gram.y

PostgreSQL 8.1.4的源码 适用于Linux下的开源数据库系统

}

/*
 * Read a SQL construct and build a PLpgSQL_expr for it.
 *
 * until:		token code for expected terminator
 * until2:		token code for alternate terminator (pass 0 if none)
 * expected:	text to use in complaining that terminator was not found
 * sqlstart:	text to prefix to the accumulated SQL text
 * isexpression: whether to say we're reading an "expression" or a "statement"
 * valid_sql:   whether to check the syntax of the expr (prefixed with sqlstart)
 * endtoken:	if not NULL, ending token is stored at *endtoken
 *				(this is only interesting if until2 isn't zero)
 */
static PLpgSQL_expr *
read_sql_construct(int until,
				   int until2,
				   const char *expected,
				   const char *sqlstart,
				   bool isexpression,
				   bool valid_sql,
				   int *endtoken)
{
	int					tok;
	int					lno;
	PLpgSQL_dstring		ds;
	int					parenlevel = 0;
	int					nparams = 0;
	int					params[MAX_EXPR_PARAMS];
	char				buf[32];
	PLpgSQL_expr		*expr;

	lno = plpgsql_scanner_lineno();
	plpgsql_dstring_init(&ds);
	plpgsql_dstring_append(&ds, sqlstart);

	for (;;)
	{
		tok = yylex();
		if (tok == until && parenlevel == 0)
			break;
		if (tok == until2 && parenlevel == 0)
			break;
		if (tok == '(' || tok == '[')
			parenlevel++;
		else if (tok == ')' || tok == ']')
		{
			parenlevel--;
			if (parenlevel < 0)
				ereport(ERROR,
						(errcode(ERRCODE_SYNTAX_ERROR),
						 errmsg("mismatched parentheses")));
		}
		/*
		 * End of function definition is an error, and we don't expect to
		 * hit a semicolon either (unless it's the until symbol, in which
		 * case we should have fallen out above).
		 */
		if (tok == 0 || tok == ';')
		{
			plpgsql_error_lineno = lno;
			if (parenlevel != 0)
				ereport(ERROR,
						(errcode(ERRCODE_SYNTAX_ERROR),
						 errmsg("mismatched parentheses")));
			if (isexpression)
				ereport(ERROR,
						(errcode(ERRCODE_SYNTAX_ERROR),
						 errmsg("missing \"%s\" at end of SQL expression",
								expected)));
			else
				ereport(ERROR,
						(errcode(ERRCODE_SYNTAX_ERROR),
						 errmsg("missing \"%s\" at end of SQL statement",
								expected)));
		}

		if (plpgsql_SpaceScanned)
			plpgsql_dstring_append(&ds, " ");

		switch (tok)
		{
			case T_SCALAR:
				snprintf(buf, sizeof(buf), " $%d ",
						 assign_expr_param(yylval.scalar->dno,
										   params, &nparams));
				plpgsql_dstring_append(&ds, buf);
				break;

			case T_ROW:
				snprintf(buf, sizeof(buf), " $%d ",
						 assign_expr_param(yylval.row->rowno,
										   params, &nparams));
				plpgsql_dstring_append(&ds, buf);
				break;

			case T_RECORD:
				snprintf(buf, sizeof(buf), " $%d ",
						 assign_expr_param(yylval.rec->recno,
										   params, &nparams));
				plpgsql_dstring_append(&ds, buf);
				break;

			default:
				plpgsql_dstring_append(&ds, yytext);
				break;
		}
	}

	if (endtoken)
		*endtoken = tok;

	expr = palloc(sizeof(PLpgSQL_expr) + sizeof(int) * nparams - sizeof(int));
	expr->dtype			= PLPGSQL_DTYPE_EXPR;
	expr->query			= pstrdup(plpgsql_dstring_get(&ds));
	expr->plan			= NULL;
	expr->nparams		= nparams;
	while(nparams-- > 0)
		expr->params[nparams] = params[nparams];
	plpgsql_dstring_free(&ds);

	if (valid_sql)
		check_sql_expr(expr->query);

	return expr;
}

static PLpgSQL_type *
read_datatype(int tok)
{
	int					lno;
	PLpgSQL_dstring		ds;
	PLpgSQL_type		*result;
	bool				needspace = false;
	int					parenlevel = 0;

	lno = plpgsql_scanner_lineno();

