ccm/lib/parser.c

#include "parser.h"

#define CCM_TRY(rule) parser_try_new_rule(self, rule)
#define CCM_TRY_LL1(rule, node) \
parser_try_new_rule_ll1(self, rule, node)

void parser_init(parser_t* self, lexer_t* lexer)
{
    assert(self);
    assert(lexer);

    self->lexer = lexer;
    err_init(&self->err);
    self->current = NULL;
}

void parser_free(parser_t* self)
{
    err_free(&self->err);
}

node_t* parser_try_new_parse(parser_t* self)
{
    assert(self);
    node_t* res = CCM_TRY(parser_try_new_module);

    if (self->lexer->cursor < (ssize_t)strlen(self->lexer->source))
    {
        str_t s;
        str_init(&s);

        while (self->lexer->cursor
            < (ssize_t)strlen(self->lexer->source)
            && !isspace(self->lexer->source[self->lexer->cursor]))
        {
            str_push(&s, self->lexer->source[self->lexer->cursor]);
            self->lexer->cursor++;
        }

        err_push(&self->err, res ? res->line : 0,
                 "unexpected end after '%s'", s.value);
        str_free(&s);

    }

    return res;
}

node_t* parser_try_new_rule_ll1(parser_t* self,
                                rule_ll1_t rule,
                                node_t* node)
{
    if (!err_is_ok(&self->err))
    {
        return NULL;
    }

    lexer_state_t state = lexer_state(self->lexer);

    node_t* result = rule(self, node);
    if (result) { return result; }

    lexer_restore(self->lexer, state);

    return (void*) NULL;
}

node_t* parser_try_new_rule(parser_t* self, rule_t rule)
{
    if (!err_is_ok(&self->err))
    {
        return NULL;
    }

    lexer_state_t state = lexer_state(self->lexer);

    node_t* result = rule(self);
    if (result) { return result; }

    lexer_restore(self->lexer, state);

    return (void*) NULL;
}

int parser_ensure(parser_t* self, node_t* node, NodeKind kind)
{
    assert(self);

    if (!node)
    {
        err_push(&self->err, self->lexer->line,
                 "expected token '%s', got nothing",
                 NodeKindStr[kind] + strlen("NODE_"));
        return 0;
    }

    if (node->kind != kind)
    {
        err_push(&self->err, self->lexer->line,
                 "expected token '%s', got '%s'",
                 NodeKindStr[kind] + strlen("NODE_"),
                 NodeKindStr[node->kind] + strlen("NODE_"));
        return 0;
    }

    return 1;
}

node_t* parser_try_new_module(parser_t* self)
{
    assert(self);
    node_t* module = malloc(sizeof(node_t));
    node_init(module, NODE_MODULE, "", self->lexer->line);

    node_t* node = NULL;

    do {
        node = CCM_TRY(parser_try_new_expr);

        if (!node)
        {
            node_free(module);
            free(module);
            return NULL;
        }

        node_push_new_child(module, node);
        lexer_skip_spaces(self->lexer);
    } while(self->lexer->cursor < (ssize_t) strlen(self->lexer->source));
    return module;

}

node_t* parser_try_new_expr(parser_t* self)
{
    assert(self);

    if (lexer_peek_kind(self->lexer, NODE_VAR, 0))
    {
        return CCM_TRY(parser_try_new_decl);
    }

    if (lexer_peek_kind(self->lexer, NODE_ASSERT_EQ, 0)
        || lexer_peek_kind(self->lexer, NODE_ASSERT_NE, 0))
    {
        return CCM_TRY(parser_try_new_assert);
    }

    return CCM_TRY(parser_try_new_or);
}

node_t* parser_try_new_decl(parser_t* self)
{
    assert(self);
    if (!lexer_consume_next(self->lexer, NODE_VAR))
    {
        return NULL;
    }

    if (!lexer_peek_kind(self->lexer, NODE_IDENT, 0))
    {
        return NULL;
    }

    node_t* ident = lexer_try_new_next(self->lexer);

    if (!lexer_consume_next(self->lexer, NODE_ASSIGN))
    {
        node_free(ident); free(ident);
        return NULL;
    }

    node_t* expr = CCM_TRY(parser_try_new_expr);
    if (!expr)
    {
        node_free(ident); free(ident);
        return NULL;
    }

    node_t* node = malloc(sizeof(node_t));
    node_init(node, NODE_VARDECL, "", self->lexer->line);

    node_push_new_child(node, ident);
    node_push_new_child(node, expr);

