Bug: Delete Dataset with 2-parent-image

[wmoore]

If you have a Dataset, containing an Image that is also in another Dataset, then you can't delete the Dataset or move to another group.
Clients don't give you any useful message or suggested workaround.

Existing test for this is in /components/tools/OmeroPy/test/gatewaytest/test_chgrp.py .

@pytest.mark.xfail(reason="ticket 11610")
def testTwoDatasetsChgrpToProject(gatewaywrapper):

[mtbcarroll]

Fixed by ​http://github.com/openmicroscopy/openmicroscopy/pull/2180.

[jamoore]

Re-opening for client investigation of the message until this can be fixed further. See PR 2180 for status update.

[mtbcarroll]

The above PR has some good integration test work by Colin once this ticket gets a server-side fix.

[jburel]

work done by Dominik in insight see ​https://github.com/openmicroscopy/openmicroscopy/pull/2160
Passing it to Will.

[dlindner]

Just checked again, unfortunately PR 2160 does not cover the deletion of Datasets, just the explicit deletion of images linked to multiple Datasets.

[wmoore]

You mean deleting a Dataset that is in more that one Project?
If you meant deleting a Dataset containing images that are in more than one Dataset, this should be OK, since images will not be deleted.

[dlindner]

Currently deletion of a Dataset containing images which are linked to other Datasets, fails (without further explanation to the user). What should actually happen in this case: Deletion of a Dataset which contains this kind of images but also images which are not part of other Datasets?

[wmoore]

If you delete a Dataset, any images that are ONLY in that Dataset should be deleted. Any images that are also in another Dataset should simply remain in that Dataset. But we don't need a warning in either case, since an image wouldn't get deleted accidentally. So the warning PR at ​https://github.com/openmicroscopy/openmicroscopy/pull/2160 doesn't have anything to do with this ticket. Probably JM put the comment about that PR on the wrong ticket?

[dlindner]

Ok, sorry, was my fault, I thought this issue is related to PR 2160. I'll fix this in Insight.

[wmoore]

I think this got passed to me accidentally.

[jamoore]

I don't see us able to do something safely for this for 5.0.2 after all of Colin's and Mark's attempts at coming up with something.

Pushing to 5.0.3 and added under "Limitations" under the 5.0.2 testing card.

[hflynn]

Adding myself to the CC list so I have a reminder to remove this from the limitations page once it is fixed.

[mtbcarroll]

Now fixed by ​​https://github.com/openmicroscopy/openmicroscopy/pull/3228.

[drussell-x]

Probably worth mentioning that ​PR #3141 addresses this new functionality on the client side. E.g. if dataset1 and dataset2 both contain the same image are are both visible and that image is chgrp'd or deleted, then it will correctly be removed from both locations.