[prev in list] [next in list] [prev in thread] [next in thread] 

List:       boost
Subject:    [boost] Re: octonion_test and quaternion_test failures on Intel
From:       Markus_Schöpflin <markus.schoepflin () comsoft ! de>
Date:       2005-05-31 12:54:29
Message-ID: d7hmkv$s48$1 () sea ! gmane ! org
[Download RAW message or body]

Hubert Holin wrote:

> A few scattered problems remain as well, COMSOFT apparently fails to 
> define the "float atan(float)" signature

I did a modification to the compiler toolset which fixes this. The
(undocumented) macro __CNAME_OVERLOADS needs to be defined to get all the 
overloads required by the standard.

> (and provides useless output for the quaternion and special functions 
> tests),

This one was due to the fact that during the regression test runs a wrong 
version of the shared C++ standard library was used. This is due to the 
fact that several compiler versions are installed side-by-side on the 
machine. I changed my local toolset to link with the static version of this 
library.

All math tests are now passing on Tru64 with this toolset.

Markus


[prev in list] [next in list] [prev in thread] [next in thread]