Also, ignore HID++1.0 replies from HID++2.0 devices and use the correct
function ID for getDfuStatus to avoid an error.

Using the default value of '0' is what the kernel is using (related, libratbag
is using 0x08) and makes hotplug unreliable.

Apparently, '4.3' is a perfectly valid version number.

This reduces log spew for disconnected-but-paired devices.

This will be more representative of what the build actually is
rather than a mismarked tag.

If the checkout doesn't include a tag, fallback to parsing meson.build
though.

* Use sed to adjust package version while running script rather than `makepkg`
* Run script as root so it can modify `/build`
* Run `makepkg` as nobody rather than "build"
* Separate build/install steps

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=865363 has been
fixed

This means that if someone wants to use the test inteface with
packages, no hacky launching needed.

Just:

Most people won't be running the installed-tests,  but make it
easy enough for them to get at them if they want to

We used to do this dance to avoid reading the Option ROM on hardware by default
(some faulty hardware would crash...) but now we're doing the verify update in
the daemon there's no need to split this into two steps.

Fixes: https://github.com/hughsie/fwupd/issues/149

Applications always need a valid source with newer appstream-glib versions;
even in this case where the source is the hardware itself.

If we happen to poll when the device is in the *process* of closing then the
open() fails in a weird way. If we're flashing it's probably a good idea to not
be too clever anyway...