Re: isnan in fortran

[Mike Miller]

On Wed, Apr 11, 2012 at 03:46:47PM -0400, John W. Eaton wrote:
> On 11-Apr-2012, c. wrote:
> 
> | 
> | On 11 Apr 2012, at 18:43, Mike Miller wrote:
> | 
> | > On Wed, Apr 11, 2012 at 12:19 PM, c. <address@hidden> wrote:
> | >> what about the attached changeset?
> | > 
> | > Looks good to me, just read through it quickly for now.  For my part,
> | > I won't be able to test it until later, but when I do I will test
> | > against at least g77 3.4.6 and gfortran 4.1.2.  And now for some
> | > comments.
> | > 
> | > Related to another thread started by jwe today, the new file names
> | > should be *.in.f instead of *.f.in.
> | > 
> | > I would add "implicit none" to the configure check test program and
> | > define "real x", just to be extra pedantic and make sure the compiler
> | > doesn't quietly accept "isnan" as an undefined external symbol.
> | > 
> | > -- 
> | > mike
> | 
> | OK, here is an updated version of the changeset.
> 
> I applied it with a few small changes.  The most significant change
> was to use hg rename instead of hg remove+hg add.

Works perfectly for me with g77 3.4.6 and gfortran 4.1.2.  Great work.
I even coerced the build to not use the C erfc functions and made sure
this change works as expected with both NaN and non-NaN args.

The attached changeset makes it so 'make clean' deletes the files and
'make dist' doesn't include them in the tarball.  OK to push?

-- 
mike