Closed addisoncrump closed 11 months ago
I cannot find a version for which the behaviour is consistent, so I assume this bug has been around since the JIT compiler.
re> /\1+?()/match_unset_backref data> a No match data> a\=no_jit 0: 1:
In this case I believe the JIT is wrong, as ungreedy here should just be 1 match of empty.
@zherczeg , this is another one where I think the interpreter's behaviour is the right one.
Fixed in 9de4d53cf850e0fca625ce9d80c12bea5b2a5ab9
I cannot find a version for which the behaviour is consistent, so I assume this bug has been around since the JIT compiler.
In this case I believe the JIT is wrong, as ungreedy here should just be 1 match of empty.