Updated function declarations in triangulate.c and adssub.c, to address warnings from CRAN Checks page. Incremented version to 1.5-8.

Deleted 3rd example in triangulate function example (segmentation fault on my laptop)
Set in.poly parameter bdry as FALSE in both code and documentation.

Updated branches/stable-1.5 from ads-1.5-4 and updated NAMESPACE and DESCRIPTION for with new spatstat packages family

Yet another cosmetic change in authors attributes in DESCRIPTION file and deleted unecessary comments in NAMESPACE

Deleted transpose function in utils.R since the C function is absent from utils.c. Deleted it as well from internal.Rd
Deleted utils.c and moved function triangulate to triangulate.c
Renamed C function triangulate into c_triangulate and updated triangulate.R
Added new file init.c to declare C and Fortran functions as recommanded by outputs of tools::package_native_routine_registration_skeleton(".")
Updated NAMESPACE "useDynLib(ads, .registration=TRUE)" as recommanded for R >= 3.4
Updated MD5

In src/*.h cosmetic changes to declare the same way all functions. In fads.R there was a wrong call to ripley_tr_disq_ic instead of coor_tr_disq_ic. Deleted comments en plot.vads.R that were problematic according to devtools:check() function. Updated DESCRIPTION, NAMESPACE and MD5 files

This commit adds a new option to the ads::mimetic function: the user may provide a raster in the arguments for defining a mask. 
mimetic(..., mask=NULL)
mask must either inherit from RasterStack or RasterLayer at the moment (we might change this in future commit). Every non NA values are considered as allowed values for positioning new points.
It only works for rectangular windows without triangle at the moment (we will extend the option to other window type in future commits).

In the C sources, I updated the functions mimetic_rect(), exchange_point_rect() and created a new function s_alea_rect_mask based on s_alea_rect. The main differences is that instead of 
generating new points anywhere within the sampling window, it only generates points within the raster limits inside the sampling window.
The precision parameter might not be implemented properly for the masked point process, it should be double checked.
The updated functions take the additional arguments nmask, xmask, ymask, dx and dy:
nmask: number of non-NA points of the raster
xmask: x-coordinates of the non-NA points in the raster
ymask: y-ccordinates of the non-NA points in the raster
dx: dimension of one raster cell along the x-axis
dy: dimension of one raster cell along the y-axis

The ads::mimetic function does not check whether the spatial point pattern and the raster are consistent in terms of CRS. It assumes at the moment that they are expressed in the same CRS. The function gets rid of any point from the raster that are outside the rectangular sampling window.

Formatted the code using Netbeans formatting tool with Whitesmith style.

Updated all the mimetic_* functions to handle properly spatial point patterns with negative coordinates.