Re: Change password from Oracle Forms 4.0

In article <cj10.9.000AFB21_at_ucs.cam.ac.uk>, cj10_at_ucs.cam.ac.uk (C.J.Jardine) writes:
>In article <radams.5.000FB3D3_at_msgate.mdhc.mdc.com> radams@msgate.mdhc.mdc.com (Ron Adams) writes
:
>>From: radams_at_msgate.mdhc.mdc.com (Ron Adams)
>>Subject: Change password from Oracle Forms 4.0
>>Question: Is there a method in Forms to support a user changing their own
>>password without having to deploy SQL*Plus?
>
>The problems are as follows:
>
>1. ALTER USER xx IDENTIFIED BY yyy si a DDL statement. It therefore
> has implicitly commits. For this reason, you can't issue this
> statement in the session which Forms is using (as only Forms
> should commit).

Issuing your own commits in Forms, either directly or implicitly via DDL, is not a problem, as long as you know what you're doing and are aware of how it will impact any locks acquired by or changes pending in the form or any calling forms (ie the Forms transaction model).

To say that only Forms itself should issue commits is a bit OTT.

If you bear in mind the issues I mentioned, there's no reason why you shouldn't use the Forms_DDL() built-in to issue the 'ALTER USER...' statement. If you do this from a form whose sole purpose is to enable the user to change their password, and which has no base table blocks, and are careful about where you let the form be called from, I don't see a problem.

> Charles Jardine
> University of Cambridge (UK) Computing Service.

Tony Rothwell                                      rothwelt_at_heac004.gb.ec.ps.net