Bug: hibernate fails on time+acquisitions

[jamoore]

See:

• ​https://github.com/openmicroscopy/bioformats/pull/1265
• ​https://github.com/openmicroscopy/openmicroscopy/pull/3099
• ​https://github.com/joshmoore/hibernate-orm/tree/3.5-PATCHED
• ​https://hibernate.atlassian.net/browse/HHH-5280 (possibly)

The combination of a number of plates with planes with timeIncrement or deltaT set caused Hibernate to fail with odd excepts. Before I patched Hibernate for HHH-5280 a PropertyAccessException was being thrown since Hibernate was Trying to use Well.getWellSamples() on an Image instance. After the patch, a PropertyValueException is being thrown since WellSample.image is null. The well sample passed in by OMERO is not null, but likely the complexity of the graph is causing some lookup to fail.

For the moment, 4 tests which match the above description have been disabled. But to prevent odd exceptions for users in the future, the deeper issue here must be resolved before release. (A Hibernate upgrade may suffice)

[jamoore]

I assume this will be the one dangling item in m2.

[jmoore <josh@…>]

(In [a9f2e17ad03817262d73c122ff14994b560a7a19/ome.git] on branch develop) Temporarily disable broken tests (See #12601)

4 tests were failing due to the addition of the Time
elements on planeInfo. This is a deeper Hibernate issue
which will require further investigation.

[jamoore]

Updating Hibernate to 3.6.10 did not solve this issue, though the work for the upgrade has been opened as a separate PR -- ​https://github.com/openmicroscopy/openmicroscopy/pull/3181

Ultimately, the solution chosen was to move to proper Java (ome.model.enums.*), Ice (omero/model/Units.ice), and SQL (psql-header.vm) enums across the board in the units PR: ​https://github.com/openmicroscopy/openmicroscopy/pull/3146