Bug in Apex 2.2.1.00.04 in overload call of function?

Question

We are having custom table defined as:
CREATE TABLE HTMLDB_USERS (
  ID             INTEGER               CONSTRAINT HTMLDB_USERS_ID_NN NOT NULL,
  USERNAME       VARCHAR2(64 BYTE)     CONSTRAINT HTMLDB_USERS_USERNAME_NN NOT NULL,
desc htmldb_users Name                                      Null?    Type
ID                                        NOT NULL NUMBER(38)
USERNAME                                  NOT NULL VARCHAR2(64)
...Where ID is primary key and on USERNAME there is unique index defined.
In package we have two overload functions defined as:
FUNCTION update_password (
            p_id       IN HTMLDB_USERS.ID%TYPE
           ,p_password IN HTMLDB_USERS.PASSWORD%TYPE
          ) RETURN PLS_INTEGER;and later declared:
FUNCTION update_password (
            p_username       IN HTMLDB_USERS.USERNAME%TYPE
           ,p_password IN HTMLDB_USERS.PASSWORD%TYPE
          ) RETURN PLS_INTEGER;which should be correctly determined in call according different data types. In our page we had a call like:
:P102_STATE :=#OWNER#.pkg_auth.update_password(:P102_ID,:P102_NEW_PASSWORD2);
...Call was going to correct function (the one with ID in parameter).
In trace (have special pragma function for writing values to database) we get that passed values all the time was "-2", regarding that "P102_ID" was having correct INTEGER value ("3", "4", "7" etc...).
Interesting thing was that call with v('P102_ID') in function results OK!
When we place call by reference
p_id=>:P102_ID,p_password=>:P102_NEW_PASSWORD2);Then all was OK!
We are pointing to this bug because it may raise many problems to users....
select * from v$version;BANNER
Oracle Database 10g Enterprise Edition Release 10.2.0.1.0 - Prod
PL/SQL Release 10.2.0.1.0 - Production
CORE    10.2.0.1.0      Production
TNS for 32-bit Windows: Version 10.2.0.1.0 - Production
NLSRTL Version 10.2.0.1.0 - ProductionHope this was worth of logging....and even this may not be treated as bug worth ow warning...

Answer 1

Vikas,
all clear but that v('P102_ID') call is bothering me!
And other thing....I was debugging in first function (with number parameter)...and if API is making VARCHAR2 value, then in no circumstances first function (with number) cannot be called but only second one (with two VARCHAR2)...which was not called !!!!
To make more clear, create two function (in this order in package):
FUNCTION update_password (
            p_id       IN HTMLDB_USERS.ID%TYPE
           ,p_password IN HTMLDB_USERS.PASSWORD%TYPE
          ) RETURN PLS_INTEGER;
BEGIN
  dbms_output.put_line(v('P102_ID'));
  dbms_output.put_line('Number called function "'||to_char(p_id)|'"');
  return null;
END;
FUNCTION update_password (
            p_username       IN HTMLDB_USERS.USERNAME%TYPE
           ,p_password IN HTMLDB_USERS.PASSWORD%TYPE
          ) RETURN PLS_INTEGER;
BEGIN
  dbms_output.put_line(v('P102_ID'));
  dbms_output.put_line('VARCHAR2 called function "'||p_username||'"');
  return null;
END;and make a call like in my case.
This is most interesting of all! Intersting which function will be called!
Maybe this note helps someone to be mislead...
Message was edited by:
Funky