tests/pthread_barrier: fix test script

[maribu]

Contribution description

Previously the test script relied on the exact sequence of numbers
returned by the used PRNG. This resulting e.g. in

$ USEMODULE=prng_musl_lcg make -C tests/pthread_barrier flash test

to fail, only because the order in which the children completed is
slightly different due to different sleep durations. This fixes the
issue.

Testing procedure

The command above should succeed, no with and without USEMODULE=prng_musl_lcg before the invocation of make.

Issues/PRs references

Useful for #17188

[maribu]

I disabled compile tests (as this doesn't change C code). I think this will also skip builds and test runs for native, so reviewers should make sure that make test not only works on my machine ;-)

[maribu]

The race condition hit and Murdock is already almost through a full build. So at least on native the test is run by the CI.

[maribu]

Thanks for the review

[chrysn]

I think the use of fstrings here broke CI hardware tests, possibly because the pi workers still use an old version of Python 3:

-- running on worker pi-ef3229ad thread 1, build number 10729.
-- executing tests for tests/pthread_barrier on samr21-xpro (compiled with gnu toolchain):
- executing hook run_test_pre
- resetting USB power...
- waiting for tty to re-attach...
[...]
make: Entering directory '/tmp/dwq.0.7924475090058434/6e6f7168dc09d6596896571abdc7edba/tests/pthread_barrier'
  File "/tmp/dwq.0.7924475090058434/6e6f7168dc09d6596896571abdc7edba/tests/pthread_barrier/tests/01-run.py", line 12
    child.expect(f'Start {i + 1}')
                                ^
SyntaxError: invalid syntax
/tmp/dwq.0.7924475090058434/6e6f7168dc09d6596896571abdc7edba/makefiles/tests/tests.inc.mk:22: recipe for target 'test' failed
make: *** [test] Error 1
make: Leaving directory '/tmp/dwq.0.7924475090058434/6e6f7168dc09d6596896571abdc7edba/tests/pthread_barrier'

(From https://ci.riot-os.org/RIOT-OS/RIOT/17256/c360b13e6096968cb675e0c00d2e9afe803ff1d5/output/run_test/tests/pthread_barrier/samr21-xpro:gnu.txt)

[maribu]

Thanks for reporting. I'll wait a bit before PR'ing a fix to see if the Pis can be updated instead

[chrysn]

For reference, fix is in #17283