Hey folks,
I just wondered if you were aware of / had any insights into something we've had reported in oapi-codegen:
This leads to cases where we have two different resulting JSON representations of the specification (i.e. oapi-codegen/oapi-codegen#1572 (comment)) which in our case can cause noise for folks, but may be an actual bug we want to look at here.
This only appears to have been reported with oapi-codegen v2.1.0 onwards, in which we bumped from v0.118.0 to v0.120.0.