If you invest 2m to make a search in this very topic you will find out that the Archer C5 1.2 is exactly the same HW as the C7 v2, you only have to "convert" the C5 to a C7.
The reason the package install failed is as the error shows. The kernel version in the build you flashed is 4.4.92 and it was bumped to 4.4.93 yesterday in trunk.
As the readme states, package selection was intended for my own use and largely to trim down the selection by r00t to free up space. Otherwise, the patches and kernel config used are nearly the same aside from using a different sfe patch so that sqm works on the wan interface. This disables sfe on wan but I don't mind since I and most others really only benefit from it on lan. Other minor inclusions and changes can be easily viewed from the commit history and diff of config.seed as you did. I like the idea of adding nfs server support though, so I ran a new build with it just now.
Meanwhile there is no @r00t version, I've manually updated the two packages (wpad and hostapd) and added: option wpa_disable_eapol_key_retries '1' in the /etc/config/wireless to mitigate the WPA2 bug.
Manually upgrade them via opkg install packagename.ipk
And optionally but recommended: put the new option @Klingon mentioned via text editor under all config wifi-iface 'default_radioX' instances in /etc/config/wireless. If some clients are unstable after this, you may need to remove the option line.
Tested on r00t's last stable build. shunjou's contains the latest versions of hostapd (common + utils) and wpad.
In the most recent versions of this firmware (as well as latest trunk nightlies), 5Ghz Wifi is pretty unstable for me on some client devices (multiple Android phones specifically).
I only use this router as a dumb WiFi AP connected by LAN cable to my main router (DLink DIR860L).
I'm on channel 44 if that helps. Does adjusting transmit power actually do anything, or is it recommended to leave on auto?
Last question: Is there a reason why this router is still on kernel version 4.4 while other devices are well into the 4.9 version?
This was my latest test on C7_v2, still running SNAPSHOT r4967-5508510e74 I found to be very stable. With SQM Queue Discipline set to cake and test_triple_isolated_IIt_cake_qos
Klingon, please tell me what was your reasoning behind the action to point me to that firmware?
I never saw such a buggy firmware in my life on a router!
speedtest returned error message that cannot perform test because of a verryyyy big latency test
luci web interface extremelly laggy
I never saw such a bad and instable 5 ghz wifi ... etc
I'm using the lastest stable version and in my case is working like a charm.
I have 3 wired devices: 1 Smart TV, 1 Computer, 1 Videoconsole.
Wireless Clients: 3 Phones on 2,4 Ghz., 2 Tablets on 5 Ghz., 1 SmartTV on 5 Ghz., 1 Tablet on 2.4 Ghz. and I have no problems at all.
My internet speed test over the wired computer is 300 MB. symmetric (SQM OFF), and over 5 Ghz. I get 1300 Mbps (163 MB/s).
but that firmware version which you pointed me to I think it's from the nightly builds; anyway you're lucky and I'm not the only one reporting extremelly instable 5 ghz on the latest 2 firmwares; I am back now to the august version but soon I'll flash back the stock one because nothing compares to it when it comes to the 5ghz speed and stability; I have tested already few reccomended dd-wrt builds which had also their version of fast path enabled and I was hoping to find improvements from the LEDE but unfortunatelly is not the case , one big minus using openwrt based firmwares would be (at least for me) that there is no easy option to configure a transparent client bridge on the same subnet, only gargoyle had that option but there are many more inconveniences too; so far , for a regular user, nothing beats the stock firmware ...
I (also) use a r00t version from August on my C7 V2s in production, but I find it extremely stable with both radios. Note, I have reconfigured the WAN port to a VLAN interface that ends up in an OPNsense firewall, so no worries about ppoe or sqm for me.
Regarding transparent bridging, that is achieved through the Physical Settings tab in the Interface. So you take the LAN interface, give it an IP address that is used to access the AP, then you go to Physical Settings and add the Wireless Network interface. The two will be bridged. I have never seen anything so elegant for defining bridges.
Glad i found this thread. Have installed the 2017-09-10 release last night, and i have doubled my wireless - LAN throughput over the standard lede i ran previously. Throughput is now the same as i got with the TP link firmware - 50 Mb/sec.
Thanks for the effort, and please keep up the good work.