[scrieve]

Adds a test beside test_find in Lib/test/test_str.py that drives the
adaptive search path with non-BMP characters.

Objects/stringlib/fastsearch.h compiles adaptive_find once per character
width, and only the ASCII and UCS1 copies ran under the whole suite:

Function 'ucs4lib_adaptive_find'   Lines executed:0.00% of 39
Function 'ucs2lib_adaptive_find'   Lines executed:66.67% of 39
Function 'ucs1lib_adaptive_find'   Lines executed:71.79% of 39
Function 'asciilib_adaptive_find'  Lines executed:79.49% of 39

The adaptive branch needs a haystack of at least 2500 characters, a needle of at
least 100, and a needle at least a third of the haystack. test_adaptive_find
meets all three with 'a' and 'b', so the UCS4 instantiation stayed cold.

The new test repeats that shape with '\U0001d11e' and '\U0001d11f', checking
both a miss and a hit, through find and count. With it,
ucs4lib_adaptive_find goes from 0.00% to 89.74% of 39 lines; the branches left
are the count and reverse-search modes, which are separate entry points and are
covered elsewhere.

Test-only change, no behaviour change, so no NEWS entry.

./python -m test test_str -m "*adaptive*" passes. The new test adds about
0.85 s to test_str on a -O0 --coverage build, which is slower than a normal
one; the whole module runs in roughly 14 s there.

[bedevere-app]

Most changes to Python require a NEWS entry. Add one using the blurb_it web app or the blurb command-line tool.

If this change has little impact on Python users, wait for a maintainer to apply the skip news label instead.

[scrieve]

No NEWS entry needed here: the guide lists "test changes" as exempt in "Changes that require NEWS entries" (devguide, pull-request-lifecycle). This PR only adds a test.

[scrieve]

The ASan job fails in test_profiling, not in anything this PR touches:

FAIL: test_sampling_with_collapsed_export (...)
AssertionError: ':' not found in '<unreadable frame>'

That is the same assertion and the same line as gh-160334, which records it as an ASan-only flake: since gh-154194 the profiler degrades a frame it cannot read to <unreadable frame> with no location, so the collapsed export line carries no colon. gh-160542 is already working on the profiler checks.

This PR only adds coverage in test_str for the UCS4 adaptive search path, so it is unrelated. I cannot rerun the job from a fork, so a maintainer rerun or gh-160542 landing should clear it.