I just tried to build the newest CVS version of CS on my AMD64 Gentoo Linux box and a small bug prevents it from building.
The file include/csutil/hash.h yields the following error messages:

Code:

./include/csutil/hash.h:101: error: redefinition of `class csHashComputer<long int>'
./include/csutil/hash.h:94: error: previous definition of `class csHashComputer<long int>'
./include/csutil/hash.h:104: error: redefinition of `class csHashComputer<long unsigned int>'
./include/csutil/hash.h:97: error: previous definition of `class csHashComputer<long unsigned int>'

I looked at the appropriate section of the file:

Code:

CS_SPECIALIZE_TEMPLATE
class csHashComputer<long> : public csHashComputerIntegral<long> {};
CS_SPECIALIZE_TEMPLATE
class csHashComputer<unsigned long> :
  public csHashComputerIntegral<unsigned long> {};
   
CS_SPECIALIZE_TEMPLATE
class csHashComputer<longlong> :
  public csHashComputerIntegral<longlong> {};
CS_SPECIALIZE_TEMPLATE
class csHashComputer<ulonglong> :
  public csHashComputerIntegral<ulonglong> {};

I don't have much experience programming C++ or C but it looks like long and longlong might be identical on AMD64.
I didn't know where to report this bug (didn't want to join the mailing list only for that) but I thought I should at least document
it here as AMD64 is not that common as a build platform so it might slip through for quite a while otherwise.

The bug was probably introduced (or "activated" somewhere in the newest code added (according to the cvs log) by res2002
two days ago as he mentions CS_SPECIALIZE_TEMPLATE in the commit message and I remember building a week or so ago without running into
this problem (though I had other build errors, this might have been after those in the build process).

If you need any additional information about my build environment please ask. I will be checking back here in the next days
when I find the time to try again and the build error is not gone.

Thanks. This script was exactly what I was looking for. Although the tutorial does not yet run (segfaults) I believe I can handle it from here as the hard (and boring) part was the build process and that works now.

Edit: Works with --enable-debug.

Thanks for the quick answer.

I installed both manually from CVS as the ebuilds don't work correctly on AMD64. The celtutorial.zip I downloaded contain only a pdf and an sxw file of the tutorial text, a file switch.sh, level data and nine folders p1 to p9 with a src subfolder containing app.h, app.cpp and main.cpp and some of them behave.h and behave.cpp. I downloaded it from http://www.crystalspace3d.org/tikiwiki/tiki-index.php?page=CEL+Documentation (under Local Ressources).

Is there a different tutorial including configure and/or a jamfile?
If not, could you post the one you use (or at least the part needed to compile and link one executable from the respective .cpp and .h files).

I want to learn about Crystal Space to see wether it is suitable (and easy enough to use) for a game I am planning.
It should run on Linux. I tried compiling Crystalspace and CEL on Linux which worked fine.
Then I tried the tutorial and there seem to be neither make nor jamfiles included. I have not much experience with C++ beyond fixing the occasional compile error on
my Gentoo systems and reading Bjarne Stroustroups book a few years back.

My question is now: How do I compile the CEL Tutorial(s) on Linux without any GUI-development environment (I want to use emacs instead)?

I didn't install cel and CS systemwide, just used the CRYSTAL and CEL environment variables. Is there a simple way to achieve this?