A couple things. Not 100% sure what you mean by accumulators. I'm also guessing you are using strategies. Any reason? Finally an example of what isn't working would be super helpful for knowing how we can help. Not sure if someone has reimplemented drstructuring already using meander, but should definitely be possible.

Thank you Here's an example where the final pattern fails

(let [s
      (m*/rewrite
       [[!b ..?n '& ?bs] [!x ..?n & ?xs]]
       [!b !x ..?n
        ?bs ?xs])]

  [(s '[[x & xs] [1 2 3]])
   (s '[[x y & xs] [1 2 3]])
   (s '[[x y z & xs] [1 2 3]])
   (s '[[x y z & xs] [1 2]])])

Yeah, that fails because you are saying that the repeats for both vectors should be the same. That is what using ?n twice does. So on that last example, there are 3 elements before the & and there are only two on the right hand side. So they don’t match.

Well, it's pretty clear they don't match, I just provided some examples of stuff I tired. I'm not sure what would work correctly, though :thinking_face:

Well, it depends on what you want. I’m guessing you want to assign z to nil and xs to nil in that least example? If so, you can do something like this

(let [s
      (m*/rewrite
       [[!b ..?n '& ?bs] [!x ..?n & ?xs]]
       [!b !x ..?n
        ?bs ?xs]

       [[!b ..?n '& ?bs] [!x ...]]
       [!b !x ..?n
        ?bs nil])]
  [(s '[[x & xs] [1 2 3]])
   (s '[[x y & xs] [1 2 3]])
   (s '[[x y z & xs] [1 2 3]])
   (s '[[x y z & xs] [1 2]])])

Ah, the second pattern is for when the first doesn't match then it all gets accumulated to !x?

Yep exactly right.