Fix the expected DDL output value

[hyoklee]

The test source code uses -99.0:

Particle1 fill_data[1] = {{"no data", -1, -1, -99.0F, -99.0}};

I386 Alpine CI test needs this patch to pass all tests.

This change may fail some other 64-bit systems.
If so, I suspect a bug in the HDF5 library.

[byrnHDF]

This is not the correct approach - why did this not fail until testing i386. First investigate the why of the failure.
And if i386 testing is correct and the 64-bit tests are correct - then add a special case.

[byrnHDF]

Looking at the example code the fill date is defined:
Particle1 fill_data[1] = {{"no data", -1, -1, -99.0F, -99.0}};

So why is the test not always failing?

[byrnHDF]

Closing this in favor of: HL TB examples are incorrect #5332