As other poster hinted - to get right "8-bits Unicode Transformation Format" (commonly known as "UTF-8") output results it is not sufficient to output it right - it is also needed to output to an environment that is able to interpret UTF-8 properly. In 2019 it takes considerable efforts to find an environment that is not able to do it. Yet as usual Matthew rulez - "and whoever seeks, finds" (Mat. 7:8)

And yes, it is important do not screw up things by doing wrong file savings. Me personally, I perl with Notepad++ and with its option on "Encoding -> UTF-8" plus flag on "apply to open ANSI files". That keeps me safe for my purposes.

Because as of the year 2019 Unicode/UTF-8 is not yet another encoding to deal with. It is the encoding, the only one to deal with. Anything else - only for very special occasions when the soul is really asking for adventures. :-)

(update) With that coming back to my original question. Just try to replace rather convoluted POST calls with LWP native ->get() and ->post() methods - and we are back to the circle one with all response like �‘�˜�‘�›�˜�žТ�•�šАР� and crap. So LWP gets the right response headers (that it is UTF-8). Program says use utf8. Print says output utf8. Yet everything seems to be staying on the logic "if there is a slightest possibility it is not that damn utf-8, then don't use it; if there is not such possibility, still don't use it". This is what I said in my OP: "in order to achieve it I had to make my script like a drunk buddy I have to get from the bar back home :-) - once I lower attention, he tries to fell on the ground and sleep". It may be funny first but it gets highly irritating soon.