I had read Jeff's earlier request to file a JIRA bug and did so
(DERBY-357) prior to reading this email. Should I go ahead and
cancel/close this request?
Thanks,
David
Satheesh Bandaram wrote:
I am not sure if we should document this syntax. This is highly unusual
syntax and definitely not standard.
An interesting enhancement to Derby would be to convert these diagnostic
VTIs into pseudo system tables. Derby has SYSCS_UTIL and other system
schemas and it would be great to provide a pseudo system table interface
to all diagnostic VTIs as system tables in one of these system schemas.
Would be a great project to learn and enhance Derby SQL compilation and
schema management areas.
If there is any interest, I could file a Jira ENHANCEMENT request and
provide some information....
Satheesh
The way to get it to work currently I found in the tests
(org.apache.derbyTesting.functionTests.tests.lang.errorCode.java)
you do: (in ij:)
select * from new org.apache.derby.diag.ErrorMessages() c
and that dumps out all messages.
To get just one particular message, you'd do:
select * from new org.apache.derby.diag.ErrorMessages() c where
sql_state='07000';
I am wondering though, is it ok to document this?
It's very non-standard way of doing things...
At the same time, to recreate your doc with additional comments, it
must be possible to write a program that uses the ErrorMessages vti in
combination with an extra file(/extra files) & attaches the extra
hints for each errormessage...And build the error message doc up that
way. So you'd miss a manual step to keep error messages and debugging
hints in sync.
