toke.c: S_intuit_more: More effectively use S_scan_ident

[khwilliamson]

People have avoided looking at this routine for a very long time. As a result it hasn't known until recently about lexical variables, for example. This p.r. rectifies that. The new enhancements committed today to scan_ident are now used by this function to examine potential identifiers better.

If 'strict vars' is in effect, and a variable is not found, it now assumes this is an error, or a character class was meant. This was done for the variable in effect at the time of the call to this function (external to it) recently via 3450d19250e29cc6fc0f4161db65d5995435c2ed. But is here extended to potential variables it encounters in its parse.

• This set of changes does not require a perldelta entry.