Thanks for the PR - I think this is a good improvment.

It seems to me that this PR doesn't change the wait-behavior. Instead it seem that now only stdout is used to read the password from while stderr is not catched anymore and stdin can now be used to manually input something into the command (instead of forcing the command to abort, which was the previous behavior). Is this correct?
Can you please also add a changelog entry?

now only stdout is used to read the password from while stderr is not catched anymore and stdin can now be used to manually input something into the command (instead of forcing the command to abort, which was the previous behavior)

Yes. Also, stderr, for example in keepassxc-cli, is used to show password prompt.

I think maybe it's a good idea to exit, if command exited with non-zero code, will do