quoted

The matching algorithm would parse the array as a finite automation
eating up byte by byte in the text.

Looks good but Boyer-Moore algorithm doesn't need this at all. BM 
doesn't support wildcards like '*' because of the shiftings that it uses 
to look for matches. This issue together with memory consumption are two 
limitations that we have to live with if we want to use BM. Anyway I 
think it's worth it because we can perform search in O(n/m).

This would be _additional_ to BM/KMP/naive/finite automata/you name it.
The benefit of libqsearch is hide various algorithms behind a single
interface, isn't it? It's argueable whether it would make sense to try
and rule out bad shifts while parsing a chain of non-wildcard regular
expression tokens but I guess that would be over the top.