Your GZIP file seems to be corrupted.

gzip: Cycle.tar.gz: unexpected end of file

Perhaps Wikifortio cannot handle well GZIP files. Could you gzip it again and send it to ufec@atlas.cz? Or try to upload it again?

Never mind.
I downloaded it again and it works.

So I understand, that you managed to run it and put it on your server. If it does not work, please send me a PM, that we can find a solution. Thanks Lesan.

Both in virtual machine and using a live CD I only get a framerate of about 1 FPS. When I have 8 GB of RAM, do you think using a non-live non-virtual Linux will increase the framerate?

Wow, thats not good. Hmm, the meshes for the background are big, but not so really huge, that it should reduce framerate in such a way. Is it possible to launch the graphics driver from your hardware vendor? E.g. nVidia has its own Ubuntu Graphics Driver. Maybe Live CD and Virtual Box make use of software rendering. This would be bad.

If you can link me to a copy of the source I could take a look at producing a build for Windows. I've ported my own software between Windows and Linux before, using roughly the same sets of technologies.

I tried the whole week to make it run on Windows. Here is my attemp. There seems to be a problem in the strategy mode - but only on Windows. It is not recognized on Linux. Also the stl string library seems to have problems with solving the path of the mtl libraries of the wavefront meshes. Lighting is also a bit strange on Windows. I have no answers.

Would be great if you manage to compile it. You should download the resources from the other link.

Thanks in advance.

Source
http://www.wikifortio.com/754277/Cycle_Source.zip

Assets
http://www.wikifortio.com/839120/Cycle.tar.gz

I've gotten it to compile and link, but you appear to have some rather serious bugs in the program. MSVCs vector debugging caught at least two instances where you stepped out of bounds of a vector<>.

I can see about posting the version I built, but it is rather useless unless you can fix these bugs. One possible approach would be to riddle the code base with assert statements, so if you are looping from i = 0 to N on a vector, assert(vector.size() >= N); That, and you can use the function std::vector::at() rather than operator[] for accessing elements, you will get an exception but at least it will alert you to places where you need to look at your logic.

It was rather easy to get working. If you are interested in improving the portability of your code I highly suggesting using a higher warning level when compiling, most of the compile messages I got were warning about implicitly converting doubles to floats (always use the "f" modifier on floating-point literals unless you really want a double) or signed/unsigned integer arithmetic (std::vector::size() is unsigned, prefer to use vector::size_type over int when looping across a vector).

On a more positive note, the loading bar looks really good!

Thanks for the feedback. I try to find some bugs. However, I tried to avoid extending std::vector arrays. Its strange that exatly this causes an error...
K, the LINUX version is available (and tested by some collegues...). I'm going to invest some time in debugging on Windows.