I'm not sure whether that Bruno's example should signal an error. CLHS 2.4.6
says
"`basic is the same as 'basic, that is, (quote basic), for any expression
basic that is not a list or a general vector."
#1A(...) represents a general vector. So one could perhaps construct
an argument which says that Bruno's example should be considered
to be equivalent to `#(1 2 ,(+ 2 2) 4).
OTOH, 2.4.6 talks about syntax and gives a, presumably, exhausted
list of permitted syntaxes after backquote.
I'm not sure whether that Bruno's example should signal an error. CLHS 2.4.6
says
"`basic is the same as 'basic, that is, (quote basic), for any expression
basic that is not a list or a general vector."
#1A(...) represents a general vector. So one could perhaps construct
an argument which says that Bruno's example should be considered
to be equivalent to `#(1 2 ,(+ 2 2) 4).
OTOH, 2.4.6 talks about syntax and gives a, presumably, exhausted
list of permitted syntaxes after backquote.
So I think signaling an error is indeed ok.