[geos-devel] 3.4.1, stray patch from pkgsrc

[Greg Troxel]

Sandro Santilli <strk at keybit.net> writes:

> On Sat, Aug 17, 2013 at 09:09:31PM -0400, Greg Troxel wrote:
>> 
>> pkgsrc is carrying the following patch.  It strikes me that probably
>> there should be a configure-time feature test instead.
>
> Indeed configure is already checking for isnan, but the block
> you see, full of OS checks, is run when HAVE_ISNAN is undefined.
> The correct fix would indeed be with defining a possibly different
> macro for HAVE_STD_ISNAN or fix the configure test to use full
> qualification when trying.
>
> I'll love to see that ugly block in platform.h removed, but I'm afraid
> we don't have a large enough set of builders to catch a bug as soon
> as it is introduced. If you have a patch for configure that fixes the 
> *BSD case I'm happy to apply it.

http://trac.osgeo.org/geos/ticket/650

And for the larger isnan issue:

http://trac.osgeo.org/geos/ticket/651


From reading the g++ headers, it seems that it's wrong to use bare isnan
in C++.  But I could well be misinterpreting that.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 194 bytes
Desc: not available
URL: <http://lists.osgeo.org/pipermail/geos-devel/attachments/20130819/23b8d7a6/attachment.pgp>