Refine BackupManager

[koppor]

Alleviates https://discourse.jabref.org/t/no-content-in-table-after-blue-screen/3479

• We call backup files .bak and temporary writing files now .sav.
• JabRef keeps 10 older versions of a .bib file in the user data dir (instead of a single .sav (now: .bak) file in the directory of the .bib file)
• In case a backup is found, the filename of the backup is shown.

---

• Change in CHANGELOG.md described in a way that is understandable for the average user (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 developer's documentation: Is the information available and up to date? If not, I outlined it in this pull request.
• Checked documentation: Is the information available and up to date? If not, I created an issue at https://github.com/JabRef/user-documentation/issues or, even better, I submitted a pull request to the documentation repository. --> https://docs.jabref.org/advanced/autosave

[koppor]

Follow-up: #2961 (comment)

[calixtus]

looks good to me.
Is the backupmanager running in a seperate thread?

[calixtus]

Is #8750 superceded by this PR?

[koppor]

Is #8750 superceded by this PR?

No, it is not. This PR is an extract of #8750 to keep things going and to enable easier reviewing.

[koppor]

Ideas (if they are quick to implement)

• Introduce needsBackup flat
• Listen to database changes. If a change occurred, set needsBackup flag to true.
• In performBackup: If needsBackup is true, really perform backup. If not, no action

In this way, we only create backups if necessary.

I am not sure, however, if we the current change infrastructure trigger changes in all changes (e.g., change of encoding, change of groups, add a BibTeX string). At the event bus, I was pretty sure. Does anyone know about the new infrastructure?

[Siedlerchr]

We still have the EventBus in place... But I would stick with creating backups always

[koppor]

Converted to draft, since I have to investigate the creation of BackupManger in the two cases i) restore at JabRef start and ii) file opening. Before the PR, the BackupManager was started and afterwards (!) it was checked if the backup file differed from the current file. Here, that lead to 0 byte backup files.

[koppor]

Refs #8298 (comment): Partially fixes the issue regarding openDatabses(), because (at least) the already opened database use the same code. The behavior for databases passed by the command line is not changed, because this is too difficult for now (the user could trigger and import, opening a bib file, etc. One would need to distinguish that).