Erik Wright
2004-05-28 17:09:09 UTC
Hello,
We're new to middlegen, but pretty happy with it.
We're using the head of the sourceforge tree ("2.1") on Windows and
Linux, and we're a bit confused about how the oracle sequence stuff is
supposed to work.
There are two options, as I see it: using the 'oracleSequence' subtask
and using the 'weblogic' subtask with pkgen="VENDOR_SEQ" (we're
deploying to Weblogic).
I've seen in this list that the second choice is preferred, as that
makes better use of the builtin WL CMP stuff. But I'm not quite sure how
it should work, since it only allows you to specify one sequence name
for all of your tables, and (naturally, I think) we want to use a
separate sequence for each table.
The oracleSequence subtask seems a bit smarter - it uses
<TABLE_NAME>_SEQ as the sequence name. But we had other troubles with
that. The OracleSequenceSessionBean generated is abstract(?!) and can't
be deployed. Is there a reason for that? Is there something big missing
that we need to fill in afterwards? Or is this just a bug. If it's a
bug, I'm happy to fix the template and submit a patch on Jira (we
already submitted another small patch yesterday for another bug).
For now I guess we'll patch Middlegen and use the oracleSequence task -
but if anyone has insight on how to make the weblogic VENDOR_SEQ stuff
work we'd appreciate it.
Thanks,
Erik
We're new to middlegen, but pretty happy with it.
We're using the head of the sourceforge tree ("2.1") on Windows and
Linux, and we're a bit confused about how the oracle sequence stuff is
supposed to work.
There are two options, as I see it: using the 'oracleSequence' subtask
and using the 'weblogic' subtask with pkgen="VENDOR_SEQ" (we're
deploying to Weblogic).
I've seen in this list that the second choice is preferred, as that
makes better use of the builtin WL CMP stuff. But I'm not quite sure how
it should work, since it only allows you to specify one sequence name
for all of your tables, and (naturally, I think) we want to use a
separate sequence for each table.
The oracleSequence subtask seems a bit smarter - it uses
<TABLE_NAME>_SEQ as the sequence name. But we had other troubles with
that. The OracleSequenceSessionBean generated is abstract(?!) and can't
be deployed. Is there a reason for that? Is there something big missing
that we need to fill in afterwards? Or is this just a bug. If it's a
bug, I'm happy to fix the template and submit a patch on Jira (we
already submitted another small patch yesterday for another bug).
For now I guess we'll patch Middlegen and use the oracleSequence task -
but if anyone has insight on how to make the weblogic VENDOR_SEQ stuff
work we'd appreciate it.
Thanks,
Erik