Hi,

The below mail was send to the gphoto2 mailinglist a couple of days ago, I've not checked but F-14 has the involved gvfs and libgphoto2 versions so this likely needs to be fixed in F-14 too:

###

Hello list!

As you probably know, GVFS has its gphoto2 backend which detects USB
cameras by listening to (g)udev events having ID_GPHOTO2 property set.
When a camera gets plugged in, it runs this exact code:

#v+
port = g_strdup_printf ("usb:%d,%d", bus_num, device_num);

/* Connect to the camera */
context = gp_context_new ();
if (gp_camera_new (&camera) != 0)
  goto out;
if (gp_port_info_list_new (&il) != 0)
  goto out;
if (gp_port_info_list_load (il) != 0)
  goto out;
n = gp_port_info_list_lookup_path (il, port);
if (n == GP_ERROR_UNKNOWN_PORT)
  goto out;
if (gp_port_info_list_get_info (il, n, &info) != 0)
  goto out;
if (gp_camera_set_port_info (camera, info) != 0)
  goto out;
gp_port_info_list_free (il);
il = NULL;
if (gp_camera_init (camera, context) != 0)
  goto out;
#v-

I've just spent quite some time investigating why gp_camera_init()
fails. Boy, that was fun! It turns out that since svn revision 13242 and
13245 libgphoto2 doesn't like when the specified port string differs in
format from what libgphoto2 autodetects: "usb:3,21" vs "usb:003,021".

Do you think it's a GVFS "bug" or not necessarily? My current workaround
is to patch gvfs ("usb:%03d,%03d").

BTW, I find it ugly that libgphoto2 autodetects what was already
autodetected by udev (rules generated by gphoto's own
`print-camera-list`).

Cheers.

PS. GNOME users running libgphoto2 >= 2.4.10 and gvfs <= 1.6.3 are going
    to be PISSED :)

###

Regards,

Hans