gh-158585: Optimize bytes.join() for non-empty separator

[vstinner]

• Issue: C API: Rework and enhance the PyBytesWriter implementation #158585

[vstinner]

Benchmark:

import pyperf
runner = pyperf.Runner()
for list_size in ('10', '10**3', '10**6'):
    runner.timeit(
        f'short items: {list_size}',
        setup=f"data=[b'x']*{list_size}; sep=b'.'",
        stmt="sep.join(data)")

for list_size in ('10', '10**2', '10**4'):
    runner.timeit(
        f'long items: {list_size}',
        setup=f"data=[b'x' * 100]*{list_size}; sep=b'.'",
        stmt="sep.join(data)")

Results on Fedora 44 with CPU isolation:

Benchmark	ref	change
short items: 10	215 ns	206 ns: 1.04x faster
short items: 10**3	16.7 us	15.5 us: 1.08x faster
short items: 10**6	63.0 ms	62.6 ms: 1.01x faster
long items: 10	294 ns	282 ns: 1.04x faster
long items: 10**2	1.83 us	1.75 us: 1.05x faster
Geometric mean	(ref)	1.03x faster

Benchmark hidden because not significant (1): long items: 10**4

[vstinner]

See also #158693 which optimizes PyUnicode_Join().

[eendebakpt]

Small change and a nice improvement. While investigating this I found there are larger gains possible in case of exact bytes (the common case?), but also a bug in the FT build (I created #158803).