ndk-gdb passes -D to am start to make the process wait until the debugger is launched and ready, instead of possibly crashing the app before the debugger can catch it (paired with a nice "waiting for debugger" dialog and "Force close" button in Android). While we could also set the flag ourselves, this change simplifies the workflow on our end as well now that we don't have to run our own command, potentially get stuck on pidof returning multiple PIDs, and having to override no_logcat (#342).
(For the multiple-PID case though, we can optionally pass -f to make ndk-gdb kill the debugger and app first, if this ever becomes an issue.)
ndk-gdb
passes-D
toam start
to make the process wait until the debugger is launched and ready, instead of possibly crashing the app before the debugger can catch it (paired with a nice "waiting for debugger" dialog and "Force close" button in Android). While we could also set the flag ourselves, this change simplifies the workflow on our end as well now that we don't have to run our own command, potentially get stuck onpidof
returning multiple PIDs, and having to overrideno_logcat
(#342).(For the multiple-PID case though, we can optionally pass
-f
to makendk-gdb
kill the debugger and app first, if this ever becomes an issue.)