[libgfortran, patch] C99 floating-point macros: isinf, isnan, ...

Tobias Burnus burnus@net-b.de
Thu Sep 9 12:48:00 GMT 2010


  On 09/09/2010 12:24 PM, FX wrote:
> The attached patch removes the configure checks and hackish libgfortran.h definitions for C99 floating-point classification macros: isinf, isnan, isfinite, isnormal and signbit. Instead, I define all of them using the GCC built-ins (__builtin_isinf, etc.).
>
> Bootstrapped and regtested on x86_64-linux, OK to commit?

OK.

  * * *

Regarding __float128 support, what is the current status?

I think with regards to http://gcc.gnu.org/ml/fortran/2010-08/msg00499.html
- Most front-end changes have been committed
- isnan problem is solved (cf. above)
- number of digits in I/O: diagnosed
- cygwin cbrtl/sqrtl: diagnosed
- SC: pending
- array support: Pending - though, I fear the descriptor update will not 
make it for 4.6. If so, will you introduce GFC_DT_FLOAT128 - or defer it 
to 4.7?

Is this more or less the __float128/quad status?

How about posting a new patch without the merged parts and with the I/O 
digits/precision fixed?

Tobias



More information about the Gcc-patches mailing list