[Libcdio-devel] New MMC routines and thoughts on eject

[R. Bernstein]

I looked at libunieject to try to get a better understanding of
issues. A large bit of libunieject_traytoggle() looked like it should
be a libcdio MMC command, so I've added that as
mmc_get_tray_status(). And since that shares a bit of code with
another mmc_get_media_changed(), a common routine was added which
returns the raw status bits from an MMC GET_EVENT_STATUS command.

In testing this out (a new example program, mmc3, was added), it looks
like in some cases I do notice that some of the routines don't work on
GNU/Linux or need to be run as root or probably some device-grouped
user. I seem to recall the media changed status routine used to work
and now it doesn't, but maybe it was just that I was using the
GNU/Linux ioctl rather than the MMC routine.

Been thinking more about the eject problem. Would it make sense to
have two kinds of eject, one that does OS magic and another that just
works at drive level? Might it not be the case that some programs want
to eject only if the OS doesn't have to get involved (which may also
require special permission), while other applications might want an
unconditional eject? I don't know. What do people think?