    return node;
}

node_t* parser_try_new_assert(parser_t* self)
{
    assert(self);
    node_t* node = malloc(sizeof(node_t));

    if (lexer_peek_kind(self->lexer, NODE_ASSERT_EQ, 0))
    {
        lexer_consume_next(self->lexer, NODE_ASSERT_EQ);
        node_init(node, NODE_ASSERT_EQ, "", self->lexer->line);
    }
    else if (lexer_peek_kind(self->lexer, NODE_ASSERT_NE, 0))
    {
        lexer_consume_next(self->lexer, NODE_ASSERT_NE);
        node_init(node, NODE_ASSERT_NE, "", self->lexer->line);
    }
    else
    {
        free(node);
        return NULL;
    }


    node_t* tuple = CCM_TRY(parser_try_new_tuple);

    if (!tuple)
    {
        node_free(node);
        free(node);
        return NULL;
    }

    node_push_new_child(node, tuple);

    return node;
}

node_t* parser_try_new_or(parser_t* self)
{
    assert(self);
    node_t* lhs = CCM_TRY(parser_try_new_and);
    if (!lhs) { return NULL; }

    while (lexer_peek_kind(self->lexer, NODE_OR, 0))
    {
        lexer_consume_next(self->lexer, NODE_OR);
        node_t* node = malloc(sizeof(node_t));
        node_init(node, NODE_OR, "", lhs->line);

        node_push_new_child(node, lhs);
        node_t* rhs = CCM_TRY(parser_try_new_and);

        if (!rhs)
        {
            node_free(lhs);
            free(lhs);
            node_free(node);
            free(node);
            return NULL;
        }
        node_push_new_child(node, rhs);
        lhs = node;
    }

    return lhs;
}

node_t* parser_try_new_and(parser_t* self)
{
    assert(self);
    node_t* lhs = CCM_TRY(parser_try_new_eqne);
    if (!lhs) { return NULL; }

    while (lexer_peek_kind(self->lexer, NODE_AND, 0))
    {
        lexer_consume_next(self->lexer, NODE_AND);
        node_t* node = malloc(sizeof(node_t));
        node_init(node, NODE_AND, "", lhs->line);

        node_push_new_child(node, lhs);
        node_t* rhs = CCM_TRY(parser_try_new_eqne);

        if (!rhs)
        {
            node_free(lhs);
            free(lhs);
            node_free(node);
            free(node);
            return NULL;
        }
        node_push_new_child(node, rhs);
        lhs = node;
    }

    return lhs;
}

node_t* parser_try_new_eqne(parser_t* self)
{
    assert(self);
    node_t* lhs = CCM_TRY(parser_try_new_cmp);
    if (!lhs) { return NULL; }

    if (lexer_peek_kind(self->lexer, NODE_EQ, 0)
    || lexer_peek_kind(self->lexer, NODE_NE, 0))
    {
        node_t* node = lexer_try_new_next(self->lexer);

        node_t* rhs = CCM_TRY(parser_try_new_cmp);

        if (!rhs)
        {
            node_free(node); free(node);
            node_free(lhs); free(lhs);
            return NULL;
        }

        node_push_new_child(node, lhs);
        node_push_new_child(node, rhs);
        lhs = node;
    }

    return lhs;
}

node_t* parser_try_new_cmp(parser_t* self)
{
    assert(self);
    node_t* lhs = CCM_TRY(parser_try_new_term);
    if (!lhs) { return NULL; }

    if (lexer_peek_kind(self->lexer, NODE_LT, 0)
    || lexer_peek_kind(self->lexer, NODE_LE, 0)
    || lexer_peek_kind(self->lexer, NODE_GT, 0)
    || lexer_peek_kind(self->lexer, NODE_GE, 0))
    {
        node_t* node = lexer_try_new_next(self->lexer);

        node_t* rhs = CCM_TRY(parser_try_new_term);

        if (!rhs)
        {
            node_free(node); free(node);
            node_free(lhs); free(lhs);
            return NULL;
        }

        node_push_new_child(node, lhs);
        node_push_new_child(node, rhs);
        lhs = node;
    }

    return lhs;
}

node_t* parser_try_new_term(parser_t* self)
{
    assert(self);
    node_t* lhs = CCM_TRY(parser_try_new_factor);
    if (!lhs) { return NULL; }

    while (lexer_peek_kind(self->lexer, NODE_ADD, 0)
    || lexer_peek_kind(self->lexer, NODE_SUB, 0))
    {
        node_t* node = lexer_try_new_next(self->lexer);
        node_t* rhs = CCM_TRY(parser_try_new_factor);

