[plt-scheme] attaching object names to values

require/contract is given a name tho, isn't it? Why not use that name?

Robby

On Sat, Dec 6, 2008 at 4:57 PM, Sam TH <samth at ccs.neu.edu> wrote:
> On Sat, Dec 6, 2008 at 11:55 AM, Robby Findler <robby at cs.uchicago.edu> wrote:
>> I think there isn't one, but if you see how to make one, please let me know.
>
> It seems like a procedure that just changes the `object-name' of value
> would be valuable.
>
>> FWIW, provide/contract does a better job than just `contract' and I
>> believe would do what you want.
>
> Unfortunately, this is in the implementation of a `require/contract'
> form, so I can't use `provide/contract'.
>
> Thanks,
> sam th
>
>>
>> Robby
>>
>> On Fri, Dec 5, 2008 at 12:44 PM, Sam TH <samth at ccs.neu.edu> wrote:
>>> The contract system uses the `object-name' of predicates to print out
>>> their names in error messages.  But this can be useless if the
>>> predicate itself is wrapped in a contract.  This program shows the
>>> problem pretty well:
>>>
>>> #lang scheme
>>>
>>> (define x? (contract (-> any/c boolean?) (lambda (x) #t) 'here 'there))
>>>
>>> (define my/c (-> x? x?))
>>>
>>> (contract my/c 5 'foo 'bar)
>>>
>>>
>>> It gives a contract error referencing "unsaved-editorx:y:z", instead
>>> of "x?".  Unfortunately, there seems to be nothing that I can do to
>>> change the object-name of the contract value (except eta-expand).
>>> Notably, the inferred-name syntax property doesn't work, since the
>>> actual lambda expression isn't available (it's in the contract
>>> library). Is there a solution here I'm missing?
>>>
>>> Thanks,
>>> --
>>> sam th
>>> samth at ccs.neu.edu
>>> _________________________________________________
>>>  For list-related administrative tasks:
>>>  http://list.cs.brown.edu/mailman/listinfo/plt-scheme
>>>
>>>
>>
>
>
>
> --
> sam th
> samth at ccs.neu.edu
>
>