[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[gnugo-devel] Current CVS
From: |
Daniel Bump |
Subject: |
[gnugo-devel] Current CVS |
Date: |
Mon, 28 Jan 2002 07:06:59 -0800 |
I just added the patches heikki_1_23.3, arend_1_23.[45] and
inge_1_23.4 to the CVS.
The current CVS version does nicely on the regressions
with 25 PASS and only 3 FAIL.
There are two outstanding patches. We could consider
doing a release now and dealing with those later, or
we could wait. Let me explain why these are still
not in the CVS.
The semeai patch is still crashing after bringing it into
sync with Gunnar's revision. The patch as posted still
clears the reading cache but the patch should not crash
if this is taken out, and it does.
Trevor's patch should be made contingent on a configure
option, so the definition of EXPERIMENTAL_READING is
moved to config.h. I intend to revise this patch and
use it but I am having trouble getting the optional file
read_attack.c into GGBUILTSOURCES in patterns/Makefile.
This is what I tried in configure.in:
AC_ARG_ENABLE(experimental-reading,
[--enable-experimental-reading use experimental reading],
[if test ${enableval} = yes; then
read_attack_c=read_attack.c
else
read_attack_c=
fi],
[read_attack_c=])
AC_SUBST(read_attack_c)
AH_TEMPLATE([EXPERIMENTAL_READING],
[Experimental Reading. 0 standard.])
if test "$enable_experimental_reading" = "yes" ; then
AC_DEFINE(EXPERIMENTAL_READING, 1)
else
AC_DEFINE(EXPERIMENTAL_READING, 0)
fi
Then in patterns/Makefile.am the definition of
GGBUILTSOURCES contains @address@hidden The idea
is that this would be replaced by read_attack.c if
you configure --experimental-reading but not
otherwise.
I don't see what's wrong with this scheme but
autoconf refuses to do it.
Any ideas?
Dan
- [gnugo-devel] Current CVS,
Daniel Bump <=