Open danielnorberg opened 3 months ago
Thank you for this library!
We use the rust tracing crate and ran into an issue where the panic hook itself panics on invoking log::error! when tracing is in a broken state.
log::error!
First a tracing macro would panic, causing the panic hook to get invoked and then the panic hook panicked, yielding:
thread panicked while processing panic. aborting.
Could it make sense to fall back to e.g. eprintln! if log::error! panics?
eprintln!
How would that be implemented?
Not sure, is it possible to catch a panic in the panic hook or does that second panic cause an immediate abort?
It is an immediate abort.
Thank you for this library!
We use the rust tracing crate and ran into an issue where the panic hook itself panics on invoking
log::error!
when tracing is in a broken state.First a tracing macro would panic, causing the panic hook to get invoked and then the panic hook panicked, yielding:
Could it make sense to fall back to e.g.
eprintln!
iflog::error!
panics?