TODO.txt

Issues:
- are there any more hard-coded constants? I have removed the one from the forces.c

- as I understand, STRETCHING_FORCE, BENDING_FORCE and AREA_FORCE_LOCAL are
  only bonded interactions, and do not cost anything to be permanently
  compiled in. Maybe, we can remove the #ifdef-guards?

- we don't need guards for bonded forces, as they only act
- ESPResSo has global variables this_node and n_nodes, no need for
  Comm_rank etc.

- too much hard-coded PATHS etc in the samples, fix the scripts

- we need a test case. Should run a couple of secs and check that no
  future changes to the code break the computations. Normally,
  checking that the forces in a given setup are correctly calculated
  is sufficient.

Currently resolved issues, please check my changes!
- the contents of mathtools.c/h is general, and should be merged into
  utils.h; also check which functionality already exists.
  - some don't make sense, e.g. double2double should better be done
    by a simple memcpy.

- there is a lot of commmented out code, can we just remove it?
  Commented out test printfs is ok, but not whole functions.


- hard-coded constant, e.g. "for (int i=0;i<100000;i++)"
  in forces.c. Make the limit configurable, at least as a C constant
  in config.h (compare SHAKE_MAX_ITERATIONS)


- we should get out all the comments "new code", as this won't be new soon,
  and for checking, diff is also a better tool.

- removed warnings:
  - changed img data type to correct data type int, to prevent warnings.
  - removed unused variables

- prepare_vmd_connection/writepdb were changed. That has nothing to do
  with FSI, so why is it there? Also, the code is actually wrong,
  and breaks e.g. being able to fold coordinates. So, I took it out.

- same in imd, taken out

- the documentation is currently only a PDF, we need the sources.
  That can be part of the UG, as it is short enough, or we could
  move all to the tutorials directory. Anyways, all new Tcl
  commands / interaction need to be documented in the UG.