Once again some help from Pedro Rivera. He has suggested to add space mapping for the white space characters 0, 9, 10, 12 and 32 (is already present) to the the StandardEncoding, MacRomanEncoding and WinAnsiEncoding tables . I had to really thing about this proposal as it is far reaching. It appears to be valid for all font types except CID types as the suggest characters codes would get mapped to other values via the toUnicode map. When a toUnicode is used we don't use the cmap tables defined in Encoding so this change should only improve glyph association. I check the code and we use the origional character code to calculate the width the associated width and not the mapped character value.

The main reason for this proposed change is that many fonts don't have glyphs for the character codes mentioned and will put in box or '?' as a substitution, it all depends on the font program being used.

The QA framework should show any big issue with this change.