[racket] TR Regression: Types for regexp operations

From: Robby Findler (robby at eecs.northwestern.edu)
Date: Mon Jun 20 01:41:09 EDT 2011

FWIW, there is a bug report form as a part of DrRacket that is the
preferred means of submission (in the Help menu).

Robby

On Mon, Jun 20, 2011 at 1:15 PM, Richard Lawrence
<richard.lawrence at berkeley.edu> wrote:
> Hi Typed Racket folks,
>
> I checked out the development version of Racket today, and noticed what
> I believe to be a regression: TR thinks that regexp-match and friends
> return a String when given an Input-Port, but in fact they return Bytes:
>
> ; as expected: string input returns String
> (regexp-match #rx"^[0-9]" "100")
> - : (U False (Pairof String (Listof (U String False))))
> '("1")
>
> ; wrong: bytes typed as String
> (regexp-match #rx"^[0-9]" (open-input-string "100"))
> - : (U False (Pairof String (Listof (U String False))))
> ;                    ^^^^^^
> '(#"1")
>
> ; in 5.1.1:
> (regexp-match #rx"^[0-9]" (open-input-string "100"))
> - : (U False (Listof (U Bytes False)))
> '(#"1")
>
> Also affected:
> regexp-match*
> regexp-try-match
> regexp-match-peek
> regexp-match-peek-immediate
> regexp-split
> (possibly others)
>
> I have not confirmed this but I suspect commit #
> 029651b9eeaed6b4cecb800da2c01fa6cbed1270.
>
> By the way, is there a more official way to submit a bug report?  I
> couldn't see one on racket-lang.org...should I have sent this to the dev
> list?
>
> Best,
> Richard
>
> _________________________________________________
>  For list-related administrative tasks:
>  http://lists.racket-lang.org/listinfo/users
>



Posted on the users mailing list.