Build testing on GNU Chess 5.06

[Simon Waters]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Okay, Anders and I have been busy building (or failing to build)
GNU Chess 5.06 - more volunteers welcome!

I can supply a tar ball by e-mail for those who don't want to
"make dist" out of CVS.

I have committed a number of small but relevant changes to CVS
this evening.

In particular;

1) my multiline string literals in cmd.c are deprecated, and now
replaced with one printf statement (as Mike had suggested for
his build with non-Cygwin Windows compilers).

2) The Makefile entry to include "getopt_long" source on
platforms without getopt_long now works (finally).

3) Obselete code that was preventing detection of correct flags
for Posix threads has been removed from configure.ac.

So far I have smooth builds on;

        Redhat Linux GCC 2.95.something

        Tru 64 5.1 with the native C compiler

HP-UX 11.11 builds native C compiler but the Posix threading
code appears bust, it sits consuming CPU without producing a
prompt. GDB claims it to be in the "while" loop  at line 142 in
input.c. I'm suppose to be the HP-UX guru, but I have no idea on
this, nor if I can easily trace what both threads are doing (if
two threads are running on the test server).

Noted that the "prompt" output on Tru64 looks messy when "post"
is enabled, but I don't think I can run X back from the test
server to test with Xboard and analysis - I'll check with HP.

Tru64 C compiler complains that input_thread seems to be both
external and locally declared.

HP-UX complains that Macros "MIN" and "MAX" are already declared
in one of it's header files, next thing to check is if this is
the cause of the hangs.

How are we suppose to keep CPP macro namespace from having
collisions BTW?

 Simon
-----BEGIN PGP SIGNATURE-----
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQE+ZoD1GFXfHI9FVgYRAqSxAKCOeK7/rd9o5zaTc+rzx7bdldrlUwCggrGC
eLiDEtC8JT3lls6r6w5fWcg=
=KKl2
-----END PGP SIGNATURE-----