[racket-dev] (round), etc. in Typed Racket

From: Vincent St-Amour (stamourv at ccs.neu.edu)
Date: Sun Dec 12 20:52:01 EST 2010

At Sun, 12 Dec 2010 20:33:08 -0500,
Eli Barzilay wrote:
> -1 for a `racket2',

I agree.

> but why not change racket to that?

Sounds good to me. #lang racket is young enough that I expect most of
the existing code we're afraid of breaking is still written in #lang
scheme (or mzscheme).

> -- In a way
> that doesn't change `scheme', so it's still available for legacy
> code.  Does anyone have an idea how bad such a breakage is?

I guess it depends on what we end up changing. If we're willing to
break backward compatibility like that, we should probably get rid of
as many warts as possible at the same time.

As for integer? vs exact-integer?, I wouldn't be surprised if this
change fixes latent bugs. I wouldn't be surprised if there's code out
there that uses integer? as if it was exact-integer? and could break
because of a stray float.


Posted on the dev mailing list.