You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Decommision support for 11, 12 and 18 databases. #1309
We always had our own DB builds images long before Docker was a thing that everyone had.
So we have our own images created.
For some reason, the 19c image that we have is not working well.
Maybe I need to rebuild it.
There is some strange issue with connecting to that image.
I can reproduce it locally.
Connections work well when I connect from my local Windows.
When I use Ubuntu 24 on Windows (WSL) , I get the same error as we see in pipeline.
When doing the same thing with Oracle 23 image - all works fine.
Well. I reported the problem with Oracle 23 image to Gerald - related to some non-deterministic error.
The other problem with our own 19c Image should now be resolved with this commit
Well. I reported the problem with Oracle 23 image to Gerald - related to some non-deterministic error. The other problem with our own 19c Image should now be resolved with this commit
I did some digging a little bit and turn out that on 23 error was due to our code.
On Oracle 23free, treat(a_event_item as ut_run) is triggering ORA-21779 because a_event_item is passed as IN (not IN OUT NOCOPY), and 23c is stricter about object durations when casting/treating objects passed by value.
I have put a PR there with a fix as well as the switch to fast start.
In that case we should bump the major.
While I think stop <19 builds and new features or bugs is good, removing conditional compilation might be a breaking change.
I'm up for it just I think it has to be a very clear distinction.
I did some digging a little bit and turn out that on 23 error was due to our code. On Oracle 23free, treat(a_event_item as ut_run) is triggering ORA-21779 because a_event_item is passed as IN (not IN OUT NOCOPY), and 23c is stricter about object durations when casting/treating objects passed by value.
I have put a PR there with a fix as well as the switch to fast start.
Can you share what you found (sites/pages)? I would like to learn a bit more about this "improvement"
Activity
[-]Decommision 11g support.[/-][+]Decommision support for 11, 12 and 18 databases.[/+]jgebal commentedon Mar 8, 2026
We should only keep 19 and above.
!9c goes out of premier support in 1 month. Extended support ends in 1 year.
https://www.itechtics.com/eol/oracle-database/
jgebal commentedon Mar 8, 2026
jgebal commentedon Mar 8, 2026
I'm refreshing the github repo and b ringing the project back to stability now. I see that you also started to do the same.
lwasylow commentedon Mar 8, 2026
Yes @jgebal , almost there apart from 19c.
I think looking on the images there was never 19c XE. I wonder if there is an issue with that image
jgebal commentedon Mar 8, 2026
We always had our own DB builds images long before Docker was a thing that everyone had.
So we have our own images created.
For some reason, the 19c image that we have is not working well.
Maybe I need to rebuild it.
There is some strange issue with connecting to that image.
I can reproduce it locally.
Connections work well when I connect from my local Windows.
When I use Ubuntu 24 on Windows (WSL) , I get the same error as we see in pipeline.
When doing the same thing with Oracle 23 image - all works fine.
jgebal commentedon Mar 8, 2026
Well. I reported the problem with Oracle 23 image to Gerald - related to some non-deterministic error.
The other problem with our own 19c Image should now be resolved with this commit
lwasylow commentedon Mar 8, 2026
Nice I will pull the changes to bring my branch up to date.
jgebal commentedon Mar 8, 2026
I think I have solved the problem with 19c.
At least the tests are now running 🤞
https://github.com/utPLSQL/utPLSQL/actions/runs/22830170857/job/66216998893
The problem was with how the
sqlnet.orawas configured in the container. Similar to issue that existed in 21-XE.gvenzl/oci-oracle-xe#43
lwasylow commentedon Mar 8, 2026
Nice.
For 21 and 23 we can switch also to slim faststart versions.
lwasylow commentedon Mar 8, 2026
@jgebal
I did some digging a little bit and turn out that on 23 error was due to our code.
On Oracle 23free, treat(a_event_item as ut_run) is triggering ORA-21779 because a_event_item is passed as IN (not IN OUT NOCOPY), and 23c is stricter about object durations when casting/treating objects passed by value.
I have put a PR there with a fix as well as the switch to fast start.
jgebal commentedon Mar 9, 2026
We need to review the list of supported DB versions.
We should also remove conditional compilation that references versions < 19.
That should make installation a bit more clean.
lwasylow commentedon Mar 9, 2026
In that case we should bump the major.
While I think stop <19 builds and new features or bugs is good, removing conditional compilation might be a breaking change.
I'm up for it just I think it has to be a very clear distinction.
jgebal commentedon Mar 9, 2026
Can you share what you found (sites/pages)? I would like to learn a bit more about this "improvement"