Finish inkey()/inkey_ex() migration

[flambard-took]

The work done by Billsey to support newer V Z-Term (#944) left one feature in a TODO state: the newer inkey_ex() functions are mostly discarded in favor of our old, trusty inkey().

This was nice, because it kept everything working, and if we're to migrate completely we'll likely break some things. But we do need to migrate, because we do want to actually use the new functions.

Here's what's going on (I'm comparing to V309):

function	In V:	In MAng:
char inkey()	thin wrapper around inkey_ex	main function to grab input
event_type inkey_aux()	main function to grab input	same, but mostly unused
event_type inkey_ex()	return next input event, as a struct	same, but only returns char

In other words, in V, inkey() calls inkey_ex(), which calls inkey_aux(). All the way, event_type is returned, only to be discarded at inkey(). If you want single char, call inkey(), if you want extended event, call inkey_ex(). In MAng, inkey_ex() can't return extended information, so there is no way to get it; inkey() doesn't call inkey_ex(), but does it's own processing.

Also, the macro complete char used in V is 30, yet it is 29 in MAng.

Also, the parse_slash mode is completely removed from V, so there's no special handling of char 28.

For this ticket, I'm keeping the char 28 hack, yet am planning to upgrade macro complete char from 29 to 30, to match V.