	/* Often there will be a lookahead token, but if not, get one */
	if (tok == YYEMPTY)
		tok = yylex();

	if (tok == T_DTYPE)
	{
		/* lexer found word%TYPE and did its thing already */
		return yylval.dtype;
	}

	plpgsql_dstring_init(&ds);

	while (tok != ';')
	{
		if (tok == 0)
		{
			plpgsql_error_lineno = lno;
			if (parenlevel != 0)
				ereport(ERROR,
						(errcode(ERRCODE_SYNTAX_ERROR),
						 errmsg("mismatched parentheses")));
			ereport(ERROR,
					(errcode(ERRCODE_SYNTAX_ERROR),
					 errmsg("incomplete datatype declaration")));
		}
		/* Possible followers for datatype in a declaration */
		if (tok == K_NOT || tok == K_ASSIGN || tok == K_DEFAULT)
			break;
		/* Possible followers for datatype in a cursor_arg list */
		if ((tok == ',' || tok == ')') && parenlevel == 0)
			break;
		if (tok == '(')
			parenlevel++;
		else if (tok == ')')
			parenlevel--;
		if (needspace)
			plpgsql_dstring_append(&ds, " ");
		needspace = true;
		plpgsql_dstring_append(&ds, yytext);

		tok = yylex();
	}

	plpgsql_push_back_token(tok);

	plpgsql_error_lineno = lno;	/* in case of error in parse_datatype */

	result = plpgsql_parse_datatype(plpgsql_dstring_get(&ds));

	plpgsql_dstring_free(&ds);

	return result;
}

static PLpgSQL_stmt *
make_select_stmt(void)
{
	PLpgSQL_dstring		ds;
	int					nparams = 0;
	int					params[MAX_EXPR_PARAMS];
	char				buf[32];
	PLpgSQL_expr		*expr;
	PLpgSQL_row			*row = NULL;
	PLpgSQL_rec			*rec = NULL;
	int					tok;
	bool				have_into = false;

	plpgsql_dstring_init(&ds);
	plpgsql_dstring_append(&ds, "SELECT ");

	while (1)
	{
		tok = yylex();

		if (tok == ';')
			break;
		if (tok == 0)
		{
			plpgsql_error_lineno = plpgsql_scanner_lineno();
			ereport(ERROR,
					(errcode(ERRCODE_SYNTAX_ERROR),
					 errmsg("unexpected end of function definition")));
		}
		if (tok == K_INTO)
		{
			if (have_into)
			{
				plpgsql_error_lineno = plpgsql_scanner_lineno();
				ereport(ERROR,
						(errcode(ERRCODE_SYNTAX_ERROR),
						 errmsg("INTO specified more than once")));
			}
			tok = yylex();
			switch (tok)
			{
				case T_ROW:
					row = yylval.row;
					have_into = true;
					break;

				case T_RECORD:
					rec = yylval.rec;
					have_into = true;
					break;

				case T_SCALAR:
					row = read_into_scalar_list(yytext, yylval.scalar);
					have_into = true;
					break;

				default:
					/* Treat the INTO as non-special */
					plpgsql_dstring_append(&ds, " INTO ");
					plpgsql_push_back_token(tok);
					break;
			}
			continue;
		}

		if (plpgsql_SpaceScanned)
			plpgsql_dstring_append(&ds, " ");

		switch (tok)
		{
			case T_SCALAR:
				snprintf(buf, sizeof(buf), " $%d ",
						 assign_expr_param(yylval.scalar->dno,
										   params, &nparams));
				plpgsql_dstring_append(&ds, buf);
				break;

			case T_ROW:
				snprintf(buf, sizeof(buf), " $%d ",
						 assign_expr_param(yylval.row->rowno,
										   params, &nparams));
				plpgsql_dstring_append(&ds, buf);
				break;

			case T_RECORD:
				snprintf(buf, sizeof(buf), " $%d ",
						 assign_expr_param(yylval.rec->recno,
										   params, &nparams));
				plpgsql_dstring_append(&ds, buf);
				break;

