Fix invalid CSS selector test

[schoekek]

Fix invalid CSS selector test

The test_css_selector_invalid_syntax test currently fails because the
test was changed to call root.css("[invalid") without handling the
exception raised by the CSS selector parser.

This change was introduced in commit 54c4818 ("Make ruff linter
happy"), which removed the broad except Exception block to satisfy
Ruff's BLE001 rule. However, root.css("[invalid") is expected to
raise SelectolaxError for invalid CSS syntax.

Use pytest.raises(SelectolaxError) instead of catching Exception.
This keeps the test compatible with Ruff while explicitly testing the
expected exception.

The CSS selector implementation itself is unchanged; this PR only
corrects the test.

[rushter]

[invalid] is now a valid selector for the most recent lexbor, let's use something like [invalid&] so it fails for both versions.

[schoekek]

Unfortunately, a simple root.css("[invalid&") doesn't work like that—or what do you think?

=================================== FAILURES ===================================
_______________________ test_css_selector_invalid_syntax _______________________

    def test_css_selector_invalid_syntax():
        html = "<div><p>Test</p></div>"
        parser = LexborHTMLParser(html)
        root = parser.root
        assert root is not None

>       root.css("[invalid&")

tests/test_lexbor.py:771:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
selectolax/lexbor/node.pxi:374: in selectolax.lexbor.LexborNode.css
    return self.parser.selector.find(query, self._get_node())
selectolax/lexbor/selection.pxi:70: in selectolax.lexbor.LexborCSSSelector.find
    cpdef list find(self, str query, LexborNode node):
selectolax/lexbor/selection.pxi:71: in selectolax.lexbor.LexborCSSSelector.find
    return self._find(query, node, 0)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

>   raise SelectolaxError("Can't parse CSS selector.")
E   selectolax.lexbor.SelectolaxError: Can't parse CSS selector.

selectolax/lexbor/selection.pxi:88: SelectolaxError
=========================== short test summary info ============================
FAILED tests/test_lexbor.py::test_css_selector_invalid_syntax - selectolax.le...

[rushter]

Unfortunately, a simple root.css("[invalid&") doesn't work like that—or what do you think?

=================================== FAILURES ===================================
_______________________ test_css_selector_invalid_syntax _______________________

    def test_css_selector_invalid_syntax():
        html = "<div><p>Test</p></div>"
        parser = LexborHTMLParser(html)
        root = parser.root
        assert root is not None

>       root.css("[invalid&")

tests/test_lexbor.py:771:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
selectolax/lexbor/node.pxi:374: in selectolax.lexbor.LexborNode.css
    return self.parser.selector.find(query, self._get_node())
selectolax/lexbor/selection.pxi:70: in selectolax.lexbor.LexborCSSSelector.find
    cpdef list find(self, str query, LexborNode node):
selectolax/lexbor/selection.pxi:71: in selectolax.lexbor.LexborCSSSelector.find
    return self._find(query, node, 0)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

>   raise SelectolaxError("Can't parse CSS selector.")
E   selectolax.lexbor.SelectolaxError: Can't parse CSS selector.

selectolax/lexbor/selection.pxi:88: SelectolaxError
=========================== short test summary info ============================
FAILED tests/test_lexbor.py::test_css_selector_invalid_syntax - selectolax.le...

We just needed to update selector, but keep the exception handling. I've applied the fix. Thanks.