[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: when GC is permitted

On Fri, 09 Jan 2004 10:04:39 -0500, Richard Kelsey <kelsey@xxxxxxx> wrote:

Date: Fri, 09 Jan 2004 07:42:48 +0100
From: Felix Winkelmann <felix@xxxxxxxxxxxxx>

Now, I'm all for maximum performance, if possible. But that's not
(IMHO) the proper goal for this SRFI.

I am curious as to what you mean by 'this SRFI'.  Do you think that
maximum performace is not a proper goal for any FFI SRFI?  Or only
for the first one?  Would you be less displeased by SRFI-50 if it
appeared after a JNI-style SRFI was in place?

Maximum performance is secondary here, period. When writing code for
a Scheme FFI that is too low-level and exposes too many implementation-
specific details, you are walking on very thin ice. It requires
implementation-specific knowledge that a user of the FFI normally
shouldn't be bothered with. If the necessary abstractions have
a certain performance impact, then it should be accepted.

I would be less displeased having a single, portable, robust
SRFI for accessing external libraries. It's as simple as that,
and probably a not completely unreasonable desire.

I can understand wanting the first FFI SRFI being a safer, more general
one, perhaps based on JNI or Pika.  This SRFI isn't that SRFI because
that isn't the type of FFI that Mike and I needed.

Sorry, Richard, but I get the impression that you have not really
grasped the intention behind the SRFI process.