We are currently using serialization to json file (using serde) and writing to disk using the atomicwrites crate.
Cons of the current design:
- Very inefficient, as the whole database should be written for every mutation.
- Possible atomicity issues. sqlite3 is probably more battle tested than the
atomicwrites crate.
Required steps:
- Designing an SQL schema.
- Refactoring Funder's mutations to contain less logic.
- Translating Funder's mutations to SQL statements.
We are currently using serialization to json file (using
serde) and writing to disk using theatomicwritescrate.Cons of the current design:
atomicwritescrate.Required steps: