search

networkupstools / nut

The Network UPS Tools repository. UPS management protocol Informational RFC 9271 published by IETF at https://www.rfc-editor.org/info/rfc9271 Please star NUT on GitHub, this helps with sponsorships!
https://networkupstools.org/
Other
1.99k stars 349 forks source link

nut-scanner network scans should provide a "-m auto" option #2244

Closed aquette closed 3 months ago

aquette commented 9 months ago

when scanning network devices (SNMP, XML, NUT), nut-scanner should provide a "-m auto" option to automatically provide the local network information.

Example: when your IP is 192.168.1.1 on a class C network, "-m auto" should be equivalent to "-m 192.168.1.1/24"

jimklimov commented 9 months ago

I think this logic could be already done in 42ity fork? port-back is still piecemeal and pending for the bigger features (for diligent separated reviewability and bug dissection).

aquette commented 9 months ago

I told myself the same 42ITy, but not sure. IIRC, it's uses libcidr and I don't want such deps.

For now, in HA SmartNUT add-on, found a clean workaround ;) (bashio returns exactly what is expected) https://github.com/aquette/addon-smartnut/blob/main/smartnut/rootfs/etc/cont-init.d/smartnut.sh#L89

jimklimov commented 8 months ago

BTW, what is wrong with libcidr? :) Not that it's hard to parse a few string patterns and do bit maths ourselves...

jimklimov commented 3 months ago

FWIW, I've recently had another pass at porting stuff from the FTY fork and checking the remaining differences (by now two big things are packaging recipes and DMF, and also docs/nut-names.txt, see #1316 for the detailed list of that effort).

There were no mentions of libcidr there, and no different code about CIDR handling, so I actually looked deeper and found nutscan-ip.c which does the related magic (including CIDR conversion to a start/end IP address pair) since your and Frederic's time. I guess we dealt with libcidr elsewhere in the 42ITy project, in some other components.

So here your issue is indeed new and pristine after all, with the challenging part being about portably detecting what non-loopback IP address(es) we have at disposal on the current system, with what netmasks, and checking perhaps all of the resulting ranges (if there's more than one).