Re: New 8.1.6 bug of the day

If the previous posts are completely true (and why wouldn't they be) then yes I now think there's a real bug there all right and apologies for rushing to judgement before.

Spoke to Oracle Support in Colorado Springs a short while ago.

Until Oracle can re-create the bug, and until they have it reported to them by the actual person who is suffering because of the bug, as per their official policy, they will not commit resources to looking at it solely because it was reported to them by me.

Of course I don't think the support analyst I was talking to took the issue as seriously as I and others are.

This bug is hardly in the same class as the typical ""-- TAR 523522.34 --- 'Designer 2000 doesn't display all the referential integrity constraints on a schema if you have 45 tables and 23 sequences and a) you're using NT 4.0 SP 9.2, b) there is a full moon, c) the tide is out, d) it's snowing outside and e) you were born in the Chinese Year of the Dragon...... etc. etc. . -----'"".

I think its pretty "production-core-critial"

(or any other collection of serious sounding words all hyphenated together ;-) )

...The key thing here is that someone actually able to reproduce the bug has to get a fully documented representation of their installation. This means I suppose, the server version, patch number, OS, OS version, OS patches (if any), full copy of init****.ora , full copy of the install log for the server, EXPLAIN_PLANS (or SQL_Traces or both) of the statement as it is running a) wrongly and b) correctly after some tuning, alter.log and anything else vaguely related to the instance.

Then if they can reasonably claim that they are suffering a true-life loss or disadvantage due to this problem, then they can ask Oracle to try and recreate the environment and reproduce the bug.

Hope someone unfortunate to have met this bug can make the effort....

  D. P. O.C.P.(DBA)