ITRANS ## toggle interprets some characters

[vpysb]

Thanks for adding the ITRANS toggle ## support which is meant to toggle transliteration.

There is an issue with some text that is with in ## xxxxxx ##

A test with all alphabets shows the issues:
Input:
##abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ01234567890~`!@#$%^&*()_+{}[]:";'<>,.?/|##

Output: when output language is devanagari
abcdefghijklmnopqrstuvvkShyzABCD^eFGHIzKLMNoPQRSTUVWXYZ01234567890`!@#$%^&*()_+{}[]:";'<>,.?/.\

The following alphabets are interpreted and converted (which should not be the case):
w converts to v
x converts to kSh
E converts to ^e
N converts to ~N
O converts to o
| converts to .