Reduce memory footprint

[tobiasdiez]

by removing some caching of ids and keys in databases.

Realized to late that the numbers are actually quite small. So this doesn't give much of a boost, but anyway it also doesn't hurt. Moreover, in my opinion, the code is more readable now.

• Change in CHANGELOG.md described (if applicable)
• Tests created for changes (if applicable)
• Manually tested changed features in running JabRef (always required)
• Screenshots added in PR description (for UI changes)
• Checked documentation: Is the information available and up to date? If not created an issue at https://github.com/JabRef/user-documentation/issues or, even better, submitted a pull request to the documentation repository.

[Siedlerchr]

You seem to have committed on top of an old branch. I had a déjà vu seeing some changes

[tobiasdiez]

Yep, now it should be fine. Sorry

[koppor]

I like intelligent models. No need for "artificial" separation in model and logic. However, we wanted to keep the model package very slim ("dumb"?).

I'd like to hear @stefan-kolb's take on this.

[Siedlerchr]

Could you please check the performance of import? The duplication checker is heavily used on import

[tobiasdiez]

I didn't experience any performance problems with it. Adding a new entry is now slightly slower because we have to traverse a ArrayList instead of looking up a key in a hashmap. But I guess the difference should be negligible to average size databases of <= 10k entries.

@koppor not sure what you mean with model vs logic. All my changes concern only the model package.

[koppor]

@koppor not sure what you mean with model vs logic. All my changes concern only the model package.

Here a screenshot of the diff. I marked the package in green. I read model. OK, I have to confess that the functionality did not change. Furthermore, I learned that our thinking is called AnemicDomainModel, which is thought as being wrong by @martinfowler. So my comment (#6298 (review)) is not applicable any more. Still want to discuss that wiht @stefan-kolb. Maybe at #JabCon?