Re: tail recursion (was Re: a (revised) draft of R5RS)

[Matthias Blume <blume@CS.Princeton.EDU>]

From: Kent Pitman <kmp@harlequin.com>
Subject: tail recursion (was Re: a (revised) draft of R5RS)
Date: Tue, 27 May 97 16:51:40 EDT

I wrote:
>    tricky cases. For example:
> 
>       (call-with-input-file file f)
> 
>    Intuitively the call to f will not be tail-recursive,

Kent answers:

> I don't agree.  Of course the call will be tail recursive.  FROM ITS
> POINT OF CALL, which is not the caller of CALL-WITH-INPUT-FILE. [ ... ]

Nobody knows that without additional information (such as seeing the
source code for call-with-input-file).  That was my point.

And in most implementations of Scheme I expect the loop

 (define (loop) (call-with-input-file some-file (lambda (p) (loop))))
 (loop)

to run either out of file handles or out of memory.  This is what I
mean by "intuitively".  The implementation must jump through quite a
few hoops in order to make the above not run out of resources (but it
could be done as I outlined before).

-Matthias