If the cmap parser was the one to cause _getGlyphNamesFromCmap() to be called, no unicode cmap was found as it was just starting to get loaded. This resulted in different glyph names, depending on when the cmap parser was invoked. Also added a bunch of comment describing what this method does. git-svn-id: svn://svn.code.sf.net/p/fonttools/code/trunk@204 4cde692c-a291-49d1-8350-778aa11640f8