Shrared state Async ubus interface

[javierbrk]

This pull request adds shrare-state-async ubus interface. Get method returns sharedstate "get" json data or reported error in case no json is found. Errors are bubbled up as they come from sharedstate inside a json structure with the name "error".

Since there is no Shared State async binary available in the testing environment, testing possibilities are reduced. Manual testing can be done on a router with bat-hosts package using this commands:

Get bat-hosts

# ubus -S call shared-state-async get "{'data_type': 'bat-hosts'}"
{"a8:40:41:1f:73:ab":"lrsegundo_eth1_2","02:95:39:1f:73:aa":"lrsegundo_eth0_250","a8:40:41:1c:85:c3":"lrsegundo_wlan2_mesh","a8:40:41:1c:85:16":"lrsegundo_wlan1_mesh","02:bb:ed:1f:73:aa":"lrsegundo_eth1_250","a8:40:41:1f:73:a8":"lrsegundo_wlan0_ap","aa:40:41:1f:73:a8":"lrsegundo_wlan0_apname","02:db:d6:1f:73:aa":"lrsegundo_eth0_1_250","02:29:0f:1f:73:aa":"lrsegundo_eth1_2_250","c2:10:20:5e:7f:b3":"lrsegundo_bat0","a8:40:41:1f:73:aa":"lrsegundo_eth0"}

Get an invalid data type

# ubus -S call shared-state-async get "{'data_type': 'bat-hosss'}"
{"error":53248}

Sync valid data type

# ubus -S call shared-state-async sync "{'data_type': 'bat-hosts'}"
{"error":0}

Sync invalid data type

ubus -S call shared-state-async sync "{'data_type': 'bat-hoss' }"
{"error":53248}

Sync valid data type with unreachable ipv4 addresses

# ubus -S call shared-state-async sync "{'data_type': 'bat-hosts' ,'peers_ip':['10.0.0.1','10.0.0.2']}'"
{"error":32768}

Sync valid data type with invalid ipv4 address

# ubus -S call shared-state-async sync "{'data_type': 'bat-hosts' ,'peers_ip':['10.0.0.1','10.0.2']}'"
{"error":61952}

Sync invalid data type with specified ipv4 address

ubus -S call shared-state-async sync "{'data_type': 'bat-hosdts' ,'peers_ip':['127.0.0.1','127.0.0.1']}'"
{"error":53248}

Sync valid data type with reachable ipv4 addresses

# ubus -S call shared-state-async sync "{'data_type': 'bat-hosts' ,'peers_ip':['127.0.0.1','127.0.0.1']}'"
{"error":0}

[G10h4ck]

missing new line at the end of file

[G10h4ck]

See my inline comments

[G10h4ck]

why do we explicitely mention lime-app here? other components could benefit to access shared-state-async via ubus too

[javierbrk]

While logged in via ssh, you have direct, full access to ubus. This file is to enable lime app to use it. other users should enable other profiles.

[G10h4ck]

This way of handling errors doesn't seems robust to me :-/, what if a few calls are performed in parallel and some fails and some not?
shared-state-async command return the error code to the calling shell if some error happens, in that case the error should be bubbled up without more parsing

[G10h4ck]

if this prove too cumbersome to do in Lua, consider rewriting the whole ubus plugin using ucode, or ash which are supported by default by OpenWrt

[javierbrk]

commit e891527 implements a new way of handling, what do you think?

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 87.50000% with 2 lines in your changes are missing coverage. Please review.

Project coverage is 79.64%. Comparing base (63242c2) to head (5336c53).
Report is 44 commits behind head on master.

❗ Current head 5336c53 differs from pull request most recent head 5356cbd. Consider uploading reports for the commit 5356cbd to get more accurate results

Files	Patch %	Lines
...s/lime-system/files/usr/lib/lua/lime/proto/lan.lua	87.50%	2 Missing ⚠️

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1086      +/-   ##
==========================================
+ Coverage   79.55%   79.64%   +0.08%     
==========================================
  Files          53       53              
  Lines        4564     4579      +15     
==========================================
+ Hits         3631     3647      +16     
+ Misses        933      932       -1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[G10h4ck]

That the error get printed only on failure happens implicitely, please add at least a comment documentig this behaviour, so we cna understand what's happening when we look at this code in a few months or years ;)

[javierbrk]

Done ! it is a weird behavior that gives a lot of headaches. Any output that is printed in the console is used as output. If its a json everything goes fine, if it is not a json you get "invalid input" as a response.