[boost] Re: boost/math octonion/quaternion failures?

From: Hubert Holin (Hubert.Holin_at_[hidden])
Date: 2003-08-14 04:10:21

Somewhere in the E.U., le 14/08/2003

   Bonjour

In article <4.3.2.7.2.20030813111042.02734b18_at_[hidden]>,
 Beman Dawes <bdawes_at_[hidden]> wrote:

> At 09:57 AM 8/13/2003, Hubert Holin wrote:
> 
>  >Somewhere in the E.U., le 13/08/2003

[SNIP]

>  >(in other words,
>  >Gcc 2.95.x might be out).
>  >
>  >      At any rate, I only have CodeWarrior locally available for tests
>  >(but I will read the results of the various platform runs to fix
>  >things), so it might prove inconvenient for the "imminent" release.
> 
> If there are new failures, is it OK if I try to fix them? I've got 8 
> compilers, will have a pretty good idea if changes break anything.
> 
> --Beman

      No problems at all!

      I have checked in yesterday a modified version (added guarded 
using directives at namespace level and ifdef them out of functions; did 
that a bit fast and perhaps not entirely in the best possible way) of 
quaternion_test.cpp and of octonion_test.cpp, where I believe lies the 
problem, but did *not* check in the modified special functions files, as 
the modification on these later files where not needed for the 
quaternions and octonions and might break things for Gcc 2.95.x.

      I will only check in the modified special functions files after 
the imminent release. This will pose a problem too, as a bug report 
filled via sourceforge for one of these files will therefore not be 
addressed in this release, but I find it a lesser evil, and it has the 
advantage of allowing me time (without causing any further delay to the 
impending release) to give valarray support to all existing special 
functions, beyond the one I have done (which is something I want before 
embarking on the new fonctions I promissed a while back), and hopefully 
to ifdef out the bits that could possibly upset Gcc 2.95.x.

   Merci

HH