Re: pari-2.2.6.alpha released

[Alexander Shumakovitch <shurik@math.unibas.ch>]

On Sun, Jun 22, 2003 at 09:03:39PM +0200, Karim BELABAS wrote:
> This is a bug in the benchmark files: they compute total time using the
This is something I didn't think about ;-) I thought the benchmarks are
written in stone.

> I have updated all benchmark files in CVS to do the right thing.
> Is it better now ?
Everything look coherent now. Thanks a lot! I wasn't sure whether to
install 2.2.6 system-wide or not, because of its "slowness" ;-)

In the meantime, I've stumbled upon another misbehavior:
  (22:32) gp > Mat(5) * [0]~
  time = 0 ms.
  %1 = [0]~
  (22:33) gp > [Mat(5)] * [[0]~]~
  time = 0 ms.
  %3 = 0

The second result is not a column, but a number. It's not the case if
[0] is replaced with [2], say. No problems with versions <= 2.2.5.

Thanks,

  --- Alexander.