    build_parser
-    compile("spec/test_lexer_unknown_character.d")
-    results = run
-    expect(results.stderr).to eq ""
-    expect(results.status).to eq 0
-  end
+        end
+        build_parser(language: language)
+        compile("spec/test_lexer_unknown_character.#{language}", language: language)
+        results = run
+        expect(results.stderr).to eq ""
+        expect(results.status).to eq 0
+      end
 
-  it "generates a parser" do
-    write_grammar <<EOF
+      it "generates a parser" do
+        write_grammar <<EOF
 token plus /\\+/;
 token times /\\*/;
 token zero /0/;
@@ -124,11 +152,65 @@ E -> B;
 B -> zero;
 B -> one;
 EOF
-    build_parser
-  end
+        build_parser(language: language)
+      end
 
-  it "generates a parser that does basic math - user guide example" do
-    write_grammar <<EOF
+      it "generates a parser that does basic math - user guide example" do
+        case language
+        when "c"
+          write_grammar <<EOF
+<<
+#include <math.h>
+>>
+
+ptype size_t;
+
+token plus /\\+/;
+token times /\\*/;
+token power /\\*\\*/;
+token integer /\\d+/ <<
+  size_t v = 0u;
+  for (size_t i = 0u; i < match_length; i++)
+  {
+    v *= 10;
+    v += (match[i] - '0');
+  }
+  $$ = v;
+>>
+token lparen /\\(/;
+token rparen /\\)/;
+drop /\\s+/;
+
+Start -> E1 <<
+  $$ = $1;
+>>
+E1 -> E2 <<
+  $$ = $1;
+>>
+E1 -> E1 plus E2 <<
+  $$ = $1 + $3;
+>>
+E2 -> E3 <<
+  $$ = $1;
+>>
+E2 -> E2 times E3 <<
+  $$ = $1 * $3;
+>>
+E3 -> E4 <<
+  $$ = $1;
+>>
+E3 -> E3 power E4 <<
+  $$ = (size_t)pow($1, $3);
+>>
+E4 -> integer <<
+  $$ = $1;
+>>
+E4 -> lparen E1 rparen <<
+  $$ = $2;
+>>
+EOF
+        when "d"
+          write_grammar <<EOF
 <<
 import std.math;
 >>
@@ -179,25 +261,26 @@ E4 -> lparen E1 rparen <<
   $$ = $2;
 >>
 EOF
-    build_parser
-    compile("spec/test_basic_math_grammar.d")
-    results = run
-    expect(results.stderr).to eq ""
-    expect(results.status).to eq 0
-  end
+        end
+        build_parser(language: language)
+        compile("spec/test_basic_math_grammar.#{language}", language: language)
+        results = run
+        expect(results.stderr).to eq ""
+        expect(results.status).to eq 0
+      end
 
-  it "generates an SLR parser" do
-    write_grammar <<EOF
+      it "generates an SLR parser" do
+        write_grammar <<EOF
 token one /1/;
 Start -> E;
 E -> one E;
 E -> one;
 EOF
-    build_parser
-  end
+        build_parser(language: language)
+      end
 
-  it "distinguishes between multiple identical rules with lookahead symbol" do
-    write_grammar <<EOF
+      it "distinguishes between multiple identical rules with lookahead symbol" do
+        write_grammar <<EOF
 token a;
 token b;
 Start -> R1 a;
@@ -205,14 +288,14 @@ Start -> R2 b;
 R1 -> a b;
 R2 -> a b;
 EOF
-    build_parser
-    compile("spec/test_parser_identical_rules_lookahead.d")
-    results = run
-    expect(results.status).to eq 0
-  end
+        build_parser(language: language)
+        compile("spec/test_parser_identical_rules_lookahead.#{language}", language: language)
+        results = run
+        expect(results.status).to eq 0
+      end
 
