Thanks for your patches.
I decided to just rip out the broken fbink and udev functionality from the script, as I want to trigger it from nickelmenu anyway. But I found it would always complain about the database being locked, unless I wrapped my script:
Code:
echo usb plug add >> /tmp/nickel-hardware-status ; my-autoshelf.sh ; echo usb plug remove >> /tmp/nickel-hardware-status
Is there someway to unlock the database without resorting to that USB plug simulation?
Quote:
Originally Posted by vdanjean
Hi,
fbink from https://www.mobileread.com/forums/sh...d.php?t=299110 is not recent enough for Kobo Sage. But recompiling it from github sources works ;-) You can see my messages at the end of the just cited thread.
I've been hit by another problem with AutoShelf. To detect user action, it monitors /dev/input/event1. But, on Kobo Sage, this is not (always?) the good file.
I wrote a patch to detect the good eventX file based on /proc/bus/input/devices.
To be fully correct, we must know the name of the touch screen. I put the one I know in the script (ie for Kobo Sage and Kobo Forma).
If nothing is found, it failback to /dev/input/event1 as before but also generate a file in /mnt/onboard/AutoShelf-bugreport.txt (with the contents of /proc/bus/input/devices) so that my patch can be improved. |