Hi,
One point of my bug report is that what you're trying to do
(examine a block device for the content stored on it) is not
anything new. We've been doing this for many many years
and the way to do it is
SUBSYSTEMS=="block", ACTION=="add|change", <more checks here>, <import/run etc here>
and this is not something that "needs testing" or needs to be
second-guessed or (dare I say) even tested since it's such a
fundamental operation in how udev and the kernel works. Then
there's the thing that the existing rules are obviously run -
for example it really makes no sense anywhere to run programs
accessing a device node on a "remove" event.
That said, I do appreciate that you are testing things (we all
should be testing more, really) but please realize that "success"
or "failure" rarely means anything absolute when it comes to
event driven system - often means that you just either hit or
avoided some race condition :-)
David