        if (!rhs)
        {
            node_free(lhs);
            free(lhs);
            node_free(node);
            free(node);
            return NULL;
        }

        node_push_new_child(node, lhs);
        node_push_new_child(node, rhs);
        lhs = node;
    }

    return lhs;
}

node_t* parser_try_new_factor(parser_t* self)
{
    assert(self);
    node_t* lhs = CCM_TRY(parser_try_new_usub);
    if (!lhs) { return NULL; }

    while (lexer_peek_kind(self->lexer, NODE_MUL, 0)
    || lexer_peek_kind(self->lexer, NODE_DIV, 0)
    || lexer_peek_kind(self->lexer, NODE_MOD, 0))
    {
        node_t* node = lexer_try_new_next(self->lexer);

        node_t* rhs = CCM_TRY(parser_try_new_usub);

        if (!rhs)
        {
            node_free(lhs);
            free(lhs);
            node_free(node);
            free(node);
            return NULL;
        }

        node_push_new_child(node, lhs);
        node_push_new_child(node, rhs);
        lhs = node;
    }

    return lhs;
}

node_t* parser_try_new_usub(parser_t* self)
{
    assert(self);

    if (lexer_peek_kind(self->lexer, NODE_SUB, 0))
    {
        lexer_consume_next(self->lexer, NODE_SUB);
        node_t* node = malloc(sizeof(node_t));
        node_init(node, NODE_SUB, "", self->lexer->line);

        node_t* rhs = CCM_TRY(parser_try_new_usub);

        if (!rhs)
        {
            node_free(node);
            free(node);

            return NULL;
        }

        node_push_new_child(node, rhs);

        return node;
    }
    else
    {
        return CCM_TRY(parser_try_new_not);
    }

    return NULL;
}

node_t* parser_try_new_not(parser_t* self)
{
    assert(self);

    if (lexer_peek_kind(self->lexer, NODE_NOT, 0))
    {
        lexer_consume_next(self->lexer, NODE_NOT);
        node_t* node = malloc(sizeof(node_t));
        node_init(node, NODE_NOT, "", self->lexer->line);

        node_t* rhs = CCM_TRY(parser_try_new_not);

        if (!rhs)
        {
            node_free(node);
            free(node);

            return NULL;
        }

        node_push_new_child(node, rhs);

        return node;
    }
    else
    {
        return CCM_TRY(parser_try_new_pow);
    }

    return NULL;
}

node_t* parser_try_new_pow(parser_t* self)
{
    assert(self);
    node_t* lhs = CCM_TRY(parser_try_new_in);
    if (!lhs) { return NULL; }

    if (lexer_peek_kind(self->lexer, NODE_POW, 0))
    {
        if (!lexer_consume_next(self->lexer, NODE_POW))
        {
            node_free(lhs);
            free(lhs);
            return NULL;
        }

        node_t* rhs = CCM_TRY(parser_try_new_in);

        if (!rhs)
        {
            node_free(lhs);
            free(lhs);

            return NULL;
        }

        node_t* node = malloc(sizeof(node_t));
        node_init(node, NODE_POW, "", self->lexer->line);

        node_push_new_child(node, lhs);
        node_push_new_child(node, rhs);
        lhs = node;
    }

    return lhs;
}

node_t* parser_try_new_in(parser_t* self)
{
    assert(self);
    node_t* lhs = CCM_TRY(parser_try_new_literal);
    if (!lhs) { return NULL; }

    if (lexer_peek_kind(self->lexer, NODE_IN, 0))
    {
        lexer_consume_next(self->lexer, NODE_IN);
        node_t* node = malloc(sizeof(node_t));
        node_init(node, NODE_IN, "", lhs->line);

        node_t* rhs = CCM_TRY(parser_try_new_literal);

        if (!rhs)
        {
            node_free(lhs); free(lhs);
            node_free(node); free(node);

            return NULL;
        }

        node_push_new_child(node, lhs);
        node_push_new_child(node, rhs);
        lhs = node;
    }

    return lhs;
}

node_t* parser_try_new_literal(parser_t* self)
{
    assert(self);

    if (lexer_peek_kind(self->lexer, NODE_OSQUARE, 0))
    {
        node_t* array = CCM_TRY(parser_try_new_array);

        if (lexer_peek_kind(self->lexer, NODE_OSQUARE, 0))
        {
            return CCM_TRY_LL1(parser_try_new_index, array);
        }

