[racket] Determining type from a syntax object

On Jun 16, 2014, at 2:53 PM, Sam Tobin-Hochstadt <samth at cs.indiana.edu> wrote:

> This is exactly what predicates are for, in general.  However, you
> can't have predicates that check things about functions -- there's no
> information there at runtime to look at.

I was kind of wondering if there was anything that could make the type checker put something there for runtime to look at.  

> Typed Racket functions are
> just plain Racket functions. We could add some extra metadata to every
> value that held its type,

It wouldn’t have to be every value, it could just put that information in for :has-type expressions, right?  
Would that still require fundamental changes to Racket?  

> and then implement this operation, but that
> would require fundamental changes to Racket.
> 
> Even if we wanted to do that, what if `f` came in from untyped Racket?

If f comes from untyped Racket, then the type checker says to use require/typed to import it.  

> 
> Sam
> 
> On Mon, Jun 16, 2014 at 2:49 PM, Alexander D. Knauth
> <alexander at knauth.org> wrote:
>> Would there be a way to test the type for making decisions at run-time (not
>> just with predicates)?
>> 
>> For example:
>> (cond [(:has-type? f (Number -> Any))
>>       (f 1)]
>>      [(:has-type? f (String -> Any))
>>       (f "1")])
>> 
>> On Jun 16, 2014, at 6:54 AM, Matthias Felleisen <matthias at ccs.neu.edu>
>> wrote:
>> 
>> 
>> That's what I thought you wanted it for so my answer stands -- Matthias
>> 
>> 
>> 
>> On Jun 15, 2014, at 12:55 AM, Spencer Florence wrote:
>> 
>> This is about making decisions at compile time. Specifically I have a
>> sequence of expressions I want to partition into expressions of some type T
>> and expressions of other types.
>> 
>> 
>> On Sat, Jun 14, 2014 at 7:34 PM, Robby Findler <robby at eecs.northwestern.edu>
>> wrote:
>>> 
>>> Would it be enough to expand into an 'ann' expression? Or do you need
>>> to make decisions at compile time based on whether or not the types
>>> worked?
>>> 
>>> Robby
>>> 
>>> On Sat, Jun 14, 2014 at 2:04 PM, Matthias Felleisen
>>> <matthias at ccs.neu.edu> wrote:
>>>> 
>>>> No, TR expands first, then checks. -- Matthias
>>>> 
>>>> 
>>>> 
>>>> 
>>>> On Jun 14, 2014, at 2:59 PM, Spencer Florence wrote:
>>>> 
>>>>> Hey All,
>>>>> 
>>>>> I'm trying to take advantage of typed/racket in a few macros. Is there
>>>>> any way to check the type of an expression from its syntax object? something
>>>>> like:
>>>>> 
>>>>> (:has-type? (-> Void) #'expression)
>>>>> 
>>>>> 
>>>>> --Spencer
>>>>> ____________________
>>>>> Racket Users list:
>>>>> http://lists.racket-lang.org/users
>>>> 
>>>> 
>>>> ____________________
>>>>  Racket Users list:
>>>>  http://lists.racket-lang.org/users
>> 
>> 
>> 
>> ____________________
>> Racket Users list:
>> http://lists.racket-lang.org/users
>> 
>> 
>> 
>> ____________________
>>  Racket Users list:
>>  http://lists.racket-lang.org/users
>>