Open jayvdb opened 1 year ago
There are a lot less failures when I use https://github.com/GREsau/schemars/pull/182
I threw together a proof-of-concept for a fix that should preserve arbitrary precision when a schema is de/serialized using serde_json, and falls back to serializing as a u64/i64/f64 (potentially losing precision) when serialized to other formats. But before I put a proper fix in, I'm curious what your use-case is for enabling the arbitrary_precision
serde_json feature if you're not using JSON for serializing?
When I enable
serde_json
feature "arbitrary_precision" in my project, I get schemas that look likeIt seems like this is coming from schemars, as when I enable this feature in all my other dependencies, their tests pass, but schemars tests fail when this feature is enabled.