bshell/bshell/parse/lex/statement.c

#include "lex-internal.h"

static enum bshell_status statement_hyphen(struct lex_ctx *ctx)
{
	fx_wchar c = peek_char(ctx);
	if (!fx_wchar_is_alnum(c)) {
		push_symbol(ctx, SYM_HYPHEN);
		handle_lex_state_transition(ctx, SYM_HYPHEN);
		return BSHELL_SUCCESS;
	}

	struct lex_token *tok = NULL;
	enum bshell_status status = read_word(
		ctx,
		READ_NO_SET_TOKEN_START | READ_APPEND_HYPHEN,
		&tok);
	if (status != BSHELL_SUCCESS) {
		return status;
	}

	unsigned int token_type = TOK_WORD;

	if (convert_word_to_int(tok)) {
		token_type = TOK_INT;

		/* because of APPEND_HYPHEN (which is needed to ensure operator
		 * tokens are detected properly), the resulting number will be
		 * negative.
		 * this token will be preceded by a HYPHEN token, so the number
		 * must be positive */
		tok->tok_int *= -1;
		push_symbol(ctx, SYM_HYPHEN);
	} else if (convert_word_to_operator(ctx, tok)) {
		token_type = TOK_OPERATOR;
	}

	handle_lex_state_transition(ctx, token_type);
	enqueue_token(ctx, tok);

	return BSHELL_SUCCESS;
}

static enum bshell_status statement_symbol(struct lex_ctx *ctx)
{
	const struct lex_token_def *sym = NULL;
	enum bshell_status status = read_symbol(ctx, &sym);

	if (status != BSHELL_SUCCESS) {
		return status;
	}

	handle_lex_state_transition(ctx, sym->id);

	struct lex_token *tok = NULL;
	switch (sym->id) {
	case SYM_DQUOTE:
		return BSHELL_SUCCESS;
	case SYM_HYPHEN:
		return statement_hyphen(ctx);
	case SYM_SQUOTE:
		status = read_literal_string(ctx, &tok);
		if (status != BSHELL_SUCCESS) {
			return status;
		}
		enqueue_token(ctx, tok);
		return BSHELL_SUCCESS;

	case SYM_HASH:
		return read_line_comment(ctx);
	case SYM_DOLLAR:
		status = read_var(ctx, TOK_VAR, &tok);
		if (status != BSHELL_SUCCESS) {
			return status;
		}

		enqueue_token(ctx, tok);
		return status;
	case SYM_AT:
		status = read_var(ctx, TOK_VAR_SPLAT, &tok);
		if (status != BSHELL_SUCCESS) {
			return status;
		}

		enqueue_token(ctx, tok);
		return status;
	case SYM_DOLLAR_LEFT_BRACE:
		status = read_braced_var(ctx, TOK_VAR, &tok);
		if (status != BSHELL_SUCCESS) {
			return status;
		}

		enqueue_token(ctx, tok);
		return status;
	default:
		break;
	}

	push_symbol(ctx, sym->id);
	return BSHELL_SUCCESS;
}

static enum bshell_status statement_word(struct lex_ctx *ctx)
{
	struct lex_token *word = NULL;
	enum bshell_status status = read_word(ctx, 0, &word);
	if (status != BSHELL_SUCCESS) {
		return status;
	}

	struct lex_state *state = lex_state_get(ctx);

	bool enable_keywords = !(state->s_flags & STATEMENT_F_DISABLE_KEYWORDS);
	unsigned int token = TOK_WORD;

	if (enable_keywords && convert_word_to_keyword(word)) {
		token = word->tok_keyword;
	} else if (convert_word_to_int(word)) {
		token = TOK_INT;
	}

	handle_lex_state_transition(ctx, token);

	enqueue_token(ctx, word);
	return BSHELL_SUCCESS;
}

static enum bshell_status statement_pump_token(struct lex_ctx *ctx)
{
	fx_wchar c = peek_char(ctx);
	bool newline = false;

