On Thu, Jul 18, 2013 at 3:51 PM, Matthew Flatt <mfl...@cs.utah.edu> wrote:
> At Thu, 18 Jul 2013 15:45:47 -0400, Sam Tobin-Hochstadt wrote:
>> On Thu, Jul 18, 2013 at 3:43 PM, Matthew Flatt <mfl...@cs.utah.edu> wrote:
>>
>> > With this change, does `raco exe' still work? I think `lazy-require'
>> > expands in a way that would require the `racket/place/distributed'
>> > module to be present when creating an executable that uses
>> > `racket/place'.
>>
>> What's the right way to test this?
>
> Try `raco exe' on
>
> #lang racket/base
> (require racket/place)
>
> in an installation without distributed places.
This gives the error
create-empbedding-executable: expansion mismatch when getting external paths
which, while spelled wrong, I think indicates that I did indeed break things.
>> > Probably the right approach is to have a `#:node' argument effectively
>> > carry `supervise-place-at' with it, so that `lazy-require' is not
>> > needed.
>>
>> I don't understand how it would "carry `s-p-a` with it". What do you mean?
>
> It looks like a valid `node' value is an object, and so the object
> could have a method that is called instead of passing the object to
> `supervise-place-at'.
It appears that `supervise-place-at` is almost exactly like the
`launch-place` method of `remote-node%`, except for the issue of named
vs unnamed places, which I don't fully understand yet.
Sam
_________________________
Racket Developers list:
http://lists.racket-lang.org/dev