-  it "handles reducing a rule that could be arrived at from multiple states" do
-    write_grammar <<EOF
+      it "handles reducing a rule that could be arrived at from multiple states" do
+        write_grammar <<EOF
 token a;
 token b;
 drop /\\s+/;
@@ -220,14 +303,29 @@ Start -> a R1;
 Start -> b R1;
 R1 -> b;
 EOF
-    build_parser
-    compile("spec/test_parser_rule_from_multiple_states.d")
-    results = run
-    expect(results.status).to eq 0
-  end
+        build_parser(language: language)
+        compile("spec/test_parser_rule_from_multiple_states.#{language}", language: language)
+        results = run
+        expect(results.status).to eq 0
+      end
 
-  it "executes user code when matching lexer token" do
-    write_grammar <<EOF
+      it "executes user code when matching lexer token" do
+        case language
+        when "c"
+          write_grammar <<EOF
+<<
+#include <stdio.h>
+>>
+token abc <<
+  printf("abc!\\n");
+>>
+token def;
+Start -> Abcs def;
+Abcs -> ;
+Abcs -> abc Abcs;
+EOF
+        when "d"
+          write_grammar <<EOF
 <<
 import std.stdio;
 >>
@@ -239,21 +337,35 @@ Start -> Abcs def;
 Abcs -> ;
 Abcs -> abc Abcs;
 EOF
-    build_parser
-    compile("spec/test_user_code.d")
-    results = run
-    expect(results.status).to eq 0
-    verify_lines(results.stdout, [
-      "abc!",
-      "pass1",
-      "abc!",
-      "abc!",
-      "pass2",
-    ])
-  end
+        end
+        build_parser(language: language)
+        compile("spec/test_user_code.#{language}", language: language)
+        results = run
+        expect(results.status).to eq 0
+        verify_lines(results.stdout, [
+          "abc!",
+          "pass1",
+          "abc!",
+          "abc!",
+          "pass2",
+        ])
+      end
 
-  it "supports a pattern statement" do
-    write_grammar <<EOF
+      it "supports a pattern statement" do
+        case language
+        when "c"
+          write_grammar <<EOF
+<<
+#include <stdio.h>
+>>
+token abc;
+/def/ <<
+  printf("def!\\n");
+>>
+Start -> abc;
+EOF
+        when "d"
+          write_grammar <<EOF
 <<
 import std.stdio;
 >>
@@ -263,21 +375,39 @@ token abc;
 >>
 Start -> abc;
 EOF
-    build_parser
-    compile("spec/test_pattern.d")
-    results = run
-    expect(results.status).to eq 0
-    verify_lines(results.stdout, [
-      "def!",
-      "pass1",
-      "def!",
-      "def!",
-      "pass2",
-    ])
-  end
+        end
+        build_parser(language: language)
+        compile("spec/test_pattern.#{language}", language: language)
+        results = run
+        expect(results.status).to eq 0
+        verify_lines(results.stdout, [
+          "def!",
+          "pass1",
+          "def!",
+          "def!",
+          "pass2",
+        ])
+      end
 
-  it "supports returning tokens from pattern code blocks" do
-    write_grammar <<EOF
+      it "supports returning tokens from pattern code blocks" do
+        case language
+        when "c"
+          write_grammar <<EOF
+<<
+#include <stdio.h>
+>>
+token abc;
+/def/ <<
+  printf("def!\\n");
+>>
+/ghi/ <<
+  printf("ghi!\\n");
+  return $token(abc);
+>>
+Start -> abc;
+EOF
+        when "d"
+          write_grammar <<EOF
 <<
 import std.stdio;
 >>
@@ -291,19 +421,44 @@ token abc;
 >>
 Start -> abc;
 EOF
-    build_parser
-    compile("spec/test_return_token_from_pattern.d")
-    results = run
-    expect(results.status).to eq 0
-    verify_lines(results.stdout, [
-      "def!",
-      "ghi!",
-      "def!",
-    ])
-  end
+        end
+        build_parser(language: language)
+        compile("spec/test_return_token_from_pattern.#{language}", language: language)
+        results = run
+        expect(results.status).to eq 0
+        verify_lines(results.stdout, [
+          "def!",
+          "ghi!",
+          "def!",
+        ])
+      end
 
