Re: xsltproc floating-point exception

On 4/29/26 10:20, Martin Husemann wrote:
> On Tue, Apr 28, 2026 at 10:52:16PM +0300, Monica Vaida wrote:
>> When trying to build libxslt I get many errors during the compilation
>> of two files, `libexslt/math.c` and `libexslt/date.c`, both part of
>> the libxslt extension library, libexslt. The errors are caused by
>> `NAN` being undefined in `<math.h>` due to the VAX's nonstandard
>> floating-point.
> 
> You need to patch xsltproc to cope with architectures that do not define
> NAN. That is why the define is optional in the C standard - some machines
> just do not have it.

You could argue that VAX have NaN, but it only have signalling NaN. 
Which probably is a problem here.

   Johnny