Closed firelizzard18 closed 1 month ago
It looks like this https://github.com/go-delve/delve/blob/master/pkg/proc/gdbserial/rr.go#L144 line is where the arguments for rr record
are constructed.
It looks like this https://github.com/go-delve/delve/blob/master/pkg/proc/gdbserial/rr.go#L144 line is where the arguments for rr record are constructed.
That's replay, record is above in RecordAsync.
🤦 of course
@aarzilli Would you accept a PR for this? Do you have a preferred flag name or other mechanism for setting the RR flag?
Do you have a preferred flag name or other mechanism for setting the RR flag?
That's the problem. The simple way to do this is to add a flag to debug
, test
and exec
but that sucks, they already have too many flags and adding another one that only works when --backend
is rr
is ugly. The implementation is trivial.
I would prefer to have backend specific --backend-opts
parameter to avoid confusion and, probably futher using for other backend.
For example, I can't use rr without --bind-to-cpu=0
option on my CPU and OS combination.
Thanks in advance.
@firelizzard18 @dmvolod I have made a PR about this using environment variables: #3726
@aarzilli That works for me, I should be able to set those via VSCode so that takes care of my use case.
Please provide a way to pass the
--chaos
flag to therr
backend. Either by a dedicated flag such as--rr-chaos
or as a general--rr-flags
,--backend-flags
,--backend-opts
, etc.Chaos mode is an extremely useful tool for reproducing weird bugs.