-  it "supports lexer modes" do
-    write_grammar <<EOF
+      it "supports lexer modes" do
+        case language
+        when "c"
+          write_grammar <<EOF
+<<
+#include <stdio.h>
+>>
+token abc;
+token def;
+tokenid string;
+drop /\\s+/;
+/"/ <<
+  printf("begin string mode\\n");
+  $mode(string);
+>>
+string: /[^"]+/ <<
+  printf("captured string\\n");
+>>
+string: /"/ <<
+  $mode(default);
+  return $token(string);
+>>
+Start -> abc string def;
+EOF
+        when "d"
+          write_grammar <<EOF
 <<
 import std.stdio;
 >>
@@ -324,22 +479,42 @@ string: /"/ <<
 >>
 Start -> abc string def;
 EOF
-    build_parser
-    compile("spec/test_lexer_modes.d")
-    results = run
-    expect(results.status).to eq 0
-    verify_lines(results.stdout, [
-      "begin string mode",
-      "captured string",
-      "pass1",
-      "begin string mode",
-      "captured string",
-      "pass2",
-    ])
-  end
+        end
+        build_parser(language: language)
+        compile("spec/test_lexer_modes.#{language}", language: language)
+        results = run
+        expect(results.status).to eq 0
+        verify_lines(results.stdout, [
+          "begin string mode",
+          "captured string",
+          "pass1",
+          "begin string mode",
+          "captured string",
+          "pass2",
+        ])
+      end
 
-  it "executes user code associated with a parser rule" do
-    write_grammar <<EOF
+      it "executes user code associated with a parser rule" do
+        case language
+        when "c"
+          write_grammar <<EOF
+<<
+#include <stdio.h>
+>>
+token a;
+token b;
+Start -> A B <<
+  printf("Start!\\n");
+>>
+A -> a <<
+  printf("A!\\n");
+>>
+B -> b <<
+  printf("B!\\n");
+>>
+EOF
+        when "d"
+          write_grammar <<EOF
 <<
 import std.stdio;
 >>
@@ -355,20 +530,21 @@ B -> b <<
   writeln("B!");
 >>
 EOF
-    build_parser
-    compile("spec/test_parser_rule_user_code.d")
-    results = run
-    expect(results.status).to eq 0
-    verify_lines(results.stdout, [
-      "A!",
-      "B!",
-      "Start!",
-    ])
-  end
+        end
+        build_parser(language: language)
+        compile("spec/test_parser_rule_user_code.#{language}", language: language)
+        results = run
+        expect(results.status).to eq 0
+        verify_lines(results.stdout, [
+          "A!",
+          "B!",
+          "Start!",
+        ])
+      end
 
-  it "parses lists" do
-    write_grammar <<EOF
-ptype uint;
+      it "parses lists" do
+        write_grammar <<EOF
+ptype #{language == "c" ? "uint32_t" : "uint"};
 token a;
 Start -> As <<
   $$ = $1;
@@ -380,15 +556,15 @@ As -> As a <<
   $$ = $1 + 1u;
 >>
 EOF
-    build_parser
-    compile("spec/test_parsing_lists.d")
-    results = run
-    expect(results.status).to eq 0
-    expect(results.stderr).to eq ""
-  end
+        build_parser(language: language)
+        compile("spec/test_parsing_lists.#{language}", language: language)
+        results = run
+        expect(results.status).to eq 0
+        expect(results.stderr).to eq ""
+      end
 
-  it "fails to generate a parser for a LR(1) grammar that is not LALR" do
-    write_grammar <<EOF
+      it "fails to generate a parser for a LR(1) grammar that is not LALR" do
+        write_grammar <<EOF
 token a;
 token b;
 token c;
@@ -401,13 +577,29 @@ Start -> b E d;
 E -> e;
 F -> e;
 EOF
-    results = build_parser(capture: true)
-    expect(results.status).to_not eq 0
-    expect(results.stderr).to match %r{reduce/reduce conflict.*\(E\).*\(F\)}
-  end
+        results = build_parser(capture: true, language: language)
+        expect(results.status).to_not eq 0
+        expect(results.stderr).to match %r{reduce/reduce conflict.*\(E\).*\(F\)}
+      end
 
