[racket-dev] Overly general types for mutable containers

On 07/07/2012 10:28 PM, Sam Tobin-Hochstadt wrote:
> On Sun, Jul 8, 2012 at 12:58 AM, Neil Toronto <neil.toronto at gmail.com> wrote:
>> It runs directly counter to what I expect from immutable containers, which I
>> use most of the time:
>
> This is the problem.  Immutable containers are very different from
> mutable ones, and your expectations shouldn't be expected to carry
> over.  Mutability is a communications channel, not just a data storage
> mechanism, and you should expect it to be different.

Yes, I'm seeing that now. A lot. :D

(I apologize for my recent negative tone. I've made up for it by 
submitting bug reports for you. Or something.)

I keep trying to come up with better rules for generalizing the 
containee types in mutable containers. I haven't found any that don't 
have problems, so I'll have to be content with your choices on this.

Neil ⊥