fix: read config files that start with a UTF-8 byte order mark - #2388

A config file saved as "UTF-8 with BOM" (Notepad's default on older Windows, and an option in VS Code and other editors) can't be read at all:

$ coverage run m.py          # .coveragerc, setup.cfg or tox.ini starts with a BOM
File contains no section headers.
file: '.coveragerc', line: 1
'[run]\n'

$ coverage run m.py          # pyproject.toml starts with a BOM
Couldn't read config file pyproject.toml: Invalid statement (at line 1, column 1)

Both readers open the file with encoding="utf-8", which keeps the BOM as a  character at the start of the text, and neither configparser nor tomllib skips it. This switches both to utf-8-sig, which drops a leading BOM and otherwise reads the same as utf-8.

test_config_file_with_bom covers .coveragerc, setup.cfg, tox.ini and pyproject.toml; all four fail before the change and pass after it. tests/test_config.py passes. I also ran test_api.py, test_cmdline.py and test_process.py: the only failures are two test_pth_and_source_work_together cases that fail the same way on master here.