-  it "provides matched text to user code blocks" do
-    write_grammar <<EOF
+      it "provides matched text to user code blocks" do
+        case language
+        when "c"
+          write_grammar <<EOF
+<<
+#include <stdio.h>
+#include <stdlib.h>
+>>
+token id /[a-zA-Z_][a-zA-Z0-9_]*/ <<
+  char * t = malloc(match_length + 1);
+  strncpy(t, (char *)match, match_length);
+  printf("Matched token is %s\\n", t);
+  free(t);
+>>
+Start -> id;
+EOF
+        when "d"
+          write_grammar <<EOF
 <<
 import std.stdio;
 >>
@@ -416,18 +608,31 @@ token id /[a-zA-Z_][a-zA-Z0-9_]*/ <<
 >>
 Start -> id;
 EOF
-    build_parser
-    compile("spec/test_lexer_match_text.d")
-    results = run
-    expect(results.status).to eq 0
-    verify_lines(results.stdout, [
-      "Matched token is identifier_123",
-      "pass1",
-    ])
-  end
+        end
+        build_parser(language: language)
+        compile("spec/test_lexer_match_text.#{language}", language: language)
+        results = run
+        expect(results.status).to eq 0
+        verify_lines(results.stdout, [
+          "Matched token is identifier_123",
+          "pass1",
+        ])
+      end
 
-  it "allows storing a result value for the lexer" do
-    write_grammar <<EOF
+      it "allows storing a result value for the lexer" do
+        case language
+        when "c"
+          write_grammar <<EOF
+ptype uint64_t;
+token word /[a-z]+/ <<
+  $$ = match_length;
+>>
+Start -> word <<
+  $$ = $1;
+>>
+EOF
+        when "d"
+          write_grammar <<EOF
 ptype ulong;
 token word /[a-z]+/ <<
   $$ = match.length;
@@ -436,53 +641,56 @@ Start -> word <<
   $$ = $1;
 >>
 EOF
-    build_parser
-    compile("spec/test_lexer_result_value.d")
-    results = run
-    expect(results.stderr).to eq ""
-    expect(results.status).to eq 0
-  end
+        end
+        build_parser(language: language)
+        compile("spec/test_lexer_result_value.#{language}", language: language)
+        results = run
+        expect(results.stderr).to eq ""
+        expect(results.status).to eq 0
+      end
 
-  it "tracks position of parser errors" do
-    write_grammar <<EOF
+      it "tracks position of parser errors" do
+        write_grammar <<EOF
 token a;
 token num /\\d+/;
 drop /\\s+/;
 Start -> a num Start;
 Start -> a num;
 EOF
-    build_parser
-    compile("spec/test_error_positions.d")
-    results = run
-    expect(results.stderr).to eq ""
-    expect(results.status).to eq 0
-  end
+        build_parser(language: language)
+        compile("spec/test_error_positions.#{language}", language: language)
+        results = run
+        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.propane"))
-    build_parser
-    compile(["spec/test_parsing_json.d", "spec/json_types.d"])
-  end
+      it "allows creating a JSON parser" do
+        write_grammar(File.read("spec/json_parser.propane"))
+        build_parser(language: language)
+        compile(["spec/test_parsing_json.#{language}", "spec/json_types.#{language}"], language: language)
+      end
 
-  it "allows generating multiple parsers in the same program" do
-    write_grammar(<<EOF, name: "myp1")
+      it "allows generating multiple parsers in the same program" do
+        write_grammar(<<EOF, name: "myp1")
 prefix myp1_;
 token a;
 token num /\\d+/;
 drop /\\s+/;
 Start -> a num;
 EOF
-    build_parser(name: "myp1")
-    write_grammar(<<EOF, name: "myp2")
+        build_parser(name: "myp1", language: language)
+        write_grammar(<<EOF, name: "myp2")
 prefix myp2_;
 token b;
 token c;
 Start -> b c b;
 EOF
-    build_parser(name: "myp2")
-    compile("spec/test_multiple_parsers.d", parsers: %w[myp1 myp2])
-    results = run
-    expect(results.stderr).to eq ""
-    expect(results.status).to eq 0
+        build_parser(name: "myp2", language: language)
+        compile("spec/test_multiple_parsers.#{language}", parsers: %w[myp1 myp2], language: language)
+        results = run
+        expect(results.stderr).to eq ""
+        expect(results.status).to eq 0
+      end
+    end
   end
 end