        return array;
    }

    if (lexer_peek_kind(self->lexer, NODE_OPAR, 0))
    {
        node_t* tuple = CCM_TRY(parser_try_new_tuple);

        if (tuple)
        {
            if (lexer_peek_kind(self->lexer, NODE_OSQUARE, 0))
            {
                return CCM_TRY_LL1(parser_try_new_index, tuple);
            }

            return tuple;
        }


        if (!lexer_consume_next(self->lexer, NODE_OPAR))
        {
            return NULL;
        }

        node_t* expr = CCM_TRY(parser_try_new_expr);

        if (!lexer_consume_next(self->lexer, NODE_CPAR))
        {
            if (expr)
            {
                node_free(expr);
                free(expr);
            }

            return NULL;
        }

        return expr;
    }

    if ((lexer_peek_kind(self->lexer, NODE_STR, 0)
        || lexer_peek_kind(self->lexer, NODE_IDENT, 0))
        && lexer_peek_kind(self->lexer, NODE_OSQUARE, 1))
    {
        node_t* target = CCM_TRY(parser_try_new_builtin);
        return CCM_TRY_LL1(parser_try_new_index, target);
    }

    return CCM_TRY(parser_try_new_builtin);
}

node_t* parser_try_new_array(parser_t* self)
{
    assert(self);

    node_t* node = malloc(sizeof(node_t));
    node_init(node, NODE_ARRAY, "", self->lexer->line);

    lexer_consume_next(self->lexer, NODE_OSQUARE);

    int first = 1;

    while (!lexer_peek_kind(self->lexer, NODE_CSQUARE, 0))
    {
        if (!first)
        {
            lexer_consume_next(self->lexer, NODE_COMMA);
        }

        node_t* expr = CCM_TRY(parser_try_new_expr);

        if (!expr)
        {
            node_free(node); free(node);
            return NULL;
        }

        node_push_new_child(node, expr);

        first = 0;
    }

    lexer_consume_next(self->lexer, NODE_CSQUARE);

    return node;
}

node_t* parser_try_new_index(parser_t* self, node_t* target)
{
    assert(self);

    lexer_consume_next(self->lexer, NODE_OSQUARE);

    node_t* node = malloc(sizeof(node_t));
    node_init(node, NODE_INDEX, "", target->line);

    node_push_new_child(node, target);

    while (!lexer_peek_kind(self->lexer, NODE_CSQUARE, 0))
    {
        node_t* element = CCM_TRY(parser_try_new_expr);
        if (!element)
        {
            node_free(node); free(node);
            return NULL;
        }

        node_push_new_child(node, element);

        if (!lexer_peek_kind(self->lexer, NODE_CSQUARE, 0))
        {
            lexer_consume_next(self->lexer, NODE_COMMA);
        }
    }

    lexer_consume_next(self->lexer, NODE_CSQUARE);

    return node;
}

node_t* parser_try_new_tuple(parser_t* self)
{
    assert(self);
    node_t* node = malloc(sizeof(node_t));
    node_init(node, NODE_TUPLE, "", self->lexer->line);

    if (!lexer_consume_next(self->lexer, NODE_OPAR))
    {
        node_free(node);
        free(node);
        return NULL;
    }

    node_t* lhs = CCM_TRY(parser_try_new_expr);

    if (!lhs)
    {
        node_free(node);
        free(node);
        return NULL;
    }

    node_push_new_child(node, lhs);

    int contains_more_than_one_expr = 0;

    while (lexer_peek_kind(self->lexer, NODE_COMMA, 0))
    {
        lexer_consume_next(self->lexer, NODE_COMMA);

        node_t* child = CCM_TRY(parser_try_new_expr);

        if (!child)
        {
            node_free(node);
            free(node);
            return NULL;
        }

        node_push_new_child(node, child);
        contains_more_than_one_expr = 1;
    }

    lexer_consume_next(self->lexer, NODE_CPAR);

    if (!contains_more_than_one_expr)
    {
        node_free(node);
        free(node);
        return NULL;
    }

    return node;
}

node_t* parser_try_new_builtin(parser_t* self)
{
    assert(self);

    node_t* node = lexer_try_new_next(self->lexer);

    if (node &&
        (
            node->kind == NODE_NUM
            || node->kind == NODE_BOOL
            || node->kind == NODE_STR
            || node->kind == NODE_IDENT
        )
    )
    {
        return node;
    }

    if (node)
    {
        node_free(node);
        free(node);
    }

    return NULL;
}