Two-part patch for hppa64-hpux libstdc++ testsuite failures

[Paolo Carlini]

Phil Edwards wrote:

>Yes; somebody else pointed out, last year I think?, that building the .so
>with -f*-sections is not only /not/ a win, but can actually be a lose on
>certain platforms.  It really makes sense to only build the .a with those
>flags, but this hasn't been enough of a priority for anybody to write up
>the needed (fairly invasive) patch.  Someday...
>
If the needed patch is fairly invasive and non-trivial (almost 
everything seems
non trivial to me, in this area ;) couldn't we just remove the switches 
completely
for now, then carry out extensive tests to decide whether having them 
for the
static build only?

Paolo.