[00200200]

Related to #370 (the documentation clarification agreed in the discussion).

Document the existing encoding rules without changing matching behavior:

• string/bytes expectations compare the encoded query byte-for-byte;
• mapping expectations use decoded names and values, including + as a space;
• literal plus signs and percent sequences must be distinguished from their encoded forms.

Add executable HTTP examples to the how-to and focused tests for both mapping types and string/bytes expectations. Parameter-order guidance is limited to distinct names; repeated names retain the documented first-value/MultiDict semantics.

Validation on Python 3.13.4:

• Query tests and new examples: 32 passed, 1 expected xfail.
• Non-release test suite: 191 passed, 3 skipped, 2 expected xfails.
• All applicable pre-commit hooks pass.
• Sphinx generated HTML containing the new section and examples. Strict -W builds report the same 50 pre-existing warnings on the unchanged base and this branch, with no added warnings.

This does not implement the separate diagnostic/API changes discussed in #370.

[csernazs]

Thanks!