Regarding chi2_quality.cc test in libstdc++-v3 testsuite for gcc

[Renlin Li <renlin dot li at arm dot com>]

Hi all,

@austern, sorry to brother you, I forgot to bring the discussion to the 
mailing list.

Has anyone noticed that, this test fails for 32bit architecture linking 
with glibc? The VERIFY statement ( VERIFY( chi2 < k*1.1 );) in 
test_uniform_random() and test_bit_flip_set() fail.

I have tested it on x86_32 and arm-none-linux-eabi, they both fail with 
the same result. However it passes for x86_64 and aarch64-none-linux-gnu.

Basically the random number generated using rand() are the same, it's 
the std::hash<> makes the difference.

If the target linked with newlib, this test case for 32-bit arm target 
passes. However, it's because  the implementation of rand() function is 
different from glibc. If I feed the same set "random" data to the test 
case(instead of calling rand() function), the results are identical, 
which means the compiler behaves correctly.

Do you know how to address this? Is the condition for the assertion too 
strict?

Any information is appreciated!

Kind regards,
Renlin Li