	set_token_start(ctx);
	while (fx_wchar_is_space(c)) {
		if (c == '\n') {
			newline = true;
		}

		set_token_end(ctx);
		advance_char_noread(ctx);
		c = peek_char_noread(ctx);
	}

	if (newline) {
		struct lex_token *tok = lex_token_create(TOK_LINEFEED);
		enqueue_token(ctx, tok);
		handle_lex_state_transition(ctx, TOK_LINEFEED);
		return BSHELL_SUCCESS;
	}

	if (char_can_begin_symbol(ctx, c)) {
		return statement_symbol(ctx);
	}

	if (char_has_flags(ctx, c, LEX_TOKEN_UNARY_ARITHMETIC)) {
		lex_state_change(ctx, LEX_STATE_ARITHMETIC);
		return BSHELL_SUCCESS;
	}

	return statement_word(ctx);
}

static const struct lex_state_link links[] = {
	LINK_PUSH(SYM_DQUOTE, LEX_STATE_STRING, 0),
	/* arithmetic tokens */
	LINK_CHANGE(TOK_KEYWORD, LEX_STATE_ARITHMETIC),
	LINK_CHANGE(TOK_INT, LEX_STATE_ARITHMETIC),
	LINK_PUSH(SYM_DOLLAR, LEX_STATE_ARITHMETIC, 0),
	LINK_PUSH(SYM_DOLLAR_LEFT_BRACE, LEX_STATE_ARITHMETIC, 0),
	LINK_CHANGE(SYM_AT_LEFT_BRACE, LEX_STATE_ARITHMETIC),
	LINK_PUSH(SYM_AT_LEFT_BRACE, LEX_STATE_HASHTABLE, 0),
	LINK_PUSH(SYM_AT, LEX_STATE_ARITHMETIC, 0),
	LINK_CHANGE(SYM_LEFT_PAREN, LEX_STATE_ARITHMETIC),
	LINK_CHANGE(SYM_BANG, LEX_STATE_ARITHMETIC),
	LINK_PUSH_WITH_TERM(
		SYM_LEFT_PAREN,
		LEX_STATE_STATEMENT,
		STATEMENT_F_DISABLE_KEYWORDS,
		SYM_RIGHT_PAREN),

	/* statement tokens */
	LINK_PUSH(SYM_LEFT_BRACE, LEX_STATE_STATEMENT, 0),
	LINK_PUSH_WITH_TERM(
		SYM_DOLLAR_LEFT_PAREN,
		LEX_STATE_STATEMENT,
		0,
		SYM_RIGHT_PAREN),

	/* command tokens */
	LINK_CHANGE(KW_FUNC, LEX_STATE_COMMAND),
	LINK_CHANGE(SYM_AMPERSAND, LEX_STATE_COMMAND),
	LINK_CHANGE(TOK_WORD, LEX_STATE_COMMAND),
	LINK_END,
};

static const unsigned int keywords[] = {
	KW_FUNC,
	KW_IF,
	KW_ELSEIF,
	KW_ELSE,
	KW_NONE,
};

static const unsigned int operators[] = {
	TKOP_BNOT,
	TKOP_NOT,
	TKOP_NONE,
};

static const unsigned int symbols[] = {
	SYM_AMPERSAND,
	SYM_BANG,
	SYM_SQUOTE,
	SYM_DQUOTE,
	SYM_HASH,
	SYM_AT,
	SYM_AT_LEFT_BRACE,
	SYM_PIPE,
	SYM_COMMA,
	SYM_SEMICOLON,
	SYM_LEFT_BRACE,
	SYM_RIGHT_BRACE,
	SYM_LEFT_BRACKET,
	SYM_RIGHT_BRACKET,
	SYM_LEFT_PAREN,
	SYM_RIGHT_PAREN,
	SYM_NONE,
};

const struct lex_state_type lex_statement_state = {
	.s_id = LEX_STATE_STATEMENT,
	.s_pump_token = statement_pump_token,
	.s_links = links,
	.s_keywords = keywords,
	.s_operators = operators,
	.s_symbols = symbols,
};