SqlSelectBuilder from块toString时on后面的部分显示引用地址且表别名丢失

------sql------
SELECT SUM(e.cost_total_new) AS sumCost_total_new, COUNT(DISTINCT USER.userid) AS custCount
FROM cf_cust e join app_dbm USER
on e.userid=USER.userid WHERE USER.day >= 1 AND USER.day <= 30 AND USER.city
IN ('北京市', '杭州市', '上海市', '广州市', '深圳市') AND e.gmt_create >= '2019-04-04' AND e.gmt_create <= '2019-04-04'

---------
已知以上sql,调用以下方法创建sqlBuilder
SQLSelectBuilder builder = SQLBuilderFactory.createSelectSQLBuilder(sql, JdbcConstants.MYSQL);
SQLSelectStatement statement = builder.getSQLSelectStatement();
SQLSelect sqlSelect = statement.getSelect();
MySqlSelectQueryBlock sqlQueryBlock = (MySqlSelectQueryBlock)sqlSelect.getQuery();
String from = sqlQueryBlock.getFrom().toString();

得到的from语句中，on之后的部分为一个引用地址，且表别名丢失，toString后的 from语句如下
cf_cust JOIN app_dbm
ON com.alibaba.druid.sql.ast.expr.SQLBinaryOpExpr@1a4c3d21

期望结果是：
cf_cust e join app_dbm USER
on e.userid=USER.userid

Reproduce the issue with the supplied SQL using SQLBuilderFactory.createSelectSQLBuilder(sql, JdbcConstants.MYSQL), then inspect MySqlSelectQueryBlock.getFrom().toString(). Check how the JOIN source, aliases, and ON expression are rendered. Done when the output preserves cf_cust e, app_dbm USER, and the e.userid=USER.userid condition instead of an object reference.

Written by the indexing model from the issue text.