gh-137821: Convert _json module to use Argument Clinic

[hnnynh]

• Issue: Convert _json to Argument Clinic #137821

[hnnynh]

• Benchmark Result

encode_basestring_medium: Mean +- std dev: [base] 2.11 us +- 0.05 us -> [ac] 2.33 us +- 0.04 us: 1.11x slower
encode_basestring_medium_escapes: Mean +- std dev: [base] 15.3 us +- 0.4 us -> [ac] 17.6 us +- 0.2 us: 1.15x slower
scanstring_medium: Mean +- std dev: [base] 158 ns +- 2 ns -> [ac] 110 ns +- 2 ns: 1.43x faster

Benchmark hidden because not significant (4): encode_basestring_ascii_medium, encode_basestring_ascii_medium_escapes, encode_basestring_ascii_medium_unicode, encode_basestring_medium_unicode

Geometric mean: 1.02x faster

• Script

import pyperf
import _json

# Test data - medium objects (typical API responses)
medium_string = "The quick brown fox jumps over the lazy dog. " * 20  # ~900 chars
medium_string_with_escapes = '''Line 1: "status": "success"
Line 2: Response received at 2024-01-15T10:30:00Z
\tError code: null
\tMessage: "Operation completed successfully"
''' * 50
medium_unicode = "User: 김철수 | Email: [email protected] | Status: ✓ Active | Location: 東京都 | Rating: ★★★★☆ | " * 30
medium_json_string = '"' + '{"id": 12345, "name": "John Doe", "email": "[email protected]"}' * 20 + '"'


def bench_encode_basestring_ascii(s):
    return _json.encode_basestring_ascii(s)

[hnnynh]

I've checked the memory leak test.

>  ./python.exe -m test test_json -R3:3

Using random seed: 2482685647
0:00:00 load avg: 5.30 Run 1 test sequentially in a single process
0:00:00 load avg: 5.30 [1/1] test_json
beginning 6 repetitions. Showing number of leaks (. for 0 or less, X for 10 or more)
123:456
XX. ...
0:00:30 load avg: 4.92 [1/1] test_json passed in 30.6 sec

== Tests result: SUCCESS ==

1 test OK.

Total duration: 30.6 sec
Total tests: run=203 skipped=3
Total test files: run=1/1
Result: SUCCESS

[corona10]

LGTM, parsing overhead does not exists and it becomes better!