ref(types): make U128 a string again #18
Conversation
FranklinWaller
force-pushed
the
ref/u128
branch
from
0c1e461 to
9dabc0a
Compare
There was a problem hiding this comment.
This shouldn't be necessary. AFAIK we've been using this version for a while, and the u128's are serialzed as strings, despite being u128's.
https://github.com/sedaprotocol/seda-common-rs/blob/main/src/msgs/data_requests/types.rs#L132
It looks like the actual problem was I missed adding that in some spots in the one struct :c:
https://github.com/sedaprotocol/seda-common-rs/blob/main/src/msgs/data_requests/types.rs#L66.
and
https://github.com/sedaprotocol/seda-common-rs/blob/main/src/msgs/data_requests/types.rs#L68
All we should need to do is add that line and fix it. Or the better way might be to not alias the type so we don't need to add this line everywhere.
As far as the testing this... that might be hard cause we'd have to enable the cosmwasm feature mid way through a test. Not sure if that's doable but I can look into it.
|
@gluax I'm not sure if that version was pushed to devnet, because staking is broken complaining about the JSON number type. Checked a bit deeper and its broken. Only serialisation works but deserialisation doesn't. |
|
Discussed internally with @gluax , we have to implement deserialise conversion of a u128. Closing this PR for now |
Motivation
The overlay node broke because u128 is not supported as a JSON type. It must be a string (added a comment for future references)
Explanation of Changes
Reverted changes to go back to a string instead of a u128
Testing
cargo test works