🚀 go-pugleaf

"Lukas Geyer" <geyer@math.uni-dortmund.de> wrote in message
news:67elj91neu.fsf@math.uni-dortmund.de...
> Simon Waters <Simon@wretched.demon.co.uk> writes:
>
> > Dann Corbit wrote:
> > >
> > > "Simon Waters" <Simon@wretched.demon.co.uk> wrote:
> > > > Dann Corbit wrote:
> > > > >
> > > > > Here is my tweaked version (the stock version won't even build on
my
> > > PC):
> > > > >
ftp://cap.connx.com/pub/chess-engines/new-approach/GnuChess-503.exe
>
> Thanks for providing that Windows version. However, the corresponding
> sources look _very_ Windows-specific, we should come up with something
> portable.
>
> > > > Did you try using gcc?
> > >
> > > This is with the unmodified version (straight from the tar file):
> > >
> > > $ gcc --version
> > > 2.95.3-5
> > >
> > > dcorbit@DANNFAST l:/gnuchess-5.03
> > > $ ./configure
> > ....
> > > checking for inttypes.h... no
> > > checking for stdint.h... no
> >
> > The great thing about standards is they keep changing.
>
> Well, C99 has been around for some time, unfortunately implementations
> still are lacking behind.
>
> > Probably the simplest workaround is just before....
> >
> > typedef uint64_t BitBoard;
> > typedef uint64_t HashType;
> > typedef uint32_t
> > KeyType;
> >
> > to add;
> >
> > typedef u_int16_t uint16_t;
> > typedef u_int32_t uint32_t;
> > typedef u_int64_t uint64_t;
> >
> > I used "#define uint16_t u_int16_t" etc in testing under cygwin.
>
> Well, we could try to solve this with autoconf magic. However, I do
> not like a solution with special windows sources which use C++ and
> some windows-specific headers.

But the UNIX specific headers are fine of course.
;-)

The version I made will compile just fine under UNIX too.

> (This is fine to be distributed by Dan
> but the mainstream gnuchess should remain portable C, I think.)
> Unfortunately I am not able to test with Cygwin due to a local lack of
> MS software, but someone should maybe test with Cygwin before
> releases.

I also fixed the hash table allocation in my latest release.

> Lukas
>
> P.S.: There was another report on problems with the Cygwin readline
> DLL, can someone confirm that?

Here are the results for my latest compile, together with the opening book
that I built

    Program            Elo    +   -   Games   Score   Av.Op.  Draws

  1 GNU Chess 5.03  : 2306   31  27   480    51.7 %   2294   18.3 %
  2 GNUChess-414    : 2294   27  31   480    48.3 %   2306   18.3 %

--
C-FAQ: http://www.eskimo.com/~scs/C-faq/top.html
 "The C-FAQ Book" ISBN 0-201-84519-9
C.A.P. FAQ: ftp://cap.connx.com/pub/Chess%20Analysis%20Project%20FAQ.htm