On Thu, 10 Sep 2009 12:45:14 -0500, Andre van Tonder <[email protected]>  
wrote:

> I'm not sure that is the case.  At least in R6 Larceny, you can write the
> portable
>
>    (import (foo))
>
> and let the implementation worry about finding the library for you.  In
> other words, finding the library is not part of the programmer's
> responsibility.  Now, there may be some implementation-specific  
> restrictions on
> where you may /put/ the library in the file system, and what you may  
> call it,
> but that is not reflected in any way in the syntax of the portable  
> program, and
> IMO it is not the place of a spec to really nail this down.

Sure, I can write (import (foo)) in any implementation, and then also  
write a chart of N different methods I need to use to get an  
implementation to actually *do* something sensible with that. It's an  
unnecessary bother, and I don't think it requires importing all of POSIX  
just to talk about generalized notions of "files" and "directories" that  
could be used to nail down how the library is found a little bit more. And  
if all else fails, I'd love just to be able to concatenate the libraries  
my program requires together with a top-level program and hand the whole  
kit-and-kaboodle to my implementation.

Ultimately the implementation is responsible for reading a top-level  
program and its dependent libraries in some fashion, compiling, and  
executing the result. Unless the concatenate-the-world solution is  
endorsed, they must be separated in some fashion in the input. I don't  
think it's too much to speak loosely about directories and files and leave  
the specifics of how that actually works up to the implementation. The  
result will be that implementations sharing a platform will do the same  
thing where the concepts of "directories" and "files" have clear  
definitions, and that your interpretive dancers will have to wear placards  
describing the locations of the libraries they implement.

-- 
Brian Mastenbrook
[email protected]
http://brian.mastenbrook.net/

_______________________________________________
r6rs-discuss mailing list
[email protected]
http://lists.r6rs.org/cgi-bin/mailman/listinfo/r6rs-discuss

Reply via email to