			default:
				plpgsql_dstring_append(&ds, yytext);
				break;
		}
	}

	expr = palloc(sizeof(PLpgSQL_expr) + sizeof(int) * nparams - sizeof(int));
	expr->dtype			= PLPGSQL_DTYPE_EXPR;
	expr->query			= pstrdup(plpgsql_dstring_get(&ds));
	expr->plan			= NULL;
	expr->nparams		= nparams;
	while(nparams-- > 0)
		expr->params[nparams] = params[nparams];
	plpgsql_dstring_free(&ds);

	check_sql_expr(expr->query);

	if (have_into)
	{
		PLpgSQL_stmt_select *select;

		select = palloc0(sizeof(PLpgSQL_stmt_select));
		select->cmd_type = PLPGSQL_STMT_SELECT;
		select->rec		 = rec;
		select->row		 = row;
		select->query	 = expr;

		return (PLpgSQL_stmt *)select;
	}
	else
	{
		PLpgSQL_stmt_execsql *execsql;

		execsql = palloc(sizeof(PLpgSQL_stmt_execsql));
		execsql->cmd_type = PLPGSQL_STMT_EXECSQL;
		execsql->sqlstmt  = expr;

		return (PLpgSQL_stmt *)execsql;
	}
}


static PLpgSQL_stmt *
make_fetch_stmt(void)
{
	int					tok;
	PLpgSQL_row		   *row = NULL;
	PLpgSQL_rec		   *rec = NULL;
	PLpgSQL_stmt_fetch *fetch;

	/* We have already parsed everything through the INTO keyword */

	tok = yylex();
	switch (tok)
	{
		case T_ROW:
			row = yylval.row;
			break;

		case T_RECORD:
			rec = yylval.rec;
			break;

		case T_SCALAR:
			row = read_into_scalar_list(yytext, yylval.scalar);
			break;

		default:
			yyerror("syntax error");
	}

	tok = yylex();
	if (tok != ';')
		yyerror("syntax error");

	fetch = palloc0(sizeof(PLpgSQL_stmt_select));
	fetch->cmd_type = PLPGSQL_STMT_FETCH;
	fetch->rec		 = rec;
	fetch->row		 = row;

	return (PLpgSQL_stmt *)fetch;
}


static void
check_assignable(PLpgSQL_datum *datum)
{
	switch (datum->dtype)
	{
		case PLPGSQL_DTYPE_VAR:
			if (((PLpgSQL_var *) datum)->isconst)
			{
				plpgsql_error_lineno = plpgsql_scanner_lineno();
				ereport(ERROR,
						(errcode(ERRCODE_ERROR_IN_ASSIGNMENT),
						 errmsg("\"%s\" is declared CONSTANT",
								((PLpgSQL_var *) datum)->refname)));
			}
			break;
		case PLPGSQL_DTYPE_ROW:
			/* always assignable? */
			break;
		case PLPGSQL_DTYPE_REC:
			/* always assignable?  What about NEW/OLD? */
			break;
		case PLPGSQL_DTYPE_RECFIELD:
			/* always assignable? */
			break;
		case PLPGSQL_DTYPE_ARRAYELEM:
			/* always assignable? */
			break;
		case PLPGSQL_DTYPE_TRIGARG:
			yyerror("cannot assign to tg_argv");
			break;
		default:
			elog(ERROR, "unrecognized dtype: %d", datum->dtype);
			break;
	}
}

/*
 * Given the first datum and name in the INTO list, continue to read
 * comma-separated scalar variables until we run out. Then construct
 * and return a fake "row" variable that represents the list of
 * scalars.
 */
static PLpgSQL_row *
read_into_scalar_list(const char *initial_name,
					  PLpgSQL_datum *initial_datum)
{
	int				 nfields;
	char			*fieldnames[1024];
	int				 varnos[1024];
	PLpgSQL_row		*row;
	int				 tok;

	check_assignable(initial_datum);
	fieldnames[0] = pstrdup(initial_name);
	varnos[0]	  = initial_datum->dno;
	nfields		  = 1;

	while ((tok = yylex()) == ',')
	{
		/* Check for array overflow */
		if (nfields >= 1024)
		{
			plpgsql_error_lineno = plpgsql_scanner_lineno();
			ereport(ERROR,
					(errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
					 errmsg("too many INTO variables specified")));
		}

		tok = yylex();
		switch(tok)
		{
			case T_SCALAR:
				check_assignable(yylval.scalar);
				fieldnames[nfields] = pstrdup(yytext);
				varnos[nfields++]	= yylval.scalar->dno;
				break;

			default:
				plpgsql_error_lineno = plpgsql_scanner_lineno();
				ereport(ERROR,
						(errcode(ERRCODE_SYNTAX_ERROR),
						 errmsg("\"%s\" is not a variable",
								yytext)));
		}
	}