spec/test_basic_math_grammar.c

@@ -0,0 +1,29 @@
+#include "testparser.h"
+#include "testutils.h"
+#include <string.h>
+
+int main()
+{
+    char const * input = "1 + 2 * 3 + 4";
+    p_context_t context;
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert_eq(P_SUCCESS, p_parse(&context));
+    assert_eq(11, p_result(&context));
+
+    input = "1 * 2 ** 4 * 3";
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert_eq(P_SUCCESS, p_parse(&context));
+    assert_eq(48, p_result(&context));
+
+    input = "(1 + 2) * 3 + 4";
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert_eq(P_SUCCESS, p_parse(&context));
+    assert_eq(13, p_result(&context));
+
+    input = "(2 * 2) ** 3 + 4 + 5";
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert_eq(P_SUCCESS, p_parse(&context));
+    assert_eq(73, p_result(&context));
+
+    return 0;
+}

spec/test_error_positions.c

@@ -0,0 +1,39 @@
+#include "testparser.h"
+#include <assert.h>
+#include <string.h>
+
+int main()
+{
+    char const * input = "a 42";
+    p_context_t context;
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_SUCCESS);
+
+    input = "a\n123\na  a";
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_UNEXPECTED_TOKEN);
+    assert(p_position(&context).row == 2);
+    assert(p_position(&context).col == 3);
+    assert(context.token == TOKEN_a);
+
+    input = "12";
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_UNEXPECTED_TOKEN);
+    assert(p_position(&context).row == 0);
+    assert(p_position(&context).col == 0);
+    assert(context.token == TOKEN_num);
+
+    input = "a 12\n\nab";
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_UNEXPECTED_INPUT);
+    assert(p_position(&context).row == 2);
+    assert(p_position(&context).col == 1);
+
+    input = "a 12\n\na\n\n77\na   \xAA";
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_DECODE_ERROR);
+    assert(p_position(&context).row == 5);
+    assert(p_position(&context).col == 4);
+
+    return 0;
+}

spec/test_error_positions.d

@@ -33,6 +33,5 @@ unittest
     input = "a 12\n\na\n\n77\na   \xAA";
     p_context_init(&context, input);
     assert(p_parse(&context) == P_DECODE_ERROR);
-    writeln(p_position(&context));
     assert(p_position(&context) == p_position_t(5, 4));
 }

