Handle bracketed URL netloc sanitization

[0xTaoZ]

Summary

• restore URL validation for legacy feed URLs whose network location contains bracketed markers rejected by modern urllib
• keep the existing urlsplit path for normal URLs and use a small fallback only when urlsplit raises ValueError
• unskip the HTML table parser regression test for issue Regression on parsing invalid URLs #2382 and add direct URL harmonization coverage

Closes #2382.

Validation

• git diff --check
• INTELMQ_TEST_EXOTIC=1 .venv/bin/python -m pytest -q -o addopts='' intelmq/tests/bots/parsers/html_table/test_parser_column_split.py intelmq/tests/lib/test_harmonization.py intelmq/tests/lib/test_message.py
• .venv/bin/python -m compileall -q intelmq/lib/harmonization.py intelmq/tests/lib/test_harmonization.py intelmq/tests/bots/parsers/html_table/test_parser_column_split.py

Risk

Low scope: the fallback only runs when urllib.parse.urlsplit() raises ValueError, while ordinary URL handling continues through the existing parser-backed path.