regnodes for won't match unless target string is UTF-8?

Consider

qr/foo...bar\x{100}/

This cannot match a target string that isn't in UTF-8.  We can determine 
this at pattern compilation time.  We could introduce new EXACTish nodes 
that indicate that no match is possible for that node if the target 
isn't UTF-8.

An attempted match of such a node could be rejected immediately if at 
execution time the target string isn't UTF-8 instead of having to test 
each byte in the node against the corresponding byte in the target until 
we come to the failing one (potentially having to fold them).

This would be a big win if cases like this are common in the real world. 
  Otherwise, it's probably not worth it.

This is a pretty trivial change, and we are not running out of available 
node-types.  I'm unsure of its potential value.

• regnodes for won't match unless target string is UTF-8? by Karl Williamson
• Re: regnodes for won't match unless target string is UTF-8? by demerphq
• Re: regnodes for won't match unless target string is UTF-8? by Aaron Crane
• Re: regnodes for won't match unless target string is UTF-8? by Karl Williamson