feat: add SNMP support for Baytech RPC3-NC PDUs

[noahajac]

Some models of Baytech RPC3-NC PDUs do not have a value at ".1.3.6.1.4.1.4779.1.3.5.2.1.24.1", causing the sysOID match to fail. Added a 2nd MIB that looks at ".1.3.6.1.4.1.4779.1.3.5.5.1.3.2.1" instead.

This patch would fall under the "various RPC" entry in the hardware compatibility list, thus the list does not need updating.

I was unable to run make distcheck, it appears to require quite a few dependencies that are not on my system.

[jimklimov]

Cheers, and thanks for the contribution. If the mappings are all the same, you can get away defining an additional mib2nut_info_t in the same old source, see how e.g. ietf-mib.c is used for Tripplite MIB as well (matched by its own TRIPPLITE_SYSOID).

As for the distcheck, we also have a make distcheck-light which auto-detects the available dependencies instead of requiring them, and should not insist on building docs either (I think).

[noahajac]

@jimklimov

Cheers, and thanks for the contribution. If the mappings are all the same, you can get away defining an additional mib2nut_info_t in the same old source, see how e.g. ietf-mib.c is used for Tripplite MIB as well (matched by its own TRIPPLITE_SYSOID).

As for the distcheck, we also have a make distcheck-light which auto-detects the available dependencies instead of requiring them, and should not insist on building docs either (I think).

Hi there. I initially tried to implement a 2nd mib2nut_info_t as you described, however match_model_OID() was failing. The OID for device.model/ups.model is different on the device I have and the one in the original MIB. I could change device.model/ups.model to a generic OID such as .1.3.6.1.2.1.1.1.0, which I assume should be present on any Baytech device, but wasn't sure how proper that would be.

make distcheck-light gives me errors regarding libusb drivers. Installing the devel package for those then leads to missing neon lib errors. If it is strictly necessary, I can go through the deps and install them.

Thank you!

[jimklimov]

Well, I guess passing distcheck on your side is not strictly required - it would try and might fail on CI side, after a few hundred CPU hours across different platforms, which is the authoritative answer for the ultimate merge anyway.

Passing (or especially failing) it locally in one use-case just speeds up that part.