Ignore negative code points instead of erroring

[ronnodas]

From FIGfont:

In more extensive Unicode fonts, you can assign a negative character code (other than -1) to one or more translation tables, similar to #2 above.

This is a minimal change to ignore any codetag with a negative character code. This allows parsing e.g. big.flf without erroring out.