Re: Exposure of "direct" arrays

--6c2NcOVqGQ03X4Wi
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

rif wrote:
> Under CMUCL, there is functionality to make an array accessed by a C
> program and CL have the "same data" by getting the address of the
> array from CMUCL.  This is obviously risky (the manual suggests you
> should turn off garbage collection when you do this to avoid your
> thing getting moved out from udner you, for instance), but I'm finding
> that at this point, a substantialy fraction of my total runtime (maybe
> 20%) is spend basically copying my data between my CMUCL array and the
> foreign array which is passed to C.  So is this functionality exposed
> at all in UFFI, or do I have to switch to using CMUCL's FFI directly
> in order to get at it?

UFFI doesn't support this. It is a feature under consideration for a
rewrite of UFFI that would be based on higher-level macros that could
encapsulate this behavior.

--=20
       Kevin Rosenberg        |  .''`.  ** Debian GNU/Linux **
  http://b9.com/debian.html   | : :' :      The  universal
  GPG signed and encrypted    | `. `'      Operating System
     messages accepted.       |   `-    http://www.debian.org/

--6c2NcOVqGQ03X4Wi
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.0 (GNU/Linux)

iD8DBQE+ZSG4ES7N8sSjgj4RAgVBAJ9nQ+hyN33HzxHYyC2B1u06NBHGDACggSKf
uVOmPYIXq8Y+/eb+4rZFUiM=
=ma2q
-----END PGP SIGNATURE-----

--6c2NcOVqGQ03X4Wi--