0.4-5	Corrected C code underlying the reading of DBF files to allow character fields to be arbitrarily wide (bug report Trondheim diamonds with fields over 200 characters wide - buffer overflow from unchecked max width 49); Added access to CHANGELOG through maptools() function.
0.4-4	Corrected C code for reading SHP/SHX files to reduce danger of mismatched PROTECT/UNPROTECT pairs, and hopefully resolve issues with crashes on Windows platforms in some cases.
0.4-3	Back out of a too-aggressive heuristic for guessing ring-directions and plot-ordering
0.4-2	Introduced changes including warnings in Map2poly() and plot.polylist() with regard to rind-direction and plot-order heuristics - Map2poly() tries to determine which polygons and sub-polygons are wholly contained within each other, to note the order in which they should be plotted in plot.polygon() to avoid over-plotting when painting polygons, and, if the raw= argument is FALSE, to correct possible wrong ring directions (lakes are not usually found in the sea etc.). plot.polylist() then tries to use this information to avoid over-painting polygons by painting enclosed units after their enclosing units, and, if the forcefill= argument is FALSE, will paint anti-clockwise rings (aka holes) with background colour rather than fill colour. At the present, raw= and forcefill= defaults are TRUE, and warnings are issued that this behaviour will change. The underlying problem is that most shapefiles in circulation do not follow ESRI specifications for polygons, many have at least some rings in the wrong direction, over and above slivers and dangles. This is unlikely to change, so defensive code is required.
