On 14/08/14 15:11, Gioacchino Mazzurco wrote:
On Thursday 14 August 2014 12:08:50 Gui Iribarren
wrote:
in this particular case, the errors are due to
trying to get network.get_mac
from a wlan0_ap interface which doesn't exist yet on firstboot. So, this
new debug possibility is already proving veeery useful :)
The bug is fixed in hardware detection branch
Yes, i saw it, it checks for ieee80211/macaddress ; still, i think it's
just a workaround and possibly the proper thing to do is to separate the
"runlevels" like we discussed last time, and this hwd module would then
go into "postinit" time :)
P.S. have you seen last commit there is a general and more upstream solution
for wan interface
yeah, much better!! :D i'm really happy that we are finally having time
together, at the same moments, to have this kind of feedback - it's
invaluable!
i've just pushed
https://github.com/libre-mesh/lime-packages/tree/sandbox/release/14.08-rc1
since it's a release branch, i only concentrated on fixing details and
regressions
check it out!
in particular, i've left the wlanX_ap out of batman and bmx6 :( since
a) currently it'd qualify as a feature
b) it's not thoroughly tested
comments more than welcome :D
it's a sandbox branch since i might still rebase it,
but i'll compile a lime-build set of binaries right now, and flash all
the hardware i have at hand, which is fortunately varied:
* ubnt rocket m5
* tl-mr3220
* tl-mr3040
* tl-mr3020
* tl-wr741nd
* tl-wr842nd
* tl-wdr3500
* tl-wdr3600
i'll report back later
cheers!