crossref field must not contain multiple entries

[sfo]

JabRef 5.0-dev
Linux 4.4.0-159-generic amd64
Java 11.0.4

Build from latest master at 99a0a7a

---

The entry editor allows to specify a comma separated list for the crossref field. However, this is not allowed, an entry can have only one "parent" entry.

[tobiasdiez]

I think this would be best implemented as an integrity check (as we usually don't restrict input in the fields).

[sfo]

Actually, the GUI needs refinement as well, I think. The crossref field explicitly allows to set multiple values, which is not correct (see screenshot). This might be changed to allow only one entry or change to a simple text field combined with some integrity checks.

[ajsaraujo]

Can I take this one? In which classes I should work? Which integrity checks should the program perform?

[lberetti]

Hello,

We are a team of three students from France and we would be glad to try to help solving this issue.
It will be our first contribution to an OpenSource project.

[Siedlerchr]

@LucasB98 Bonjour et merci de l'intérêt que vous portez à JabRef 😸

Please follow our contribution guide for getting started and for setting up the workspace ( I would suggest you use Intellij as there are currently some difficulties with eclipse has some problems atm).
If you have all setup, the relevant classes for this behaviour are FieldEditor and LinkedEntriesEditor.
They are located under org.jabref.gui.fieldeditors
It seems that at the moment both FieldProperty.SINGLE_ENTRY_LINK and FieldProperty.MULTIPLE_ENTRY_LINK are handled in the same way.
A possible solution would be to separate the handling of multiple entry links (allowing multiple values) and single entry links (allowing only one link).

[lberetti]

We noticed that in crossref field, you can enter your own bibtexkey in crossef.
An exemple: if I create an entry in a library with John as Author and John2019 as bibtexkey, I noticed we can put John2019 as crossref. But i think that's not normal. What do you think about it ?

[tobiasdiez]

That's a good observation. I think, we shouldn't prevent people from entering the same key (maybe he wants to change the key immediately after), but there should indeed be a warning (realized using an integrity check).

[Vencesdab]

About the warning you propose, should it be made a new issue ?

[tobiasdiez]

I think it would be best to directly implement it as part of your PR, if that is possible.