spec/test_lexer.c

@@ -0,0 +1,92 @@
+#include "testparser.h"
+#include <assert.h>
+#include <string.h>
+
+int main()
+{
+    size_t result;
+    p_code_point_t code_point;
+    uint8_t code_point_length;
+
+    result = p_decode_code_point((uint8_t const *)"5", 1u, &code_point, &code_point_length);
+    assert(result == P_SUCCESS);
+    assert(code_point == '5');
+    assert(code_point_length == 1u);
+
+    result = p_decode_code_point((uint8_t const *)"", 0u, &code_point, &code_point_length);
+    assert(result == P_EOF);
+
+    result = p_decode_code_point((uint8_t const *)"\xC2\xA9", 2u, &code_point, &code_point_length);
+    assert(result == P_SUCCESS);
+    assert(code_point == 0xA9u);
+    assert(code_point_length == 2u);
+
+    result = p_decode_code_point((uint8_t const *)"\xf0\x9f\xa7\xa1", 4u, &code_point, &code_point_length);
+    assert(result == P_SUCCESS);
+    assert(code_point == 0x1F9E1u);
+    assert(code_point_length == 4u);
+
+    result = p_decode_code_point((uint8_t const *)"\xf0\x9f\x27", 3u, &code_point, &code_point_length);
+    assert(result == P_DECODE_ERROR);
+
+    result = p_decode_code_point((uint8_t const *)"\xf0\x9f\xa7\xFF", 4u, &code_point, &code_point_length);
+    assert(result == P_DECODE_ERROR);
+
+    result = p_decode_code_point((uint8_t const *)"\xfe", 1u, &code_point, &code_point_length);
+    assert(result == P_DECODE_ERROR);
+
+
+    p_token_info_t token_info;
+    char const * input = "5 + 4 * \n677 + 567";
+    p_context_t context;
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_lex(&context, &token_info) == P_SUCCESS);
+    assert(token_info.position.row == 0u);
+    assert(token_info.position.col == 0u);
+    assert(token_info.length == 1u);
+    assert(token_info.token == TOKEN_int);
+    assert(p_lex(&context, &token_info) == P_SUCCESS);
+    assert(token_info.position.row == 0u);
+    assert(token_info.position.col == 2u);
+    assert(token_info.length == 1u);
+    assert(token_info.token == TOKEN_plus);
+    assert(p_lex(&context, &token_info) == P_SUCCESS);
+    assert(token_info.position.row == 0u);
+    assert(token_info.position.col == 4u);
+    assert(token_info.length == 1u);
+    assert(token_info.token == TOKEN_int);
+    assert(p_lex(&context, &token_info) == P_SUCCESS);
+    assert(token_info.position.row == 0u);
+    assert(token_info.position.col == 6u);
+    assert(token_info.length == 1u);
+    assert(token_info.token == TOKEN_times);
+    assert(p_lex(&context, &token_info) == P_SUCCESS);
+    assert(token_info.position.row == 1u);
+    assert(token_info.position.col == 0u);
+    assert(token_info.length == 3u);
+    assert(token_info.token == TOKEN_int);
+    assert(p_lex(&context, &token_info) == P_SUCCESS);
+    assert(token_info.position.row == 1u);
+    assert(token_info.position.col == 4u);
+    assert(token_info.length == 1u);
+    assert(token_info.token == TOKEN_plus);
+    assert(p_lex(&context, &token_info) == P_SUCCESS);
+    assert(token_info.position.row == 1u);
+    assert(token_info.position.col == 6u);
+    assert(token_info.length == 3u);
+    assert(token_info.token == TOKEN_int);
+    assert(p_lex(&context, &token_info) == P_SUCCESS);
+    assert(token_info.position.row == 1u);
+    assert(token_info.position.col == 9u);
+    assert(token_info.length == 0u);
+    assert(token_info.token == TOKEN___EOF);
+
+    p_context_init(&context, (uint8_t const *)"", 0u);
+    assert(p_lex(&context, &token_info) == P_SUCCESS);
+    assert(token_info.position.row == 0u);
+    assert(token_info.position.col == 0u);
+    assert(token_info.length == 0u);
+    assert(token_info.token == TOKEN___EOF);
+
+    return 0;
+}

spec/test_lexer_match_text.c

@@ -0,0 +1,15 @@
+#include "testparser.h"
+#include <assert.h>
+#include <string.h>
+#include <stdio.h>
+
+int main()
+{
+    char const * input = "identifier_123";
+    p_context_t context;
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_SUCCESS);
+    printf("pass1\n");
+
+    return 0;
+}

spec/test_lexer_modes.c

@@ -0,0 +1,20 @@
+#include "testparser.h"
+#include <assert.h>
+#include <string.h>
+#include <stdio.h>
+
+int main()
+{
+    char const * input = "abc \"a string\" def";
+    p_context_t context;
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_SUCCESS);
+    printf("pass1\n");
+
+    input = "abc \"abc def\" def";
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_SUCCESS);
+    printf("pass2\n");
+
+    return 0;
+}

spec/test_lexer_result_value.c

@@ -0,0 +1,19 @@
+#include "testparser.h"
+#include <assert.h>
+#include <string.h>
+
+int main()
+{
+    char const * input = "x";
+    p_context_t context;
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_result(&context) == 1u);
+
+    input = "fabulous";
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_result(&context) == 8u);
+
+    return 0;
+}

spec/test_lexer_unknown_character.c

@@ -0,0 +1,18 @@
+#include "testparser.h"
+#include <assert.h>
+#include <string.h>
+
+int main()
+{
+    char const * input = "x";
+    p_context_t context;
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_UNEXPECTED_INPUT);
+
+    input = "123";
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_result(&context) == 123u);
+
+    return 0;
+}

