[Fwd: [PATCH] PR21915, Add acosh, asinh, and atanh to intrinsics]

[Jerry DeLisle]

Roger Sayle wrote:
> Hi Jerry,
> 
> On Fri, 24 Jun 2005, Jerry DeLisle wrote:
> 
>>2005-06-24  Jerry DeLisle <jvdelisle@verizon.net>
>>
>>	PR libfortran/21915
>>	* Makefile.am: include intrinsics/hyper.c
>>	* c99_protos.h: add prototypes for single precision versions of
>>	acosh, asinh, and atanh for platforms that do not have these
>>	* config.h.in: add #undef for wrappers
>>	* configure.ac: add checks for single precision versions
>>	* aclocal.m4: generated
>>	* Makefile.in: generated
>>	* configure: generated
>>	* intrinsics/hyper.c: New file, adds wrapper functions
> 
> 
> I appreciate that this has already been committed to mainline following
> Steve Kargl's approval, but I have two comments/suggestions for a
> follow-up patch (even though I'm not a gfortran maintainer).
> 
> The first is that in your ChangeLog entries (such as the libgfortran
> portion above) descriptions should be full sentences starting with a
> capital letter and ending in a period/full-stop.
> 
Thankyou for this comment.  This was my first patch of this type.  I 
will clean that up and resubmit.  It really helps to get some feedback 
to make inprovements.

> Secondly, it's a strange decision to place these three replacement
> functions (acoshf, asinhf and atanhf) in their own new source file,
> intrinsics/hyper.c, and then update Makefile.am and Makefile.in to
> build it.  Instead it would have made more sense to add these to
> the existing instrinsics/c99_functions.c which already serves exactly
> this purpose and contains near identical functions for many of the
> C99-required libm transcendental and trigonometric functions.  Indeed,
> the prototypes for these three float variants were added by your patch
> to the existing c99_protos.h file.
> 

Thankyou Steve for explainung that.

Jerry