UnicodeDecodeError "invalid continuation byte" when Tribler GUI reads Core stdout on Windows

[kozlovsky]

After #6567, Tribler core stdout assumes to be in UTF8 encoding. On Windows machines, the actual encoding of stdout is non-UTF8. With the introduction of auto-generated tags, tag names are logged into the core stdout and sometimes contain non-ASCII symbols. As a result, the Tribler GUI process crashed with the following error when trying to read Core output:

ErrorHandler.gui_error(): tribler.gui.utilities.CreationTraceback: 
  File "run_tribler.py", line 97, in <module>
  File "tribler\gui\start_gui.py", line 77, in run_gui
  File "tribler\gui\utilities.py", line 410, in trackback_wrapper
  File "tribler\gui\core_manager.py", line 95, in start_tribler_core


The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "tribler\gui\utilities.py", line 413, in trackback_wrapper
  File "tribler\gui\utilities.py", line 410, in trackback_wrapper
  File "tribler\gui\core_manager.py", line 110, in on_core_stdout_read_ready
UnicodeDecodeError: 'utf-8' codec can't decode byte 0xf1 in position 81: invalid continuation byte

This bug affects all Windows machines and is guaranteed to crash Tribler on Windows after a long enough period.

It may be hard to detect the correct Core stdout encoding on all systems. I think there is a simpler solution - try to read the core stdout as UTF8, and in case of the non-UTF8 output, read it as ASCII with escaping of all non-ASCII symbols. With this approach, all Core output will be preserved. There is a possibility that some artificial non-UTF8 sequences will be mistakenly read as UTF8, but it is non-critical, and for the normal output, it should not happen.

[sentry-for-tribler]

Sentry issue: TRIBLER-TESTS-22

[kozlovsky]

Fixed in #6917