On Wed, Jun 20, 2012 at 02:59:47PM +0200, anonym wrote:
> First of all, awesome work!
>
> I've been fairly tied up with other things the past few days and and
> will continue to be so until the next week. After that I will take a
> real look at your work. At first glance looks very promising (my own
> investigation also ended up with sikuli as the best candidate testing
> backend), and I'd love to work on it in tandem with you as we discussed
> some months ago. Is that still on, you think?
Yeah, didn't forgot that part, and it's completely my intend. Intrigeri
told me you were planning to work on this todo item soon, and I suspect it
was one of the reason for me to vortex on this : put down what I had in
mind.
I know that you have an agenda, so on my side I can take care to have
spare time when you'll be ready. I guess we should maybe setup an IRC
meeting to discuss this issue, at least both of us who intend to go on on
this task, but also anyone who wants to be part of the design process.
> Just so no one wastes time examining this, I just want to point out that
> this is exactly what we need; I played with the dummy_hcd and
> g_file_storage (which also is needed) modules some weeks ago with this
> exact use case in mind. For instance, here's how to emulate a 2 GiB USB
> storage device:
>
> dd if=/dev/zero of=usb_drive bs=1M count=2048
> modprobe dummy_hcd
> modprobe g_file_storage file=$PWD/usb_drive
>
> Works like a charm.
That is a great news! It makes one of the top-most hard-to-implement
feature easy like a breeze it seems. I believe we'll have hard time to get
this into the official Debian kernel though, meaning we'll probably have
to maintain a custom kernel. All in all I wonder how it behaves with kvm,
and in nested environnment.
bert
