EDIT

This is fixed, also for pp synonyms.

I also adapted the changeSynonymToAcceptedTaxon, the newly created taxon has the same publish state as the synonym, is this the wanted behaviour?

And the clone method of TaxonBase also clones the publish flag now.

Also swapping synonym with accepted has strange behavior: only the names swap but the publish flag stays. E.g. if old accepted is published the new accepted is also published and if the old synonym is unpublished the new synonym is also unpublished. This I think is unwanted.
In the rare case that the both status differ, we should keep the flag with the according taxon but warn that the flags differ.

Accordingly we should check ALL operations to handle the publish flag correctly.

Andreas Müller wrote:

The message pops up 2x now

With the actual nightly build I can not reproduce this.

Andreas Müller wrote:

The message pops up 2x now

I can not reproduce this anymore.

Andreas Müller wrote:

Accordingly we should check ALL operations to handle the publish flag correctly.

E.g. also move synonym to another accepted taxon should do the same check as synonymToAccepted. If the old accepted taxon has a different publish flag then the new accepted taxon a warning should appear.

By the way, currently this is handled as an info (i) but I think a warning would be better.

Changing an unpublished accepted to a synonym of an published accepted taxon, how should we handle this? Keep the publish flag from the old accepted taxon and warn?

Katja Luther wrote:

Changing an unpublished accepted to a synonym of an published accepted taxon, how should we handle this? Keep the publish flag from the old accepted taxon and warn?

Yes, I think a warning is needed here. In general we should always warn if not all related taxa (taxon/syn itself, old parent/accepted, new parent/accepted) do have the same pub. status.

Katja, can you please test correct functionality for each of the above operations and leave a message in this ticket what you have tested.

E.g.

swap: parent 0, taxon 0, syn 1 => message appears (OK)
swap: parent 1, taxon 0, syn 1 => no message (FALSE)
swap: parent 1, taxon 1, syn1 => no message (OK)
...

For operations like swap I do not see the need to test for the parent if the old and the new accepted have the same publish status, because the status of the child is still the same.

Katja Luther wrote:

For operations like swap I do not see the need to test for the parent if the old and the new accepted have the same publish status, because the status of the child is still the same.

OK, you are right. In this special case it is not really needed to check all 3 related taxa.

Tested Syn => MAN/ppSyn

acc 1, syn 0 => MAN 0 (OK)
acc 0, syn 1 => MAN 1 (OK)
acc 0, syn 0 => MAN 0 (OK)

Tested MAN/ppSyn => Syn

acc 1, MAN 0 => Syn 0 (OK)
acc 1, MAN 1 => Syn 0 (FALSE)
acc 0, MAN 0 => Syn 0 (OK)
acc 0, MAN 1 => Syn 0 (FALSE)

The new synonym always looses its state when publish flag was checked. But it should keep the state (without warning in this case).

Tested change accepted to synonym

Old parent 1, old acc 1, new acc 1 => new syn 0, w/o warning (FALSE, should be 1 w/o warning)
Old parent 1, old acc 1, new acc 0 => new syn 0, with warning (OK)
Old parent 1, old acc 0, new acc 1 => new syn 1, w/o warning (FALSE, should be 1 with warning)
Old parent 1, old acc 0, new acc 0 => new syn 0, with warning (FALS, should be 0 w/o warning)

Andreas Müller wrote:

Tested change accepted to synonym

Old parent 1, old acc 1, new acc 1 => new syn 0, w/o warning (FALSE, should be 1 w/o warning)
Old parent 1, old acc 1, new acc 0 => new syn 0, with warning (OK)
Old parent 1, old acc 0, new acc 1 => new syn 1, w/o warning (FALSE, should be 1 with warning)
Old parent 1, old acc 0, new acc 0 => new syn 0, with warning (FALS, should be 0 w/o warning)

The evaluation assumes that the state of the old parent does not matter as discussed via email.

change synonym to accepted taxon

Old acc 1, old syn 1, new parent 1 => new acc 1 w/o warning (OK)
Old acc 1, old syn 0, new parent 1 => new acc 0 with warning (OK)
Old acc 1, old syn 1, new parent 0 => new acc 1 with warning (OK)
Old acc 1, old syn 0, new parent 0 => new acc 0 w/o warning (OK)
Old acc 0, old syn 1, new parent 1 => new acc 1 w/o warning (OK), old acc should have no influence

This operation looks ok.

Andreas Müller wrote:

Tested MAN/ppSyn => Syn

acc 1, MAN 0 => Syn 0 (OK)
acc 1, MAN 1 => Syn 0 (FALSE)
acc 0, MAN 0 => Syn 0 (OK)
acc 0, MAN 1 => Syn 0 (FALSE)

The new synonym always looses its state when publish flag was checked. But it should keep the state (without warning in this case).

this is fixed.

