nut-scanner seems to not really handle IPv6 addresses?

[jimklimov]

During work on PR #2509, nut-scanner console log got filled by these lines:

upscli_splitaddr: no port specified after ':' separator

My first guess is that colon-separated IPv6 addresses are not really recognized by upscli?.. (CC @aquette : I would only get to look at this in detail later in the summer)

[jimklimov]

At least, with upsc the square brackets around the address are needed:

:; ./clients/upsc -l 'xxxx:xxxx:xxxx:xxxx:da5e:d3ff:fe5c:5fd6'
Error: No such host

:; ./clients/upsc -l '[xxxx:xxxx:xxxx:xxxx:da5e:d3ff:fe5c:5fd6]'
dummy
eco650

# On another VM with an isolated IPv6 subnet island, the attempt reasonably reports:
#   Error: Connection failure: Network is unreachable

Scanning a single address in a different build than that reported from PR preparation above, no errors were logged but no discovery either:

:; LD_LIBRARY_PATH=`pwd`/clients/.libs ./tools/nut-scanner/nut-scanner -DDDDDD -O -s xxxx:xxxx:xxxx:xxxx:da5e:d3ff:fe5c:5fd6

...
   0.130177     [D1] nutscan_init: succeeded to load the library for NUT Client library
   0.130557     [D5] add_ip_range: only start address was provided, setting end to same: xxxx:xxxx:xxxx:xxxx:da5e:d3ff:fe5c:5fd6
   0.130567     [D1] Recorded IP address range #1: [xxxx:xxxx:xxxx:xxxx:da5e:d3ff:fe5c:5fd6 .. xxxx:xxxx:xxxx:xxxx:da5e:d3ff:fe5c:5fd6]
...
   0.130635     [D2] Entering run_nut_old for 1 IP address range(s)
   0.130649     [D1] nutscan_scan_nut: Scanning "Old NUT" bus for single IP address: xxxx:xxxx:xxxx:xxxx:da5e:d3ff:fe5c:5fd6
   0.130988     [D2] nutscan_scan_nut: all planned scans launched, waiting for threads to complete
   0.131007     [D2] Entering list_nut_devices for xxxx:xxxx:xxxx:xxxx:da5e:d3ff:fe5c:5fd6
   0.133573     [D2] nutscan_scan_nut: all threads freed
...

...and with the square brackets, the address is treated as invalid (so at least got a new inconsistency here):

...
   0.022613     [D1] nutscan_init: succeeded to load the library for NUT Client library
   0.022633     [D5] add_ip_range: only start address was provided, setting end to same: [xxxx:xxxx:xxxx:xxxx:da5e:d3ff:fe5c:5fd6]
   0.022638     [D1] Recorded IP address range #1: [[xxxx:xxxx:xxxx:xxxx:da5e:d3ff:fe5c:5fd6] .. [xxxx:xxxx:xxxx:xxxx:da5e:d3ff:fe5c:5fd6]]
...
   0.022705     [D2] Entering run_nut_old for 1 IP address range(s)
   0.022759     [D1] nutscan_scan_nut: Scanning "Old NUT" bus for single IP address: [xxxx:xxxx:xxxx:xxxx:da5e:d3ff:fe5c:5fd6]
   0.022754     [D1] NUT bus (old) SCAN: join back the pthread
Invalid address : [xxxx:xxxx:xxxx:xxxx:da5e:d3ff:fe5c:5fd6]
   0.151301     [D2] Finished run_nut_old loop
...

[jimklimov]

Also problematic is a query for upsc -L upsname@[IPv6:add:ress]{:port}.

UPDATE: also happens to IPv4, due to handling of ups-listing query if(varlist) { upscli_splitaddr(...) } assuming the argv[0] is hostname[:port]. Oh well, a user's shot in the foot. Normal queries for readings from a device do work, for [IPv6]{:port} also.