substrate-archive
substrate-archive copied to clipboard
Archive 1.0 Tracking Issue
- [ ] Integration with Desub
- [x] Desub should seek some kind of integration with the upcoming scale-info changes
- [x] Desub needs to be updated to the latest types/metadata
- [ ] Actor for type-decoding
- [ ] storage schema for type-decoding
- [ ] rel #159 #132 #26
- [x] Desub should seek some kind of integration with the upcoming scale-info changes
- [ ] Re-Do SQL Schema migrations #326
Bug Fixes needed:
- [ ] Handle Forked Blocks Bug (#229)
- [x] Fix 'Tasks Timed Out' bug
Desub issues
- [x] https://github.com/paritytech/desub/issues/61
- [ ] https://github.com/paritytech/desub/issues/59
- [x] https://github.com/paritytech/desub/issues/57
- [x] https://github.com/paritytech/desub/issues/55
- [x] https://github.com/paritytech/desub/issues/52
- [x] https://github.com/paritytech/desub/issues/51
- [x] https://github.com/paritytech/desub/issues/49
- [x] https://github.com/paritytech/desub/issues/43
- [x] https://github.com/paritytech/desub/issues/50
Nice to Have:
- [ ] More Unit tests
- [ ] Unit tests for each actor
- [ ] Unit tests for each query
- [ ] Integration tests for archive in general
- [ ] Integration tests for Polkadot/Kusama
What's the status of
Actor for type-decoding storage schema for type-decoding
Since https://github.com/paritytech/substrate-archive/pull/394 was merged.
@xcaptain
There's an actor to decode extrinsics and it works, with some caveats:
- there may be discrepencies between the formats of newer, scale-info-produced JSON and
legacy-desubjproduced JSON (in terms of extrinsics) - once archive reaches the point at which it begins to decode extrinsics using
desub-current, i.e.scale-infothe memory tends to balloon. This could/is for a few reasons and there could be optimizations made in archive, but imo mainly because of the presence of doc-strings within the metadata (which take up a most of the extrinsic in scale-info). If we are decoding 1000-10,000 extrinsics at once, this causes significant memory issues
Other than this, indexing is fast enough and it works
storage-decoding /actor still in the works