Tested move synonym/homotypic group to another taxon

Old acc 1, old syn 1, new acc 1 => new syn 1 w/o warning (OK)
Old acc 1, old syn 0, new acc 1 => new syn 0 with warning (OK)
Old acc 1, old syn1 1, old syn2 0, new acc 1 => new syn1 1, new syn2 0 w/o warning (FALSE, the homotypic synonym syn2 should also be checked) - if any of the synonyms in the HG differs from state of the new acc taxon the warning should come
Old acc 1, old syn 1, new acc 0 => new syn 1 with warning (OK)
Old acc 1, old syn 0, new acc 0 => new syn 0 w/o warning (OK)

Generally correct, but not yet implemented for full homotypic group

Katja, can you do the same check for "move accepted to new parent". I don't if this is already fully fixed, committed and build in nightly

Andreas Müller wrote:

Andreas Müller wrote:

Tested change accepted to synonym

Old parent 1, old acc 1, new acc 1 => new syn 0, w/o warning (FALSE, should be 1 w/o warning)
Old parent 1, old acc 1, new acc 0 => new syn 0, with warning (OK)
Old parent 1, old acc 0, new acc 1 => new syn 1, w/o warning (FALSE, should be 1 with warning)
Old parent 1, old acc 0, new acc 0 => new syn 0, with warning (FALS, should be 0 w/o warning)

The evaluation assumes that the state of the old parent does not matter as discussed via email.

I think the warnings should come when the new accepted taxon and the synonym have different publish flags and I checked it again and it looks like this:

old acc 1, new acc 1 => new syn 1
old acc 1, new acc 0 => new syn 1 with warning
old acc 0, new acc 1 => new syn 0 with warning
old acc 0, new acc 0 => new syn 0

Andreas Müller wrote:

Katja, can you do the same check for "move accepted to new parent". I don't if this is already fully fixed, committed and build in nightly

moved acc 0, new parent 1 => warning
moved acc 0, new parent 0 => no warning
moved acc 1, new parent 1 => warning -> false

Katja Luther wrote:

Andreas Müller wrote:

Katja, can you do the same check for "move accepted to new parent". I don't if this is already fully fixed, committed and build in nightly

moved acc 0, new parent 1 => warning
moved acc 0, new parent 0 => no warning
moved acc 1, new parent 1 => warning -> false

this is fixed, now:

moved acc 0, new parent 1 => warning
moved acc 0, new parent 0 => no warning
moved acc 1, new parent 1 => no warning
moved acc 1, new parent 0 => warning

Katja Luther wrote:

Andreas Müller wrote:

Tested MAN/ppSyn => Syn

acc 1, MAN 0 => Syn 0 (OK)
acc 1, MAN 1 => Syn 0 (FALSE)
acc 0, MAN 0 => Syn 0 (OK)
acc 0, MAN 1 => Syn 0 (FALSE)

The new synonym always looses its state when publish flag was checked. But it should keep the state (without warning in this case).

this is fixed.

for me it does not work yet

Katja Luther wrote:

Andreas Müller wrote:

Andreas Müller wrote:

Tested change accepted to synonym

Old parent 1, old acc 1, new acc 1 => new syn 0, w/o warning (FALSE, should be 1 w/o warning)
Old parent 1, old acc 1, new acc 0 => new syn 0, with warning (OK)
Old parent 1, old acc 0, new acc 1 => new syn 1, w/o warning (FALSE, should be 1 with warning)
Old parent 1, old acc 0, new acc 0 => new syn 0, with warning (FALS, should be 0 w/o warning)

The evaluation assumes that the state of the old parent does not matter as discussed via email.

I think the warnings should come when the new accepted taxon and the synonym have different publish flags and I checked it again and it looks like this:

old acc 1, new acc 1 => new syn 1
old acc 1, new acc 0 => new syn 1 with warning
old acc 0, new acc 1 => new syn 0 with warning
old acc 0, new acc 0 => new syn 0

I agree about the handling of the warning and I tested again and everything works, except for 1x for the second case (no warning appeared) but I could not reproduce, so I guess in general it works. However, we should watch it a bit.

P.S: Looks like I mixed case 2 and 3 somehow in my original message

Andreas Müller wrote:

Katja Luther wrote:

Andreas Müller wrote:

Tested MAN/ppSyn => Syn

acc 1, MAN 0 => Syn 0 (OK)
acc 1, MAN 1 => Syn 0 (FALSE)
acc 0, MAN 0 => Syn 0 (OK)
acc 0, MAN 1 => Syn 0 (FALSE)

The new synonym always looses its state when publish flag was checked. But it should keep the state (without warning in this case).

this is fixed.

for me it does not work yet

maybe I missed to commit, I try again, when jenkins is ready.

Katja Luther wrote:

Andreas Müller wrote:

Katja Luther wrote:

Andreas Müller wrote:

Tested MAN/ppSyn => Syn

acc 1, MAN 0 => Syn 0 (OK)
acc 1, MAN 1 => Syn 0 (FALSE)
acc 0, MAN 0 => Syn 0 (OK)
acc 0, MAN 1 => Syn 0 (FALSE)

The new synonym always looses its state when publish flag was checked. But it should keep the state (without warning in this case).

this is fixed.

for me it does not work yet

maybe I missed to commit, I try again, when jenkins is ready.

tested again with nightly on test server:

acc 1, MAN 0 => Syn 1 (FALSE)
acc 1, MAN 1 => Syn 1 (OK)
acc 0, MAN 0 => Syn 0 (OK)
acc 0, MAN 1 => Syn 0 (FALSE)

but with local server with same TaxonServiceImpl it is:

acc 1, MAN 0 => Syn 0 (OK)
acc 1, MAN 1 => Syn 1 (OK)
acc 0, MAN 0 => Syn 0 (OK)
acc 0, MAN 1 => Syn 1 (OK)

Now also MAN -> Synonym works as expected:

acc 1, MAN 0 => Syn 0 (OK)
acc 1, MAN 1 => Syn 1 (OK)
acc 0, MAN 0 => Syn 0 (OK)
acc 0, MAN 1 => Syn 1 (OK)