testsuite oddity

[N8TM@aol.com]

In a message dated 11/24/98 7:01:18 AM Pacific Standard Time,
espie@quatramaran.ens.fr writes:

>  If by any chance you get
>  a regression in the libio/libstdc++ tests, the corresponding call to
runtest
>  actually returns an error...  this does stop a make check dead in its
tracks.
> 
>  
Au contraire, I was able to run the most recent testsuite on linux-gnulibc1,
irix6.4 and hpux10.20 without the previously usual -k option.  make install,
however, still requires -k on all the Intel systems, both linux and cygwin.