Recipe updates for NetBSD 11 support; fix nut-scanner builds --without-threading

[jimklimov]

• Newer python warnings
  • Closes: Python helper scripts: fix deprecation warning #3256
• Updated package dependency docs
• Weird parallel bmake behavior

UPDATE: Further fixes due to building recent code on Solaris 8 as the other range of the spectrum, and finding some bugs regarding nut-scanner initialization. It suffers somehow from faults inside malloc's during vupslog() called at random places when debugging is enabled, but passes OK e.g. for network scans which also involve quite a bit of memory work for the strings, I think. For posterity:

:; echo bt | gdb /home/jim/nut/tools/nut-scanner/.libs/nut-scanner `ls -lat /var/cores/nut-scanner.* | head -1 | awk '{print $NF}'`

...
Core was generated by `./tools/nut-scanner/.libs/nut-scanner -DDDDDD -h'.
Program terminated with signal SIGSEGV, Segmentation fault.
#0  0xdf916836 in realfree () from /usr/lib/libc.so.1
(gdb) #0  0xdf916836 in realfree () from /usr/lib/libc.so.1
#1  0xdf916dbf in cleanfree () from /usr/lib/libc.so.1
#2  0xdf9163aa in _malloc_unlocked () from /usr/lib/libc.so.1
#3  0xdf9162c9 in malloc () from /usr/lib/libc.so.1
#4  0xdf90c44e in calloc () from /usr/lib/libc.so.1
#5  0xdf9533de in dgettext () from /usr/lib/libc.so.1
#6  0xdf91ec1d in strerror () from /usr/lib/libc.so.1
#7  0xdfb7a5c8 in vupslog (priority=7, fmt=0x8047598 "[D1] \tVia %s:\t%s", va=0x80479b8 "\345%\270\337\b}\004\b^_\267\337", use_strerror=0) at common.c:3845
---Type <return> to continue, or q <return> to quit---
#8  0xdfb79cac in s_upsdebugx (level=1, fmt=0x8047598 "[D1] \tVia %s:\t%s") at common.c:4353
#9  0xdfb79eca in upsdebugx_report_search_paths (level=1, report_search_paths_builtin=1) at common.c:4961
#10 0xdfb75f5e in nutscan_init () at nutscan-init.c:171
#11 0x080512d1 in main (argc=5, argv=0x8047b34) at nut-scanner.c:1259

That exercise led to fixing a long-standing imperfection about fallback PRI<x>SIZE definitions NOT via %z<x> which is not a valid printf formatting character on older systems. Also a smaller initial buffer for vupslog() would be used on some systems, which should suffice for the majority of our log messages without realloc.

[AppVeyorBot]

❌ Build nut 2.8.4.4189-master failed (commit e43cf0d9db by @jimklimov)

[AppVeyorBot]

✅ Build nut 2.8.4.4190-master completed (commit 7a50d367e3 by @jimklimov)

[AppVeyorBot]

❌ Build nut 2.8.4.4195-master failed (commit 73665287b0 by @jimklimov)

[AppVeyorBot]

✅ Build nut 2.8.4.4197-master completed (commit 4bf47a15c2 by @jimklimov)