Writeable fields should not cause rewriting of the whole entry

[koppor]

When an entry is unmarked, the field __markedentry is removed. If all other fields are shorter, the complete entry changes as the alignment changes. I am aware that we decided that at #116. We have, however, not thought about such exceptional cases.

Example:

(I know that a little bit more changed here. My database is written as BibLaTeX database).

Suggestion: Do not include non-displayable fields in the calculation of the amount of spaces to write.

Refs #574

[oscargus]

Would it make sense to store the selected entires (which I think this particular __markedEntry means) in the metaData instead?

[tobiasdiez]

The markedEntry is not the real problem. The reformatting occurs whenever a long field (eg maintitleaddon) is removed. I wouldn't change the serialization algorithm just because of this and would live with the reformatting...

[koppor]

👍 for metadata storing. Described at the paragraph starting with "A possible solution" at #574.

Proposal: First implement my solution (#851) and then work on #574.

[koppor]

Metadata storing: Difficult because of referential integrity. Therefore, we keep the information in the entry.

Proposal:

• Rename __markedentry to mark and add a migration. Decision: not worth the effort
• Rename and simply field. The new field supports marking from a single user only. Not useful, because multiple users will mark different things.

Decision: no action.

[lenhard]

Since the decision is no action, we can mark this issue as on hold and close it.

[koppor]

No freeze anymore as there are no non-writable fields anymore