Hi Stefano,
The number of fails can be higher as LDS might explore the same parts of the
tree more than once (You probe the tree several times with increasing max
discrepancy). The details I have to check again.
Your second observation I do not really buy into: when you turn a stack into
a queue
The new version is really coming soon, the only thing left is to fix a bug
in search found by Stefano in Gecode and test Gecode/J. Otherwise we are
done.
So if you have one week/two weeks to wait then go for 2.0.1. Otherwise go
for 1.3.1 (which is also a good system...)
Cheers
Christian
--
Chris
Hi,
I know the Gecode trunk is not meant to be stable. But I encountered
some problems that might occure in the next release too.
My setup :
WinXP
Cygwin
g++ (GCC) 3.4.4
aclocal, automake 1.1.0
autoconf 2.61
> Interesting: I do develop under Cygwin with autoconf etc and haven't
> noticed any of the quirks you mention. But thanks!
Yes, I thought so too and was surprised to have these problems.
> In principle, MinGW gcc should be able to link against MSVC (that's what
> they say on their webpage). B
Hi,
Interesting: I do develop under Cygwin with autoconf etc and haven't noticed
any of the quirks you mention. But thanks!
In principle, MinGW gcc should be able to link against MSVC (that's what
they say on their webpage). But as I haven't tried myself I do not know
whether this is true.
Cheer
Martin Mann wrote:
2) your AC_CONFIG_FILES statement at the end of configure.ac doesnt
work
properly on my system. I get the following:
[...]
I fixed it by replacing the following lines
AC_CONFIG_FILES([Makefile doxygen.conf:doxygen/doxygen.conf.in
doxygen.hh:doxygen/doxygen.hh.in
