maybe fix 13465

[Toon Moene]

Richard Henderson wrote:

> It solves the quadratic problem only insofar as the user doesn't
> write out all the values by hand.  But since the problems seen
> in real code involve (large) repeat counts, this may solve the
> practical problem.
> 
> Just now running the full testsuite, but it does the right thing
> for a couple of tests I've tried.  I don't trust the testsuite to
> be very comprehensive, so please look it over for corner cases.

Your change still makes use of the multiprecision arithmetic package 
included in GMP for counting out the array elements to be filled.

My reasoning for forgoing that was that this is frontend business, and 
hence doesn't have to use any other "counting mechanism" than the host's 
one to get this correct.

Am I missing something ?

Thanks in advance,

-- 
Toon Moene - mailto:toon@moene.indiv.nluug.nl - phoneto: +31 346 214290
Saturnushof 14, 3738 XG  Maartensdijk, The Netherlands
Maintainer, GNU Fortran 77: http://gcc.gnu.org/onlinedocs/g77_news.html
A maintainer of GNU Fortran 95: http://gcc.gnu.org/fortran/