Use make-prefab-struct instead of constructor for prefab structs

[AlexKnauth]

closes #3

[samth]

I think there are tests for pconvert somewhere.

Also, you might use for/list and in-vector to avoid the intermediate list.

[AlexKnauth]

I just added a test.

[AlexKnauth]

For that I used the same pattern as https://github.com/racket/pconvert-lib/pull/4/files#diff-cb76abd3c1f365f0647a55475b35155eL475 a few lines later, but with a direct use of recur instead of a lambda.

[AlexKnauth]

What do you think?

[samth]

Looks reasonable to me. Avoiding the intermediate list would look like

(for/list ([i (in-vector (struct->vector expr)]
           [j (in-naturals)]
           #:unless (zero? j))
  (recur i))

[AlexKnauth]

Do you want me to change the https://github.com/racket/pconvert-lib/pull/4/files#diff-cb76abd3c1f365f0647a55475b35155eL475 a few lines later too? Come to think of it I could change https://github.com/racket/pconvert-lib/pull/4/files#diff-cb76abd3c1f365f0647a55475b35155eL163 to avoid intermediate lists as well. (Edit: and this https://github.com/racket/pconvert-lib/pull/4/files#diff-cb76abd3c1f365f0647a55475b35155eL160) (Further edit: and this https://github.com/racket/pconvert-lib/pull/4/files#diff-cb76abd3c1f365f0647a55475b35155eL281) (Further edit: and this https://github.com/racket/pconvert-lib/pull/4/files#diff-cb76abd3c1f365f0647a55475b35155eL394)

[samth]

Up to you. That could also be a separate change -- I think this is good to go as is.

[AlexKnauth]

I think it should be a separate change.

[AlexKnauth]

I'll make a separate pull request after.

[AlexKnauth]

Are you going to merge?

[samth]

Done.

[AlexKnauth]

Is for not provided by mzscheme?

[samth]

That wouldn't shock me. Porting the code to racket/base would also be nice ...