Program sfdiffsim now reads a velocity vector with a sample for each
reflector. Each sample of the velocity vector is assigned to a set of
(t0,m0) pairs in t0s and m0s files. Because of that, t0s and m0s files
now have n2 axis assigned to a given velocity sample in velocity vector.

So, v[0] corresponds to a velocity sample for
(t0[0][0],m0[0][0]),(t0[1][0],m0[1][0]),(t0[2][0],m0[2][0])... pairs

and  v[1] corresponds to a velocity sample for
(t0[0][1],m0[0][1]),(t0[1][1],m0[1][1]),(t0[2][1],m0[2][1])... pairs

and so on...

Modify the program sfdiffsim to receive one velocity each time for
a set of (t0, m0) pairs for a given reflector. It allows the program
to run in a loop, each iteration of the loop receive (t0,m0) pairs
picked above one reflector in the stacked section.

To simulate a stacked section with diffraction hyperbolas someone
should use half of the medium velocity in the hyperbola equation
following the model of a exploding reflector.

A division v=v/2 inside of a loop is a logic error because this
division will be done each loop iteration.

Use pm0 and pt0 directly to calculate hyperbola traveltime
instead of using intermediate steps with m0, it0 and t0.

To avoid segmentation fault errors, check if calculated traveltime
sample is outside the section, if it is the case, jump to the next
itaration. This check will allow the user to pick and generate
hyperbolas closer to the section boundaries

Feature/issue/17

The previous SConstruct used deprecated modules and atributes.

Use macros to show debug logging messages in the terminal.
If the macros are defined the code with logging messages will
be compiled, it will help during debug of the for loops

The program Mdiffsim.c should be able to receive a stacked section
with a diferent axis origin in time and CMP, and its origin should
be consider during the index calculations in the main loop of the
program to generate the simulated hyperbolas and determine if the
simulated samples and traces are inside the time and CMP window.

This commit message template shows several especial marks to
be added to the begining of the commit title to allow a script
to set this commit messages in a changelog file. Those marks
can be used during deploy to set relevant commits and the type
of modification

Define a Macro N_SAMPLES_RICKER that establishes the number of samples
of the ricker pulse to be generated before diffraction simulation. If
the number of samples is low, the convolution that builds the ricker
pulse will fail, causing the bug described in #20