WIP: 64bit long double tests

[Blackhex]

Tests accompanying Windows-on-ARM-Experiments/gcc-woarm64#3. So far %q* and %Q* printf formats do not work, but I am not sure they actually should. Further investigation of this is ongoing.

The tests output:

expected: sizeof(float) == 4
actual: sizeof(float) == 4
expected: sizeof(double) == 8
actual: sizeof(double) == 8
expected: sizeof(long double) == 8
actual: sizeof(long double) == 8
expected: float %f => 1.234560
actual: float %f => 1.234560
expected: float %F => 1.234560
actual: float %F => 1.234560
expected: float %e => 1.234560e+00
actual: float %e => 1.234560e+00
expected: float %E => 1.234560E+00
actual: float %E => 1.234560E+00
expected: float %g => 1.23456
actual: float %g => 1.23456
expected: float %G => 1.23456
actual: float %G => 1.23456
expected: float %a => 0x1.3c0c2p+0
actual: float %a => 0x1.3c0c2p+0
expected: float %A => 0X1.3C0C2P+0
actual: float %A => 0X1.3C0C2P+0
expected: double %f => 1.234560
actual: double %f => 1.234560
expected: double %F => 1.234560
actual: double %F => 1.234560
expected: double %e => 1.234560e+00
actual: double %e => 1.234560e+00
expected: double %E => 1.234560E+00
actual: double %E => 1.234560E+00
expected: double %g => 1.23456
actual: double %g => 1.23456
expected: double %G => 1.23456
actual: double %G => 1.23456
expected: double %a => 0x1.3c0c1fc8f3238p+0
actual: double %a => 0x1.3c0c1fc8f3238p+0
expected: double %A => 0X1.3C0C1FC8F3238P+0
actual: double %A => 0X1.3C0C1FC8F3238P+0
expected: long double %f => 2.718282
actual: long double %f => 2.718282
expected: long double %F => 2.718282
actual: long double %F => 2.718282
expected: long double %e => 2.718282e+00
actual: long double %e => 2.718282e+00
expected: long double %E => 2.718282E+00
actual: long double %E => 2.718282E+00
expected: long double %g => 2.71828
actual: long double %g => 2.71828
expected: long double %G => 2.71828
actual: long double %G => 2.71828
expected: long double %a => 0x1.5bf0a8b145769p+1
actual: long double %a => 0x1.5bf0a8b145769p+1
expected: long double %A => 0X1.5BF0A8B145769P+1
actual: long double %A => 0X1.5BF0A8B145769P+1
expected: long double %Lf => 2.718282
actual: long double %Lf => 2.718282
expected: long double %LF => 2.718282
actual: long double %LF => 2.718282
expected: long double %Le => 2.718282e+00
actual: long double %Le => 2.718282e+00
expected: long double %LE => 2.718282E+00
actual: long double %LE => 2.718282E+00
expected: long double %Lg => 2.71828
actual: long double %Lg => 2.71828
expected: long double %LG => 2.71828
actual: long double %LG => 2.71828
expected: long double %La => 0x1.5bf0a8b145769p+1
actual: long double %La => 0x1.5bf0a8b145769p+1
expected: long double %LA => 0X1.5BF0A8B145769P+1
actual: long double %LA => 0X1.5BF0A8B145769P+1
expected: long double %lf => 2.718282
actual: long double %lf => 2.718282
expected: long double %lF => 2.718282
actual: long double %lF => 2.718282
expected: long double %le => 2.718282e+00
actual: long double %le => 2.718282e+00
expected: long double %lE => 2.718282E+00
actual: long double %lE => 2.718282E+00
expected: long double %lg => 2.71828
actual: long double %lg => 2.71828
expected: long double %lG => 2.71828
actual: long double %lG => 2.71828
expected: long double %la => 0x1.5bf0a8b145769p+1
actual: long double %la => 0x1.5bf0a8b145769p+1
expected: long double %lA => 0X1.5BF0A8B145769P+1
actual: long double %lA => 0X1.5BF0A8B145769P+1
expected: long double %llf => 2.718282
actual: long double %llf => 2.718282
expected: long double %llF => 2.718282
actual: long double %llF => 2.718282
expected: long double %lle => 2.718282e+00
actual: long double %lle => 2.718282e+00
expected: long double %llE => 2.718282E+00
actual: long double %llE => 2.718282E+00
expected: long double %llg => 2.71828
actual: long double %llg => 2.71828
expected: long double %llG => 2.71828
actual: long double %llG => 2.71828
expected: long double %lla => 0x1.5bf0a8b145769p+1
actual: long double %lla => 0x1.5bf0a8b145769p+1
expected: long double %llA => 0X1.5BF0A8B145769P+1
actual: long double %llA => 0X1.5BF0A8B145769P+1
expected: long double %Qf => 2.718282
actual: long double %Qf => %Qf
Assertion failed: strncmp (expected, actual, SIZE) == 0, file test-math.c, line 27

[ZacWalk]

This PR looks good to me just wondering about the .c thing.
I like this thorough testing as opposed to just an exe that builds or not.

[Blackhex]

This PR looks good to me just wondering about the .c thing. I like this thorough testing as opposed to just an exe that builds or not.

Thank you. I've just unified the naming convention of the executables produced by the tests. Please, let me know if I misunderstood the convention or if we should move this change to a separate PR.

[ZacWalk]

I would prefer not to have the '.c' or '.cpp' in the test output files.
Originally there were some naming conflicts but I don't think there should be any more.

[Blackhex]

Moving to Draft mode until it will be known whether %q, %Q should be supported and tested.