	/*
	 * We read an extra, non-comma character from yylex(), so push it
	 * back onto the input stream
	 */
	plpgsql_push_back_token(tok);

	row = palloc(sizeof(PLpgSQL_row));
	row->dtype = PLPGSQL_DTYPE_ROW;
	row->refname = pstrdup("*internal*");
	row->lineno = plpgsql_scanner_lineno();
	row->rowtupdesc = NULL;
	row->nfields = nfields;
	row->fieldnames = palloc(sizeof(char *) * nfields);
	row->varnos = palloc(sizeof(int) * nfields);
	while (--nfields >= 0)
	{
		row->fieldnames[nfields] = fieldnames[nfields];
		row->varnos[nfields] = varnos[nfields];
	}

	plpgsql_adddatum((PLpgSQL_datum *)row);

	return row;
}

/*
 * When the PL/PgSQL parser expects to see a SQL statement, it is very
 * liberal in what it accepts; for example, we often assume an
 * unrecognized keyword is the beginning of a SQL statement. This
 * avoids the need to duplicate parts of the SQL grammar in the
 * PL/PgSQL grammar, but it means we can accept wildly malformed
 * input. To try and catch some of the more obviously invalid input,
 * we run the strings we expect to be SQL statements through the main
 * SQL parser.
 *
 * We only invoke the raw parser (not the analyzer); this doesn't do
 * any database access and does not check any semantic rules, it just
 * checks for basic syntactic correctness. We do this here, rather
 * than after parsing has finished, because a malformed SQL statement
 * may cause the PL/PgSQL parser to become confused about statement
 * borders. So it is best to bail out as early as we can.
 */
static void
check_sql_expr(const char *stmt)
{
	ErrorContextCallback  syntax_errcontext;
	ErrorContextCallback *previous_errcontext;
	MemoryContext oldCxt;

	if (!plpgsql_check_syntax)
		return;

	/*
	 * Setup error traceback support for ereport(). The previous
	 * ereport callback is installed by pl_comp.c, but we don't want
	 * that to be invoked (since it will try to transpose the syntax
	 * error to be relative to the CREATE FUNCTION), so temporarily
	 * remove it from the list of callbacks.
	 */
	Assert(error_context_stack->callback == plpgsql_compile_error_callback);

	previous_errcontext = error_context_stack;
	syntax_errcontext.callback = plpgsql_sql_error_callback;
	syntax_errcontext.arg = (char *) stmt;
	syntax_errcontext.previous = error_context_stack->previous;
	error_context_stack = &syntax_errcontext;

	oldCxt = MemoryContextSwitchTo(compile_tmp_cxt);
	(void) raw_parser(stmt);
	MemoryContextSwitchTo(oldCxt);

	/* Restore former ereport callback */
	error_context_stack = previous_errcontext;
}

static void
plpgsql_sql_error_callback(void *arg)
{
	char *sql_stmt = (char *) arg;

	Assert(plpgsql_error_funcname);

	errcontext("SQL statement in PL/PgSQL function \"%s\" near line %d",
			   plpgsql_error_funcname, plpgsql_error_lineno);
	internalerrquery(sql_stmt);
	internalerrposition(geterrposition());
	errposition(0);
}

static void
check_labels(const char *start_label, const char *end_label)
{
	if (end_label)
	{
		if (!start_label)
		{
			plpgsql_error_lineno = plpgsql_scanner_lineno();
			ereport(ERROR,
					(errcode(ERRCODE_SYNTAX_ERROR),
					 errmsg("end label \"%s\" specified for unlabelled block",
							end_label)));
		}

		if (strcmp(start_label, end_label) != 0)
		{
			plpgsql_error_lineno = plpgsql_scanner_lineno();
			ereport(ERROR,
					(errcode(ERRCODE_SYNTAX_ERROR),
					 errmsg("end label \"%s\" differs from block's label \"%s\"",
							end_label, start_label)));
		}
	}
}

#include "pl_scan.c"