Excessive resource consumption when compiling cs_pyth.cpp

[anonymous]

When I compile Crystal Space with jam everything runs smoothly until it hits cs_pyth.o. Then it just sits their and ends up taking up more than 300 MB of memory. If I had let it run any longer my system would have crashed so I shut it down with the System Monitor. It skips it and the compilation of the Python modules. The rest of it compiled okay but I doubt it has Python support.

[sunshine]

The current plan is to break apart the bindings into smaller modules which can be built more easily. (The breaking-apart is an implementation detail. From the client perspective, the bindings will still appear as usual -- i.e. 'cspace'.)

See this thread for a discussion of the issue:  ML-main-thread:10725

[tupone@…]

I got the same problem compiling CS with -O3 A build with -O2 still takes ages to compile but at last it builds successfully.

[sunshine]

Resolution of this ticket may also indirectly resolve #210.

[sunshine]

New split-binding architecture merged into trunk with r27769, r27770, r27771, CEL:r3012, CEL:r3013.