clsql/postgres works with CLISP -- MOP works fine

Hi,

this is a short notice that I now successfully tested clsql-3.1.15(?) =
with postgres and my UFFI wrappers with a recent cvs-clisp.

In other words, the MOP support in CLISP, which clsql makes use of, =
works fine. I hadn't had a change to test this since last year, when =
Bruno Haible and Kevin Rosenberg made changes to CLISP's MOP and clsql =
to this effect.

I got 3 warning from loading sql/metaclasses.fas, about redefining =
metaclasses having no effect(?).

This also means that my latest UFFI-wrapper macros, modified for =
suitability with cl-sdl, did not break the clsql postgres code (but =
maybe clsql/db-postgres does not use uffi:get-slot-value, so the =
changes did not affect it).

Only 4 of 233 testcases failed: the 4 involving use of the LOOP =
extension (e.g. add-loop-path of MIT-LOOP), which is not available in =
CLISP's LOOP.

My UFFI wrapper macros of CLISP are in Sourceforge's Patches section =
about clisp:
http://sourceforge.net/tracker/index.php?func=3Ddetail&aid=3D1028683&gro=
up_id=3D1355&atid=3D301355

[They are not yet in UFFI itself, because there are still several =
conceptual mismatches between how UFFI uses C type declarations and =
what CLISP's FFI expects about types, not to mention lack of time to =
merge my code into cl-uffi]

I forgot that I could also have tested the postgres-socket backend, but =
then IIRC, that one does not use UFFI (except maybe for 64bit int =
conversions).

Regards,
	J=F6rg H=F6hle


-------------------------------------------------------
This SF.Net email is sponsored by Yahoo.
Introducing Yahoo! Search Developer Network - Create apps using Yahoo!
Search APIs Find out how you can build Yahoo! directly into your own
Applications - visit http://developer.yahoo.net/?fr=offad-ysdn-ostg-q22005