flang win: use long double instead of __float128 on Windows platform

win: use long double instead of __float128 on Windows platform

Open tdusnoki opened this issue 3 years ago • 1 comments

128 bit float is not supported on Windows platform and it causes build error.

Aug 23 '22 14:08 tdusnoki

long double is not 128 bits in size on windows? Does it make sense to do typedef long double float128_t?

By default long double is 8 byte on Windows. It seems it needs to use '-mlong-double-128' compiler option to force long double to be 128 bits.

Aug 24 '22 11:08 kaadam

@bryanpkc @kiranchandramohan gentle ping?

Oct 07 '22 08:10 kaadam

@isuruf Do you approve this PR now? Was your comment sufficient addressed?

Oct 18 '22 17:10 bryanpkc