Allow decoding Events containing BitVecs #408
Conversation
|
I wonder whether it's worth trying to simplify the decode logic; if bitvec contains its byte length, I could just ignore the order and store things entirely since I no longer need to try and re-encode it (I haven't checked yet, but should do so before committing to this). |
|
I had a look into the encode impl and reworked the handling to cover more cases and not care about ordering |
| decode_and_consume_type(type_id, &self.metadata.runtime_metadata().types, input)?; | ||
| // count how many bytes were consumed based on remaining length: | ||
| let consumed_len = all_bytes.len() - input.len(); | ||
| // move those consumed bytes to the output vec unaltered: |
There was a problem hiding this comment.
Nice, much more efficient 💪
|
|
||
| // Given a type Id and a type registry, attempt to consume the bytes | ||
| // corresponding to that type from our input. | ||
| fn decode_and_consume_type( |
There was a problem hiding this comment.
TODO: If we are happy and merge this, we could create a new "good first issue" to review the tests, perhaps adding some like the bitvec ones to test this method more specifically?
There was a problem hiding this comment.
Yes this would be a good method to unit test
|
|
||
| // We just need to consume the correct number of bytes. Roughly, we encode this | ||
| // as a Compact<u32> length, and then a slice of T of that length, where T is the | ||
| // bit store type. So, we ignore the bit order and only care that the bit store type |
There was a problem hiding this comment.
| // bit store type. So, we ignore the bit order and only care that the bit store type | |
| // bit store type. We ignore the bit order and only care that the bit store type |
| @@ -404,6 +449,19 @@ mod tests { | |||
| EventsDecoder::<DefaultConfig>::new(metadata) | |||
| } | |||
|
|
|||
| fn decode_and_consume_type_consumes_all_bytes< | |||
* Handle bitvec decoding * clippy * rework bitvec decoding; don't care about order and can handle other store primitives * cargo fmt
Also a little refactoring to make it easier to test the core decode logic, and remove the need to re-encode.
closes #405