spec/test_multiple_parsers.c

@@ -0,0 +1,19 @@
+#include "testparsermyp1.h"
+#include "testparsermyp2.h"
+#include <assert.h>
+#include <string.h>
+
+int main()
+{
+    char const * input1 = "a\n1";
+    myp1_context_t context1;
+    myp1_context_init(&context1, (uint8_t const *)input1, strlen(input1));
+    assert(myp1_parse(&context1) == MYP1_SUCCESS);
+
+    char const * input2 = "bcb";
+    myp2_context_t context2;
+    myp2_context_init(&context2, (uint8_t const *)input2, strlen(input2));
+    assert(myp2_parse(&context2) == MYP2_SUCCESS);
+
+    return 0;
+}

spec/test_parser_identical_rules_lookahead.c

@@ -0,0 +1,17 @@
+#include "testparser.h"
+#include <string.h>
+#include <assert.h>
+
+int main()
+{
+    char const * input = "aba";
+    p_context_t context;
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_SUCCESS);
+
+    input = "abb";
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_SUCCESS);
+
+    return 0;
+}

spec/test_parser_rule_from_multiple_states.c

@@ -0,0 +1,24 @@
+#include "testparser.h"
+#include <assert.h>
+#include <string.h>
+
+int main()
+{
+    char const * input = "a";
+    p_context_t context;
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_UNEXPECTED_TOKEN);
+    assert(p_position(&context).row == 0);
+    assert(p_position(&context).col == 1);
+    assert(context.token == TOKEN___EOF);
+
+    input = "a b";
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_SUCCESS);
+
+    input = "bb";
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_SUCCESS);
+
+    return 0;
+}

spec/test_parser_rule_user_code.c

@@ -0,0 +1,13 @@
+#include "testparser.h"
+#include <assert.h>
+#include <string.h>
+
+int main()
+{
+    char const * input = "ab";
+    p_context_t context;
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_SUCCESS);
+
+    return 0;
+}

spec/test_parsing_lists.c

@@ -0,0 +1,24 @@
+#include "testparser.h"
+#include <assert.h>
+#include <string.h>
+
+int main()
+{
+    char const * input = "a";
+    p_context_t context;
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_result(&context) == 1u);
+
+    input = "";
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_result(&context) == 0u);
+
+    input = "aaaaaaaaaaaaaaaa";
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_SUCCESS);
+    assert(p_result(&context) == 16u);
+
+    return 0;
+}

spec/test_pattern.c

@@ -0,0 +1,20 @@
+#include "testparser.h"
+#include <stdio.h>
+#include <assert.h>
+#include <string.h>
+
+int main()
+{
+    char const * input = "abcdef";
+    p_context_t context;
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_SUCCESS);
+    printf("pass1\n");
+
+    input = "defabcdef";
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_SUCCESS);
+    printf("pass2\n");
+
+    return 0;
+}

spec/test_return_token_from_pattern.c

@@ -0,0 +1,13 @@
+#include "testparser.h"
+#include <assert.h>
+#include <string.h>
+
+int main()
+{
+    char const * input = "defghidef";
+    p_context_t context;
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_SUCCESS);
+
+    return 0;
+}

spec/test_user_code.c

@@ -0,0 +1,20 @@
+#include "testparser.h"
+#include <assert.h>
+#include <stdio.h>
+#include <string.h>
+
+int main()
+{
+    char const * input = "abcdef";
+    p_context_t context;
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_SUCCESS);
+    printf("pass1\n");
+
+    input = "abcabcdef";
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_SUCCESS);
+    printf("pass2\n");
+
+    return 0;
+}

spec/testutils.c

@@ -0,0 +1,12 @@
+#include <stdio.h>
+#include <assert.h>
+#include <stdbool.h>
+
+void assert_eq_size_t_i(size_t expected, size_t actual, char const * file, size_t line)
+{
+    if (expected != actual)
+    {
+        fprintf(stderr, "%s:%lu: expected %lu, got %lu\n", file, line, expected, actual);
+        assert(false);
+    }
+}

spec/testutils.h

@@ -0,0 +1,7 @@
+#pragma once
+
+void assert_eq_size_t_i(size_t expected, size_t actual, char const * file, size_t line);
+
+#define assert_eq(expected, actual) \
+    assert_eq_size_t_i(expected, actual, __FILE__, __LINE__)
+