feat: Change timestamp field in transactions

[gfyrag]

Change timestamp field in transactions from string to time.Time

This change the type of the field 'Timestamp' inside transactions.

Both sqlite and Postgres can handle encoding of the field without additional efforts.
But, at least sqlite, if the db column is not defined as a date type, is not able to decode the value directly to a time.Time struct.
So the decoding is still manual for reading transactions (ok on logs since the column has a date type).

Both SQLite and Postgres encode time.Time as rfc3339nano.

So, to keep backward compatibility, we :

• Round the dates to the second internally (on transactions, and logs)
• Add a custom marshal function on transactions to format dates as rfc3339

We could change the precision in a later release.

Also the PR set the same date for the log, and the generated transactions.
It was not the case previously. I don't think it is a breaking change since the logs is not exposed anywhere actually.

The PR also remove verbose logging on tests because the tests give so many line and the CI refuse to display all the lines.

Type of change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Refactoring / Technical debt

What parts of the code are impacted ?

• pkg/core
• pkg/storage

Checklist:

• My code follows the style guidelines of this project
• I have performed a self-review of my code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes

[codecov]

Codecov Report

Merging #269 (0bb66f8) into main (3e0f91b) will decrease coverage by 0.36%.
The diff coverage is 100.00%.

❗ Current head 0bb66f8 differs from pull request most recent head bc4438c. Consider uploading reports for the commit bc4438c to get more accurate results

@@            Coverage Diff             @@
##             main     #269      +/-   ##
==========================================
- Coverage   71.85%   71.48%   -0.37%     
==========================================
  Files          82       79       -3     
  Lines        4302     4230      -72     
==========================================
- Hits         3091     3024      -67     
- Misses        936      939       +3     
+ Partials      275      267       -8     

Impacted Files	Coverage Δ
pkg/core/log.go	83.13% <100.00%> (+0.63%)	⬆️
pkg/core/transaction.go	75.51% <100.00%> (+5.51%)	⬆️
pkg/ledger/process.go	84.21% <100.00%> (ø)
pkg/storage/sqlstorage/log.go	78.70% <100.00%> (+1.78%)	⬆️
pkg/storage/sqlstorage/transactions.go	85.79% <100.00%> (ø)
pkg/storage/cache.go	53.84% <0.00%> (-23.08%)	⬇️
pkg/storage/sqlstorage/store.go	78.87% <0.00%> (-11.61%)	⬇️
pkg/storage/sqlstorage/flavor.go	53.19% <0.00%> (-4.26%)	⬇️
pkg/api/internal/testing.go	90.16% <0.00%> (-1.10%)	⬇️
pkg/storage/sqlstorage/migrate.go
... and 3 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 3e0f91b...bc4438c. Read the comment docs.

[jdupas22]

why is this removed ? shouldnt it have been replaced by a NullTime ?

[gfyrag]

The Scan() method is called directly with a pointer on the date :
err := rows.Scan(&l.ID, &l.Type, &l.Hash, &l.Date, &data)
No need to handle NULL values as the date is required.