tgmath

[Javier Múgica]

Hello:

I suppose all of you in this list are also in the cfe-dev list, so you probably know about this. Given that I defined auxiliary macros an then used them all throughout the file, a diffs listing would be pointless, virtually everything has changed. Therefore, I also send the corresponding preprocessed files for both the old and the new one, just in case someone wants to compare them.

The file I send here is the same I sent a few days ago to the cfe-dev list but for the correction of a silly-obvious mistake and the splitting of some long lines, following the llvm programming style guide.

For those of you who will be reviewing my changes, here is an explanation:

1st. Guard against libraries not supporting complex types:

#if defined(__STDC_NO_COMPLEX__) || !__has_include_next(<complex.h>)
#define _TG_NO_COMPLEX
#endif

And if you search for _TG_NO_COMPLEX you wil find where it is needed. The uses are self-explanatory.


2nd. Helper macros for defining the functions. These are the definitions from

// Macros for defining the functions

till

// End of macros for defining the functions

And then come the function definitions. Where the old file had, for instance,

static float
    _TG_ATTRS
    __tg_fabs(float __x) {return fabsf(__x);}

static double
    _TG_ATTRS
    __tg_fabs(double __x) {return fabs(__x);}

static long double
    _TG_ATTRS
    __tg_fabs(long double __x) {return fabsl(__x);}

I have

__define_function_real(fabs)

And the whole file is like that, so you see how useless a diffs report would be.

Regards
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: tgmath-new.h
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20150610/69c90666/attachment.h>