Cannot delete last Bitstream using new "Edit" menu

[tdonohue]

Describe the bug

This issue was found during testing of #5145, the PR which adds this new "Edit" menu (via submission form).

If you use the new Edit (via form) menu to modify an existing Item, and then attempt to delete all the Item's bitstreams, you will find that you cannot delete the final Bitstream (i.e. the Item cannot be left with no bitstreams). While the reason for this behavior is that every Item requires a Bitstream (in the default form), the Edit Form will not display any useful information or errors. This makes the behavior confusing to users.

Workaround: To work around this problem, just add another Bitstream to the Item before deleting the final Bitstream. Bitstream deletions work fine as long as there are multiple Bitstreams.

To Reproduce

Steps to reproduce the behavior:

1. Modify any archived Item or Entity using the new "Edit" menu. Find an Item that has multiple Bitstreams if possible.
2. On the Edit Form, delete all Bitstreams from the Item/Entity except the final Bitstream. These deletions will all be successful.
3. Now, attempt to delete the final Bitstream. This deletion request will never complete (the delete button will show an in-progress icon) because it fails silently behind the scenes because of validation errors (as the Item requires at least on Bitstream exist).

Expected behavior

Any one of these behaviors would be acceptable:

1. Either display a useful error to the user that says: "All Items require at least one File. To delete this File, please first add another File to the Item."
2. Or, disable the Bitstream "Delete" button if there is only one Bitstream. Provide a hover-over which says the same as 1.
3. Or, allow the Bitstream deletion to succeed, but disable the "Save" buttons until the form is valid again (i.e. another Bitstream has been added)

Related work

This "Edit" menu (via submission form) was introduced in #5145.

[guillermo-escire]

Hi @tdonohue, I'm a partner at @jtimal, could you please assign me this issue to work on?

[guillermo-escire]

Hi @tdonohue, just following up on my previous comment—I'd be happy to take this issue on. Could you please assign it to me when you get a chance? Thanks!

[tdonohue]

@guillermo-escire : This bug describes an issue in a brand new feature which was just merged today, which is why I didn't respond yesterday. However, the new feature is now on the main branch and the bug can now be reproduced there.

I'll assign it to you now. Thanks for volunteering. If you run into issues finding a solution, please let us know. Because this is a bug in a feature coming from DSpace-CRIS, we'd like to get this resolved in 10.0 if possible.