Description
Currently, we only get stack traces when running in development mode or using RUST_BACKTRACE=1 on panics, but since the idiomatic way of handling errors relies on using Result<T, E> we often miss the context on why an error has been propagated, specifically the exact line of code where the error is raised.
Acceptance criteria
Include stack traces in our Result types in both release and development profiles, without instrumenting all our code. The goal is to have traces "for free" with minimal maintenance cost, and in case an error is propagated to main we should have a full stack trace which shows the exact line where the error was raised.
Documentation link(s)
Potential libraries:
Description
Currently, we only get stack traces when running in development mode or using
RUST_BACKTRACE=1on panics, but since the idiomatic way of handling errors relies on usingResult<T, E>we often miss the context on why an error has been propagated, specifically the exact line of code where the error is raised.Acceptance criteria
Include stack traces in our
Resulttypes in both release and development profiles, without instrumenting all our code. The goal is to have traces "for free" with minimal maintenance cost, and in case an error is propagated tomainwe should have a full stack trace which shows the exact line where the error was raised.Documentation link(s)
Potential libraries: