Torbjorn Granlund wrote:
> We won't be sending any more code replacement blobs to this address; it
> is most surely the wrong place.

Hi Torbjorn,

I guess you're saying that because there's been too little feedback?
IMHO, this is great work.
I've been reviewing the latest and had prepared several patches.
Just hadn't made time to send them.

> Please get our suggested factor.c replacement from
> <http://gmplib.org:8000/factoring/>.
>
> I plan to spend no more time on this project now.  Should the
> contribution be accepted, I will make the necessary amendments to the

You may consider it accepted.  That was clear in my mind from
the beginning.  Sorry if I didn't make that clear to you.
Now, it's just a matter of integrating it.

> GNU copyright paperwork.  I am certainly willing to answer questions
> about the code, of course.

Here are some suggested changes -- I made these against
a temporary local git repository using your -005 tarball.
That was before I learned (just now) that you have a mercurial
repository.

I made the Makefile parallelization changes mostly to avoid
waiting too long when I run "make check" -- in the very short run.
Obviously we cannot use its GNU make features in coreutils/tests.

In coreutils, we are pretty strict on warnings, so most of these
changes are to avoid the few that remained.  I've also moved
some declarations "down" to be nearer their point of first
initialization.  That's another style issue in coreutils.  We
are no longer required to declare all variables at the top of
each block.