Should I just build it or we do some more fixes?

Removing 'set -e' from firstrun would be a requirement for me (and maybe thinking about fixing the green led blinking issue on sun8i) but apart from that I would go with Zador's suggestion from today's github issue comment 'chatting'

Edited May 17, 201610 yr by wildcat_paris
nothing... just tkaiser eyes blinking behind his eyeglasses

What issue do you have with green led? You mean heartbeat trigger set by default? If that is so, you can take my patch...

What issue do you have with green led?

 

I have none since I've not booted Armbian 5.10 on H3 devices so far (no access to H3 devices currently). But according to forum posts there are issues like constantly heartbeat blinking starting with 5.10 release.

I have none since I've not booted Armbian 5.10 on H3 devices so far (no access to H3 devices currently). But according to forum posts there are issues like constantly heartbeat blinking starting with 5.10 release.

I have that too:

130			SwapState="$(grep swap /etc/fstab)"
131			if [ -f "X${SwapState}" != "X" ]; then
132				echo heartbeat >/sys/class/leds/*green*/trigger
133				[ -f "/root/.not_logged_in_yet" ] && BlinkTime=120 || BlinkTime=10
134				(sleep ${BlinkTime} && echo default-on >/sys/class/leds/*green*/trigger) &
135			fi

root@orangepione:~# systemctl status armhwinfo -l
â— armhwinfo.service - LSB: Armbian gathering informations about hardware
   Loaded: loaded (/etc/init.d/armhwinfo)
   Active: active (exited) since Tue 2016-05-17 10:41:09 MSK; 1min 24s ago
  Process: 476 ExecStart=/etc/init.d/armhwinfo start (code=exited, status=0/SUCCESS)

May 16 21:50:22 orangepione armhwinfo[476]: [ ok ] Setting noop I/O scheduler for mmcblk0
May 16 21:50:23 orangepione armhwinfo[476]: /etc/init.d/armhwinfo: line 131: [: too many arguments
May 16 21:50:23 orangepione armhwinfo[476]: [ ok ] Starting ARM hardware info: Orange Pi One
May 17 10:41:09 orangepione systemd[1]: Started LSB: Armbian gathering informations about hardware.
root@orangepione:~#

So extra "-f" on line 131 breaks whole conditional and LED doesn't turn off.

So extra "-f" on line 131 breaks whole conditional and LED doesn't turn off.

 

Ah, I introduced this bug on Apr 2 already when trying to get led behaviour right for both legacy and vanilla kernel (due to different sysfs node names). Thx for fixing it already!

 

Hmm... don't understand really why I didn't take notice (maybe fixed it locally but forgot to adopt fix in repo and send correction). Anyway IMO we can release 5.11 now.

Also I think some of GitHub issues (including old ones) should be closed with new release, i.e.

https://github.com/igorpecovnik/lib/issues/297(fixed)

https://github.com/igorpecovnik/lib/issues/211(toolchain autodetection works)

https://github.com/igorpecovnik/lib/issues/115(disk activity LED was added to some boards)

https://github.com/igorpecovnik/lib/issues/103(provided kernel parameter should work)

thanks for reintroducing SUBVERSION (it was LOCALVERSION once) so I can compile my kernel without getting the "apt upgrade" when version (package "5.11") collides (ok updated to 5.12)

=> so I don't need to modify the lib source

linux-dtb-next-sunxi_5.12-golfromeo_armhf.deb             linux-image-next-sunxi_5.12-golfromeo_armhf.deb
linux-firmware-image-next-sunxi_5.12-golfromeo_armhf.deb  linux-trusty-root-next-lamobo-r1_5.12-golfromeo_armhf.deb
linux-headers-next-sunxi_5.12-golfromeo_armhf.deb         linux-u-boot-next-lamobo-r1_5.12-golfromeo_armhf.deb

yet after dpkg -i *.deb

I have to

sudo dpkg -i frandom-dkms_1.1-0centrych4_armhf.deb but ok then reboot!

This was actually introduced for some other purpose but yes it's usable this way as well. 

 

I am slowly preparing on weekly auto build for those who wanted to be on the edge.

 

"5.11-$weeknumber"